Methodological Individualism

viiiContents5 Society as subjectively meaningful interaction 126Symbolic interactionism 127Phenomenological sociology 137Existentialism 144Ethnomethodology 150Social constructionism 1606 Positivism in philosophy and social science 167Positivist philosophy 167British empiricism 168Logical positivism 173Positivist social science 179Systematic empiricism 179Positivist social theory 189Homans’s theory of exchange 190German Verhaltenstheorie 1967 Popperian methodological individualism 200Karl Popper 200J.W.N. Watkins 211Institutional individualism 218Joseph Agassi 218Ian C. Jarvie 221John O. Wisdom 224Conclusion 2268 Economics: the individualist science 228Macroeconomics and microfoundations 230General equilibrium theory 241Game theory 250Conclusion 2549 The new institutional economics 255Social organisations 260The household 261The firm 265The state 272Interest groups 275Conclusion 275Social rules 276Property rights 276Law 278Constitutions 281Evolutionary economics 281Conclusion 286

Contents10 Rational choice individualism 288Rational choice sociology 290James Coleman 292Raymond Boudon 306Analytical Marxism 309Jon Elster 310John Roemer 316Structural individualism 31811 Why methodological individualism? 320Philosophical background 321Reductionism 324Scientific reduction 324Psychologism 331Microfoundations 336Normative individualism 336Political individualism 337Humanism 339Conclusion 34512 Methodological individualism restated 346Strong and weak methodological individualism 346Versions of strong methodological individualism 349Explanatory methodological individualism 354ixNotes 357Bibliography 388Index of author names 437Index 441

Figures1.1 A graphic representation of methodological individualism 52.1 The theory of the social contract 92.2 Adam Smith’s theory of the invisible hand 112.3 Emile Durkheim’s rules of sociological method 342.4 Social wholes and collectives 403.1 John Stuart Mill’s psychologistic methodological individualism 494.1 Carl Menger’s theory of social institutions 914.2 Actions and institutions according to Robert Nozick 1254.3 Individuals in society according to Weber, Mises and Hayek 1255.1 Society according to symbolic interactionism 1365.2 Society according to phenomenological sociology 1445.3 Society according to ethnomethodology 1605.4 Berger and Luckmann’s dialectic of individual and society 1615.5 A mapping of some sociological theories in a two-dimensional space 1656.1 Mario Bunge’s methodological systemism 1866.2 Roy Bhaskar’s transformational model of the society/personconnection 1876.3 Hedström and Swedberg’s typology of social mechanisms 1887.1 Popperian institutional individualism 2278.1 The elements of the economic system according to Keynes 2328.2 The methodological individualism of general equilibrium theory 2459.1 Social institutions as endogenous variables 2869.2 Social institutions as endogenous and exogenous variables 28710.1 Coleman’s micro–macro scheme 29910.2 Methodological individualism according to Coleman 30010.3 Coleman’s micro–macro scheme extended in time 30110.4 Social structure as a determinant of individual action 30310.5 Social structure as positions to be filled 30510.6 Structural individualism 31812.1 The individualistic research programme, or versions of strongmethodological individualism 35312.2 Deductive–nomological explanations 355

PrefaceThis book is partly based on my doctoral thesis, Methodological Individualism – ACritical Appraisal (1987). More exactly, chapters 1, 4–7 and 11–12 are revisedversions of parts of my dissertation, while chapters 2–3, and 8–10 are entirelynew. I estimate the new material to make up about half of this book.Of my previous work, I have used only the first part, which consists of apresentation and clarification of methodological individualism, but not thesecond, critical, part. The reason for this is not that I am less critical of the originaldoctrine of methodological individualism today, than I was in 1987, but thatI wanted to expand my presentation and explication of this doctrine. One thing,however, makes a critical appraisal of methodological individualism morecomplicated today than it was in 1987. It is now more obvious than it was beforethat methodological individualism exists in a number of weak versions, in additionto the strong version that still dominates in economics. These weak versionsof methodological individualism are, in my opinion, far less objectionable thanthe original strong version, which I criticised in 1987. I don’t believe, however,that they are unobjectionable, but I will save my objections to a later work.The present book is part of a larger research project, which is going toinclude three more volumes. The first of these is a history of the individualistictheory of ‘man’ and society from Greek Antiquity to the rise of social scienceand of methodological individualism. My work on this volume is almost finished,and I hope that it will be published in about a year or so. A second volume willdeal with the obverse of methodological individualism, usually called methodologicalcollectivism, or holism. Also this volume is well on its way and I plan tocomplete it within the next couple of years. In a third volume, finally, I willreturn to the critique of methodological individualism, which I began in mydissertation. My intention is to try to carve out a position of my own, which isbased on the viable elements of both individualism and holism, but avoids theirobjectionable features.

AcknowledgementsI have been working on this book for quite some time. My first research on thetopic of methodological individualism goes back to the end of the 1970s. Overthe years I have been helped in various ways by lots of people and I am afraidthat today I do not even remember all who made some contribution to the endresult which is this book.My first and greatest debt is to Thomas Coniavitis, who suggested that Ishould write my doctoral thesis on methodological individualism and who influencedmy thinking about it over many years of teaching and discussion. Withouthim, this book would most probably not have been written at all, and had it beenwritten, it had been different from what it is. My intellectual exchange withThomas has shaped my own thinking about social theory and methodology inimportant ways.My thesis was written at the Department of Sociology, Uppsala University,and my supervisors were Björn Eriksson and Ulf Himmelstrand. I am grateful toboth of them, not only for the help they gave as part of their duty, but also forhelp I received after they had finished their assigned tasks. Peter Johnson readthe whole manuscript and helped me improve upon many arguments andformulations. I am particularly grateful for his efforts to improve my English. Aspecial thanks goes also to Tommy Törnqvist for intellectual stimulation and, notleast, for unfailing moral support. Lola Billås wrote the whole thesis on a PC,with scrupulous exactitude, before I had myself switched to this new technology.Other people, then at the department of sociology in Uppsala, who read parts ofmy thesis and made valuable suggestions include Göran Ahrne, Tom Burns,Magdalena Czaplicka, Peter Ekegren, Mats Franzén, Kaj Håkansson, Bo Lewin,Peter Sohlberg, Richard Swedberg and Börje Svensson.A special thanks goes to Wodek Rabinowicz and Lars Bergström, then at theDepartment of Philosophy in Uppsala, who generously offered their philosophicalexpertise to a dilettante in this field. While helping me to avoid somephilosophical blunders, they have no responsibility for those that remain. I am,finally, indebted to Margareta Bertilsson and Sven Eliaeson, who also read andcommented on the manuscript to my dissertation.My doctoral thesis was presented and defended at a disputation in Uppsala in1987. The faculty opponent was Steven Lukes, who is one of the main authoritieson the subject of individualism, including methodological individualism. I

1 IntroductionThere has been in the history of social thought a constant battle over the truenature of society and about the best way to understand and explain it. A majordivide goes between those who see society as an aggregate, collection, orcomplex of individuals and those who see society as some kind of ordered wholeand/or unitary collective. The former try to explain social phenomena in termsof individuals and their interaction, while the latter maintain that this is notpossible without essential reference to the social wholes of which they are partand/or the collectives to which they belong. 1 The opposition between theseconceptions of society was inherited by the social sciences and divided them intwo conflicting camps. With the emergence of the social sciences, however, themetaphysical issue was increasingly turned into a methodological issue. As weshall see, this does not mean that the metaphysical issue disappeared, only that itreceded into the background.There have been many names used to designate the two camps and theirrespective doctrines. In the twentieth century two (or three) names have beenselected as the most common. The battle has been increasingly waged in termsof methodological individualism, and its transmutations, versus methodological collectivismand/or holism. My interest, in this book, is in the former, but in order tounderstand one side, it is necessary to take a look also at the other side. Aboveall, it is necessary that there is a genuine divide separating the two doctrines, orelse this book would be very much ado about nothing.There are those who see, in this issue, the most fundamental and most importantproblem of the social sciences: that of the relation between individual andsociety. According to others, however, individualists and holists are engaged in asham battle. 2 I believe the first view is more correct. Anyone the least acquaintedwith the social sciences, knows that it matters which view you adopt in thismatter. Methodological collectivists and holists do tend to ask different questionsand provide different answers than do methodological individualists. There areimportant differences also within the two camps, but this is another matter. I alsofind it hard and a little bit odd to believe that the best minds in the history ofsocial thought should really have engaged, and with so much energy, in somethingwhich turns out to be a sham battle. Didn’t they notice?When I first started working on the topic of methodological individualism, I

2 Introductionwas often told that the debate about it was over. Several sociologists tried topersuade me that arguments against methodological individualism advanced bySteven Lukes and others are so strong as to render methodological individualisma mere curiosity, or at least harmless. The main point of Lukes was that it is vainto discuss methodological individualism without making clear what conceptionof ‘individuals’ you are using (Lukes, 1968; see also Burman, 1979). His argumentwas addressed in part to the methodological individualist, J.W.N Watkins,who had argued that methodological individualism is, or follows, from the ontological‘truism’ that ‘[a]ll social phenomena are, directly or indirectly, humancreations’ (Watkins, 1952a: 28).Today, we know that the debate about methodological individualists was notover. With the recent upsurge of rational choice, a new wave of methodologicalindividualism has swept the social sciences. One of the most influential advocatesof this approach, Jon Elster (1986b: 66; 1989b: 13), 3 has recently repeatedWatkins’s claim that methodological individualism is trivially true, but since it isimpossible for a methodology to be at all true, I suppose that Elster really meansmetaphysical, or ontological, individualism. 4 J.W.N. Watkins was more correcton this point, at least, since he recognised that his truism is an ontological thesisrather than a methodological rule. But, as he later himself admitted (1952b:186f), he was nevertheless wrong to assume that the latter follows from theformer. Even if ontological individualism is trivially true, it does not follow thatmethodological individualism is the only, or even the best, way to explain allsocial phenomena (Lukes, 1968; Kinkaid, 1997: 4, 16f). As Ernest Gellner (1956:176) put it some time ago, with respect to history: ‘History is about chaps. It doesnot follow that its explanations are always in terms of chaps’. Today, it is fairlycommon to accept ontological individualism, but deny methodological individualism.5I am not going to argue against ontological individualism, here, but I denythat it is trivially true. It is only if stated in a trivial enough way, that ontologicalindividualism is true, but this says little, or nothing, about the real issues involvedin the debate about it (cf. Miller, 1978; 1987: 115). If, for instance, methodologicalindividualism is only intended to deny that society is literally an organismendowed with a mind, or consciousness, which exists apart from the minds ofindividuals, then, it is of course true, but as Miller points out, even the archholist,Hegel, maintained that the world spirit is manifested in the actions ofindividual human beings and nowhere else. 6 Also, to suggest that the ‘truth’ ofmethodological individualism is secured by the fact that social wholes are madeup of individuals and their relations to one another is to beg the fundamentallyimportant questions: ‘What is an individual?’ and ‘What is a social relation?’ 7The issue between methodological individualists and their critics, then, isgenuine, but this does not mean that the debate has always been about the realissues involved. Far from it. There has been too much confusion surrounding themeaning and implications of the two positions, for a really fruitful debate to takeplace. This sad fact, has, no doubt, contributed to create the impression of asham battle. To an astonishing degree, the disputants have argued at cross-

Introduction 3purposes, and without a manifest intention to understand the opposite point ofview. There has been a marked tendency among representatives of both sides tomisinterpret and misrepresent the ideas and arguments put forward by the otherside.The reason for this is, probably, that the issue of methodological individualismversus collectivism and holism is felt to be important, not only for purelyscientific, but for extra-scientific reasons as well (cf. Kinkaid, 1997: 2ff). First ofall, it seems to be inextricably mixed up with some of people’s most entrenched,and most strongly-held beliefs about human nature and society. Second, thesebeliefs seem to be closely linked to their moral and political convictions. Third,there is clearly a connection with fundamental beliefs about science and itsgrowth. Finally, there is a more crass reason for social scientists to have strongopinions about methodological individualism. It seems to have territorial implications.Many social scientists, especially sociologists, have, no doubt, rejectedmethodological individualism, because they believed that it implies psychologism,or the reduction of sociology to psychology. If so, methodologicalindividualism would rob sociologists of their discipline. For all these reasons, thedebate between methodological individualists and their critics has been moreconfused than usual in social science and philosophy.It may be maintained – and sometimes I have been inclined to think so myself– that the debate between methodological individualists and methodologicalholists concerns one of those eternal issues which will never be settled by socialscience, philosophy, or by any form of argument. If so, it is, of course, merewaste of energy to write a book about it. Now, obviously, this is not what Ibelieve. While, I still think that the issue between individualists and holists mightnever be finally settled, once and for all, I do believe that it is possible to makesome progress.The aim of this study is to bring some clarity about the meaning of one sideof the divide: methodological individualism. 8 This is much needed, since, asDavid-Hillel Ruben (1985: 132) has maintained: ‘methodological individualismhas never been stated with enough clarity and precision to permit its proper evaluation’.Now, I do not cherish any illusions about what can be accomplished byway of remedy, but I do hope to be able to shed, at least, some light on thiscontroversial doctrine. I will try to do so by writing the history of methodologicalindividualism. As far as I can see, there is no other way to contribute to ourunderstanding of a doctrine than by looking at the various statements and usesof it in the history of ideas; in this particular case, in social science and philosophy.Since no doctrine can reasonably be interpreted as the sum total of all statementsabout it, at first I imposed two criteria of adequacy: Statements aboutmethodological individualism must be consistent, and the doctrine stated must besignificant. By the latter criterion, I mean that methodological individualism mustbe stated in such a way that it fits those social scientific theories and approacheswhich are generally considered individualistic.While working with this book, I have decided to drop both criteria. The

4 Introductionreason is that the idea of methodological individualism has developed in curiousways, and is hard to recognise these days. It would, of course, be possible to stickto the criterion of consistency and rule out what is not consistent with the originalversion of methodological individualism as being something else. When anincreasing number of people begin to believe that this else is methodologicalindividualism, however, this strategy becomes problematic. In this situation, Ihave decided to take a less essentialist approach and accept the development ofnew versions of methodological individualism, different from the original one,but, nevertheless, versions of methodological individualism. The criterion ofsignificance is, of course, dependent upon that of consistency, but it is stillpossible to conceive of some social scientific theories as being paradigm cases ofmethodological individualism and other theories as being more or less individualistic,compared to these paradigmatic cases.There are several possible ways of approaching the task I have set myself. Themost obvious way to proceed is, probably, to scan the literature for explicit statementsof the doctrine of methodological individualism. Another way is toconcentrate on the uses of methodological individualism in social science andhistory. There might be other ways, as well, but these are the ways I havefollowed. In order to clarify and, perhaps, justify the second route, I make adistinction between methodological individualism as a principle, or programme,of social research, and methodological-individualism-in-use; social science that is inline with this principle, or programme, without being explicitly based upon it. Itis, of course, possible to suggest, or accept, theories that comply with the stricturesof methodological individualism, without being committed to, or even beaware of this principle, or programme. If we have to do with theories ratherthan particular explanations, it is possible to speak of ‘theoretical individualism’,and if it is a general theory, of the ‘individualistic theory of society’. When I usethese expressions, in this book, I understand pieces of social science, whichqualify as methodological-individualism-in-use.This study, then, seeks the meaning of methodological individualism in theshort history of this doctrine. Methodological individualism goes back to thenineteenth century, and the various attempts to lay the foundation of the socialsciences. The individualist theory of society is much older, of course, but this is atopic I will treat in another book. 9How do I choose what to include in this history? Who are the main representativesof programmatic methodological individualism and which theories areparadigms of methodological-individualism-in-use? There is no certain way todecide these things and, in the end, you have to rely on previous knowledge andrules-of-thumb, as your main guides. A first rule I have tried to follow, is toconcentrate on those social scientists and philosophers, who themselves areadherents of methodological individualism, to the exclusion of those who merelycomment upon it, or criticise it. A second rule I have followed, is to pay mostattention to those methodological individualists who are singled out as mostimportant by the scientific and philosophical communities. ‘Most important’, inthis case, means most cited in the literature. Equally, when it comes to method-

Introduction 5ological-individualism-in-use. I have presented those theories and approacheswhich are commonly cited as individualistic and actually there is more agreementon this than on the exact meaning of methodological individualism.In order to illustrate the differences there are between various versions ofmethodological individualism, I am going to make use of some graphic representations(see Figure 1.1). More specifically, I will use a scheme I haveborrowed from the philosopher Arthur Danto (1965a). As we shall see, manyothers have used similar schemes to represent the relations between individualand society, action and structure, or micro and macro. As I use Danto’sscheme there are two levels, representing the individual and society, respectively.The arrows represent the direction of explanation according tomethodological individualism proper, or of causality in the case of ontologicalindividualism.Figure 1.1 A graphic representation of methodological individualismSource: Adapted with modification from Danto (1965a: 269)A word of warning should be voiced immediately against interpreting thissimple scheme too literally. Individualists and collectivists usually understandboth individual and society in different ways. To assume that we can talk aboutthe individual and society in a neutral way is, therefore, an illusion (Lukes, 1968:123ff; Archer, 1995: 34ff). If necessary qualifications are made, however, I thinkthe Danto’s two-level scheme is a useful tool for making the differences betweenvarious versions of methodological individualism visible.Two more clarifications need to be made before I get on to my real task. Thisstudy is all about the meaning of methodological individualism, and not at allabout its merits, if any. This does not mean that I am, somehow, neutral in thedebate between methodological individualism and its critics. I belong to thecritics, even if I am far less critical of the recent, weak versions of methodologicalindividualism than I am of the original, strong version. Even so, I haverefrained from passing explicit judgement on methodological individualism inthis work and I have tried, seriously, to give a fair account of its background,history and meaning. I hope that I have succeeded.Finally, I am a sociologist, interested in the other social sciences and in philosophy,but, nevertheless, a dilettante, outside my own discipline. I have tried to

6 Introductionunderstand the intricacies of economics and philosophy, in particular, and Ihope I have succeeded reasonably well, but I may, of course, have failed in somerespect. If so, I hope that my failure does not seriously damage the overallquality of this work.

2 BackgroundThe individualist theory of society has a long history in Western thought. Since Iwill try to tell this history in another volume, I will not go into details here. Mypurpose in this chapter is merely to provide a brief background to the rise of thespecifically methodological version of individualism. For this purpose, I will firstpresent the two main versions of the individualist theory of society whichpreceded methodological individualism; the theory of the social contract and thetheory of spontaneous order, as it took shape in Adam Smith’s idea of themarket as an invisible hand. This idea laid the foundation of economics and it isoften assumed that classical economics was not only the first social science, butalso the first example of methodological individualism. I am going to arguebelow that, as such, it was not an altogether clearcut example, although I agree,of course, that it was more individualistic than the views of most of its critics.Classical economics was mainly a British phenomenon, but in Germany, and toa lesser degree in France, there was a massive reaction against the individualismof the Enlightenment, of utilitarianism and of classical economics. The mainexpression of this reaction is the cultural movement known as Romanticism,whose most significant manifestation in the human sciences is German historicism.In France there emerged a doctrine which shared some elements with bothBritish and German thinking, namely the positivist sociology of Auguste Comte.Methodological individualism emerged, I believe, as an individualist counterreactionto the anti-individualist reaction of German historicism and thesociologism of positivist sociology. If this belief is correct, it becomes necessaryto include the latter two intellectual currents as important parts in any account ofthe background of methodological individualism.The social contractThe individualist theory of society goes back, as far as we know, to GreekAntiquity, where it was advanced, in particular, by the Sophists and by theEpicureans. The former invented the theory of the social contract and saw allsocial institutions as man-made conventions. The latter adopted the theory of thesocial contract and added to it an atomist metaphysics and a hedonist psychology.The individualist theory of society disappeared with Antiquity and was

8 Backgroundreplaced by a more holistic and collectivist view of society in the Middle Ages. Itreappeared in the Renaissance and culminated with the Enlightenment. Themost important figures are Thomas Hobbes (1588–1679) and John Locke(1632–1704), at least from an individualist point of view. Of these, I believeHobbes is the most important as a representative of a theoretical, and perhapsalso methodological, individualism, while Locke is more important as a representativeof political individualism.The point of departure of most theories of the social contract, and Hobbes’stheory is no exception, is the ‘state of nature’. In Hobbes’s version, the state ofnature is characterised by a war of each individual against all other individuals.In such condition, there is no place for industry; because the fruit thereof isuncertain; and consequently no Culture of the Earth; no navigation, nor useof the commodities that may be imported by Sea; no commodious Building;no Instruments of moving, and removing such things as require much force;no Knowledge of the face of the Earth; no Account of Time; no Arts; noLetters; no Society; and which is worst of all, continual feare, and danger ofviolent death; and the life of man, solitary, poore, nasty, brutish and short.(Hobbes, [1651] 1968: 186).The reason for this sad state of things is that there is no law and no commonpower to fear. But why is there need for a power to fear? Do people abide by lawonly for fear of consequences? According to Hobbes: Yes! Human nature is suchthat individuals seek only their own gain, and above all, glory, without regard forothers. Without a law to prevent them, they will invade one another in order totake what they want, or to prevent others from taking what they want.Happily, human beings are not only self-interested, they are also rational, andthis is their salvation. Hobbes’s state of nature is no place you would choose toinhabit, if there were an alternative. Now, since human beings are rational theyrealise that they would all be better off in a state of society where there is law andjustice and, therefore, peace. Hence, they enter a contract where they give up theirnatural right to everything and authorise an absolute sovereign to institute justice.This is a schematic version of Hobbes’s theory of the social contract. It is atheory, which is extremely individualistic, in the sense that it starts with natural, orpre-social individuals and explains the institution of society, or the state, solely interms of the human nature of these individuals (Peacock, 1986: 11–13; Pizzorno,1991). Since human beings, according to Hobbes, are rational egoists, it is possibleto see his theory of the social contract as a piece of rational choice analysis (see, e.g.,Hechter, 1989: 60). It has been suggested by Steven Lukes (1968: 119; 1973: 110)that Hobbes was also the first to articulate the principle of methodological individualist(see also Watkins, 1955b: 132f; [1965] 1973: 34), but this is more doubtful. Theclosest you get to an articulation of methodological individualism, in the writings ofHobbes, is his advocacy of the method of resolution and composition, which issimilar to the synthetic method suggested by Friedrich von Hayek (see p. 117).After Hobbes, the theory of the social contract was propagated by Benedict

Background 9Spinoza (1632–77), Samuel Pufendorf (1632–94), John Locke (1632–1704) andJean-Jacques Rousseau (1712–78), among others. Of these, Spinoza andRousseau were not individualists at all. Pufendorf was an individualist, but withsome reservations. Locke, however, was certainly an individualist, but he wasprimarily a political individualist and only secondarily a theoretical individualist.Like Hobbes, Locke took his point of departure in the state of nature, definednegatively as the absence of government. He did not share Hobbes’s view of thestate of nature as a state of war, however. When Locke turns to the state ofnature in the second volume of Two Treatises of Civil Government ([1690] 1960), hedepicts it as a state of freedom, equality and reason. According to Locke, reasonis the law of nature telling every man ‘not to harm another man in his Life,Health, Liberty and Possessions’ (p. 311). When Locke returns to the state ofnature later on in Two Treatises, however, he paints another picture, closer to thatof Hobbes. It is now a state of uncertainty, where the individual is ‘constantlyexposed to the Invasion of others’.This makes him willing to quit a Condition, which however free, is full offears and continual dangers; And ’tis not without reason, that he seeks out,and is willing to joyn in Society with others who are already united, or havea mind to unite for the mutual Preservation of their Lives, Liberties andEstates, which I call by a general name, Property.(Locke [1690] 1960: 395)Once in the condition of society, individuals see its advantages and consent toput themselves ‘under an obligation to every one of that society’, and ‘to submitto the determination of the majority’ (p. 376). The social contract, or compact, ofLocke, is not limited to the original contract and it need not be explicit. Peoplemay also give their tacit consent to a Government, which fulfils its chief end; ‘thepreservation of their property’ (p. 395).The theory of the social contract, in one form or another, is the main theoryof society before the eighteenth century. It is also the first paradigm of an individualisttheory of society, or of methodological-individualism-in-use. Using a,nowadays, common, if problematic, way of representing the relation of individualsto society, as two distinct levels of reality and/or analysis, I suggest thefollowing representation of the theory of the social contract.Figure 2.1 The theory of the social contract

10 BackgroundThe invisible handAfter Locke, there was a break with the theory of the social contract, at least as atheory of society. Social theory developed in a less rationalist and less individualistdirection. In France and Scotland, there emerged the insight that socialinstitutions are not conscious inventions, but have developed gradually and unintentionally.In the terminology, recently made popular by Friedrich von Hayek,most social institutions are ‘spontaneous orders’. Following the Scottish philosopherAdam Ferguson (1723–1816), he maintains, that they are the result ofhuman action, but not of human design (Hayek, 1948: 7). Among the mostimportant critics of the theory of the social contract was another Scottishphilosopher, David Hume (1711–76), who dismissed this idea as a mere fictioninvented for political purposes ([1741/2] 1963: 452ff).The most important application of the idea that social institutions are theunintended consequences of the intentional actions of individuals would turnout to be the market. In this case there was a special twist to the idea of spontaneousorder, because of its somewhat miraculous power of turning self-interestedindividual action into a collective good. The first to make this idea well knownwas the Dutch born physician Bernard Mandeville (c. 1670–1733), who movedto England and wrote his infamous Fable of the Bees (1714/29/32), with the moretelling subtitle, Private Vices, Public Benefits. The argument of Mandeville is thatmany selfish and even some vicious actions, such as crimes, are turned intopublic benefits.The most famous use of the idea of spontaneous order, transforming privateself-interest into collective benefit, is of course Adam Smith’s theory of themarket working as if governed by an invisible hand. (1723–90). In The Wealth ofNations (1776), he suggested that an individual, while intending only his owngain, is in many cases ‘led by an invisible hand to promote an end [the publicinterest] which was not part of his intention’ ([1776] 1937: 423).It is common among more recent methodological individualists to trace thisdoctrine back to the idea of spontaneous order, as suggested by BernardMandeville, David Hume, Adam Ferguson and Adam Smith. The most wellknownexample is Friedrich von Hayek, but he is not alone. 1 It is my impressionthat this interpretation is most common among social scientists and philosopherswho share a commitment to classical liberalism, which suggests a relationbetween methodological and political individualism. Be that as it may, those whotrace the origin of methodological individualism to Mandeville and the Scottishphilosophers have been anxious to point out that this version of individualism is,not only less rationalistic, but also less radical than that of the theory of thesocial contract. It is pointed out that the Scottish version of individualism takesits point of departure in the human individual as a social being, with a languageand other social institutions. I agree with this interpretation, but suggest that thisversion of methodological individualism is different, not only from that implicitin the theory of the social contract, but also from all explicitly stated versions ofthis principle before Joseph Agassi suggested the principle of institutional individualism(cf. Song, 1995).

Background 11Society SocialinstitutionsPricesIndividualMarketexchangeFigure 2.2 Adam Smith’s theory of the invisible handClassical economicsThe individual and isolated hunter and fisherman, with whom Smith andRicardo begin, belongs among the unimaginative conceits of the eighteenth-centuryRobinsonades, which in no way express merely a reactionagainst over-sophistication and a return to a misunderstood natural life …Smith and Ricardo still stand with both feet on the shoulders of the eighteenth-centuryprophets, in whose imaginations this eighteenth-centuryindividual … appears as an ideal whose existence they project into the past.Not as a historic result but as history’s point of departure. As the NaturalIndividual appropriate to their notion of human nature, not arising historically,but posited by nature. This illusion has been common to each newepoch to this day.(Marx, 1857–8: 83)This quotation is from the first page of Marx’s Grundrisse. A similar charge wasmade in the famous section on commodity fetishism in Capital, vol. 1 ([1867]1976), where he observes that ‘political economists are fond of Robinson Crusoestories’ (p. 169). 2 Marx was not the only one and not the first to accuse politicaleconomists of excessive individualism. Similar accusations were part of the standardcritique directed at classical economics from the quarters of German andEnglish Romanticism. 3 In his Past and Present (1843), Thomas Carlyle(1795–1881) had called economics ‘The Dismal Science’ preaching the ‘Gospelof Mammonism’ and reaching strange conclusions:We call it a Society; and go about professing openly the totalest separation,isolation. Our life is not mutual helpfulness; but rather, cloaked under duelaws-of-war, named ‘fair competition’, and so forth, it is mutual hostility. Wehave profoundly forgotten everywhere that Cash-payment is not the sole relationbetween human beings; we think, nothing doubting, that it absolves andliquidates all engagements of Man. ‘My starving workers?’ answers the richmill-owner: ‘Did not I hire them fairly in the market? Did I not pay them, to

12 Backgroundthe last sixpence, the sum covenanted for? What have I to do with themmore?’ – Verily Mammon worship is a melancholy creed.(Carlyle [1843] 1965: 148f)Carlyle was a conservative, but he was cited with approval by Marx andEngels in the Communist Manifesto ([1848] 1967: 82), and it is not hard to see why.There is a close affinity between the conservative and the radical, socialistcritique of captalism in the nineteenth century. This is even more obvious in thewritings of John Ruskin (1819–1900), who was also a conservative, but exerted astrong influence on the guild socialists. Ruskin took a more scholarly approach toeconomics than did Carlyle, and made a detailed, if not very pertinent, critiqueof the theories of Ricardo and John Stuart Mill. The ideal economy, for Ruskin,was that of the household, which is a moral economy, governed by moral principlesof the organisation of labour and of the distribution of its fruits. Politicaleconomy, therefore, ought to be a moral science, rather than a catallactics, ortheory of exchange (Ruskin, 1857–62). 4In the critique of political economy by Carlyle and Ruskin are mixed theoretical,political and moral considerations. Economics is both theoretically andmorally individualistic and by moral individualism, is understood self-interest.This is typical of the nineteenth century. Usually, no distinction was madebetween theoretical and methodological considerations on the one hand, andpolitical, economic and moral considerations on the other hand (cf. Sowell,1974: ch. 1 and Vanberg, 1975: ch. 1). As we shall see, the term ‘methodologicalindividualism’ was introduced to make the distinction between methodologicaland political individualism. 5The view that classical economics was individualistic is not a story made upby its critics, however. As we have already seen, it is common among defenders ofmethodological individualism to trace the origin of this doctrine to the ScottishEnlightenment and this tradition included Adam Smith, the acknowledgedfounder of classical economics and, therefore, the father of economic science.But were Adam Smith and the other classical economists methodologicalindividualists? There is no clear answer to this question, for several reasons. Firstof all, the classical economists, before John Stuart Mill, did not engage verymuch in methodological discussion and not at all systematically (Sowell, 1974:112). By the time of John Stuart Mill, however, classical economics had becomeindividualistic. Smith, however, was not a radical individualist, like Mill, and itwas only later that his model of the market as a spontaneous order was turnedinto a paradigm of strong methodological individualism. In the theory ofgeneral equilibrium, first suggested by Walras and more recently developed byArrow and Debreu (among others), we find an economic theory without anysocial institutions.The individualism of classical economics was more political and economicthan theoretical and methodological. The classical economists favoured a freemarket with limited government, but their view of society was not radically individualisticand they did not advocate an individualistic methodology. Their

Background 13approach to society is more aptly described by the term ‘institutionalism’. It istrue, though, that Adam Smith’s analysis of the market as if steered by an invisiblehand, when stripped of the institutional element, has turned into aparadigm of an individualist theory of society, or of methodological individualism-in-use.Thomas Malthus’s (1766–1834) An Essay on the Principle of Population ([1798]1970) seems to be rather different from Smith’s Wealth of Nations. The pessimisticargument of the book is based on two postulates: ‘First, That food is necessary tothe existence of man. Second, That the passion between the sexes is necessaryand will remain nearly in its present state’ (p. 70). From ‘these fixed laws of ournature’, together with some additional assumptions, Malthus arrives at thefollowing conclusion: ‘Population, when unchecked, increases in a geometricalratio. Subsistence increases only in an arithmetic ratio’ (p. 71). This is anextremely individualistic argument, without a trace of social institutions, orhistorical relativity. But appearances are deceptive. The word ‘unchecked’ isimportant, indeed. The laws about the increase of population and food, aresubject to a ceteris paribus clause (pp. 112f). As a matter of fact, Malthus’ discussionis unusually rich with observations on social institutions, which counteractthe tendency to shortage of food.It may also be pointed out that, despite appearances to the contrary, Malthuswas not an advocate of an abstract-deductive approach to political economy. Hewas very careful to support his laws by empirical evidence and in his SummaryView of the Principle of Population (1830) he maintains that they are inductive generalisations,derived from experience. But this was after the intellectual exchangewith his friend David Ricardo.On the other side of the Channel, Jean-Baptiste Say (1776–1832) alsodefended a more inductive and more ‘holistic’ approach to economics. Withoutdenying the utility of the abstract-deductive method, which he himselffrequently used, he maintained that all principles of economics must be based,not on hypotheses, but on observation of facts and derived from these facts byinduction ([1803] 1971: xxvf, lii). Like Malthus, Say criticised Ricardo forstarting from abstract hypotheses and then reasoning in a straight line, withoutever comparing his conclusion with observed facts. ‘From that instant nothing inthe author’s work is represented as it really occurs in nature’ (p. xlvii). In order tobe of practical utility, however, the science of political economy must be basedon observed facts. Say’s economist par préference was Adam Smith and, like thelatter, he was some kind of methodological institutionalist. He seems to havegone further in a holistic direction, however. Not only did he advocate theapproach of social economy, but he obviously conceived of society as anorganism of interrelated parts (Forget, 1999: 53ff). He did not, however, wish toinclude politics in the science of political economy. Economics is a science of theproduction, distribution and consumption of wealth, but not of the government,or state, conceived of as a household, writ large. Jean-Baptiste Say clearly didnot contribute to making classical economics more methodologically individualistic.He was, however, an advocate of laissez-faire and Say’s law of the market,

14 Backgroundwhich says that markets are always cleared (Sowell, 1974: ch. 2), 6 is an article offaith for most libertarian economists.It is fair to say, I believe, that economics became more individualistic withDavid Ricardo (1772–1823). At least, it became more abstract, and this led to amore individualistic approach, simply because it was social institutions and relationswhich disappeared in the process of abstraction. In the case of Ricardo,there is some justification for Marx’s complaint in the quotation above. Ricardo frequentlyplaces his abstract reasoning in ‘the early stages of society’ (Ricardo,[1817] 1973: passim), as if it were a fully fledged market society. It is this lack ofhistorical insight which Marx objects to with some justice. 7 It should be pointedout, though, that Ricardo is still looking at the distribution of wealth, as a matterof its division among the three classes in society: ‘the proprietor of the land, theowner of the stock or capital necessary for its cultivation, and the labourers bywhose industry it is cultivated’ (p. 3). This makes it possible to talk about ‘theholism of the Classicals’ (Greaves, 1996: 2). Economics would definitely becomemore individualistic with the rise of the neoclassical school of economics.In 1820 Malthus published his Principles of Political Economy, partly as a reply toRicardo, who had criticised his theory of rent. This is not my business, butbesides this disagreement there is a more principal one concerning approach andmethod. Malthus sides clearly with Smith against Ricardo ([1820] 1989: 22f),and this means that he is sceptical about the excessive simplification and generalisation,which follows with using mathematics, or geometry, as a model foreconomic theory (pp. 1ff). He is also critical of the unwillingness of someeconomists to test their theories, but, nevertheless, to use them as guides to politicalpractice. Malthus repeats these complaints in Definitions in Political Economy([1827] 1971 1ff), where he also criticises Ricardo for breeding confusion byusing familiar economic terms in unfamiliar ways (pp. 28–36).This critique of Ricardo’s way with words was echoed by Nassau Senior(1790–1864) in An Outline of the Science of Political Economy ([1836] 1965: 5), wherehe also repeated some other complaints, made by Malthus, against Ricardo.Senior shared Ricardo’s view of economics as an abstract-deductive science,resting on a small number of indubitable general propositions, from which inferencesare drawn, which, if our logic is impeccable, are as true as the premises.According to Senior, inferences about the nature and production of wealth areuniversally true, whereas those about its distribution are affected by ‘disturbingcauses’, in the form of ‘the peculiar institutions of particular countries’ (p. 3).But no matter how general and how true the economist’s conclusions, theydo not authorize him in adding a single syllable of advice. That privilegebelongs to the writer or the statesman who has considered all the causeswhich may promote or impede the general welfare of those he addresses,not to the theorist who has considered only one, though among the mostimportant, of those causes. The business of a Political Economist is neitherto recommend nor to dissuade, but to state those general principles, which it

Background 15is fatal to neglect, but neither advisable, nor perhaps practicable, to use asthe sole, or even the principal, guides in the actual conduct of affairs.(Senior [1836] 1965: 3)The voices of Malthus and Senior were two in a choir. The critique againstRicardo’s use of the abstract-deductive approach was massive, even from thosewho used such an approach themselves. There were mainly two charges: (1)Ricardo failed to see that abstraction is also simplification and (2) he made illegitimateuse of his abstract theory as a basis for political recommendations, latercalled the ‘Ricardian Vice’ (Schumpeter, 1954: 472).It is commonly assumed that Ricardo owed his method to his close friendJames Mill (1773–1836), while Mill owed his knowledge about economic mattersto Ricardo (Hutchison, 1978: ch. 2). Against this view, Samuel Hollander (1985:1) has argued that ‘[a] sharp distinction should be made between the methodologicalorientations of James Mill and David Ricardo’. While Ricardo neverpublished anything substantial about the method of economics, Hollander maintainsthat he had a fairly sophisticated view of the nature and limitations ofeconomic analysis in his unpublished writings (pp. 15–36). That, in fact, he didfully recognise that abstraction is simplification. According to Hollander, then,Mill is the one to blame, not Ricardo. In a forceful reply to Hollander, Hutchison(1994: 91) describes this view as ‘odd’, since it is to deny ‘plain fact’, ‘for whichthere is abundant evidence’. Allen Oakley (1994) grants that Ricardo may havebeen a more sophisticated methodologist than commonly believed (p. 132), but itremains the case that the public impression given by Ricardo to his contemporariesand immediate successors was that he espoused a severely abstractdeductivemethodology as the one most appropriate to political economy’ (p.134).There is no doubt whatsoever that James Mill was a champion of an abstractdeductivemethodology, based on the model of mathematics, and used it withouthesitation, as a basis of political recommendation. In the Preface to The Elementsof Political Economy (1820b), it is written: ‘My object has been to compose aschool-book of Political Economy, to detach the essential principles of thescience from all extraneous topics, to state the propositions clearly and in theirlogical order, and to subjoin its demonstrations to each’ (p. 204). He goes on todraw an analogy between demonstrations in political economy and in mathematics,which says something about his view of economic theory. In theintroduction he makes another analogy: ‘Political economy is to the State, whatdomestic economy is to the family’ (p. 210). This view of economics was laterdeemed seriously wrong by the Austrian methodological economists. Friedrichvon Hayek called it ‘false individualism’ (pp. 119f ).If political economy is a matter of state, there must be a sharing of subjectmatter between economics and political science. But this is not all. According toMill, there is also a sharing of approach. In his famous ‘Essay on Government’(1820a), Mill used an economic approach to politics, much like the recent theoryof public choice. 8 Mill’s main source of inspiration was Jeremy Bentham, who

16 Backgroundprovided his view of human nature: ‘We may allow, for example, in generalterms, that the lot of every human being is determined by his pains and pleasures;that his happiness corresponds with the degree in which his pleasures aregreat, and his pains are small’ (pp. 55f). This is Bentham’s first assumption. Hissecond assumption is that it is the business of government to maximise happinessin society.A more remote source of inspiration was Hobbes, who provided Mill withtwo more assumptions: (1) that the principal means used to maximise happinessare wealth and power, and (2) that individuals, unless prevented, will use theirpower to appropriate the wealth of others. The task of government, then,becomes that of preventing individuals from interfering with the property ofothers. 9The question Mill tried to answer in his Essay was this: What is the best formof government, on the assumption that it is made up of individuals, who pursuetheir own selfish ends? The answer he reached differed from that of Hobbes.Instead of an absolute sovereign, Mill recommended a representative system.But then the original question reappears.There can be no doubt, that, if power is granted to a body of men, calledRepresentatives, they, like any other men, will use their power, not for theadvantage of the community, but for their own advantage, if they can. Theonly question is, therefore, how they can be prevented? In other words, howare the interests of the Representatives to be identified with those of thecommunity?(Mill [1820a] 1978: 75)The answer Mill came up with was a form of representative democracy, which,however, excluded women on the ground that their interests are included in thatof their fathers or their husbands, as the case may be (p. 79).Mill’s essay was bound to create controversy. The conservatives disliked it,because they rejected any form of democracy and those more radical than Millhimself – eventually including his own son John Stuart Mill – were disappointedwith his exclusion of women. The most effective critique came from the historianThomas B. Macaulay (1800–59), in a famous article in Edinburgh Review(1829). Macaulay’s critique was directed, not at this or that conclusion, but at theapproach as such – an approach which, according to Macaulay, makes it possibleto reach any conclusion whatsoever: ‘Our objection to the Essay of Mr Mill isfundamental. We believe that it is utterly impossible to deduce the science ofgovernment from the principles of human nature’ ([1829] 1978: 124).What proposition is there respecting human nature which is absolutely anduniversally true? We know of only one: and that is not only true but identical[tautological]; that men always act from self-interest. This truism theUtilitarians proclaim with as much pride as if it were new and as much zealas if it were important. But in fact, when explained, it means only that men,

Background 17if they can, will do as they choose. … If the doctrine that men always actfrom self-interest, be laid down in any other sense than this – if the meaningof the word self-interest be narrowed so as to exclude any one of themotives which may by possibility act on any human being, – the propositionceases to be identical; but at the same time it ceases to be true.(Macaulay [1829] 1978: 124f)Macaulay’s argument made a deep impression upon the younger Mill, whoalso was disappointed with the reply of his father (Mill [1873] 1961: 96f). By thetime of Macaulay’s attack, John Stuart was already in a state of mental crisis,due, at least partly, to increasing doubts about his inherited worldview. Thetheory, method and politics of Benthamite utilitarianism no longer seemed toprovide all the answers. Macaulay’s article was the last nail in the coffin of hisfather’s absolute intellectual authority. What escaped the notice of bothMacaulay and John Stuart Mill, however, is the fact that James Mill did notderive his conclusions from assumptions of human nature alone. Equally importantfor his argument are the institutional arrangement; monarchy, aristocracy,democracy, different voting rules, etc., within which the play of individualinterest takes place. The approach of James Mill was that of institutional individualism.It is arguable that John Stuart’s psychologism was more individualisticthan the geometric method of James Mill.Eventually John Stuart would reject the geometric method altogether, andespecially in politics, where he came to believe that the assumption of economicman simply does not work (see p. 47). Before he reached this conclusion,however, he wrote his classic article ‘On the Definition of Political Economy andthe Method of Investigation Proper to It’ (1836). In this article, the younger Millstill believed that the abstract-deductive method of economics, and the othersocial sciences, is strictly analogous to the geometric method of Euclides. Unlikethe elder Mill, however, he did realise, with perfect clarity, that abstraction is alsosimplification and, as a corollary, that abstract-deductive theories are inadequateguides to political action, and the more abstract, the less adequate.In the article ‘On the Definition of Political Economy’ Mill made a cleardistinction between economics as a science and as an art: ‘Science is a collectionof truths; art, a body of rules or directions for conduct’ (1836: 410). Politicaleconomy is a science, not to be confused with the art of domestic economy.More specifically, it is a moral – or psychological – science (p. 414). As such, itdepends on the laws of mind, or human nature, but not upon these alone.According to Mill, we must distinguish between the individual (1) in isolationfrom other individuals, (2) in interaction with other individuals and, (3) ‘as livingin a state of society, that is forming part of a body or aggregation of humanbeings, systematically co-operating for common purposes’ (p. 417). Mill calls thescience of (3) ‘social economy’, or the ‘science of politics’. Political economy isnot social economy, however, ‘but a branch of that science’.

18 BackgroundIt does not treat of the whole of man’s nature as modified by the social state,nor of the whole conduct of man in society. It is concerned with him solelyas a being who desires to possess wealth, and who is capable of judging ofthe comparative efficacy of economic means for obtaining that end. Itpredicts only such of the phenomena of the social state as take place inconsequence of the pursuit of wealth. It makes entire abstraction of everyhuman passion or motive, except those which may be regarded as perpetuallyantagonizing principles to the desire of wealth, namely, aversion tolabour and desire of the present enjoyment of costly indulgences. These ittakes, to a certain extent, into its calculations, because these do not merely,like other desires, occasionally conflict with the pursuit of wealth, butaccompany it always as drag or impediment, and are therefore inseparablymixed up in the consideration of it.(Mill [1836] 1950: 420)This is the first clear statement of the character and methodological role ofeconomic man in the history of economic thought. Mill’s next step is to maintainthat this is the procedure of political economy, not because ‘any politicaleconomist was ever so absurd as to suppose that mankind are really thus constituted,but because this is the mode in which science must necessarily proceed’ (p.421). Even so, the assumption of wealth maximisation, is a fairly good approximationto the truth about men and women in their economic endeavours.Mill goes on to deny that political economy is based on induction. It is, on thecontrary, essentially ‘an abstract science’ and its method is ‘the method a priori’(p. 424). Just like geometry starts from an arbitrary definition of a line, politicaleconomy starts from an arbitrary definition of man. They are both equallyremoved from reality and their conclusions are only true ‘in the abstract’ (p.425). Mill defends the use of the geometric method in social science, by pointingout that experiments are not feasible. I still find it a bit surprising that he saw thegeometric method as the only alternative to experiment, especially, since he goeson to acknowledge the existence of disturbing causes (pp. 429ff) and to arguethat because of them, the operation of laws in society give rise to tendenciesonly. In geometry, there are neither causes, nor tendencies.Except for the misleading analogy with geometry, Mill’s characterisation ofpolitical economy stands out as the first really adequate attempt to grasp thenature of economics as a science. In his most important work on the method ofscience, A System of Logic (1843), Mill abandoned the view that political economyuses a geometric method and in his Principles of Political Economy (1848), hedeparted from a strict use of the abstract-deductive method. Instead ofneglecting disturbing causes, he frequently introduced them in his analysis. Oneexample is his discussion of the division of the produce, which is said to be ‘theresult of two determining agencies: Competition and custom’ ([1848] 1976:242).The final statement of the method of classical economics is J.E. Cairnes’s TheCharacter and Logical Method of Political Economy (1875), which is a work largely

Background 19based on the position of Mill. Cairnes defends political economy againstenemies, such as Comte, and he defends the Ricardian abstract-deductiveversion of economics against the alternative conceptions of Say and Senior. Heis very emphatic about the importance of the ceteris paribus clause in the case ofall economic laws ([1875] 1965: 61ff) and he seems to believe that the laws ofeconomics depend upon those of physics and psychology (pp. 52, 72).Sometimes he also includes institutions among the facts that are given toeconomic analysis.In the considerations just adduced, we may perceive what the proper limitsare of economic inquiry – at what point the economist, in tracing thephenomena of wealth to their causes and laws, may properly stop toconsider his task as completed, his task as solved. It is precisely at that pointat which in the course of his reasonings he finds himself in contact withsome phenomenon not economic, with some physical or mental fact, somepolitical or social institution. So soon as he has traced the phenomena ofwealth to causes of this order, he has reached the proper goal of hisresearches; and such causes, therefore, are properly regarded as ‘ultimate’ inrelation to economic science. Not that they may not deserve and admit offurther analysis and explanation, but that this analysis and explanation is notthe business of the economist – is not the specific problem which he undertakesto solve.(Cairnes [1875] 1965: 53f)This quotation is interesting mainly because it is a fairly clear – maybe thefirst – statement of what economists today would call exogenous variables, orgivens. From my point of view, however, it presents a problem: like John StuartMill, Cairnes clearly senses the importance of social institutions for the functioningof the market, but, like the former, he is not equally clear about their rolein economic analysis and explanation. As I have already said, he usuallymentions only the physical and mental facts and laws on which, not onlyeconomics, but ‘historical, political and, in general, social investigation’ depend.Political economy belongs ‘neither to the department of physical nor to that ofmental inquiry’, but occupies ‘an intermediate position’. The phenomena itinvestigates depend on ‘physical, physiological, and mental causes’ and areexplained by ‘the concurrence of physical, physiological, and mental laws’ (p.52). The conclusion I draw from these statements is that Cairnes believed thatsocial institutions can be explained by the physical, physiological and mentallaws. If this interpretation is correct, Cairnes – like Mill – was a psychologistic,rather than an institutional individualist (see pp. 218–27).German historicismIt has been pointed out, that the French and the British saw history as theadvance of civilisation, whereas the Germans, with characteristic disdain for

20 Backgroundscience and capitalism, saw themselves as the carriers of culture (Elias [1939]1968: 3–34). This particular manifestation of German nationalism, wouldsurvive well into the twentieth century and reached a peak during the FirstWorld War (Kusch, 1995: 212ff). As a nation characterised by its culture, theGermans attached great value to the humanities and especially to history.Because of this, the dominating approach to man and society in Germany washistoricism.There is a certain confusion surrounding the term ‘historicism’ (Iggers, 1995).According to the well-known view of Karl Popper, it denotes ‘an approach to thesocial sciences which assumes that historical prediction is the principal aim, andwhich assumes that this aim is attainable by discovering the “rhythms” or the“patterns”, the “laws” or the “trends” that underlie the evolution of society’(Popper, 1957: 3). But this is not what most people understand by ‘historicism’.One problem with Popper’s view, is that it includes many philosophers and historiansfrom Antiquity to the Enlightenment, who also saw history as a law-boundprocess, but who did not manifest the most typical marks of ‘historicism’, such ashistorical relativism. The solution to this puzzle is that Popper made a distinctionbetween ‘historicism’ and ‘historism’ (p. 17). What most people today understandby ‘historicism’ is what Popper, in accordance with an earlier convention, called‘historism’. In this second sense of ‘historicism’, is denoted a doctrine whichemphasises the historicity and individuality of human beings and societies. ‘Theessence of historism is the substitution of a process of individualising observationfor a generalising view of human forces in history’ (Meinecke, [1959] 1972: lv). 10All social phenomena, such as languages, myths, laws, customs, social institutions,even man herself, are products of history. As such they differ between times andplaces. Since human beings also change with history, historicism rejects the idea,defended most strongly in Great Britain, of a constant human nature.In this chapter, I want to add a third meaning to the term ‘historicism’.‘Historicism’, as I use this term, also means a reduction, or subordination, ofother disciplines to history. This form of ‘historicism’ was a distinctive feature ofthe human sciences in nineteenth-century Germany. Philosophy, economics,jurisprudence, linguistics, were dominated by historical schools and, more or less,transformed into historical sciences. Theoretical economics had no place at all inthis intellectual milieu. Sociology was also resisted, but less so than economics.As I have already mentioned, German historicism did not share the Frenchadoration of science and technology, but, above all, it rejected the positivistclaim of unity of method between the human and the natural sciences. But whatabout psychology? Before I turn to this question, I will give a brief sketch of thedevelopment of German historicism, because it was this doctrine, or some of itselements, that methodological individualism was created to defeat.It should be pointed out at the beginning, however, that methodological individualism,as an explicitly stated principle of social science, emerged in a milieudominated by German historicism, and adopted many of its ideas. In a rich andfascinating study of Carl Menger and the Origins of Austrian Economics (1990), MaxAlter argues that the largely unknown roots of Austrian Economics are in

Background 21German historicism and Romanticism. I find his argument, on the whole,convincing and recommend Part I of his book as a somewhat richer presentationof German historicism than the one I give here.The roots of historicism are to be found in Romanticism, and in the beginningthere was an almost complete overlap between the two movements.Romanticism, of course, was broader, since it included literature and the arts. Itis a commonplace to see Romanticism as a reaction against the Enlightenment.If the latter was characterised by rationalism, mechanism, individualism andutilitarianism, the former was a manifestation of irrationalism, vitalism, collectivismand culturalism. Against rationalism it emphasised emotion and will,against mechanism it posited life and organicism, against individualism it cherishedthe reality and value of nation, state and community and againstutilitarianism it posited the Greek ideals of autonomy, virtue and self-realisation.A common source of Romanticism and historicism is the philosophy ofhistory of Johann Gottfried Herder (1744–1803). In a prize-winning ‘Essay onthe Origins of Language’ (1772), he criticised the Enlightenment view thatlanguage is the result of an agreement to use words in certain ways ([1772]1966: 100f). Language is the result of organic growth (p. 113). ‘Least of all is itagreement. An arbitrary convention of society’ (p. 119). In his first importantcontribution to the philosophy of history, Auch eine Philosophie der Geschichte zurBildung der Menscheit (1774), he poured scorn and irony on the philosophy ofhistory of the French Enlightenment and its belief in linear development,progress and ultimate superiority of European civilisation ([1774] 1967 102).Against this view, Herder maintains that each people, like each individual is aunique individuality, which cannot be compared to other peoples (pp. 41, 61). Allpeoples, however, are part of humanity and contribute to its culture. Despite hishumanism, Herder is most famous as the father of nationalism. In his mainwork, Reflections on the Philosophy of the History of Mankind (1784–91), he maintainsthat ‘every nation is one people, having its own national form, as well as its ownlanguage: the climate, it is true stamps on each its mark … but not sufficient todestroy the original national character’ ([1784–91] 1968: 7). ‘Time, place, andnational character alone, in short the general cooperation of active powers intheir most determinate individuality, govern all the events that happen amongmankind’ (p. 159).Herder is often mentioned together with the Italian philosopher GiambattistaVico (1668–1744), because of the similarity of their views of history and society.Vico was the first to grasp clearly the differences between nature and historyand, therefore, also between the methods used in the two sciences. Among otherthings, Vico maintained that we understand what is created by human beings, ina way different from the way we understand nature (see Berlin, 1976: xvi–xix).The reason for this is that human beings express themselves in their creations.There is no constant human nature, however, but humanity changes with thecourse of history. This means that we must understand each human creation asthe expression, not only of a particular human being, but above all, of a humanbeing belonging to a particular culture at a particular time and in a particular

22 Backgroundplace. The fact that social institutions, like law, language, government, etc, arehuman creations, does not mean that they are constructions, made by design.Vico was a critic of the theory of the social contract. Like Mandeville, thoughindependent of him, he argued that social institutions and nations are the unintendedconsequences of intentional human activity. Following Aristotle, hemaintained that they grow, like organisms; that they are natural growths.All these ideas were shared by Herder, especially the idea of ‘expressionism’,to borrow a useful term from Isaiah Berlin. ‘Expressionism [is] the doctrine thathuman activity in general, and art in particular, express the entire personality ofthe individual or the group, and are intelligible only to the degree to which theydo so’ (Berlin, 1976: 153). What led Herder to adopt and develop this idea, washis interest in the origin and development of languages and their various uses,especially in poetry (pp. 169–72). It is in language, above all, that we find thecharacter, spirit, or soul, of a people (Volksseele), or what we, nowadays, call itsculture (pp. 194–9). 11The idea of natural growth was also to become of tremendous importance inGerman historicism. It was shared by the conservative founder of the HistoricalSchool of Law, Friedrich, Karl von Savigny (1779–1861), who insisted that lawis, or should be, the result of historical development, not of rational constructionand by his student Karl Marx, who constantly maintained that society is anatural growth (naturwüchsig). In The Vocation of Our Age for Legislation andJurisprudence (1814), written in the wake of French occupation, Savigny rejects allattempts to introduce the Code Napoleon in German legislation. He is alsoconcerned about the foreign element of Roman law. German law must be basedon German legal tradition. The reason is that, like in the case of language, thereis an ‘organic connection of law with the being and character of the people …Law grows with the growth, and strengthens with the strength of the people,and, finally dies away as the nation loses its nationality’ ([1814] 1975: 27). Themain task of German jurisprudence is the study of German customary law,which must be the basis of all German law. ‘We shall then possess a trulynational law; and a powerful expressive language will not be wanting to it. Wemay then give up the Roman law to history, and we shall have, not merely afeeble imitation of the Roman system, but a truly national and new system ofour own’ (p. 154f). It may be pointed out that this, originally holistic, view ofsocial institutions, was adopted by the methodological individualists Carl Mengerand Friedrich von Hayek, who both suggested individualistic models of thegrowth, or evolution of, social institutions.An important role in the development of German historicism must also beassigned to German idealist philosophy and, in particular, to the philosophy ofHegel (1770–1831). In his Philosophy of Right (1821), Hegel maintained that ‘Thestate is the actuality of the ethical idea’ ([1821] 1967: 155) and of ‘concretefreedom’ (p. 160). As such, it ‘has supreme right against the individual, whosesupreme duty is to be a member of a state’ (p. 156). The theoretical justificationfor this subordination of the individual to the state is that the state is anorganism, such that each member is a part of this organic whole and depend

Background 23upon the other parts for his/her existence and thriving. This is the ‘organic’, or,as it is sometimes called, the ‘metaphysical’ theory of the state, which wascommon to Romanticism and idealist philosophy, and which was eventuallyadopted also by most historians.Hegel’s theory of the state was part of his philosophy of history, which owes alot to the ideas of Herder. Before the state, there is the people and a people is a‘spiritual individual’, which expresses itself in ‘its religion, its cult, its customs, itsconstitution and political laws, the whole scope of its institutions, its events anddeeds. This is its work: This one people. People’s are what their deeds are’ (Hegel,[1837] 1953: 89f). Each people has a history of its own and passes throughdifferent stages, according to a principle of differentiation, which turns eachspirit into a particular national spirit. Each national spirit, in its turn, is only amoment in the development of the ‘world spirit’ towards self-consciousness,absolute knowledge and freedom.The idealist theory of the state as an ethical idea was adopted by the mostfamous of all German historians and historicists, Leopold von Ranke(1795–1886). Most people probably think of him as an ‘empiricist’ historian whointroduced the critical method to the discipline in the form of new and moresevere standards of documentary evidence and who saw the task of the historianas finding out how it really was (wie es eigentlich gewesen). But Ranke was also adefender of the idealist and metaphysical theory of the state as ‘a spiritualreality’ (Ranke, [1836] 1973: 110), ‘a living thing, an individual, a unique self ’ (p.112). Because of the role of historians and of historicism in the creation andconsolidation of the German nation-state, it is possible to see the former as itsservants and the latter as its ideology (Iggers [1968] 1983: 17–28). Ranke was thehead of its conservative branch.Due to the importance attached to the state in nineteenth-century Germany,economics took the form of political economy, or Nationalökonomie. ‘For theGerman economists the symbiosis of state and economy was self-evident’ (Tribe,1988: 5). Because of the strength of historicism in the German universities,economics assumed the form of history. In Germany, the Historical Schooldominated economics well into the twentieth century, or until the Nazis seizedpower. The founder of the Historical School of Economics, was WilhelmRoscher (1817–94). Other members of the Older Historical School, were BrunoHildebrand and Karl Knies. The Older Historical School adopted some of thetypical ideas of historicism, in general. According to Roscher, for instance, thenational economy is an organism, and the task of political economy(Staatswirtschaft) is to find the law of development of this organism (1843: 4).Towards the end of the nineteenth century, there emerged a Younger HistoricalSchool in Economics, with Gustav Schmoller as its head. Other members wereLujo Brentano and G.F. Knapp. The Younger Historical School had thrown offmost of the romanticist baggage, but retained the view of economics as politicaleconomy. It was characterised, above all, by detailed empirical investigation ofeconomic institutions. The German Historical School of Economics was to

24 Backgroundbecome the immediate object of attack for the Austrian methodological individualists.12To sum up my presentation of historicism, so far, I suggest that it comprisesthe following holistic doctrines:IOrganicism: The state and/or society is conceived of as an organism. This ideahas two elements: (1) Organisms are wholes, where each part depends for its existenceand functioning upon the other parts and upon its place in the whole. Thisdoctrine is the traditional opposite of the individualist theory of the state andsociety. (2) As other organisms, social institutions and societies grow. The implicationsof this idea are not clear, however. According to one idea of organicgrowth (2a), it implies that history is irrational and unpredictable, because thereis novelty in the development of life. According to a second idea of organicgrowth (2b), it implies Aristotelian teleology. Like natural organisms, societiesoriginate as germs, or embryos, and grow until they reach their fully developedform. For some organicists, development is arrested at the stage of perfection, forothers, societies also decay and die. Still others, understand by organic growth(2c) only spontaneous development. This is the thesis that social phenomena arethe unintended consequences of human activity.There is also the idea of philosophy of history that history has a goal and/ormeaning. This idea is not necessarily associated with organicism, but is moretheological in origin. While organicism, tends to be materialistic, philosophy ofhistory tends to be idealistic. It conceives of history as the development of mind,spirit, or reason. This view of history, brings me to the second important idea ofearly historicism:IIObjective idealism: Like the natural organism of an individual human being has amind, the social organism has a mind of its own. The ideas and values of apeople (Volk) are expressed in their language, customs, laws, constitution, myths,literature, etc. All these social institutions are but various expressions of the soul,or spirit, of the people. In the well-known case of Hegel, objective idealism takesthe form of a teleological philosophy of history, but this is an exception. Mostversions of objective idealism, are content to point out the historical andgeographical variability of folk spirits, but refrain from forcing them into apattern of development, or succession.The main legacy of Romanticism, then, was holistic, but there was also, especiallyin the early phase, a fairly strong tendency of individualism. Germanindividualism, however, was different from French and British individualism andthe difference is caught very well by Georg Simmel ([1917] 1950: 81), who callsthe latter quantitative and the former qualitative. As the term ‘qualitative’ suggests,

Background 25German individualism highlights the individuality and uniqueness of humanbeings, rather than their equality. Because it tends to give pride of place to greatindividuals, or personalities, it has also been called ‘aristocratic’ individualism(Köhler, 1922: 48). Aristocratic individualism is both methodological and ethical:as a methodology, it assigns an important role to great individuals in historicalexplanation; as an ethic, it assigns more value to great individuals than to lesserones. In Germany, qualitative individualism was represented, in particular, byGoethe, Schiller and Wilhelm von Humboldt, and, later, by Friedrich Nietzsche,but it also influenced some British philosophers and historians, such as Coleridgeand Carlyle, and, to some extent, also John Stuart Mill (see p. 44). I find it convenientto divide qualitative individualism in three forms: political, historical andexpressive.IPolitical individualism: While liberalism in Great Britain and France was characterisedby quantitative individualism, the German version of liberalism wasdistinguished by qualitative individualism. In The Limits of State Action (1791–2)Humboldt saw political individualism, in the form of a minimal state, as aprecondition for the free development of individuals, as did many other leadingfigures of German Romanticism. According to George G. Iggers ([1968] 1983:7ff, 40ff), it was the experience of French dominion and the wars of liberation(1792–1815), which turned most German intellectuals, including Humboldt, intonationalists and defenders of a strong state. The historian Johann GustavDroysen is a case in point, but also Max Weber would be an example.IIHistorical individualism: Historicism inherited, from Romanticism, the admirationof great individuals and personalities, and a corresponding contempt for, or atleast lack of interest in, the destiny of the common people, or ‘masses’, as theywere commonly called. This historical individualism originated in Germany, butit is best epitomised by the Briton Thomas Carlyle, who wrote in his On Heroes,Hero-Worship and the Heroic in History: ([1841] 1975: 239, 235, 266) that ‘the historyof the world is but the biography of great men’. Historical individualism wasopposed by many historians – especially Marxists – and there ensued an intensedebate about the role of the individual in history.IIIExpressive individualism: I will end my short presentation of historicism, with a noteabout method. Historicism was, from the very beginning, closely associated withthe study of language and this is probably the reason for the importanceattached to the hermeneutical method, or art of interpretation. Vico hadsuggested that we understand human creations better than nature. Also Herder,

26 Backgroundhad mentioned Einfühlung as a way to understand human expressions. The mostimportant founder of hermeneutics, however, is the great theologian FriedrichSchleiermacher (1768–1834). He developed hermeneutics into an astonishinglysophisticated art of interpretation, to which I cannot here, of course, do justice.An important point for my purposes is that he made a distinction betweensubjective, psychological interpretation and objective grammatical interpretationof language. ‘Understanding has a dual direction, towards the language andtowards the thought’ and ‘[n]either language nor the individual as productivespeaking individual can exist except via the being-in-each-other of both relationships’(Schleiermacher [1809–10] 1998: 229). Both types of understanding arecontextual: ‘every utterance has a dual relationship, to the totality of language and to thewhole thought of its originator’ (Schleiermacher [1809–10] 1998: 8). The consequenceof this is the famous hermeneutic circle: ‘The vocabulary and the history ofthe era of an author relate as the whole from which his writings must be understood as part, andthe whole must, in turn, be understood from the part’ (p. 24). There is a hermeneuticcircle also on the psychological side (pp. 90ff), but both moments rest on knowledge,not on intuition, or empathy, even though he admits that some people havea ‘talent for knowledge about individual people’ (p. 11).Wilhelm von Humboldt was also an important pioneer in the study oflanguage, but his idea of hermeneutics is more romantic than was that ofSchleiermacher. Humboldt pays lip service to Ranke’s view of the historians’task, which is to ‘present what actually happened’ ([1821] 1967: 57), but actually,he is against this view. According to Humboldt, literal truth is strictly impossible.The reason for this is that the essence of truth is hidden from observation. Whenall the facts are collected, there remains the intuitive understanding of their‘inner causal nexus, on which, after all, their inner truth is solely dependent’ (p.58). Humboldt’s account of the historians’ task is suggestive, but impressionistic.It is not clear, for instance, what ‘inner truth’ refers to: if it is the subjective truthof the individuals acting in history, if it is some kind of objective truth inherentin their creations; a work of art, a system of ideas, or a historical event, or if it isthe inner truth of the historian. One thing is clear, however: according toHumboldt, the task of the historian involves an important element of artisticcreativity.Among professional historians, it was Johann Gustav Droysen (1808–84) who,more than anyone else, tried to develop the method of verstehen in history.Droysen was engaged in a two-front war against the critical method of Ranke,on the one hand, and the scientific history of Buckle, on the other hand. Aliberal like Humboldt, Droysen was opposed to the conservative Ranke for politicalreasons. In the eyes of Droysen, Ranke’s critical method would reducehistory to a sterile concern with mere facts, thereby robbing it of its role in theeducation (Bildung) and emancipation of mankind (MacLean, 1982). A similarconcern, made him reject also the scientific history of Comte and Buckle. Theproblem with scientific history is that it is deterministic and aims at prediction. 13Once again like Humboldt, Droysen saw history as an art, not as a science. 14The method of verstehen was launched in opposition to both critical and scientific

Background 27history, as a way of introducing moral agents, or subjects, in history (cf. Burger,1978).Droysen made a distinction between nature and history in terms of the categoriesof space and time. The method of the natural sciences is explanation(erklären), that of history, understanding (verstehen) (Droysen [1858] 1977: 22,150f). History deals with the traces left by human beings. Every such trace left isthe expression of an individual “I”, but, nevertheless, an individual possible tounderstand. ‘With human beings … we have an essential kinship and reciprocityof nature: every “I” enclosed in itself, yet each in its utterances disclosing itself toevery other’ (pp. 121f). The impression created by this translation is thatDroysen’s method of verstehen is radically individualistic. This is not the case,however. For Droysen, psychological interpretation is only the third, of four,forms of interpretation. The other three, are (1) pragmatic interpretation, (2)interpretation of conditions and (4) interpretation of moral forces and ideas([1858] 1977: 149ff). Of these, the first and third are clearly individualistic, whilethe second and fourth are not. Pragmatic interpretation is concerned with instrumentalaction; what the historical remains reveal about the ends of historicalactors, while psychological interpretation digs deeper into their personalities.The conditions of action include, among other things, social institutions, whilemoral forces and ideas include cultural entities like religions, ideologies andworldviews. There is nothing to indicate that Droysen saw social institutions andcultural phenomena as psychic, or, in any other way, individualistic, as did somelater advocates of the method of verstehen. On the contrary, Droysen is safelyanchored in the tradition of historicism, collectivism and holism (Iggers [1968]1983: 110ff; MacLean, 1982: 355).Positivist sociologyIn France, the dominating approach to the scientific study of ‘man’ and societywas positivist sociology. The first positivist sociologist, of some importance, wasHenri St Simon, who is the source of many ideas, including that of a comingindustrial society. 15 More important, however, was his secretary Auguste Comte,who introduced, not only the terms ‘positivism’ and ‘sociology’, but also that of‘altruism’. By ‘positivism’, Comte understood a science based upon observationof facts, as distinguished from religion based upon faith and metaphysics basedon speculation. According to his famous law of the three stages, religion is thefirst, metaphysics the second, and positive science the third stage in the developmentof the human mind. ‘Sociology’ is the highest, but also the least developedand least exact, in the hierarchy of the sciences: mathematics, astronomy,physics, chemistry, biology and sociology. The reason sociology is the least developedand the least exact is that it deals with the most complex phenomena.Except for this, however, there is no gulf between natural science and socialscience. Their methods are essentially alike. Comte’s positivism, then, is amethodological monism.It may be noted that, neither economics, nor psychology is part of Comte’s

28 Backgroundladder. The reason economics is missing is that Comte saw it as part of sociology;the all-encompassing science of society as a whole. The mistake ofeconomics is to analyse the economy in isolation from the rest of society andfrom its history (Comte [1836–42] 1974: 446–50). It may be noted that Comte’scritique of economics is not directed at Adam Smith, who is mentioned withgreat admiration. Since I am partly engaged in a terminological survey, it mayalso be noted that Comte uses the term ‘spontaneous order’, made famous byFriedrich von Hayek and in a similar manner, to designate those socialphenomena, including the market, which are not consciously designed. 16 Theterm ‘altruism’ was also introduced with an eye on economics and the assumptionof rational egoism, or self-interest. If the altruistic instincts did not exist inus,the mutual services that this activity of industry calls out would certainly notbe able to create them. Were this dismal hypothesis a true one, the reciprocalassistance of man to man would never become gratuitous, and theonly moral influence that this aid would have, would be to develop aconstant prudence of interest.(Comte [1851–4] 1975: 407)Psychology fares even worse than economics, since it has no place at all inComte’s system. A first reason for this is that the mind is unobservable and thereforeinaccessible to positive investigation. A second reason is that the self ismultiple rather than unitary. The idea of an I, or a self is a fiction created byreligion and metaphysics. According to Comte ([1836–42] 1974: 3, 380ff), thereis nothing in between the biological and social organism. Psychology is anideology and, as such, superfluous. The affective and intellectual faculties it isused to explain is better explained by (phrenological) physiology. Psychology,then, is reducible to physiology. But only partly. Human beings are social beingsand the main bulk of human action is part of society. What cannot be explainedby physiology, therefore, must be explained by sociology. In addition to being aphysiological reductionist, then, Comte was a sociological reductionist; thefounder, not just of sociology, but of sociologism.Like many other social theorists in the nineteenth century, Comte conceivedof society as an organism and like most of them, he saw the social organism as awhole which must be investigated as such: ‘viewing each element in the light ofthe whole system’ (p. 462). Sociology is divided in two branches: statics anddynamics, the former concerned with the mutual interdependency of the partsof the social organism and the latter with its development or growth. Socialstatics, or the theory of spontaneous order, seeks laws of coexistence, orharmony, while social dynamics seeks the laws of succession of socialphenomena.The raison d’être of sociology lies in the historicity of social phenomena. At thebeginning of history, social phenomena may be explained in terms of humannature alone. As soon as human beings have been modified by social develop-

Background 29ment, however, this is no longer possible. Biological reductionism, therefore, iswrong. ‘The consequence of this error is that social modifications proper tocertain periods, and passing away with them, are too often supposed to beinherent in human nature, and therefore indestructible’ (Comte [1836–42] 1974:488). This argument has lost none of its topicality. It seems to be the mainmessage also of the various versions of contemporary social constructionism.The reason, adduced by Comte, for the variability of human nature is the moredated. Human beings have become social beings and must be understood as partof the social organism. Comte does not reject all use of the laws of humannature in sociology. Since human nature does set limits to the variability of socialphenomena, no law arrived at by way of historical induction can be allowed tocontradict the ‘known laws of human nature’. This argument, in particular,appealed to John Stuart Mill, who made it the basis of his inverse-deductive, orhistorical, method.Comte’s organicism and his argument about the historicity of socialphenomena is similar to that of German historicism (see pp. 20–22), but he isless interested in the spiritual, or cultural, side of society. His law of the threestages is about the development of knowledge, but there is no place for conceptslike ‘folk soul’ and ‘national spirit’ in his sociology. The main difference, betweenthem is that Comte was a positivist, preaching unity of method, whereas theGerman historicists saw the human sciences as entirely different from the naturalsciences. Because of this, Comte was actually a main target of attack fromhistoricist quarters.For the purposes of this book, however, the similarities are more importantthan the differences. Like the German historicists, Auguste Comte was amethodological holist, and as such both were common targets for the attacks ofmethodological individualists. It may be added that he was not an economic, orpolitical individualist either and this contributed to his unpopularity amongmethodological individualists. 17 According to Comte, the market may be spontaneous,but it is not self-regulating. If left to itself, it leads to chaos rather thanorder. This was one reason why Comte supported a big government, engaged inlarge-scale social engineering; in technocracy, if not in bureaucracy. 18A final note on positivism and the science of statistics. 19 Comte, himself, sawno use for statistical, or other mathematical, methods in positivist sociology. Hebelieved that social phenomena are too complex for mathematical treatment,and criticised Condorcet for failing to see the limits of mathematical analysis,including the probability calculus, in social science ([1819–26] 1974: 166–74;[1836–42] 1974: 492–4). But there were others, such as the Belgian statisticianLambert A.J. Quetelet (1796–1894) and the English historian Henry ThomasBuckle (1821–62), who saw the possibility of using statistics for the purposes of apositivist sociology and history.Quetelet was a pioneer in statistics and especially in its use to register moralphenomena. Following St Simon and Comte, he called his moral statistics, socialphysics. Quetelet is famous for his idea of the Average Man, which was muchcriticised, because it was believed that he failed to see that the Average Man is a

30 Backgroundfiction. Actually, Quetelet was aware that the Average Man is a ‘fictitious being’([1835] 1842: 8), but he obviously believed that averages tell us somethingimportant about ‘man in general’ and about society. Quetelet was also amongthe first to observe the astonishing constancy of many statistical phenomena,such as crime and suicide rates. Above all, he was among the first to collect alarge body of statistical data to demonstrate such regularities. Quetelet was notsatisfied, merely to register statistical regularities, he wanted to explain them, andhis explanation was sociological:Society includes within itself the germs of all the crimes committed, and atthe same time the necessary facilities for their development. It is the socialstate, in some measure, which prepares these crimes, and the criminal ismerely the instrument to execute them. Every social state supposes, then, acertain number and a certain order of crimes, these being merely the necessaryconsequences of its organisation.(Quetelet [1835] 1842: 6)According to Quetelet, the crime rate depends upon social institutions, and thisis encouraging, since it is possible to change social institutions in a way thatreduces the number of crimes.Buckle’s views on statistics can be found in his History of Civilization in England(1858). In this work he appears as a follower of Comte and Quetelet, attemptingto demonstrate that history, like the natural sciences is subject to necessity and,therefore, law-governed. His main argument for this thesis is the regularitiesfound in moral statistics. These regularities, according to Buckle (pp. 20ff), forceus to conclude that the actions of men are determined by the state of societyand, therefore, subject to social laws. His main example is the apparently individualact of suicide, which exhibits such an astonishing regularity in theaggregate – a fact which led Buckle to reach the following dubious conclusion:In a given state of society, a certain number of persons must put an end totheir own life. This is the general law; and the special question as to whoshall commit the crime depends of course upon special laws; which,however, in their total action, must obey the large social law to which theyare subordinate. And the power of the larger law is so irresistible, thatneither the love of life nor the fear of another world can avail any thingtowards even checking its operation.(Buckle, 1858: 25f)Buckle was not alone to be fascinated by statistical regularities and the suiciderate, in particular, had a powerful grip on the minds of nineteenth-century socialscientists. The most famous study is, of course, Emile Durkheim’s Suicide (1897),‘the masterpiece of nineteenth century statistical sociology’ (Hacking, 1990:64). 20 Before I turn to Durkheim, however, I will stay in Britain for a while.Early positivist sociology was mainly a French affair, but it is common to

Background 31include also the British sociologist Herbert Spencer among its ranks. Spencer,himself, protested against being associated with Comte and published a list ofReasons for Dissenting From the Philosophy of M. Comte (Spencer [1864] 1984: 23). Hedid not deny, however, being a ‘positivist’ in a more general sense of that term.Spencer’s main quarrel with Comte was political. Being a radical individualist inmatters political and economic, Spencer did not at all fancy the political part ofComte’s positivist programme (pp. 18f, 23f). In sharp contrast to Comte, Spencerwas a supporter of a free market society watched by a minimal state. 21 In thisrespect, Spencer had much more in common with another critic of Comte,Friedrich von Hayek.Another cause of dissension was Comte’s dismissal of subjective psychologyas an important science in itself, and as an indispensable basis of sociology(Spencer [1864] 1984:10). 22 According to Spencer, ‘psychological truths underliesociological truths, and must therefore be sought by the sociologist’ ([1873] 1961:348).No one indeed, who is led to dwell on the matter, can fail to see how absurdis the supposition that there can be a rational interpretation of men’scombined actions, without a previous rational interpretation of thosethoughts and feelings by which their individual actions are prompted.Nothing comes out of society but what originates in the motive of an individual,or in the united similar motives of many individuals, or in theconflict of the united similar motives of some having certain interests, withthe diverse motives of others whose interests are different.(Spencer [1873] 1961: 348f; see also [1864] 1984: 15)This seems to be a fairly clear statement of psychologism cum methodologicalindividualism, but perhaps it is not clear enough. A problem with Spencer is thathe also emphatically endorsed the idea of society as an organism (Spencer,1860). 23 As I have already indicated, this idea is almost unanimously conceivedof as holistic and as the very opposite of methodological individualism. 24 In theliterature about Spencer, it is a recurrent allegation that there is a conflictbetween his individualism on the one hand and his organicism on the other. Anextensive survey of these allegations can be found in Gray (1985) and Taylor(1992), who both argue that it is possible to save Spencer from contradiction. Ina second and much extended work, Gray (1996) accepts that there are a numberof apparent tensions both between his methodological individualism and organicism(ch. 2) and between atomism/mechanism and organicism in his theory ofsociety (ch. 5). Even so, he argues (ch. 14) that most of these tensions are, indeed,more apparent than real, since they are due to conceptual ambiguity.A problem with this controversy is that the authors do not always distinguishclearly between ontological, methodological and political issues. For my part, Iagree that there is no necessary conflict between political individualism andorganicism. Even if most organicists are political collectivists, it is possible tocombine organicism with political and economic individualism, and Herbert

32 BackgroundSpencer is, indeed, the primary example of this unusual combination. 25 It isimportant to observe the difference between those who say that the state is anorganism and those who say that society is an organism. The individualists are tobe found among the latter and they usually have a leaning towards conservatism.There is an unintentional ‘wisdom’ in the organically grown traditions and in theworking of the spontaneous order of the market, which cannot be achieved, onlyruined, by political means. Of special importance for Spencer’s pro-market attitudewas his view that the social organism lacks a consciousness of its own. Onlyindividuals have consciousness (Spencer [1860] 1982: 397). The particularwisdom of the market lies in the spontaneous co-operation of individuals withspecialised knowledge. This view foreshadows Hayek’s famous argument aboutthe role of knowledge in a market economy (cf. Taylor, 1992: 143). 26Concerning the conflict between methodological individualism and organicism,I see more of a problem, even if I agree that it is always possible toformulate these doctrines in a way that makes the conflict disappear (see Gray,1985: 246–53; 1996: ch. 14). I am not convinced that Spencer can be saved inthis way, but my interest is not in the consistency of his position. I am interestedin the question of whether or not he was a methodological individualist, and onthat point I find myself in agreement with Ellen Frankel Paul (1988): ‘ThatSpencer was a methodological holist is, I think, beyond question’ (p. 443). Sheadds that ‘elements of methodological individualism can be discerned in some ofSpencer’s writings but it must be considered a minor element’ (p. 444).After Spencer, positivist sociology returned to France. The most importantfigure was Emile Durkheim, who praised Spencer for going further than Comtein spelling out the details of the organism analogy. Like Spencer, Durkheim sawsociety as a spontaneous order, but he credits the economists with being first tosuggest this idea.They [the economists] were the first to sense all that is living and spontaneousin societies. They understood that collective life could not brusquelybe established by some clever artifice; that it did not result from an externalmechanical impulsion but is slowly elaborated in the very heart of society. Inthis way they were able to ground a theory of liberty on a foundation moresolid than a metaphysical hypothesis. In effect, it is obvious that if collectivelife is spontaneous, it must not be deprived of its spontaneity. Any fetterwould be absurd.(Durkheim [1888] 1978: 48)Here agreement ends, however. According to Durkheim, Spencer made themistake of believing that self-interested individuals could engage in peacefulexchange, without a system of rights and duties to regulate their behaviour. I donot believe, this is an altogether fair critique of Spencer, but it is certainly fair tosay that Spencer was much more pro-market than Durkheim, who saw animportant role for state in economic life (see Udehn, 1996: 354–7).Durkheim’s agreement with the economists is also limited to the idea of spon-

Background 33taneous order. Economics is the individualist social science, par excellence, and,like most sociologists, Durkheim is the very opposite: a methodological holist,who sees methodological individualism as a serious defect. 27 I will quote apassage at some length, because I believe it gives a clear and concise picture ofeconomics that is fairly typical among sociologists.According to them [the economists], there is nothing real in society but theindividual. Everything emanates from him, and it is toward him that everythingconverges. A nation is only a nominal entity; it is a word which servesto designate a mechanical aggregate of juxtaposed individuals. But it has nospecific properties that distinguish it from other things. Its properties arethose of the elements that compose it, merely enlarged and amplified. Theindividual is, therefore, the only tangible reality that the observer can get at,and the only question which science can set for itself is to investigate howthe individual should conduct himself in the principal circumstances ofeconomic life, given his nature. Economics and, more broadly, social laws,are not, then, very general facts which the scholar induces from the observationof societies, but logical consequences deduced from the definition ofthe individual. The economist does not say, ‘Things happen in this waybecause experience has established that fact.’ Rather he says, ‘Things shouldhappen in this way because it would be absurd if it were otherwise.’ Theword natural should therefore be replaced by the word rational, which is notthe same thing. If only this concept of the individual, which is supposed tocontain in itself the entire science, were adequate in reality! But in theattempt to simplify things, the economists have artificially impoverished theconcept. Not only have they ignored the circumstances of time, place, andcountry in order to conceive of man’s abstract type in general, but in thisideal type itself they have neglected everything which does not bear uponstrictly individual life to such an extent that, passing from abstraction toabstraction, nothing is left but the sad portrait of an isolated egoist.(Durkheim [1888] 1978: 49)Durkheim’s description of economics is, I believe, representative of the view of amajority of sociologists and (economic) historians in the nineteenth century. Tosome extent Durkheim’s view may be based on that of Comte, but it owes muchmore to his reading of Adolf Wagner, Gustav Schmoller and other economistsbelonging to the German Historical School (see Durkheim [1887] 1993: 58–77;[1888] 1978: 59ff).But is it a fair description, or is it the prejudice of an ignorant sociologist?The answer, I believe, is that Durkheim paints a largely correct picture ofeconomics, at least at the end of the nineteenth century. It should be added,though, that what is a vice to the typical sociologist is a virtue to the typicaleconomist.Durkheim seems to have shared Comte’s sociologistic view that economics isbest conceived of as a part of ‘general sociology’ (Durkheim [1909] 1978: 83).

34 BackgroundHe also shared some of the latter’s dismissive attitude towards psychology. 28 Hedid not, however, deny it a place in the hierarchy of sciences. ‘In succession,physics and chemistry, then biology and finally psychology have been established’(Durkheim [1888] 1978: 47). He even admitted that sociology depends uponpsychology. He emphatically denied, however, that sociology is reducible to,psychology. It was the merit of Comte to have established the fact that society islike a living organism and, as such, ‘something more than a collection of individuals.A whole is not identical to the sum of its parts, even though without them itwould be nothing’ (p. 50). Or in other words: ‘for Comte society was a being suigeneris’ (p. 54).Durkheim adopted this view of society. For him, social facts are sui generis and,therefore, irreducible to facts about individuals. More specifically, social factshave two characteristics: (1) they are external to individuals and (2) have coercivepower over them (p. 51; see also Durkheim, [1895] 1982: Ch. 1 and 1898). Thisdoes not mean, however, that social facts exist outside all individuals taken collectively.Durkheim’s argument is that they are external to each single individual.Nevertheless, from these characteristics of social facts, Durkheim derives themethodological rule that social facts must be explained in terms of other socialfacts, not in terms of facts about human nature ([1895] 1982: 134).Using the scheme I borrowed from Arthur Danto (see Introduction, p. 5), it ispossible to represent Emile Durkheim’s methodology in the following way:Figure 2.3 Emile Durkheim’s rules of sociological methodVertical arrows represent Durkheim’s methodological rule that social facts areexternal constraints on individual behaviour, while horizontal arrows representthe rule that social facts must be explained in terms of other social facts. Hisfamous explanation of the differences in suicide rates between countries isdescribed, by himself, as horizontal, but James Coleman (1986a: 347f) hasargued that it is better represented by the vertical arrows. I tend to agree withColeman. Since Durkheim wants to explain an aggregated macro-phenomenon– namely differences in suicide rates, in terms of social facts such as solidarityand social control, it is clear that these social facts really act upon individuals(Durkheim [1897] 1951: 297ff), who decide to commit, or not commit suicide.In comparison with Comte and Spencer, Durkheim moved away from

Background 35organicism towards idealism. What makes social facts sui generis is not that societyis an organism, but that it has a collective consciousness (Durkheim [1900] 1973:13; [1911] 1974: 93). The source of this change is probably Durkheim’s visit toGermany in the 1880s, where he was exposed to Völkerpsychologie and to WilhelmWundt’s empirical approach to ethics (Durkheim, 1887). There is, in fact, strongreason to believe that Durkheim derived most of his view of social facts fromWundt (Gisbert, 1959; Hall, 1993: 37–51). One idea, in particular, has beenassumed to derive from Wundt; that of ‘creative synthesis’ (Ginsberg, 1956: 45).Just as life emerges out of the association of molecules in the cell and mind fromthe combination of physiological processes in the brain, social facts emerge as aresult of the interaction of individual minds. 29 Before Wundt, the idea ofcreative synthesis was used by John Stuart Mill, under the name of ‘mentalchemistry’, to argue that new ideas emerge in the mind of individuals, but it wasWundt who used it to explain the emergence of social facts (see p. 61).Another source of Durkheim’s holism was the view of the statisticianQuetelet, who led him to see in the stability of various statistical measures overtime a sign of the existence of social facts. His own study of Suicide (1897)strengthened him in this belief and led him to argue that the suicide rate ‘is notsimply the sum of independent units, a collective total, but is itself a fact suigeneris, with its own unity, individuality and consequently its own nature – anature, furthermore, dominantly social’ ([1897] 1951 46).Before leaving Durkheim, it is important to point out that when he maintainsthat sociology is irreducible to psychology, he intends individual psychology, ofwhich he has a rather narrow view. It seems to be restricted to a concern withhuman beings as unaffected by society. As soon as they turn into social beings,they become the subject matter of sociology. Durkheim has no objection at all tocalling sociology a form of ‘social’, or ‘collective psychology’ and therefore, in abroader sense, a proper part of psychology (see, e.g. Durkheim [1895] 1982:249f; [1897] 1951: 312; [1898] 1974: 34). Because of this, and some otherconsiderations, Lorenzo Infantino (1998: ch. 5) asks the question: ‘Is an“Individualistic” reading of Durkheim possible?’ Given the fact that Durkheim isprobably the most often cited representative of collectivism and holism (see, e.g.Pratt, 1978: 112ff; Rosenberg 1988: 118ff; Pettit [1993] 1996: 132ff), this questionmight seem absurd, but I do not think it is. At least not with respect toontological individualism. Having made this admission, I had better add that Ialso believe that an individualistic reading of Durkheim, even if not absurd, isultimately wrong, and it is not possible with respect to methodology. Durkheimwas certainly not a methodological individualist.Not all French sociologists joined the sociologistic camp. Durkheim’s fiercestcritic, Gabriel Tarde was among the founders of social psychology and adefender of psychologism in sociology. On this issue he sided with Mill againstComte (Tarde [1898] 1969: 77ff). He was also critical of Quetelet’s social physics(Tarde [1890] 1962: 114ff) and the organicism of Spencer. His main target,however, was Durkheim. According to Tarde, Durkheim’s argument that socialfacts are sui generis, implies their existence in some Platonic heaven over and

36 Backgroundabove individual human beings ([1894] 1969: 114ff). In a debate with Durkheimin 1903, he is reported to have said:Does Mr. Durkheim think that social reality is anything other than individualsand individual acts or facts? ‘If you believe that,’ said Mr Tarde, ‘Iunderstand your method, which is pure ontology. Between us is the debatebetween nominalism and scholastic realism. I am a nominalist. There canonly be individual actions and interactions. The rest is only a metaphysicalentity, mysticism’.(Tarde, [1904] 1969: 140)In contradistinction to Durkheim, Tarde maintains that ‘it is imitation which isthe elementary and universal social fact’ ([1898] 1969: 54). All social phenomenaare transmitted not collectively, but individually, ‘from one individual – parent,teacher, friend, neighbour, comrade – to another’. This is not to deny, however,that the collective result often is fairly constant. According to Tarde, it is this factwhich ‘gives rise to Mr Durkheim’s ontological illusion. For there is no doubt thatit is veritable scholastic ontology that the learned writer is undertaking to injectinto sociology in place of the psychology he opposes’ ([1894] 1969: 115). ‘Tosum up the question which I began by asking: What is society? I have answered:Society is imitation. We have still to ask: What is imitation? Here the sociologistmust yield to the psychologist’ ([1890] 1962: 76). Tarde, then, reaches the sameconclusion as did Mill, and it is possible to see in the laws of imitation anattempt to realise Mill’s unfinished project of a science of ethology (see pp. 45f ).Tarde, then, is highly critical of Durkheim’s idea of social facts sui generis and,it might be added, of his associated idea of a ‘collective consciousness’, which isdistinct from the individual consciousness of human beings. According to Tarde,there is nothing but individual consciousness. This does not imply, however, thatthere is nothing but individual psychology. Sociology is collective, or social,psychology, even if it depends on individual psychology.Collective psychology, inter-mental psychology, that is sociology, is thuspossible only because individual psychology, intra-mental psychology, includeselements which can be transmitted and communicated from one consciousnessto others, elements which, despite the irreducible hiatus betweenindividuals, are capable of uniting and joining together in order to formtrue social forces and quantities, currents of opinion or popular impulses,traditions or national customs.(Tarde [1898] 1969: 95)In the debate between Durkheim and Tarde, it eventually turned out that alsothe former saw sociology as a kind of social psychology. Does this mean thatthere is no important difference between their respective versions of sociology? Ibelieve there is, but I am not going to argue this point here. It may be added,though, that Tarde was critical of other aspects of positivist sociology, as well,

Background 37and, especially, of its organicism and historicism ([1898] 1969: 77ff). While itwould be absurd to call Emile Durkheim a ‘methodological individualist’, itseems fair enough to attach this epithet to Gabriel Tarde.I have suggested that sociology emerged, at least in part, as a holistic reactionto the individualistic discipline of economics. If so, what was the position ofVilfredo Pareto, who was both a neoclassical economist and a classical sociologist?It is often assumed (see e.g. Boland, 1982: 27ff), as more or less self-evident,that Pareto was a methodological individualist – maybe because it is simplytaken for granted that a neoclassical economist has to be a methodological individualist– but I wish to cast some doubt upon this seemingly self-evident truth.Some things suggest that Pareto was, indeed, a methodological individualist:(1) his denial of the possibility of interpersonal comparisons of utility is typical ofmethodological individualists, at least in the Austrian tradition. (2) His famousPareto optimality has become part and parcel of political individualism. (3) Hiscritique of conceptual realism and leaning toward nominalism (Pareto, [1916]1963: 31f, 1094ff), point in the direction of ontological individualism. But thereare other indications pointing in the opposite direction.In his first important book, Cours d’Economie Politique ([1896] 1976), Paretodresses his theory explicitly in organicist language, usually associated with socialholism, and social evolution is an important theme. In his later works, organicismgives way to a theory of the economic system, as part of a larger social system,and evolution becomes a less prominent theme. Still, there is much that points inthe direction of social holism.Pareto was a fierce critic of theories dealing with the origins of social institutions(see, e.g., Pareto [1909] 1972: 17f), as did the theory of the social contractand some of the Austrian economists. Instead of the causal-genetic approach, hefavoured theories of functional interdependence. Pareto’s main units of analysiswere the economic system and the larger social system. The main elements ofthese systems are ‘residues’ (motives) and ‘derivations’ (beliefs) inhering in individualhuman beings (pp. 1433ff), but residues and derivations are partly socialfacts (pp. 33, 71–4). According to Pareto, they are individual, from one point ofview, and social from another.Investigating whether moral sentiments have an individual or social origin isuseless. The man who does not live in society is a very unusual man, onewho is almost, or rather entirely, unknown to us. And a society distinct fromindividuals is an abstraction which does not correspond to anything real.Consequently, all the sentiments which are observed in a man living insociety are individual from one point of view, and social from another.(Pareto [1909] 1972: 71)The fact that we deal with individua by no means implies that a number ofindividua taken together are to be considered a simple sum. They form

38 Backgroundcompounds which, like chemical compounds, may have properties that arenot the sum of the properties of their components.(Pareto, [1916] 1963: 32)This is the doctrine of emergence, typically used by holists, such as Wundt andDurkheim, to defend their position. It recurs in Pareto’s discussion of thecomposition of social forces, where he suggests that some composition effects areanalogous to mechanical resultants, while others are analogous to thecompounding of chemical elements (pp. 1446ff). Much like Durkheim, Paretoargues that these emergent social facts are objective and external to individualhuman beings. He also accepts the reality of social groups and makes a distinctionbetween individuals and the positions they occupy in the social structure.There is, thus, quite a lot that goes against an interpretation of Pareto as aprogrammatic methodological individualist.Holism and collectivismMethodological individualism has been opposed to a large number of nominallydifferent doctrines, but most commonly to methodological collectivism and holism.In the early stages of its development, especially in the case of Austrian methodologicalindividualism, it was usually opposed to collectivism. In the later stagesof its development, methodological individualism has been increasinglycontrasted with methodological holism. This change, which is due mainly to KarlPopper, was not possible until J.C. Smuts introduced the term ‘holism’ in hisbook Holism and Evolution (1926). I think it is possible to make a distinctionbetween collectivism and holism, which may be worth keeping in mindthroughout the rest of this book.In the doctrine of emergent evolution, which was a source of inspiration forboth Smuts and Durkheim, these two root ideas are combined. The suggestion isthat new phenomena, such as life, individual consciousness, and ‘folk souls’ emergeas a result of the organisation of parts into wholes, by way of creative synthesis.Holism is the term here coined (from olos = whole) to designate that wholewardtendency in Nature, this fundamental feature of ‘wholes’ in theuniverse … Every organism, every plant or animal is a whole with a certaininternal organisation and a measure of self-direction, and an individualspecific character of its own … beyond the ordinary domain of biology itapplies in a limited sense to human associations like the state, and to thecreations of the human spirit in all its greatest and most significant activities.(Smuts [1926] 1936: 96)Genuine wholes differ from mere aggregates by being more than the sum oftheir parts. ‘A whole which is more than the sum of its parts, has somethinginternal, some inwardness of structure and function, some specific inner relations,some internality of character or nature which constitutes more’ (Smuts

Background 39[1926] 1936: 101f). The difference, then, is that the parts of genuine wholes areinternally related. This makes them different also from mechanical composites,whose parts are only externally related. ‘It is the very essence of a whole thatwhile it is formed of its parts, it in turn influences the parts and affects their relationsand functions. This reciprocal influence underlies the internality or interiorcharacter of the whole’ (p. 103).The picture of the world painted by the doctrine of emergent evolution is that ofthe universe as a hierarchic order of different levels, where each level has emergedfrom, but is supervenient upon, and irreducible to, the level immediate below.We have a progressive superposition of level on level. Higher kinds of relatedness– chemical, vital and conscious – are each in turn supervenient on thosethat stand lower in the scale; but they do not supersede them in the sense that,when some higher kind of relatedness comes, the lower kinds go.(Morgan, 1923: 278)According to the idea of social holism, society is the highest level of organisationin this hierarchic order. It is supervenient upon, but irreducible to, humanindividuals and their actions. This is the meaning of Emile Durkheim’s claimthat society is sui generis. Actually, however, there are two distinct, though relatedclaims made by Durkheim and by most social holists and both claims can betraced back to German Romanticism.In the section on German historicism, above, I made a distinction betweenorganicism and objective idealism. According to the first doctrine, societies arewholes made up of interdependent and functional parts. If we assume that theparts of societies are individuals, the relation between individuals, according toorganicism, is that of part-whole. According to objective idealism, on the otherhand, social organisms have minds, or souls, which are expressed in theirlanguage, customs, laws, literature, etc. In this case, the relation between individualsand folk souls is not that between part and whole, but rather something theyhave in common, something they share. The relation here is more akin to that ofparticular-universal. Emile Durkheim’s distinction between mechanic and organicsolidarity was an early attempt to catch this difference (Durkheim, 1893: 70ff).More recently, Philip Pettit ([1993] 1996: chs 3–4) has suggested that the oppositeof individualism is collectivism, but not holism, which is instead the oppositeof atomism. I agree with Pettit that there is a difference between collectivism andholism, but I do not agree with his rendering of this difference. 30The ideas of social organisms and social minds are metaphors, which weresoon abandoned by social scientists, although they remain as root ideas aboutsocial life, even in contemporary social science. Organicism turned into structural-functionalism,which combines two holistic ideas: structure and function.The ideas about folk souls, national spirits and collective consciousness haveturned into the social scientific idea of culture.Now, I suggest that it is possible to make a distinction between holism andcollectivism. Holism is the idea that there are wholes made of social parts, which

40 Backgroundare interrelated in a way that makes them different from a mere aggregate andirreducible to the parts taken in isolation. Collectivism is the idea that individualshave something in common which turns them into a collective. The mostcommon form of collectivism is ‘idealistic’. Individuals are supposed to share aculture that is common to the group to which they belong. But collectivism mayalso derive from structuralism. In this case, individuals are supposed to form acollective by virtue of the fact that they occupy the same position in the socialstructure. This is the Marxist case of class analysis.Holism and collectivism often combine into entities which are both wholes and collectives.All organisations, for instance, tend to be both: the wholes made up of parts (this is themeaning of organisation), but they also tend to develop a common culture and typicallythey act in common. To this extent, they are also collectives. A simple illustrationof the distinction between wholes and collectives, would be this:Figure 2.4 Social wholes and collectivesHaving made this distinction between collectivism and holism, I immediatelyannounce that I intend not to use it, unless it is called for by the circumstances orpurpose at hand. Throughout this book, I am going to use the term ‘holism’ torefer both to collectivism and holism.

3 Psychologism in early socialscienceMethodological individualism, as a programme in social research, goes back tothe nineteenth century, and the various attempts to lay a foundation for thesocial sciences. The rise of a specifically methodological version of individualismat this particular time probably had much to do with the divorce of the varioussocial sciences from philosophy, and their establishment as academic disciplinesat the universities of Europe and the USA. To turn the individualist theories ofsociety into a methodological rule, principle, or programme for social scientificanalysis answered a deeply felt need for philosophical justification.As we have seen in chapter 2, there existed at the beginning of the nineteenthcentury one social science; economics, and this science was largely individualistic.Other approaches to society, such as German historicism and Frenchpositivist sociology developed, partly, as reactions to the individualism ofeconomics. Methodological individualism is, I suggest, a reaction to this reaction.In the nineteenth century, philosophers and social scientists started to reflect onthe proper methodology of the social sciences, and the issue of individualismversus collectivism in social science was one of the most hotly debated.The nineteenth century was also characterised by the emancipation of thevarious human sciences from philosophy and their further differentiation into anumber of academic disciplines fighting for a territory of their own. In particular,the relation of psychology first to philosophy and second to the varioushuman sciences was a matter of much concern, and it was closely intertwinedwith the issue of individualism versus collectivism. Since psychology is thescience of human individuals, it is only natural that methodological individualismshould take the form of a requirement or suggestion that the humansciences must be based on psychology. It is the thesis of this chapter that individualism,as a methodological principle or programme for social research, arose inthe nineteenth century and made its first appearance in the form of psychologism.The term ‘methodological individualism’ did not appear until thetwentieth century.There is little agreement about the meaning of the term psychologism. 1 Formy purposes, it is a form of reductionism. Some other theory or discipline isreducible to, or at least heavily dependent upon, some form of psychology. Ileave the term ‘reduction’ undefined in this chapter (see chapter 11 for a more

42 Psychologism in early social sciencedetailed discussion), except to say that I understand it in a broad sense to includeboth complete and partial reduction. Most often, in the nineteenth and earlytwentieth century, it was philosophy that was supposed to be reducible topsychology. We may call this version ‘philosophical psychologism’. My interest,however, is in psychologism with respect to the human sciences: the thesis thatthey are reducible to psychology. This thesis might be called ‘scientific psychologism’.These two versions of psychologism are analytically distinct, but notindependent. Many psychologicists have endorsed both and some anti-psychologicistshave defended the autonomy of both philosophy and social science andfor a similar reason. The methodological individualist Karl Popper, for instance,would be one example (see pp. 202–10). For him, the existence of a third world,or world 3, of objective ideas makes both philosophy and the social sciencesautonomous vis-à-vis psychology.‘Psychology’, in its turn, may also mean different things (see pp. 331–36), butfor the purposes of this chapter, I suggest that psychology is whatever putativeknowledge we have of the minds and behaviour of individual human beings. Onthe basis of this broad view of psychology, I suggest that there are three forms ofpsychologism in social science: (1) scientific psychologism, which claims that all socialphenomena can be explained by and all social laws deduced from, psychologicallaws, existing or potential; (2) folk psychologism, which maintains that all socialphenomena can be explained by mental phenomena and/or human actions, and(3) ontological psychologism, which maintains that social phenomena are made up of,or constituted by, mental phenomena and/or human actions.Psychologism as an approach to philosophy, including social philosophy, is asold as individualism and, for obvious reasons, closely related to it. Psychology, forall its diversity, is a theory about the mental states and/or behaviour of individualhuman beings. Virtually all individualist theories of society, therefore,have been based upon some ‘psychological’ view of human nature. Actually,there is one view of human nature, in particular, which has dominated thehistory of individualism: the view of man as a rational egoist. This theory isreally the conjunction of views of human nature: the Aristotelian idea of ‘man’as a rational animal and Epicurean hedonism. This view of human nature wasthe psychological foundation of the theory of the social contract. It was eventuallyadopted and baptised as economic man (see Myers, 1983).A second psychological theory of considerable importance in the history ofsocial thought is the suggestion that our beliefs result invariably from the associationof ideas. This theory was held by most British empiricists and is animportant part of David Hume’s theory of knowledge, in particular. The mostelaborated version of associationist psychology is to be found in James Mill’sAnalysis of the Phenomena of the Human Mind (1829). 2As we shall see in later chapters, these two psychological theories, or views ofhuman nature, are far from dead. On the contrary, they are very much part ofthe short history of methodological individualism; the first, as part of economicsand eventually of rational choice theory in all social sciences; the second becauseit is part of John Stuart Mill’s psychologism and the forerunner of behaviourist

Psychologism in early social science 43psychology, which was the psychological foundation of George Homans’smethodological individualism (see pp. 190–96).The doctrine of psychologism made its first appearance in the nineteenthcentury along with the discipline of psychology, and immediately became amatter of heated controversy. An important reason for this was, no doubt, thecompetition between representatives of the various academic disciplines forrecognition and resources at a time when these disciplines were institutionalisedat the universities in Europe and the USA (Kusch, 1995). Important as it is,however, this is not the only reason. There were also a number of genuine questionsconcerning the proper content and methods of the emerging socialscientific and humanistic disciplines at issue.My interest in this chapter is mainly in psychologism as a doctrine whichmaintains that the foundation of the human sciences is to be found inpsychology. As I have already suggested, this version of psychologism may beseen as the first form of the principle of methodological individualism in thesocial sciences. Following the methodological individualist Karl Popper, we maycall it psychologistic methodological individualism. Its origin is in Great Britainand the first social science it conquered was economics. In the latter half of thenineteenth century, the centre of psychologism was in Germany, where its mainimpact was on the humanities and eventually on sociology.British psychologismI have mentioned two psychological theories in the history of social thought:hedonism and associationism. Both of them had their strongest hold in GreatBritain. I think it is fairly safe to suggest that before the nineteenth century atleast Great Britain was the main seat of psychologism. In no other country wasthe interest in human nature more intense. Psychology was at the centre ofdiscussion, both among the British moralists and of the early social scientists.The conviction of most British social philosophers was that there is a universalhuman nature, constant over time and all over the world. The only problem wasto find out exactly what this universal human nature is.John Stuart MillJohn Stuart Mill is often mentioned as an exponent of psychologism, both inphilosophy and in the social sciences. My interest is in the latter, viz., his viewthat social laws can and should be reduced to psychological laws. This view isgenerally considered a form of methodological individualism. 3 Indeed, themethodological individualist Karl Popper ([1945] 1966: 88–99) has appointedMill to being the main representative of ‘methodological psychologism’, which isaccording to him, an extreme and untenable form of methodological individualism.Popper’s follower Joseph Agassi (1960) has suggested the term‘psychologistic individualism’, to designate the position Popper ascribed to Mill.Lorenzo Infantino (1998: 45–7) makes a distinction between the two, but I

44 Psychologism in early social sciencesuppose he would not deny that psychologism is a form of methodological individualism.George Homans, however, is a methodological individualist whoequates the two doctrines. Maybe that is because he is the only recent methodologicalindividualist who was clearly indebted to John Stuart Mill?As a philosopher, Mill was deeply rooted in the British empiricist tradition.Like Locke, Berkeley and Hume, he believed that all knowledge, not self-evident,is based on induction from sensory experience. Like all empiricists, he alsorejected conceptual realism, but he was not a nominalist, like Hobbes andBerkeley. With Locke, he accepted the existence of abstract ideas, which is to saythat he was a conceptualist. Above all, however, John Stuart Mill was a defenderof the associationist psychology, used by David Hume and his own father toexplain the acquisition of belief. 4 Mill was not a naive empiricist, though. Hedefended the deductive method in economics and he was a pioneer of the hypothetical-deductive,or covering-law, model of scientific explanation.Mill was also a utilitarian of sorts, but not in the form given to this doctrineby Jeremy Bentham and his father. At the end of the 1820s, he was exposed tocertain alternative ways of looking at human beings in society and he was led ina direction away from simple-minded utilitarianism, both in ethics and socialscience. In his Autobiography (Mill [1873] 1961: 94ff), he mentions three people asparticularly important in modifying his utilitarian view of ‘man’ and society: thehistorian Thomas B. Macaulay (1800–59), the poet and philosopher SamuelTaylor Coleridge (1772–1834) and the philosopher and sociologist AugusteComte (see pp. 27–29).I have already mentioned the strong impression that Macaulay’s critique ofJames Mill’s ‘Essay on Government’ made on his son. The younger Mill wasshaken in his belief in the excellence of utilitarianism and of the ‘geometric’, orabstract-deductive, method. His doubts were reinforced by his reading ofColeridge, who introduced him to the ideas of German Romanticism (Mill[1840a] 1969). A first insight gained from reading Coleridge was that there arecertain essential requisites of civil society, not noticed by Enlightenment thinkers.Mill mentions three: (1) ‘a system of education’ incurring a ‘restraining discipline’; (2)‘a feeling of allegiance, or loyalty’, and (3) ‘a principle of cohesion among themembers of the same community or state’ (Mill [1840a] 1969: 133–5). A secondinsight, gained from acquaintance with the work of Wilhelm von Humboldt, inparticular, concerned the existence and value of individuality; of differencesbetween individuals, ignored by utilitarian individualism (Mill [1859] 1975:69–91; Hinchman, 1990). According to Mill, Bentham himself was entirelyinsensitive to spiritual ends and failed to recognise the fundamental importanceof self-formation, or self-culture, in a system of ethics. But formation of self, orcharacter is also important in social science, where it is the subject matter of theall-important science of ethology (see below). Self-formation is, finally, importantin Mill’s solution to the problem of freedom. While it is the task of ethology toexplain individual character as the necessary consequence of circumstances, it ispossible for individuals to influence their own circumstances and, thus, theircharacter (Mill [1843–72] 1974: 836–43).

Psychologism in early social science 45The extent of Mill’s debt to Comte is a matter of some controversy.According to Hayek, ‘Mill himself, in the sixth book of his Logic, which dealswith the methods of the moral sciences, became little more than an expounderof Comtian doctrine’ (1955: 186). Alan Ryan, points out that Mill himself ‘wasanxious not to be thought to have learned everything from Comte’ and adds that‘there is every reason to believe his declaration on this point’ (1974: 228). So far,I agree with Ryan, but there is, on the other hand, no reason at all to believeMill, if the following statement is a correct rendering of his view of the matter:‘[t]he only area in which Mill recognised a clear debt to Comte was that of theso-called method of inverse deduction’. Mill’s debt went further than this, but hedid not share Comte’s view of psychology and political economy and, above all,he positively disliked Comte’s political thinking, not to speak of his later ideas ofpositivism as a religion of humanity. Probably, most of the differences betweenthem boils down to the fact that Comte was a holist and collectivist, whereas Millwas an individualist.John Stuart Mill’s System of Logic (1843) was probably the most influentialwork written on scientific method in the nineteenth century and this is especiallytrue of the part devoted to the ‘social’ or ‘moral’ sciences. 5 Mill did not use theterm ‘methodological individualism’, but he did address the issue and advancedcertain ideas that are clear manifestations of ‘methodological individualism’, inthe sense eventually attached to this term.Like Comte, Mill was a methodological monist, who saw but one remedy ofthe backwardness of the social sciences: the use of the method that had alreadyproved so successful in the natural sciences. He also agreed with Comte that thecomplexity of social phenomena makes this an exceedingly difficult task. Millrecognised three types of law in social science: (1) laws of the mind; (2) laws ofthe formation of character, and (3) laws of society.Laws of the mind are, of course, psychological law belonging to the science ofpsychology ([1843–72] 1974: 849–60). The main example of such laws are thoseof the laws of association, identified by Mill’s father James Mill (p. 852). Sincethere are no innate ideas, all our beliefs are a result of the association of ideas.Of our desires, on the other hand, some may be natural, even though most ofthem are acquired (p. 856). It is interesting that Mill sees a place for ‘mentalchemistry’, as distinguished from mechanics. Just as two substances like hydrogenand oxygen, when mixed, give rise to a new substance, water, which has propertiesof its own, so association of ideas may give rise to new ideas. 6 This is theidea of emergence, which was eventually to play such an important role in thedoctrine of holism. It is significant, though, that Mill recognised emergence onlywithin the individual mind, but not in the collective consciousness, as did theholist Emile Durkheim (see p. 35).Laws of the formation of character belong to the science of ethology, whichwas intended to be Mill’s own contribution to social science. Ryan has suggestedthat it corresponds to social psychology and I believe that this is largely correct.The subject matter of ethology is the way human beings are socialised andeducated to become new members of society. Because the circumstances in

46 Psychologism in early social sciencewhich individuals are brought up are never exactly the same, each individual willbecome to some extent unique. Also, since circumstances differ between nations,there will be different national, or collective, characters. This phenomenon usedto be the main object of investigation by German Völkerpsychologie, but it is notwithin the scope of mainstream social psychology today.The laws of ethology are either empirical or causal. As examples of theformer, Mill mentions the maxims of common wisdom that exist in allcountries. 7 Invaluable as they are in everyday life, these approximate generalisationsare not scientific. To the extent that they are true, they need to beexplained by causal laws about the formation of character. The science ofethology does not exist, but it is an absolutely necessary precondition for thedevelopment of social science. The causal laws of ethology are the axiomata mediawhich connects the laws of society with the laws of mind. 8The laws of the formation of character are derivative laws, depending uponthe laws of mind and the specific circumstances. They tell us what will be theinfluence of circumstances on the formation of character. ‘In other words,Ethology, the deductive science is a system of corollaries from Psychology, theexperimental science’ (p. 872).Mill went on to make clear that the science of ‘[e]thology is still to be created’(p. 872) and he saw it as his own task to do so. His efforts in this direction failed,however, and he turned to other projects (Feuer, 1976; Collini, et al., 1983: 156ff).Among other things he wrote a book on political economy and an autobiography.Laws of society, finally, may be divided into laws of coexistence and laws ofsuccession. The possibility of laws of coexistence are based upon the fact thatthere is a consensus among the different parts of society; that only certaincombinations of social phenomena can coexist. Mill even uses the holisticmetaphor of an organism to express this fact ([1843–72] 1974: pp. 899, 912).‘The fundamental problem’ of social science, however, ‘is to find the lawsaccording to which any state of society produces the state which succeeds it andtakes its place’ (p. 912). The discussion, in which this passage occurs, has ledPopper to accuse Mill of both holism and historicism (Popper, 1957: 72). Ibelieve this accusation is a bit unfair. First of all, I think it is pretty obvious thatMill makes no serious use of the organism metaphor. He talks about ‘states ofsociety’ succeeding one another, but this is ‘holism’ only in an extensive sense, tobe carefully distinguished from true holism, which asserts that there are socialwholes, which cannot be explained in terms of individuals. According to Mill,however, all laws about social states must be explained in terms of psychologicallaws. The same goes for Mill’s alleged historicism. He does suggest that there arehistorical laws of development, even if they are hard to come by, but they havethe status of empirical laws, or trends, only. Like the laws of coexistence, they areonly derivative law, which must be deduced from the ultimate laws of humannature ([1843–72] 1974: p. 916).A problem with finding the laws of society is that experiments are notfeasible. The inductive or ‘chemical’ method suggested by Bacon is, therefore,

Psychologism in early social science 47inapplicable. Nor is the abstract or ‘geometric’ method used by Hobbes and theBenthamist school appropriate, except, perhaps, in the case of economics. Oneproblem with the geometric method is that it relies entirely on the assumption ofself-interest, but does not admit of conflicting forces in the form of mixedmotives (pp. 887ff). The method of the social sciences is the ‘physical’ orconcrete deductive. The following statement from Mill’s Autobiography is interesting,because it reveals the connection, first, between the method of concretededuction and a mechanistic view of society and, second, between the latter andindividualism. It is a characteristic feature of mechanical forces that they can beadded, or aggregated, in a way that makes the whole equal to, instead of morethan, the sum of its parts.My practice (learnt from Hobbes and my father) being to study abstractprinciples by means of the best concrete instances I could find, theComposition of Forces, in dynamics, occurred to me as the most completeexample of the logical process I was investigating. On examining, accordingly,what the mind does when it applies the principle of the Compositionof Forces, I found that it performs a simple act of addition. It adds the separateeffects as the joint effect. But is this a legitimate process? In dynamics,and in all the mathematical branches of physics, it is; but in some othercases, as in chemistry, it is not … I now saw that a science is either deductiveor experimental, according as, in the province it deals with, the effects whenconjoined, are or are not the sums of the effects which the same causesproduce when separate. It followed that politics must be a deductive science.It thus appeared that both Macaulay and my father were wrong; the one inassimilating the method of philosophizing in politics to the purely experimentalmethod of chemistry; while the other, though right in adopting adeductive method, had made a wrong selection of one, having taken as thetype of deduction, not the appropriate process, that of the deductivebranches of natural philosophy, but the inappropriate one of pure geometry,which, not being a science of causation at all, does not require or admit ofany summing up of effects.(Mill [1873] 1961: 98)In the case of the general science of society, or sociology, however, the onlypossible method is that of inverse deduction, suggested by Comte (see p. 29).Since only such social states can exist that are compatible with human nature,laws of society must be subjected to ‘constant verification by psychological andethological laws’ (p. 917).An important vehicle for tracing and verifying empirical laws of society is theemerging field of statistics. In an addition to the 1862 edition of a System of Logic,Mill (pp. 931–42) comments upon the recently published History of Civilization inEngland (1857) by Henry Thomas Buckle. In this book, as we have seen (p. 30)Buckle argued, like Quetelet before him, that the constancy of certain statisticalphenomena, such as suicide and crime rates, is proof that there are statistical

48 Psychologism in early social sciencelaws. It is significant that Mill rejects this argument in favour of a more individualistinterpretation (p. 934). He also gives qualified support to the ‘aristocratic’individualism of Carlyle, which assigns a special importance to great men inhistory. While agreeing that ‘the varieties of character among ordinary individualsneutralize one another on any large scale, exceptional individuals inimportant positions do not in any given age neutralize one another’ (p. 937).According to Mill, laws of society are empirical laws, or generalisations, only.As such they must be explained by the causal laws of the mind and of the formationof character on which they ultimately depend (pp. 896, 908, 914, 924f). 9The laws of the phenomena of society are, and can be, nothing but the lawsof the actions and passions of human beings united together in the socialstate. Men, however, in a state of society are still men; their actions andpassions are obedient to the laws of individual human nature. Men are not,when brought together, converted into another kind of substance, withdifferent properties, as hydrogen and oxygen are different from water, or ashydrogen, oxygen, carbon, and azote, are different from nerves, muscles,and tendons. Human beings in society have no properties but those whichare derived from, and may be resolved into, the laws of the nature of individualman. In social phenomena the Composition of Causes is theuniversal law.(Mill [1843–72] 1974: 879)This passage is usually quoted as proof that Mill was a methodological individualistand, as such, it seems conclusive. It seems equally clear that hismethodological individualism was psychologistic. According to Mill, all causallaws in social science are psychological laws about human nature. Social scienceis turned into large-scale psychology. More precisely, however, it is a metaphysical,or ontological, justification for methodological individualism. Mill explicitlydenies that social phenomena have emergent properties and make up a reality ofits own, or sui generis. Instead, he affirms the mechanist view of social phenomenaas resultants, or composite effect, of the actions of individuals as governed by thelaws of human nature.It seems clear that Mill was a psychologistic methodological individualist, asKarl Popper maintained. But Popper went further and maintained that Mill’spsychologism ‘forces him to adopt a historicist method’ (Popper [1945] 1966: 92)and ‘to operate with the idea of a beginning of society’ (p. 93). Alan Ryan rejectsPopper’s argument and maintains that Mill’s psychologism is not different fromPopper’s methodological individualism (Ryan, 1970: 156ff). Since it is not myaim to criticise psychologism here, I will not consider Popper’s argument, eventhough I believe that he is basically right (Udehn, 1987: 192–201). I also believe,however, that Ryan is right to maintain that Mill’s psychologism is not differentfrom Popper’s methodological individualism. It takes Popper’s institutionalism toget rid of psychologistic historicism and Mill assigns no explicit role to socialinstitutions in his methodology of the social sciences. There are only the laws of

Psychologism in early social science 49individual and social psychology (ethology). It may be argued that the science ofethology makes room for social institutions, but being a science of the formationof character, it is bound to give a psychologistic interpretation of them, and thisis not enough to escape historicism in the sense of a quest for origins. It is quiteanother matter that Mill often treats institutions as real factors given to analysisand treated as causal factors in his empirical investigations. 10According to Mill, then, the social sciences are based on psychology. Thisview was in contradistinction to that of Comte, who saw no role at all for thisscience. But Mill took exception also to Comte’s view of economics. Whereas thelatter argued that the economy must be analysed as part of society and inductively,rather than deductively, Mill defended the abstract-deductive approach ofhis father and his friend David Ricardo.Mill’s philosophy of the social sciences made a tremendous impact on theearly human sciences. Even those who did not share his views, like WilhelmDilthey, advanced their own views in opposition to the positivism of Mill andComte. The Austrians read Mill and were probably to some extent influenced byhis psychologism, even if other influences may have been stronger. Karl Popperdeveloped his institutionalism in contrast to Mill’s psychologism. The most clearexample of Mill’s influence is the sociologist George Homans, who suggests aversion of methodological individualism, which looks almost exactly like that ofMill (see pp. 190–96). But it is possible to see similarities also between Mill’spsychologism and the quest for microfoundations, advanced recently by manysociologists (see pp. 187–89). Like the latter, Mill maintained that empiricalgeneralisations; both those of common sense and those of official statistics, mustbe provided by microfoundations in the form of causal laws of psychology.Figure 3.1 John Stuart Mill’s psychologistic methodological individualismUtilitarian economicsIn chapter 2, we saw that classical economics was largely if not wholly individualistic.There was also an important element of institutionalism in several of theclassical economists. In my opinion, economics became definitely individualisticin 1871, when Stanley Jevons and Carl Menger made the marginalist revolution,together with Leon Walras (1874). The idea that revolutionised economics wasthat economic decisions about what to buy depend, not upon the total utility, but

50 Psychologism in early social scienceon their marginal utility to the consumer. What matters is not the utility of thetotal quantity I possess of a certain good, but the utility of an additional unit ofthat good. An important part of this revolution was also the switch from anobjectivist to a subjectivist theory of value. A consequence of this switch was ashift in economic theory, from a concern with the problems of production anddistribution to an almost exclusive preoccupation with consumption. All marketphenomena were explained, in the last instance, as the result of the consumer’ssubjective valuation and choice of goods.In this section, I am going to concentrate on the British branch of neoclassicaleconomics, which is the clearest example of psychologism, at least in theearly phase of its development. The psychological theory, used by the earlyBritish marginalists was, of course, the utilitarianism of Jeremey Bentham andJames Mill.Stanley Jevons (1835–82) was first to introduce marginalism in Britisheconomics. He did not use the term ‘marginal utility’, however, but called it ‘finalutility’. Jevons was also first to build economics on the foundation of utilitarianism,and described his theory as ‘the mechanics of utility and self-interest’ ( Jevons[1871] 1970: 90). He took utilitarianism from Bentham, but turned it from anethical theory into an explanatory theory of human behaviour. ‘In this work Ihave attempted to treat economy as a calculus of pleasure and pain, and havesketched out, almost irrespective of previous opinions, the form which thescience, as it seems to me, must ultimately take’ ( Jevons [1871] 1970: 44).Pleasure and pain are undoubtedly the ultimate objects of the calculus ofeconomics. To satisfy our wants to the utmost with the least effort – toprocure the greatest amount of what is desirable – in other words, tomaximise pleasure is the problem of economics.( Jevons [1871] 1970: 101)It may be noticed that, for Jevons, utility was not a psychic phenomenon, but aquality of objects with the capacity of producing a feeling of pleasure in itsutilisers. However, in order to maximise pleasure or happiness we have tomaximise utility.Jevons did not state any principle of methodological individualism, althoughthere are some indications of individualism in his major work. He rejected thepossibility of comparing the feelings of pleasure in different minds and, hence, ofarriving at a utility function for a collectivity of individuals. According to Jevons,‘the weighing of motives must always be confined to the bosom of the individual’(p. 85).I must here point out that, though the theory presumes to investigate thecondition of a mind, and bases upon this investigation the whole ofeconomics, practically it is an aggregate of individuals which will be treated.The general form of the laws of economics are the same in the case of individualsand nations; and, in reality, it is a law operating in the case of

Psychologism in early social science 51multitudes of individuals which gives rise to the aggregate represented in thetransactions of a nation.( Jevons [1871] 1970: 86)Jevons here points to a practical problem which always did beset methodologicalindividualism and which has led some social scientists to adopt a more collectivisticmethodology. He continues:Practically, however, it is quite impossible to detect the operation of generallaws of this kind in the actions of one or more individuals. The motives andconditions are so numerous and complicated that the resulting actions havethe appearance of caprice, and are beyond the analytic powers of science.( Jevons [1871] 1970: 86)The solution to this problem is to use averages, or aggregates, assuming that,in the long run, accidental and disturbing causes will neutralise each other andproduce a more predictable aggregate outcome. ‘Accordingly, questions whichappear, and perhaps are, quite indeterminate as regards individuals, may becapable of exact investigation and solution in regard to great masses and wideaverages’ (p. 86). Jevons’s solution is not, however, to follow Quetelet and give upindividualism, but to insist that laws of the aggregate depend upon laws of individualbehaviour:The laws which we are about to trace out are to be conceived as theoreticallytrue of the individual; they can only be practically verified as regardsthe aggregate transactions, productions, and consumptions of a large bodyof people. But the laws of the aggregate depend of course upon the lawsapplying to individual cases.( Jevons [1871] 1970: 108f. See also p. 135)Jevons is no more explicit about the assumption of rationality. Of course it maybe objected that the assumption of rationality is implicit in that of utilitymaximisationor, even, that they are identical. This is a common view today, butit is a view that I reject for at least two reasons. First of all, rationality presupposesconscious intention, whereas utility-maximisation does not, and second,individuals may try to maximise utility, but fail because they are irrational. Itmay be argued that Jevons talks about the calculus of pleasure and pain and thatcalculation certainly is part of rational choice. He also mentions that we derivepleasure from anticipation of future pleasure (p. 99). This is a further indicationof conscious intention. On the other hand, he immediately adds that, in ourordinary affairs, we make these calculations almost unconsciously (p. 100).Francis Y. Edgeworth followed Jevons in leaning on Bentham’s principle ofgreatest happiness, but considered two different versions of this principle: theethical and the economic. He called the former ‘utilitarian’ and the latter ‘egoistic’.According to Edgeworth the scope of economics is limited to self-interested

52 Psychologism in early social scienceutility-maximisation. In his well-known statement, ‘the first principle ofEconomics is that every agent is actuated only by self-interest’ ([1881] 1967: 16).As things turned out, however, also the utilitarian version would become aconcern of economists. Edgeworth’s utilitarian calculus may be conceived of asan early contribution to welfare economics. More important, from a narrowlyscientific point of view, Edgeworth was first to introduce the idea of indifferencecurves (pp. 21ff), thereby creating the possibility for an alternative mathematicalrepresentation of economic theory.Edgeworth’s distinction between an economical and a utilitarian calculus wasbased on a distinction between ‘egoistic’ and ‘universalistic’ hedonism made byHenry Sidgwick in his monumental and influential The Methods of Ethics (1874).There is no exact correspondence, however, since Sidgwick conceived of both asethical doctrines to be distinguished from ‘psychological hedonism’. The lattercorresponds to Edgeworth’s egoistic principle, which is also the first principle ofthe science of economics.Sidgwick was of another opinion, however. He defended universal hedonism,or utilitarianism, as an ethical theory, but rejected egoistic hedonism.Benevolence, not self-interest is the centrepiece of common sense morality and,therefore, of ethics. But there would be little point in defending universal hedonismif human nature were entirely dominated by psychological hedonism.Because of this, Sidgwick went to great length in order to rebut the latterassumption (Sidgwick [1874] 1981: 42ff, 199ff, 497ff). It would still be possible,of course, to use self-interest as a simplifying assumption for the purposes ofeconomics, but Sidgwick rejected even this use of psychological hedonism. Whileadmitting that economics has to use simplifying assumptions (Sidgwick, 1885:37), he denied that it has to use the assumption of strict self-interest.All that the deductive reasonings of English economists supply is a methodof analysing the phenomena and a statement of the general causes thatgovern them, and of the manner of their operation. In this analysis, nodoubt, the assumption is fundamental that the individuals concerned in theactual determination of the economic quantities resulting from freeexchange will aim, ceteris paribus, at getting the most they can for what theysell and giving the least they can for what they buy.(Sidgwick, 1885: 34)The reason this assumption does not imply self-interest is that individualstrying to get the most they can when selling and giving the least they can whenbuying, may do so for the benefit of their family, their friends, or some otherpeople, rather than for their own benefit (p. 29). Sidgwick, then, appears as aprecursor of Wicksteed’s principle of non-tuism. Instead of self-interest,economics relies on the assumption of wealth-maximisation: Other things beingequal, they always prefer more to less (Sidgwick [1883] 1901: 46).

Psychologism in early social science 53The first and most fundamental [assumption] is that in a state of economicfreedom, all persons engaged in industry will, in selling or lending goods orcontracting to render services, endeavour, other things being equal, to get asmuch wealth as they can in return for the commodity they offer. This ismore briefly expressed by saying the Political Economy assumes the universalityand unlimitedness of the desire for wealth.(Sidgwick [1883] 1901: 41)Sidgwick does not express any manifest belief in methodological individualism.He does talk about ‘individualism or economic egoism’ to designate thenature of economic man ([1883] 1901: 34) and he often refers to a free marketeconomy as the ‘individualistic organization’ of industry, or society (pp. 44, 591);still this does not imply methodological individualism. This principle seems to beimplicit, however, in the first quotation from Sidgwick, above, and it mayunderlie his critique of organicist sociology (Sidgwick, 1885: 41–57).Philip Wicksteed was less committed to utilitarianism, and closer, in somerespects, to Austrian than to British marginalism. In his first contribution toeconomics, The Alphabet of Economic Science (1888), he was still exploring the fieldopened up by Jevons, Walras and Marshall and, consequently, much concernedwith ‘marginal utility’, yet he was much more interested in the rational choicesmade by individuals comparing the marginal utility of various goods (Wicksteed[1888] 1970: 51ff). 11 His paradigm of a rational economic agent is the householder,which, as it happens, is usually a housewife:The clever housekeeper has a delicate sense for marginal utilities, and canbalance them with great nicety. She is always on the alert and free from theslavery of tradition. She follows changes of condition closely and quickly,and keeps her system of expenditure fluid, so to speak, always ready to riseor fall on any one of innumerable and ever shifting, expanding andcontracting channels through which it is distributed, and so always keepingor recovering the same level everywhere. She keeps her marginal utilitiesbalanced, and never spends a penny on A when it would be more effective ifspent on B; and combines the maximum of comfort and economy with theminimum of ‘pinching’.(Wicksteed [1888] 1970: 126)Wicksteed is also more individualistic than the other British marginalists. Likethe Austrians, he insists that ‘we must make the relative intensity of the desiresand wants of the individual our starting-point’ (p. 77). Consequently, he starts hisanalysis with the case of Robinson Crusoe. From this individualistic point ofview, it is not legitimate to compare the desires and wants of different individualsand, hence, to arrive at a utility function for the whole community, conceived asa unity. Economics is a science of a ‘catallactic community … in which the individualsfreely exchange commodities one with another, each with a view to

54 Psychologism in early social sciencemaking the enjoyment he derives from his possessions a maximum’ (p. 79,note). 12In Wicksteed’s major work, The Common Sense of Political Economy (1910), utilitymaximisationis pushed a bit further into the background. Economics is nowbased on a ‘psychology of choice between alternatives’ ([1910] 1933: 2f). For thepurposes of economic theory, each individual is endowed with a single scale ofpreferences encompassing all objects of desire (pp. 32–6). The economicproblem confronting each individual is how to administer limited resources inthe choice between alternatives according to his/her scale of preferences.Wicksteed is aware that choice is not always deliberate and not always rational,but maintains that the principle of price is always active (p. 28). Wicksteed’s realpoint of departure is the household, not the individual, but he assumes that it isadministered by an individual in the form of a housewife. The end of economicanalysis, however, is to explain the organization of industry and commerce. As inhis first work, he calls it a ‘catallactic community’, but he also describes it as a‘spontaneous organisation’, made up of ‘spontaneous relations’ (pp. 15f). The‘aggregate’, or combined, result of the actions of all individuals in this catallacticcommunity is called, with a term borrowed from mechanics, a resultant. (see, e.g.pp. 5, 162, 167).Wicksteed is critical of the idea of economic man acting from a specificeconomic motive. Economics is not based on the assumption of a specific motiveat all, but on a specific, impersonal, economic relation called ‘non-tuism’. For thepurposes of economic science there is no need to assume that individuals areself-interested. They may very well act from altruism or benevolence. The onlyassumption needed is that individuals are not altruistic or benevolent towardsparties to exchange, but that they try to strike the best bargain they can in theireconomic transactions. ‘The economic relation does not exclude from my mindeveryone but me, it potentially includes every one but you’ ([1910] 1933: 174).Wicksteed knows, of course, that economic relations are usually ‘embedded’ insocial relations of various sorts, but defends the simplifying assumption ofisolated economic relations for the purposes of economic science (pp. 194ff). 13Economics, then, is not based on assuming a particular motive such as selfinterest,or utility-maximisation, still this does not mean that it is free frompsychology (Wicksteed [1914] 1933: 780): ‘If political economy is the science ofwealth, then it deals with efforts made by man to supply wants and satisfydesires. “Want,” “effort,” “desire,” “satisfaction,” are each and all psychicphenomena’ (Wicksteed [1914] 1933: 766). It is equally obvious, however, thatthe economist is not engaged in doing psychology. Following John NevilleKeynes, Wicksteed argues that economics takes psychological principles as itspoint of departure, rather than subjecting them to investigation. But the ultimatelaws of economics are psychological and ‘it may be argued that politicaleconomy is largely, or even prevailingly, applied psychology, so that the economistmust from first to last realise that he is dealing with psychological phenomena,and must be guided throughout by psychological considerations’ (p. 767).Alfred Marshall (1842–1924) was probably the most influential of the British

Psychologism in early social science 55neoclassical economists, at least in the long run, but he was not altogetherorthodox in his outlook. Marshall was less prone than other neoclassicaleconomists to engage in abstraction and insisted that economics is about realmen and women as members of society and not about an abstract economicman acting in splendid isolation (Marshall [1890] 1920: 12–23). More than otherneoclassical economists, Marshall took an interest in the German HistoricalSchool of Economics and in sociology, and tried to make use of some of its ideas(Coase [1975] 1982: 412). One obvious sign of this influence is that Marshallrejected the orthodox view of human nature as a constant ([1890] 1920: 3ff, 622,630f). Another sign is that Marshall conceived of his own theory as holistic, ororganicist, rather than as individualistic (pp. 20f, 637). 14 Unlike the other Britishmarginalists, he also advocated a development of economic theory in a moreevolutionary direction (Foster, 1993).Marshall wished to play down the break with classical economics and accusedJevons for exaggerating discontinuity between classical and marginalisteconomics. He denied that value is altogether determined by utility and not at allby cost of production. ‘We might as reasonably dispute whether it is the upper orthe under blade of a pair of scissors that cuts a piece of paper, as whether valueis governed by utility or cost of production’ (Marshall [1890] 1920: 290). He alsorejected Jevons’s claim that the theory of consumption is the scientific basis ofeconomics. Economics is a science of human action and of choice, not of wants,or of utility. Wants are of interest, only to the extent that they manifest themselvesin the use of money for their satisfaction. The existence of the measuringrod of money is the basis of economics as a science and the reason for its superiorityto the other social sciences. Marshall, then, might be seen as a progenitor ofthe idea of starting from revealed preferences, rather than from the psychologicalbasis of choice.Economics, then, is a science of action and of choice as manifested in the useof money on the market. But this does not mean that Marshall conceives of allaction as the result of deliberate choice.When we speak of the measurement of desire by the action to which itforms the incentive, it is not to be supposed that we assume every action tobe deliberate, and the outcome of calculation. For in this, as in every otherrespect, economics takes man just as he is in ordinary life: and in ordinarylife people do not weigh beforehand the results of every action, whether theimpulses to it come from their higher nature or their lower.(Marshall [1890] 1920: 17)According to Marshall, much of ordinary life is governed by habit andcustom, but he also suggests that most customs have been adopted and survivebecause they are rational, or functional, especially in the modern world ([1890]1920: 18f; 1885: 48ff). It is a thesis of Marshall, that there has been a historicaldevelopment from a society dominated by custom, to a society increasinglydominated by deliberate choice (Marshall [1890] 1920: 1–11, 602–23). A

56 Psychologism in early social scienceconsequence of this thesis is that the theory of rational choice is of wider scopein modern than in traditional society.With all these provisos, it should finally be admitted that the main legacy ofAlfred Marshall is as a founder of individualist neoclassical economics. It may beargued that Marshall’s methodology rested upon two pillars: the principles ofsubstitution and continuity (cf. Boland, 1992: chs 2–3). According to the first,every economic agent will try ‘to obtain better results with a given expenditure,or equal results with a lesser expenditure’ (Marshall [1890] 1920: 295). In otherwords: individuals are constrained maximisers. The principle of continuity saysthat everything is relative and a matter of degree (314f). Most importantly, theprinciple of continuity applies to time. Thus there is only a difference of degreebetween the long run and the short. In the short run, both technology and socialinstitutions are exogenously given as constraints on maximisation. In the longrun, however, they themselves may be treated as endogenous variables subject toconstrained maximisation. Which variables are conceived of as exogenous andendogenous is entirely relative to the time-perspective and to their stability overtime. Those ‘variables’ that are stable over long periods of time, are treated asexogenously fixed conditions.Although the proximate causes of the chief events in history are to be foundin the actions of individuals, yet most of the conditions which have madethese events possible are traceable to the influence of inherited institutionsand race qualities and of physical nature.(Marshall [1890] 1920: 602)It seems to me that Marshall represents what has been later called institutionalindividualism (see chapter 6). His way of tackling the problem remindsone of the new institutionalism in economics (see chapter 8), which is notsurprising, since he has been a major influence on at least one of its branches,namely the Chicago School. I believe there is a difference, however. WhereasMarshall suggested that we should try to maximise the number of exogenouslyfixed conditions (see Boland, 1992: 27ff), the Chicago School tries to endogeniseas much as possible. The strategy of the former may be described as institutionalisticand that of the latter as individualistic.In the beginning of the nineteenth century, economics was gradually to freeitself from its dependence upon utilitarianism, and turn into a theory of rationalchoice, more or less drained of psychological content (see Stigler, 1950). Part ofthe responsibility for this development lies with the other branches ofmarginalism: the Austrian and Lausanne schools of economics. I believe it wasmainly the Austrians, who turned economics from utilitarianism to rationalchoice (see chapter 4), while the Lausanne school paid virtually no attention atall to the individual, except as a cog in the lifeless machine of the market (see pp.241–50). Of great importance in this development was also themathematical–technical side of the matter. The first important steps were apparentlytaken by Irving Fischer and Vilfredo Pareto in the beginning of the 1890s.

Psychologism in early social science 57Other figures who contributed to this development were Gustav Cassel andEugen Slutsky. The final step was taken by John R. Hicks and Roy G.D. Allen intheir article ‘A Reconsideration of the Theory of Value’ (1934). According tothem (p. 52), the most important figure in the replacement of the concept ofutility by that of a scale of preferences was Vilfredo Pareto, the successor ofLeon Walras in Lausanne (see also Stigler, 1950: 386ff, and Marchionatti andGambino, 1997).It is a bit surprising, therefore, that Lawrence Boland (1982: 27ff) shouldsingle out Pareto’s position, among all economists, as an example of psychologism.However, this is not to suggest, that he is entirely wrong to do so. InPareto’s main sociological treatise, The Mind and Society ([1916] 1963: 1442), weread: ‘All human conduct is psychological and, from that standpoint, not only thestudy of economics but the study of every other branch of human activity is apsychological study and the facts of all such branches are psychological facts’.This passage is quoted by Lawrence Boland (1982: 27) at the beginning of achapter on methodological individualism, where he suggests that Pareto was aproponent of psychologism. But was he? It depends, of course, upon what youunderstand by ‘psychologism’, but I do not think that Pareto is a very goodexample of psychologism in economics. The passage quoted by Boland is from anote, where Pareto is really more concerned to play down the importance ofpsychology for the science of economics (see also pp. 20f, 324), but he might alsohave quoted the following passage from his main work on economics, Manual ofPolitical Economy:Clearly psychology is fundamental to political economy and all the socialsciences in general. Perhaps a day will come when the laws of social sciencecan be deduced from principles of psychology, just as some day perhaps theprinciples of the composition of matter will give us all the laws of physicsand chemistry by deduction; but we are still very far from that state ofaffairs, and we must take a different approach.(Pareto [1909] 1972: 29)This is the programme of reductionism, in general, and of psychologism, inparticular, but it is not a programme that Pareto recommends, as can be seenfrom the last sentence. And it is not his own procedure. Being a follower ofWalras, Pareto’s psychologism is minimalistic indeed. According to Pareto, theeconomist is not interested in the tastes of individuals (Pareto [1909] 1972: 95).Individuals may be represented by their indifference curves. ‘The individual candisappear, provided he leaves us this photograph of his tastes’ (p. 120). 15The relation of economics to psychology has been a constant bone ofcontention. While most economists deny that economics has anything to learnfrom psychology, there have always been dissenters who maintain the opposite.Some of the former go as far as to deny that economics in any way depends uponpsychological knowledge. I believe that this view is mistaken (see pp. 331–36). It israther a matter of the extent to which economics depends on psychological

58 Psychologism in early social scienceknowledge and what kind of psychological knowledge it is. Whether it should bebased on detailed knowledge about human behaviour derived from the science ofpsychology, or upon some simple assumptions derived from common sense.I suggest that we make a distinction between thin and thick versions of psychologism.Thin psychologism is minimalistic, and makes use of very simplepsychological assumptions, like those about economic man. It does not rely onany acquaintance with the science of psychology. Thick psychologism, on theother hand, makes use of more detailed knowledge about the motives andreasoning of human beings. An important source of this more sophisticatedknowledge is, of course, the science of psychology. The psychologism of Paretowas a thin version of this doctrine.Pareto’s main contribution to individualism, was not his psychologism, but hisdenial of the possibility of interpersonal comparisons of utility and his famousidea of optimality, commonly called Pareto optimality, or Pareto efficiency. Thelatter idea is one of the central pillars of the individualist tradition in normativeeconomics and political philosophy. Its basis, the denial of the possibility ofcomparing the utilities of different individuals is also a characteristic mark ofindividualism, but not of utilitarianism (despite the fact that Jevons also deniedthis possibility). Among economists, it is typically the most outspoken methodologicalindividualists, the Austrians, who deny the possibility of comparing theutility of different individuals. The reason, of course, is that it blocks the possibilityof constructing a collective utility-function and, more importantly, of acollectivistic politics, based on such a utility-function. The intellectual roots ofthis view are, I believe, in German qualitative individualism, which is a source,not only for Austrian individualists, but also for Pareto (cf. Carroll, 1973).German psychologismIn chapter 1, we saw that thinking about the individual and society in nineteenth-centuryGermany was dominated by historicism. It is perhaps paradoxicalthat the science of psychology should first emerge in this intellectual context, andthat it should take the form of a natural, rather than a human, science.Nevertheless, it did, but not without a fight. The natural and experimentalscience of psychology met with strong resistance from the start and, in the end,there was a split between an experimental and explanatory psychology, on theone hand, and a phenomenological and descriptive psychology, on the other.The latter was conceived of as the foundation upon which all human sciencesrest. The implication was a kind of psychologism, that was the most commonversion of methodological individualism in sociology (see chapter 5) before therise of rational choice sociology (see chapter 10).The rise of psychologyIt is commonly agreed that psychology as a science, or at least as an academicdiscipline, originated in Germany in the middle of the nineteenth century. It had

Psychologism in early social science 59two roots; one in philosophy, the other in natural science. Of these, the latter wasthe more important in turning psychology into a science, but institutionallyspeaking, psychology emerged as a sub-discipline of philosophy. There were, forquite some time no separate chairs in psychology. Psychologists competed withphilosophers for chairs in their discipline. Like a young cuckoo, the science ofpsychology threatened to crowd out the philosophers from their nest. Or, so itseemed to philosophers proper, who mounted a counterattack in the form ofanti-psychologism. Armed with the mighty weapon of logic, the philosopherseventually succeeded in throwing the psychologists out of their discipline. Thisstory is told in much detail, and in a persuasive manner, by Martin Kusch (1995).In a larger perspective, it is just another chapter in the play of group interestswhich drives the process of the specialisation and differentiation of scientificknowledge.In Great Britain, as we have seen, philosophy was largely psychologistic. Thiswas not the case in Germany: not even at the start. German philosophy in thenineteenth century was largely anti-psychologistic, and this explains some of theanimosity, but even so, it did contribute to the emergence and development ofthe new discipline.It is virtually impossible to say anything at all about German philosophywithout going back to Immanuel Kant. Unlike John Stuart Mill in England,Kant denied that philosophy depends upon psychology. Pure logic, for instance,‘has nothing to do with empirical principles, and does not … borrow anythingfrom psychology … Pure logic is a body of demonstrated doctrine, and everythingin it must be certain a prior’ ([1781/7] 1965: 95). He also denied thatpsychology is possible as a science in its own right. In Kant’s terminology, onlyan ‘empirical’ psychology, but not a ‘rational’ psychology is possible (328ff). ‘Thisis because mathematics is inapplicable to the phenomena of the internal senseand their laws, unless one might want to take into consideration merely the lawsof continuity in the flow of this sense’s internal changes’ (Kant [1786] 1970: 8).It might be added that, for Kant, science means natural science and quantificationis its distinguishing mark. The reason why psychic phenomena are notquantifiable, in its turn, is that psychic experience exists only in time, but not inspace. Perhaps, this says more about Kant’s concept of science, than about theprospects of psychology. According to Kant, chemistry also fails to qualify as ascience. I suspect that Kant’s view of mental phenomena continued to play somerole in the discussions about the possibility of measuring and comparing theutility of different individuals, which has been so important in welfareeconomics.Despite Kant’s own doubts, some of his disciples tried to establish psychologyas a science (see Leary, 1978), and, ironically, two of them, Jacob Friedrich Fries(1773–1843) and Friedrich Eduard Beneke (1798–1854), may have been first toadvance and defend psychologism in philosophy (Notturno, 1985: 12). More wellknown is Friedrich Herbart (1776–1881), who argued, against Kant, that psychologicalexperiences are, indeed, quantifiable. His most famous thesis was that the

60 Psychologism in early social scienceintensity of various sensations tend towards equilibrium. A change in the intensityof one sensation is offset by changes in other sensations.At the beginning of the nineteenth century, Kant was replaced by theGerman idealists, Johann Gottlieb Fichte (1762–1814), Friedrich Wilhelm vonSchelling (1775–1854) and Georg Wilhelm Friedrich Hegel (1770–1831), as thedominant philosophers of the epoch. The tenor of their philosophy was evenmore anti-psychologistic than that of Kant. While psychology is a science of thesubjective experience of individual human beings, German idealism was holisticand objectivistic. From the latter point of view, the approach of psychologyappeared as limited, at best. According to the German idealists, the mind of theindividual is but a moment in the development of the absolute spirit. Even so, itmade some contributions to the burgeoning science of psychology. First of all, itgave philosophical legitimacy to social psychology as a special branch of the newscience. In Germany, social psychology took the special form of folk psychology(Völkerpsychologie). 16 Second, Fichte, in particular, emphasised the active side ofhuman beings, as a manifestation of will. This theme was developed by otherphilosophers, such as Arthur Schopenhauer and by Friedrich Nietzsche, in amore individualistic direction, but it also made an impact on psychology (Leary,1980).The main root of scientific psychology, however, was natural science. Twoimportant figures were Ernst Heinrich Weber (1795–1878) and Gustav Fechner(1801–87), who both – but especially the latter – contributed to the famousWeber–Fechner law of psychophysics. According to this law, the smallest noticeabledifference in the intensity of sensations increases in proportion to thisintensity. The explanation for this fact is that the intensity of our sensationsincreases at a slower rate than the increase in the stimuli which causes the sensation.The most important natural scientist behind the development of scientificpsychology, however, was Hermann von Helmholtz (1821–94), who specialisedin the investigation of the functioning and activity of our senses and muscles.Part of Helmholz’s importance lies in the fact that he was the teacher ofWilhelm Wundt, who is considered by many the real founder of psychology as ascience.Mention should, finally, be made of Hermann Lotze (1817–81), who was aphilosopher and successor of Herbart, but who was trained in medicine andtried to accomplish a marriage between psychology and physiology. AmongLotze’s pupils was Franz Brentano, who might be considered a co-founder ofscientific psychology.The emergence of psychology was gradual, but there is reason to assign aspecial importance to the year 1874. In this year two books of particular importancefor the establishment of psychology as a science were published (seeTitchener, 1921), namely Wilhelm Wundt’s Grundzüge der physiologischen Psychologie(Lectures on Human and Animal Psychology) and Franz Brentano’s Psychologie vomempirischen Standpunkte (Psychology from an Empirical Standpoint).Wilhelm Wundt (1832–1920) was an incredibly productive writer and, in

Psychologism in early social science 61addition, a good organiser, who created the first laboratory for experimentalpsychology in Leipzig in 1879. He did more than anyone to establish psychologyas an academic discipline and therefore is often mentioned as the founder ofpsychology. As I have already mentioned, his background was in physiology andhe always considered psychophysics an important part of psychology. Thecentral part, however, was the study of mental phenomena themselves, andfor the grounding of this part, he relied heavily on the associationist psychologyof James and John Stuart Mill. It might be noted that he borrows the ideaof mental chemistry from the later Mill, but calls it ‘psychic synthesis’(Wundt, 1874: 484). Also this part, however, was to become part ofexperimental psychology. Eventually, Wundt would become much interested inVölkerpsychologie. 17In the first stage of his career, Wundt appears as a follower of Mill also inmatters concerning the logic of the human sciences (Geisteswissenschaften). In thefirst edition of his own Logik (1883), Wundt consistently maintains that (individual)psychology is the foundation of the human sciences (pp. 478ff, 572). Thelatter establishes facts of a singular, or generic, nature, but it takes psychologicallaws to explain them and the reason is that only psychological laws are causal(pp. 500, 514–18, 541). Statistical regularities, for instance, are not causal, butonly empirical, and, therefore, in need of a causal explanation in terms ofpsychological laws (pp. 77–9). A significant difference between Wundt and Mill isthat the former makes an important distinction between history and socialscience. In the case of sociology, he follows Mill closely, except in his opinion ofComte, which is much more negative. Wundt sees no justification for a science ofsociology, à la Comte, which he dismisses as philosophy of history (p. 568f). Likemost Germans, he is also somewhat less impressed by the achievements ofabstract-deductive economics (pp. 586ff). As a German, finally, he assigns animportant place to hermeneutics, or the art of interpretation, in the humanities(pp. 518ff).In his Ethics, however, written only three years later (1886), Wundt has turnedinto a critic of ‘the one-sided individualism of the Enlightenment’ (p. 394) and ofutilitarianism (p. iv). He still maintains that psychology is an important aid to thesocial sciences and history, but now it is Völkerpsychologie, not individualpsychology, which provides the foundation. As we have seen (p. 35), Wundt’sEthics had a profound effect upon the sociologist Emile Durkheim and its place isin the history of holism, rather than in that of methodological individualism.Although Durkheim tried to deny it, Wundt is most probably the source of hisidea of ‘creative synthesis’. As I have already suggested, Wundt took the idea ofpsychic synthesis from Mill and used it to explain novelty in the minds of individuals.This idea is also hinted at in the first edition of his Logik (1883: 523f),where he talks about ‘the creation of psychic energy’ in individuals. It is only inthe second edition of this work (1895: 267–97), however, that creative synthesis(schöpferische Synthese) becomes an important topic and Wundt now applies thisidea to cultural phenomena that originate in a community of individuals, ratherthan in an individual mind.

62 Psychologism in early social scienceFrom the 1880s to the end of his life, Wundt moved progressively away fromindividual psychology and individualism to Völkerpsychologie and social holism. Inthe preface to the posthumously published fourth edition of Logik (1921), his sonMax Wundt writes that, his father’s last concern was to overcome individualism(Wundt, 1921: viii).Franz Brentano was a scholar of a somewhat different bent from Wundt. Histraining was in classical philosophy rather than natural science. As a Catholic hehad a strong leaning towards Aristotle and the scholastics, but he was also wellversed in modern philosophy, including Comte and Mill. Of these two, hepreferred Mill, if for no other reason than that he defended psychology againstthe critique of Comte (Brentano [1874] 1973: 33–5). For Brentano, himself,psychology is the science of mental phenomena (pp. 19, 100) and it occupies animportant place of its own in the hierarchy of the sciences. Brentano goes tosome length in rebutting Comte’s reduction of psychology to physiology (pp.48ff, 128f), but follows Mill in attempting to reduce philosophy and social scienceto psychology. ‘Along with aesthetics and logic, ethics and politics also stem fromthe field of psychology. And so psychology appears to be the fundamental conditionof human progress in precisely those things which, above all, constitutehuman dignity’ (p. 21). With direct address to Comte, he also writes:The observation of mental phenomena in human society undoubtedly shedslight upon the mental phenomena of the individual; the opposite, however,is even more true. Indeed, in general it is a more natural procedure to try tounderstand society and its development on the basis of what has beendiscovered about individuals than to proceed the other way around and tryto shed light on the problems of individual psychology by means of theobservation of society.(Brentano [1874] 1973: 43)Brentano, then, seems to have been a psychologicist and, consequently, amethodological individualist, but this is not a central part of his argument. Hismost well-known contribution to philosophy and psychology is his argument thatmental phenomena are characterised by intentionality; by ‘reference to acontent, direction toward an object (which is not to be understood here as athing), or immanent objectivity’.Every mental phenomenon includes something as object within itself,although they do not all do so in the same way. In presentation something ispresented, in judgement something is affirmed or denied, in love loved, inhate hated, in desire desired and so on.(Brentano [1874] 1973: 88)Brentano’s conception of psychology was somewhat modified towards theend of the century. In his later writings he made a sharp distinction between‘genetic’, or explanatory, and ‘descriptive’ psychology, akin to, but not identical

Psychologism in early social science 63with, that made by Wilhelm Dilthey (see pp. 68, 70f). The former is an inductivescience seeking laws of ‘mental’ phenomena. It is a ‘natural science’connecting mental with physiological phenomena, in the manner of psychophysics.The descriptive psychologist, on the other hand, ‘investigates theconstituents of human consciousness; he seeks out its elements and attempts todetermine as exhaustively as possible their modes of combination’ (Brentano,1890–1, quoted in Chisholm, 1976: 92). Descriptive psychology is also an‘empirical’ though not an inductive science, since it aims at discovering apodictic,or certain, truths about psychic phenomena. It is a posteriori, to the extentthat it is based on experience, but a priori, to the extent that its truths are independentof experience. In this respect, Brentano’s descriptive psychologycomplies with Kant’s synthetic a priori. Most of its statements, however, areanalytic. Brentano’s descriptive psychology, therefore cuts across Kant’s distinctionbetween analytic and synthetic judgements (see Chisholm, 1976; de Boer,1976). To conclude: unlike Dilthey’s descriptive psychology, that of Brentano isan exact science.Brentano’s ideas of a descriptive psychology and of the intentionality ofmental phenomena would become important elements in the burgeoningphenomenology of Edmund Husserl, who had attended Brentano’s lectures as astudent (Husserl 1981: 342–8). More important, for my purposes, is the suggestionthat Brentano influenced members of the Austrian School of Economics. Itis likely that Carl Menger got some of his Aristotelianism from Brentano, andmore than likely that he got his idea of economics as an exact science from him(see Gordon, 1993: 17ff and Smith, 1986: 8ff). It has also been argued thatLudwig von Mises praxeological version of Austrian Economics pointed out thatit owes a lot to Brentano (Chisholm, 1986: 191f). If so, it is a bit odd that Misesdid not recognise the similarities between his own view and that of Menger, butbelieved that the latter was a follower of Mill (Mises, 1969: 27f).Wundt in particular, but also Brentano had a large number of students, whoeventually became famous psychologists. I am not, however, aware of anyonewho made an important contribution to psychologism in the human sciences.Wundt’s experimental psychology also made a great impact outside Germanyand, especially, in the USA. An important link between Germany and the USAwas the British psychologist E.B. Titchener (1867–1927), who had been a studentof Wundt and organised his own laboratory at Cornell University.This is not to suggest, however, that psychology in the USA was importedfrom Germany. The Unites States had a tradition of its own closely associatedwith pragmatist philosophy, which exerted an important formative influence onthe further development of psychology and philosophy, not only in the USA, butalso in Europe. The pioneer was William James, who wrote one of the classics ofthe discipline, The Principles of Psychology (1890), which includes the famouschapter IX on ‘The Stream of Thought’. In this chapter, James rejects the analysisof the mind suggested by associationist psychology. The mind, orconsciousness, is not a passive receptor of sense-impressions, as in Locke’s tabularasa, but a function of human activity: hence, the label ‘functionalist

64 Psychologism in early social sciencepsychology’. 18 Equally wrong is the atomism of associationist psychology. Themind does not consist of bits and pieces put together into complex ideas.Whatever goes on in the mind is a process, or a stream.Consciousness, then, does not appear to itself chopped up in bits. Suchwords as ‘chain’ or ‘train’ do not describe it fitly as it presents itself in thefirst instance. It is nothing jointed; it flows. A ‘river’ or a ‘stream’ are themetaphors by which it is most naturally described. In talking of it hereafter, letus call it the stream of thought, of consciousness, or of subjective life.( James, [1890] 1950: 239)Also important is James’s theory of the self, and on this point he is more inagreement with associationist psychology, especially that of Hume. Like thelatter, he rejects the idea of the self as some kind of substance. James’s analysisof the self is in terms of an ‘I’ and, especially, of a ‘me’, or empirical self. Thisempirical self is inclusive, indeed:In its widest possible sense … a man’s Self is the sum total of all that he cancall his, not only his body and his psychic powers, but his clothes and hishouse, his wife and children, his ancestors and friends, his reputation andworks, his lands and horses, and yacht and bank-account.( James, [1890] 1950: 291)It is evident from this list that the self is, to a large extent, a social self.According to James, ‘[a] man’s social Self is the recognition which he gets from hismates’ ( [1890] 1950: 293). Or, properly speaking, ‘a man has as many social selves asthere are individuals who recognize him and carry an image of him in their mind’ (p.294). James goes on to suggest that these individuals fall into different classes,which means that an individual ‘has as many different social selves as there aredistinct groups of persons about whose opinion he cares’ (p. 294). For each ofthese groups, an individual shows a different side of his/her self. What James ishinting at here is the idea of a social role, which would later become an importantpart of social psychology and sociology.William James influenced the symbolic interactionists, and also HenriBergson and Alfred Schutz, to mention only some of the more important. Healso influenced John Dewey, who adopted his functionalism with respect both tomind and society.Around the turn of the century, a sub-discipline called social psychology wasestablished as a field of its own. Of particular importance in the history of socialpsychology is the year 1908. In that year two books with that title were published– one written by the American sociologist E.A. Ross (1908), the other by theBritish psychologist William McDougall (1908). Of these works, the former waslargely based upon the ideas of Tarde, while the latter exhibits important influencesfrom him in the parts that deal with social psychology rather than generalpsychology. Both authors claim that social psychology provides a much-needed

Psychologism in early social science 65and necessary basis for the social sciences, and sociology in particular (Ross,1908: 2; McDougall [1908] 1924: 1–3, 16), but neither in Ross, nor inMcDougall, is it possible to find an explicit avowal of psychologism. 19The roots of social psychology are to be found both in Europe and the USA.In Europe, Völkerpsychologie was one source, even if it was closer to ethnology andsocial anthropology than to social psychology, as conceived of today. Two othersources were the sociologists Gabriel Tarde and Georg Simmel, who both wrotetreatises on social psychology (Tarde, 1898; Simmel, 1908c). In the USA, Jamesand Dewey had already been moving in the direction of a social psychology andDewey’s follower Charles Ellwood wrote a ‘Prolegomena to Social Psychology’(1899), which suggested an extremely holistic version of social psychology,derived from German Romanticism and from Hegel.In his later writings, however, Ellwood retreats from social holism into a moreindividualistic version of psychology and even into psychologism. In ThePsychology of Human Society (1925), he has reached the conclusion that sociologydiffers from psychology in subject matter, but its explanations are ‘psychological’(pp. 14ff). Unlike psychology, sociology is about groups, but since groups consistof individuals interacting with one another, and individuals consist of body andmind, or psyche, explanations of group life must be psychological. Ellwoodclaims that ‘psychological sociology’ is only a part of sociology, but his broadview of psychology makes it hard to see what would not be ‘psychological’.According to Ellwood (pp. 11, 461–6), the category of the ‘psychic’ includesculture. Ellwood’s later sociology, or social psychology is launched as a third waybetween the two untenable theories of society which had dominated earliersocial thought: the theory of the social contract and the theory of society as anorganism (pp. 453ff).Ellwood’s psychology of society is a synthetic work based upon the contributionsmade to social psychology in the first quarter of the nineteenth century andupon contributions to sociology, which are psychologistic, or, at least individualistic.Among social psychologists, Ellwood was influenced by William James, JohnDewey, James Baldwin, William McDougall, E.A. Ross and Floyd H. Allport.Among sociologists, he mentions Georg Simmel and Gabriel Tarde, but seems tobe most influenced by Franklin H. Giddings, Charles Horton Cooley, W.I.Thomas, L.T. Hobhouse and Robert M. MacIver. Of these, the sociologists L.T.Hobhouse and Robert M. MacIver, in particular, may be mentioned, since theywere important critics of collectivism in social science. Hobhouse wrote a wellknowncritique of The Metaphysical Theory of the State (Hobhouse, 1918), directedat the political philosophy of Hegel and, especially, at a book by his Britishfollower Bernard Bosanquet: The Philosophical Theory of the State (1899). MacIverwrote a number of articles (MacIver, 1911; 1913; 1914; 1915), where he criticisedboth organicism and the idea of society as some kind of super-person witha social mind or collective consciousness.From Tarde, Baldwin and Ross, Ellwood took the idea of imitation as a formof learning, 20 from McDougall, the notion of instincts, and from Giddings thetheory of socialisation. From Simmel and Ross he borrowed the idea that

66 Psychologism in early social sciencesociology is about groups and from the interactionists Cooley and Thomas hisconcept of culture, as something subjective, or psychic. From Hobhouse,MacIver and Allport, finally, he derived his critical stance towards organicismand the concept of social mind.Among the most influential works in the early history of the discipline isFloyd H. Allport’s Social Psychology from 1924. In this work, Allport appears as apsychological reductionist, and as a passionate critic of all forms of the ‘groupfallacy’ in social science.Social psychology, according to Allport is a science of the individual and, assuch, a part of the psychology of the individual. There is no such thing as apsychology of the group, as distinguished from the psychology of individuals.Theories about a social mind, group mind, collective consciousness, classconsciousness, general will and the like, are but various forms of the groupfallacy, as is also the idea of a social organism (Allport, 1924b: 4–10). ‘Behavior,consciousness, and organic life, belong strictly to individuals’, which is also thereason why, according to Allport, social psychology is the foundation science ofsociology (pp. 10f). ‘Since all behavior phenomena of groups are reducible tomechanisms of individual behavior in the social environment, the relation ofsocial psychology to the disciplines which treat of these higher aggregates is afundamental one’ (p. 382).Allport’s psychologism is even more explicit, and its identity with methodologicalindividualism more evident, in the article ‘The Group Fallacy in Relation toSocial Science’. As in Social Psychology, Allport argues that social phenomena mustbe explained by social psychology. The reason, once again, is that the cause ofsocial phenomena is invariably to be found in the behaviour of individuals.Social institutions exist in the attitudes and consciousness of individuals (Allport,1924a: 690ff). But psychologism is also justified as part of a more general reductionprogramme for scientific explanation. In the hierarchy of the sciences, eachscience describes phenomena at their own level, but they are only explained by adescription of phenomena at the level immediately below.Turning now to the sociologist, we find that the data which he describesreach the highest point of breadth and complexity. They embrace collectionsof individuals in organised societies, the products of suchorganisations, and the changes which they undergo. This is indeed a vastfield for descriptive analysis. Yet for explanation sociology is in its turndependent upon the descriptive formulas of the science just below it,namely psychology. Just as psychology has to seek its causation within theunits (reflex arcs) of which its material, individual behavior is composed; sosociology must find its explanatory principles in the units (individuals) ofwhich society is composed.(Allport, 1924a: 700)When we turn to the individual for causation, we only follow the rule of theother sciences in explaining the complex in terms of the simple, the whole in

Psychologism in early social science 67terms of its parts (p. 703). In the case of sociology, this means behaviour that isstimulated by the behaviour of others, that is social behaviour, as investigated bysocial psychology. More specifically, then, when we want to explain socialphenomena, as described by sociology, it is to social psychology that we shouldturn. ‘The work of sociology, therefore, would be to describe social aggregatesand social change in terms of the group, but to explain these phenomena interms of the social psychology of the individual’ (p. 703). 21Allport’s psychologism bears a striking resemblance to that of John StuartMill and Gabriel Tarde. Like the latter two, he argues that theories about socialwholes and their development only describes, whereas a causal explanation mustbe sought at the level below, that is in the laws of psychology and socialpsychology. As we shall see later on, this is a view shared by many methodologicalindividualists, who argue that any theory about social wholes, or aggregates,must be provided with individualistic microfoundations. The similarity goesfurther than a mere quest for microfoundations. Like some recent methodologicalindividualists (see pp. 187–89), Allport identifies the causal mechanism ofsocial phenomena in the behaviour of individuals. 22All theories which partake of the group fallacy have the unfortunate consequenceof diverting attention from the true locus of cause and effect,namely, the behavior mechanism of the individual. They place the groupprior to this mechanism in order of study, and substitute description ofsocial effects in place of true explanation.(Allport, 1924b: 9)Wilhelm DiltheyThe new science of psychology met with much resistance, from historians andsocial scientists, but, above all, from the philosophers. As I have already hintedat, the reason for this was that the philosophers had most to loose by the emergenceof psychology. While historians and social scientists risked only theautonomy of their disciplines, philosophers also risked loosing their jobs.Psychologists were appointed as philosophers. In addition, philosophy wasthreatened, not only by psychology, but by positivism and historicism, as well.There was no place for philosophy in Comte’s hierarchy of the sciences and aseeming implication of his law of the three stages is that it belongs to the metaphysicalstage and, thus, to history. The future belongs to the positive sciences.The effect of historicism was to turn philosophy into a kind of history of mind,or spirit, best exemplified by the philosophy of Hegel. It was in this situation thatsome philosophers turned to Kant as their saviour. Under the battle cry: ‘Back toKant’, they fought back against all enemies trying to obliterate philosophycompletely, or annex it to their own province (Willey, 1978). According to Kant,philosophy is a theory of knowledge, assuming the form of a critique of knowledge.This is an activity which cannot be reduced to any one of the scientificdisciplines, not even to all of them together. Neo-Kantians showed the strongest

68 Psychologism in early social scienceresistance against positivism, historicism and psychologism, but they were notalone. They were joined by logicians, such as Gottlob Frege, and by EdmundHusserl, the founder of phenomenology.The most common reaction, from philosophers, then, was to deny that philosophyis, in any way, dependent upon psychology, or any other discipline. Somephilosophers, however, admitted that philosophy does depend on psychology, butnot upon experimental psychology. The main example of this reaction wasWilhelm Dilthey, who was a pupil of one of the founders of Völkerpsychologie,Moritz Lazarus. Dilthey wanted to build a new foundation for the humansciences in the form of descriptive psychology, to be distinguished from theexplanative psychology of the professional psychologists.Dilthey’s first important attempt to lay a new foundation for the humansciences was in his Introduction to the Human Sciences (Einleitung in dieGeisteswissenschaften, 1883), a work that was not far behind Comte’s Cours dePhilosophie Positive and Mill’s System of Logic in terms of influence upon the humansciences. To some extent, it was also written as a response to the works of Comteand Mill. The main thesis of Dilthey is that there is a fundamental differencebetween the natural and the human sciences ([1883a] 1989: 60ff, 80f). While theformer deal with external reality, which can only be inferred from experience,the latter deal with inner experience itself. Because we have direct access to ourinner experience, the human sciences are epistemologically privileged, relative tothe natural sciences. One consequence of this is that also their methods differ. Inthe human sciences, the method of understanding (vertstehen) plays an importantrole (pp. 88, 158), even if this is not at all an important theme in Introduction to theHuman Sciences. 23But Dilthey was critical also of many aspects of historicism. He explicitlyrejected the notions of Volksseele and Volksgeist ([1883a] 1989: 83, 92), which wereused by his teacher Moritz Lazarus, and also the idea of a social organism (pp.83, 87, 121). There is no unity of consciousness in a people and the analogybetween a society and an organism is taken much too far by historicists and sociologistsalike. Dilthey, finally, denied unity also to history. There is no meaning orgoal of history and there are no laws of historical development. The attempts todetect such unilinear patterns by sociology and the philosophy of history aremisconceived (pp. 142ff).The rejection of the most holistic aspects of historicism and sociology, doesnot lead Dilthey to run to the other extreme and jump right into the camp ofradical individualism. ‘The fundamental error of the ancient natural-law schoolwas to isolate individuals and then to connect them mechanistically as themethod of constructing society’ ([1883a] 1989: 82). Dilthey wants to steer amiddle course between the extremes of individualism and holism, but it is mybelief, that he stays on the individualistic side of the divide, at least in his earlywritings. It is possible to understand his project as that of providing historicismwith a new individualistic foundation (see pp. 48–52).The main idea of Dilthey was that the human sciences must be provided witha new foundation in the form of psychology. ‘Analysis designates the life-unit,

Psychologism in early social science 69i.e., the psycho-physical individual, as the element from which society andhistory are formed; and the study of these life-units constitutes the most fundamentalgroup of the human sciences’ ([1883a] 1989: 80). As we have alreadyseen, however, Dilthey rejected the extreme individualism of natural law, whichtakes its point of departure in the isolated, or abstract, individual. It is part of hisfoundation of the human sciences that individuals interact:The units that act on one another in the marvellously complex totality ofhistory and society are individuals, psychophysical wholes, each of whichdiffers from every other, each of which constitutes a world, for the worldexists nowhere else but in the representations of such an individual.(Dilthey [1883a] 1989: 80f)There is, I believe, an obvious reference to Leibniz’s ‘monadology’ in thisquotation, and there is no doubt about the influence of Leibniz on Dilthey, but ifLeibniz’s monad has no windows, Dilthey’s individual has. ‘Thus the individualis found within it [society] in interaction with other elements’ ([1883a] 1989: 87).This interaction, moreover, is not ‘mechanical’ as in the case of neoclassicaleconomics, but social, and for Dilthey this means symbolic or at least meaningful.It would probably not be too wide off the mark to conceive of Dilthey as the firstsymbolic interactionist, but as we shall see later, he was also a progenitor ofphenomenology.A consequence of seeing society as meaningful interaction between individualsis that each individual can only be understood in the context of systems ofinteraction. ‘The individual is … an element in the interactions of society, apoint of intersection of the various systems of these interactions, reacting to theinfluences of that society with conscious intentions and actions’ (p. 89).According to Dilthey, then, the human sciences are concerned with the intentionalactions of human beings. He goes on to point out that intentions, ormotives, are different from other causes, but he does not deny that they arecauses. ‘The interaction which occurs here, namely, the motive which arises inconsciousness, differs from every other cause’ ([1883a] 1989: 89). On this pointhe takes a position similar to that of Max Weber.The interaction of individuals in history gives rise to three classes of objects,which the human sciences need to study: ‘the external organization of society; thecultural systems within it, and individual peoples. Of all these enduring entitiesthat of a people as a whole is the most complex and difficult’ (Dilthey [1883a]1989: 92f). The sciences devoted to the study of these objects have to introduce,what Dilthey calls, ‘second-order concepts’ and the theories arrived at are, consequently,‘second-order theories’, relative to individual psychology (pp. 92, 96). Bysecond-order concepts, Dilthey understands all concepts developed within thehuman sciences, such as ‘work’, ‘value’, ‘price’, ‘status’, ‘role’, etc. There is noquestion, however, but that the second-order concepts and theories, used bysecond-order human sciences, such as philology, economics and ethnology, makesense only to the extent that they refer to psychic facts about the lived experience

70 Psychologism in early social scienceof individuals (pp. 114ff). All human studies become scientific ‘by establishingthe reference of their concepts and propositions to psychological andpsychophysical facts’ (p. 129). When this is achieved, ‘the sciences of the externalorganization of humanity would be based upon concepts of psychic andpsychophysical facts and upon propositions about them which correspond to theconcepts and propositions on which the sciences of the cultural systems arebased’ (p. 118).I hope it is clear by now that Dilthey saw the basis of the human sciences inpsychology. At the beginning of my presentation of Dilthey, I also mentionedthat this psychology is called descriptive, as distinguished from the explanativepsychology of Mill and Wundt. What the former amounts to is not made clear,except by what is indicated by the label ‘descriptive’. The latter, however, is characterisedby its attempt to explain the whole of human reality by means of a fewsimple assumptions, or hypotheses, about human nature ([1883a] 1989: 84).Dilthey’s idea of a ‘descriptive’ psychology was a bit more developed in IdeasConcerning a Descriptive and Analytic Psychology (1894). Once again, it is contrastedwith the explanatory psychology of associationism and various types of materialism,thatseeks to explain the constitution of the psychic life [Seelenleben] with the helpof its components, energies and law, just as physics and chemistry explainthose of the corporeal world … By explanatory science is to be understoodevery subordination of a domain of experience to a system of causality[Kausalzusammenhang] by means of a limited number of well-determinedelements (i.e., the components of the system).(Dilthey [1894] 1977: 23)A first characteristic of explanatory psychology, then, is that, like the naturalsciences, it seeks to give a causal explanation of psychic life with the help of acombination of hypothetical laws. ‘In the human studies, to the contrary, thenexus of psychic life constitutes originally a primitive and fundamental datum.We explain nature, we understand psychic life’ ([1894] 1977: 27). A second characteristicof explanatory psychology is that it is ‘constructive’, or ‘synthetic’. Withanother term, equally apt, and often used to describe British associationistpsychology, we might call it ‘atomistic’. In sharp contrast to explanatorypsychology, descriptive psychology is holistic. It is holistic because experience,immediately given, lived experience is holistic. ‘[L]ife exists everywhere only as anexus or coherent whole’ (p. 28).Psychic life-process is originally and above all, from its most elementaryforms to the highest, a unity. Psychic life does not grow together from parts;it is not composed of elements; it is not a composite nor is it a result of thecollaboration of sensory or affective atoms: it is originally and always acomprehensive unity.(Dilthey [1894] 1977: 92)

Psychologism in early social science 71That psychic life is a whole does not imply, however, that social life or cultural lifeis also a whole, and it is not at all unusual that methodological individualists arepsychological holists in the manner of Dilthey (cf. Rickman, 1979: 9–11). Thereason for this is, of course, that psychic life – even if a whole – is contained inindividual human beings. There are purposive systems of social life, such aseconomic life, law, art and religion, which may be studied apart. ‘But theensemble of such a system is nothing other than the psychic nexus, of the peoplewho cooperate in it. It is therefore, in the end, psychological’ (Dilthey [1894]1977: 40). For Dilthey, the human sciences start and end in the psychic life ofindividuals. This is, in the first instance, an epistemological thesis. But it is also anontological thesis. While the natural sciences point towards an external reality,existing independently of the lived experience of individuals, the human sciencesdo not. Society and culture exist nowhere, except in the immediately given, livedexperience of individual human beings. This is subjective idealism, at least withregard to society and culture, and to that extent congruent with phenomenology.I mentioned above that Dilthey saw explanatory psychology as synthetic.Consequently, descriptive psychology is analytic. If psychic life is given to experienceas a whole, it is the task of descriptive psychology to analyse this whole inits various parts ([1894] 1977: 51ff). ‘In the service of this descriptive activitystand the logical operations of comparing, distinguishing, establishing degrees,separating degrees, separating and associating, abstracting, connecting parts intoa whole, deducing similar relationships from particular cases, analyzing individualevents and classifying’ (p. 56). It is possible to see, in these tasks ofdescriptive psychology, a point of contact with phenomenology. 24It is a common observation among students of Dilthey that his view ofpsychology and its relation to the human sciences changed towards the end ofhis life. The direction of this change was away from subjectivism and individualismtowards a more objectivistic and holistic view; from psychologicalexperience to cultural hermeneutics. 25A clear sign of change can be found in the article on ‘The Development ofHermeneutics’ (1900). The source of knowledge about culture is no longer thelived experience of individuals, but the expressions of others and, especially ofwritten records. Hermeneutics is defined as ‘the methodology of the interpretationof written records’ ([1900] 1976: 249). Such interpretation invariably takesthe form of a circle:The whole of a work must be understood from individual words and theircombination, but full understanding of an individual part presupposesunderstanding of the whole. This circle is repeated in the relation of anindividual work to the mentality and development of its author, and itrecurs again in the relation of such an individual work to its literary genre.(Dilthey, [1900] 1976: 259)What is only hinted at in the article on hermeneutics, however, is more fullydeveloped in ‘The Understanding of Other Persons and Their Life-Expressions’

72 Psychologism in early social science(c. 1910). Dilthey here makes a distinction between two classes of lifeexpressions:(1) concepts, judgements and thought structures and (2) actions([1927] 1986: 153). He also makes a distinction between elementary and higherforms of understanding of life-expressions. By elementary understanding,Dilthey means understanding of ‘single life-expressions’, for instance an assertionor a facial expression. A third example would be ‘elementary acts of whichcontinuous activities are composed, such as picking up an object, letting thehammer drop, cutting wood with a saw’ (p. 154). 26 The higher form of understandingalso goes in two directions: either it seeks the relation between productand producer, or it is oriented to the relation between expression and what isexpressed, which is not at all reducible to what the producer intended to express.When understanding is oriented to the producer of a life-expression, understandingtakes the form or ‘re-creation’, or ‘re-living’ (Nacherlebnis). This is not,however, a psychological form of understanding. According to the later Dilthey,all understanding rests on the objective mind. ‘It is the medium in which theunderstanding of other persons and their life-expressions takes place’ (p. 155).A sentence is intelligible because a language, the meaning of words and ofinflections, as well as the significance of syntactical arrangements iscommon to a community. The fixed order of behaviour within a culturemakes it possible for greetings or bows to signify, by their nuances, a certainmental attitude to other people and to be understood as doing so. Indifferent countries the crafts developed particular procedures and particularinstruments for special purposes; when, therefore, the craftsman uses ahammer or saw, his purpose is intelligible to us. In this sphere the relationbetween life-expression and mental content is always fixed by a commonorder. This explains why this relation is present in the apprehension of anindividual expression and why – without conscious inference based on therelation between expression and what is expressed – both parts of theprocess are welded into a unity in the understanding.(Dilthey [1927] 1986: p. 156)Dilthey’s commitment to the idea of an objective mind is clearly a move awayfrom psychologism, at least in epistemology and methodology, but is it also abreak with ontological and methodological individualism? H.P. Rickman, one ofthe authorities on Dilthey, argues that he made no metaphysical claims about theentities of the objective mind (1962: 42; 1979: 116f). They are human creationsand have an existence independent of individuals.I am not at all convinced that this is a correct interpretation of Dilthey, but Iwill not argue that point here. I will call attention, instead, to a decisive differencebetween Dilthey and the methodological individualist Max Weber. Thelatter made a distinction, similar to that of Dilthey between elementary andhigher understanding, but he called them direct and explanatory understanding.Weber, even used exactly the same example of a woodcutter, as did Dilthey,before him. For Weber, however, explanatory understanding, does not include

Psychologism in early social science 73reference to culture and social structure, which it certainly did to Dilthey. ForWeber, explanatory understanding is exclusively understanding in terms of themotive of the producer of the product or, more generally, of the acting individual(see pp. 100f ). If the later Dilthey was a methodological individualist,therefore, he was not a methodological individualist in the sense of Weber andprobably not at all.Georg SimmelIt should be clear by now that in Germany it was history that dominated thehuman sciences. Psychology was also established as an academic discipline, firstas an explanatory natural science, but in the second half of the nineteenthcentury also as a descriptive science of mind and as a science of culture(Völkerpsychologie). It might be added that philosophy was still strong and encyclopaedic.In this situation, there is little wonder that sociology met with muchresistance – if less than psychology – before it was accepted as an academic discipline,and that the first German sociologists had somehow to adapt to thissituation. Of the two most famous German sociologists, Max Weber was also ahistorian, and Georg Simmel a philosopher, making important contributions tothe philosophy of history. Both of them rejected French and British positivismand created their own versions of sociology, better suited to German conditions.In the case of Weber, it was an interpretative (verstehende) sociology, in the case ofSimmel a formal sociology. The impression created is that Weber owed most toDilthey, whereas Simmel owed more to neo-Kantianism. The truth, I believe, isthat it was the other way around, but certainly both Weber and Simmel wereinfluenced by both Dilthey and neo-Kantianism. 27 In the case of Weber,however, there was a third important influence, from the Austrian School ofEconomics. It was this influence, in particular, which turned Weber into aprogrammatic methodological individualist. I have found it convenient, therefore,to present Weber in the next chapter, which is devoted to Austrianmethodological individualism. Here I will concentrate on Simmel, who was akind of psychologicist, at least for a start.Like Dilthey, Simmel was a pupil of Moritz Lazarus and, therefore, certainlyfamiliar with his Völkerpsychologie. It is also pretty certain that Simmel owed someimportant ideas to the latter, especially in his early writings, which include quitea lot on the theme of psychology (Frisby, 1992: ch. 2). 28 Simmel’s debt to Diltheywas even larger and it is possible to detect many interesting similarities betweenthem. The idea that the individual is situated at the point of intersectionbetween many social circles, or groups, is, perhaps, the most obvious one. BothDilthey and Simmel made important contributions to the philosophy of historyand made much to clarify the problem of interpretation (verstehen). 29 It has evenbeen suggested that there is a parallel in the development of the thinkers: frompsychologism to the idea of an objective mind, or culture (Oakes, 1980: 57–60).One difference between them is that Simmel tried to lay the foundation of sociology,a discipline that Dilthey rejected. But, then, it should be mentioned that

74 Psychologism in early social scienceSimmel’s sociology was not at all like that of Comte and much more in line withDilthey’s view of the human sciences.Simmel is sometimes classified as a methodological individualist (Infantino,1998: 95–9, 106–13, 131–34) and even as a psychologicist (Moscovici [1988]1993: 236). I do not deny that there are some good reasons for doing so, butnevertheless I believe that it is wrong. Already in an early version of his article‘The Problem of Sociology’ (1895) he wrote: ‘the overthrow of the individualistpoint of view may be considered the most important and fruitful step whichhistorical science and the moral sciences (Geisteswissenschaften) generally havemade in our time’ (p. 52). Of course this is not proof enough, since Simmel mayhave had some extreme version of individualism in mind, but it is an indication.In my view, Simmel attempted to steer a middle course in between the extremesof atomistic individualism and romanticist collectivism, but he ended up as moreof a holistic structuralist, than as a methodological individualist.Simmel’s sociology is called ‘formal sociology’ because he was interested inthe forms of social life, as distinguished from its content. The distinction betweenform and content he borrowed from Kant. With forms of social life, Simmelintended things like superiority, subordination, hierarchy, competition, opposition,division of labour, representation, etc. ([1908] 1909: 299). There are, Ibelieve, two possible reasons for conceiving of Simmel as a psychologisticmethodological individualist.The first obvious reason is that he maintained that the content, or stuff, ofsociety is mental, or psychic (Simmel, 1895: 54; [1908] 1950: 24). The other,equally obvious reason is that he maintained that the form of society takes theshape of forms of interaction, or of social relationships (Simmel, 1895: 54;[1908] 1950: 10, 21, 40). Society, according to Simmel, is the sum-total of ‘intersubjectiverelations which bring it to pass that individuals become societies’(1895: 61). Simmel’s concept of ‘society’, then, is a synthesis of two views ofsociety:1 The intersubjectivist theory of society as being made up solely of ideas inthe minds of individuals.2 The interactionist theory of society, which says that society is made up ofthe sum total of interactions between individuals.If this were the end of it, Simmel would clearly be a psychologistic methodologicalindividualist, but it is not. First of all, it does not follow from the fact that thecontent of social phenomena is psychic, if that is a fact, that the explanation ofsocial phenomena must be psychological.In The Problems of the Philosophy of History ([1892] 1977: 113–17) Simmeldiscusses the role of psychological laws in history. He takes his point of departurein the thesis that the ultimate constituents of society are human beings. Onlyhuman beings are real entities, with causal efficacy, whereas groups exist only ina secondary sense. Does it follow from this that the laws of history must be lawsof individual psychology? Simmel does not want to draw this conclusion: first of

Psychologism in early social science 75all because the minds of individuals are extremely complex entities, which mightbe resolved into simpler elements. ‘Simplicity and complexity are relativeconcepts’. Simmel, himself seems to take an agnostic attitude to ontology.‘Suppose there are no absolutely basic units or entities. Suppose that the definitionof the unit of investigation is simply a consequence of the special criteria ofeach science’. If this is so, it is ‘certainly not prima facie self-evident’ that the‘composite phenomena of historical life are nomologically comprehensible onlyif they can be deduced from the established laws of individual psychology … thetheoretical interests of history could permit the societal group to function as itsbasic “unit” of inquiry’ (p. 114).Simmel agrees with the individualists that ‘at least in one sense’ the ultimatesource of knowledge about history is ‘the individual mind or personal psyche’. Itis therefore, also the basic unit of historical inquiry. ‘The mind is a basic unit orentity for this reason: what we call unity or uniformity is possible only as a consequenceof the inner, first-person experience of the self ’ ([1892] 1977: 115). Buthe adds something potentially important: ‘It is not psychic unity or the mentalentity as such that constitutes the historical event; on the contrary, it is thecontents of mind’ (p. 117). This conclusion points in the direction of Simmel’sHegelian idea of objective spirit, or in modern terminology, of culture.The development of the social freedom of the individual is correlated withthe formation of an objective spirit, a wealth of superpersonal culturalproducts, scientific, aesthetic, and technological. These propositions andothers like them may be regarded as precursors of proto-forms of structuresthat will eventually be the objects of exact nomological knowledge.(Simmel [1892] 1977: 145f)Simmel returned the theory of objective spirit and objective culture in hisPhilosophy of Money ([1900] 1978: 446ff), where his break with psychologistic individualismis fairly evident. Products of the objective mind (cultural products,such as buildings, machines, books, works of art, etc.) have an objective existenceindependent of each particular individual and exert a profound influence onindividual human beings. Interaction taking place in this created cultural environmentcrystallises into more or less permanent social structures (pp. 170ff).These facts cast serious doubts on Infantino’s attempt to find convergencesbetween Simmel and Carl Menger. Not that it is difficult to find such convergences,but the divergences seem, to me, more significant. Take their respectivetheories of money. For Carl Menger, as we shall see in the next chapter, the allimportanttask is to explain the origin of money. For, Simmel, on the contrary, themain concern is to analyse the effects of money (pp. 228ff).In ‘The Problem of Sociology’ ([1908] 1909), Simmel admits that his discussionof this science ‘appear[s] to be nothing but chapters of psychology, or atmost of social psychology’ (p. 313), but adds ‘that the scientific treatment ofpsychical facts still by no means needs to be psychology’ (p. 314). I interpret thesepassages as indicating that Simmel was an ontological individualist, but not a

76 Psychologism in early social sciencemethodological individualist. At least not a psychologistic methodologicalindividualist.But there are some clear indications that, in the end, Simmel gave up even hisontological individualism. Forms of interaction between individuals ‘crystallise’into social structures, which take on a life of their own.More specifically, the interactions we have in mind when we talk about‘society’ are crystallized as definable consistent structures such as the stateand the family, the guild and the church, social classes and organizationsbased on common interests … The large systems and super-individual organizationsthat customarily come to mind when we think of society, arenothing but immediate interactions that occur among men constantly, everyminute, but that have become crystallized as permanent fields, asautonomous phenomena. As they crystallize, they attain their own existenceand their own laws, and may even confront or oppose spontaneous interactionitself.(Simmel [1908] 1950: 9f)In addition to individual human beings, then, there also exist social structures,which are autonomous relative to individuals. Thus, according to Simmel,society takes on a life of its own: ‘“Society” develops its own vehicles and organsby whose claims and commands the individual is confronted as by an alien party’([1917] 1950: 58). The result is a kind of split between individual and society,which is inconceivable to an ontological individualist.A society … is a structure which consists of beings who stand inside andoutside of it at the same time. This fact forms the basis for one of the mostimportant sociological phenomena, namely, that between a society and itscomponent individuals a relation may exist as between two parties.(Simmel [1908] 1971: 14f)The conclusion of Simmel’s investigation is that ‘[s]ociety is a structurecomposed of unequal elements’ ([1908] 1971: 18). He offers the phenomenon ofbureaucracy as a small-scale analogue of society. Typical for a bureaucracy isthat it consists of a structure of positions with a predetermined function assignedto each position, irrespective of the particular person occupying that position.Society ‘is a system of elements each of which occupies an individual place, aco-ordination of functions and function-centers which have objective existence’(p. 20).There is much in Simmel that reminds us of Marx and Durkheim, but whichhas no place in an individualist theory, as represented by symbolic interactionism,phenomenology, existentialism and ethnomethodology (see chapter 5).There are certain similarities between Simmel’s notion of sociation(Vergesellschaftung) and Anthony Gidden’s theory of structuration, but Simmel’sdualistic conception of individual and society, make him clearly less individual-

Psychologism in early social science 77istic than Giddens. Because of this dualism, Simmel is closer to the later Sartre,Berger and Luckmann and Pierre Bourdieu (see pp. 147f, 161ff, 186f ), perhapsalso to Roy Bhaskar.To conclude: Georg Simmel was not a psychologistic methodological individualistand probably not even an ontological individualist. It is still possible toclassify him as a less radical methodological individualist. Perhaps, it is possibleto conceive of him as a structural individualist (see chapter 10).Psychologism on trialDilthey’s move away from psychologism was, at least in part, a reaction to objectionsraised against this doctrine by the neo-Kantian philosophers WilhelmWindelband and Heinrich Rickert, and by the phenomenological philosopherEdmund Husserl. The most destructive critique, however, was directed atDilthey’s particular form of descriptive psychology, by the experimental psychologistHermann Ebbinghaus. Georg Simmel was influenced both by Dilthey andneo-Kantianism, but was not, himself, an object of much critique. Most critiquewas directed at experimental psychology, but this is not my business here. In thelast section of this chapter, I will focus on some reactions to Dilthey, and also onsome positive contributions to psychologism in social science.At the University of Strasbourg in 1894 in his famous Rectorial Address on‘History and Natural Science’, the leader of the Southwest School of neo-Kantian philosophy, Wilhelm Windelband, took issue with Dilthey’s way ofmaking a distinction between the natural and the human sciences ([1894] 1980:173). 30 The main problem with Dilthey’s dichotomy seems to be that psychologybelongs to both types of science (p. 174). Instead of making this distinction onthe basis of a substantive dichotomy of nature and mind, we should use an epistemologicaldistinction, in terms of cognitive interest. From this point of view,the empirical sciences can be divided into those that seek general laws and thosewhich are interested in specific historical facts. In the well-known terminologysuggested by Windelband, ‘scientific thought is nomothetic in the former case andideographic in the latter case’. In terms of this dichotomy, ‘psychology falls unambiguouslywithin the domain of the natural sciences’ (p. 174).With Windelband’s distinction between history and natural science, itbecomes necessary to reject also the positivist attempt to turn history into anatural science, seeking laws of historical development. Against this view, ‘it isnecessary to insist upon the following: every interest and judgement, everyascription of human value is based upon the singular and unique’ (p. 182). Therole of values was to become even more prominent in the constitution of the‘cultural sciences’ by Windelband’s pupil Heinrich Rickert.Rickert’s much more ambitious attempt to establish a distinction betweenhistory and natural science can be found in his most important work, The Limitsof Concept Formation in the Natural Science (Die Grenzen der naturwissenschaftlichenBegriffsbildung, 1902) and in the shorter Science and History (also 1902). 31 LikeWindelband, Rickert was much concerned to combat various reductionisms and,

78 Psychologism in early social sciencein the first instance, those that threaten philosophy. Thus, he rejected historicismand positivism in philosophy. Psychologism was also rejected, because it is anatural science and psychologism, therefore, a form of positivism. ‘In all of itsessential aspects, however, ours is a logical, or a methodological and epistemological,investigation’ ([1902] 1986: 19). Rickert’s main task, however, was to savehistory from being reduced to natural science, in the manner of Comte’s positivism(pp. 23ff). Hence, the project of identifying the limits to natural science,once again including psychology.Rickert’s distinction between history and natural science is similar to that ofWindelband, but much more elaborated and refined. History and naturalscience are based on different cognitive interests: ‘Empirical reality becomesnature when we conceive of it with reference to the general. It becomes historywhen we conceive it with reference to the distinctive and the individual’ (Rickert[1902] 1986: 54). For Windelband this distinction was absolute. Rickert,however, saw the difference as relative and preferred to talk about the method ofhistory as individualising and that of natural science as generalising ([1902] 1986:xii).Rickert’s point of departure was Kantian: there is an unbridgeable gapbetween concept and reality and the latter is inaccessible to our knowledge.Every form of epistemological realism, and especially the picture theory ofmeaning is therefore excluded from the outset. Reality is reality as perceived, orexperienced, and as constructed by means of concepts. ‘Reality itself, the infinitemanifold of which scorns every conception, can best be called “irrational,” andeven this designation could be applied to it only on the grounds that it resistsevery conception’ ([1902] 1986: 52). According to Rickert, then, empirical realityis indefinite and, therefore, practically infinite. It is extensively infinite, in the sensethat there are an infinite number of possible objects of knowledge and it is intensivelyinfinite, in the sense that there are infinitely many aspects of every singleobject of knowledge. In order to cognitively master this infinite manifold, we useabstraction. The procedure of the natural sciences is to go searching for similarities,while abstracting from the differences of objects, including their position inspace and time. Positivism assumes that this is the only scientific procedure. But,according to Rickert, there are limits to natural scientific concept formation,since it prevents us from seeing what is unique and individual about objects ofknowledge. But this is exactly the cognitive interest of the historical sciences.They are interested in individual events, languages, works of art, cultures, etc, inspace and time. The real problem that Rickert set out to solve was that ofconcept formation in history. If empirical reality is an infinite manifold and themethod individualising, what kind of abstraction is used in history?Obviously, history is only interested in some of the (extensively infinite) manifoldof individuals there are in empirical reality, and only in some aspects ofthese (intensively infinite) individuals. Rickert calls them ‘historical individuals’and they comprise, not only human individuals, or personalities, but all sorts ofhistorical particulars, such as the Renaissance, the French revolution and WorldWar I (Rickert [1902] 1986: 78–98). Rickert’s solution to this problem is the so-

Psychologism in early social science 79called value relevance, or value relation (Wertbeziehung). History concentrates onthose individuals and those aspects of individuals to which a value, or meaning isattached by people in general, that is individuals who are considered interestingand/or important in terms of cultural values.The thesis that the knowledge interest of history is directed at ‘historical individuals’,then, does not imply a focus on human beings and Rickert denies thathistory is ‘individualistic’ in this sense. First of all, he rejects the ‘rationalisticteleological’historiography in which ‘historical events are shown to be intendedconsequences produced by rational and purposive beings’ ([1902] 1986: 103).Those who see the setting of a conscious purpose and the conduct thatfollows from it as the motive force of all historical movements, and in consequenceregard purpose as the explanatory principle of history, must not onlysee individual personalities as the chief objects of history – because thesetting of conscious purpose can be confirmed in them alone; as the resultof this, they must also take the view that single individuals make history, sothat everything becomes a product of individual intention. But we are farfrom advocating an individualistic conception of history in this sense.(Rickert [1902] 1986: 104)Second, Rickert denies that history is ‘individualistic’, even if by individualswe understand historical individuals. The reason is that no historical individual,whether a person or some larger individual, can be understood in isolation fromthe historical nexus, or whole, of which it is a part (pp. 107ff).Like Dilthey, then, Rickert argued that historical understanding is, in a sense,‘holistic’. He also agreed with Dilthey (and Hegel) that understanding is ofobjective mind, or spirit ([1902] 1986: 138ff). Unlike Dilthey, however, Rickerttook great care to point out that objective mind is ideal and nonreal. Cultureconsists both of that part of reality to which meaning is attached, and thisnonreal meaning, or spirit, itself. According to Rickert ‘Dilthey confuses thenonreal meaningful content of culture that is situated in the realities of historywith the real psychic existence that actually occurs in the mental life of singleindividuals’ (p. 146). This accusation would have been entirely justified, hadRickert intended Dilthey’s early writings, but his target is actually the theory ofthe human sciences that Dilthey developed in his last writings. This does notmake sense to me, but the whole issue is too complex to be treated here.It seems extremely important for Rickert to distinguish culture from thepsychic, or mental. A first reason for this is, no doubt, that his distinctionbetween the natural and the cultural sciences must not be reducible to Dilthey’sdistinction between natural and human sciences. A second, and related, reason, Ipresume, is that his logical/epistemological distinction must not be reducible to amaterial distinction between nature and mind, or spirit. Even so, there is a sensein which Rickert falls back upon an ontology of individuals and their minds.First, his view of objective mind as nonreal, makes it possible to avoidconceptual realism and maintain that only human individuals exist ([1902] 1986:

80 Psychologism in early social science150f ). The fact that ‘[m]eaningful wholes can be understood only as unities ortotalities’ (p. 159), does not imply that there are real wholes, irreducible to theirparts.The object of understanding as nonreal meaning configuration alwaysremains a whole or a unity. Realities, on the other hand, are decomposedinto their parts for the purpose of explanation. Or, in explanation, the pathleads from the parts to the whole; in understanding, it proceeds in the oppositedirection, from the whole to the parts.(Rickert [1902] 1986: 159)Since, for Rickert, the parts of real ‘wholes’ are individuals, this view is in agreementwith the methodological individualism of Weber’s interpretive sociology(see pp. 101f ), which admits that holistic understanding is necessary as a preparationbefore the real sociological investigation of motives begins. Unlike Weber,however, Rickert never allowed subjective meaning to play an important role inhistorical methodology. While agreeing in substance, their emphasis is entirelydifferent. Rickert admits that historical reality is meaningful because individualsattach meaning to it and values are valid because accepted as valid by individuals,but he never attaches much importance to the subjective meaning andcausal explanation. For Weber, on the other hand, the subjective meaning ofhuman action and causal explanation in terms of motives are the core of interpretivesociology.To conclude: Rickert was a metaphysical individualist, since he conceived ofobjective mind as nonreal. He was not a methodological individualist, if this isunderstood as a principle of understanding – which, according to him, is themain task of the cultural sciences – but to the extent that the cultural sciencesengage in explanation, they are individualistic. The reason for this differencemay be that Rickert’s methodology was intended for history, whereas Weber wasengaged in the attempt to create an interpretive sociology.The most famous, and most influential, critique of psychologism came fromEdmund Husserl (1859–1938), the founder of phenomenology. It can be foundin the first volume – the so-called ‘Prolegomena’ – of his Logical Investigations(1900). Space does not admit of any detailed presentation of Husserl’s argumentand it is not necessary for my purposes, since his argument was directed atpsychologism in philosophy. 32 Simply put, Husserl maintained that psychologyhas no bearing on normative disciplines such as mathematics and ethics, butleads to subjectivism, scepticism and relativism. In the second volume of (theGerman edition of) Logical Investigations, Husserl set out his first version ofphenomenology, which, paradoxically, owed much to the descriptive psychologyof his teacher Franz Brentano. ‘The Logical Investigations … are a full developmentof Brentano’s suggestions, as is a matter of course, since I was animmediate student of Brentano’ (Husserl, 1977: 24). The most importantborrowing is the idea that mental phenomena are intentional. It also had some-

Psychologism in early social science 81thing in common with Dilthey’s epistemology and descriptive psychology, at leastthat is what Dilthey, himself, thought. 33The mature version of phenomenology, however, took shape in the firstdecade of the twentieth century. A first groping attempt to lay a new foundationfor this philosophy was made by Husserl in a series of lectures given in 1907,later published as The Idea of Phenomenology (1950). It is not easy to understandfrom these lectures what phenomenology really is, but it is clear that it has littleto do with psychology, including descriptive psychology ([1950] 1973: 5, 33ff).Phenomenology is a priori (p. 41), but not in the sense of mathematics.‘Phenomenology proceeds by “seeing,” clarifying, and determining meaning, and by distinguishingmeanings’ (p. 46). Husserl’s method is that of Descartes. Like the latter, heseeks absolute certainty and finds it in cognition.A few years later, Husserl made a new attack on psychologism in ‘Philosophyas a Rigorous Science’ (1911), but it is clear that he conceives of psychology as anatural science. The main fault of this naturalistic psychology is ‘to set aside anydirect and pure analysis of consciousness’ (1981: 174). This time Husserl’scritique is extended to include also historicism (pp. 185ff), which ‘if consistentlycarried through, carries over into extreme sceptical subjectivism’ (p. 186). Thetarget is Wilhelm Dilthey’s ‘Weltanschauung philosophy’, which, according toHusserl, is bound to lead to relativism. It is doubtful whether this is really thecase, but what is more interesting is that Husserl makes no recognition ofDilthey’s version of a descriptive and analytic psychology, despite his knowledgeabout it. 34The first comprehensive expression of Husserl’s mature version ofphenomenology is to be found in Ideas Pertaining to a Pure Phenomenology and to aPhenomenological Philosophy (First Book 1913). This is not the place to assume thedemanding task of doing justice to Husserl’s pure phenomenology. I will do nomore than indicate some of the ideas, which may help us better to understandphenomenological sociology (see pp. 137–44). In our everyday life, we take theexistence of an external world of physical things and other human individualsfor granted. I suppose this is roughly what is understood by ‘common sense’ or‘naive realism’. Husserl calls it the ‘natural attitude’, but, according to him, itcharacterises not only common sense, but also the empirical, or ‘experiential’sciences. The distinguishing mark of the natural attitude is that it posits a worldof external, or ‘transcendent’, objects in space and time. In addition to cognitionof matters of fact, however, there is something Husserl calls ‘eidetic seeing’, orintuition of essences (Wesensschau), which is about a new sort of ideal universalobjects (Eidos). The idea of a straight line in geometrics would be one example.The sciences about these objects are called the eidetic sciences and they includelogic and mathematics, but also the ‘pure theories of time, space, motion, and soforth’ (p. 16).Pure phenomenology, as created by Husserl, is an eidetic science aboutessences – if different from mathematics and logic – and, like all such sciences, itis also a science about phenomena, or objects of intention, but not about a worldof things beyond them, or things-in-themselves. In order to be able to attend

82 Psychologism in early social scienceonly to phenomena, Husserl maintains that we must free ourselves from thenatural attitude of presupposing things-in-themselves; we must put the objectiveworld ‘within brackets’. This is the method of phenomenological reduction,which is similar to Descartes’s methodical doubt, except that it excludes alsodoubt. We simply disregard everything objective. Husserl calls it the phenomenologicalepoché, and it is a first preparatory step in order to reach the world of purephenomena.The next step towards pure, transcendental, phenomenology is the intuitionof essences, which I have already mentioned. This is called the eidetic reduction,because it leads from particular phenomena, such as this house, to universalphenomena and the necessary relations between them, but it is not yet the transcendentalreduction which leads to the transcendental ego of purephenomenology.In Ideas Husserl juxtaposed phenomenology to psychology and denied anycommon ground between them. Of course, they have something in common:both are theories of phenomena of consciousness; he even admitted that there isan eidetic psychology, but pure phenomenology is distinct from psychology andthere is no phenomenological psychology. In his ‘Inaugural Lecture at Freiburgin Breslau’ in 1917, he told his audience that ‘pure phenomenology is too separatedsharply from psychology at large and, specifically, from the descriptivepsychology of the phenomena of consciousness’ (Husserl 1981: 14). 35 In themiddle of the 1920s, however, there is a new revision of phenomenology.An interesting document from this period is Husserl’s lectures on phenomenologicalpsychology, given in Paris in 1925 (Husserl, 1977). In these lecturesHusserl begins by paying tribute to Dilthey and Brentano. After hailing Diltheyas a genius, he goes on to state his main critique, which is that Dilthey was moreof an intuitive than analytical thinker and therefore failed to lay the foundationof a new psychology. His main complaint, however, is that Dilthey was too muchof a historicist, seeking the unique and culturally relative. A phenomenologicalpsychology, on the contrary, seeks essences and universal laws. As firmly as Mill,he believes in a universal human nature, which it is the task of phenomenologyto lay bare, and which is at the bottom of all social life. All concepts of thehuman sciences, therefore, rest on the foundation of phenomenologicalpsychology.Having originated in the living nexus of the human psyche, they can beunderstood only on this same basis. Only because there is uniformity andregularity in psychic life can they arise as forces overreaching the individualand making the same order possible for all individuals. Another significantpoint is their combination with the theory of knowledge.(Husserl, 1977: 10)In the famous 1927 article on ‘Phenomenology’ in Encyclopaedia Britannica(repr. in Husserl, 1981), Husserl divides phenomenology into ‘phenomenologicalpsychology’ and ‘transcendental phenomenology’. What is somewhat confusing,

Psychologism in early social science 83however, is that phenomenological psychology seems to have almost everythingin common with pure phenomenology, as construed in Ideas, except that it operateswithin the natural attitude (p. 29). Like pure phenomenology, in Ideas, it isthe result of both a phenomenological and an eidetic reduction (pp. 24ff). Purephenomenology is now pushed back into the realm of the transcendental egoand is consequently called ‘transcendental phenomenology’. I will not even try toaccount for the ontology of possibilities that can be dimly perceived. Husserlhimself is perfectly aware that to grasp the distinction between transcendentaland phenomenological psychology is difficult, indeed. In his preface to theEnglish edition of Ideas, he admits that it ‘is a problem that as a rule brings greatdifficulties with it … It has led to misunderstandings, to which even thinkers whosubscribe to the phenomenological line of thought are subject’ (1981: 44).Despite these acknowledged problems, however, Husserl is confident that transcendentalphenomenology is the universal science and the foundation of allother sciences and of all knowledge. 36It should be obvious by now that phenomenology is a form of radical subjectivism.What is perhaps less obvious, as yet, is that it is a form ofintersubjectivism. ‘The reductive method is transferred from self-experience tothe experience of others’. The application of the method of phenomenologicalreduction to ‘community experience’, or ‘community mental life’ is called ‘intersubjectivereduction’ (Husserl, 1981: 25). ‘[T]the reduced intersubjectivity, inpure form and concretely grasped is a community of pure “persons” acting inthe intersubjective realm of the pure life of consciousness’ (p. 25). Althoughthese quotations referred to phenomenological psychology, transcendentalphenomenology also is intersubjective.The psychic subjectivity, the concretely grasped ‘I’ and ‘we’ of ordinaryconversation is experienced in its pure psychic ownness through the methodof phenomenological-psychological reduction. Modified into eidetic form itprovides the ground for pure phenomenological psychology. Transcendentalsubjectivity, which is inquired into in the transcendental problem, and whichsubjectivity is presupposed in it as an existing basis is none other than again‘I’ myself and ‘we’ ourselves; not, however, as found in the natural attitudeof everyday or of positive science; i.e., apperceived as components of theobjectively present world before us, but rather as subjects of conscious life,in which this world and all that is present – for ‘us’ – ‘makes’ itself throughcertain apperceptions’.(Husserl, 1981: 30)Husserl’s claim that phenomenology implies intersubjectivity has alwaysappeared as a problem to his critics. The subjectivist point of departure wouldrather seem to imply solipsism. In the fifth of the Cartesian Meditations (1929),Husserl set out to solve the problem of transcendental intersubjectivity. Hisproposed solution is in terms of Leibnizian monadology. According to Leibniz,monads have no windows; which means that they do not influence one another,

84 Psychologism in early social scienceat least not in a substantial way. The order of the world is a pre-establishedharmony. Now, according to Husserl, the transcendental ego is similarly unaffectedby any interaction between ego and others. Transcendentalintersubjectivity is egological and derived by analogy. ‘It is clear from the verybeginning that only a similarity connecting, within my primordial sphere, thatbody over there with my body can serve as the motivational basis for the “analogizing”apprehension of that body as another animate organism’ ([1929] 1973: 111).Husserl’s term for this form of analogy is ‘pairing association’. It may beobjected that this solution does not explain intersubjectivity, but this would beunfair, since phenomenology is not intended to be explanatory. In the terminologyof Husserl, phenomenology ‘constitutes’ the world. Whether Husserl’ssuggested solution to the problem of transcendental intersubjectivity is asuccessful constitution, or not, is beyond my power to judge.I will end this short presentation by mentioning the last development ofHusserl’s thinking. Towards the end of his life, Husserl was working ferventlywith a new foundation for phenomenology and knowledge. The result was TheCrisis of European Sciences and Transcendental Phenomenology (written between 1935–7and publihsed in 1954), which remained unfinished at his death in 1937. In thiswork, Husserl proceeds in the unusual way of a history of knowledge.Transcendental phenomenology now appears as the end-product and consummationof this history. 37 The most interesting novelty, for my purposes, however,is the introduction of the concept of ‘life-world’ [Lebenswelt] as a new point ofdeparture for phenomenological reduction.In their everyday life, human beings live together in a pregiven intersubjectiveworld, which they take for granted ([1954] 1970: 103ff). ‘The life-world is a realmof original self-evidences’ (p. 127) or ‘common-sense’. The life-world is pre-scientific,both historically and epistemologically. Science emerges from the life-world,as the result of the rise of the ‘theoretical attitude’ in Greek Antiquity, but it isforever anchored in the life-world, to the extent that science is a human creationand a social activity (pp. 269ff). 38 The dominant form of the theoretical attitudein the history of Western rationalism has been the objectivism of natural science.Also psychology has been snared in this unfortunate conception. According toHusserl, not even Brentano and Dilthey were successful in their attempts toescape this trap (pp. 222f, 245f). Closest to a breakthrough was Brentano, with hisimportant insight that mental phenomena are intentional (pp. 233f). 39The only way to establish psychology as a rigorous science is subjectivism andphenomenological reduction ([1954] 1970: 212). The point of departure is thelife-world, but stripped of any trace of objectivism.Why does not the whole flowing life-world not figure at the very beginningof a psychology as something ‘psychic,’ indeed as the psychic realm which isprimarily accessible, the first field in which immediately given psychicphenomena can be explicated according to types? … Naturally there aredifferences in the manner of life-world experience, depending on whetherone experiences stones, rivers, mountains or, on the other hand, reflectively

Psychologism in early social science 85experiences one’s experience of them or other ego-activity, one’s own or thatof others … But does this change the fact that everything about the lifeworldis obviously ‘subjective’? Can psychology as a universal science, haveany other theme than the totality of the subjective?(Husserl [1954] 1970: 220)As I have already pointed out, The Crisis is an unfinished work. It is risky,therefore, to use it as a basis for drawing conclusions about Husserl’s final position.It does seem to me, however, as if Husserl had been moving in the directionof psychologism. It is pretty obvious, I believe, that descriptive psychology nowoccupies an important place in the phenomenological reduction ([1954] 1970:191ff, 235–41). Husserl even concludes that ‘pure psychology in itself is identicalwith transcendental philosophy as the science of transcendental subjectivity’ (p.258). This conclusion looks like philosophical psychologism. But Husserl seemsto be a psychologicist also with respect to the human sciences.what is required for the regional sciences of man, then, is obviously first ofall what is sometimes called (by contrast to social psychology) individualpsychology … This individual psychology must … be the foundation for asociology and likewise for a science of objectified spirit (of cultural things),which after all refers, in its own way, to the human being as person, i.e., tothe life of the soul.(Husserl [1954] 1970: 228)Husserl’s phenomenology, especially his phenomenological psychology andhis idea of a life-world, made a tremendous impact on continental philosophyand on the social sciences, especially sociology. Among others, it influenced,directly and indirectly, Dilthey, Scheler, Jaspers, Heidegger, Schutz, Merleau-Ponty, Sartre, Ethnomethodolgy, the Frankfurt School, especially Habermas,Bourdieu and Giddens. Indirectly, through Schutz, it may even have influencedthe Austrian School of Economics. I will return to some of these figures in laterchapters. Here I will make a brief remark on the immediate reception ofHusserl’s phenomenology.It has been maintained, by Kusch (1995: 224ff), that Husserl’sphenomenology emerged victorious in the battle between philosophical schoolsin Germany, at the beginning of the nineteenth century. The most importantreason for this was the success of Husserl’s critique of psychologism. Anotherreason was that phenomenology, unlike neo-Kantianism, could be assimilated bythe emerging philosophy of life (Lebensphilosophie). Kusch mentions three philosophersand three texts as especially important for the development of philosophyin the Weimar Republic. One of them, Oswald Spengler, was a historicist andholist, who is of no interest for my purposes.More interesting is Max Scheler, who was an influential philosopher – eventuallysociologist of knowledge – who started as a phenomenologist, but lateroriented himself towards the new philosophy of life. In an influential article

86 Psychologism in early social sciencefrom 1915, he hailed Friedrich Nietzsche, Wilhelm Dilthey and Henri Bergson asthe most important representatives of this new philosophy (Scheler, 1915). Allthree have a place in the history of individualism. I have already discussedDilthey and I am going to discuss Nietzsche in another work on the history ofindividualism. This leaves the French philosopher Henri Bergson (1859–1941),who suggested a view of consciousness similar to that of German idealism andof William James.In his Time and Free Will (1889), Bergson rejected any attempt to quantify andmeasure psychic states, including the utilitarian states of pleasure and pain andthe sensations of the Weber–Fechner law of psychophysics. So far, he agreedwith Kant. He did not, however, accept Kant’s theory of the forms of understanding.In particular, he disagreed with Kant’s view of time and free will.Much like his friend and ally, William James, Bergson saw consciousness as aflow. Consciousness takes the form of duration (durée) of heterogeneous states,but not of a succession of measurable, homogeneous, time-units. The latterconcept of time is really a misleading derivation from our concept of space. Asindicated by the title of the English translation, Bergson’s book is about free will,which he accepted as a matter fact. We experience free will, and the reason,though philosophers and scientists often deny it, is the idea of time, modelled onthat of space and fostered by science, or, at least, by mechanicism.Bergson was not an individualist and his main legacy was holistic. In his mostfamous work, Creative Evolution (1907), he advanced the idea of emergence ofnovel phenomena, which played an important part in the development of thedoctrine of holism (see Udehn, 1987: 85ff). In his last book, The Two Sources ofMorality and Religion (1932), Bergson adheres to an organicist view of society,which is clearly holistic. But he also introduces the distinction between closed andopen societies ([1932] 1977: 30ff), which Karl Popper used in his famous book TheOpen Society and Its Enemies (1945), and which includes one of most cited statementsof the principle of methodological individualism (see pp. 200ff ). But evenif Bergson was not a methodological individualist, himself, he did play a role inthe development of an individualist theory of society. Bergson exerted a considerableinfluence on the phenomenology, not only of Max Scheler, but also ofAlfred Schutz, and he also influenced the existentialism of Jean Paul Sartre andMaurice Merleau-Ponty (see chapter 5).The third philosopher, mentioned by Kusch (1995: 228, 236ff) is Karl Jaspers,and the text is Psychologie der Weltanschauungslehre (1919), where Jaspers makes a‘psychological’ analysis of worldviews. Jaspers mentions some precursors of thistype of analysis, such as Hegel and Dilthey, but rejects the approach of theformer. His own approach is based mainly on two of Hegel’s individualisticcritics: Nietzsche and Kierkegaard. Among other things Jaspers discusses individualismand collectivism/holism, but only as Weltanschauungen, as psychologicalfacts that is. Jaspers is often conceived of as the founder of existentialism and,the germ of this doctrine can be clearly detected in his Psychologie, with itsfrequent references to the Existenz of human beings. 40

4 Austrian methodologicalindividualismThe new Austrian methodology, which traced market prices to the actions ofindividuals was known as methodological individualism. It pointed out thatmarket prices, as well as all other economic phenomena, rested unequivocallyon the decisions, preferences, and subjective values of individuals.(Greaves, 1996: 2)Methodological individualism, as an explicitly stated principle for social scientificresearch, originates with the Austrian School of Economics. It is also themembers and sympathisers of this school, who have made most to propagate thisprinciple in the social sciences, at least in the early stages of its development.The Austrian School has little in common with the British and Lausanneschools of economics besides the principle of marginal utility. Its roots are not inutilitarianism and it does not make much use of mathematics and the idea ofequilibrium. Its hallmark is subjectivism and it uses a genetic approach to socialinstitutions, prices included. The Austrian School of Economics is, I believe, bestunderstood against the background of German and Austrian thinking aboutmethodology, economy and society. 1Carl MengerThe founding father of the Austrian School in Economics was Carl Menger(1840–1917), one of those three economists who made the marginalist revolution.The other two were W. Stanley Jevons and Leon Walras. Menger’scontribution to the rise of neoclassical economics was Principles of Economics(Grundsätze der Volkswirtschaftslehre, 1871). The main message of this work was thatthe value and the price of an economic good is determined, neither by thelabour it takes to produce it, nor by the total costs of production, but by thesubjective evaluations of economising individuals. The ultimate determinant ofeconomic value and of price is the needs of individual human beings. 2Already in the Preface to Principles of Economics, Menger makes the followingproclamation:

88 Austrian methodological individualismIn what follows I have endeavoured to reduce the complex phenomena ofhuman economic activity to the simplest elements that can still be subjectedto accurate observation, to apply to these elements the measure correspondingto their nature, and constantly adhering to this measure, toinvestigate the manner in which the more complex economic phenomenaevolve from their elements to definite principles.(Menger [1871] 1976: 46f)This, according to Menger, is the method proposed by Francis Bacon andcommonly accepted in the natural sciences. 3 Applied to economics, it amountsto demonstrating ‘that man with his needs and his command of the means tosatisfy them, is himself the point at which human economic life both begins andends’ (p. 108). The ultimate explanation of all economic phenomena, therefore,is in terms of the behaviour of economising individuals. The starting-point ofMenger’s analysis is the isolated individual, represented by Robinson Crusoealone on his island (pp. 133ff), but the same principles of economising govern thebehaviour of individuals exchanging goods on a market. This is Menger’s ‘atomistic’method, which would later become known as ‘methodologicalindividualism’ (see Hayek, 1973: 8; 1978: 276f).Principles of Economics was a work on substantive economic theory, containingonly a few scattered remarks on methodology. The latter subject was treatedmore fully some ten years later in Menger’s second major work, Problems ofEconomics and Sociology (Untersuchungen über die Methode der Socialwissenschaften und derPolitischen Oekonomie insbesondere, 1883). The reason Menger turned to methodologywas the relative neglect of his work on economic theory in an academicatmosphere dominated by the German Historical School in economics. Hismethodological treatise would become all but neglected. It became the ignitingspark of the famous Battle of Methods (Methodenstreit), which divided Germanand Austrian economists of different scientific and political persuasions forseveral decades. 4 The main combatants in this battle were Carl Menger andGustav von Schmoller, the undisputed leader of the Younger Historical Schoolin German economics.One, largely latent, source of animosity between the two combatants was nodoubt political-ideological (cf. Cubeddo, 1993: 21ff). Menger was a liberal, witha certain conservative leaning, defending a free market economy against stateintervention. Schmoller was a conservative state socialist, a so-called ‘AcademicSocialist’, or ‘Socialist of the Chair’ (Kathedersozialist), who saw the economy as anaffair of the state and advocated social reform. Simply put, Menger was relativelymore pro-market, Schmoller more pro-state. More manifestly, however, thebattle raged over the nature of economic science and the methods proper to it.Menger defended a theoretical economics proceeding, like classical economics,by abstraction and deduction. First, rational and self-interested economic man isabstracted from concrete human beings. Second, the market of pure exchangerelations is abstracted from the society in which it is embedded. The lawsgoverning the universe of theoretical economics have the status of exact natural

Austrian methodological individualism 89laws. Schmoller conceived of economics as a historical science proceeding byinduction from concrete historical reality, in order, eventually, to arrive at laws ofeconomic phenomena. From this point of view, the postulates of theoreticaleconomics are simply false – figments of the imagination. An important part ofthe argument concerned the relation between theory and practice, but this is notmy business here. The most important difference, for my purposes, is thatMenger defended an atomistic method focusing on individuals, while Schmollersuggested a holistic approach focusing on social institutions. For Schmoller, theeconomy is but one aspect of the life of nations and must be analysed as part ofsociety as a whole. 5In his Problems of Economics and Sociology (1883), Menger made a division of thefield of economics into three groups: the historical, the theoretical and the practicalsciences. His main concern was to defend theoretical economics against theattempts to turn economics into a historical science. Theoretical economics, inits turn, was divided into a realistic-empirical and an exact orientation. Withoutdenying the utility of the former, Menger saw, as the crowning achievement oftheoretical economics, its development into an exact science. 6One of the historical school’s main criticisms of theoretical economics wasthat it is atomistic. This charge of atomism ‘is supposedly based on the fact thateconomic phenomena theoretically are reduced ultimately to individualeconomic efforts or to their simplest constituent elements, and are thusexplained’ (Menger [1883] 1963: 90f). According to Menger, however, this criticismis based upon a misunderstanding, since all exact sciences are necessarilyatomistic. Arguing for an exact economics, therefore, is tantamount to arguingfor an atomist method (pp. 93f).Menger’s defence of atomism in economics is combined with a critique of thealternative interpretation of ‘national economy’ as concerned with the nation as‘a large subject that has needs, that works, practices economy, and consumes’([1883] 1963: 92f).Thus the phenomena of ‘national economy’ are by no means direct expressionsof the life of the nation as such or direct results of an ‘economicnation.’ They are, rather the results of all the innumerable individualeconomic efforts in the nation, and they therefore are not to be broughtwithin the scope of our theoretical understanding from the point of view ofthe above fiction … Whoever wants to understand theoretically thephenomena of ‘national economy,’ those complicated human phenomenawhich we are accustomed to designate with this expression, must for thisreason attempt to go back to their true elements, to the singular economiesin the nation, and to investigate the laws by which the former are built upfrom the latter. But whoever takes the opposite road fails to recognize thenature of ‘national economy.’ He moves on the foundation of a fiction, butat the same time he fails to recognize the most important problem of the

90 Austrian methodological individualismexact orientation of theoretical research, the problem of reducing complicatedphenomena to their elements.(Menger [1883] 1963: 93)Menger does not, however, deny the existence of unitary social wholes. Whilecritical of the excesses of organicism in nineteenth-century social theory, he doesadmit that there are social ‘organisms’, in the sense of unitary wholes. He alsoadmits that they may be subjected to functional analysis of social institutions asparts of such wholes. But the existence of organic wholes is not something to beassumed a priori, it is something to find out by research, and even unitary wholesare amenable to an exact atomistic analysis. ‘The acknowledgement of anumber of social phenomena as “organisms” is in no way in contradiction to theaspiration for exact (atomistic!) understanding of them’ ([1883] 1963: 141).The defining characteristic of an ‘organism’ is that it has grown, withoutbeing anyone’s deliberate creation. Menger, thus, sides with the critics of thetheory of the social contract, such as Hegel and Savigny. While many social institutionsand structures are mechanisms of pragmatic origin, others are organismswhich have ‘grown’, in the sense of being the unintended result of the actions ofmany men. This is the immediate source of Hayek’s celebrated distinctionbetween two types of order: organisations and spontaneous orders (see p. 120).The most well known example of an organism or spontaneous order, in thesense of Menger and Hayek, is the market. At least, as analysed by Adam Smith,where the it appears as if guided by an invisible hand. Prices and othereconomic phenomena are the unintended results of the economic action ofmany individuals and, as such, amenable to atomistic analysis. More original andinteresting, however, is Menger’s suggestion that other social institutions also,such as the family, the state, and law can be explained in similar manner. Simplystated, the procedure is to show how social institutions emerge as the unintendedend-result of a sequence of actions, many of which are intended and rational,others mere custom. Menger himself uses this type of atomistic analysis in hiswell-known explanation of the origin of money (Menger [1871] 1976: ch. 8;[1883] 1963: 152–5; 1892).In the beginning, human beings exchanged goods in the form of barter. Butsome goods are more useful for this purpose than others, since they fulfil needsthat are common to many people. These goods, therefore, are more saleable.Cattle, for instance, is a good which virtually all individuals find useful and thereforetend to accept as payment. This praxis turns into custom and this customseems to have been the embryo of money in many cultures. Though useful,cattle have some disadvantages. They are big and heavy and not perfectly divisible,at least not when alive. Other goods such as shells and pieces of metal haveobvious advantages in this respect. How it comes about that money takes exactlythe form it has today may be hard to tell, but the principle of explanation isclear:It is clear … that the origin of money can truly be brought to our full under-

Austrian methodological individualism 91standing only by our learning to understand the social institution discussedhere as the unintended result, as the unplanned outcome of specifically individualefforts of members of society.(Menger [1883] 1963: 155)I suggested earlier (p. 9) that Hobbes’s theory of the social contract was a firstparadigm of methodological individualism-in-use. I also suggested that AdamSmith’s idea of the market, as governed by an invisible hand, was a secondparadigm of theoretical individualism (p. 12). I now wish to suggest that CarlMenger’s explanation of the origin of money is a third paradigm of methodologicalindividualism-in-use. 7 It is an explanation which has precedents in theconservative thinking of German historicism, especially that of Savigny and inHume’s analysis of convention. It is similar to Smith’s theory of the market, tothe extent that it is an invisible-hand explanation of a spontaneous order, but it isdifferent in being an evolutionary, or genetic, explanation. Menger’s explanation,forebodes, not only the game-theoretic analysis of conventions as solutions togames of co-ordination (Wärneryd, 1990: ch. 4), but also explanations in termsof path-dependence.Figure 4.1 Carl Menger’s theory of social institutionsThe German economist Gustav von Schmoller reviewed Menger’s book onmethodology in the year of its appearance, but in a tone that Menger did notlike at all. From his pedestal in German academic life, Schmoller looked downupon Menger as an ‘acute dialectician, with a logical mind’, but lacking ‘theuniversal philosophical and historical training’ necessary for a reformation ofeconomics (Schmoller, 1883: 251). It must have been the tone that Mengerreacted to, since Schmoller was very predictable in his review and actuallyagreed with some of Menger’s views. Anyway, Menger became absolutelyfurious and wrote a piece called Irrthümer des Historismus (Delusions of Historism,1884), which consists of little more than invectives and ironies directed atSchmoller. The final act was melodramatic: Schmoller (1884: 677) returned hiscopy of Menger’s Irrthümer to its author with an open letter explaining that thealternative would have been to throw it in the waste-paper basket. 8The Battle of Methods raged for decades and included several contributionsby Menger himself and by his pupils. Nothing of great importance was added,

92 Austrian methodological individualismhowever, to the topic of methodological individualism, or atomism. Mengermade an addition to his earlier tri-partition of economics to comprise also amorphological orientation (1889) which may be seen as a departure frommethodological individualism. 9 Emil Sax (1884: ch. 5) made a distinctionbetween individualism and collectivism as tendencies in the motives and actionsof human beings and went on to argue that corresponding to these tendenciesare different economic theories (ch. 6). Nothing follows from Sax’s suggestion,however, for the issue of methodological individualism versus collectivism.Eugen von Philippovich (1886: 50, n 3) rejected Sax’s distinction and restatedMenger’s original position, which he clarified in some respects. He maintained(pp. 38f), for instance, that economics accepts individuals’ wants (preferences) asgiven, without asking the further question why they happen to have these particularwants.Friedrich von Wieser (1851–1926) agreed with Menger that there is no suchthing as a social interest (1884: 25). Society is not a person and has no interestother than the interests of its members. He also defended the atomistic methodof marginalist economics. The fundamental law of ‘marginal utility’ – a termthat Wieser introduced (p. 128) – is true of the isolated individual, no less than ofindividuals in exchange (pp. 210–14). This is in contrast to the theory of price,which is a theory of economic exchange (Wieser [1893] 1971: 50–3). While criticalof the ‘naive formulation’ of the ‘individualistic theory of society’, typical ofpolitical individualism, he defended methodological individualism:one cannot get away from its fundamental concept, that the individual is thesubject of social intercourse. The individuals who comprise society are thesole possessors of all consciousness and of all will. The ‘organic’ explanation,which seeks to make society as such, without reference to individuals,the subject of social activity, has patently proved a failure. One must holdhimself aloof from the excesses of the individualistic exposition, but theexplanation must still run in terms of the individual. It is in the individualthat one must look for those tendencies that make the social structure, thatdove-tail (if we may use that expression) in such a manner as to give the firmcohesion of social unity and at the same time provide the foundation for theerection of social power.(Wieser [1914] 1967: 154)Eugen Böhm-Bawerk (1851–1914) similarly distinguished between twoeconomic processes: one involving only the isolated Robinson Crusoe and thegoods he uses, the other involving exchange of goods with other individuals(1891: 379). In the individualist imagery of Böhm-Bawerk, atomism is replacedby the idea of the individual as a microcosm, reflecting the macrocosm of theeconomy as a whole (pp. 380f). In his famous critique of Marxian economics,Böhm-Bawerk (1896: 258ff, 278f, 287) maintains that the root of the trouble isMarx’s use of Hegelian dialectics instead of adequate psychological microfoundations.

Austrian methodological individualism 93Like early British neoclassical economics, the early Austrian School wasclearly and explicitly psychologistic (Endres, 1997: 33–7). 10 As we have alreadyseen, the psychological foundation of Menger’s economics was a theory ofneeds. Von Wieser and Böhm-Bawer dropped the psychology of needs, butretained psychologism in the form of satisfaction of wants. According to vonWieser (1884: 39), ‘the theory of value, rightly conceived, is applied psychology’and Böhm-Bawerk agreed, as we have seen. Increasingly, however, psychologismwas replaced by new microfoundations in the form of rational choice (Endres,1997: 213–15), conceived of as free from psychology.Menger was much less explicit about rationality than he was about methodologicalindividualism. The individualist basis of his theory is individualsengaged in the satisfaction of their needs. No doubt, however, Menger’sreasoning is based on the assumption that individuals do so in a rational way.This becomes evident when he makes the qualification that individuals do notalways act rationally in the satisfaction of their needs:Even individuals whose economic activity is conducted rationally, and whotherefore certainly endeavour to recognize the true importance of satisfactionsin order to gain an accurate foundation for their economic activity, aresubject to error. Error is inseparable from all human knowledge.(Menger [1871] 1976: 148)It seems to me that the element of rational choice in early marginalisteconomics was pointed out most clearly by economists outside the Austriancamp. One of them was Heinrich Dietzel who argued, in effect, that economicsis a theory of rational action. The economic principle (Das wirtschaftliche Prinzip)does not depend upon any particular motive, such as self-interest, but states onlythat individuals are instrumentally rational in the choice of means to satisfy theirneeds (Dietzel, 1883: 26ff; 1884: 17–44). 11The Austrian economist Emile Sax responded to Dietzel’s suggestion bypointing out that his formulation of the economic principle fails to exclude technologyfrom the scope of economic theory. Following the German sociologistAlbert Schäffle, Sax suggested that the economic principle is what the Austrianscalled economising (wirtschaften): the use of scarce resources to achieve amaximum of satisfaction (Sax, 1884: 9ff). This is, I believe, the main contributionof Austrian Economics to a theory of rational choice: they all conceived ofeconomic theory as essentially concerned with individuals rationally engaged ineconomising on scarce resources, with alternative uses. This view of economicswas eventually made famous by Lionel Robbins in his The Nature and Significance ofEconomic Science (1932).The Austrians of the first generation were aware and admitted thatmarginalist economics relies on an assumption of purposeful and rationalbehaviour on the part of economic agents, but this assumption was never thefocus of their attention. It is my conjecture that the work of Max Webercontributed significantly to make rationality a central assumption of Austrian

94 Austrian methodological individualismEconomics – it was certainly an important theme of Max Weber’s own work inhistory and sociology.To conclude: Carl Menger may be considered the founder of ‘methodologicalindividualism’, but, like Mill, he did not use this term himself. Borrowing a labelused by the critics of classical economics, he called it ‘atomism’, which meansthat complex phenomena should be explained in terms of their simplestelements, or parts. It is not absolutely clear to me whether Menger’s atomismnecessarily implies strict methodological individualism, or if there may be atomslarger than individual human beings. From the point of view of economictheory any economising unit – a household, a firm, even a state – may be treatedas an atom of analysis (see Menger [1871] 1963: 193–6). On the other hand, it isperfectly clear that, according to Menger, the ultimate explanation of alleconomic phenomena is in terms of needs, something attributable only to individuals.Menger’s plea for atomist analysis is not dogmatic and it is strictlylimited to the exact orientation of theoretical economics. History and the empirical-realisticorientation in social science make legitimate and necessary use of acollectivist method. 12Menger’s immediate followers accepted his atomistic method, as well as itslimits, but did not contribute to its development, or clarification. AmongAustrian economists, it is above all Friedrich von Hayek who has followedMenger in matters methodological. But there were others. Two Germaneconomists, Heinrich Dietzel and Max Weber, were swimming against the tideand professed a heretical belief in abstraction and atomism against the orthodoxyof the Historical School. Of these two, Weber is by far the most important,and will be treated in a separate section below, but the now forgotten HeinrichDietzel was not without influence in his own lifetime and deserves a fewcomments.Dietzel was read by the Austrian economists and may have had some influenceon their methodological views. Like them, he made a distinction betweenthe theory of man’s relation to economic goods, which he called economics(Wirtschaftslehre), and the theory of economic exchange, which he called socialeconomics (Socialwirtschaftslehre, later Socialökonomik). Like Menger, Dietzel wascritical of the idea of a ‘national economy’ (Volkswirtschaft or Nationalökonomie)because of the misleading connotations of this term (Dietzel, 1895: 51ff). Hisown preferred term ‘social economics’ was suggested as a way to avoid theRomantic ideas of a people (Volk) and a nation (Nation), which only obscured thetrue nature of the economy as marketplace. Dietzel was a defender of classicaleconomics and later also of marginal economics against the critique of theHistorical School. He defended the abstract or, as he preferred to call it, theisolating method and also atomism, or individualism, at least for the purposes ofsocial economics. Dietzel (1882: 55) referred to individualism as a ‘premiss’ ofsocial economics and, indirectly, as a ‘methodological assumption’ (p. 62) whichcomes close to suggesting the term ‘methodological individualism’, although hedid not actually do so. Methodological individualism is also implicit in Dietzel’sconception of economics as a theory of instrumental action, more specifically

Austrian methodological individualism 95economic action, aimed at the satisfaction of needs by means of economic goods(Dietzel, 1883: 59ff; 1884: 17ff). Social phenomena are caused and constitutedby the actions of individuals, which are caused by the motives of action, whichare caused by the needs of individuals (Dietzel, 1895: 18). Dropping needs as theultimate causes of social phenomena, we get something close to Max Weber’sview in a nutshell.Max WeberThe Austrians of the first generation were aware and admitted that marginalisteconomics relies on an assumption of purposeful and rational behaviour on thepart of economic agents, but this assumption was never the focus of their attention.It is my conjecture that the work of Max Weber contributed significantly tomake rationality a central assumption of Austrian Economics – it was arguablythe central concern of Max Weber’s own work in history and sociology. As iswell known, Max Weber saw history, or at least the history of the West, assubject to a tendency of increasing rationalisation. Not only economic life, butalso politics, social life and culture, are subject to the irresistible force of humanreason. In economic life, we see a culmination of instrumental rationality withthe development of modern capitalism and the market. For this reason,economics, in particular, must be based on the assumption of instrumental rationalityon the part of agents. (Udehn, 1991: 134–6).The first source of Weber’s view of the role of rationality in economics is insome notes he made to a course on theoretical economics in 1898 (Weber [1898]1990: 29ff). Weber here conceives of economising as means-ends rationality, butdoes not yet make the distinction between economic and technical rationality.He is careful to point out that economics is not about real human beings, butabout a constructed economic subject with certain ideal characteristics. Amongthe most important features of this ideal figure are those that define her/hisrationality: perfect knowledge of the situation (the alternatives) and perfectunderstanding of the most adequate means to realise the various alternatives.Weber returns to the ideal typical assumption of rationality in his earlymethodological essays on “‘Objectivity” in Social Science and Social Policy’(1904) and on the German economist of the Historical School Karl Knies(Weber, 1903–6). In the first essay, Weber arrives at the common interpretationof economics as bound up with the fundamental fact of scarcity (1949: 64f). Thisarticle also contains Weber’s first extensive discussion of ideal types, illustratedby abstract economic theory: ‘It offers us an ideal picture of events on thecommodity-market under conditions of a society organized on the principles ofan exchange economy, free competition and rigorously rational conduct’ (pp.89f). In the second article, Weber’s starting-point is Knies’s argument thatfreedom of will is somehow irrational and, therefore, escapes social scientifictreatment ([1903–6] 1975: 95–101). Against this argument, Weber correctlymaintains that freedom is related to rationality rather than to irrationality, andalso that free and rational actions are most eminently suited to social scientific

96 Austrian methodological individualismtreatment (pp. 120–9). Weber presents rational interpretation as the most importantform of interpretive understanding. Rational action, which consists inchoosing the most adequate means to reach an end, exhibits a unique degree ofself-evidence, which makes it methodologically privileged. If we know that anaction X is related to Y as cause to effect, then we can explain X as a means toreach the end Y. Such rational interpretation can be used either as a hypothesisto be verified by the facts, or as an ideal type, with which to compare actualactions. Actions, which are not rational, may then be understood as non-rationalactions, as deviations from the ideal type of rational action. In both cases,according to Weber, rational interpretation is of ‘extraordinary heuristic value’(pp. 186–191). With this argument, Weber appears as an early advocate of the‘principle of charity’, later defended by Donald Davidson and Jon Elster, amongothers.A fourth source of Weber’s views of rationality in economics is his article on‘Marginal Utility and “The Fundamental Law of Psychophysics”’ (Weber,1908a). His main business in this article is to deny that the idea of marginalutility is in any way dependent on the Weber–Fechner law of psychophysics orany other psychological laws for that matter. Economics is based on certaincommon-sense observations of economic behaviour and on the equallycommon-sense assumption that individuals are rational in their economicactivity. Economic theory rests on the assumption of perfect rationality and itsheuristic value rests on the socio-cultural fact that economic life in the Westincreasingly approximates this assumption.Weber’s last treatment of the methodology of economics, is in his article on‘The Meaning of “Ethical Neutrality” in Sociology and Economics’ and inchapter II of Economy and Society (1922). In the first, you will find Weber’s secondextensive treatment of ideal types, once again exemplified by economics:Economic theory makes certain assumptions which scarcely ever correspondcompletely with reality but which approximate it in various degrees andasks: how would men act under these assumed conditions, if their actionswere entirely rational? It assumes the dominance of pure economic interestsand precludes the operation of political or other non-economic considerations.(Weber 1949: 44)In Economy and Society, Weber returns to the nature of economics. Economicaction is now defined as ‘any peaceful exercise of an actor’s control overresources which is in its main impulse oriented towards economic ends’.Economic action is not necessarily rational. Rational economic action ‘requiresinstrumental rationality in this orientation, that is deliberate planning’ ([1922]1978: 63). But resources are necessarily scarce, ‘the most essential aspect ofeconomic action for practical purposes is the prudent choice between ends’ (p. 65).Weber has finally arrived at the distinction between economics and technology.The difference between them is this: ‘Economic action is primarily oriented to

Austrian methodological individualism 97the problem of choosing the end to which a thing shall be applied; technology, tothe problem, given the end, of choosing the appropriate means’ (pp. 66f). 13Max Weber was not from Austria and he was not only an economist. He wasa jurist, historian, economist and sociologist, with a special interest in methodology.To posterity Max Weber is known as a sociologist, but it has been arguedthat he was first of all an economist, albeit of a kind that is extinct today; amember of the German Historical School in economics (Hennis, 1988: ch. 3).More common is the view that Weber was primarily a historian, using sociologyas an auxiliary science, providing history with a system of basic concepts (Burger,1976: 138; 1994; Roth, 1976). This view is often coupled with the suggestionthat he developed from being a historian, to becoming more of a sociologist. 14 Ithink that Max Weber defies classification in terms of disciplinary belonging, butthat is probably the least important thing about him. Whatever he was, MaxWeber was a great scholar and a great thinker.Weber’s attention to methodological issues was, at least partly, occasioned bythe Methodenstreit and resulted in his intervention in that battle. It is generallyagreed that Weber tried to steer a middle course between the positions of themain combatants Menger and Schmoller. The famous ideal type is often interpretedas Weber’s way to avoid both horns of a dilemma. 15 Weber agreed withMenger that theoretical economics is, and should be, an abstract and atomisticscience, but agreed with Schmoller and the Historical School that economics is acultural science and a Wirklichkeitswissenschaft; a science of reality (von Schelting,1922: 701–26; Tenbruck, 1986). The simplifying assumptions of economics;homo economicus, perfect competition, etc., are ideal types used for heuristicpurposes, but should not be mistaken for natural laws. As Schumpeter (1954:819) suggests, Weber ‘saw no objection of principle to what economic theoristsactually did, though he disagreed with them on what they thought they weredoing, that is on the epistemological interpretation of their procedure’. Even so,Joseph Schumpeter classified Weber as a member of the ‘Youngest HistoricalSchool’ in economics. His reason for doing so is Weber’s adherence to thedoctrine of Verstehen (cf. also Lachmann, 1970: 17–48). I suggest that Webershared the substantive concerns of the Historical School, but was closer to theAustrian School in methodology. 16There are many sources of Weber’s methodology, but three, in particular, aregenerally recognised: Heinrich Rickert’s neo-Kantianism, Austrian Economicsand Dilthey’s doctrine of Verstehen (Eliaesson, 1990). 17 For my present purposes,the influence of Austrian Economics is most important, because this is the mostimmediate source of his methodological individualism. By way of extremesimplification, I suggest that Weber took methodological individualism from theAustrian School of Economics, supported it by a neo-Kantian view of conceptformation and turned it into subjectivist methodological individualism with thehelp of Dilthey and Simmel. 18The suggestion that Weber took his methodological individualism fromeconomics gains support from a letter he wrote to the economist RobertLiefmann in 1920:

98 Austrian methodological individualismI do understand your battle against sociology. But let me tell you: If I nowhappen to be a sociologist according to my appointment papers, then Ibecame one in order to put an end to the mischievous enterprise which stilloperates with collective notions (Kollektivbegriffe). In other words, sociology,too, can only be practised by proceeding from the action of one or more,few or many, individuals, that means, by employing a strictly ‘individualist’method.(Quoted in Roth, 1976: 306)This is the first time Weber makes explicit mention of an ‘individualistmethod’. His critique of the use of collective concepts, however, goes back to hisearly methodological essays. Already in his articles on Roscher and Knies(1903–6) – economists of the German historical school – Weber criticises, asmetaphysical, the use of concepts such as ‘Volk’ and ‘general will’. He is also criticalof the view of societies as organic wholes, and the attempt by Roscher toarrive at the laws of development of such wholes (Weber [1903–6] 1975: 60–91,202–7). Weber sees in the use of such concepts and doctrines the influence ofGerman Romanticism and especially of Hegelian ‘panlogism’. What he objectsto, is the view that particular cultural phenomena are explained as ‘emanations’of a volksgeist and as parts of an organic whole. Weber obviously prefers theopposite route. To the extent that there are such things as Volksgeists and socialwholes at all, they have to be explained as the results of particular cultural valuesand beliefs entertained by individual human beings (cf. Burger, 1994: 88ff).Weber, then, sides with Austrian atomism against the collectivism and organicismof the German Historical School. 19In his article “‘Objectivity” in Social Science and Social Policy’ (1904), Webercriticises the conflation of concept with reality. The failure to realise thatconcepts are theoretical constructs of the mind, has particularly unwholesomeeffects in the case of collective concepts taken from the language of everyday life.Weber mentions as particularly problematic the ascription of interests to collectiveentities such as ‘agriculture’, ‘classes’ and the ‘state’. ‘The use of theundifferentiated collective concepts of everyday speech is always a cloak forconfusion of thought and action’ (Weber, 1949: 107–10). With respect to thescientific conception of the ‘state’, Weber makes clear that it is ‘always asynthesis which we construct for certain heuristic purposes … abstracted fromthe unclear syntheses which are found in the minds of human beings’ (p. 99).When we inquire as to what corresponds to the idea of the ‘state’ in empiricalreality, we find an infinity of diffuse and discrete human actions, bothactive and passive, factually and legally regulated relationships, partlyunique and partly recurrent in character, all bound together by an idea,namely, the belief in the actual or normative validity of rules and of theauthority-relationships of some human beings towards others.(Weber, 1949: 99)

Austrian methodological individualism 99Weber’s distrust of collective concepts was also voiced in his contribution tothe meeting of the German Sociological Association in 1910. He attacked acertain Dr Ploetz for making unscientific use of the concepts ‘race’ and ‘society’.Biological traits do not explain social phenomena. The social use of the conceptof ‘race’ is invariably ideological; invented and used to serve the interest of somegroup in the oppression of another group (race). The concept of ‘society’,according to Weber ([1910] 1971: 36), is purely conventional and may bereplaced by ‘social relationships and social institutions’. What Weber objected towas Ploetz’s suggestion that society is a living creature; his organicism. From asociological point of view ‘nothing useful ever emerges from the joining togetherof several precise concepts with indefinite notions. And so it is here. We have thepossibility of understanding the rational behavior of single human individualsthrough intellectual empathy’ (p. 39). In these short remarks are contained twoimportant seeds of Weber’s emerging sociology: (1) the belief that all complexsocial phenomena can be defined in terms of social relationships and (2) theconviction that all social phenomena can be ultimately understood andexplained in terms of individuals’ motives for their own actions.Weber’s first positive statement of methodological individualism, or‘atomism’, can be found in his article ‘Some Categories of InterpretiveSociology’ (1913). This article, written in preparation of his magnum opusEconomy and Society (Wirtschaft und Gesellschaft, 1922), also for the first time containsan outline of his interpretive (verstehende) sociology. Methodological individualism,in fact, is presented as the logical outcome of verstehen.The object of the discussion, ‘Verstehen,’ is ultimately also the reason whyinterpretive (verstehende) sociology (as we have defined it) treats the single individualand his action as its basic unit, as its ‘atom,’ if a questionable analogyis allowed here.(Weber [1913] 1981: 158)Weber contrasts his sociology with other approaches, which treat the individualeither as a complex of smaller (chemical or psychic) parts, or as parts ofsocial wholes. Interpretive sociology, then, is atomistic not only in the sense oftreating the human individual as its elementary unit, but also in the sense oftreating the individual as itself indivisible. Most relevant for my purposes is thestandpoint of interpretive sociology with respect to collective concepts.It is peculiar not only to language but also to our thinking that the conceptsthat comprehend action let it appear in the guise of a persistent structure,either of a material entity, or of a personified force leading a life of its own.This is true particularly in sociology. Concepts such as the ‘state,’ ‘association,’‘feudalism,’ and the like generally indicate for sociology categories ofcertain kinds of joint human action; it is therefore the task of sociologyto reduce these concepts to ‘understandable’ action, meaning, without

100 Austrian methodological individualismexception, the action of the participating individuals. This is not necessarilytrue for other disciplines.(Weber [1913] 1981: 158)Weber mentions jurisprudence, which is justified in treating some collectivesas wholes, without bothering much about the component individuals. Thus, forthe purposes of legal science, the state might be treated as a ‘legal person’, but‘[f]or sociological analysis, the word “state” – if it is used at all – signifies only acourse of human action of a particular kind’ (Weber [1913] 1981: 159).In his essay on ‘The Meaning of “Ethical Neutrality” in Sociology andEconomics’ (1917), Weber writes about the ‘individualistic orientation’ ofeconomic theory and maintains that this individualistic orientation is ‘apolitical’and for ‘analytical purposes only’ (Weber, 1949: 44)The term ‘individualistic method’ does not appear in Weber’s writings untilthe posthumous publication of Economy and Society (1922). This work also containsWeber’s most mature and most extensive treatment of methodological individualismand is, therefore, the main source for an understanding of his ideas on thissubject. What he has explicitly to say about methodological individualismappears in the negative form of contrasts and oppositions to other doctrines,from which a positive statement has to be extracted. But one thing emergesclearly: Weber’s ‘individualistic method’ is inseparable from his definition ofsociology as concerned only with subjectively meaningful behaviour, that is,action.Sociology (in the sense in which this highly ambiguous word is used here) isa science concerning itself with the interpretive understanding of socialaction and thereby with a causal explanation of its course and consequences.We shall speak of ‘action’ insofar as the acting individual attaches asubjective meaning to his behavior – be it overt or covert, omission or acquiescence.Action is ‘social’ insofar as its subjective meaning takes account ofthe behavior of others and is thereby oriented in its course.(Weber [1922] 1978: 4)It is to be observed that interpretive understanding is not an end in itself, buta means for arriving at a causal explanation of action. To this effect, Webermakes the distinction between ‘direct observational understanding’ and ‘explanatoryunderstanding’, which was part and parcel of the hermeneutic traditionbefore Weber. Direct observational understanding is achieved when an action isidentified as being of a certain type, explanatory understanding when the motiveof the action is known. ‘A correct causal interpretation of a concrete course ofaction is arrived at when the overt action and the motives have both beencorrectly apprehended and at the same time their relation has become meaningfullycomprehensible’ ([1922] 1978: 12).It may also be pointed out that Weber’s social action makes sociologydifferent from economics, which starts with the isolated individual and sees the

Austrian methodological individualism 101market as a play against nature (see Swedberg, 1998: ch. 2). For Weber, then,rationality is not parametric. Individuals do take the actions of others intoaccount, but Weber did not have the means (game theory) at his disposal toarrive at the idea of strategic rationality (Elster, 1979a; 68f; 2000: 38f). Unlikeboth mainstream economics and game theory, however, Weber focuses his attentionon meaning and this meaning is cultural. For Weber, therefore, botheconomics and sociology are ‘cultural sciences’ dealing with cultural phenomena(Weber, 1949: 66ff). 20The reason why, according to Weber, his interpretive sociology implies anindividualist method, is that only individuals act. ‘Action in the sense of subjectivelyunderstandable orientation of behavior exists only as the behavior of oneor more individual human beings’ ([1922] 1978: 13). This is also the reason whyWeber contrasts the sociological use of collective concepts with that of jurisprudence.In juridical contexts social collectives, such as the state, businesscorporations, associations and foundations are treated like individual persons –legal persons – and according to Weber, this is perfectly legitimate.But for the subjective interpretation of action in sociological work thesecollectivities must be treated as solely the resultants of and modes of organizationof the particular acts of individual persons, since these alone can betreated as agents in the course of subjectively understandable action.(Weber [1922] 1978: 13)Weber’s conclusion is not, however, that concepts denoting social collectivesshould – or could? (Weber is not absolutely clear on this point) – be entirely eliminatedfrom sociological analysis. Such elimination is unrealistic for severalreasons, including their use in other disciplines and in everyday speech. Theimportant thing, according to Weber, is to keep in mind that ‘for sociologicalpurposes there is no such thing as a collective personality which “acts”’, and insociological contexts collective concepts refer only to ‘a certain kind of developmentof actual or possible social actions of individual persons’ ([1922] 1978: 14).Weber also opposes methodological individualism to the functionalist or‘organic’ school of sociology, most typically represented by Albert Schäffle andhis ‘brilliant work’ Bau und Leben des sozialen K örpers (1878). Weber does not,however, dismiss the functional approach as of no value at all to sociology. Heassigns to it the role of something like a heuristic device, and this is yet anotherreason why, according to Weber, collective concepts cannot be altogether eliminatedfrom sociological analysis. The functional frame of reference is useful, firstof all, for illustration and for a provisional orientation. Second, the functionalapproach is sometimes the only means of identifying the processes at work in agiven social phenomenon. Weber warns us, however, against overestimating thecognitive value and, thereby, illegitimately ‘reifying’ the collective concepts used.Functional analysis is only the beginning of sociological analysis, as understoodby Weber. In sociological analysis it is possible to ‘go beyond merely

102 Austrian methodological individualismdemonstrating functional relationships’. It is possible to attain a ‘subjectiveunderstanding of the action of the component individuals’ ([1922] 1978: 15).Additional light is thrown upon Weber’s view of sociology, its individualistmethod and relation to functionalism, by looking at his encounter with the functionalistor ‘universalist’ method of Othmar Spann. Weber grants to Spann theinsight that we must know what kind of action is necessary for the ‘survival’ ormaintenance of a cultural type before we can even pose the question of whatmotives determine the action. 21It is necessary to know what a ‘king’, an ‘official’, an ‘entrepreneur’, a‘procurer’, or a ‘magician’ does, that is, what kind of typical action, whichjustifies classifying an individual in one of these categories, is important andrelevant for an analysis, before it is possible to undertake the analysis itself… But it is only this analysis itself which can achieve the sociological understandingof the actions of typically differentiated human (and only human)individuals, and which hence constitutes the specific function of sociology.(Weber [1922] 1978: 18)Weber concludes his discussion of methodological individualism by criticising,as mistaken, the belief that an individualist method somehow presupposes individualistvalues, or is applicable only to a free market economy. According toWeber, even a socialist economy has to be understood sociologically in the same‘individualistic’ terms. Once again:The real empirical sociological investigation begins with the question: Whatmotives determine and lead the individual members and participants in thissocialist community to behave in such a way that the community came intobeing in the first place and that it continues to exist? Any form of functionalanalysis which proceeds from the whole to the parts can accomplish only apreliminary preparation for this investigation – a preparation, the utility andindispensability of which, if properly carried out, is naturally beyond question.(Weber [1922] 1978: 18)This exhausts what Weber has to say about the ‘individualistic method’. Anypresentation of Weber’s version of methodological individualism would beincomplete, however, without a mention of his definition of the basic concepts ofhis sociology in terms of the social actions of individuals ([1922] 1978: 22ff).Weber’s verstehende sociology starts from the concept of ‘social action’, which, aswe have seen above, is action oriented to the behaviour of others. On thisconcept is based the concept of ‘social relationship’ which is social actioninvolving a plurality of actors, each taking account of the actions of the others.Of special importance is Weber’s utilisation of the concept of ‘probability’ in hisdefinition of ‘social relationship’. A social relationship, according to Weber‘consists entirely and exclusively in the existence of a probability that there will

Austrian methodological individualism 103be a meaningful course of social action – irrespective, for the time being, of thebasis for this probability’ (pp. 26f). It has been argued (see, e.g., Wrong, 1970:24f; Weiss, 1975: 88–90), correctly I believe, that Weber’s use of the concept of‘probability’, in this and in other definitions, should be understood as an attemptto avoid reification and, thereby, to secure methodological individualism (cf. alsoRinger, 1997: 158f). Social orders or structures, being complexes of social relationships,do not exist as entities sui generis, but only as the probability that certainindividuals will act in certain ways.Weber, then, recognises the ‘existence’ of organisations – such as firms andstates – and of social structures – such as markets, religions, systems of law, cities,classes, status groups, nations, races, etc. – but not of society. There is no suchthing as society. Society exists neither as an entity, nor as a ‘level of reality’ inaddition to individuals and their actions. Weber’s sociology, therefore, has no usefor the concept of ‘society’ (cf. Tyrell, 1994). Sociology is a science of individualsand their actions, not of society. 22To conclude this presentation of Weber’s methodological individualism:Weber advanced his ‘individualistic method’ in opposition to the use of collectiveconcepts in sociology. Collective concepts are used in everyday life andinfluence the actions of individuals. As such, they are relevant for the sociologist.But sociologists should not turn collective concepts used in everyday lifeinto their own tools of analysis. Collective concepts, such as class, race, etc.,are also used by social scientists, but should then be treated with utmost suspicion,because they are often used for dubious ideological purposes. Weber didnot, however, want to abolish collective concepts from sociological analysisaltogether, provided one remembers that they denote nothing but complexes ofactions of individual persons. To this end, collective concepts must be definedin terms of individuals. Weber’s definitions at the beginning of Economy andSociety of the basic concepts of sociology in terms of social action is hisattempt to achieve this end. But Weber’s methodological individualism is notonly about concepts. It is also inseparable from his definition of sociology asexclusively concerned with the interpretation of the subjective meaning individualsattach to their actions and the causal explanation of these actions interms of motives. From this definition follows that Weber did not allow the useof collective concepts in sociological explanation. A highly plausible interpretationof Weber’s methodological individualism, therefore, is that it forbids theuse of collective concepts, if not in sociological analysis, then at least in sociologicalexplanation (cf. Runciman, 1972: 24f). It is not altogether clear to me ifWeber intended to suggest that all social phenomena should be explained interms of motive alone and, if so, if he conceived of this as a peculiarity of hisown interpretive sociology, or as imperative for all sociology. My guess is thatthe answer to the first question is yes, but that he saw interpretive sociology asone of several possible types of sociology.

104 Austrian methodological individualismJoseph SchumpeterIn 1909, when Max Weber was planning a handbook in economics calledGrundriss in Sozialökonomik (Outline of Social Economics), he asked Friedrich vonWieser to write about economic theory and his pupil Joseph Schumpeter to writea piece on the history of economic theory. Wieser’s part was published in 1914as Theorie der Gesellschaftlichen Wirtschaft (Social Economics, 1927). Schumpeter’scontribution was published the same year (1914) as Epochen der Dogmen- undMethodengeschichte (Economic Doctrine and Method, 1954). Eventually, Weber was tomake a contribution of his own to this project; Wirtschaft und Gesellschaft,published posthumously in 1922 (Economy and Society, 1978). When Weber, in1918 was offered a chair in economics in Vienna, he was asked to give hisopinion on the candidates to another chair in theoretical economics. Weberunhesitatingly recommended Joseph Schumpeter to the post – the ‘greatest theoreticaltalent’, who is also ‘an excellent teacher’ (quoted in Hennis, 1991: 49).Weber and Schumpeter knew and respected each other, but little is knownabout their personal relationship. 23 Intellectually, influence was probably asymmetric,running mainly from the older man to the younger. It has been argued byRichard Swedberg (1991: 2) that Schumpeter was heavily influenced by Weber’sidea of a transdisciplinary economics, called Sozialökonomik (social economics), andeventually made it his own. There is a presumption of a connection betweenSchumpeter’s theory of economic development and Weber’s theory of capitalism.Schumpeter’s innovative entrepreneur seems to have something incommon with Weber’s charismatic leader (Macdonald, 1965). Schumpeterdefended and adopted Weber’s doctrine of value neutrality (Osterhammel, 1987:109). Schumpeter’s theory of democracy, finally, bears a clear imprint ofWeber’s ditto (Udehn, 1996: 138). Weber, on the other hand, got at least onething from Schumpeter: the term ‘methodological individualism’.Schumpeter was probably first to use the term ‘methodological individualism’(see Machlup [1951] 1978: 471f). He introduced it in Das Wesen und derHauptinhalt der theoretischen Nationalökonomie (The Nature and Essence of TheoreticalEconomics, 1908) in order to make the distinction between methodological andpolitical individualism and, above all, to argue that they are altogether independent.The ultimate purpose of the distinction was to defend the theory ofmarginal utility against the recurrent accusations that, like classical economics, itpropagates political individualism (Schumpeter, 1908: 88–98). The argument isrepeated in Economic Doctrine and Method, which also includes a rather odd characterisationof political individualism:This method of starting from a fact of individual psychology led to twogroups of objections. Firstly, the general objections to individualism andatomism were levelled especially against this school. In this respect anadequate distinction between political individualism, the view that individualsare independent causes of social phenomena which represent merely a

Austrian methodological individualism 105resultant of these causes, and the mere method of starting from the individualfor purposes of pure economics was not made.(Schumpeter [1914] 1954: 190)In the History of Economic Analysis (1954), Schumpeter distinguishes threeforms of individualism: political, sociological and methodological. By the first,Schumpeter now means Manchester liberalism. What in the above quotationwas called ‘political individualism’, is now called ‘sociological individualism’:the view widely held in the seventeenth and eighteenth centuries, that theself-governing individual constitutes the ultimate unit of the social sciences;and that all social phenomena resolve themselves into decisions and actionsthat need not or cannot be further analyzed in terms of superindividualfactors. This view is, of course, untenable so far as it implies a theory of thesocial process. From this, however, it does not follow that, for the specialpurposes of a particular set of investigations, it is never admissible to startfrom the given behavior of individuals without going into the factors thatformed this behavior … In this case we speak of MethodologicalIndividualism.(Schumpeter, 1954: 888f)With ‘methodological individualism’, then, Schumpeter understood themethod-in-use of theoretical economics: ‘the point of departure when describingeconomic phenomena is the actions of individuals’ (dass man bei der Beschreibunggewisser wirtschaftlicher Vorgänge von dem Handeln der Individuen ausgehe [Schumpeter,1908: 90f] my translation). Unlike Weber, however, Schumpeter himself was notcommitted to ‘methodological individualism’, and he did not believe that thismethod was of much use in other theories and disciplines. In the theory oforganisations, for instance, and in sociology more generally, Schumpeter did notbelieve that you would get very far with an individualist method (pp. 94f;539–41). 24In a speech Schumpeter gave in Japan in 1931, but published only recently(1991), Schumpeter’s instrumental attitude to methodology in general, and tomethodological individualism in particular, is most clearly expressed.It is no use before approaching a problem to form a preconceived idea ofhow to proceed with it. As you know, at the beginning of the eighteenthcentury all economists used to start in their analysis with the individual. Thenineteenth-century economists also took the individualistic standpoint. It is[the creed of methodological universalism] that it is quite wrong to startfrom the individual. The individual is only a product of the social environmentand therefore it is wrong to start from him. We know that everyindividual is fashioned by the social influences in which he grows up. In thissense he is the product of the social entity or class and therefore not a freeagent. That is certainly so, but it is quite uninteresting as long as it is stated

106 Austrian methodological individualismin this general form, where it is nothing but a phrase. It is interesting onlywhen we can build a concrete knowledge on the basis of that view, and mygreat objection to it is that we do not make anything out of preachinguniversalism as a creed. In some problems of sociology or political life andso on we have no choice but to start from the social whole. In other cases,such as market phenomenon and most problems of modern industry, thereis no choice but to start from the individual. In one class universalism and inthe other class individualism is the indicated method. Therefore we ought tobe neither individualistic nor universalistic. It is a matter of convenience;neither individualism nor universalism is an eternal truth.(Schumpeter, 1991: 286f)This passage shows clearly that Schumpeter, like the early Austrians, was amethodological pluralist. It also indicates that ‘methodological individualism’meant something different for Schumpeter, than it does for many contemporarymethodological individualists (see pp. 112, 210). Finally, it shows that Schumpeterdid not follow Weber in his attempt to build an individualistic sociology. Withrespect to sociology, Schumpeter sided with Weber’s methodological oppositeEmile Durkheim.In his own sociological works, Schumpeter did not practise methodologicalindividualism. 25 The main sources of inspiration for his theory of social classeswas Gustav Schmoller, Emile Durkheim and Othmar Spann (Schumpeter[1927] 1951: 135). As Schroeter has pointed out (1985: 163), this makes for ‘avery holistic and deterministic approach’, far from the individualistic method ofMax Weber. And, indeed, according to Schumpeter ‘every social class is a specialsocial organism, living, acting, and suffering as such and in need of being understoodas such’ (Schumpeter, 1927: 137). Weber would have disapproved stronglyof this definition, had he been alive. Another deviation from methodologicalindividualism (cf. Swedberg, 1991: 104) is that Schumpeter saw the family, notthe individual, as the unit of class analysis (Schumpeter [1927] 1951: 148). 26The term ‘methodological individualism, then, was coined by JosephSchumpeter. However he himself did not adhere to any doctrine with this name.Schumpeter was a methodological pluralist with an exceptional tolerance foralternative approaches. Schumpeter used the term ‘methodological individualism’descriptively, to designate the method of theoretical economics, butwithout normative implications for other theories. He also used it to distinguishbetween political, sociological and methodological individualism. I believe thatanother distinction would help to clarify his own position. I suggest that a fruitfuldistinction can be made between substantive and procedural methodological individualism.Substantive methodological individualism says something about thecontent of concepts and/or explanations. This is how Weber understoodmethodological individualism and, as we shall see, it is the most commonversion. Procedural methodological individualism, on the other hand, says somethingabout the order of analysis and/or presentation. For Schumpeter,methodological individualism did not say anything about social reality, about the

Austrian methodological individualism 107causes of social phenomena, or about how to explain them. It was merely anacknowledgement of the fact that marginal economics actually starts with thewants of individual human beings – something he calls its modus procedendi(Schumpeter, 1909: 214). Methodological individualism, according toSchumpeter, is procedural methodological individualism.Ludwig von MisesMethodological individualism was also proposed by the Austrian economistLudwig von Mises. Although never his pupil, Mises was, as an economist, muchinfluenced by Menger, 27 and eventually became a leading representative of theAustrian, or the New Austrian, School in economics. He did not, however, shareall of Menger’s ideas on methodology. According to Mises, ‘Menger was toomuch under the sway of John Stuart Mill’s empiricism to carry his own point ofview to its full logical consequences’ (Mises, 1969: 27f). In matters methodological,Mises was also influenced by Weber (Selgin, 1988: 22), and this includes thematter of methodological individualism (Rothbard, 1979: 57). He was personallyacquainted with Weber and frequently discussed methodological issues with him(Mises, 1978: 4). It has also been testified that Weber’s interpretive sociology wasa favourite topic at Mises’s private seminar in the 1920s, which includedmembers such as F.A. von Hayek, Fritz Machlup, Alfred Schutz and FelixKaufmann (Haberler [1961] 1974: 191). There is no reason to doubt this testimony.Mises’s first substantial contribution to social science methodology,‘Sociology and History’ (Soziologie und Geschichte, [1929] 1933), was a criticaldiscussion of Weber’s ideal type.Like Schumpeter, Mises was sympathetic to Weber’s doctrine of valuefreedom.Unlike the former, however, he also agreed with Weber that sociologyshould conform to the principle of methodological individualism. It should benoted, though, that at least initially Mises had a different conception of ‘sociology’.He conceived of it as a general theory of action, which includedeconomics, as its most developed branch. As soon as he realised that this was notin conformity with common usage, he replaced the term ‘sociology’ by the term‘praxeology’. What Mises liked best about Weber’s sociology was its subjectivismand the fact that it is defined as a theory of action. He did not, however, agreewith Weber on the status of economic laws as ideal types and on his concept of‘rationality’ (Mises, 1933: 68ff; 1961: 124–7). For Mises, all action, as distinguishedfrom (reactive) behaviour, is rational. Action is a category ultimatelygiven to our experience, and praxeology is an a prioristic science resulting frompondering on this ultimate experience. It is sometimes maintained that Mises’spraxeology should be understood along Kantian lines (Rothbard, 1979: 35f;Parsons, 1990: 296ff), as made up of a priori statements that are either analytic(Smith, 1990: 279–82), or synthetic (Selgin, 1988: 21), but this is not the onlypossibility. Praxeology may also be understood as an exact science in the sense ofphenomenology (Chisholm, 1986: 191; Smith, 1990: 279ff).Mises’s first discussion and defence of methodological individualism was in

108 Austrian methodological individualismSocialism. An Economic and Sociological Analysis (Die Gemeinwirtschaft, 1922). Firstamong Austrian economists, Mises explicitly insisted that economic action isrational action. Methodological individualism follows as consequence of this fact:All rational action is economic. All economic activity is rational action. Allrational action is in the first place individual action. Only the individualthinks. Only the individual reasons. Only the individual acts. How societyarises from the actions of individuals will be shown in a later part of ourdiscussion.(Mises [1922] 1951: 113)Mises’s next step is to argue that economics is based on principles governing theaction of the isolated individual. Mises goes further than his predecessors andargues that it is ‘quite correct to regard all economic activity, even the economicactivity of isolated man, as exchange’ (p. 117). Exchange, for Mises, reduces tochoice. It is to prefer (exchange) A to (for) B. This does not mean, however, thatMises operates with an abstract, or asocial, individual. Rational man is a socialbeing.Man is already a member of a social body when he appears as a thinking,willing creature, for the thinking man is inconceivable as a solitary individual… The development of human reason and the development ofhuman society are one and the same process. All further growth of socialrelations is entirely a matter of will. Society is the product of thought andwill. It does not exist outside thought and will. Its being lies within man, notin the outer world. It is projected from within outwards.(Mises [1922] 1951: 291)The prerequisite of all social life is language (pp. 321f). First of all, languageis a precondition for thinking and rational action. Second, it is a precondition forco-operation and, hence, for society. Society, according to Mises, is co-operationand takes the form of division of labour. Therefore, to ask ‘how society arisesfrom the actions of individuals’, is to ask for an explanation of division of labour.The explanation is simple: ‘Progress in the division of labour depends entirelyon a realization of its advantages, that is of its higher productivity’ (p. 302). Thisseems like a rationalistic account of society, but Mises does not accept rationalismwholesale. He rejects the theory of the social contract.The doctrine of natural law has erred in regarding this great change, whichlifts man from the state of brutes into human society, as a conscious process;as an action, that is, in which man is completely aware of his motives, of hisaims and how to pursue them.(Mises [1922] 1951: 43)

Austrian methodological individualism 109Like Menger, Mises recognises two types of social order: organisations andorganisms; the former being conscious creations, the latter a result of growth.Mises is also influenced by Herbert Spencer, and it is probably from him, that hepicks up a second way of drawing the above distinction: ‘Organization is anassociation based on authority, organism is mutuality’ ([1922] 1951: 296). TheOrganisation is the state, the Organism is the market and do not let the formerinterfere with the latter.Mises’s first explicit mention of methodological individualism is, probably, inthe article ‘On the Development of the Subjective Theory of Value’ (Vom Weg derSubjektivistischen Wertlehre), first published in 1929 and included in his EpistemologicalProblems of Economics (Grundprobleme der Nationalökonomie, 1933). In this article,Mises ([1933] 1976: 153) identifies individualism with nominalism and opposes itto universalism (Universalismus). He returns to this subject, at greater length, inthe main essay in this book, ‘The Task and Scope of the Science of HumanAction’ (Aufgabe und Umfang der allgemeinen Wissenschaft von menschlichen Handeln,1933), which is a preparatory work for his magnum opus Human Action.For the purposes of science we must start from the action of the individualbecause this is the only thing of which we can have direct cognition. Theidea of a society that could operate or manifest itself apart from the actionof individuals is absurd. Everything social must in some way be recognizablein the action of the individual. What would the mystical totality of theuniversalists be if it were not alive in every individual? Every form of societyis operative in the actions of individuals aiming at definite ends. Whatwould a German national character be that did not find expression in theGermanism of individuals? What would a church be that did not expressthe faith of individuals?(Mises [1933] 1976: 43)In this quotation, Mises uses an epistemological argument to back up a kindof procedural methodological individualism. Social science takes the individualas its point of departure, since only the individual human being is immediatelygiven to experience. Mises differs from Schumpeter, however, by suggesting thatthe individual is the ‘empirical’ point of departure. For Schumpeter, the individualwas the theoretical point of departure. Second, this epistemologicalargument is supported by the ontological thesis that only individual humanbeings exist, or that collective entities exist only in the actions of individuals.Methodological individualism is associated with nominalism and collectivismwith conceptual realism. According to Mises, the development of methodologicalindividualism in the eighteenth and nineteenth centuries, which supplantedthe older conceptual realism or universalism was the Copernican revolution ofsocial science (Mises [1933] 1976: 153).Mises returns to the principle of methodological individualism in his mostimportant work, Human Action (1949), but little is added to his earlier treatment of

110 Austrian methodological individualismthis topic. Methodological individualism now appears as an implication of praxeology,the science of human action.Praxeology deals with the actions of individual men. It is only in the furthercourse of its inquiries that cognition of human cooperation is attained andsocial action is treated as a special case of the more universal category ofhuman action as such.(Mises [1949] 1966: 41)As in Socialism, Mises argues that economics starts with the isolated individualand also that action is exchange. He now makes a distinction, however, betweenautistic exchange and interpersonal exchange ([1949] 1966: 97f). 28 Once again,‘Society is concerted action, cooperation’. As such, it ‘is the outcome ofconscious and purposeful behavior’ (p. 143). More surprising, Mises alsoendorses the following statement, if reluctantly: ‘Individual man is born into asocially organized environment. In this sense alone we may accept the saying thatsociety is – logically and historically – antecedent to the individual’ (p. 143). Thisis a statement typically made by collectivists against individualism. The presumptionis, of course, that the way society is organised to some extent determines thebehaviour of individuals. This is how Schumpeter saw the matter and the reasonhe was a methodological collectivist (universalist) in sociology (see p. 366, note26). But how does Mises avoid this conclusion. By shifting from methodology toontology. ‘[S]ociety is nothing but the combination of individuals for cooperativeeffort. It exists nowhere else than in the actions of individuals’ (p. 142).As in Epistemological Problems of Economics, Mises introduces the topic ofmethodological individualism by discussing its relation to the issue of nominalismversus realism, but now with a more uncertain relation betweenmethodological individualism and nominalism.No less inappropriate with regard to our problem is the reference to theantagonism of realism and nominalism, both these terms being understoodin the meaning which medieval scholasticism attached to them. It is uncontestedthat in the sphere of human action social entities have real existence.Nobody ventures to deny that nations, states, municipalities, parties, religiouscommunities, are real factors determining the course of human events.Methodological individualism, far from contesting the significance of suchcollective wholes, considers it as one of its main tasks to describe and toanalyze their becoming and disappearing, their changing structures, andtheir operation.(Mises [1949] 1966: 42)Mises now seems to suggest that the problem of individualism versus collectivismhas nothing to do with that of nominalism versus realism, but later on in thesame work he once again lumps them together and declares that the doctrines ofcollectivism, conceptual realism, universalism and holism are the arch-enemies

Austrian methodological individualism 111of individualism ([1949] 1966: 44, 145). The inconsistency disappears when it isrealised that Mises uses the term ‘collectivism’ in two senses, rejecting it in onesense and accepting it in another: Insofar as collectivism asserts that collectiveentities have an existence independent of, or apart from, individual humanbeings, and insofar as it endows these collective entities with minds, purposes andinterests of their own, it is wrong. If it merely asserts that individuals associateand co-operate for certain purposes, collectivism is quite legitimate and compatiblewith methodological individualism (pp. 145–53).But Mises not only accepts the existence of collective entities; in the quotationabove, he maintains that collective wholes, such as ‘nations, states, municipalities,parties, religious communities, are real factors determining the course of humanevents’. How can this statement be compatible with methodological individualism?The answer is that his collectivism is illusory. Collective wholes do notexist ‘out there’. They exist in the minds and in the acts of individuals, and assuch they determine the course of human events.It is the meaning which the acting individuals and all those who are touchedby their action attribute to an action, that determines its character. It is themeaning that marks one action as an action of an individual and anotheraction as the action of the state or the municipality. The hangman, not thestate, executes the criminal. It is the meaning of those concerned thatdiscerns in the hangman’s action an action of the state.(Mises [1949] 1966: 42)The reason why Mises can hold that collective wholes determine the course ofhuman events and still be a methodological individualist is that, according tohim, a ‘collective whole is a particular aspect of the actions of individuals and assuch a real thing determining the course of human events’ ([1949] 1966: 43).Mises, then, does not at all accept the existence of collective wholes. LikeWeber, he only accepts that individuals use collective concepts in their thinkingand talking about ‘society’. Collective concepts, not real collectives, determinethe actions of individuals. Because of this ontology of the social world, epistemologicalconsiderations suggest that methodological individualism is theappropriate procedure.If we scrutinize the meaning of the various actions performed by individualswe must necessarily learn everything about the actions of collectivewholes … Thus the way to a cognition of collective wholes is through theanalysis of the individual’s actions.(Mises [1949] 1966: 42)It is one thing, however, to say that something determines the meaning orcharacter of an action, and another thing to say that it determines the course ofevents. Mises does not clearly separate these two cases, but actually holds thatideas ‘determine’ the actions of individuals in both ways. It would have been

112 Austrian methodological individualismbetter, had Mises said that ideas about collectives determine the course of events,for this is what, in the end, it all comes down to. Mises’s methodological individualismturns out to be some kind of ‘subjective idealism’ opposed, primarily, toMarxian materialism.Society is a product of human action. Human action is directed by ideologies.Thus society and any concrete order of social affairs are the outcomeof ideologies; ideologies are not as Marxism asserts, a product of a certainstate of social affairs.(Mises [1949] 1966: 187)Methodological individualism is also an important theme in Theory and History(1957), and in The Ultimate Foundation of Economic Science (1962). Very little isadded, however, to his earlier views on this matter. In the first work, Mises makesthe somewhat startling statement that ‘collectivist philosophy denies that thereare such things as individuals and actions of individuals’ ([1957] 1985: 256). Inthe latter work, he warns us against the mistake of hypostatisation, that is, theascription of substance or existence to mental constructs. As an example of thisfallacy Mises takes the way the term ‘society’ is used by the ‘various schools ofpseudo science’. Against these schools – Mises’s example is Marxism – he holdsthat ‘society itself is neither a substance, nor a power, nor an acting being’ (1962:78–80). Once again methodological individualism appears as a form of ‘subjectiveidealism’ advanced in direct opposition to Marxist ‘materialism’.For if one realizes that what sets action in motion is ideas, one cannot helpadmitting that these ideas originate in the minds of some individuals andare transmitted to other individuals. But then one has accepted the fundamentalthesis of methodological individualism, viz., that it is ideas held byindividuals that determine their group allegiance.(Mises, 1962: 82)So far I have concentrated on the meaning of Mises’s methodological individualism,but said nothing about its status and range of application. Theimmediate impression is that Mises is more dogmatic about it than were theearlier Austrians. I believe that Mises conceives of methodological individualismas a principle a priori, which the sciences of human action have to follow, becauseepistemology and ontology dictate that it is the only correct way to proceed. Ialso think it is safe to suggest that, for Mises, methodological individualism isuniversally applicable to the sciences of human action. According to Mises, thereare two sciences of human action: praxeology and history (sociology is dividedbetween them). Praxeology, as we have seen is certainly individualistic. But whatabout history? It is individualistic too. Methodological individualism, therefore, isuniversal in scope.

Austrian methodological individualism 113Individualism, as a principle of the philosophical, praxeological, and historicalanalysis of human action means the establishment of the facts that allactions can be traced back to individuals and that no scientific method cansucceed in determining how definite external events, liable to a descriptionby the methods of the natural sciences, produce within the human minddefinite ideas, value judgements, and volitions. In this sense the individualthat cannot be dissolved into components is both the starting point and theultimate given of all endeavours to deal with human action.(Mises, 1962: 82)Mises rejects the idea of a social science distinct from praxeology and history.There is no science of society, only sciences of human action (Mises [1957]1985: 256ff; 1962: 105–8). In this respect, Mises is a follower of Weber, who sawsociology as a science of action, but not of Menger. He also agrees with Weberthat sociology, if it is not a part of praxeology, is a science auxiliary to history or,rather, a part of it. We saw, at the beginning of this section, that Mises criticisedWeber’s doctrine of Verstehen and his notion of ideal types. But what Misesrejected was only Weber’s understanding of economics. Mises actually acceptedWeber’s methodology, tout court, for the purposes of history. 29Mises’s version of methodological individualism derives partly from Menger,but mainly from Weber. For Mises, as for Weber, methodological individualismfollows from a theory of action and from a subjectivist interpretation of socialreality. There is a difference, however. For Weber, subjectivism is bound up withhis idea of an interpretive sociology; subjective reality is the subject matter ofsociology, but it is not necessarily the only social ‘reality’. For Mises, subjectivereality is the exclusive social reality. A consequence of this difference is a changeof emphasis, from the methodological level in the case of Weber, to the ontologicallevel in the case of Mises. The reason for this change of emphasis is,probably, that Mises was influenced by the phenomenology of Edmund Husserland Alfred Schutz (see Mises 1943/44: 530).When it comes to explicit formulations of methodological individualism,there is not much to touch upon in the case of Mises. One of its main tasks issaid to be that of describing and analysing collective wholes in terms of individuals.This methodological principle is underpinned by the epistemologicalargument that all we can know about collective wholes is the actions of individuals,and the ontological thesis that collective wholes exist only in the minds andin the actions of individuals. But Mises also formulates methodological individualismas a thesis about social causation: it is ideas held by individuals thatdetermine their group allegiance. More generally, Mises holds the ontogeneticthesis that collective wholes, including society at large, are the products ofhuman action, which in turn are the outcomes of ideas held by individuals, sothat, indirectly, society is the result of the ideas of individuals.

114 Austrian methodological individualismFriedrich von HayekThe Austrian branch – actually it is more like the root and trunk – of methodologicalindividualism culminates in the writings of Friedrich August von Hayek.His particular version of it is a development of the ideas of Max Weber, AlfredSchutz and of his teacher Ludwig von Mises, but it owes, perhaps, even more tothe original version of Carl Menger. With Hayek, the circle of Austrian methodologicalindividualism comes to a close. 30 It is a form of methodologicalindividualism, which is less radical than that of the theory of the social contractand also of mainstream neoclassical economics. Austrian methodological individualismis based on the premise that humans are social and cultural beings (cf.Madison, 1990: and Lange-von Kulessa, 1997).In his inaugural lecture at the London School of Economics in 1933, Hayekintroduced some of the main themes of his future work (cf. Caldwell, 1988). Oneof them was his version of methodological individualism: ‘By combiningelementary conclusions and following up their implications he [the economist]gradually constructs, from the familiar elements, a mental model which aims atreproducing the working of the economic system as a whole’ (Hayek, 1933b:128). The year after, he wrote an introduction to a reprint of Menger’s Grundsätzeder Volkwirtschaftslehre published by the London School of Economics. This piecewas also published as an article in Economica (1934) presenting Carl Menger to anEnglish public. A point to notice, because unusual, is that Hayek devotes considerablespace also to Menger’s methodological work, Problems of Economics andSociology, and maintains that it was ‘hardly less an achievement than theGrundsätze’ (Hayek, 1934: 405). ‘Of the central contentions of the book one maybe singled out for further comment; his emphasis on the necessity of a strictlyindividualistic or, as he generally says, atomistic method of analysis’ (p. 406). Ifmethodological individualism is now separated from ethical individualism and ifthe subjective element has been fully developed only by Austrian Economics,‘this is largely due to Menger’s brilliant vindication in this book’ (p. 406).In 1937 Hayek’s now classic article ‘Economics and Knowledge’ appeared: itquestioned not only the theory of general equilibrium, but also Mises’s praxeology,and eventually led to a split within the Austrian movement. Hayek arguedthat equilibrium analysis and praxeology are both tautological and fail to sayanything about reality. The general remedy for this is to work with verstehen andideal types, as suggested by Max Weber (Hayek, 1948: p. 47, n. 12 and p. 52, n.18). A particular problem is the assumption of isolated individuals, which fails toaccount for the interdependence of people’s plans. For equilibrium to be possiblein the real world, individuals would have to know the plans of other individuals.Game theory is one attempt to deal with this interdependence, but it does notsolve the problem of knowledge (see p. 252). 31The main source for an understanding of Hayek’s ideas on methodologicalindividualism is ‘Scientism and the Study of Man’, first published as articles inEconomica (1942–4) and later included as Part One in The Counter-Revolution ofScience (1955). 32 This work is first of all an attack on ‘scientism’; the mechanicaland uncritical application of habits of thought, taken from the natural sciences,

Austrian methodological individualism 115in the social sciences. In contrast to the objective character ofphenomena, Hayek stresses the subjective nature of human actions.naturalIn fact, most of the objects of social or human action are not ‘objectivefacts’ in the special narrow sense in which this term is used by the Sciencesand contrasted to ‘opinions’, and they cannot at all be defined in physicalterms. So far as human actions are concerned the things are what the actingpeople think they are.(Hayek, 1942–4: 277f; 1955: pp. 26f)Hayek’s subjectivism derives from his Austrian view of economics, but also fromGerman hermeneutics and phenomenology. Therefore, since all human action issubjective in the sense of the above quotation, every science dealing with theresults of conscious human action must follow the course taken by economics.The points which we want to stress are that in all such attempts we muststart from what men think and mean to do, from the fact that the individualswhich compose society are guided in their actions by a classification ofthings and events in a system of sense qualities and concepts which has acommon structure and which we know because we, too, are men, and thatthe concrete knowledge which different individuals possess will differ inimportant respects. Not only man’s action towards external objects but alsoall the relations between men and all social institutions can be understoodonly in terms of what men think about them.(Hayek, 1942–4: 283; 1955: 33)So far, Hayek’s concern is with our understanding of human action. His aimseems to be the justification of a subjectivist methodology by epistemologicalarguments. Hayek is talking about knowledge. But his subjectivism goes deeperand turns out to be of an ontological nature, like that of Mises. Not only is ourknowledge of society limited to knowledge about the beliefs of individuals.Society, itself, is made up of the beliefs of individuals. This is the reason knowledgeof society is knowledge about the beliefs of individuals.Society as we know it is, as it were, built up from the concepts and ideas heldby the people; and social phenomena can be recognized by us and havemeaning to us only as they are reflected in the minds of men.(Hayek, 1942–4: 283; 1955: 33f)This does not imply, however, that the aim of social science is limited to investigatingthe minds of individual human beings. Social science is interested insocial structures, but the true elements of social structures are beliefs existing inthe minds of a plurality of individuals.

116 Austrian methodological individualismThe structure of men’s minds, the common principle on which they classifyexternal events, provide us with the knowledge of the recurrent elements ofwhich different social structures are built up and in terms of which we canalone describe and explain them. While concepts or ideas can, of course,exist only in individual minds, and while, in particular, it is only in individualminds that different ideas can act upon another, it is not the whole of theindividual minds in all their complexity, but the individual concepts, theviews people have formed of each other and of things, which form the trueelements of the social structure.(Hayek, 1942–4: 283f; 1955: 34)In order to illustrate his subjectivism, Hayek gives some examples from thesocial sciences. The first example, presumably taken from functionalist anthropology,concerns the concepts of ‘tool’ and ‘instrument’. Such concepts – that of‘hammer’, for instance – cannot be defined by their physical attributes, but mustbe defined in terms of their function, or the ‘purpose for which men think theycan be used’ (1942–4: 278; 1955: 27). From archaeology Hayek takes theexample of the problem of determining ‘whether what looks like a stone implementis in truth an “artifact”, made by man, or merely a chance product ofnature’. According to Hayek, the only way to answer this question is ‘by trying tounderstand the working of the mind of prehistoric man, of attempting to understandhow he would have made such an implement’ (1942–4: 278; 1955: 27). Afinal example concerns a problem that belongs to economics, but would occur asa problem only in economic history or economic anthropology.That the objects of economic activity cannot be defined in objective termsbut only with reference to a human purpose goes without saying. Neither a‘commodity’ or an ‘economic good’, nor ‘food’ or ‘money’, can be definedin physical terms but only in terms of the views people hold about things.Economic theory has nothing to say about the little round disks of metal aswhich an objective or materialist view might try to define money. It hasnothing to say about iron or steel, timber or oil, or wheat or eggs as such.The history of any particular commodity indeed shows that as humanknowledge changes the same material thing may represent quite differenteconomic categories. Nor could we distinguish in physical terms whethertwo men barter or exchange or whether they are playing some game orperforming some religious ritual.(Hayek, 1942–4: 281f; 1955: 31)Now, the subjectivism of Hayek should not be interpreted as implying thatthe social scientist must use only such concepts as are used in everyday life, andin the same way. Hayek is careful to guard himself against such misunderstandingof his standpoint (1942–4: 284f; 1955: 36). The ideas and conceptsused in everyday life are only the starting-point, the elements from which thesocial scientist gradually builds up his model of the social structure. The reason

Austrian methodological individualism 117why the social scientist must not and cannot confine himself to the use of suchconcepts as guide individuals in their acting, is that social phenomena are ‘theunintended or undesigned results of the actions of many men’ (1942–4: 276;1955: 25).In order to make clear his own idea of a scientific model of social structure, asdistinguished from a naive collectivist approach, Hayek makes a distinctionbetween, on the one hand, those ideas which motivate people in their actions andwhich are constitutive of social phenomena and, on the other hand, those speculativeideas people form about these phenomena in order to explain them (1942–4:285; 1955: 37). He is aware that the same idea may occur in both capacities, butinsists that the social scientist starts his analysis from ideas in their constitutivecapacity, and takes speculative ideas for what they are; popular myths.It is the ideas which the popular mind has formed about such collectives as‘society’ or the ‘economic system’, ‘capitalism’ or ‘imperialism’, and othersuch collective entities, which the social scientist must regard as no morethan provisional theories, popular abstractions, and which he must notmistake for facts. That he consistently refrains from treating these pseudoentitiesas ‘facts’, and that he systematically starts from the concepts whichguide individuals in their actions and not from the results of their theorizingabout their actions, is the characteristic feature of that methodological individualismwhich is closely connected with the subjectivism of the socialsciences.(Hayek, 1942–4: 286; 1955: 37f)The subjectivism and individualism of the social sciences make them in animportant respect different from the natural sciences. The difference is epistemologicaland Hayek uses it in support of his methodological individualism. In thenatural sciences complex phenomena are observed; material things which, fortheir explanation, must be analysed into their component parts. Their method,therefore, is ‘analytic’. In the social sciences the opposite is the case. We know bydirect acquaintance only the parts; the individuals, from which the complexphenomena must be constructed. The method of the social sciences, therefore is‘compositive’, or ‘synthetic’ (1942–4: 287; 1955: 39). The complex phenomenato which the social scientist must give his sole attention are those phenomenawhich are the unintended results of the conscious action of many men.Consciously designed social phenomena, though sometimes in need of a psychologicalexplanation, are in no need of a social scientific explanation. Hayek givesan example of how order is produced by the actions of many men without beingintended by any one of them.The way in which footpaths are formed in a wild broken country is such aninstance. At first everyone will seek for himself what seems to him the bestpath. But the fact that such a path has been used once is likely to make iteasier to traverse and therefore more likely to be used again; and thus

118 Austrian methodological individualismgradually more and more clearly defined tracks arise and come to be used tothe exclusion of other possible ways. Human movements through the regioncome to conform to a definite pattern which, although the result of deliberatedecisions of many people, has yet not been consciously designed byanyone. This explanation of how this happens is an elementary ‘theory’applicable to hundreds of particular instances … It makes no difference forour present purpose whether the process extends over a long period of timeas it does in such cases as the evolution of money or the formation oflanguage, or whether it is a process which is constantly repeated anew as inthe case of the formation of prices or the direction of production undercompetition.(Hayek, 1955: 40; cf. 1942–4: 289)This example shows the imprint of Menger on the methodology of Hayek. It isan attempt to illustrate the ‘exact’ and atomistic analysis of institutions suggestedby Carl Menger in Problems of Economics and Sociology and used by Menger in histheory of money (see pp. 90f ). The particular example of the footpath, however,is borrowed from Descartes. 33Like Mises, Hayek sees a clear connection between individualism and nominalism,and collectivism and realism, respectively. Methodological collectivism isan outgrowth of conceptual realism. It makes the mistake of treating collectiveentities, such as society, economy, capitalism, or class as definite objects when, infact, they are ‘models constructed by the popular mind to explain the connectionbetween some of the individual phenomena which we observe’. The fundamentalobstacle to the assumption made by naive realism is that ‘the wholes assuch are never given to our observation but are without exception constructionsof our mind’ (1942–4: 43; 1955: 54). But as with Mises, there is a certain ambiguityon this point. Hayek denies that collective entities exist as definite objects,but at the same time takes the view that wholes – the market, for instance – existas structures of relationships (1942–4: 44; 1955: 55). The fallacy Hayek wants toavoid is also partly that of Mises. He wants to avoid the danger, inherent incollectivism, of treating collective entities as some kind of ‘superpersons’endowed with ‘social’ or ‘collective’ minds (1942–4: 45; 1955: 57).In support of his view that collective entities are not definite objects, Hayekpoints to the fact that social wholes are not ‘natural units’ like flowers and butterflies;unlike physical things, they are not continuous in space and time, but‘selections of certain elements of a complex picture on the basis of a theoryabout their coherence’ (1942–4: 43; 1955: 55).The social sciences, thus, do not deal with ‘given’ wholes but their task is toconstitute these wholes by constructing models from the familiar elements –models which reproduce the structure of relationships between some of themany phenomena which we always simultaneously observe in real life.(Hayek, 1942–4: 44; 1955: 56)

Austrian methodological individualism 119Hayek’s methodological individualism is also expounded in an article fromthe same period, ‘The Facts of the Social Sciences’ (1943), which is a concisestatement of his views on the methodology of the social sciences. Social facts,including not only human actions and written language, but also ‘things’ such astools, food, medicine and weapons, are subjective in contrast to the objectivecharacter of natural facts. ‘Collective wholes’ are not observable, and do notexist as definite objects, independent from the ‘theories’ people have about them.Collective wholes are ‘created by an act of construction or interpretation’. Termsdenoting collective wholes refer to certain activities of individuals.This is all the theories of the social sciences aim to do. They are not aboutthe social wholes as wholes; they do not pretend to discover by empiricalobservation laws of behaviour or change of these wholes. Their task israther, if I may so call it, to constitute these wholes, to provide schemes ofstructural relationships which the historian can use when he has to attemptto fit together into a meaningful whole the elements which he actually finds.(Hayek [1948] 1972: 72)In this quotation, methodological individualism is opposed to the search forempirical laws about social wholes. This is another return to Carl Menger, whoopposed the atomism of the exact orientation to the collectivism of the realisticempiricalapproach. But it is also an adjustment to the views of Weber andMises. Hayek does not accept a social science based on the realistic-empiricalapproach. This type of research belongs to history. But social science is an aid tohistory by providing individualistic models of collective wholes. Or, in the terminologyof Max Weber, sociology is an auxiliary discipline to history, providingindividualistic ideal types of collective phenomena.In its early development, especially in Weber and Schumpeter, methodologicalindividualism was clearly distinguished from political individualism. InMises, this distinction is totally blurred. His allegiance to value-freedom is merelip-service. Methodological individualism appears as inseparable from, sometimeseven identical with, the political doctrine of liberalism. But Hayek is firstamong the proponents of methodological individualism to treat it as part of awider system of ideas, something like a world view, including, or related to, acertain ideology. In his article ‘Individualism: True and False’ (1945), Hayek iscareful to distinguish the ‘true individualism’ of the Scottish Enlightenment, andmen such as Alexis de Tocqueville and Lord Acton, from the ‘false individualism’,which originates with Réné Descartes. The most significant expressions ofthe latter are the philosophy of the French Enlightenment, especially Jean-Jacques Rousseau, and the English philosophical radicalism of Jeremy Benthamand the two Mills. The most mistaken element of ‘false individualism’, is the‘design theory’ of social institutions; the belief that society can be constructedaccording to a plan, or blueprint.

120 Austrian methodological individualismWhat, then, are the essential characteristics of true individualism? The firstthing that should be said is that it is primarily a theory of society, an attemptto understand the forces which determine the social life of man, and only inthe second instance a set of political maxims derived from this view ofsociety … its basic contention … is that there is no other way toward anunderstanding of social phenomena but through our understanding of individualactions directed toward other people and guided by their expectedbehavior. This argument is directed against the properly collectivist theoriesof society which pretend to be able directly to comprehend social wholeslike society, etc., as entities sui generis which exist independently of the individualswhich compose them.(Hayek [1948] 1972: 6)The main insight of true individualism is that most social phenomena are theunintended and unforeseen results of the actions of individuals; that like livingorganisms they have grown, step by step, without being consciously aimed at andconstructed. This insight is used by Hayek, as by many other conservatives andliberals, to argue against social planning and social engineering. But Hayek hasprobably done more than anyone else to develop this argument in great detail,including the theory of society on which it is based. Like Menger, Hayek distinguishestwo types of order: organisations and spontaneous orders (organisms).The main part of Hayek’s theory is about spontaneous orders, and the way theyhave developed in the course of history. The most important contributions tothis theory are from the 1960s onwards. From about this time, Hayek also stopsmentioning methodological individualism in his writings. I interpret this fact asan indication that Hayek no longer adheres to this principle, and for a goodreason. His theory of society is no longer compatible with methodological individualism(see pp. 282–4).To sum up: Hayek’s methodological individualism is, like that of Weber andMises, inseparable from subjectivism. Society exists only in the minds andactions of individual human beings. Social wholes do not exist as definite objects,on a par with natural units, like physical things and biological organisms.Concepts such as ‘society’, ‘capitalism’ and ‘imperialism’ are popular abstractions,and the concepts used by social science, such as ‘government’, ‘trade’ or‘army’, are theoretical constructions. But neither the collective concepts ofeveryday life, nor those of social science, denote collectives that are real, independentlyof any idea, or theory, about them. I think it is possible to give thefollowing explication of Weber’s, Mises’s and Hayek’s individualist analysis ofcollective concepts: their intension is ideas in people’s mind, their extension is theactions of individuals.But the subjective meaning individuals attach to actions and things is only thepoint of departure. The real task of social science is to construct and explainlarge-scale social (complex) phenomena. This is done by constructing models ofindividual actions and their intended and unintended consequences. The subjectivemeaning individuals attach to their actions (and to physical things) are the

Austrian methodological individualism 121elements out of which the social scientist constitutes ‘social wholes’. Hayek’smethodological individualism is both procedural and substantive. The individualis the point of departure, but also the end of social science. The ultimate explanationof all social phenomena is in terms of individuals, their actions and theconsequences of these actions. Like Menger, Hayek is especially interested in theevolution of social institutions, which he tends to see as conventions, in themanner of Hume.The Austrian influenceAlthough part of neoclassical marginalism, Austrian Economics has never reallybeen part of mainstream economics. After the first generation, it disappeareddefinitely from the mainstream and formed a small marginal stream all of itsown. At an early stage of its development, however, it made an impact on someimportant non-Austrian economists, who helped to propagate methodologicalindividualism. I am going to end this chapter by mentioning three of them:Frank H. Knight, Lionel Robbins and the early Gunnar Myrdal.Frank H. Knight was an American economist and a founding father of thefamous Chicago School of Economics, which counts among its members suchprominent economists as George J. Stigler, Milton Friedman, Armen A. Alchian,Harold Demsetz, Richard Posner and Gary S. Becker. Knight was clearly influencedby all three versions of marginalism. He mentions Wicksteed, Pareto andMarshall, as especially important, but there is little doubt that also the Austriansloom large in his writings. His emphasis on uncertainty and his view thateconomics is an exact science, for instance, indicate an Austrian influence. 34In his classic treatise Risk, Uncertainty and Profit (1921), Frank H. Knight wasfirst to make the important distinction between risk and uncertainty, where theformer implies knowledge of the probabilities of different alternative outcomes,whereas the latter means that we are left completely in the dark concerningfuture states of affairs. According to Knight, it is the existence of uncertaintythat gives rise to profit. If economic life were characterised by perfect knowledge,or even by risk, competition would eliminate profits.Despite the obvious problem of defining rationality in a state of uncertainty,Knight may have been the first to explicitly suggest that economics is a theory of‘rational choice’ (Knight, 1921: 89). 35 Following Wicksteed, he maintains thateconomics is a theory of choice, not of pleasure and pain. The term ‘rationalchoice’ is probably introduced to avoid the psychologism implicit in Wicksteed’sexpression ‘psychology of choice’. The theory of conduct, relevant foreconomics, is the ‘Law of Choice’, stated by Knight in the following way:In the utilization of limited resources in competing fields of employment,which is the form of all rational activity in conduct, we tend to apportionresources among alternative uses that are open in such a way that equalamounts of resource yield equivalent returns in all fields.(Knight, 1921: 65)

122 Austrian methodological individualismThis is the Austrian view of economics, stated most succinctly by Weber andlater by Lionel Robbins. It is possible to recognise an Austrian element also inKnight’s distinction between a Robinson Crusoe and a social economy.Economics is a theory of choice and exchange, but not of choice alone. Thisview is stated most clearly in The Economic Organization ([1933] 1971; see also1961: 185–93, 273–82), where Knight draws the obvious conclusion thateconomics is limited in scope to systems of free market economics, but isrepeated in The Ethics of Competition (1935/6).In its second aspect or stage, economic theory deals with social, in the senseof interindividual, relations. But the ‘economic man’ is not a ‘social animal’,and economic individualism excludes society in the proper human sense.Economic relations are impersonal. The social organization dealt with ineconomic theory is best pictured as a number of Crusoes interactingthrough the markets exclusively.(Knight, 1935/6: 282)Knight was not in any sense a radical individualist even though he did favouran individualistic organisation of the economy. He was fully aware of the significanceof social institutions and social organisation for an understanding of socialphenomena in general, and of economic phenomena in particular (see, e.g.,Knight 1956: chs 1 (1951) and 5 (1941). It is difficult to tell, therefore, whetherKnight was a programmatic methodological individualist, or not; and if so, whatkind of methodological individualist he was. One thing is certain, however: likeSchumpeter, he did conceive of theoretical economics as an individualisticscience.By the nature of its fundamental conceptions, theoretical economics is anindividualistic science. As explained at length in an earlier part of this essay,the ‘economic man’ is not a social man, and the ideal market dealings oftheory are not social relations. The science takes its economic individual as adatum, in his three aspects of wants, resources and technical knowledge,ignoring all questions of origin; and it abstracts from all his social relationswith other human beings, except those of the perfect market, which arereally relations to commodities as such.(Knight, 1935/6: 337)Lionel Robbins was an English economist working at the famous LondonSchool of Economics in its most glorious days, from 1929 to 1961. He was aneconomist of somewhat similar bent as Frank Knight, but even more influencedby the Austrians and by Wicksteed. Robbins was the author of one of the mostcited books on economic theory and methodology, his classic An Essay on theNature and Significance of Economic Science (1932). In this book you find, I am sure,the most often cited definition of economics ever produced: ‘Economics is thescience which studies human behaviour as a relationship between ends and

Austrian methodological individualism 123scarce means which have alternative uses’ ([1932] 1935: 17). It may be addedthat, in this situation, ‘behaviour necessarily assumes the form of choice’ (p. 14).Robbins made no secret of the fact that this is the Austrian conception ofeconomics, or at least half of it. Robbins did not follow most of the Austriansand Frank Knight in conceiving of economics as a theory of exchange, or catallactics.Without this limitation, economics becomes a general theory of humanaction applicable to various social phenomena other than market exchange.Robbins was thus a precursor of theoretical economic imperialism, and alsomade some suggestions in this direction: ‘There is an important sense in whichthe subject-matter of political science can be conceived to come within the scopeof our definition of the economic. Systems of government, property relationships,and the like, can be conceived as the result of choice’ (Robbins [1932]1935: 134.). Unlike some later proponents of economic imperialism, however,Robbins was keenly aware also of the limitations of the economic approach (pp.131–5; see also Udehn, 1991; 1996).Robbins was one of the main figures in the project to free economics from itsdependence on psychological hedonism and turn it into a theory of rationalchoice. Economics does not in any way indulge in speculation about the psychological‘causes’ of human action. It takes people’s motives and wants as givenand treats them as data in its own theories. The psychological content ofeconomics boils down to two simple assumptions: (1) ‘that individuals canarrange their preferences in an order, and in fact do so’ (pp. 78f); and (2) ‘thateach final choice is consistent with every other, in the sense that if I prefer A to Band B to C, I also prefer A to C’ (p. 92). This is the meaning of the term‘rational’ in economics (p. 91).Another important strain in Robbins’s thought, is his (individualist) denial ofthe possibility of interpersonal comparisons of utility. ‘There is no means oftesting the magnitude of A’s satisfaction as compared with B’s’ (pp. 139f). Hence,the impossibility of a scientific concept of ‘social utility’. The only possible significanceof such a concept is as part of ethical theory, which must be clearlydistinguished from the positive theory of economics. ‘It is simply the accidentaldeposit of the historical association of English Economics with Utilitarianism:and both … will be the better if this is clearly recognised’ (p. 141).There are many similarities between Robbins’s An Essay on the Nature andSignificance of Economic Science (1932) and Gunnar Myrdal’s The Political Element inthe Development of Economic Theory (1929), even though they came to differentconclusions and Myrdal, eventually, took off in another direction. 36 Myrdal waseven more critical, than Robbins, of the dependence of early Britishmarginalism on utilitarian psychological hedonism and on the idea of ‘socialvalue’, or ‘social utility’, as a sum of the utilities of individuals. Such addition ofindividual utilities is only possible by the scientifically illicit procedure ofcomparing the utilities of different individuals.The utilitarian economists were individualists, but they were also radicals andsocial reformers. As such they had a strong inclination to develop their doctrinein a politically relevant direction. By adding the utilities of all members of

124 Austrian methodological individualismsociety it was possible to arrive at the politically relevant notion of ‘social value’or ‘social utility’, which is at the basis of welfare economics. From the very beginning,methodological individualists have been sceptical about welfare economics,and especially about notions such as ‘social value’ or ‘social utility’. A first objectionis that it presupposes the possibility of comparing the utilities of differentindividuals. The most serious defect of concepts such as ‘social value’ and ‘socialutility’, however, is that they seem to suggest some kind of collective subject,engaged in ‘collective’, or ‘social housekeeping’ (Myrdal [1929] 1953: ch. 6).There is a clear echo of Carl Menger’s critique of the concept of ‘nationaleconomy’ in Myrdal’s objections to the concepts of ‘social value’ and ‘socialutility’. The worst sinners are, of course, the German political metaphysicians,with their organicism and their speculation about the essence of the state. Butalso the utilitarian economists – especially when they draw political conclusionsfrom their doctrine – ‘tend to be forced into an untenable “communistic fiction”about the unity of society’ (p. 54).Like Robbins, Myrdal observes that economics has progressively freed itselffrom its dependence upon psychological hedonism, and from any other psychologicaltheory for that matter. Economics has become a ‘behaviourist’ theory ofchoice (Wahlhandlungstheorie). Unlike Robbins, however, Myrdal did not at allwelcome this development. From his point of view, a ‘behaviourist theory’ ofchoice, ‘is an attempt to construct a subjective theory of value without psychologicalcontent. Such an attempt is bound to lead to empty mathematicalscholasticism’ ([1929] 1953: 97f ). For Myrdal, the solution is not to reintroducepsychological hedonism in the science of economics – this theory is provenwrong by the science of psychology – but to find some better psychologicaltheory to replace it. Myrdal, then, appeared as one of the first in a long line ofcritics, who have tried to reintroduce the programme of psychologism ineconomics, but in vain.ConclusionThe Austrian School of Economics is the single most important source ofmethodological individualism in modern social science. Its founder Carl Mengerdefended the use of an atomistic method in theoretical economics againstmembers of the German Historical School and Joseph Schumpeter coined theterm ‘methodological individualism’ to refer to this method. An important partof Menger’s individualist programme was his suggestion that not only prices, butall social institutions are amenable to individualist analysis.A possible problem with this extension of the individualist programme hasbeen pointed out by Robert Nozick, who notices that Menger provides an explanationof the creation and maintenance of institutions, but adds the seeminglyobvious point that ‘[e]xisting institutions also shape and affect the actions ofindividuals’ Nozick, 1977: 357). According to Nozick, therefore, the interplay ofindividuals and institutions should rather look like this:

Austrian methodological individualism 125Figure 4.2 Actions and institutions according to Robert NozickSource: Nozick (1977: 358)The pertinent question, now, is this: ‘If each time we explain the current situationas arising from previous actions in a certain institutional setting, then whyare actions prior? In this apparent chicken and egg situation, why aren’t weequally methodological institutionalists?’ As we shall see in chapter 6 onPopperian methodological individualism, Nozick was not first to ask this question,but he may have been first to ask it explicitly about Menger’smethodological individualism.I do not know if the later Austrians, Weber, Mises and Hayek, were aware ofthis problem, or saw it as a problem, but they did suggest a way of escape. WhileNozick implies that social institutions are irreducible to individuals and theiractions, Weber, Mises and Hayek adopted an intersubjectivist theory of society,which situates social institutions in the minds of individuals. Social institutionsare the subjective meaning individuals attach to social actions or social thingslike money. Instead of conceiving of individual and society as two levels, we get ahorizontal, or flat picture of individuals interacting with one another. Since thereis a shift from methodology to ontology involved in this common route ofescape, I call it the ‘ontological turn’ or ‘twist’.Figure 4.3 Individuals in society according to Weber, Mises and Hayek

5 Society as subjectivelymeaningful interactionIn the beginning, sociology was mainly holistic. Most sociologists saw society as areality sui generis, with laws of its own, irreducible to the acting individuals. Butthere were also some early sociologists, who took a more individualistic view ofsociety: Gabriel Tarde, George Simmel and Max Weber, for instance. The earlySimmel and Weber conceived of society as constituted by the subjectively meaningfulaction of individuals. The most important sources of this conception ofsociety were German Völkerpsychologie and the descriptive psychology of WilhelmDilthey, who also saw society as interaction. Simmel, who, like Dilthey, had intellectualroots in Völkerpsychologie, saw society as both subjective meaning andinteraction. Borrowing a distinction from Kant, he saw society as both form andcontent. Society consists of forms of interaction, but its content is subjectivemeaning. This is, I believe, the main sociological version of the individualistictheory of society. It differs from the theory of the social contract and theeconomic theory of the market by taking a broader view of social action, but,above all, by its interest in the meaning of social phenomena. 1In the further development of the individualist theory of society in sociology,there was an important input from two philosophical schools: American pragmatismand German phenomenology. Of these, the former was not reallyindividualistic, but nevertheless it exerted a profound influence on one particularversion of the individualist theory of society, namely symbolic interactionism.The philosophical phenomenology of Edmund Husserl, finally, is the source, notonly of phenomenological sociology, but also of existentialism – which is atheory somewhere in between philosophy, psychology and sociology – and ofethnomethodology.Symbolic interactionism, sociological phenomenology, existentialist sociology,and ethnomethodology are all sociological theories, which conceive of society asinteraction and intersubjective meaning, even if they do not attach the sameimportance to both elements. Because these theories typically focus on interactionbetween human beings, in co-presence, or face-to-face relations, they areusually considered as ‘microsociology’ (Knorr-Cetina, 1981: 1ff) or as belongingto the ‘microinteractionist tradition’ (Collins, 1985: ch. 3).According to the American sociologist Jeffrey Alexander, the theoriesbelonging to microinteractionist tradition are caught in an ‘individualist

Society as subjectively meaningful interaction 127dilemma’. I agree with Alexander (and Parsons) that there is such an individualistdilemma, but my interest, in this chapter, is in individualism per se, withoutthe dilemma.Symbolic interactionismSymbolic interactionism is an American movement of thought usually associatedwith the name of George Herbert Mead, but really a creation of his pupil andfollower, the sociologist Herbert Blumer. A philosopher and social psychologist ofthe pragmatist school, Mead was influenced by the work of Charles Peirce,William James and John Dewey. 2 Also important for Mead’s theory of the individualand society were some ideas of the sociologist Charles Horton Cooley. Athird influence – in this case it is probably better to talk about cross-fertilisation –was the so-called Chicago School of Sociology, founded by William I. Thomasand Robert E. Park. Although mainly an American product (Alexander, 1987:196ff; Joas, 1987: 94ff), symbolic interactionism was also influenced by Europeanthinking. Among the more remote sources of symbolic interactionism we findAdam Smith and other representatives of the Scottish Enlightenment, theGerman Romantics and Hegel. 3 A more immediate influence came fromVölkerpsychologie, Dilthey and Simmel. 4I take Cooley, Thomas and Mead to be the most important progenitors ofsymbolic interactionism. I dispute, however, that any one of them was a‘symbolic interactionist’, in the present-day sense of that term. I also doubt thatany one of them held the intersubjectivist theory of society, or was a theoreticalindividualist (cf. Joas, 1987: 94ff). 5 I do agree, however, that all of them developedcertain ideas, which, if taken in isolation, were important contributions tothe further development of symbolic interactionism into an individualist theoryof society.Cooley’s contribution to symbolic interactionism was his theory of thelooking-glass self and the theory of society that it suggests. The looking-glass selfhas ‘three principal elements: the imagination of our appearance to the otherperson; the imagination of his judgement of that appearance, and some sort ofself-feeling such as pride and mortification’ (Cooley [1902] 1964: 184). Animportant part of our selves, then, is the way we imagine we appear in other peoples’minds. From this, it is but a short step to suggest that people exist primarily or,exclusively, as ideas in the minds of one another.Cooley’s conception of society, as advanced in his first and most influentialbook Human Nature and Social Order (1902), is probably the most extreme version ofan idealist and intersubjectivist theory of society ever proposed by a sociologist.Society, according to Cooley, is mental: ‘The immediate social reality is the personal idea… Society, then in its immediate aspect, is a relation among personal ideas … Societyexists in my mind as the contact and reciprocal influence of certain ideas named‘I’ Thomas, Henry, Susan, Bridget, and so on’ ([1902] 1964: 119). A consequenceof this view is ‘that there is no separation between real and imaginarypersons; indeed to be imagined is to become real, in a social sense’ (p. 95).

128 Society as subjectively meaningful interactionAnother consequence is that corporeal existence is immaterial for the question ofsocial existence. Historic persons, characters of fiction and mythological figuresare as real as corporeal persons, since they exist as ideas in the minds of people.Indeed, they may be more real, sincea corporeally existent person is not socially real unless he is imagined. If thenobleman thinks of the serf as a mere animal and does not attribute to hima human way of thinking and feeling, the latter is not real to him in thesense of acting personally upon his mind and conscience.(Cooley [1902] 1964: 123)W.I. Thomas shared Cooley’s subjectivism, at least in methodology. His maincontribution to symbolic interactionism was his stress upon the importance forsocial scientific investigation of taking account of the individual’s own definitionof his situation. ‘There is always a rivalry between the spontaneous definitions ofthe situation made by the member of an organized society and the definitionswhich his society has provided for him’ (Thomas [1923] 1967: 42).On the other hand, this definition of the situation is from one standpoint asgood as if it were true. It is a representation of the situation as appreciatedby the subject, ‘as if ’ it were so, and this is for the behavior study the mostimportant phase of reality.(Thomas [1927] 1966: 160)The subjectivist strain in Thomas’s thought is best epitomised in his famousdictum, sometimes called Thomas’s theorem; ‘if men define situations as real,they are real in their consequences’ (Thomas and Thomas, 1929: 572). Thomas’subjectivism, then, was more methodological than that of Cooley. The reason forthis is probably that he was much more engaged in empirical investigation thanthe latter. His most famous study is The Polish Peasant in Europe and America (1918),written together with the Polish sociologist Florian Znaniecki. In the methodologicalintroduction to this work, they state a methodological principle which,according to them, is fundamental both in social psychology and sociology: ‘Thecause of a social or individual phenomenon is never another social or individual phenomenonalone, but always a combination of a social and an individual phenomenon’ ([1918] 1927:44). This principle is of course written as a direct reply to Durkheim’s famousdictum (see p. 34), which denies the role of individual phenomena and individualpsychology altogether. Thomas and Znaniecki opt for a less extreme alternative,which recognises the need for microfoundations, but not reduction, or, at least,only a partial reduction of sociology to psychology.George Herbert Mead’s main contribution to symbolic interactionism is hissocial psychological theory of the self and the main source for this theory isMind, Self and Society (1934). It must be counted as an original and importantcontribution to social psychology, even if it was foreshadowed by Adam Smith,William James and Charles Horton Cooley. Mead’s theory deals with the rise of

Society as subjectively meaningful interaction 129the self, both phylogenetically, in the human species, and ontogenetically, in thehuman individual, but above all, ontogenetically. As such, it is a theory aboutsocialisation. The main contention of this theory is that the self can only arise ina human society. In very brief outline Mead’s theory of the self amounts to this:The self arises in communication by means of significant symbols, i.e.,symbols that evoke the same kind of response in the self and in the other. Therise of the self, thus, presupposes a universe of discourse, a system of common orsocial meanings, that is, a common language. The self arises in the individual byhis/her taking the attitude, or the role, of the other. It is only by taking the roleof the other that the individual can become an object to him/herself, and it isonly by becoming an object to him/herself that the individual acquires a self.Mead distinguishes two stages in the development of a social self in the child.The first stage is that of play, where the individual takes the role of a particularor ‘significant other’; a mother, a father, a policeman or a teacher. The secondstage is that of game, where the child takes the role of all the others taking partin a certain activity. A game is an organised activity governed by a set of rules.Mead’s term for all others involved in a group activity is the ‘generalised other’.More generally, the generalised other is the form in which the individualconfronts society. 6Closely related to the concept of the ‘generalised other’ is Mead’s distinctionbetween the ‘I’ and the ‘me’. Both the ‘I’ and the ‘me’ are selves, or rather,belong to the self, as phases of the undivided self. ‘The “I” is the response of theorganism to the attitudes of the others; the “me” is the organized set of attitudesof others which one himself assumes’ ([1934] 1962: 175).This process of relating one’s own organism to the others in the interactionsthat are going on, in so far as it is imported into the conduct of the individualwith the conversation of the ‘I’ and the ‘me’, constitutes the self.(Mead [1934] 1962: 179)The ‘I’, then, is the spontaneous element in action. In most cases the ‘I’ isaware of the ‘me’ and acts according to its strictures, but it sometimes happensthat it acts entirely on its own, as when we are surprised by our own actions. The‘I’ exists in the specious present. As soon as an act is completed and becomespart of the past, as soon as we can reflect upon it, it sinks down into the ‘me’.The ‘me’ is the conventional, the habitual, the institutional element in action.The ‘me’ sets limits to the actions of the ‘I’ and is thus the means for the exerciseof social control in the individual. The ‘me’ is then, it would seem, the internalisedgeneralised other. 7Mead’s theory of the self could be interpreted as implying an individualistand intersubjectivist theory of society. This is the interpretation made by thesymbolic interactionists. But it is not the only possible interpretation and it ishardly the correct interpretation. In the recent literature about Mead andsymbolic interactionism, it has been argued convincingly that Mead’s socialbehaviourism was not individualistic. According to Lewis and Smith (1980: ch. 5),

130 Society as subjectively meaningful interactionMead was a social realist, who saw society as an objective reality sui generis. AsJeffrey Alexander has pointed out, Mead saw symbolic structures and social institutionsas, in a sense, prior to the actions of individuals (see Alexander, 1987:205–14). This view is shared by Wiley (1994: chs 2 and 6), who argues that Meadhad a semiotic theory of the self, which means that it is irreducible to the subjective,or psychic, level. It may be added that Mead frequently referred to societyas a whole, and not just to a meaningfully integrated whole, but also a functionallyintegrated whole. The implication is that also the method of socialpsychology must be holistic:We are not in social psychology, building up the behavior of the social groupin terms of the behavior of the separate individuals composing it; rather weare starting out with a given social whole of complex social group activity,into which we analyze (as elements) the behavior of each of the separateindividuals composing it. We attempt, that is, to explain the conduct of theindividual in terms of the organized conduct of the social group, ratherthan to account for the organized conduct of the social group in terms ofthe conduct of the separate individuals belonging to it. For socialpsychology, the whole (society) is prior to the part (the individual), not thepart to the whole; and the part is explained in terms of the whole, not thewhole in terms of the parts.(Mead [1934] 1962: 7)This is a perfectly clear statement of methodological holism, taken fromMind, Self, and Society, and yet Mead has been interpreted as an individualist bythe symbolic interactionists. According to Anselm Strauss, for instance, Mead’ssocial psychology is a form of ‘socialized individualism’. But this interpretation ispossible only on the basis of a certain interpretation of the concept of the‘generalised other’. It is possible only if the generalised other, that is society, issituated in the self. This interpretation is far from self-evident, but it is the interpretationmade by most ‘symbolic interactionists’. According to Strauss ([1956]1964: xxiii), ‘The generalized other is society’s representative in the individual’.Mead, although undoubtedly its most important progenitor, is not the creatorof symbolic interactionism. Mead’s ideas found their way to the sociologistsmainly through the influence of Ellsworth Faris and, above all, through HerbertBlumer (Fischer and Strauss, 1979: 459). Of these, the former was closer toMead’s social behaviourism (Lewis and Smith, 1980: 170f, 187–9). 8 The realarchitect of symbolic interactionism is, I suggest, Herbert Blumer. It was Blumerwho coined the term ‘symbolic interactionism’ and who turned Mead’s socialpsychology into a ‘sociological’ theory of society; ‘society as symbolic interaction’(Blumer, 1969: ch. 3). 9 But Blumer’s symbolic interactionism is clearly differentfrom Mead’s social behaviourism, in important respects. Above all, it is certainlymore individualistic. 10In a presentation of social psychology as a discipline (1937), Blumer recognises,besides the instinct and stimulus-response approaches, a third important

Society as subjectively meaningful interaction 131view of human nature, represented by a group of social psychologists ‘who maybe conveniently labelled “symbolic interactionists”’. The theory of societyascribed to this group by Blumer is, in my terminology, typically ‘intersubjectivist’(p. 153).A third view of human group life is that held by those social psychologistswhom we have termed the symbolic interactionists. They, also, recognizethat the life of human groups presents itself in the form of a body ofcustoms, traditions, institutions, and so on, but they do not regard theseforms of culture as consisting merely of so many ways of acting. Instead,they believe that these forms of culture consist of common symbols, which aremutually shared and possessed by the members of the groups. Individualways of acting are alike because these individuals are guiding their behaviorby a symbol that they share in common. Thus, individuals wear clothing, inaccordance with the custom of their group, because each of them shares thecommon understanding that he is supposed to wear clothing. In the same way,any custom, folkway, or way of acting common to a group of individuals istraceable back to their common symbol or understanding.(Blumer, 1937: 158f)Contemporary symbolic interactionism is sometimes divided in two branches;the Chicago school, associated with the name of Herbert Blumer, and the Iowaschool centred around Manfred H. Kuhn (Manis and Meltzer, 1967: vi). 11 Ofthese schools, it is undoubtedly the Chicago school, including names such asEverett C. Hughes, Tamotsu Shibutani, Anselm Strauss and Howard S. Beckerwhich can make the most legitimate claim to calling itself ‘symbolic interactionist’(Charon, 1979: 289). 12 It is also the Chicago school which is mostrelevant for my purposes, since it was within this school that Mead’s socialpsychology was turned into an individualist and intersubjectivist theory ofsociety. 13 Having said this much, some qualifications are necessary. All Chicagointeractionists do not adhere to an intersubjectivist theory of society, nor doBlumer and his followers explicitly deny the existence of social wholes. It is rathera matter of most of them positively fostering an individualist and intersubjectivisttheory of society.In his article ‘Society as Symbolic Interaction’, Blumer postulates ‘that humansociety is made up of individuals who have selves’ (1969: 82). From this point ofdeparture, he goes on to attack sociological thought for reducing human beingsto organisms responding to various forces acting upon them, whether in the formof ‘social system’, ‘social structure’, ‘culture’, ‘status position’, ‘social role’,‘custom’, ‘institution’ or whatever. The main fault with traditional sociologicaltheory, then, is that it seems to deny or, at least, to ignore the fact that individualshave selves. Blumer recognizes two fundamental sociological views: symbolicinteractionism, which lodges social action in individuals and sees society asconsisting of acting people, and traditional sociology, which ‘lodges social action

132 Society as subjectively meaningful interactionin the action of society or in some unit of society’, and which tends to viewhuman society in terms of social structure or social organisation. 14The same message is repeated in the article on the ‘Sociological Implicationsof the Thought of George Herbert Mead’. Blumer renews his attack on traditionalsociology for resorting to ‘societal factors, such as cultural prescriptions,values, social roles, or structural pressures’, thereby missing ‘the central pointthat human interaction is a positive shaping process in its own right’ (1969: 66).Against the structuralist view of traditional sociology, Blumer posits Mead’s viewof society as an ongoing process of social action.Mead’s scheme definitely challenges this conception. It sees human societynot as an established structure but as people meeting their conditions of life;it sees social action not as an emanation of societal structure but as a formationmade by human actors; it sees this formation made by human actors; itsees this formation of action not as societal factors coming to expressionthrough the medium of human organisms but as constructions made byactors out of what they take into account; it sees group life not as a releaseor expression of established structure but as a process of building up jointactions … accordingly, it sees society not as a system, whether in the form ofa static, moving, or whatever kind of equilibrium, but as a vast number ofoccurring joint actions, many closely linked, many not linked at all, manyprefigured and repetitious, others being carved out in new directions, and allbeing pursued to serve the purposes of the participants and not the requirementsof a system.(Blumer, 1969: 74f)The same points are repeated and expanded upon in Blumer’s introductoryessay to his Symbolic Interactionism. In opposition to a ‘structuralist’ conception ofsociety, Blumer insists that ‘human groups or society exists in action and must beseen in terms of action’. More specifically, ‘a society consists of individuals interactingwith one another’. Even more specifically, such interaction in humansociety is characteristically and predominantly on the symbolic level’ (1969: 6–7).This is the view of society as symbolic interaction. The necessary preconditionfor this view, however, is intersubjectivity. In addition, therefore, Blumer suggeststhat the nature of an object, including social objectsconsists of the meaning that it has for the person for whom it is the object… The meaning of objects for a person arises fundamentally out of the waythey are defined to him by others with whom he interacts … Out of aprocess of mutual indications common objects emerge – objects that havethe same meanings for a given set of people and are seen in the samemanner by them.(Blumer, 1969: 11)

Society as subjectively meaningful interaction 133Concerning joint or collective action, the action of groups, institutions, organisationsand social classes, Blumer takes the view that they ‘consist of individualsfitting their lines of action to one another … The joint action of the collectivityis an interlinkage of the separate acts of the participants’ (1969: 16f).Blumer ends his introductory essay by drawing the methodological implicationsof the symbolic interactionist view of society. Being a subjectivist theory ofsociety, the methodology proper to that theory is also subjectivist. In order tounderstand social life, it is necessary that the social scientist sees the actions ofpeople, the objects towards which these actions are directed and the situation inwhich the action take place, in the same way as the acting individuals themselves.It is also necessary that the social scientist analyses the process in which humanaction is formed or constructed and refrains from explaining human action interms of antecedent conditions or causes. Concerning the activity of organisations,symbolic interactionism ‘seeks explanation in the way in which theparticipants define, interpret, and meet the situations at their respective points’(Blumer, 1969: 58).Symbolic interactionism, then, is Herbert Blumer’s interpretation of GeorgeHerbert Mead’s social psychological theory of the self and the view of society itsuggests. Typical for this interpretation is the view of society as an ongoingprocess of symbolic interaction or communication, together with an insistencethat society is continually constituted and reconstituted by the actions of individuals.A typical representative of this view is Tamotsu Shibutani who sees societyas a ‘communicative process’ based upon consensus in the form of shared meanings(1961, passim; 1962: 134ff). This consensus is not such as to give rise to anythinglike a stable social structure. On the contrary, Shibutani underscores the transientcharacter of society.Society consists of the recurrent adjustment and cooperation of associatedpersons through which action patterns of all kinds are formed, sustained,modified, evaded, or contravened. Sometimes the coordinated activitiesbecome highly organized, but there are also transient forms of interaction.All this suggests that human society might best be regarded as an on-goingprocess, a becoming rather than a being. Society might be viewed most fruitfullyas a succession of events, a flow of gestural interchanges among people… In this sense society can be said to exist only in the give and take amonghuman beings.(Shibutani, 1961: 174f)Another illustration of the symbolic interactionist view of society is AnselmStrauss’s image of society as a ‘negotiated order’. Strauss, at first takes overShibutani’s view of society as a ‘communicative process’ (Strauss [1969] 1977:148ff, 162), but later, while investigating the organisation of work in the hospital,develops his own vision of the social order. 15

134 Society as subjectively meaningful interactionIn short, the bases of concerted action (social order) must be reconstitutedcontinually; or as remarked above, ‘worked at’. Such considerations have ledus to emphasize the importance of negotiation – the process of give-andtake,of diplomacy, of bargaining – which characterizes organizational life.(Strauss et al., 1963: 148)Shibutani’s and Strauss’s views of society seem to me like new wine in oldbottles. They are, in all essentials, new versions of the old theory of society as asocial contract. What is new, is the emphasis upon change and renewal. Peopleare constantly engaged in ‘signing’ new contracts. It is the theory of society as anongoing social contract. This is the reason why consensus over rules does notgive rise to a stable structure.When professionals are brought together and enjoined to carry out theirwork in the same locale, concepts of structure (formal or informal) as relativelyset systems of norms and expectations are inadequate to explainresulting activity. The activity of interacting professionals is, we submit, largelygoverned by continual reconstitution of bases of work through negotiation.(Strauss et al., 1964: 375)The notion of ‘negotiated order’ was initially developed by Strauss and hiscolleagues in order to capture the nature of work organisation in the hospital,but was extended in its use to cover the nature of social organisation generally.Society becomes an arena on which individuals meet and carry out theircomplex negotiations.Pushing our logic to its extreme, we might even argue that the very idea of‘nation’ or ‘society’ is only a fiction and that, if the sociologist subscribes tothis common-sense fiction, rather than viewing a nation or society as anexceedingly complex arena, he may fall into the deadly trap of merelystudying the fiction as if it were a fact.(Strauss et al., 1964: 377)Admittedly this is an extreme example. Symbolic interactionism, I have said, isindividualistic not so much by explicit denial of the existence of social wholes, asby a positive fostering of the alternative view of society as an intersubjectivecommunity based on shared meanings.There are also many symbolic interactionists who do not present any generalpicture of society at all, and some who do not deny the utility or even the truthof a holist approach to society, but who have chosen to concentrate on the more‘microscopic’ processes of face-to-face interaction, thereby unwittinglycontributing to the spread and maintenance of the intersubjective theory ofsociety. An obvious example of this kind of attitude is Erving Goffman, who isusually included among the symbolic interactionists, even if he did not includehimself in this category.

Society as subjectively meaningful interaction 135Goffman has, in his various studies, focused consistently on face-to-face interaction,especially in the form of so-called ‘encounters’ or focused gatherings,using the models of the theatre and the game as his main analytical instruments.Most well known is the model of the theatre and the dramaturgical approach,which he used in his path-breaking study of The Presentation of the Self in EverydayLife (1959). 16 It is not at all that Goffman denies the existence of social structures;quite the contrary. He explicitly affirms their existence and also discusses theirrelation to the kind of ‘situated activity systems’ and the kind of social interactionhe is interested in (Verhoeven, 1985). I think it is fair to say that Goffman’smain concern is with human behaviour as a manifestation of the ‘degrees offreedom’ there are within social structures. A particularly clear example of this ishis discussion of ‘role distance’ in Encounters (1961: 85ff), which deals with therelative autonomy of the individual with respect to social structure. Sometimes,Goffman’s focus is on the autonomy of actors, who are then depicted as strategicallyrational, more or less, in the manner of rational choice. The most obviousexample of this approach is his book with the title Strategic Interaction [1969] 1970.At other times his focus is on social action as a manifestation of social structure.The essays assembled in Interaction Ritual (1967) exemplify this Durkheimian viewof social life. Both approaches are used in Relations in Public (1971), whereGoffman points to the problematic nature or concept of ‘individual’ and castsdoubt on the idea of the individual as the unit of analysis in sociology. The individualturns out to be different things in different systems of activity (pp. 23ff).Goffman, then, is an individualist, if at all, in practice only, but not always inpractice, and not at all in principle. The following disclaimer comes from the‘Introduction’ to his Frame Analysis (1974), but is typical for his position:This book is about the organization of experience – something that an individualactor can take into his mind – and not the organization of society. Imake no claim whatsoever to be talking about the core matters of sociology– social organization and social structure. Those matters have been and cancontinue to be quite nicely studied without reference to frame at all. I amnot addressing the structure of social life but the structure of experienceindividuals have at any moment of their social lives. I personally holdsociety to be first in every way and any individual’s current involvements tobe second; this report deals only with matters that are second. This bookwill have weaknesses enough in the areas it claims to deal with; there is noneed to find limitations in regard to what it does not set about to cover.(Goffman, 1974: 13)Recent symbolic interactionism seems to go in several directions. Some partsare more individualistic, others more structuralistic (Fine, 1990). If we acceptthat Blumer is the founder of ‘symbolic interactionism’, we may call the individualistictendency ‘orthodox’ symbolic interactionism. This tendency unitesaround some version of the intersubjectivist theory of society. In his introductorycontribution to the first reader on symbolic interactionism Arnold M. Rose

136 Society as subjectively meaningful interactiondepicts society as ‘a network of interacting individuals – with its culture – therelated meanings and values by which individuals interact’ (1972: 13). In theintroductory contribution to another reader on symbolic interactionism BernardN. Meltzer suggests that ‘human society rests upon a basis of consensus, i.e., thesharing of meanings in the form of common understandings and expectations’(1967: 8). The clearest and most elaborated statement of symbolic interactionismas an intersubjectivist theory of society is probably to be found in Joel M.Charon’s Symbolic Interactionism (1979: ch. 11). Charon bases his view of societymainly on the writings of Blumer, but also on those of Shibutani, Strauss andMeltzer. Society is characterised by two things: symbolic interaction and cooperation.‘Society, then, is defined here as individuals in symbolic interactionwith each other, aligning their acts, and acting cooperatively to resolve problemsin situations’ (p. 161).The year after Charon’s book was published, there appeared another bookwith the same concise title, Symbolic Interactionism (1980), by Sheldon Stryker, butwith a different message. Stryker tried to launch a more structuralist version ofsymbolic interactionism. In order to do so, he had to return to the pragmatist rootsof symbolic interactionism and to replace Blumer by Ralph Turner, as the mostimportant figure in recent symbolic interactionism. The main interest of Turnerhas been in a theory of collective behaviour, but he has also made an importantcontribution to sociological role theory (Turner, 1956; 1962; 1978). Like Goffman,Turner has been interested in the difference and interrelation between role andperson, and, especially, in the process of role-taking, first anlysed by Mead. I havenothing to object to Stryker’s attempt to create a more structuralist traditionwithin symbolic interactionism – the conditions are at hand – but the fact remainsthat there is a more well-known individualist tradition in symbolic interactionism,associated with Herbert Blumer, Anselm Strauss, Tamotsu Shibutani and others.This tradition has created an individualist theory of society which, like that of theearly Simmel, is both intersubjectivist and interactionist. The combination ofthese elements, may be represented as in Figure 5.1. (Horizontal arrows representintersubjectivism and vertical arrows represent interactionism).Figure 5.1 Society according to symbolic interactionism

Society as subjectively meaningful interaction 137As has been noticed by two symbolic interactionists, Chad Gordon andKenneth J. Gergen, there are certain affinities between symbolic interactionismand the phenomenological sociology of Alfred Schutz. According to the latter,‘the cardinal feature of the social world is its intersubjectivity. This view emphasizesthe shared-in-common and mutually sanctioned set of expectations andinterpretative categories within a culture’ (Gordon and Gergen, 1968: 33).Phenomenological sociology[S]ociety as such is a concept, an abstraction – What exists in reality are individualsin whose minds society exists as a factor determining certain types of behavior. If themental attitude no longer exists, society does no longer exist either. If peoplewere not aware of each other’s existence, society would not exist, even if allthe same people were still in existence … Thus, society is an attitude in themind of the individual which is subject to X changes each second. It isunstable and undetermined, although it may appear constant and concreteon the surface during long periods, or made to appear this way by socialtheorists.(Landheer, 1952: 22)This is one of the clearest statements of the intersubjectivist theory of society Ihave seen. It is written by a phenomenological sociologist, who is also a methodologicalindividualist, or so it seems. Landheer writes: ‘The individual has historicaland logical priority in regard to the group; or quite simply, a group develops out of individualsbecause no group can exist otherwise while the individual can, at least theoretically’ (p. 38).Two methodological rules follow from this ontological thesis: (1) ‘Social science dealswith processes of change which must be explained in terms of the individual’ (p. 38), and (2)‘Social science deals with those changes in the individual which occur in a plurality of individualsand which form, consequently, the basis for what is termed social patterns’ (pp. 38f).According to Landheer, ‘the individual is the basic unit of social science, although thechanges in the individual may in turn be influenced by social processes’ (p. 39).Landheer, then, is both an ontological and a methodological individualist. 17As such he is typical of phenomenological social science. In this section I amgoing to substantiate this claim, by a presentation of the founder of phenomenologicalsociology, Alfred Schutz.As we have seen in chapter 3 (pp. 80–5), phenomenology is the name of aphilosophical school, founded by Edmund Husserl. Other philosophers associatedwith the phenomenological movement are Max Scheler, Karl Jaspers,Martin Heidegger, Jean-Paul Sartre, Maurice Merleau-Ponty and AronGurwitch. This list of names reveals a close, but controversial, relationshipbetween phenomenology and existentialism. Sociologists, inspired byphenomenology, include, besides Schutz, himself, Maurice Natanson, ThomasLuckmann, Peter Berger, John O’Neill, Kurt H. Wolff, Edward A. Tiryakianand, more recently, those who call themselves ‘ethnomethodologists’. 18

138 Society as subjectively meaningful interactionSchutz claims to draw his ideas on the foundations of the social sciences fromthe work of Husserl. This claim is not uncontested, but very difficult to assess,due to the diverging interpretations of the development of Husserl’s ownthought. 19 The most radical critique comes from Barry Hindess (1973a: 5ff;1977: 49–77), who argues that Schutz’s claim to base his phenomenology uponthat of Husserl is totally unfounded. That, in fact, Schutz’s phenomenology is aserious distortion of Husserl’s views. According to Hindess, the basic premiseand project of Schutz is instead to be found in the methodology of Max Weber.This project is to reduce the world of objective mind to that of the actions ofindividuals. As such, Schutz’s ‘phenomenology’ is a continuation of thehumanism and individualist idealism of the German Geisteswissenschaften tradition.I believe that the accusation of a distortion of Husserl’s views isconsiderably exaggerated, if not entirely unfounded, but there is little doubtabout Schutz’s place in the tradition of German individualist idealism, a factwhich is readily acknowledged also within the phenomenological camp. WhatHindess does not seem to realise, however, is that also Husserl, especially thelater Husserl, was close to the German Geisteswissenschaften tradition (see pp.82–5). 20 Schutz’s phenomenological sociology is based on the ideas of the laterHusserl (Schutz [1932] 1972: 31–44), ideas which Schutz knew well from hisfrequent discussions with Husserl himself (Schutz, 1940: 164ff; 1959: 86ff). 21I will not enter the debate about the relationship between Schutz’sphenomenological sociology and Husserl’s philosophical phenomenology, butaccept as a fact that there is something called ‘phenomenological sociology’ andthat Schutz is its founding father and central figure. First, however, I will drawattention to some ideas and notions of Husserl which have been important forthe phenomenological sociology of Schutz.Schutz’s ontological and methodological individualism was grounded beforehe became acquainted with the work of Edmund Husserl. The first influence inthis direction, probably came from his teacher in economics, the methodologicalindividualist Ludwig von Mises. In the 1920s Schutz was a participant in the‘Mises seminar’, where the sociology and methodology of Max Weber was afavourite topic of discussion (Wagner, 1983: 11ff; Prendergast, 1986). It ispossible to see both Schutz’s first book The Phenomenology of the Social World (1932)and Mises’s first treatment of philosophical topics in Epistemological Problems ofEconomics (1933), as outcomes of these discussions. Both works are encounterswith the ideas of Max Weber.In the first phase of his intellectual development Schutz was a Weberian. Hewas, especially impressed by the first part of Economy and Society, where Weberlays the conceptual foundation for his individualist and subjectivist approach tosociology (Wagner, 1983: 14). Schutz adopted this approach, himself, andremained a subjectivist methodological individualist for the rest of his life. In hisPreface to the posthumously published The Structures of the Life-World (1973),Schutz’s co-author Thomas Luckmann makes the following authoritative statementabout Schutz: ‘He shared Max Weber’s methodological individualism and

Society as subjectively meaningful interaction 139realized the strategic importance of an adequate theory of human action for themethodology of social science’ (p. xix).Alfred Schutz’s theory of the social world and the method of its study is laiddown in all essentials in his The Phenomenology of the Social World (1932). Thecentral problems of this work are set by the interpretive sociology of Max Weberand his predecessors. In the first paragraphs of his Phenomenology, Schutz makesclear that the fundamental problem of the social sciences is that of the relationshipbetween individual and society.Is society prior to the individual, so that apart from the social whole theindividual does not exist at all? Or should we put it quite the other way andsay that the individual alone exists and that social organizations, includingsociety itself are mere abstractions – ‘functions’ of the behavior of separateindividuals?(Schutz [1932] 1972: 4)In trying to answer these questions and improve upon the answers given byWeber, Schutz appeals first of all to the philosophy of Husserl, but also to that ofHenri Bergson. Schutz makes clear, however, that in grounding his analysis ofWeber on the phenomenology of Husserl, he leaves aside all problems of transcendentalphenomonology and stays at the level of phenomenologicalpsychology (Schutz [1932] 1972: 43f). 22 There are, in particular, two elements ofWeber’s methodology which Schutz wishes to improve upon. First, there isWeber’s analysis of the subjectively intended meaning of action. According toSchutz, Weber is not absolutely clear about whose subjective meaning, theactor’s or the observer’s, he is talking about (ch. 1). Weber also fails to make theimportant distinction between because motives and in order to motives, the firstreferring to the conditions in the past which make an individual act in a certainway and the latter referring to what the individual wants to achieve with hisaction in the future, the goals of his projected action (pp. 86ff). Second, there isWeber’s notion of ideal type, which Schutz analyses as a matter of common senseand of scientific typifications of differing degrees of anonymity (pp. 176ff).There is, however, one element of Weber’s methodology which Schutz acceptswithout reservations; his methodological individualism.we frequently use sentences in which ideal types like ‘the state’, ‘theeconomy’, ‘the nation’, ‘the people’, or perhaps the ‘working class’ appear asgrammatical subjects. In doing this, we naturally tend to personify theseabstractions, treating them as if they were real persons known in indirectsocial experience. But we are here indulging in an anthropomorphism.Actually these ideal types are absolutely anonymous … From the sociologicalpoint of view, therefore, the term ‘state’ is merely an abbreviation for ahighly complex network of interdependent personal ideal types.(Schutz [1932] 1972: 198f)

140 Society as subjectively meaningful interactionAccording to Schutz, the starting-point for the social sciences is to be found inthe intersubjective world of ordinary life, but this is only the starting-point. Thesocial scientist does not stay in this intersubjective world, he/she makes it theobject of study, and this study is objective. The social sciences aim at an objectiveinvestigation of the intersubjective world of ordinary life, or in Schutz’s ownwords: ‘All social sciences are objective meaning-contexts of subjective meaningcontexts’(p. 241).An excellent summary of the main themes of The Phenomenology of the SocialWorld can be found in Schutz’s article on ‘The Social World and the Theory ofSocial Action’ (1949). 23 I will quote extensively from this article, since, in myopinion, it is a superbly clear and concise statement of the intersubjective theoryof society and its methodology, but also because, written in 1940, it revealsclearly the influence of Weber and of Ludwig von Mises on the thinking ofSchutz, as well as the influence of Schutz on the thinking of Friedrich vonHayek. Schutz attacks those social scientists who admit that social phenomenaconsist of the activities of human beings, but who nevertheless deny that it isnecessary to go back to the subjective activities of human beings for theirdescription and explanation. He admits that, on a certain level, social sciencecan proceed without entering into the problems of subjectivityBut then – and this is an important point – this reference to the subjectivepoint of view always can be performed and should be performed. As thesocial world under any aspect whatsoever remains a very complicatedcosmos of human activities, we can always go back to the “forgotten man”of the social sciences, to the actor in the social world whose doing andfeeling lies at the bottom of the whole system … The safeguarding of thesubjective point of view is the only sufficient guarantee that the world ofsocial reality will not be replaced by a fictional non-existing worldconstructed by the scientific observer.(Schutz, 1960: 207, 209)In contrast to the world of natural things stands the world of social things.I cannot understand a social thing without reducing it to the human activitywhich has created it and, beyond it, without referring this human activity tothe motives out of which it springs. I do not understand a tool withoutknowing the purpose for which it was designed, a sign or a symbol withoutknowing what it stands for, an institution if I am unfamiliar with its goals, awork of art if I neglect the intentions of the artist which it realizes.(Schutz, 1960: 211f)‘Summing up, we come to the conclusion that social things are only understandableif they can be reduced to human activities; and human activities are onlymade understandable by showing their in-order-to or because motives’ (p. 214).

Society as subjectively meaningful interaction 141Social relations consist of the intersubjective adjustments of individual’smotives in their interaction with each other.It can be proved that all social relations as they are understood by me, ahuman being living naively in the social world which is centered aroundmyself, have their prototype in the social relation connecting myself with analter ego with whom I share space and time. My social act, then, is orientednot only to the physical existence of this alter ego but to the other’s actwhich I expect to provoke by my own action. I can, therefore, say that theother’s reaction is the in-order-to motive of my own act. The prototype ofall social relationship is an intersubjective connection of motives.(Schutz, 1960: 215)In 1940 Schutz also wrote a review of Talcott Parsons’s recently publishedThe Structure of Social Action (1937) on the request of Friedrich von Hayek andintended for publication in Economica. Before sending the review to Hayek,Schutz sent it to Parsons for a reaction, which turned out to be such that hedecided to abstain from publication. The review is now published (Schutz andParsons, [1940–1] 1978), together with the correspondence it occasioned.Obviously Parsons and Schutz did not understand each other, despite theircommon interest in the sociology of Max Weber. There are several possiblereasons for this. One reason may be that Parsons was a neo-Kantian, whereasSchutz was a phenomenologist. Another reason may be that Parsons was notvery interested in a philosophical foundation, whereas Schutz was convinced thatit is necessary. A third reason is probably that, despite their common point ofdeparture in Weber’s theory of action, they were moving from there in oppositedirections: Parsons in the direction of a holistic structural-functionalism; Schutzin the direction of a more firm individualist foundation of social theory.The main point, made by Schutz ([1940] 1978: 25ff), is that Parsons’sapproach is too objectivistic, but lacks an adequate subjectivist foundation, especiallya theory of motives, which is necessary to all social science. A main part ofthe review consists of a summary of Schutz’s own attempt to provide this foundation(pp. 33ff). ‘Summing up, we come to the conclusion that social things areunderstandable only if they can be reduced to human activities; and humanactivities can be made understandable only by showing their in-order-to orbecause motives’ (p. 53). Parson’s reaction was negative, and one possible reasonfor this is that he was already on his way to developing an even more objectivistictheory of the social system. More recently, Jürgen Habermas has combined thesubjectivist approach of Schutz and the objectivist approach of Parsons in histheory of society as both life-world and system (Habermas [1981b] 1987).Schutz’s later writings are elaborations of the views put forward in ThePhenomenology of the Social World, with only minor modifications, inspired mainlyby his reading of William James and the symbolic interactionists. Schutz nowrecognises the ‘existence’ of multiple realities or sub-universes; the world ofdreams, of ‘imageries and phantasms’, the world of art, the world of religious

142 Society as subjectively meaningful interactionexperience and of science, but of these worlds the world of daily life isparamount. All other worlds are derived from, and refer back to, the intersubjectivelife-world within which we work and think and which is pre-given to us all(Schutz, 1945a: 549ff; 1955: 186–93). This intersubjective world of everyday lifeis a world of meaningful objects created by individuals in their interaction withone another. ‘Thus, in these reciprocal acts of positing meaning, and of interpretationof meaning, my social world of mundane intersubjectivity is built; it is alsothe social world of Others, and all other social and cultural phenomena arefounded upon it’ (1940: 182). Social objects, such as nations, governments,markets, prices, etc. and cultural objects, such as books, tools, works of art, etc.are meaningful phenomena created and maintained by individual human beings.According to Schutz, however, social objects – collectivities and institutions –‘are as such not entities within the province of meaning of everyday reality butconstructs of common-sense thinking which have their reality in another subuniverseof ideal relations’ (1955: 198). These ideal relations have their locus inthe minds of individuals. Schutz’s idealism is an individualist idealism and he iscareful to dissociate himself from Durkheim’s notion of a ‘collective consciousness’and also from ‘Husserl’s conception of the collectivity as a subjectivity of ahigher order’ (1959: 92). Of special importance for the social sciences is Schutz’sview that all social relations derive from the ultimate social reality of interactionin face-to-face relationships. The related phenomenon of social role is reducedto the typical expectations people have of each other’s behaviour. 24Turning now to the reality of ‘scientific contemplation’, we find that Schutzrefers to the concepts of social science as concepts of the ‘second degree’, asconcepts referring to an already conceptualised world (1953: 3; 1954: 267ff). 25We also find that, according to Schutz, these concepts do not refer to concretehuman beings, but to ideal types in the form of ‘puppets’ or ‘homunculi’ (1945a:572; 1960: 219). The concepts of social science, like those of common sense, aresubjected to the postulate of subjective interpretation; a variant of methodologicalindividualism. According to this postulate, concepts such as ‘savings’,‘capital’, ‘business cycle’, ‘unemployment’ and ‘monopoly’ are intellectual shorthandfor human actions.Correctly understood, the postulate of subjective interpretation as appliedto economics as well as to all the other social sciences means merely that wealways can – and for certain purposes must – refer to the activities of thesubjects within the social world and their interpretation, motives, relevancesand so on.(Schutz, 1953: 27)If there is a central theme of Schutz’s theory of society, it is intersubjectivity(cf. Rogers, 1983: ch. 8). The life-world we all inhabit and experience, is an intersubjectiveworld we share in common. It has been suggested, by Prendergast(1986: 3, 11), that the origin of this theme, in Schutz’s theory of society, shouldbe sought in his early attachment to the Austrian School of Economics, which,

Society as subjectively meaningful interaction 143according to Schutz, lacked, but needed, an adequate understanding of intersubjectivity.In the phenomenology of Husserl, Schutz found a theory seriouslyconcerned with the problematic fact of intersubjectivity. In the natural attitudewe take it for granted that we live in an intersubjective world, shared, to somedegree, by all of us. In his last book on The Crisis of European Sciences, Husserlcalled the common-sense part of this world, the life-world. The main challenge,for Husserl, however, was to provide a transcendental constitution of intersubjectivity.His main attempt to do so was in Cartesian Meditations (see pp. 83f ).By the end of his life, Schutz voiced serious objections to Husserl’s treatmentof transcendental intersubjectivity (Schutz [1957] 1970). I neglect the details andturn to the main conclusion: according to Schutz, it is simply impossible toconstitute an intersubjective world, or community, from the transcendental ego,since ‘reciprocal understanding and communication already presuppose acommunity of knowledge’ (p. 72). Many unsettling questions arise concerningthe transcendental ego. The transcendetal ego, we recall, is a Leibnizian monad,‘[b]ut how can my full monad in its concretization enter into a transcendentalwe-relationship with yours? (p. 76). Is there such a thing as a transcendental we?‘[I]s it conceivable and meaningful to speak of a plurality of transcendentalegos?’ (p. 77). ‘Are the Sumerians and the pygmies of the African bush, who areunknown to me, actually constituted in my meditating ego? Do I constituteSocrates or does he constitute me?’ (pp. 81f).As a result of these considerations we must conclude that Husserl’s attemptto account for the constitution of transcendental intersubjectivity in terms ofthe operations of the consciousness of the transcendental ego has notsucceeded. It is to be surmised that intersubjectivity is not a problem ofconstitution which can be solved within the transcendental sphere, but israther a datum (Gegebenheit) of the life-world. It is the fundamental ontologicalcategory of human existence in the world and, therefore, of allphilosophical anthropology. As long as man is born of women, intersubjectivityand the we-relationship will be the foundation of all other categoriesof human existence. The possibility of reflection on the self, discovery of theego, capacity for performing the epoché, and the possibility of all communicationand of establishing a communicative surrounding world as well, arefounded on the primal experience of the we-relationship … It can … besaid with certainty that only such an ontology of the life-world, not a transcendentalconstitutional analysis, can clarify that essential relationship ofintersubjectivity which is the basis of all social science – even though, as arule, it is there taken for granted and accepted without question.(Schutz [1957] 1970: 82)When Schutz died in 1959, he was engaged in the project of creating such anontology of the life-world. The project was finished by his student ThomasLuckmann and published as a joint work called The Structures of the Life-World(1973: 19). In this work, there are some openings towards objectivism and

144 Society as subjectively meaningful interactionstructuralism, but since it is partly written by Luckmann, it is difficult to knowwhether these openings really reflect Schutz’s views. Since they point in thedirection of the views expressed by Berger and Luckmann (1966), they might bethe contribution of Luckmann, more than of Schutz. I conclude, therefore, withwords written by a Popperian (objectivist) methodological individualist:Of the several forms of individualism, which is a basic schema in the socialsciences, the phenomenological approach is one of the most significant andextreme. The overriding aim is to interpret all sociological happenings inrelation to the individual experiences of individual people.(Wisdom, 1973: 257)The main idea of the phenomenological view of society is that of intersubjectivity.Unlike symbolic interactionism, it is almost exclusively concerned withintersubjective meaning, but hardly at all with social interaction. The view ofsociety contained in phenomenology, may be represented like this (the continuoushorizontal arrows indicate that intersubjectivism is more important thaninteraction, which is represented by broken arrows):Figure 5.2 Society according to phenomenological sociologyExistentialismExistentialism first arose in Germany, and it is common to see Karl Jaspers as thefounder of the movement. The most important forerunners are SörenKierkegaard and Friedrich Nietzsche. Another important figure in the earlyhistory of the movement was Martin Heidegger. Today we think of existentialismas mainly a French movement of thought associated, above all, withnames such as Jean-Paul Sartre, Simone de Beauvoir and Maurice Merleau-Ponty.Existentialism developed in close connection with the phenomenology ofEdmund Husserl. Because of this, it is sometimes called ‘existentialphenomenology’. Maurice Merleau-Ponty, above all, is a phenomenologicalphilosopher, but also Jean Paul Sartre was heavily indebted to Husserl. Equallyimportant, however, is the French heritage. Existentialism is, likephenomenology, a philosophy of consciousness and, like the latter, it takes its

Society as subjectively meaningful interaction 145point of departure in the cogito of Descartes. Another important source of inspirationis the early philosophy of Henri Bergson.The influence of Bergson is most obvious in the case of Gabriel Marcel(1889–1973), the first French existentialist philosopher, who was also a Catholic,interested, above all, in the prospects of Christian freedom and morality in themodern world. Marcel was the first to introduce existentialism in France, but itwas Jean Paul Sartre who made it the most popular philosophy in post-warFrance.Jean-Paul Sartre was an atheist, with a different attitude to the world, but heshared Marcel’s focus on the freedom of the individual subject. Sartre started asa phenomenological philosopher, addressing various problems in Husserl’s analysisof consciousness. In the latter half of the 1930s, he wrote some minor works,where he criticised Husserl’s analysis of the transcendental ego and laid thefoundation for a phenomenological psychology of emotions. The phenomenologicalbasis of existentialism is evident also in Sartre’s first magnum opus Being andNothingness (1943), which bears the subtitle: An Essay on Phenomenological Ontology,but now he orients himself, also to Heidegger and Hegel.Ontology means a theory of being and, according to Sartre, there are tworealms of being: ‘being-in-itself ’ and ‘being-for-itself ’. Being-in-itself is the realmof non-conscious being, or of things, passive and inert. Being-for-itself is therealm of consciousness, of activity and freedom. Having established these tworealms of being, Sartre goes on to consider their relation. Like Heidegger, heconceives of it as the ‘being-in-the world’, on the part of being-for-itself. Sinceonly human beings have consciousness, this relation takes the form of ‘man-inthe world’. According to Sartre, the characteristic attitude of man-in-the-world isinterrogation, to pose questions. And to questions there are two possible replies:yes and no. The very possibility of a negative reply creates ‘nothingness’. ‘Man isthe being through whom nothingness comes to the world’ ([1943] 1956: 24).Now, if being is what is, nothingness is what is not, but which could be; nothingnessis possibility. It is up to you and me to realise the possible. ‘Man’s relation withbeing is that he can modify it’ (p. 24).Existentialism is a philosophy of freedom. Human beings are free to choose,not only this or that, but their own lives, or essences. This freedom in the face ofan open future of possibilities, however, creates anguish. But since anguish is nota pleasant feeling, the immediate reaction is flight from freedom ([1943] 1956:40ff). This reaction is a manifestation of what Sartre calls ‘bad faith’. Bad faith isa form of self-deception. We delude ourselves into believing that we must do allthe things we are expected to do. We renounce ‘being-for-itself ’ and enter‘being-in-itself ’. According to Sartre, the traditional sociological picture of sociallife, as the enactment of social roles, is a life in the attitude of bad faith (pp. 59f).Existentialism, then, is a form of radical individualism. This becomes evenmore evident when Sartre turns to an analysis of ‘being-for-others’. On thispoint, he sides with Husserl’s monadology and solipsism, against Hegel’s holismand Heidegger’s collectivism ([1943] 1956: 223–52). The view of society thatemerges is Hobbesian. If I am free to pursue my own project, it is only natural

146 Society as subjectively meaningful interactionthat my freedom should clash with that of Other. In asserting my freedom, I tryto enslave Other and Other tries to enslave me. ‘Conflict is the original meaningof being-for-others’ (p. 364). The reason for this situation is that each is the limitto the freedom of the other. ‘From the moment that I exist I establish a factuallimit to the Other’s freedom. I am this limit, and each of my projects traces theoutline of this limit around the Other’ (p. 409). In the end, this destructive relationgives rise to hate, first of Other, and eventually of all others (pp. 410–12). AsSonia Kruks has pointed out, this bleak picture of the human condition hasmuch in common with Hobbes’ stature of nature, but without the escape of asocial contract.The world would be Hobbesian not only in the sense of being fundamentallyconflictual, but in the sense of being necessarily random, unstructuredand anarchic. For the main features of Hobbes’ state of nature are not onlyfear and conflict, but lack of stability, of order, of institutions, of commonlinguistic meaning – of all those regularities that permit human culture todevelop and which are made possible for Hobbes only by the contract toend the state of nature.(Kruks, 1990: 76)Sartre goes on to discuss and reject Heidegger’s idea of Mitsein andDurkheim’s idea of a collective consciousness (p. 414). His conclusion is that ‘theexperience of the We-subject is a pure psychological, subjective event in a singleconsciousness … It is a question only of a way of feeling myself in the midst ofothers’ (p. 425).Being and Nothingness did not satisfy the Marxist intellectuals of the Frenchcommunist party, who saw it as a manifestation of individualism and subjectivism.Sartre, who was sympathetic to the communist cause, and supported theparty, felt obliged to clarify his position. This he did in the short essay on TheHumanism of Existentialism (1947), which is the most accessible presentation of thelatter doctrine. In this essay, he complains, like virtually every founder of adoctrine, that the word ‘existentialism’ has become ‘so stretched and has takenon such a broad meaning, that it no longer means anything at all’ ([1947] 1965:33). Nevertheless, what all existentialists have in common, ‘is that they believethat existence precedes essence, or if you prefer, that subjectivity must be thestarting point’ (p. 34). This mean that human beings make themselves and mustassume full responsibility for what they make of their lives. ‘Subjectivism means,on the one hand, that an individual chooses and makes himself; and, on theother, that it is impossible for man to transcend human subjectivity’ (p. 37).The main message of existentialism, then, is that we are free to choose ourown lives. ‘If existence really does precede essence, there is no fixed humannature. In other words, there is no determinism, man is free … condemned to befree’ ([1947] 1965: 41). So far, there is nothing that goes beyond the message ofBeing and Nothingness, but Sartre goes on to take some small steps in a more collectivisticdirection. A first step, I believe, is his suggestion that in making yourself,

Society as subjectively meaningful interaction 147you ‘at the same time create an image of man as we think he ought to be’ (p. 37).In this way, we all contribute to the production and reproduction of society andof humanity.if I want to marry, to have children; even if this marriage depends solely onmy own circumstances or passion or wish, I am involving all humanity inmonogamy and not merely myself. Therefore, I am responsible for myselfand for everyone else. I am creating a certain image of man of my ownchoosing. In choosing myself, I choose man.(Sartre [1947] 1965: 37)This is really a piece of social constructionism, which reminds of ethnomethodologyand of Anthony Gidden’s theory of structuration. The point Sartre ismaking is that social phenomena are, in Gidden’s terminology, ‘recursive’, theyexist only in and through the actions, by which they are reproduced. An importantdifference is that, whereas Giddens believes that human beings are ableconstantly to create society only with the aid of social structures already there, ifonly as memory traces, Sartre believes that ‘man, with no support and no aid, iscondemned every moment to invent man’ ([1947] 1965: 37; my italics). Even so,Sartre rejects solipsism and I feel that there is a certain ambiguity on this point.Sartre rejects the individualistic subjectivism of Descartes and Kant on theground that ‘we reach our own self in the presence of others’ (p. 51). Others arethe condition of our own existence. ‘Hence, let us at once announce thediscovery of a world which we shall call intersubjectivity; this is the world inwhich man decides what he is and what others are’ (p. 53).The message of Sartre’s social constructionism is that we are responsible forthe kind of community we live in and recreate, but there is always the possibilityof creating another one, more to our liking; and if you want to do it by means ofcollective action you are free to do so. Existentialism is a humanism, but isMarxism a humanism? Sartre thought so and many other Marxists thought sotoo, but, not all of them. In France, there soon developed a structuralist versionof Marxism, which denied that Marxism is a humanism and went as far as toreject all forms of subjectivism and humanism. 26Sartre would soon abandon his early version of existentialism, too, and movein a more Marxist direction, but he never moved as far as Althusser in denyingthe human subject. Far from it. His final view of society is presented in hissecond magnum opus, Critique of Dialectical Reason (1960). In this work, Sartre recognisessociety as an objective reality, made up of groups, organisations andinstitutions. These objective structures are part of the situation in which humansubjects act. Life in society is conceived of as dialectic between human subjectsand objective structures, much like the view of Berger and Luckmann in theirThe Social Construction of Reality (see pp. 161f).Sartre’s move from existentialism to Marxism is usually explained in terms ofhis allegiance to the French communist party. Sartre was not the only existentialistto move in a more structuralist direction, however. Before him, Simone de

148 Society as subjectively meaningful interactionBeauvoir and Maurice Merleau-Ponty had already taken existentialism in thisdirection. Could it be that Sartre was simply following their example? (Kruks,1990: 16–18, 83ff).The second volume of Simone de Beauvoir’s classic, The Second Sex (1949),begins with the famous statement: ‘One is not born a woman, but ratherbecomes, a woman’ ([1949] 1988: 295). This statement is in line with theorthodox existentialist thesis that existence precedes essence, but the continuationis less orthodox: ‘No biological, psychological, or economic fate determinesthe figure that the human female presents in society; it is civilization as a wholethat produces this creature, intermediate between male and eunuch which isdescribed as feminine’ (p. 295). According to de Beauvoir, a woman does nothave the freedom to choose herself, she is the victim of a situation created andmaintained by men. The culture and social institutions, which constitute patriarchyoffer no room for escape. ‘Marriage is the destiny traditionally offered towoman’ (p. 445), and it leads to subordination and to ‘work within the home’,which ‘gives her no autonomy’ (p. 475).Because of her more structuralist view of society, Sonia Kruks has suggestedthat de Beauvoir was closer to Merleau-Ponty than to Sartre (Kruks, 1990: 111),but I believe that she was more of a structuralist than either of them. Merleau-Ponty was, above all, a phenomenologist, taking his point of departure inHusserl’s idea of life-world and, like Husserl and Schutz, he entertained an intersubjectivisttheory of society, or so it seems to me. 27 Two things tell against thisinterpretation of Merleau-Ponty, however.First, Merleau-Ponty is sympathetic to structuralist analysis in general and tothat of French anthropology in particular, and structuralism is not normally seenas compatible with individualism in any form. If, however, as in Merleau-Ponty,structuralism is seen mainly as a method or a form of analysis, but not as anontological thesis about social reality, the problem can be made to disappear. If,also, structure is seen as entering social reality only through the lived experienceof individuals, structuralist analysis becomes compatible with an intersubjectivisttheory of society (see Merleau-Ponty, 1974: 116, 120).Second, Merleau-Ponty is also sympathetic to Marxian social analysis. This isno less of a problem. Marxism is even more difficult to reconcile with an intersubjectivisttheory of society. But also this difficulty is, somehow, overcome byMerleau-Ponty. It is done by concentrating exclusively on the philosophicalanthropology of the young Marx. But, whereas the later Sartre interpreted thisanthropology as based on a dialectic of subjective meaning and objective socialreality, Merleau-Ponty conceives of Marx as a champion of intersubjectivismand uses him against Durkheim.Marx, unlike Durkheim, would not even agree to speak of a collectiveconsciousness whose instruments are individuals … If it is neither a ‘socialnature’ given outside ourselves, nor the ‘World Spirit’, nor the movementappropriate to ideas, nor collective consciousness, then what is, for Marx, thevehicle of history and the motivating force of the dialectic? It is man involved in a

Society as subjectively meaningful interaction 149certain way of appropriating nature in which the mode of his relationshipwith others takes shape; it is concrete human intersubjectivity, the successiveand simultaneous community of existences in the process of self-realizationin a type of ownership which they both submit to and transform, eachcreated by and creating the other.(Merleau-Ponty, 1974: 177f)Merleau-Ponty’s view of society, then, would seem to be a brand of the intersubjectivisttheory of society.Existentialism did not make much of an impact on social science. At least notfor a start and not positively. The main impact is probably negative, but as suchit was far from negligible. The existentialism of Jean-Paul Sartre was the maintarget of attack for the French structuralists, who have conquered the world,since the 1960s. The sociologist, Pierre Bourdieu, for instance, starts his maintheoretical works with a critique of Sartre’s individualism, subjectivism andvoluntarism (Bourdieu [1972] 1977: 73–6; [1980] 1990: 42–7). Bourdieu wasmore favourably disposed towards the philosophy of Merleau-Ponty. 28Among the most outspoken representatives of existential phenomenology insociology is Edward A. Tiryakian. 29 In his version, existentialist-phenomenologicalsociology becomes a hospitable theory, able to assimilate virtually all oftraditional sociology. Not only Weber, Simmel and Mead, which makes sense,but also Durkheim, Sorokin and Parsons (Tiryakian, 1965), which is more difficultto understand. Most phenomenologists obviously find it necessary to relatetheir views to Durkheim’s notion of ‘collective consciousness’. The reason forthis is not hard to find. Durkheim denied the possibility of reducing social facts,to psychological facts, and this seems to imply a rejection of the intersubjectivisttheory of society. If the collective consciousness is a reality sui generis, it ispresumably irreducible to the intersubjective social reality of phenomenology.Schutz and Merleau-Ponty were right, from a phenomenological point ofview, to reject Durkheim’s notion of a ‘collective consciousness’. Berger andLuckmann accepted it (see pp. 161f ), but in doing so, went far beyond Schutz’phenomenological foundation for the social sciences. Tiryakian represents athird reaction. He accepts the idea of a ‘collective consciousness’, but reduces itto the ‘we-pole’ of intersubjective consciousness; the ‘form of intersubjectiveconsciousness which constitutes any and every community’, that is the commonlyshared life-world (1965: 180–2; 1973: 192ff). In making this reduction, Tiryakianis true to phenomenology, but hardly to Durkheim.For Tiryakian, as for Schutz and Merleau-Ponty, the central tenet of aphenomenological social science is the intersubjectivist theory of society. ‘Aboveall, it [existential phenomenology] is an orientation that seeks to find the structureof social phenomena in their meanings, meanings which are grounded inthe experience of subjects’. The primary application of existentialphenomenology to sociology, therefore, is that of ‘sensitizing us to the nature andfundamental role of intersubjectivity in the structure of social action’ (Tiryakian,1973: 209–21).

150 Society as subjectively meaningful interactionEthnomethodologyEthnomethodology is an Anglo-American movement, like symbolic interactionism.Its founder is Harold Garfinkel. At least, it was Garfinkel who coinedthe term ‘ethnomethodology’. Before he did, however, he was a pupil of bothParsons and Schutz. Since ethnomethodology is usually considered to be a developmentof phenomenology, it would seem that he sided with Schutz againstParsons, even though he wrote his dissertation for the latter. Be that as it may,ethnomethodology is, at least to some extent, a result of Garfinkel’s criticalencounter with the structural-functional theory of his teacher Talcott Parsons. 30According to Garfinkel, ethnomethodology ‘is an organizational study of amember’s knowledge of his ordinary affairs, of his own organized enterprises,where that knowledge is treated by us as part of the same setting that it alsomakes orderable’ ([1968] 1974: 18). 31 Among the many members and sympathisersof the ethnomethodological movement are Egon Bittner, Aaron V.Cicourel, Harvey Sachs, Emmanuel Schlegloff, David Sudnow, D. LawrenceWieder, Don Zimmerman, Melvin Pollner, Thomas P. Wilson, Alan Blum andPeter McHugh. Ethnomethodology, as indicated by the name, is primarily amethodological approach to the study of social life. As such, it is, however, basedupon certain assumptions about the nature of social life. It is with ethnomethodology’s(implicit) view of society, rather than with its methodology, that thissection is concerned.Ethnomethodology is often classified as being part of or, at least, anoutgrowth of the phenomenological movement, even its most vital part. 32 Anexception is Mary F. Roger (1983: ch. 7), who points out the important differencesthere are between these movements. Most important, ethnomethodology ismuch less concerned with subjective meaning than is phenomenology, but moreconcerned with actions, or practices. It has also been argued that ethnomethodologyshows close affinities with symbolic interactionism (Denzin, 1969;Alexander, 1985: 43, 51; 1987: 272ff). If this is so, it ought to be individualisticand intersubjectivistic. 33 This conclusion has been contested, however. While noone would deny that ethnomethodology is heavily influenced by thephenomenology of Schutz, the affinities with interactionism have been denied bymembers of the ethnomethodological movement (Zimmerman and Wieder,1971). It has also been argued that ethnomethodology is close to hermeneutics,ordinary language philosophy and even French structuralism, and, therefore, notindividualistic and subjectivistic. 34I do not deny that ethnomethodology is infected by ordinary language philosophy,hermeneutics and even structuralism, but not enough to make it immuneto individualism and subjectivism. I would like to argue, instead, for the moreimportant similarities between ethnomethodology and symbolic interactionism.Ethnomethodology, like symbolic interactionism is highly critical of structural-functionalsociology with its ‘oversocialized conception of man’ (Wrong,1961). Blumer’s critique of traditional sociology is echoed in Garfinkel’s referenceto the ‘man-in-the-sociologist’s-society’ as a ‘cultural dope’ (1967: 68).Ethnomethodology goes even further than symbolic interactionism in viewing

Society as subjectively meaningful interaction 151society as a process (rather than as an order). ‘Where others might see “things”,“givens” or “facts of life”, the ethnomethodologist sees (or attempts to see)process: the process through which the perceivedly stable features of socially organizedenvironments are continually created and sustained’ (Pollner, 1974: 27). 35Society, although constantly reconstructed or reconstituted, is never there assomething ‘tangible’. It has a transient or, to borrow a term from Habermas, an‘occasional’ character (Habermas [1981b] 1987: 130). ‘The basic position is thatsocial structures, conventionally defined as patterned social relationships,become visible and viable only as practical features of concrete moments ofhuman existence’, or in other words: society is ‘the world as it happens’ (Boden,1990: 190). To conclude: according to ethnomethodology, ‘Social reality isconstantly created by the actors; it is not a preexisting entity’. This is the exactopposite of Durkheim’s holistic view of social facts, but similar to that ofsymbolic interactionism. Finally, it would be somewhat odd if ethnomethodologydid not exhibit important similarities with symbolic interactionism, since it is adevelopment of Schutz’s phenomenology, which is certainly close to symbolicinteractionism in many respects (cf. Wagner, 1983: 73f, 136–40).Against Durkheim, Garfinkel argues that social facts are accomplishments. Theyare ‘accomplishments of the concerted activities of daily life’. But for concertedaction to be possible, there must be some rationality or accountability, somecommon knowledge of these accomplishments or the way they are achieved.This common knowledge is the topic (and resource) of ethnomethodologicalstudy. ‘Ethnomethodological studies analyze everyday activities as member’smethods for making those same activities visibly-rational-and-reportable-for-allpractical-purposes,i.e., “accountable”, as organization of commonplaceeveryday activities’ (Garfinkel, 1967: vii). This double character of everydayactivities, of being both accomplishments and ‘accountable’, is called with one ofethnomethodology’s most cherished terms, their reflexivity.The second most important term of ethnomethodology is indexicality. By‘indexicality’ is understood the property of member’s talk and conduct of beingdependent for their meaning and truth upon (1) the biography and present stateof the speaker or agent and (2) upon the linguistic and social context in whichtalk and conduct take place. Ethnomethodology, then, becomes ‘the investigationof the rational properties of indexical expressions as contingent ongoing accomplishmentsof organized artful practices of everyday life’ (Garfinkel, 1967: 11).Ethnomethodology, like both symbolic interactionism and phenomenologicalsocial science, finds the basis of social order in member’s shared meanings orcommon understandings. Garfinkel is anxious to make clear, however, that thesecommon understandings are not of substantive matters, but of member’smethods for making their practical actions accountable, i.e., rational for all practicalpurposes. Common understandings have an operational structure.Garfinkel also attacks sociologists for taking the social order for granted, as apoint of departure, instead of asking how social order is possible. He finds anexception in Alfred Schutz, however, and inspired by him, goes on to investigatethe routine grounds of everyday activities, that is, ‘how the structures of

152 Society as subjectively meaningful interactioneveryday actitivies are ordinarily and routinely produced and maintained’(Garfinkel, 1967: 38). Garfinkel’s solution to the problem of order is not asradical as many ethnomethodologists would have it. It is, in fact, the typicalanswer, not only of the intersubjectivist theory of society, but of sociological roletheory in general. Social order is seen as based upon people’s expectations ofeach other’s actions and upon the ability of competent members of a languagecommunity to comply with these expectations.Earlier the argument was made that the possibility of common understandingdoes not consist in demonstrated measures of shared knowledge ofsocial structure, but consists instead and entirely in the enforceable characterof actions in compliance with the expectancies of everyday life as a morality.Common sense knowledge of the facts of social life for the members of thesociety is institutionalized knowledge of the real world. Not only doescommon sense knowledge portray a real society for members, but in themanner of a self fulfilling prophecy the features of the real society areproduced by persons’ motivated compliance with these backgroundexpectancies. Hence the stability of concerted actions should vary directlywith whatsoever are the real conditions of social organization that guaranteepersons’ motivated compliance with this background texture ofrelevance as a legitimate order of beliefs about life in society seen ‘fromwithin’ the society.(Garfinkel, 1967: 53)Ethnomethodology’s claim to originality, to having effected a radical revisionof the problem of common understandings, rests upon two distinctions: (1) ‘thedistinction between the social world as topic of, and resource for inquiry’ and (2)the related distinction between a ‘normative’ and an ‘interpretive’ paradigm insociology. While traditional sociology treats common understandings, especiallyrules, as resources used in their explanations, ethnomethodology turns them intothe topic of inquiry. The social order is turned from something taken for grantedinto a problem. This alternative conception of social inquiry hinges on ‘theoccasioned corpus of social features’.By use of the term occasioned corpus, we wish to emphasize that the featuresof socially organized activities are particular, contingent accomplishments ofthe production and recognition work of parties to the activity. We underscorethe occasioned character of the corpus in contrast to a corpus ofmember’s knowledge, skill and belief standing prior to and independent ofany actual occasion in which such knowledge, skill, and belief is displayed orrecognized. The latter conception is usually refered to by the term culture.(Zimmerman and Pollner, 1971: 94)What distinguishes ethnomethodology (and phenomenology) from traditionalsociology is its treatment of rules. According to traditional sociology (and linguis-

Society as subjectively meaningful interaction 153tics) rules are accepted as given. Social order (and language) is seen as a matter offollowing these rules. According to ethnomethodology, things are much morecomplex. Talk and conduct, far from being a simple matter of following clear-cutrules, pose tremendous problems of application in concrete situations. The operationalisationof rules into criteria for use almost always ends with an et ceteraclause. The application of rules, therefore, involves an important element of adhocing and glossing practices (Garfinkel, 1967: 18–24). This difference betweenethnomethodology and traditional sociology is the basis of the distinctionbetween an ‘interpretive’ and a ‘normative’ paradigm in sociology (Wilson,1971).The distinction between an interpretive and a normative paradigm is genuineand important. As far as I can see, the interpretive paradigm is largely coextensivewith what I call ‘the intersubjectivist theory of society’. This is why Idisagree with those ethnomethodologists who, like Zimmerman and Wieder,exaggerate the differences between ethnomethodology and symbolic interactionismand even classify symbolic interactionism as belonging to the normativeparadigm (Zimmerman and Wieder, 1971: 286–8), but tend to agree with thoseethnomethodologists who, like Wilson and Cicourel, see the important continuitiesbetween symbolic interactionism and ethnomethodology.According to Wilson, ethnomethodology belongs, together with symbolicinteractionism, to the interpretive paradigm in sociology. Typical features of thisparadigm are the view of social interaction as an interpretive, rather than as arule-governed, process and the view of complex social phenomena as ‘patternedarrangements of interactions among individual actors’ (Wilson, 1971: 66).Thus, social organization is not treated as an objectively existing structure.Rather, the question is raised how it is that the members establish repetitiveness,stability, regularity, and continuity over space and time as features oftheir social world that are taken by themselves and anyone else as objectivematters of fact rather than as creatures of fantasy or whim.(Wilson, 1971: 78f)Needless to say this view of social organisation has certain consequences forethnomethodology’s view of social roles. Drawing upon symbolic interactionistssuch as Goffman, Turner and Blumer, Wilson emphasises role-taking as an interpretiveprocess, rather than role behaviour as conformity to norms. Cicourel,who like Wilson is much influenced by Turner, emphasises role-taking and rolemakingas modifying and creative elements of role behaviour, but also followsShibutani and Strauss in their view of society as a negotiated or constructedorder. 36A corollary of the processual and occasional view of society is the ‘nominalist’view that only individuals exist, but not institutions or other social wholes. Onthis view, to treat ‘social things’ as if they existed is to reify social reality. 37 Wilson(1971: 58f), consequently, adheres to individualism as an ontological thesis aboutsociety (but not to psychological reductionism), while Cicourel (1981: 58–61)

154 Society as subjectively meaningful interactiontakes the view that macro-phenomena (concepts) can be, and should be, understoodand analysed as aggregates of micro-situations (concepts).Pointing out these similarities between ethnomethodology and symbolicinteractionism is not to deny their dissimilarities. Symbolic interactionism iscloser to traditional sociology and its interest in norms and roles, whileethnomethodology is closer to hermeneutics, ordinary language philosophy andstructuralism, and their interest in language. There has been within sociology asa whole an increasing interest in language, and ethnomethodology is one of itsmost significant manifestations. What are the implications of this ‘linguisticturn’ for the question of the individualism and subjectivism of ethnomethodology?I think it is fair to say that the main thrust of the linguistic turn in the humansciences has been holistic. There is not, however, a necessary relation between anincreasing interest in language and a holistic theory of society, or a holisticmethodology. It all depends upon the view taken of language as a socialphenomenon. My impression is that the view of language underlyingethnomethodological investigations of its uses is individualistic. 38The linguistic turn of sociology has taken two principal forms: (1) languagehas been seen as a phenomenon analogous or similar to society and (2) languagehas been seen as constitutive of social phenomena (Giddens, 1979: 4; Knorr-Cetina, 1981: 2–7). Ethnomethodology has followed both these routes, butmainly the second. There are several sources of the lingustic turn in the humansciences: hermeneutics, structuralist lingustics and ordinary language philosophy.What is the relation of ethnomethodology to these traditions?To begin with hermeneutics, it is easy to see certain superficial similaritiesbetween this tradition and ethnomethodology, but hard to find any deeperkinship between them. Above all, it is difficult to find any sign of direct influenceof hermeneutics upon ethnomethodology. Hermeneutics might, for thepresent purpose, be understood in the crystallised form of the hermeneuticcircle. As a matter of fact, there are two distinct but related circles. The firstcircle is temporal and connects understanding and pre-understanding. Thesecond circle expands and contracts in the context of meaning and connectspart to whole; word to sentence, sentence to text and social action to socialsystem. It is only when applied to the relation between social action and socialsystem that the hermeneutic circle implies a holist methodology (Udehn, 1987:220–6). Ethnomethodology has not, however, made such use of thehermeneutic circle. Its main similarity with hermeneutics lies in stressing theimportance for understanding of pre-understanding. But this is not incompatiblewith individualism or subjectivism. It is intersubjectivism over time. 39Another similarity between hermeneutics and ethnomethodology can be foundin their emphasis on the contextuality of talk and conduct – but more aboutthat below.The similarities to French structuralism are even more superficial. On closerinspection, they boil down to little more than a common interest in language, butfrom different perspectives. It is admitted that French structuralism and

Society as subjectively meaningful interaction 155ethnomethodology rely on different theories of language. While the former relieson Ferdinand de Saussure and is interested in language as a system or structure(language), the latter relies primarily on English ordinary language philosophyand is interested in the use of language (parole). 40 As a general observation, Iwould say that ethnomethodology tends to be averse to all kinds of structuralism,including linguistic structuralism (see, e.g., Douglas, 1971: 31ff and Zimmermanand Wieder, 1971: 107ff).There is one exception to this; Cicourel’s cognitive ethnomethodology. Thisversion of ethnomethodology is explicitly formulated as similar, in certainrespects, not to French structuralism, but to another member of the structuralistfamily; Chomskyan linguistics. Chomsky’s distinction between surface structureand deep structure has a direct counterpart in Cicourel’s distinction between thesurface rules of the normative order and the interpretive procedures used bymembers when applying the surface rules to concrete situations. With this notionof ‘interpretive procedures’, Cicourel provides one answer to the ethnomethodologicalquest for something universal, invariant and trans-situational amidst theotherwise haphazard nature of social reality. Interpretive procedures arecommonly shared and socially learned cognitive schemes operating upon thesurface rules of the normative order and providing a ‘sense of social structure’.The key word here is sense, so it would seem that Cicourel is safely anchored tothe individualist and subjectivist side (Cicourel, 1974: 99ff; see also Knorr-Cetina, 1981: 6f).The linguistic theory (theories) most obviously to be found behindethnomethodology’s interest in language is that of English ordinary languagephilosophy. The influence of this philosophy can be found most expressly inethnomethodology’s Wittgensteinian talk about social practices and in itsprimary concern with the contextuality of talk and conduct. Does not talk aboutsocial practices imply a transcendence of subjectivism and does not a concernwith the contextuality of talk and conduct lead to a transcendence of bothsubjectivism and individualism? Not necessarily.To begin with the second question, it is true that, on a certain analysis, aconcern with contextuality leads to holism, but I cannot find that ethnomethodologistsgo that far in their analysis. It is also true that ordinary languagephilosophers, such as Peter Winch and John R. Searle, have advanced arguments,based upon the contextuality of social action against individualism andsubjectivism. But these are exceptions. In the main, ordinary language philosophyis individualistic and subjectivistic (Strawson, 1970: 4–9), and so, I believe,is ethnomethodology.Ethnomethodology’s term for contextuality is ‘indexicality’. Or, perhapsbetter, indexicality is one form of contextuality. If we look a little bit closer at theethnomethodological analysis and use of ‘indexicality’, we find little to suggestthat it is anything but subjectivistic. The notion of ‘indexicality’ goes back toHusserl, but the immediate source of Garfinkel’s use of this term is the philosopherYehoshua Bar-Hillel. According to him, indexical sentences are such ascontain tensed verbs or expressions like ‘I’, ‘you’, ‘here’, ‘there’, ‘now’,

156 Society as subjectively meaningful interaction‘yesterday’, and ‘this’. Such sentences are dependent for their meaning and truthupon the pragmatic context of their production. By ‘pragmatic context’ Bar-Hillel understands ‘the context which the producer might have had “in hismind”’ (1954: 366, 371). According to Bar-Hillel, then, indexicality goes togetherwith a subjectivist theory of meaning. What matters is the subject’s own ‘definitionof the situation’. This is also the way most, if not all, ethnomethodologistsunderstand ‘indexicality’. The Zimmerman–Pollner–Wieder group, whichamong ethnomethodologists is most concerned with the problem of indexicality,is also the most subjectivistic. 41Garfinkel’s and Sack’s analysis of indexicality might seem a little more difficultto reconcile with individualism and subjectivism. Garfinkel and Sacks takeoff from Bar-Hillel, and as long as they follow him there is no problem to do so.According to Garfinkel and Sacks (Garfinkel, 1967: 4f; Garfinkel and Sacks,1970: 347–50), a characteristic feature of indexical expressions is that theirdenotation is relative to the speaker, or user. If this is the reason why they opposeindexical to objective expressions, it is hard to see why their view of indexicalityis anything but subjectivistic. Problems begin when they start talking about activitiesbeing organisationally situated and category-bound (Garfinkel, 1967: 11,32–4; Sacks, 1972: 221–4). It is, supposedly, this kind of talk, which has led tothe suggestion that ethnomethodology is an expression of methodological situationalism,but not of methodological individualism (Knorr-Cetina, 1981: 7–15).This suggestion is based on a misunderstanding. Methodological individualism,far from failing to refer to ‘interaction in social situations’ is explicitly stated injust those terms. Methodological individualism, then, is not characterised by aneglect of interaction, or of the social situation in which it takes place, but by theway it conceives of interaction and the social situation. The distinguishingfeature of the strong version of methodological individualism is that it sees thesocial situation as consisting of other individuals, but not of social institutions, orsocial structure. In this, there is basic agreement between ethnomethodology andmethodological individualism. As we have already seen, ethnomethodology doesnot accept the social order as something given, to be used by the social scientistas a resource for making explanations. To do this is to be guilty of reifying socialreality; of turning a mere epiphenomenon into something having causal efficacy.Karin Knorr-Cetina also mentions an argument by Maurice Mandelbaum(1955) – similar to those of Winch and Searle – to the effect that the contextualityof social action implies that reference to a social action is also, at leastimplicitly, reference to the institutional context in which it takes place. She alsomakes the observation that while it is the case that micro-sociological approaches,including ethnomethodology, make use of institutional concepts, they have nottaken notice of the holistic implications of this phenomenon (Knorr-Cetina,1981: 12f; cf. also Zimmerman, 1978: 10). The closest ethnomethodology hascome to Mandelbaum’s argument is, probably, Sacks’s analysis of categoryboundactivities. Category-bound activities, according to Sacks, are suchactivities that ‘are taken by members to be done by some particular, or severalparticular, categories of members’ (Sacks, 1972: 222). Activities are tied to cate-

Society as subjectively meaningful interaction 157gories of members by norms and these norms provide the basis for our observationand description of these same activities. ‘Crying’, for instance is done bybabies, but not by ‘big boys’. Sacks’s main interest, however, is in the sequentialordering of activities, especially conversations. A standard example is that ofsaying ‘Hello!’ as a conventional form of ‘greeting’, tied to certain individualsand situations by norms; these conventions and norms being also the basis forcategorizing a token of ‘hello’ as an instance of ‘greeting’ (pp. 224–8). 42 Certainsimilarities notwithstanding, Sacks does not go as far as Mandelbaum in hisargument. Mandelbaum’s decisive argument is not that actions, and descriptionsof actions, are governed by rules, or even that members, therefore, useholistic concepts – methodological individualism can assimilate that – but thatinstitutions make up social systems, so that reference to certain social actionsimplies reference to social systems (see Udehn, 1987: 220–31).Garfinkel’s and Sacks’s writings are also replete with talk about ‘routines’,‘practices’ and ‘social structures’. This terminology – and also more substantialconsiderations – has led some commentators to draw the conclusion thatethnomethodology has abandoned subjectivism for a more objectivist approach,directed at the analysis of language. 43 This conclusion finds support inGarfinkel’s and Sacks’s definition of ethnomethodology’s central notion of‘member’.The notion of member is the heart of the matter. We do not use the term torefer to a person. It refers instead to a mastery of natural language, whichwe understand in the following way … We offer the observation thatpersons, because of the fact that they are heard to be speaking a naturallanguage, somehow are heard to be engaged in the objective production andobjective display of common-sense knowledge of everyday activities asobservable and reportable phenomena. We ask what it is about naturallanguage that permits speakers and auditors to hear, and in other ways towitness, the objective production and objective display of common-senseknowledge, and of practical circumstances, practical actions, and practicalsociological reasoning as well. What is it about natural language that makesthese phenomena observable-reportable, that is accountable phenomena?(Garfinkel and Sacks, 1970: 342)Before discussing the issue of Garfinkel’s and Sacks’s alleged objectivism, acomment on their use of the word ‘social structure’: when Garfinkel and Sacksuse this word they usually refer to structures of everyday activities as, forinstance, the sequential ordering of conversations, but not to social structure asnormally understood by sociologists, involving relations between individuals.This use of the term ‘social structure’, as more or less synonymous with ‘regular’or ‘patterned behaviour’, does not conflict with methodological individualism.Concerning Garfinkel’s and Sacks’s objectivism, I venture the suggestion that itis, largely, a matter of terminology. This might seem too rash, but there certainly isa problem of double talk involved, at least, in the case of Garfinkel. If practical

158 Society as subjectively meaningful interactionactions are routinely performed, in what sense are they ‘accomplishments’? Ifpractical actions are objective phenomena, why this obsession with the problem ofmaking them accountable, i.e., understandable, intelligible? And is there reallyanything about natural languages that make practical actions accountable? Atother times you get the impression that the whole problem of making practicalactions accountable arises, for ethnomethodology, because of the limitations ofnatural languages, especially their indexicality. Garfinkel, for instance, talks aboutthe ‘essential incompleteness’ of any set of written instructions, while, according toGarfinkel and Sacks, indexical expressions are ‘obstinate nuisances’ to logiciansand linguists (Garfinkel, 1967: 29; Garfinkel and Sacks, 1970: 349). Is it not thecase, therefore, that members’ methods for making their actions accountable haveto be some kind of commonly shared, extra-linguistic interpretive rules or procedures,as in the case of Cicourel, or else, a complete mystery? (see Garfinkel, 1967:25–31, 38ff, 53ff, 262ff). 44 There is also the fact that Garfinkel’s writings abound,not only in objectivist talk, but in subjectivist talk as well. Talk about ‘commonunderstandings’, ‘expectancies’ and ‘definitions of the situation’ is not easilyreconcilable with an objectivist view of society and an objectivist methodology.As a matter of fact, ethnomethodology is also highly critical of methodologicalobjectivism, at least as manifested in traditional empiricist sociology. Being aradical movement, it criticises traditional empiricist sociology for not beingempiricist enough (Hindess, 1973a: 9ff; Collins, 1985: 205–11). The main chargeis that traditional empiricist sociology has not taken seriously enough the problematic,because subjective, nature of their data, and what follows from this inthe form of limitations of their measurement (Cicourel, 1964: 7–38). Differentcures have been suggested, from ceasing to do traditional empiricist sociologyaltogether, to suggestions of the repair of its data. Connected to this charge is asecond, for our purposes even more interesting one. If the data of empirical sociologyare subjective, they are, necessarily, also individualistic. The rates andmeasures arrived at by the statistical method of sociologists and officials, therefore,must not be mistaken for social facts about society.Once we follow the ‘disembodied numbers’ back to their sources to see howthey were arrived at and what, therefore, they represent, we find that theyare based on the most subjective of all possible forms of activity. This isespecially true in the case of suicide statistics, which are the result of coroners’evaluations of the ‘intentions’ of the actors involved.(Douglas, 1971: 6f) 45The address of this critique is, of course, Emile Durkheim’s famous study ofsuicide. I will not now go into any detailed discussion concerning the correctnessof this critique – I think it contains an element of truth, while misrepresentingthe position of Durkheim – only note that it raises some important questionsabout the use of statistical techniques in social research in general, and inempiricist social research in particular.Summing up the discussion in this section, it seems that the ethnomethod-

Society as subjectively meaningful interaction 159ological movement started as a continuation of phenomenology. If so, it wouldseem to be safely anchored to the subjectivist tradition of social thought. AsRogers (1983: ch. 7) has argued convincingly, however, ethnomethodology ismuch less concerned with subjective meaning than is phenomenology, and muchmore concerned with action, or practice. A possible reason for this is thatethnomethodology is also influenced by the later philosophy of LudwigWittgenstein. Nevertheless, I believe that it would be wrong to conclude thatethnomethodology takes an objectivist approach to society, or even to language.The influence from phenomenology is too strong:The phenomenological approach remains within the limits of the analysis ofconsciousness. This is why Cicourel and Garfinkel do not take the obviousstep from analysis of the lifeworld … to linguistic analysis. They are not ableto recognize the obvious rules of the grammar of language games in thestructure of consciousness. It is easy to show the systematic basis for thisincapacity in the work of Schutz; the trail leads directly to Husserl.(Habermas [1970] 1988: 116)In view of these considerations, I risk the conclusion that ethnomethodology is,indeed, the ‘Californian way of subjectivity’ (Gellner, 1975).Ethnomethodology has met with much resistance from mainstream macrosociologists.A common critique is that ethnomethodology is individualistic andsubjectivistic. This critique has been vehemently rejected by some of its adherents,who suggest that it is based on a serious misunderstanding of whatethnomethodology is really about. According to Sharrock and Anderson (1986:48ff, 99ff), for instance, it is not the concern of ethnomethodology to criticisesociology, but to engage in a different type of foundational investigation, whichleaves sociology as it is. 46 ‘Popular sociological opinion to the contrary,ethnomethodology does not contradict Durkheim’s famous proposal that weshould treat social facts as things, that we should recognize social structures asexternal and objective, environments capable of constraining our conduct’(Sharrock and Andersson, 1986: 48).Garfinkel, himself, has defended ethnomethodology against the argument ofAlexander, but his defence does not lend unambiguous support to Sharrock andAndersson.For [Parsons’s] The Structure of Social Action, Durkheim’s aphorism is intact:‘The objective reality of social facts is sociology’s fundamental principle’.For ethnomethodology the objective reality of social facts, in that, andjust how, it is every society’s locally endogenously produced, naturally organised,reflexively accountable, ongoing, practical achievement, beingeverywhere, always, only, exactly and entirely, member’s work, with not timeout, and with no possibility of evasion, hiding out, passing, postponement,or buy-outs is thereby sociology’s fundamental phenomenon.(Garfinkel, 1991: 11)

160 Society as subjectively meaningful interactionOne obvious way to read this quotation is as a statement of an individualistalternative to Durkheim’s and Parsons’s holistic theory of society. But even if thisreading is wrong, there is abundant evidence to show that ethnomethodologycomprises a view of society, which is, in the main, radically individualistic, at leastimplicitly (see Collin, 1997, ch. 1). From phenomenology, it inherited an intersubjectivistview of society, but like symbolic interactionists, ‘the ethnomethodologistsemphasize the interactional activities that constitute the social facts’ (Coulon,1995: 50, see also 71ff). In my opinion, ethnomethodology is even more focusedon action, or practices, than is symbolic interactionism and, therefore, an evenmore radical form of micro-sociology (cf. Collins, 1981b: 81–3).Figure 5.3 Society according to ethnomethodologySocial constructionismA recent vogue in the human sciences is called social ‘constructionism’, or‘constructivism’. In a broad sense, ‘social constructionism’ means simply thatsocial phenomena are human creations, rather than natural phenomena, or‘essences’. In this somewhat trivial sense, virtually all human sciences, and sociology,in particular, are forms of social constructionism. It has always been afundamental thesis of the human sciences that society is made up of ‘conventions’,or ‘customs’, or social ‘institutions’, words that imply that society ishuman-made. What distinguishes recent social constructionism from thisbroader contructionism of all human sciences, with the possible exception ofeconomics, is that it sees social phenomena as cultural; cognitive and/orlinguistic, constructions. Social phenomena are constituted by our thinking andour discourse about them. 47Social constructionism, in the above sense, is not the same type of rationalistconstructionism that Hayek objects to in his writings. The latter type is based onthe belief that it is possible to construct society intentionally and rationallyaccording to a preconceived plan. This is not at all the belief of recent socialconstructionists, who agree with Hayek that most social phenomena are unintendedconsequences of human actions. Actually Hayek, himself, was a ‘social’constructionist in the recent sense of this term. His phenomenological view thatsocial phenomena are constituted by people’s beliefs about them is the centralpillar of, at least, one form of social constructionism.

Society as subjectively meaningful interaction 161Above all, it is a central pillar of the main classic of social constructionism;The Social Construction of Reality (1966) by Peter L. Berger and Thomas Luckmann.The intellectual roots of this work are to be found mainly in the intersubjectivisttradition in social science, discussed in this chapter, but with an important dose ofobjectivism and structuralism added to it. Berger and Luckmann’s theory ofsociety is a synthesis of ideas taken from Weber, Mead, Durkheim and Marx, butbuilt mainly upon a foundation laid by the phenomenology of Schutz. Accordingto Berger and Luckmann, society is a human product. This is fully in line withtheoretical individualism and subjectivism. But it is also the case that humanbeings are social products. This is a more doubtful argument, from the point ofview of an individualist theory of society. But, of course, it all depends upon themeaning of the word ‘social’. According to Berger and Luckmann, society is anobjective reality, possessing structure, and consisting of functionally integratedinstitutions based upon division of labour ([1966] 1971: 69, 79–82, 95f, 134–7).This is not an intersubjectivist theory of society. It is not even an individualisttheory of society. Intersubjectivity, which was so important to Schutz, plays noimportant role at all in Berger’s and Luckmann’s analysis. Instead, they speak ofsociety as an objective reality, which is irreducible to the subjective reality of individuals([1966] 1971: 150–4, 183). It may be added that the particular branchof social constructionism, most influenced by Berger and Luckmann; the newinstitutionalism in the theory of organisation is decidedly holistic, even ifsome of its main representatives do recognise the need for suitable microfoundations.These are to be found in the theories treated in this chapter, or intheories akin to them, but not in the theory of rational choice. 48More specifically, Berger and Luckmann recognise two sides of society, onesubjective and one objective, standing in a dialectical relation to each other. Thisview is not strictly individualistic. Following Roy Bhaskar (1979: 40), I suggest thefollowing graphic representation of Berger and Luckmann’s view of the relationbetween individual and society.Figure 5.4 Berger and Luckmann’s dialectic of individual and societyIt is clear that Berger and Luckmann differ from Austrian methodologicalindividualism, which only admits of causal arrows leading from individuals tosociety. In the view of the Popperian philosopher J.O. Wisdom, Berger andLuckmann’s approach is closely allied to Popper’s ‘situational individualism’

162 Society as subjectively meaningful interaction(Wisdom, 1973: 263). The problem with this view is that Popper’s so-called ‘situationalindividualism’ is a creation of Wisdom, himself. Popper, as we shall see,advanced two separate methodologies: methodological individualism and institutionalism,which he never joined and could not join without changing hismethodological individualism, so as to make it compatible with his institutionalism.This was achieved, instead, by his pupil Joseph Agassi, who createdinstitutional individualism. Situational individualism is, I believe, more or less,identical with institutional individualism. After this short digression, is it possibleto pass judgement on Wisdom’s suggestion? Yes, I believe, it is possible to seeBerger and Luckmann’s theory of society as compatible with the weak version ofinstitutional, or situational individualism, but not with the strong version ofmethodological individualism, as formulated by Popper.Social constructionism today is an influential position in many areas of thehuman sciences, but especially in sociology and social psychology. The areasmost influenced by social constructionism are, probably feminism, sociology ofscience, the theory of organisations and the theory of the personality, or self.I believe that the most fertile use of social constructionism has been in feminism,where it has been part of a struggle against patriarchy, or male dominance(cf. Hacking, 1999: 7ff). I have already quoted the famous words of Simone deBeauvoir: ‘One is not born, but rather becomes a woman’. This statementcaptures the ‘kernel’ –it would be wrong to use the word ‘essence’ –of socialconstructionism. The question remains, of course, how one becomes a woman,and on this point opinions diverge. The early Sartre would probably have maintainedthat human beings of the female sex chose to become women. DeBeauvoir suggested that they had little, or no choice at all, to escape from being‘women’ in the traditional sense. It is possible to divide also recent socialconstructionism into two main streams: one flowing from the interactionist andintersubjectivist tradition I have treated in this chapter, the other from Frenchstructuralism and post-structuralism. 49 A good illustration of the differencebetween these strands of social constructionism is provided by two influentialfeminist works: S.J. Kessler and W. McKenna, Gender. An EthnomethodologicalApproach (1978) and J. Butler, Gender Trouble. Feminsim and the Subversion of Identity(1990). In the former work, the authors approach the social construction ofgender in an ethnomethodological manner, as a matter of gender attribution inthe everyday interaction between members of society. In the latter work, thetheoretical framework is provided by structuralism and post-structuralism and thesocial construction of gender is the work of culture and discourse, even if Butlerhas to find a place for agency in order to make room for the possibility of change.As expected, sociologists have tended to prefer the objectivist and structuralistversion of social constructionism, whereas social psychologists have tended tofollow the subjectivist and interactionist version (see, e.g. Sarbin and Kitsuse(eds), 1994). Most common is, perhaps, to attempt some kind of synthesis, or viamedia, between the two versions. Kenneth J. Gergen and Vivien Burr, forinstance, opt for a methodology that avoids the pitfalls of both methodologicalindividualism and social holism. The former suggests that we should concentrate

Society as subjectively meaningful interaction 163on social relationships, rather than on autonomous individuals or social wholes(Gergen, 1994: 214ff). The latter makes a distinction between ‘bottom-up’ and‘top-down’ conceptions of society. According to the first, which is that ofmethodological individualism, individuals determine society, while in the lattersociety determines individuals.both top-down and bottom-up conceptions of the relationship between theindividual and society are problematic for social constructionism. The topdownview leaves discourse as a side-effect of social structure, and ittherefore cannot be the focus for social change. The bottom-up view, worsestill, cannot accommodate any kind of social constructionism, since the individualis taken to be logically prior to the social. The individual is a ‘given’from which society arises, and which therefore cannot be said to beconstructed by that society. This methodological individualist view has all theattributes fiercely contested by social constructionists. It is humanistic andessentialist, claiming for the human being an essential nature, a coherent,unified self, and the capacity to make self-originated choices and decisions.(Burr, 1995: 96f)This argument to the effect that methodological individualism is contested bysocial constructionism is weakened by a narrow conception of the former. Whileit is true that there is a radical version of methodological individualism, whichtakes the individual as prior to society, this is not the only version. In fact, themost well-known methodological individualists – Weber, Mises, Hayek andPopper – do not at all defend this radical version of methodological individualism,represented by the theory of the social contract and the theory of generalequilibrium in economics. On the contrary, they agree that human individualsare social beings, as suggested by Burr, and as implied by the individualistictheory of society discussed in this chapter.Finn Collin is much closer to the truth when he suggests that the constructivist‘claim that social facts are determined by what agents think … points in the directionof an individualist approach to social research’ (1997: 229). He also observesthat, at first sight, there is ‘much in social constructivism to please a methodologicalindividualist and nothing to offend him’ (p. 230). His conclusion, however, is thatsocial constructionism is not a reductionist form of methodological individualism.We have seen that there are both individualistic and holistic versions of socialconstructionism and, also, that there are many attempts, either to combine thetwo, or to dissolve the dualism. I will end this chapter by a brief mention of twosuch attempts, by two well-known sociologists, who seem to take a similarapproach: Pierre Bourdieu and Anthony Giddens.Pierre Bourdieu characterises his own work as ‘constructivist structuralism’ or‘structuralist constructionism’ (1989: 14). Like Berger and Luckmann, he maintainsthat there are two sides to society: one subjective, phenomenological sideand one objective, structuralist side, both dialectically related ([1972] 1977: ch. 2;[1980] 1990: 25ff). However, if Berger and Luckmann leaned most towards the

164 Society as subjectively meaningful interactionformer, I think it is fair to say that Bourdieu leans more towards the latter. 50 Itshould also be pointed out, as a possible difference between them, that Bourdieurejects a dualistic interpretation of the objective and subjective sides of society.Bourdieu also addresses the issue of methodological individualism versusholism, both of which he rejects, as the expression of a mistaken conception ofthe relation between individual and society (Bourdieu and Wacquant, 1992:126f). Or rather, there is no such relation at all, since individual and society arenot two distinct realities. There are only social individuals related to one another.The position of Bourdieu, therefore is that of methodological relationism (1985:16ff; Bourdieu and Wacquant, 1992: 228ff).Anthony Giddens shares Bourdieu’s negative attitude towards dualisms, likeindividual–society, subject–object and action–structure. His own way out is theso-called theory of structuration, which owes more to ethnomethodology than tophenomenology. Like Garfinkel, Giddens sees society as chronically produced byknowledgeable individuals (1979; ch. 2; 1984: xxff, ch. 1). In Giddens’s terminology,human activities are ‘recursive’. Thatistosay,‘they are not brought intobeing by social actors but continually recreated by them via the very meanswhereby they express themselves as actors’ (1984: 2). It seems to me that Gidden’stheory of society is strictly individualistic. However, this does not lead him explicitlyto embrace methodological individualism. He criticises Popper, for assumingthat it is possible to state methodological individualism, without making clearwhat is to be understood by the term ‘individual’. If we assume, however, thatthe individual is a social agent, then, the collective is implicated in the individualand two methodologies become possible: (1) methodological individualism,which focuses on strategic conduct, while neglecting the institutional aspect ofaction and (2) institutional analysis, which focuses on the rules and resources thatare reproduced by individuals, but neglects the deliberations of individuals themselves(Giddens, 1979: 94f). Institutional analysis is not structuralism, however.Giddens (1984) objects strongly to the traditional sociological structuralism ofEmile Durkheim (pp. 169ff) and Peter Blau (pp. 207–13). But, once again, he isnot led by this to embrace methodological individualism, or is he?The methodological individualists are wrong in so far as they claim thatsocial categories can be reduced to descriptions in terms of individual predicates.But they are right to suspect that ‘structural sociology’ blots out, or atleast radically underestimates, the knowledgeability of human agents, andthey are right to insist that ‘social forces’ are always nothing more andnothing less than mixes of intended or unintended consequences of actionsundertaken in specifiable contexts.(Giddens, 1984: 220)This is a very weak opposition to methodological individualists, since most ofthem would readily admit that a full individualist description of society is notpossible. The important thing for most methodological individualists, thePopperians, in particular, is explanation, not description.

Society as subjectively meaningful interaction 165It seems to me, that Giddens, like Bourdieu, wants to tread some via mediabetween a radical methodological individualism and traditional sociologicalstructuralism, but ends up closer to the former than to the latter. Today, there aremany self-proclaimed methodological individualists, who adopt a more structuralistposition than does Giddens in his theory of structuration.ConclusionThe theories discussed in this chapter are united by taking an intersubjectivistand interactionist approach to society. They differ, however, in the extent towhich they emphasise either the (inter)subjective, or the (inter)active element asconstitutive of society. Phenomenology focuses on the intersubjective nature ofsociety. Ethnomethodology is most concerned with the actions, or practices,which constitute society. Symbolic interactionism, finally, seems to attach equalweight to subjective meaning and interaction. According to Hans Joas([1980] 1885: Ch. 9), it is the advantage of symbolic interactionism overphenomenology, that it explains how intersubjectivity arises in symbolic interaction,instead of taking it as given to phenomenological analysis.The theories discussed in this chapter differ from mainstream sociology,which tends to conceive of society as an objective reality made up of culturalobjects, social institutions and social structures. The theories dicussed in thischapter conceive of (reduce) culture to intersubjectively shared meanings andsocial structure to social interaction. To the extent that culture and social structurecan be said to exist, at all, they are epiphenomenal. What is real are thethoughts, feelings, desires and actions of individual human beings.Figure 5.5 A mapping of some sociological theories in a two-dimensional space

166 Society as subjectively meaningful interactionA final note of caution: it is not the case, that all representatives of the theoriesdiscussed in this chapter actually deny the existence of objective culture andsocial structure, but I agree with Jeffrey Alexander that, even if they do not, theynevertheless tend to treat them as residual categories in their theories and tointroduce them in an ad hoc manner, without theoretical grounding (Alexander,1985: 27).

6 Positivism in philosophy andsocial scienceAs we have seen (p. 27), the term ‘positivism’ was coined by Auguste Comte, whowas not a methodological individualist. Nor were the other positivists among theclassical sociologists, with the possible exception of Herbert Spencer. This indicatesthat there is not a necessary connection between these two doctrines. Morerecent positivist social science, however, is more indebted to British empiricismand the logical positivism of the Vienna Circle. This influence works in a moreindividualistic direction.Positivist philosophyBy ‘positivism’, I understand a philosophy with two basic characteristics: (1)Methodological monism, which is the doctrine that the method of all sciences is oneand the same. More specifically, methodological monism means that the socialsciences must use the same method as the natural sciences. (2) Empiricism, whichmeans that all knowledge ultimately derives from observation, or experience.From the very beginning, positivism was advanced as an alternative to metaphysicsand this remained an essential feature of this philosophy. In the case oflogical positivism, the demarcation of science from metaphysics was made withthe help of the empiricist criterion of verifiability. Methodological individualismsometimes looks like just another variant of the empiricist attack on metaphysics:a bit less radical than phenomenalism and, perhaps, also than physicalism, but,nevertheless, motivated by the same kind of epistemological considerations. Itcould be argued that flesh and blood human beings are the only directly observableentities in society – if we exclude human artefacts such as buildings,machines, books, etc. – while social wholes and collectives are not in the sameway directly given to the senses. 1 It would seem, therefore, that empiricismimplies or, at least suggests, methodological individualism. 2 This relation is asymmetric,however, and does not hold the other way around. Methodologicalindividualism does not imply empiricism and, as a matter of fact, few of themost well-known advocates of methodological individualism have been empiricists.As we have seen in previous chapters, most of them have beenhermeneuticians, neo-Kantians, phenomenologists and existentialists. As such,they have adopted a subjectivist version of methodological individualism, which

168 Positivism in philosophy and social scienceconflicts with the physicalism and behaviourism ofphilosophers. 3most empiricistBritish empiricismIn the history of philosophy, empiricism has been mainly a British affair. Thebeginning is in the Middle Ages with William of Ockham and Roger Bacon. Inthe Renaissance, there was Francis Bacon, sometimes conceived of as the ‘fatherof empiricism’, because of his advocacy of induction. On the threshold to theEnlightenment, Thomas Hobbes defended a kind of empiricism, which incorporatedan important deductive element from Descartes’s rationalism. Britishempiricism culminated with the contributions of the three Enlightenmentphilosophers, John Locke, George Berkeley and David Hume. After them, JohnStuart Mill was an important bridge between seventeenth and nineteenthcenturyempiricism. Mill, once again, introduced an important element ofdeduction in his empiricist methodology, which makes it very close to that of thelogical positivists.Common to the most important British empiricists is that they believed thatall knowledge derives from sense perception and that sense perception is atomistic.Knowledge is built out of simple ideas of sense-data. These simple ideasare combined into complex ideas, by a psychological mechanism of association.All complex ideas are ultimately collections of simple ideas of sense-data.Complex ideas of social wholes and collectives, in their turn, are collections ofideas about particulars, that is about individuals. In the nineteenth century thisanalysis is turned into the idea that social wholes and collectives are logicalconstructions of individuals – a view which is congenial to most methodologicalindividualists.The idea of logical constructions goes back to Bertrand Russell, whosuggested in Our Knowledge of the External World (1914) that matter, physicalobjects, when analysed, turn out to be logical constructions out of sense-data (seealso [1914] 1926: 106). 4 It is not that Russell denies the existence of physicalobjects. His point is rather that if we can construct them logically out of sensedata,there is no need to assume that they exist except as logical constructions.Russell’s method of analysis or logical construction, then, is used to serve themaxim called Ockham’s razor: ‘Entities are not to be multiplied without necessity’,or in Russell’s version; ‘Wherever possible, logical constructions are to besubstituted for inferred entities’ (p. 112). But the method of logical constructionis also a method of justification, since, in a way, our belief in the existence ofphysical objects is justified to the extent that it is possible to replace them bylogical constructions out of sense-data (Pears, 1968: 37f; Ayer, 1971: 40; 1972:11).The reason Russell calls physical objects logical constructions is that they canbe analysed as series of classes of appearances or sense-data, and series andclasses are things that belong to logic ([1914] 1926: 111, 128; [1917] 1963:

Positivism in philosophy and social science 169114–18, 97, 106f). Explaining how it comes that a chair, for instance, can be aseries of classes of sense-data, Russell says:A chair presents at each moment a number of different appearances. All theappearances that it is presenting at a given moment make up a certain class.All those sets of appearances vary from time to time … So you get a seriesin time of different sets of appearances, and that is what I mean by sayingthat a chair is a series of classes.(Russell [1918] 1972: 134)Russell’s theory of logical constructions cannot be understood unless you arealready acquainted with his theory of definite descriptions, his theory of knowledgeand his logical atomism. What follows is an extremely compressedpresentation of some of Russell’s main ideas, which, of course, fails to do justiceto this epoch-making contribution to twentieth century philosophy.Russell’s theory of descriptions was advanced, for the first time, in his famousarticle ‘On Denoting’ (1905), as an attempt to come to grips with denotingphrases, such as ‘the present King of France’, which lack denotation. How can adenoting phrase, which lacks denotation have a meaning? Russell’s answer is‘that denoting phrases never have meaning in themselves, but that every propositionin whose verbal expression they occur has a meaning’ (p. 480). The basicidea of the theory of definite descriptions is to replace the original propositionby a propositional function, in which the denoting phrase does not occur, and,then, to expand it into an existential statement to the effect that what is describedexists. (Russell held that descriptions imply an assertion that the described objectsexist.) Take the proposition ‘the present King of France is bald’, replace it by thepropositional function ‘x is now King of France and x is bald’, and expand it intothe existential statement: ‘There is exactly one x, such as x is now King ofFrance, and x is bald’. After this transformation, the problem with denotingphrases lacking denotation is solved. Since there is no present King of France,the expanded existential statement is simply false ([1917] 1963: 164f; [1918]1972: 99–112; 1924: 148).In his exposition of the theory of descriptions, Russell uses as examples ordinarynames, such as Scott, Bismarck, London, etc., as if they really haddenotation. But these names are used only for the purpose of illustration.Ordinary names are really descriptions, or rather, abbreviations for descriptions;truncated descriptions. They are what Russell calls, ‘incomplete symbols’ ([1917]1963: 156–8; [1918] 1972: 56, 100, 111f). But if ordinary names are reallydescriptions, what is left? The only logically proper names that Russell permitsare ‘I’ and ‘this’, or ‘this’ and ‘that’, and the only things they can name are sensedata([1917] 1963: 162; 1924: 56).The reason why logically proper names can only denote sense-data, but notobjects in the external world, must be sought in Russell’s theory of knowledge, inhis empiricism and scepticism. 5 According to Russell’s empiricism, all our knowledgederives, ultimately, from ‘knowledge by acquaintance’. ‘Every proposition

170 Positivism in philosophy and social sciencewhich we can understand must be composed wholly of constituents with whichwe are acquainted’ ([1917] 1963: 159; see also [1912] 1951: 46–59). Among thethings with which we are supposed to be acquainted, particular sense-data aremost important, but we also know universals in this way. Sense-data are suchthings as colours, noises, shapes, etc., but not as universals, only the particularnoise etc., which is the immediate data of my senses right here and right now.While the particular yellowness-of-this is a sense-datum, the yellowness that weknow through acquaintance with several yellow objects is universal. 6In his theory of knowledge, Russell admits of sense-data being complex, butaccording to his logical atomism, particulars, qualities and relations are simpleobjects ([1917] 1963: 155; [1918] 1972: 129). Simple objects are such objects ascannot be symbolised otherwise than by simple symbols, and simple symbols arethose symbols whose parts are not symbols ([1918] 1972: 48f). Logically propernames, then, are symbols for particulars, while predicates are symbols for qualities,and verbs, sometimes single, sometimes whole phrases, are symbols forrelations. Combinations of simple objects (particulars, qualities, relations) makeup atomic facts, and the propositions expressing these atomic facts, containingonly simple symbols (logically proper names, predicates and verbs), are calledatomic propositions (pp. 53–5). A language built up according to these stricturesis a logically perfect language.In a logically perfect language there will be one word and no more for everysimple object, and everything that is not simple, will be expressed by acombination of words, by a combination derived, of course, from the wordsfor the simple things that enter in, one word for each simple component.(Russell [1918] 1972: 52)The criterion of simplicity, then, is language, and it must not be assumed thatthe objects of simple symbols are simple in an absolute sense.When I speak of ‘simples’, I ought to explain that I am speaking of somethingexperienced as such, but known only inferentially as the limit ofanalysis … A logical language will not lead to error if its simple symbols …all stand for objects of some one type, even if these objects are not simple.The only drawback to such a language is that it is incapable of dealing withanything simpler than the objects which it represents by simple symbols.(Russell, 1924: 158)We have seen that Russell’s theory of descriptions leads to the elimination ofordinary names, such as Scott, Bismarck, London, etc. Such names are abbreviationsfor descriptions, and the things they name can be reduced through theanalysis of facts which are apparently about those things ([1918] 1972: 47). Ithas been argued by A.J. Ayer that Russell’s theory of descriptions leads to theelimination of bare particulars and, therefore, of logically proper names as well(Ayer, 1971: 43–7; 1972: 104–106). According to Ayer, ‘We are left with descrip-

Positivism in philosophy and social science 171tive signs, which stand for properties or groups of properties at various logicallevels; with demonstrative signs, which are neither names nor predicates butsignals, which simply do the work of orientation’ (1971: 47). I think this conclusionis correct, as evidenced by the development of Russell’s later philosophy.Logically proper names, we have seen, are symbols for particulars, as distinctfrom qualities and relations. Characteristic for these bare particulars (or particularparticulars, as distinguished from general particulars), according to Russell,is that they stand entirely alone and are entirely self-subsistent, much as the oldnotion of ‘substance’, except that they persist through a very short time ([1918]1972: 56f; 1924: 158). Logically proper names, then, are ‘words which do notassign a property to an object, but merely and solely name it’ ([1917] 1963: 162).Now, according to Russell, the simplest imaginable fact is the possession of aquality by some particular thing, as expressed by the proposition ‘This is white’([1918] 1972: 53). It is hard to see how the name ‘this’ can have any denotationas distinct from the quality white (and all other qualities possessed by this particularthing). It would seem that the substance of an object, the bare particular isexhausted by an enumeration of all facts about it. If so, logically proper names,like ordinary names, appear as abbreviations for descriptions. The function ofnames becomes demonstrative rather than denotative. Names become signalsused to point out groups of qualities, while their original task is taken over bydescriptions. ‘Logically proper names’, in the original sense, have becomedispensable (see Ayer, 1971: 43).In his later writings, Russell follows the logical implications of his theory ofdescriptions and his logical atomism and drops the commitment to a belief inthe existence of particulars. Things have no substance behind or distinct fromtheir qualities. They are nothing but bundles of coexisting qualities. The mainreason for taking this view is that ‘we experience qualities but not the subject inwhich they are supposed to inhere’ ([1940] 1973: 89–93; 1959: 120–7). But ifparticulars are abolished, logically proper names must go with them, since onlyparticulars can have names. Russell knows this, of course, but chooses to changehis theory of names, since he is doubtful about the possibility of inventing alanguage without names. This new theory is very liberal. ‘For our purpose, unlessreason should appear to the contrary, we may accept as a name whatever wouldordinarily be considered as such: Tom, Dick and Harry, the sun, the moon,England, France, etc.’ ([1940] 1973: 89–91).Russell was not interested in the logical construction of social phenomena outof individuals. His primary interest was in the logical construction of any kind ofcomplex entity out of sense-data. He did use, as one of his stock examples of acomplex entity, ‘Rumania’, but he never attempted its reduction to individuals.Nor, for that matter, did he attempt its reduction to sense-data, although he didhold that it was so reducible.It was probably John Wisdom who made it a main preoccupation amongBritish analytical philosophers to logically construct social collectives, especiallynations, and among nations, especially England, out of individuals. According toWisdom, when speaking about social collectives, such as nations, or institutions,

172 Positivism in philosophy and social sciencesuch as the Church, we are either speaking indirectly about individuals, or aboutsomething over and above individuals. He finds the first alternative both more‘attractive’ and less ‘objectionable’, and concludes that collectives, such asnations, are logical constructions of individuals. ‘For if we have said all there isto be said about Tom, Dick and Harry, including of course statements abouttheir interrelations, we have said all there is to be said about England, even ifnone of our statements mention England’ (1933b: 176). It is possible, therefore,to substitute statements about individuals for statements about nations. Thelatter can be translated or reduced by way of definition to the former. ‘If anation is reducible to its nationals (and it is) then statements about nations, i.e.predicates applicable to nations are definable in terms of predicates applicable totheir nationals’ (p. 10).Wisdom’s examples of logical construction – and this has become tradition –are all, or nearly all, about acts of hostility between nations. His first example is‘England declared war’, which is translated, without any claim of exactness, as‘Englishmen had selected a man (Minister for War) who decided that they shouldfight’ (1931: 192). The motive for this aggression was presumably that ‘EveryEnglishman fears a Frenchman’, which is substitutable for ‘England fears France’(1933a: 44ff). But confronted with such a fearful enemy as France, you needallies. Consequently, according to Wisdom, ‘Every nation invaded France’, whichis rendered ‘Every group of individuals with common ancestors, traditions andgovernors forcibly entered land owned by Frenchmen’ (1934: 77).But why should we substitute statements about individuals for statementsabout nations? In the case of Wisdom, as in that of Russell, the answer must besought in his empiricist epistemology. Philosophy is analysis and analysis is ostentation;the substitution of more ostensive for less ostensive sentences. Now,according to Wisdom, sentences about individuals are more ostensive than aresentences about nations. This is so, because individuals are, in Wisdom’s terms,(epistemically) more ‘fundamental’ or ‘ultimate’ than are nations. But most ultimateare sense-data. Another way of expressing this is to say that facts aboutindividuals are (epistemically) ‘primary’ relative to facts about nations, but‘secondary’ relative to facts about sense-data (1931: 212–16; 1933a: 187–9,195–9; 1934: 75ff). Unlike Russell, however, Wisdom seems quite satisfied toaccomplish the logical construction of nations out of individuals, and doesn’tbother much to achieve the further reduction of individuals to sense-data.The idea of logical constructions was adopted by Alfred J. Ayer in his bestsellerLanguage, Truth and Logic (1936) where he denies that they are fictitiousobjects.For while it is true that the English State, for example, is a logical constructionout of individual people, and the table at which I am writing is a logicalconstruction of out of sense-contents, it is not true that either the EnglishState or this table is fictitious, in the sense in which Hamlet or a mirage isfictitious.(Ayer, 1936: 74)

Positivism in philosophy and social science 173Ayer was a philosopher with one leg in the tradition of British empiricism andthe other leg in logical positivism. His book Language, Truth and Logic was generallyconceived of as an introduction to the latter and contributed significantly toits great success. As we shall see in the next chapter, Ayer is also the source of theidea amongst some Popperian methodological individualists that collective entitiesare logical constructions.Logical positivismLogical positivism is the name of the philosophy of the Vienna Circle. It is socalled, because it combines logical analysis of language with a positivist, orempiricist, theory of knowledge. To the original Vienna Circle belonged anumber of famous philosophers, such as Moritz Schlick, Otto Neurath, RudolfCarnap, Kurt Gödel, Herbert Feigl, Gustav Bergmann and Friedrich Waismann.Close to the Vienna Circle were many other important philosophers, such as A.J.Ayer, Gustav Hempel, Hans Reichenbach, R.B. Brathwaite and Ernest Nagel.The logical positivists also counted the early Wittgenstein, Bertrand Russell andKarl Popper as their allies, but none of them accepted the invitation to join themovement.The main objective of the logical positivists was to put an end to metaphysicsin science and philosophy. By metaphysics was understood every statement thatcannot be reduced to immediate sense experience. The implication for socialscience was that concepts and laws, which do not refer directly to the observablefeatures of human beings, are suspect and must be abolished, or redefined interms of facts about individuals (cf. Hempel, 1969a). Already in the pamphletthat constituted the Vienna Circle, it was stated that concepts, such as ‘folkspirit’, should be dropped and replaced by concepts referring to ‘groups of individualsof a certain kind … The object of history and economics are people,things and their arrangement’ (Neurath et al., 1929: 315).The most ambitious, not to say heroic, attempt to construct the world, logically,out of immediately given sense-data, or ‘elementary experiences’ was madeby Rudolf Carnap in his The Logical Structure of the World (1928). By logicalconstruction, Carnap understands ‘explicit definition’ in a broad sense,including, most importantly, definitions in use. An explicit definition is such thatall statements about the object (or concept) defined (definiendum) can be eliminatedand replaced by statements about the elements in terms of which it isdefined (definiens) ([1928] 1969: 5–10, 61ff). The objective of logical construction,according to Carnap, is the reduction of all objects to those that are epistemicallybasic (pp. 78ff). As the basic elements of his system Carnap chooses‘elementary experiences. But even more fundamental is the basic relation ‘recollectionof similarity’, out of which they are constructed. ‘Recollection ofsimilarity’ is the basic undefined concept with which Carnap wants to constructthe whole world (pp. 98ff, 122ff).Of special interest from the point of view of the present investigation isCarnap’s treatment of ‘cultural objects’. The logical construction of cultural

174 Positivism in philosophy and social scienceobjects, according to Carnap, rests upon the relations of manifestation anddocumentation. The manifestation relation is that between a cultural object andits (psychological) manifestations in the persons who are its ‘bearers’. The documentationrelation is that between a cultural object and its representation inphysical objects, in artefacts, works of art and written documents. Of these relations,the manifestation relation is primary in that documentation is reducible tomanifestation (Carnap [1928] 1969: 39–41). It is the thesis of Carnap that ‘everycultural object is reducible to its manifestations, that is, to psychological objects’. In otherwords: ‘In principle, all statements about cultural objects can be transformedinto statements about psychological objects’ (pp. 89–92).Despite this possibility in principle, Carnap admits of certain difficulties whenit comes to actually carrying out the logical construction of cultural objects fromits psychological manifestations. We have to rest content, therefore, with an indicationof the general form of such logical constructions. Carnap makes adistinction between primary and higher cultural objects. Primary cultural objectsare ‘those objects whose construction does not presuppose the construction ofother cultural objects’, but which ‘are always constructed on the basis of theirmanifestations … i.e., on the basis of those psychological events in which theyare actualized or become apparent’ (Carnap [1928] 1969: 230). As an exampleof a primary cultural object, Carnap mentions the custom of greeting, which issaid to consist in the disposition on the part of the members of a society to act ata certain time in a certain way in a certain type of situation (p. 231). It wouldseem that Carnap understands, by ‘primary cultural objects’, the institutionalisedbehaviour of individuals.The higher cultural objects are simply the remaining ones, but most importantamong them are groups and organizations. The higher cultural objects areconstructed on the basis of the primary ones, sometimes together with psychologicaland physical objects. As an example of a higher cultural object, Carnaptakes, not surprisingly, that of the state.The object ‘state’ could perhaps be constructed in the following form: arelational structure of persons is called a ‘state’ if it is characterized in suchand such a way through its manifestations, namely, the psychologicalbehavior of these persons and the dispositions toward such behavior, especiallythe disposition, on the part of some persons, to act upon the volitionsof other.(Carnap [1928] 1969: 231f)Carnap’s claim that all objects in the world are reducible, in principle if not inpractice, to the recollection of similarity between elementary experiences isadvanced as an epistemological thesis, but supposed to be metaphysicallyneutral. It is a thesis about the empirical, but not the metaphysical, reality ofobjects. Applied to cultural objects, it says that they are called real if their manifestationsbelong to real psychological objects, otherwise they are called unreal(Carnap [1928] 1969: 275).

Positivism in philosophy and social science 175In The Logical Structure of the World, Carnap chooses a phenomenalist or‘autopsychological’ basis for his system. Later, under the influence of Neurath,he comes to adopt a physicalist position. According to physicalism, all scientificsentences are translatable into the universal physicalist language, i.e. to sentencesabout physical occurrences in space and time. Applied to the social sciences,physicalism says that all its sentences (insofar as they are about people or groupsof people, including organisations and institutions) can be translated intosentences about the movements of physical bodies in space and time (Carnap,1931: 434, 451f; [1932/33] 1959: 165ff).Among the logical positivists, Otto Neurath has written most extensivelyabout the social sciences, especially sociology. There is nothing to suggest,however, that Neurath should have been a champion of an epistemologicalreduction programme implying methodological individualism (Hempel, 1969a:165–74). Like the methodological individualists, Neurath repudiated, as metaphysical,concepts such as ‘folk spirit’, ‘spirit of a nation’, ‘spirit of the age’ andthe like (1944: 4). But not because they are holistic. His main charge againstthese concepts is that they are idealistic, something which they share with individualistconcepts such as ‘mental’, ‘personality’, ‘soul’, ‘motives’, etc. (Neurath[1931a] 1973: 325, 332, 356f; [1931/32] 1959: 289, 299). Neurath, then, is notan enemy of collectivism and holism. Amongst social scientific theories, hissympathies are clearly with Marxism, and he also suggests that his own ‘aggregationalprogram’ has a metaphysical counterpart in the holism of J.C. Smuts(Neurath, 1944: 20). 7 Neurath’s arch-enemy is idealism in every conceivableform, including the method of Verstehen (1930/31: 121; 1931a: 356–8; [1931/32]1959: 295). The alternative to idealism is physicalism, which is the modern formof materialism and in social science roughly takes the form of behaviourism.According to physicalism, unified science, including of course the social sciences,deals exclusively with physical occurrences in space and time. 8Now, it might be thought that physicalism implies methodological individualism.At times, Neurath also expresses himself in a way indicating that hesubscribes to ontological individualism. ‘Peoples, states, age groups, religiouscommunication, all are complexes built up of single individuals’ (Neurath[1931a] 1973: 386). But this does not imply any commitment to methodologicalindividualism. Neurath’s Marxist leanings make him critical towards the individualismof economic theory and to prefer, instead, an institutionalist theory, usingholist and collectivist notions such as ‘classes’, ‘states’, ‘feudalism’, ‘slavery’,‘tribes’, etc. ([1931a] 1973: 366, 389–403; [1931/32] 1959: 306–15). ‘The sociologistis completely unimpeded in his search for laws. The only stipulation isthat he must always speak, in his predictions, of structures which are given inspace and time’ ([1931/32] 1959: 301). Of some relevance is Neurath’sfrequently expressed opinion that it is often easier to predict the behaviour ofgroups and collectives than to predict the behaviour of single individuals. 9Nor does Neurath advocate any radical empiricist reduction programme. Hisone and only requirement for statements to be scientific is that they be stated inthe universal ‘slang’, or ‘jargon’, of physicalism. He makes the demand that all

176 Positivism in philosophy and social sciencescientific statements, or predictions made on their basis, must be compared with,or ‘assayed’ by, observation statements, but even so, he denies a privileged statusto observation reports expressed by protocol sentences. Neurath denounces as ametaphysical fiction an ideal language constructed out of pure atomic sentences.There are, according to him, no immediate and atomic experiences, which canserve as basic elements of the system. Nor are there any atomic sentences whichare primitive relative to other sentences, which can serve as primitive protocolsentences, and which are exempted from verification. All sentences of theuniversal slang of unified science are on an equal footing.There is no way of taking conclusively established pure protocol sentences as the startingpoint of the sciences. No tabula rasa exists. We are like sailors who must rebuildtheir ship on the open sea, never able to dismantle it in dry-dock and toreconstruct it there out of the best materials. Only the metaphysicalelements can be allowed to vanish without trace. Vague linguistic conglomerationsalways remain in one way or another as components of the ship. Ifvagueness is diminished at one point, it may well be increased at another.(Neurath [1932/3] 1959: 201)This metaphor was later to become famous, and the view of scientific theory itdiscloses has won wide acclaim. The modern, more elaborated, version is the socalled‘holistic view of scientific theories’, suggested by Willard van OrmanQuine (1961: 40–2). Some version of this holistic view of scientific theories istoday accepted by most philosophers of science. It is even assimilated as a part ofthe so-called ‘standard’ (Hempel, 1970), ‘orthodox’ (Feigl, 1970), or ‘received’(Suppe [1973] 1977: 16ff), view of scientific theories, which represents the laststage in the development of logical positivism. According to this view, scientifictheories consist of axiomatic, or hypothetico-deductive, systems of scientificconcepts and laws, more or less general. At the top of the system, there are primitive(undefined) terms and axioms. From these axioms are deduced laws oflesser generality, called theorems. At the bottom of the system, some terms areoperationally defined, by means of some measurement procedure.One of the most clear and succinct statements of the logical positivist orthodoxy,by a social scientist, is Hans Zetterberg’s On Theory and Verification in Sociology(1963). 10 In this work, it is assumed that the primitive or undefined terms of sociologicaltheory are such as denote individual human beings and their actions.Pareto and Weber, as well as most contemporary social theorists, haveassumed that the building blocks of sociological definitions are terms thatdenote human beings and their actions. The rationale for this choice isfound in a suggestive analogy between the position of ‘action’ in all modernsociological theorizing, and the position of ‘primitive terms’ in anytaxonomy. The sociologists say that all social events consist of combinationsof human beings and their actions. The logicians say that all terms of atheory can ultimately be defined by combinations of primitive terms. It

Positivism in philosophy and social science 177therefore, seems useful – at least as a first approximation – to assume thatthe primitive terms of sociology should be words that denote human agentsand their actions.(Zetterberg, 1963: 52f)The analogy, pointed out by Zetterberg, may be suggestive, but it is not at allcompelling, until supplemented by the empiricist requirement that primitiveterms must refer to observables. This requirement is implicit in Zetterberg’s realreason for choosing as primitive, terms denoting human beings and their actions.The real reason is that only human beings and their actions are observable.Zetterberg, thus, complains about Rousseau’s notion of a volonté générale, that ‘wedo not know any constellation of observables that define it’ (Zetterberg, 1963:54).The same holds for any sociological conception which does not represent acombination of observable human beings and their actions. Thus, as socialtheorists, we are well advised to select primitive terms that stand for actorsand types of actions. Since these primitives then are used as building blocks,which in various combinations furnish more complex terms, we are assuredthat even very complex ideas – e.g., property, institution, feudalism, or class– will remain on this side of metaphysics.(Zetterberg, 1963: 54)Zetterberg’s advice concerning choice of primitive terms, provides an excellentillustration of how empiricism, when applied to social phenomena, becomesidentical with epistemological individualism and lends support to methodologicalindividualism.Logical positivism, in a narrow sense, never exerted much direct influenceupon the social sciences. Probably more influential, at least in sociology, was itsAmerican kin, the doctrine of operationalism. According to its founder, theAmerican physicist P.W. Bridgman, to know the meaning of a term we mustknow the conditions of its use, and this implies ‘an analysis of activity of one sortor another, or in other words, an analysis of operations. From this point of view,meanings are operational’ (1938b: 116). Originally confined to physical concepts,Bridgman extends the range of applicability of operational analysis to socialconcepts (1938a: 8). Among the social concepts analysed by Bridgman is that ofthe ‘state’ (not surprisingly, in Bridgman’s example, it declares war). Like themethodological individualists, Bridgman objects to ‘the idea of the “State” assomething intrinsically different from and superior to the people that compose it’(p. 133). Especially abhorrent is the idea of the state as an organism or superperson,endowed with a consciousness of its own. While admitting that the state,like the individual, has to take unique action (declaring war), the action of thestate is nothing but the action of its individuals;

178 Positivism in philosophy and social sciencesome one man gives an order and certain other men individually pull thetriggers of their rifles. The authority of the State derives from the superiorforce which a single man or group of men can exert by procuring unitedaction by a strategically situated group of men, each acting as an individual.The state effectively is the single man or group of men who takes theforceful action.(Bridgman, 1938a: 134)Despite Bridgman’s declaration that ‘Operational analysis really does make adifference’ (1938a: 199), I find it difficult to figure out exactly whereof this differenceconsists. This difficulty does not become less when later on he accepts thepersonification of society and the state as convenient ways of speech, and evenfinds it legitimate and harmless to ‘talk of society demanding this or that, ordoing this or that, quite in the conventional and convenient way’ (p. 211).The main problem with Bridgman’s argument is that he does not provide anyclear example of an operational analysis of a social concept. 11 Especiallyconfusing is Bridgman’s remarks about ‘verbalism’. On the one hand, ‘verbalism’stands for lack of operational meaning. On the other hand, verbal operations arerecognised as important in the operational specification of meaning. But what isthe difference between ‘verbalism’ and verbal operations?There is a strong element of verbalism in our use of the concept of state.We generalize and abstract certain aspects of the joint behaviour of groupsof people, and then cover this complicated combination with a single word.An enormous amount of circumlocution is thereby avoided, so that thesingle word thoroughly justifies itself economically in everyday usage, particularlyif the situation calling for the word arises frequently. But from thepoint of view of analysis and comprehension we have merely messed thingsup by throwing so many points of view into a single word.(Bridgman, 1938a: 136)This passage is perfectly clear if taken in isolation, but in conjunction with muchelse Bridgman has to say on the subject, confusion is bound to arise. Take forinstance the following passage:I think one must have been struck on analyzing these various social conceptsto observe how many of the operations which gave them meaning wereverbal operations, and how much of our concern that a concept should berational was a concern that we should be able to substitute it in familiarways into common verbal forms.(Bridgman, 1938a: 138)What is needed, I believe, is a distinction between an intensional and an extensionalcontext, and a more consequent upholding of the separation betweencommon-sense verbalism and scientifically relevant verbal operations.

Positivist social sciencePositivism in philosophy and social science 179I understand ‘positivist social science’, broadly, as social science based on a beliefin the fundamental similarity between the social and the natural sciences, asthese are depicted by positivist and empiricist philosophers. My use of this termis motivated solely by convenience – more specifically, my need for a rubric –and does not at all reflect any wish, on my part, to label social scientists ‘positivists’,or to say anything essential about positivist social science. My onlyinterest is to find manifestations of methodological individualism within thisimportant research tradition in social science.For the purposes of this section, I divide positivist social science in two parts:one empirical and one theoretical. The empirical part is ‘empiricist’, in the sensethat it is based on the belief that all knowledge about society derives from observation.The theoretical part is based on the belief that social phenomena canonly be explained by covering laws of varying generality, organised as hypothetico-deductivesystems, in according with the ‘orthodox’ view of scientifictheories.Systematic empiricismIt was argued above (p. 167) that methodological individualism follows naturallyfrom a radical empiricist epistemology. This might lead us to expect that empiricistsocial science is largely individualistic. As we shall see, this is not really thecase. At least not in a way that satisfies all methodological individualists.The philosophical background of empiricist social science can be found inpragmatism, logical positivism and operationalism, but above all, in Britishempiricism. Most influential have been those who are seen as representing thescientific outlook, such as Francis Bacon and John Stuart Mill, and later, thosewho developed the statistical version of the scientific outlook, such as FrancisGalton, Karl Pearson and R.A. Fisher. The important role played by statistics insocial scientific empiricism, has motivated the name ‘systematic empiricism’. 12It is true that empiricist social science is characterised, above all, by the use ofstatistical techniques, but this is historical accident more than logical necessity, aswitnessed by the ethnomethodological critique of systematic empiricism for notbeing empiricist enough. According to some critics, the main cause of the dominantposition of this particular form of empirical investigation is the need on thepart of the administration of nation-states to know and control the populationliving, as subjects of state, on its territory. 13 But while empiricism in socialscience does not necessarily resort to the use of statistical methods of investigation,the use of statistical techniques does not necessarily imply a commitment toempiricist epistemology. Statistical techniques are used by empiricists and nonempiricistsalike, but assigned a different role in the research process by the moretheoretically inclined social scientist than by the empiricist. While, according to aradical systematic empiricism, the resources of social science are exhausted bystatistical analysis, theoretical social science tends to assign to statistics a moremodest role in the research process. When, in this section, reference is made to

180 Positivism in philosophy and social science‘empiricist social science’, this term should be understood in the sense of‘systematic empiricism’.Among the resources and techniques used by systematic empiricism are theexperiment, the utilisation of official statistics in the form of census data and,most important, the survey. The experiment is, of course, the ideal weapon ofcausal and, therefore, scientific analysis. Due to the limited possibility of controllingand manipulating social phenomena, however, and the ethical problemsinvolved in the control and manipulation of human beings, the experiment hasturned out to be of limited value to the social sciences. Experiments are onlypossible with small groups of people and have, therefore, been used primarily inpsychology and social psychology. Much ingenuity has been spent on the attemptto develop research designs that simulate the experiment, or otherwise capture itslogic. Nevertheless, real experimental analysis, because of its limited applicabilityto large-scale social phenomena, tends to be individualistic (see Coleman, 1969a:99f). The subsequent discussion concerns the census and the survey but not theexperiment.Among the social sciences, systematic empiricism first gained ground in sociology,where it still has its strongest hold, but later spread to political science inthe form of behaviouralism. In economics, systematic empiricism is known as‘econometrics’ and has gained little or nothing from sociology, but was ratherahead of sociology in assimilating regression analysis as an important tool ofresearch (Coleman, 1969a: 91; Halfpenny, 1982: 40). Anthropology is interestedin cultures where census data are rare and the survey not the most suitablemethod for gaining information. Empirical anthropology relies, instead, on theless systematic method of participant observation. In what follows, there will be aconcentration upon systematic empiricism in sociology, but it is assumed that theconclusions have a more general application.There are two radically opposed views concerning the relation of systematicempiricism to methodological individualism. According to one view, systematicempiricism, in social science, is a largely holistic enterprise, because its units ofanalysis, are typically large-scale social phenomena at the macro-level. Accordingto another view, systematic empiricism tends to be individualistic, because itsunit of observation is typically individual human beings; their attitudes, opinionsand beliefs, etc.The first view dominated in the early history of social science, when methodologicalindividualists, such as John Stuart Mill, Carl Menger and Gabriel Tarde,criticised social theorists, like the statistician L.A.J. Quetelet and the historianHenry Buckle, for being too much macro-oriented and for mistaking statisticalregularities and correlations for social laws. 14 Statistical regularities and correlationsare empirical generalisations, not laws. As such, they describe, but do notexplain. 15 In order to explain social phenomena, you need causal laws and theonly causal laws in social life are psychological; or, if this is too narrow, at least,they are laws about individuals. According to most methodological individualists,there are no social laws, at least, no causal social laws. Since only individuals exist,only individuals can be causes of social phenomena.

Positivism in philosophy and social science 181The symbol and personification of macro-oriented systematic empiricism isEmile Durkheim (Little, 1991: 189f; Rosenberg 1988: 118ff), as he appears in hisclassic study of Suicide (1897). Impressed by its stability over time, Durkheimargued that the suicide rate ‘is not simply a sum of independent units, a collectivetotal, but is itself a fact sui generis, with its own unity, individuality andconsequently its own nature – a nature, furthermore, dominantly social’ ([1897]1951: 46). As we have seen in chapter 2 (pp. 29f ), Durkheim was not the first tobe impressed by the stability of many statistical measures, but more than anyother he used this fact as proof that there are social facts, irreducible to factsabout individuals. There seems to be a certain ambiguity in Durkheim’s view ofthe suicide rate, however. If he were really suggesting that the suicide rate itselfis a social fact sui generis, then, his view would be open to serious objection. Buthe is not (see below). The suicide rate, as a statistical measure is, of course, a factabout a population of individuals. As Friedrich von Hayek has pointed out,statistical collectives are not social wholes. Statistics is concerned with the propertiesof the elements of ‘collectives’ (Hayek, 1955: 61–3). In a similar veinLudwig von Mises ([1957] 1985: 260) maintains that social, or mass phenomena,as measured by statistics, ‘are not things standing outside and above individualphenomena. They are not the cause of individual phenomena. They areproduced either by the cooperation of individuals or by parallel action’. 16 It isnecessary, therefore, to make a clear distinction between a macro-approach thatis based upon the set-element or class-member relation, and one that is basedupon the whole-part relation. Statistics deals with the former relation. A typicalfeature of this relation is that facts about the set (collective) are arrived at byaggregation of facts about the elements. No statistical fact about a collective,therefore, is a social fact sui generis. Statistical facts are facts about numbers, largeor small, of elements. This leads immediately to the following conclusion: nomacro-theory which deals with relations between aggregated data, whether ofthe first, second or any order is holistic relative to its elements. A corollary of thisconclusion is that systematic empiricism is not holistic only because it deals withthe relations between aggregated data. 17No statistical information about the elements can explain to us the propertiesof the connected wholes. Statistics could produce knowledge of theproperties of the wholes only if it had information about statistical collectivesthe elements of which were wholes.(Hayek [1942–4] 1955: 62)This leads us back to the ambiguity in Durkheim’s view of the suicide rate. Oncloser inspection, it turns out to be, not so much the suicide rate per se, as its explanation,which is, and must be, social (cf. Hacking, 1990: 177). Durkheim’s pointwas that psychology is not enough to explain the relative stability of the suiciderate over time, its variability between cultures, societies and different strata of thepopulation, and its sensitivity to social disruptions. ‘The conclusion from all thesefacts is that the social suicide-rate can be explained only sociologically’

182 Positivism in philosophy and social science(Durkheim [1897] 1951: 299). It may also be pointed out that Durkheim wasperfectly aware of the phenomenon of spurious correlation. This is evidenced byhis statistical analysis of the so-called ‘egoistic suicide’ ([1897] 1951: chs 2–3)and by his explicit statement:Concomitance can occur, not because one of the phenomena is the cause ofthe other, but because they are both effects of the same cause, or indeedbecause there exists between them a third phenomenon, interposed butunnoticed, which is the effect of the first phenomenon and the cause of thesecond.(Durkheim [1895] 1982: 152)Durkheim’s conviction that the suicide rate can only be explained sociologicallywas not supposed to be limited to this particular phenomenon. Rather, thestudy of suicide was an attempt to demonstrate the validity of a general methodologicalprinciple stated by Durkheim two years earlier ([1895] 1982: 110): ‘Thedetermining cause of a social fact must be sought among the antecedent socialfacts and not among the states of the individual consciousness’. The differencebetween the methodological individualists and Emile Durkheim, then, is not thatthe latter saw, statistical regularities, per se, as social facts, sui generis. The differenceis that the individualists believe that an explanation of statistical regularitiesand correlations, must be psychological or, at least individualistic, whileDurkheim believed that they must be sociological, that is in terms of social factssuch as social institutions and social structure.According to the second view, mentioned above (p. 179), the problem withsystematic empiricism is not that it is too holistic, but that it is too individualistic.It is a common view, held both by its critics and by its defenders, that systematicempiricism has a strong individualist, atomist and subjectivist bias. 18 The individualistbias is due to the fact that much official statistics and most surveyresearch take the individual as its element and basic unit of observation. Theatomist bias is due to the further fact that official statistics and survey researchusually study the individual in isolation, without taking into account his interactionwith other individuals. The subjectivist bias, finally, concerns only surveyresearch and is the result of its traditional preoccupation with the opinions, attitudesand beliefs of individuals.The charge of individualism did not appear until the advent of the surveyand this is probably no coincidence. The survey tends to be more individualistic,or, at least more subjectivistic, than official statistics. One reason for this is thedevelopment of various instruments of measurement, which have been animportant ingredient in systematic empiricism, especially for its claim to being aScience. There is little doubt that the tendency to accept only operational definitions,as scientifically legitimate, has contributed to pushing systematicempiricism further in the direction of subjectivist individualism. 19 The reason is,probably, that the easiest ‘operation’ available to the social scientist is to ask

Positivism in philosophy and social science 183human individuals who they are, what they own and what they think aboutvarious matters.According to George A. Lundberg (their most influential spokesman in sociology)operational definitions consist of ‘words clearly designating performable andobservable operations subject to corroboration’ ([1929] 1942: 89). Lundberg, himself, wasnot a methodological individualist, however. He was an instrumentalist and, assuch, rather liberal about the use of collective concepts in sociology; if only as aconvenient construct, and as long as they prove useful ([1929] 1942: 80–112).From this point of view, he criticised the psychologism of Floyd H. Allport(Lundberg, 1939: 163–73). The most ambitious attempt to provide operationaldefinitions of the concepts of sociology is by Stuart C. Dodd, who takes as hispoint of departure the so-called ‘S-theory’, according to which, ‘Every quantitativelyrecorded societal situation can be expressed as a combination of indices oftime (T), of space (L), of population (P), and of the many characteristics (I) of peopleor their environment’. Or expressed in a simple formula: S=(T:I:L:P) (Dodd,1942: 58f; see also 1939: 628). You need but little imagination to see that thispoint of departure makes for an individualist sociology.While it seems undeniable that systematic empiricism has an individualistbias, it is difficult to estimate the exact extent to which it is individualistic. Severalfactors complicate the attempt to classify the results of systematic empiricism intothose that are individualistic and those that are holistic. First of all, there is thefact that even though the individual is usually the basic unit of investigation, thisis not always the case. In econometrics, for instance – paradoxically, sinceeconomics is usually considered the most individualistic of the social sciences –the basic units of research are usually the household and the firm. A secondcomplication is that even in those cases where the individual is the basic unit ofresearch, it might be that he/she can only be described in terms that makeimplicit reference to the institutional setting and the social structure in whichhe/she acts. Thus, it has been argued that any study of voting behaviour makesimplicit reference to the political system (Udehn, 1987: 220ff). It could also beargued that the classification of individuals according to social backgroundimplies some notion of ‘social structure’ (Lukes, 1973: 121). It is nowadays generallyagreed that all data are theory laden or theory impregnated, and the data ofsystematic empiricism is no exception to this rule. It could be, therefore, that thedata about individuals generated by systematic empiricism are impregnated with‘social theory’ that is holistic. On the other hand, systematic empiricism, to theextent that it relies on any articulated theory at all, tends to assume a subjectivisttheory of social structure, so, that even when it uses concepts with a holist ring,such as ‘social role’ and ‘social status’, these concepts are defined in terms of theexpectations and rankings of individuals, that is, in individualist terms.Confronted with the charge of individualist bias, leading representatives ofsystematic empiricism sought a remedy in the development of new techniquesfor doing justice to the structural and collective features of social life. 20 Oneexpression of this ‘structuralist’ and ‘collectivist’ tendency in systematic empiricismwas the introduction of so-called multi-level research, that is, research

184 Positivism in philosophy and social sciencewhich aims at an integration of the level of individuals with that of groups andother collectives (van den Eden and Hüttner, 1982). Among the most importantspokesmen of multi-level research – and of systematic empiricism in general – isPaul F. Lazarsfeld, who, in collaboration with Herbert Menzel, developed a wellknownclassificatory scheme for properties of individuals and of collectives(Lazarsfeld, 1968: 617–24; Lazarsfeld and Menzel, 1969). Lazarsfeld andMenzel distinguished between absolute, relational, comparative and contextualproperties of individuals, and analytical, structural and global properties ofcollectives. This classification, Lazarsfeld and Menzel point out, is purely formal(1969: 510f). As such, it is not altogether adequate for addressing the substantiveissues involved in the combat between individualists and holists.Of special interest, for our purpose, is the classification of properties ofcollectives into analytical, structural and absolute. Analytical properties areobtained by aggregation of properties of individuals and are, therefore, individualistic.‘Structural properties’ of collectives would seem, by their very name, bestsuited for catching the idea of ‘social structure’, but appearances are deceptive.These properties are exemplified, most typically, by the concentration of choicesrevealed by sociometric analysis. The study of these popularity and friendshipstructures belong, most naturally to the province of social psychology, and havelittle or nothing to do with the kind of structural analysis advocated by methodologicalholists. The global properties of collectives alone, are not based uponinformation about individuals. Examples of global properties of collectives arethe presence of certain social institutions, such as a school, an army, or a certaintype of government. The recognition of the existence of global properties ofcollectives opens the way to truly structural, hence holist, analysis of society. Itremains to be seen, whether multi-level research takes that direction. Hitherto, ithas suffered from the lack of an adequate conception of ‘social structure’, andso, remained largely individualistic (van den Eden and Hüttner, 1982: 60ff).The most popular view among those doing multi-level research, seems to bethat structural analysis consists of so-called ‘contextual propositions’, explainingthe behaviour of individuals in terms of properties of the collective or group. 21The main problem with this view is that it does not discriminate between thedifferent kinds of collective properties – analytical, structural and global –mentioned by Lazarsfeld and Menzel, and, in fact, tends to be directed mainly atthe analytical and, to some extent, the ‘structural’ properties of collectives (vanden Eden and Hüttner, 1982: 60ff). What Peter Blau at first called ‘structuraleffects’, for instance, is no more than the effects of the commonly shared valuesof individuals and of the interaction between individuals (Blau, 1960: 178–93).It should be pointed out, however, that Blau soon developed a more holisticconcept of ‘social structure’ (1974; 1977), and that Lazarsfeld came to realisethat what many sociologists have in mind when speaking of social structure is theglobal, as distinguished from the aggregative, properties of collectives(Lazarsfeld, 1970b: 312).Among the problems facing multi-level research, are the so-called ‘fallacies ofthe wrong level’. 22 These fallacies are the various mistakes you could commit

Positivism in philosophy and social science 185when making inferences from results obtained at one level of research to that ofanother level. Most well known is the ‘ecological fallacy’, first notified by W.S.Robinson (1950: 351–6). Put simply, the ecological fallacy consists in drawingconclusions from results obtained at the collective, or group level about the propertiesor behaviour of the individuals making up that collective or group. Thus, itdoes not follow from the fact that there is a high correlation between thepercentage of the population that is black and the percentage of the populationthat is illiterate that blacks are more often illiterate than are other ethnic groupsmaking up the entire population. It might be thought that ecological correlations,because of their seeming ‘irreducibility’, point to the existence of socialfacts sui generis, but this is not necessarily the case. Far from being irreduciblesocial facts, ecological correlations are sometimes spurious, and this is why inferencefrom the collective to the individual level may be invalid. But, whetherspurious or not, ecological correlations are between analytical properties ofcollectives and, as such, are based upon data about individuals. 23The argument of this section has been that systematic empiricism, even whenit engages in multi-level research, has a strong individualist bias, due to the factthat it usually takes the individual as its element and basic unit, and, for theremainder is based upon aggregation of facts about individuals. This argumenthas force only with respect to a radical systematic empiricism, which stops at thelevel of fact-finding and empirical generalisation. As soon as systematic empiricismleaves this extreme position, however – as the large majority of empiricalsocial scientists do – and asks for a theory with which to explain these facts, theproblem shifts into that of determining whether this theory is individualistic ornot. It seems a fair contention that even when systematic empiricism has soughtsome theory to explain its statistical findings, it has looked primarily in the directionof social psychology, that is, in an individualist direction (van den Eden andHüttner, 1982: 39–53; Bryant, 1985: 168–73).But how is it that opinions about the nature of systematic empiricism are sodivided? How is it possible that some believe it is a paradigm of holism andothers that it is much too individualistic? The reason, I suggest, is that there aretwo forms of individualism (introduced in the presentation of Schumpeter above(pp. 106f)): substantive and procedural methodological individualism. Systematicempiricism, to the extent that it is at all individualistic, is a form of proceduralmethodological individualism: it often takes the individual, as its point of departure,or unit of observation. Its unit of analysis, however, is usually somemacrophenomenon. It differs from the procedural methodological individualism,discussed by Schumpeter, however. While the latter is a theoretical proceduralmethodological individualism, that of systematic empiricism is, not surprisingly,an empirical procedural methodological individualism.As we have seen, methodological individualists have always complained thatstatistical, and other forms of, empirical generalisations fail to explain. Thiscritique has been most common in economics, where it takes the form of arequirement that macroeconomics must be provided with microfoundations (see

186 Positivism in philosophy and social sciencepp. 234ff ), but it has been a recurrent phenomenon in political science and sociology,as well.But methodological individualists have not been alone in this. Social scientistsof a more holistic persuasion have been no less emphatic about the need fortheory (see, e.g. Adorno, 1957 and Willer, 1967). What distinguished the latterfrom the former, then, was not the need for theory, but the kind of theory theybelieved was needed. For a long time, it was taken for granted that a theory is abody of concepts and law-like statements that are used to explain empiricalregularities. As time passed, however, it became painfully obvious that the socialsciences do not produce an abundance of generally accepted laws, as do some ofthe natural sciences. This is probably one of the reasons social scientists have,recently, suggested a reorientation of social science, away from the focus on laws,to a search for social mechanisms. 24A pioneer in this reorientation is the philosopher Mario Bunge, who hasadvocated a turn to mechanisms, for deaf ears, since the 1960s (Bunge, 1997).According to him mechanisms are needed to open the black boxes of positivistscience and to provide genuine causal explanations of phenomena, described ina ‘kinematical’ manner, for example statistically. It may be pointed out thatBunge adopts neither individualism, nor holism, but accepts both microreductive(bottom-up) and macro-reductive (top-down) explanations. He calls hisapproach systemism.Figure 6.1 Mario Bunge’s methodological systemismSource: Bunge (1997: 441)In the more recent development of social science, there are, at least, twodifferent approaches to social mechanisms. The first approach is the so-calledcritical, or new, realism in social science, stemming from the work of Rom Harréand Roy Bhaskar, but the latter, in particular (Bhaskar, 1979: ch. 2). 25 Bhaskarrejects both the macro-reductionist collectivism of Durkheim and the microreductionistindividualism of Weber. He also rejects the dialectical view ofBerger and Luckmann (see pp. 161f). His own suggestion is that both individualsand society exist, but in a symbiotic, rather than a dialectical, relation. ‘Society isboth the ever-present condition (material cause) and the continually reproducedoutcome of human agency’ (p. 43).

Positivism in philosophy and social science 187Figure 6.2 Roy Bhaskar’s transformational model of the society/person connectionSource: Bhaskar (1979: 46)As is often pointed out, Bhaskar’s model has much in common with AnthonyGidden’s theory of structuration (see pp. 164f ). I agree that there is a similarity,in that both maintain that society is constantly created, but there is also animportant difference. While for Giddens the creative aspects is in focus, Bhaskarconceives of structuration as reproduction and transformation of existing structures,with causal powers. In the end, therefore, Gidden’s theory of structurationis a theory of action, where structure appears only as an epiphenomenon,whereas Bhaskar’s theory is a form of structuralism.Other philosophers and social scientists belonging to this research tradition,include Russel Keat, John Urry, Ted Benton, Andrew Sayer, Jeffrey C. Isaac,Margaret S. Archer and Derek Layder. In this theoretical tradition, which hasimportant roots in Marxism, the causal, or generative, mechanisms used toexplain social phenomena and empirical generalisations are, typically, socialstructures. A comprehensive treatment of the individualism-holism issue can befound in Archer (1995), who also offers an interesting alternative to bothGiddens and Bhaskar (without mechanisms). Her main point, with which Iagree, is that, for the purposes of social science, we do need a dualistic approachto the relation between individual action and social structure. Among criticalrealists, Sayer (1984), and Pawson (1989: ch. 6), in particular, have tackled theproblems with systematic empiricism.The second branch of the new turn to social mechanisms is decidedly individualistic,and engaged in an attempt to provide social science withmicrofoundations. The ‘social’ mechanisms identified are those of acting individuals.The approach of this branch is similar to that of Floyd Allport (see pp.66f ), except that the latter was a behaviourist and called his mechanisms‘psychological’. The recent call for mechanisms in social science is based onrational choice and other forms of intentional action. Important sources ofinspiration for this individualistic approach are Raymond Boudon ([1977] 1982;1979a; 1979b) Thomas Schelling (1978), Jon Elster (1983a: Part I; 1985: 5ff;1989a: ch. 1) and James Coleman (1986c; 1990c: ch. 1). I will return to Boudon,Coleman and Elster in chapter 10. Here I will draw attention to the contribution

188 Positivism in philosophy and social scienceof Hedström and Swedberg, which is, to a large degree, aimed at systematicempiricism.In a first article, Hedström and Swedberg (1996b) argue that there is anunfortunate gap between theory and empirical research in sociology, and alsothat this gap can be filled by rational choice. They make the common observationthat statistical correlations and other empirical generalisations aredescriptive, not causal, but require a causal explanation. Such explanation, theysuggest, is best understood as being in terms of the mechanisms that haveproduced the correlation (p. 136). Unfortunately traditional sociological theoryhas not been very successful in providing such causal mechanisms. The hopesof Hedström and Swedberg go, instead, to rational choice, which has theadvantage, over much traditional sociology, in being (1) analytical, (2) based onmethodological individualism and (3) providing intentional explanations ofobserved phenomena (pp. 129ff). The upshot is that rational choice producesdeeper and more fine-grained explanations of empirical generalisations (p.141).Hedström and Swedberg introduce the notion of social mechanism in theirfirst article, but make it the main topic of a second article (1996b), which alsoincludes a new critical discussion of statistical variable sociology (pp. 291–3).The most important features of Hedström’s and Swedberg’s social mechanismsseem to be the following (p. 287): (1) the idea of generative mechanisms differsfrom Hempel’s famous covering-law model of causal explanations; (2) generativemechanisms are invariably at the micro-level of macro-phenomena. My interestis in the second feature, which is a manifestation of methodological individualism.The third article by Hedström and Swedberg is their ‘Introduction’ to SocialMechanisms (1998), which is something of a manifesto for a turn to individualisticmechanisms in social science. This introduction is largely a restatement of thearguments in their previous two articles, which I will not repeat here.Following Coleman, Hedström and Swedberg use the following illustration ofdifferent types of mechanism:Figure 6.3 Hedström and Swedberg’s typology of social mechanismsSource: Hedström and Swedberg (1996b: 297)

Positivism in philosophy and social science 189The first type of mechanism, mentioned by Hedström and Swedberg, is calledsituational mechanism. It is exemplified by the ways in which the social situationshapes the desires, or interests, beliefs and opportunities of individuals. Thesecond type of mechanism is called individual action mechanism and refers to the wayindividuals’ beliefs and desires generate actions. The third type of mechanism iscalled transformational mechanism and it depicts how interaction between individualsgives rise to collective outcomes. Of these mechanisms, the second and third aremost individualistic, while the first, seems to imply a holistic element.I feel that there is a certain ambiguity surrounding the situational mechanism.While the arrow runs from the macro-level to the micro-level, Hedström andSwedberg maintain that there are no macro-level mechanisms (1996b: p. 299).This is odd, since it implies that the mechanism lies in the effect rather than inthe cause. It would seem that Hedström and Swedberg follow Watkins andElster, rather than (the later) Popper and Coleman, since they maintain that ‘theelementary “causal agents” are always individual actors’ (p. 11), which impliesthat social institutions can in principle, if not always in practice, be explained interms of the actions of individuals (p. 12).It is worth noticing that one of the most renowned empirical sociologists,John Goldthorpe (1996; 1998), has recently turned to rational choice andmethodological individualism, for reasons similar to those of Hedström andSwedberg, and inspired by the same people, especially Raymond Boudon. Thismeans, among other things, that he sees the approach of generative mechanismsas the most promising way of providing a causal explanation of the results ofstatistical investigations (Goldthorpe, 2000: 149ff ). Goldthorpe is not veryexplicit about methodological individualism, but he denies being an ontologicalindividualist (1998: 167). The main thing is rational action theory, which is a lessrestricted version of rational choice, inspired mainly by Max Weber’s verstehendesociology and Raymond Boudon’s cognitivist model (see pp. 306ff ). The importantthing in rational action theory is to understand the meaning individualsattach to their actions; the motives and reasons they have for acting in certainways. These motives may be utilitarian, but they could also be of a more normativekind. An example of this can be found in his and Richard Breen’sexplanation of educational differentials in terms of class differences in resources.Cultural explanations in terms of social norms are considered less importantthan instrumental rationality, but not devoid of ‘some explanatory significance’(Breen and Goldthorpe, 1997: 297–300).Positivist social theoryThe first and main example of an individualist theory in social science iseconomics. It is not at all surprising, therefore, that social scientists, who aremethodological individualists, but not economists, should use economics as amodel for their own theoretical endeavours. As we have seen in chapter 4, thiswas the case with Max Weber. Now, economics is a theory of exchange. At least,it is seen as a theory of exchange, or catallactics, by those economists who have

190 Positivism in philosophy and social sciencebeen most eager to declare themselves methodological individualists. It is onlynatural, therefore, that social scientists, opting for an individualist theory ofsociety, should develop a theory of social exchange. The main example, I believe,is the sociologist George Caspar Homans. 26Homans’s theory of exchange 27Homans seems to be fond of writing his own intellectual biography. As he is oneof the most influential sociologists in the second half of the nineteenth century,sociologists have every reason to be grateful for this idiosyncrasy. From the fourautobiographical writings Homans has produced, we may extract the followingfacts about his professional and intellectual career. 28Homans came to sociology by chance. He was advised by his tutor in Englishliterature, his original subject, to attend a course on the Italian economist andsociologist Vilfredo Pareto held by the biochemist and historian of science, L.J.Henderson at Harvard. Thus becoming expert on Pareto, he wrote An Introductionto Pareto (1934) together with the lawyer J.P. Curtis. Circumstances conspired tomake him a sociologist and this is what he eventually became, if not intellectually,then at least professionally.At Harvard, Homans became acquainted with the anthropologist Elton Mayoand through him, with the writings of Malinowski and Radcliffe-Brown. Beingconvinced that it is the nature of individuals that determines the nature ofsociety, rather than the other way around, Homans was suspicious of functionalismfrom the very beginning. Homans recognizes three kinds of statement inwhich the word ‘function’ appears:The first kind of statement said that one institution in a society was a functionof another in the sense that the two were interrelated … The secondkind of functional statement said that an institution was functional formembers of a society in the sense of meeting their needs … The third kindof functional statement said that an institution was functional for a society inthe sense of helping to maintain society as a going concern.(Homans, 1962: 23f)The first statement was made by both Malinowski and Radcliffe-Brown, thesecond, especially by Malinowski, and the third only by Radcliffe-Brown. That iswhy Homans came to accept Malinowski’s functionalism, which he calls ‘individualistic’,but reject Radcliffe-Brown’s which is ‘societal’ (Homans 1983: 12). 29When Homans in his next work, English Villagers of the Thirteenth Century (1941),set out to seek his own roots, he did so with the conscious aim, and despite hisdoubts about it, of applying functionalist analysis to this historical material.While working on the English Villagers, however, Homans came into contact withan intellectual current very different from functionalism, P.W. Bridgman’s operationalism.The influence was indirect and came from the anthropologists EliotChapple and Conrad Arensberg, who wanted to define the chief concept of

Positivism in philosophy and social science 191anthropology in terms of the order, frequency and duration of interactionsbetween men. Homans felt that interaction was not enough to grasp the contentof social life and added sentiment and activity. This conceptual scheme; ‘interaction’,‘sentiment’ and ‘activity’ was set out for the first time in the last chapter ofEnglish Villagers and became the analytical tool of his next main work, The HumanGroup (1951: 33–44).Behind Homans’s conceptual scheme lay, besides the impact of operationalism,a conviction he had gained through his critical reading ofanthropological literature, that ‘human nature is the same the world over’. Theultimate explanatory principles in history, anthropology and sociology, therefore,are psychological propositions, propositions about men as men (1951: 443). Inhis last work, the autobiographical Coming to My Senses (1984: 329, 351), Homanssays that he reached this conclusion before the Second World War, and wasturned into a methodological individualism already at that early stage of hiscareer. He also invokes the support of John Stuart Mill, who had adopted thesame approach in his A System of Logic (see pp. 43–9).It is possible that Homans had already turned into a methodological individualistin the 1930s, but there is much even in The Human Group (1951), which ishard to reconcile with this position. Above all, there are many remnants of theorganicist influence of the philosopher Alfred North Whitehead (pp. 6–10, 86–8,272–6), which do not fit methodological individualism. Even so, The HumanGroup marks a definite change of direction in Homans’s work, from an earlierpreoccupation with social institutions to his later concentration on elementary orsub-institutional behaviour in small groups. Connected with this change wasHomans’s introduction of the distinction between the internal and the externalsystems (pp. 90, 109f). The latter would eventually be turned into the given, orboundary, conditions of Homans’s theory of exchange (Homans, 1962: 272f;1961: 230f).The Human Group was published in the same year (1951) as Talcott Parsons’s,even more influential, The Social System. Being a work of a totally different inspiration,it is only natural that Homans should have felt challenged. Mostprovocative to Homans was Parsons’s claim to have produced a sociologicaltheory. Homans read the philosophers of science, especially R.B. Braithwaite,later also Ernest Nagel and Carl G. Hempel and learned that a theory consists ofpropositions stating relationships between properties of nature. The job of ascientific theory is to explain, and in order to be able to do this job, scientificpropositions must be general. Scientific explanation is the deduction, or derivation,of lower-order generalisations from higher-order generalisations underspecified given conditions. Now, the trouble with Parsons’s ‘theory’ is that it doesnot consist of propositions, but of concepts. It is a conceptual scheme, not ascientific theory and the problem with most sociological propositions, includingHomans’s own propositions in The Human Group is that they are not generalenough. Homans is forced, therefore, to go outside the field of sociology in orderto find the most general propositions. He finds them in economics and, especially,in the behaviourist psychology of B.F. Skinner. 30

192 Positivism in philosophy and social scienceThis is a bit odd, since the behaviourist view of man is diametrically opposedto that of economics. While the latter depicts man as an autonomous beingmaking choices, behaviourism regards her/him as an automaton reactingpassively to his environment. The difference could not be greater it would seem,and yet, when it comes to application, they turn out to be very much alike. Theyare both based on motivational assumptions of a hedonist or utilitarian origin,even if expressed in somewhat different languages. The terminology of ‘reward’and ‘punishment’, however, occur in both. But while behaviourism and rationalchoice agree that men seek to maximise their rewards, rational choice assumesthat they do so by free choice, while behaviourism insists that they do so automaticallyas a result of their previous learning histories. Since social scientists, incontradistinction to some psychologists, are ignorant of the biographies of thepeople they investigate, reference to their learning histories becomes a matter ofmere faith in behaviourist doctrine, but of little consequence for actual research.Even behaviourist social science tends to treat men as if they acted purposively(see Homans 1961: 12–14; 1970a: 317–24).Behaviourism originated as a psychological theory, or perhaps we should saymethodology. It was first advanced by John B. Watson, very much as a reactionto the introspective psychology prevalent in the beginning of the twentiethcentury. Watson wanted to get rid of all ‘psychical’ concepts such as ‘soul’,‘mind’, ‘consciousness’, etc., referring to an inner ‘man’, and replace them withconcepts referring to the objective behaviour of individuals, notably those of‘stimulus’ and ‘response’. Behaviourism, according to Watson, has the twinpurpose of prediction and control of human behaviour. To predict thebehaviour of an individual we must know the way he/she has been conditionedfrom earliest childhood. To control his behaviour we must be able to control hisenvironment (Watson, 1913: 158–77; 1924/5: ch. 1).Behaviourism soon found out that there is a gap between stimulus andresponse and tried to fill this gap with an intervening variable. This interveningvariable, however, was not a very satisfactory solution since it seemed to reintroducewhat behaviourism wanted to avoid, that is, reference to some mysteriousforce within the human organism. B.F. Skinner, the leading representative ofcontemporary behaviourism, solves the problem, not by invoking some interventionbetween stimulus and response, but by looking beyond the response. Skinnerrejects the stimulus-response model and replaces it by a model centering oncontingencies of reinforcement: the when and where and how of reinforcement(1965; 1969). If behaviourism is used to explain social, and/or cultural,phenomena, it turns into a form of methodological individualism and it seemsthat Skinner approves of such use. At least, he is an ontological individualist.A species has no existence except as a collection of individuals, nor has afamily, tribe, race, nation, or class. A culture has no existence apart from thebehavior of the individuals who maintain its practices. It is always an individualwho behaves, who acts upon the environment and is changed by the

Positivism in philosophy and social science 193consequences of his action, and who maintains the social contingenceswhich are a culture.(Skinner, 1974: 204). 31Behaviourist psychology is Homans’s solution to the problem with Parsons’so-called ‘theory’. Instead of concepts, it generates general laws. The result isHomans’s theory of exchange, first presented in ‘Social Behavior as Exchange’(1962: ch. 17) and, more fully, in Social Behavior: Its Elementary Forms (1961: 51–82).It consists of five propositions and reads like this:1 If in the past the occurrence of a particular stimulus-situation has been theoccasion on wich a man’s activity has been rewarded, then the more similarthe present stimulus-situation is to the past one, the more likely he is to emitthe activity, or some similar activity, now.2 The more often within a given period of time a man’s activity rewards theactivity of another, the more often the other will emit the activity.3 The more valuable to a man a unit of the activity another gives him, themore often he will emit activity rewarded by the activity of the other.4 The more often a man has in the recent past received a rewarding activityfrom another, the less valuable any further unit of that activity becomes tohim.5 The more to man’s disadvantage the rule of distributive justice fails of realization,the more likely he is to display the emotional behavior we call anger.This theory is, with minor modifications, 32 Homans’s main contribution tosociological theory. In his later writings, Homans is satisfied to demonstrate thefruitfulness of his theory and to defend it against alternative approaches, both inits specific form as a behaviourist theory of exchange and, more generally, as anexample of psychologistic reduction and of methodological individualism.Homans for the first time proclaims himself a psychological reductionist in apaper from 1953 (Homans, 1962: ch. 16), but later denies being a psychologicalreductionist, on the ground that there are no general sociological propositions toreduce (1964a: 817; 1967b: 83–6; 1969: 15f). His explicit commitment tomethodological individualism is of later date and seems to replace hiscommitment to psychological reductionism. According to Homans, however,methodological individualism entails psychological reductionism. Homans’s individualism,like Schumpeter’s (see pp. 105f), is pragmatic and strictly methodological.It lacks every connection with ontological individualism and so, does not deny theexistence of institutions and other social wholes. Being pragmatic in character,Homans’s methodological individualism is also free from dogmatism. The issueof individualism versus holism is scientific and cannot be settled by philosophicalargument (see Homans, 1967b: 83–6; 1980: 19–21).This does not prevent Homans from expressing definite opinions about socialstructure, however; opinions that are typical for methodological individualists. If,according to Homans, there is such a thing as social structure at all, it is only as

194 Positivism in philosophy and social sciencecertain patterns of interaction among people (1975: 54). Concerning PeterBlau’s notion of ‘structural effect’, Homans suggests that it is better called‘collective effect’, since ‘it concerns the influence of a collection of individuals –but if a collection, still individuals – on another individual’ (p. 56). But ‘Whyshould things recognizable as social structures exist at all?’. Homans’s answer tothis fundamental question by Raymond Boudon is as follows:I have tried to show how relatively enduring structures, a status-system forinstance, can be created and maintained by the actions of individuals,actions of course taken under the influence and constraint of the actions ofother individuals. That is, I have tried to explain the properties of certainsimple structures using as general propositions the propositions of behavioralpsychology.(Homans, 1975: 64)Homans’s work was widely read and much debated, but he seems to haveattracted few immediate followers, at least in the USA. 33 Among later applicationsof behaviourist psychology to sociological problems in the USA, there isreason to mention John H. Kunkel’s attempt to explain social change, especiallyeconomic growth, with the help of behaviourist learning principles (Kunkel,1970). The most significant contributions to a theory of social exchange, afterHomans, were made by Richard M. Emerson, Peter M. Blau and James C.Coleman. But all of them took a more structuralist approach to social exchange,than did Homans. Most holistic was Blau, who was never a methodological individualist,but who became ever more dedicated to structuralism. 34 JamesColeman will be treated at length in chapter 10. Here I concentrate on RichardM. Emerson.Emerson appeared on the sociological scene in 1962, with a seminal articleon ‘Power-Dependence Relations’. A fundamental thesis of this article is thatpower is a property of social relations, or social structure, not of actors (pp. 32ff).From the very beginning, then, Emerson takes a more structuralist approachthan Homans and he does not give up structuralism when turning to the theoryof social exchange (Emerson, 1969; 1972a; 1972b). According to Emerson,Homans’s psychological reductionism suffers from two serious shortcomings:first, it conceives of society as the sum of individual behavior, or more accurately,it fails to offer a conception of society; and second, the overwhelmingburden of ‘explanation’ falls upon ‘given conditions.’ In short, it takes as‘given’ the very conditions sociology seeks to comprehend: the structuresurrounding ‘behaving persons,’ and lawlike changes in that structure.(Emerson, 1972a: 41)A third shortcoming of Homans’s theory of exchange is its exclusive concernwith dyadic relations. This is, of course, a serious limitation of any sociologicaltheory and it was an important step forward, when Emerson generalised the

Positivism in philosophy and social science 195theory of social exchange so as to apply also to social networks (1972b). Fromthen on, Emerson and his colleagues were mainly concerned with experimentalresearch on relations of power in exchange networks (Cook and Emerson, 1978;Cook, Emerson and Gillmore, 1983).In the meantime, a parallel development took place around the so-calledelementary theory of David Willer and his colleagues. Taking their point ofdeparture in classical structuralist sociology (Willer and Andersson, 1981; Willer,1992), rather than in social exchange theory, they arrived at an approach similarto that of Emerson and his followers. In a well-known article from 1988,Markovsky, Patton and Willer criticised the analysis and results obtained byCook, Emerson and Gilmore. This led to a further exchange between the twogroups of researchers, but despite the disagreement between them, there was ageneral feeling that both groups were united in a common project: that ofcreating a new field of social research. The name of the new field became‘network exchange theory’ (see Bienenstock and Bonacich, 1997).This new version of social exchange theory is different from that of Homansin two important respects: it includes a distinctive structuralist element and it hasreplaced behaviourist psychology by rational choice. The latest developmentseems to be the use of game theory to analyse the exchange between individualsin power relations (Willer, 1992; Willer and Skvoretz, 1997; Markovsky, 1997).This transformation of social exchange theory has led to a rapprochementwith two other theoretical traditions in the USA: the network theory of HarrisonWhite and followers, and the rational choice based theory of social exchangedeveloped by James S. Coleman. 35 The former tradition is characterised by astrong commitment to structuralism, which leaves little room for individualistmicrofoundations (White, Boorman and Breiger, 1976; White, 1992: 3, 298,passim; Wellman and Berkowitz (eds), 1988). An exception is Mark Granovetter,who makes frequent use of rational choice as a microfoundation in his analysesof networks (1973; 1978).The second tradition is more individualistic, but does not neglect social structures,as did Homans (see, e.g., Burt, 1982). Coleman, himself, is committed tomethodological individualism, but, as we will see in chapter 10, it is a weak formof individualism, including also an important structuralist element. Coleman’smethodological individualism is, I believe, identical with the structuralindividualismdefended by some Dutch sociologists (see p. 199).Network exchange theory is closer to Coleman’s theory of exchange than tothe network theory of Harrison White. Like the former, it combines actiontheory and structuralism. There is a certain ambiguity on this point in theEmerson approach. 36 According to Cook (1991: 32): ‘The goal was to constructa theory of social exchange in which social structure is the dependent variable’(see also Cook, 1987: 214). Support for this claim can be found in Emerson(1972b: 58). But there is also abundant evidence that both Emerson and Cookuse social structure as independent variable. This is fundamental to the theory ofpower dependence, which sees power as a function, or effect, of the position

196 Positivism in philosophy and social scienceindividuals occupy in social structure. Thus social structure is both dependentand independent variable:Structure is conceived, according to exchange theory, as the interconnectionof various positions in an exchange network. Framed this way, exchangetheory can illuminate not only the behavior of actors, but the structures thatemerge as the result of these exchange relations. Furthermore, by focusingon a given structure or social institution, exchange theory provides an explanationboth for the behavior of actors within structures and for structureitself.(Cook, O’Brien and Kollock, 1990: 160)Homans’s original theory of social exchange was individualistic and psychologistic.Network exchange theory differs from Homans’s theory by introducing astructural element in the theory. Indeed, for many sociologists working in thistradition, the structuralist element is the more important and most of them seemto reject methodological individualism. Karen Cook has argued that Emerson’sapproach is, at least, not a radical methodological individualism. If it is, at all,individualistic, it has to be a form of structural-individualism, since its maincharacteristic seems to be that it links micro and macro or, better, action andstructure (Cook, 1987; 1991; Willer, 1992). 37 Perhaps, it would qualify asmethodological individualism in the sense of political scientist David Knoke(1990), who defends an approach to political networks, akin to the generalapproach of James Coleman, which ‘requires that a structural theorist embrace agenuine methodological individualism’ (p. 27).German VerhaltenstheorieHomans’s theory of exchange was brought to Germany by the Polish sociologistAndrzej Malewski, who was more interested in the behaviourist foundation thanin the idea of exchange. Like Homans, Malewski wanted to integrate socialscience under the edifice of the most general of all theories about humanbehaviour. Not only sociological theories, but also economics, and certainpsychological theories, such as Festinger’s theory of cognitive dissonance, are ona lower level of generality, and can be explained by, or reduced to, behaviouristpsychology. According to Malewski ([1964] 1977: 148 [my translation]),‘Behaviourism is the psychological foundation of all behavioural sciences’ (DieVerhaltenstheorie ist das psychologische Fundament für alle Wissenschaften von menschlichenVerhalten).The most influential representatives of behaviourist social science inGermany, however, were Hans J. Hummell and Karl-Dieter Opp. Theirbehaviourism was part of a more inclusive programme of reducing sociology topsychology. For this reduction to be possible, it is necessary to make clear what ismeant by the terms ‘psychology’, ‘sociology’, and ‘reduction’. ‘We wish to characterizeas psychological all statements about and all concepts concerning absolute

Positivism in philosophy and social science 197properties and relations of individuals, and all statements and concepts concerningrelations of individuals and things.’ ‘We wish to call sociological all statements andconcepts concerning properties and relations of collectivities and concerningcollectivities and things or collectivities and individuals’ (Hummell and Opp, 1968:206f). Reduction, can be achieved in two ways: by a definition of sociological termsby psychological terms, and by a derivation of sociological hypotheses or theoriesfrom psychological hypotheses and theories. By a combination of definition andderivation Hummell and Opp arrive at four different possibilities: (1) neither definabilitynor derivability; (2) definability but not derivability; (3) derivability but notdefinability; and (4) both definability and derivability. The thesis which Hummelland Opp set out to test is the fourth one; that sociological terms are definable bypsychological terms and that sociological hypotheses or theories are derivable frompsychological hypotheses and theories. After a systematic and thorough investigation,which I have discussed and criticised elsewhere (Udehn, 1987:167f ), Hummell and Opp find their thesis corroborated; sociology isreducible to psychology (1968: 220–3; 1971: 1–12; Opp, 1977a).Reduction, however, is not to be regarded as an end in itself. Its justificationlies in the wholesome effects it has on the development of social science, i.e.,sociology: reduction is supposed to lead to more general theories, stated in moreprecise terms and more easily testable, and, therefore, to an integration andaccumulation of our knowledge (Hummell and Opp, 1968: 220–3; 1971: 81–6).Hummell and Opp at first make no mention of ‘methodological individualism’,but later recognise a close kinship between their psychologicalreductionism and Joseph Agassi’s ‘institutional individualism’ (Hummell andOpp, 1971: 8). Still later, they become sceptical about the label ‘reductionism’and suggest that ‘methodological individualist’, or simply ‘individualist’, is abetter characterisation of their programme (Hummell, 1973a: 64f; Opp, 1979:4, 47, 109). The individualist research programme, according to Opp (1972: 6[my translation]), consists in ‘the explicit use of individualistic theories to explainand predict social facts and to bring practical problems closer to a solution’ (unterexpliziter Anwendung individualistischer Theorien soziale Sachverhalten zu erklären,vorauszusagen und praktische Probleme einer Lösung näherzubringen).Opp’s individualist research programme, which is influenced by structuralindividualism,differs from Hummell’s and Opp’s psychological reductionism,above all in the following respects. (1) It aims at explaining social facts ratherthan reducing sociological hypotheses and theories. There is, on the whole, ashift of emphasis from definition to explanation (Opp, 1979: 1–3, 6, 20).Explanation is not possible, however, without a previous definition, or‘Rekonstruktion’, of collective concepts occurring in the explanation (p. 147). (2)The rules of co-ordination (Koordinationsregeln), referred to by Hummell and Oppas the means by which to provide the bridge between sociological and psychologicaltheories, are replaced by transformation functions. Like rules ofcorrespondence, transformation functions may be analytic or empirical(Hummell and Opp, 1968: 208f; 1971: 13ff). (3) The ‘requirement’ of definabilityis replaced by the reconstruction thesis (Rekonstruktionsthese), which says that

198 Positivism in philosophy and social sciencecollective concepts can, but must not, be reconstructed by individualist conceptsdenoting individuals and their properties (Opp, 1979: 116, 150).Neither Hummell’s and Opp’s psychological reductionism, nor Opp’s individualistresearch programme, says anything about the kind of psychological orindividualist theory to be used in the attempted reduction, or reconstruction, ofsociology. Several alternatives are possible, but Hummell and Opp finally settlefor B.F. Skinner’s behaviourist learning theory as the most promising alternative.Other theories that are mentioned include Leon Festinger’s theory of cognitivedissonance and different versions of the so-called ‘theory of balance’. Thebehaviourist research programme then is a sub-programme (Teilprogram) of theindividualist programme. According to Opp (1972: 15–30; 1973: 39–45; 1977b),it consists of the following claims or theses :1 Singular social phenomena can be explained by hypotheses in learningtheory. (Singuläre soziologische Ereignisse sind mittels lern-theoretischer Hypothesenerklärbar).2 Learning theory and sociology are in competition. (Hypothesen der Lerntheorienstehen mit soziologischen Hypothesen in einer Konkurrenzbeziehung).3 Learning theory explains singular social phenomena better than does sociology.(Lerntheoretische Hypothesen können soziologische singuläre Ereignisse zutreffendererklären als soziologische Theorien).4 Learning theory is superior to sociology with respect to truth, empiricalcontent and precision. (Lerntheoretische Hypothesen sind soziologischen Theorienbezüglich ihres Wahrheitsgehalts, ihres Informationsgehalts und ihre Präzision überlegen).5 Learning theoretical hypotheses are better than sociological theories forsolving pratical problems, if the latter are at all useful for that purpose.(Hypothesen der Lerntheorie sind für die Lösung praktischer Probleme besser geeignet alssoziologische Theorien, sofern diese ebenfalls anwendbar sein).6 The use of Skinner’s theory of learning in sociology leads to faster progressin this discipline than does the use of other theories of learning and othersocial psychological theories. (Die Anwendung der Skinnerschen Lerntheorie in derSoziologie führt zu einem grösseren Erkenntnisfortschritt in der Soziologie als dieAnwendung anderer Lerntheorien und die Anwendung anderer sozialpsychologischerTheorien).Hummell and Opp have done a great deal to clarify Skinner’s behaviouristlearning theory, and to turn it into a programme adapted to sociological questions.They have also made great efforts to show that their programme works. Iwill not go into details about this work, but limit myself to a mention of theirtreatment of social structure.As a point of departure Hummell and Opp adopt the orthodox sociologicalview of social structure as a set of relationships between social roles, or positions(1971: 40–4). According to Opp, however, the concept of role is of little theoreticalfertility, if used in addition to the concept of ‘expectation’, by which it isdefined. More precisely, a role is a set of expectations somehow perceived to

elong together. The only fertile use Opp can find for the concept of ‘role’ is inanswer to the question: Under what conditions will certain expectations beperceived to belong together? Typical for any answer to this question is that therole – or the criterion used to sort out the bundle of expectations constitutingthat role – appear as an independent variable (Opp, 1972: 174ff; 1973: 133–46).Hummell uses a somewhat broader concept of ‘social structure’, or ‘socialsystem’, which includes also the unintended consequences of human action. Incompliance with methodological individualism, he proposes a strategy ofresearch, which explains the social processes at the molar level in terms of individuals,acting at the molecular level (Hummell, 1973a: 65). Like manymethodological individualists before, Hummell affirms the reality (in a sense) ofsocial collectives and institutions, but demands their explanation in terms ofindividuals (Hummell, 1973b: 138f, 150). Anticipating the structuralindividualists,Hummell recommends the synthetic explanation of the systemiceffects of the actions of many individuals (1973a: 71f).In the 1980s German Verhaltenstheorie merged with the so-called ‘structuralindividualistresearch programme’, residing mainly in Holland. Among theleading figures of structural-individualism are Siegwart Lindenberg, ReinhardWippler and Werner Raub. Influenced by Homans, Coleman, Hummel andBoudon, the structural-individualists have as their main objective, the explanationof the unintended, composite, or collective, effects of the actions ofindividuals. 38 Their aim is to develop an explanatory micro-sociology – ratherthan a descriptive macro-sociology – and to use it for the explanation, mainly ofmacro-phenomena, but also of micro-phenomena. The explanation of macrophenomenais achieved, materially, by providing a causal mechanism and,formally with the aid of so-called ‘transformation functions’, or ‘transformationrules’, turning laws about individual behaviour into statements about aggregateeffects (Lindenberg, 1977; Raub and Voss, 1981: 88ff).Structural-individualism differs from German Verhaltenstheorie in being less, ifat all, individualistic. As indicated by the name, it is both individualistic andstructuralistic – if such a unity of opposites is possible. I believe that it is, but it isnot methodological individualism as originally conceived. It is a weak form ofmethodological individualism, which is considerably less vulnerable to critiquethan the original strong version (see Udehn, 1987: 202ff). I will return to structuralindividualism in chapter 10, which deals specifically with rational choicesociology.Structural-individualism, at first, differed from German Verhaltenstheorie inbeing neutral with respect to behavioural assumptions. The tendency, however,has been to replace behaviourism by rational choice. Today it seems that allstructural-individualists have adopted rational choice as the individualist part ofstructural-individualism. 39 Positivism in philosophy and social science 199

7 Popperian methodologicalindividualismMethodological individualism was until quite recently – before the renaissance ofAustrian Economics – most often associated with the name of Karl Popper. This,as we have seen, is not because he was its inventor. Nor is it, as we shall see,because he has very much to say about it that is entirely new. Popper did bringabout a major change in methodological individualism by advocating institutionalism,but he did not integrate this element in methodological individualism,himself. One reason methodological individualism used to be more often associatedwith Popper than with the Austrians, is that, being a philosopher himself,Popper was more widely read among philosophers. Another reason is that, in thedecades after the Second World War, he was probably also more widely read bysocial scientists, than were Weber, Mises and Hayek, whose influence was largelyconfined to their own disciplines of economics and sociology. Today this is nolonger true. Interest in the philosophy of Karl Popper has been on the waneamong social scientists for some time, while interest in Austrian Economics hasincreased. My conjecture is, however, that we will soon see a revival of interestalso in the philosophy of Karl Popper.Karl PopperPopper, then, was not an ‘Austrian’ in his methodology for the social sciences. Hewas, however, an Austrian by birth and, what is more, he was a student ofLudwig von Mises and a friend of Friedrich von Hayek. This fact alone makesfor a certain continuity between Austrian and Popperian methodological individualism.More important, however, is the fact that Popper read an early version ofThe Poverty of Historicism at Hayek’s private seminar in London in 1936. In his‘Intellectual Biography’, Popper also tells us that Hayek ‘saved his life’ twice.First, by helping him to find a publisher for The Open Society and Its Enemies (1945)and by writing most encouragingly about it. Second, by offering him a readershipat the University of London, tenable at the London School of Economics,and by publishing ‘The Poverty of Historicism’ as an article in Economica(1944–5), for which Hayek was then acting editor (Popper, 1974a: 95; 1976: 120).When Popper takes up the subject of methodological individualism in The Poverty

Popperian methodological individualism 201of Historicism, it is with direct reference to Hayek, but ironically, in a chapter entitled‘The Unity of Method’.Now all this, I believe, is not only true for the natural, but also for the socialsciences. And in the social sciences it is even more obvious than in thenatural sciences that we cannot see and observe our objects before we havethought about them. For most of the objects of social science, if not all ofthem, are abstract objects; they are theoretical constructions. (Even ‘the war’,or ‘the army’ are abstract concepts, strange as this may sound to some.What is concrete are the many who are killed; or the people in uniform,etc.). These objects are the result of constructing certain models (especially ofinstitutions), in order to explain certain experiences – a familiar theoreticalmethod in the natural sciences (where we construct our models of atoms,molecules, solids, liquids, etc.).(Popper, 1944–5, III: 80; see also 1957: 135)Popper is here, more or less, repeating the view of Weber and Hayek, that theobjects of social science are models or theoretical constructions. The ontologicalstatus of social objects remains obscure, however. Their being theoretical andabstract does not preclude the possibility that they refer to something real. ButPopper certainly gives the impression of denying that social objects are real.What he has to say about social objects – and about the models of naturalscience – comes suspiciously close to an instrumentalist view of scientificconcepts. This is surprising, since Popper in various contexts has appeared as anenergetic defender of a realism in science (cf. Johansson 1975: 98f ). The impressionof instrumentalism remains after reading the following passage, wherePopper makes his first mention of methodological individualism.That we are very often unaware of the fact that we are operating with theories,and that we mistake our theoretical models for things is true, but this isa kind of mistake which is only too common. This use of models explainsand at the same time destroys the claims of methodological essentialism …It explains them, for the model is of an abstract or theoretical character, andwe are liable to believe that we see it, either within or behind the changingobservable events, as a kind of permanent ghost or essence. And it destroysthem because our task is to analyse our sociological models carefully indescriptive or nominalist terms, viz., in terms of individuals, their attitudes,expectations, relations, etc. – a postulate which may be called ‘methodologicalindividualism’.(Popper, 1944–5, III: 80; see also 1957: 136)In this quotation, Popper mentions the doctrine of ‘methodological essentialism’,which he rejects. Opposed to methodological essentialism is ‘methodologicalnominalism’, which he advocates. By ‘essentialism’, Popper means conceptualrealism, or the thesis that there are universals. Nominalism, as we have seen in

202 Popperian methodological individualismchapter 4 (pp. 109f ), is the doctrine that there are only particulars. By adding theword ‘methodological’ before ‘essentialism’ and ‘nominalism’, Popper wants toavoid the metaphysical issue involved and replace it by the methodological issueof how best to achieve the objects of scientific investigation. 1 He suggests thatscientists should not concern themselves with asking essentialist questions, suchas ‘what is matter?’, or ‘what is the nature of man’ and the like, but ask, instead,‘how does this piece of matter behave?’ or ‘how do people act in a certain typeof situation?’ 2 Popper, thus, rejects the use of real definitions in science. It is notentirely clear to me why methodological nominalism is supposed to imply methodologicalindividualism.A corollary of anti-essentialism is that we should not be interested in themeaning of words. One of Popper’s favourite sayings is that ‘words do notmatter’. There is nothing to be gained in precision, or otherwise, from definitions.In defining a word, term, or concept, what we do is to substitute severalwords for one word, thus introducing new words which are left undefined. Theattempt to clarify the meaning of words by way of definition, therefore, leads toan infinite regress. 3Most early adherents of methodological individualism dissociated themselvesfrom psychologism; the doctrine that social science is reducible to psychology.But Popper is more careful, than his predecessors, to distinguish the twodoctrines – probably because he is anxious to avoid subjectivism. To this end,Popper always mentions ‘interaction’ or ‘relations’ in his statements of methodologicalindividualism. If it only mentioned ‘attitudes’, ‘expectations’, or someother ‘psychological propensity’, it would be altogether indistinguishable frompsychologism (see, Popper, 1944–5, III: 86–8; 1957: 152–9).In order to develop a methodology that is a real alternative to psychologism,Popper makes use of two methodological strategies: ‘situational logic’ and ‘institutionalism’.Situational logic is a generalisation of the method of theoreticaleconomics; the science usually cited as the very prototype, or paradigm case, ofmethodological individualism. Briefly stated, situational logic consists inconstructing models of action, based on assumptions about the ends, or goals, ofpeople and about the situation in which they act, and then to ask what theiractions would be on the fundamental assumption that they act rationally toachieve these ends. Now, it is a much debated issue, whether the method of theoreticaleconomics rests on assumptions of a psychological nature or not, butPopper’s standpoint is that it does not. ‘I should like to mention, in passing, that Iconsider neither the principle of methodological individualism, nor that of thezero method of constructing rational models, as implying the adoption of apsychological method’ (1944–5, III: 82; cf. 1957: 142).What Popper means by ‘institutionalism’ is not perfectly clear, but he seems touse the term ‘institution’ in a broad sense, to cover every man-made socialarrangement, from organisations, such as universities and other schools, tolanguage and writing (1944–5, III: 87f; 1957: 154ff). A certain ambiguity iscreated by the fact that Popper suggests that we need both ‘studies, based onmethodological individualism, of … social institutions’, and ‘individualist and

Popperian methodological individualism 203institutionalist models of such collective entities as nations, or governments, ormarkets’ (p. 154). If institutions are to be studied with an individualist method,why do we need models of collective entities that are both institutionalistic andindividualistic? This ambiguity remains when Popper concludes his own institutionalanalysis of progress.I believe that this analysis is typical, and that the human or personal factormust be generally the irrational element in most, or all, institutional socialtheories. The opposite doctrine which teaches the reduction of social theoriesto psychology, in a similar way as we try to reduce chemistry to physics,is, I believe, based on a misunderstanding. It arises from the false belief thatthis ‘methodological psychologism’ is a necessary corollary of a methodologicalindividualism – of the quite unassailable doctrine that we must reduceall collective phenomena to the actions, interactions, aims, hopes, andthoughts, of individuals. But we can be individualists without acceptingpsychologism. The ‘zero method’ of constructing rational models is not apsychological, but rather a logical method.(Popper, 1944–5, III: 88; cf. 1957: 157f)What exactly is the ‘human or personal factor’, if it is not ‘the actions, interactions,aims, hopes and thoughts of individual men’?Popper returns to the topic of methodological individualism in The OpenSociety and its Enemies (1945), where he once again separates it from psychologismand even defends the autonomy of sociology. Concerning his conception ofinstitutionalism, we are, however, still left very much in the dark. Againstpsychologism, Popper says that ‘the defender of an autonomous sociology canadvance institutionalist views’ (Popper [1945] 1966: vol 2, 89ff). Popper, then, isa defender of an autonomous sociology and the view he advances is that noaction can be explained in terms of motives alone, but that a reference to theenvironment is also needed. ‘In the case of human actions, this environment isvery largely of a social nature; thus our actions cannot be explained withoutreference to our social environment, to social institutions and to their manner offunctioning’ (p. 90). To this, the psychologist might counter that social institutionsare man-made and, therefore, to be explained in terms of the psychologicalfactors responsible for their creation and development. This was how someearlier methodological individualists saved methodological individualism fromthe threat of collectivism. But, for Popper, who rejects psychologism, this is nosolution. Or is it? Psychologism, after all, has one praiseworthy aspect:its sane opposition to collectivism and holism, its refusal to be impressed byRousseau’s or Hegel’s romanticism – by a general will or a national spirit, orperhaps, by a group mind. Psychologism is, I believe, correct only insofar asit insists upon what may be called ‘methodological individualism’ asopposed to ‘methodological collectivism’; it rightly insists that the‘behaviour’ and the ‘actions’ of collectives, such as states or social groups,

204 Popperian methodological individualismmust be reduced to the behaviour and to the actions of human individuals.But the belief that the choice of such an individualistic method implies thechoice of a psychologistic method is mistaken (as will be shown later in thischapter), even though it may appear very convincing at first sight.(Popper [1945] 1966: 91)Popper now admits that methodological individualism seems to implypsychologism, but promises to show that it does not. He advances three argumentsagainst psychologism. The first argument is that it ‘is forced to adopthistoricist methods’, and ultimately, ‘to operate with the idea of a beginning ofsociety’. This is so, because a psychological explanation of social institutionsmust start with a state where there were no social institutions, since only in sucha state can psychological factors alone be responsible for the rise of new socialinstitutions. Psychologism, according to Popper, is a form of the historical andmethodological myth of the social contract; the idea that society can beexplained in terms of a pre-social human nature (p. 93). Second, Popper repeatsHayek’s argument that social institutions, far from being ‘explicable in terms ofneeds, hopes, or motives’, are usually ‘the indirect, the unintended and often theunwanted by-product of [human] actions’ (p. 93). Popper’s third argumentagainst psychologism is that social science makes use of the method of situationalanalysis, based upon the ‘logic of the situation’ (p. 97). Popper ends hisdefence of the autonomy of sociology by, once again, reminding us of thehealthy aspect of psychologism; its methodological individualism.Also we must not overlook the great merits which psychologism has acquiredby advocating a methodological individualism and by opposing a methodologicalcollectivism; for it lends support to the important doctrine that allsocial phenomena, and especially the functioning of all social institutions,should always be understood as resulting from the decisions, actions, attitudes,etc., of human individuals, and that we should never be satisfied byan explanation in terms of so-called ‘collectives’ (states, nations, races, etc.).The mistake of psychologism is its presumption that this methodologicalindividualism in the field of social science implies the programme ofreducing all social phenomena to psychological phenomena and psychologicallaws.(Popper [1945] 1966: 98)The problem with Popper’s arguments against psychologism is that they seemto hit methodological individualism too. Popper’s main argument, implies theadoption of an institutionalist method, but this suggestion does not work as longas Popper insists that also institutions must be analysed by an individualistmethod.In Popper’s later writings on the methodology of the social sciences, thetheme of methodological individualism recedes into the background and finallydisappears altogether. After The Open Society and Its Enemies it is not even

Popperian methodological individualism 205mentioned by name. The idea behind it appears, however, in some of the articlesassembled in his Conjectures and Refutations [1962] 1968. In ‘Prediction andProphecy in the Social Sciences’ (1948), Popper sums up the argument of ThePoverty of Historicism. Methodological individualism is mentioned as the onlysound alternative to collectivism, and collectivism as one of two ‘naive theoriesof society which must be disposed of before we can understand the function ofthe social sciences’.The first is the theory that the social sciences study the behaviour of socialwholes, such as groups, nations, classes, societies, civilizations, etc. Thesesocial wholes are conceived as the empirical objects which the social sciencesstudy in the same way in which biology studies animals or plants. This viewmust be rejected as naive. It completely overlooks the fact that these socalledsocial wholes are very largely postulates of popular social theoriesrather than empirical objects; and while there are, admittedly, such empiricalobjects as the crowd of people here assembled, it is quite untrue thatnames like ‘the middle-class’ stand for any such empirical groups. What theystand for is a kind of ideal object whose existence depends upon theoreticalassumptions. Accordingly, the belief in the empirical existence of socialwholes or collectives, which may be described as naive collectivism, has tobe replaced by the demand that social phenomena, including collectives,should be analysed in terms of individuals and their actions and relations.(Popper [1962] 1968: 341)The second naive theory of society recognised by Popper is the conspiracytheory of society, the theory that behind everything that happens in society thereis the will of some individual or group of individuals. This theory is also criticisedin ‘Towards a Rational Theory of Tradition’ (1949), where Popper repeatsand develops his ideas on social institutions, taking what may be characterised asa moderate rationalist position. His main argument against a radical politicalrationalism is the same as Hayek’s: ‘that nothing ever comes off exactly asintended’ (Popper [1962] 1968: 124). As with Hayek, this fact also sets the taskfor the social sciences.It is the task of social theory to explain how the unintended consequences ofour intentions and actions arise, and what kind of consequences arise ifpeople do this that or the other in a certain social situation. And it is, especially,the task of the social sciences to analyse in this way the existence andfunctioning of institutions (such as police forces or insurance companies orschools or governments) and of social collectives (such as states or nations orclasses or other social groups).(Popper [1962] 1968: 125)Popper here makes a distinction between institutions and collectives. Themeaning of this distinction, although possible to guess, is not immediately

206 Popperian methodological individualismobvious. Above all, I find it difficult to understand the classification of states ascollectives. I would classify it as an institution, or organisation, if I had to choose.Popper makes the further distinction between institutions and traditions. Thenotion of ‘institution’, according to Popper, rests upon that of ‘social function’.What characterises a social institution is that it fulfils a social function, whereastraditions simply are. This is interesting, because functional analysis would notnormally be considered compatible with methodological individualism.Institutions and traditions have much in common; among other things thatthey must be analysed by the social sciences in terms of individual persons,their actions, attitudes, beliefs, expectations, and interrelations. But we maysay, perhaps, that we are inclined to speak of institutions wherever a(changing) body of people observe a set of norms or fulfil certain prima faciesocial functions (such as teaching, policing, or selling groceries) which servecertain prima facie social purposes (such as the propagation of knowledge, orprotection from violence or starvation), while we speak of traditions mainlywhen we wish to describe a uniformity of people’s attitudes, or ways ofbehaviour, or aims or values or tastes. Thus, traditions are perhaps moreclosely bound up with persons and their likes and dislikes, their hopes andfears, than are institutions. They take, as it were, an intermediate place, insocial theory, between persons and institutions.(Popper [1962] 1968: 133)The addition of the term prima facie before those of ‘social function’ and‘social purpose’ should probably be interpreted as an attempt to eliminate thesuspicion that Popper’s institutionalism is infected by functionalism. More interestinghowever, is Popper’s separation of persons and institutions. For, what areinstitutions made up of ? The methodological individualist is, I believe, obliged tosay: persons. But an alternative answer is to say that social institutions are madeup of functions, roles, or positions. 4In the 1950s, Popper is occupied mainly with preparing the English edition ofLogic der Forschung and does not write anything about the methodology of thesocial sciences. When he returns to this subject in the 1960s, the theme ofmethodological individualism has almost disappeared. The few scatteredremarks there are about anything resembling methodological individualism areconsistently ambiguous. Institutionalism, however, remains important, but mostimportant is situational analysis, which is now launched as the method of socialscience (see Hedström, Swedberg and Udehn, 1998).‘Situational analysis’ is Popper’s final name for a methodological tool whichhe had earlier called ‘logical’ or ‘rational’ reconstruction, the ‘zero method’, the‘logic of the situation’ and ‘situational logic’. It is a method for constructingmodels in social science, based on the assumption that people act rationally inthe pursuit of their ends. If we know people’s ends, the situation in which theyact when seeking these ends and assume that they act rationally, i.e., choose theadequate means for attaining their ends, then, we can explain why they acted as

Popperian methodological individualism 207they did. A problematic feature of Popper’s situational analysis is the status ofthe rationality principle. Popper denies that the rationality principle is psychological,or even empirical. According to him it is not a theory or a testablehypothesis, but a principle a priori. It is false, however, and therefore not valid apriori. But although false, it is a good approximation to reality. The principle ofrationality is an integral part of most theories in social science. These theoriesare tested as wholes, but in case of falsification, Popper suggests as a methodologicalrule that we keep the rationality principle and treat our analysis of thesituation as refuted (Popper [1963–4] 1994: 169f, 177f). How does this squarewith Popper’s famous principle of falsifiability. I do not intend to answer thisquestion, but I would like to suggest that the origin of Popper’s position is JohnStuart Mill and ultimately Auguste Comte; more specifically, his method ofinverse-deduction, which was borrowed by Mill (see pp. 29, 47).According to Popper, himself, situational analysis is the method of marginalisteconomics. As such, it has much, but not everything, in common with MaxWeber’s ideal type of purely rational action (cf. Jacobs, 1990). Situational analysisis also the inspiration behind Spiro J. Latsis’ critique of ‘situationaldeterminism’ as the dominating research programme within economics (Latsis,1972; 1978). More recently, it is discussed by John Goldthorpe as one version ofrational choice in social science (1998). And, indeed, Popper’s ‘situational analysis’seems to be another name for rational choice. It is possible, however, toconceive of situational analysis in a less restricted way. Among Popper’s pupilsand followers, only J.W.N. Watkins seems to equate situational analysis withrational choice, while Joseph Agassi, Ian. C. Jarvie and John O. Wisdom appearto take a more broad view of this method. But how individualistic is situationalanalysis. That, I suggest, depends upon what is included in the situation.In his contribution to a symposium arranged by the German SociologicalAssociation in Tübingen in 1961, Popper claims that situational analysis is anindividualistic method. But he also makes the following curious statement:‘Institutions do not act; rather only individuals act, in or for institutions. Thegeneral situational logic of these actions will be the theory of the quasi-actions ofinstitutions’ (Popper [1962] 1976: 333). 5 Whatever is that supposed to mean?One interpretation is that the ‘quasi-actions of institutions’ means normativeaction, which may be part of social roles, or not.In Popper’s recently published lecture on ‘Models, Instruments and Truths’(1994), already referred to, Popper mentions social institutions as the mostimportant element of the situation, besides the physical environment.In fact, I propose to use the name ‘social institution’ for all those thingswhich set limits or create obstacles to our movements and actions almost asif they were physical bodies or obstacles. Social institutions are experiencedby us as almost literally forming part of the furniture of our habitat.(Popper [1963–4] 1994: 167)Popper, then, conceives of social institutions as objective and thing-like, very

208 Popperian methodological individualismmuch like Emile Durkheim (see p. 34). This might be an indication that he nolonger adheres to methodological individualism. At least, he does not mentionthis principle in his lecture. In Popper’s latest work (from the middle of the 1960sand onwards), methodological individualism has disappeared altogether. It is noteven alluded to in his Intellectual Autobiography (1974a), and the reason is not hardto find. Popper’s theory of an autonomous world of objective ideas, capable ofacting back upon individuals and their subjective ideas is hardly compatible withthe original Austrian version of methodological individualism. 6Popper’s theory of an autonomous world of objective ideas is part of an ontologicaltheory stating that there are, at least, three distinct worlds in the universe,each of which exists as an autonomous part of reality. They are (1) ‘the world ofphysical objects or physical states, (2) the world of states of consciousness, ormental states, or perhaps behavioural dispositions to act’ and (3) ‘the world ofobjective contents of thought’ (Popper [1968] 1972: 106). Popper discussesWorld 3 mainly from an epistemological point of view, but it is clear that histheory about three worlds has implications also for the methodology of the socialsciences.To World 3 belong, first of all, such things as scientific theories and arguments,but also tools, institutions, traditions, works of art and values (Popper,1972: 113; see also 1976: 187, 195). What Popper calls world 3 is what manysocial scientists call ‘culture’. According to Popper, then, culture, or at least partof it, belongs to world 3. This world, although man-made, transcends individualhuman beings and acts back upon them.Admitting that world 3 originates with us, I stress its considerable autonomy,and its repercussions on us. Our minds, our selves, cannot exist without it;they are anchored in world 3. We owe to the interaction with world 3 ourrationality, the practice of critical and self-critical thinking and acting. Weowe to it our relation to our task, to our work, and its repercussions uponourselves.(Popper, 1976: 196)Popper’s theory of three worlds is a theory of emergent evolution. As such, itbelongs in the holistic tradition. 7 An important element in the theory of emergentevolution, including Popper’s version of it, is that each level in thehierarchic structure of the world is irreducible to the level ‘below’, but exerts acausal influence upon it. Popper calls it ‘downward causation’. In this ‘downwardcausation … the whole, the macro structure, may qua whole, act upon a photonor an elementary particle or an atom’ (Popper, 1977: 19). But there is also adownward causation from society to individuals:The most interesting examples of downward causation are to be found inorganisms and their ecological systems, and in societies of organisms. Asociety may continue to function even though many members die; but a

Popperian methodological individualism 209strike in an essential industry, such as the supply of electricity, may causegreat suffering to individual people.(Popper, 1977: 20)What has been said so far concerns only ontological individualism. It remains toshow that Popper’s theory of three worlds has implications also for methodologicalindividualism.Central to Popper’s methodology of the social sciences and to his theory ofobjective knowledge is the idea of situational analysis. By this is understood ‘acertain kind of tentative or conjectural explanation of some human action whichappeals to the situation in which the agent finds himself ’ (Popper, 1972: 179).From the very beginning, Popper has maintained that this situation is largelymade up of social institutions. What is new, in his theory of three worlds is thatsocial institutions are conceived of as belonging to a reality distinct from individualhuman beings. According to the original version of methodologicalindividualism, institutions do not belong in the situation, at least not if they areconceived of as belonging to an autonomous reality ‘above’ individuals, capableof acting back upon them. According to Austrian methodological individualism,institutions should be analysed and explained as the result of the actions of individuals,but not as a cause of human action. According to Popper, however,human action and thinking is, at least partly, caused by social institutions.According to the Austrian view, institutions exist only as ideas in people’s minds.According to Popper, they exist in an autonomous world of objective ideas ‘overand above’ individuals.Popper’s theory of world 3 may be seen as a ‘secularised’ version of the theoriesof Plato and Hegel, as he admits himself (Popper, 1972: 106, 125f), and it isno less incompatible with original methodological individualism. Popper’s theorydoes not depict society, culture, or history, as some kind of super-person, or semigod,situated over and above human beings and deciding their fates, or standingbehind the scene pulling the strings of human marionettes, but it does say thatsocial institutions, are part of a reality, separate from individual human beings.As such, it is the expression of a view of society, which methodological individualistswere always eager to combat: the hypostatisation of society as an entityindependent of individuals.Popper’s theory of a world of objective thought contents belongs clearly inthe holistic tradition. As Popper observes, himself, it has certain similarities toPlato’s theory of forms and to Hegel’s theory of objective spirit. It may be addedthat it has some similarities to the Marxist theories of objectification, alienationand reification. More obviously, it resembles the later Dilthey’s theory of objectivemind, Durkheim’s theory of collective consciousness, Cassirer’s theory ofsymbolic forms and, perhaps, most obviously, Georg Simmel’s theory of objectivemind and objective culture. It has also much in common withstructural-functionalism and with the theory of culture prevalent in socialanthropology, the most holistic of the social sciences.To sum up and conclude, Popper’s methodological individualism has much in

210 Popperian methodological individualismcommon with that of Weber, Mises and Hayek. Like his precursors, Popperadvances methodological individualism in direct opposition to the naive belief ofcollectivism or methodological essentialism that social objects exist apart from themodels used to construct and to analyse them. In a nominalist, or instrumentalist,fashion Popper holds that social objects are abstractions, or theoretical constructions.Only individual human beings are concrete objects. Methodologicalindividualism is stated as the principle that social phenomena (collectives, institutions,traditions) should be (a) analysed in terms of, (b) reduced to, or (c)understood as resulting from the attitudes, expectations, actions, interactions andrelations of individuals. In particular, it is the task of social theory to explain theexistence and functioning of institutions and social collectives in terms of theintended and unintended consequences of the social actions of individuals.One important difference between Popper and the Austrians is that he ismuch less interested in the meaning of concepts, including collective concepts.For Popper, methodological individualism is not at all about the meaning ofcollective concepts. It is about the explanation of social phenomena.Another difference is that Popper, more explicitly than his Austrian predecessors,advances methodological individualism as a categorical imperative, valid apriori and universally applicable to the social sciences and to history.Methodological individualism is stated, by Popper as the ‘postulate’, ‘demand’,or ‘unassailable doctrine’ that social phenomena ‘should’, or ‘must always’ beanalysed, or understood, in terms of individuals, etc.The main difference between Popper and the Austrians, however, is thatPopper rejects the subjectivism of the latter. Methodological individualism is nolonger a corollary of an intersubjective theory of society. According to Popper,subjectivism leads to psychologism and, ultimately, to a genetic concern with theorigins of society. In order to avoid this predicament, Popper adopts institutionalism.What is more, institutions are conceived of as objective elements of socialreality, not as ideas in the minds of individuals. This creates a tension in Popper’ssocial science methodology which is not easily resolved (cf. Bunge, 1996b: 533;1999: 107f). I have argued that methodological individualism and institutionalism,as stated by Popper, are irreconcilable doctrines. Institutionalism is anaddition to methodological individualism, which does not fit in; a young cuckoowhich eventually replaces methodological individualism as the substantive partof Popper’s social science methodology. I believe that Popper, when fully appreciatingthe consequences of his commitment to institutionalism, silently droppedmethodological individualism, while retaining institutionalism and situationalanalysis as the main elements of his social science methodology.Nevertheless, the tension between individualism and institutionalism inPopper’s thought leads his followers in somewhat different directions. J.W.N.Watkins stays on the individualist side, while Joseph Agassi and I.C. Jarvie try tosit on the fence, but finally come down on the institutionalist side. This is alsowhere Popper eventually lands with his theory of the objective mind.

J.W.N. WatkinsPopperian methodological individualism 211J.W.N. Watkins was a student of Popper, and became his successor, as professorin philosophy, at the London School of Economics. He has probably writtenmore about methodological individualism than any other adherent of thisdoctrine. The reason for this is, I believe, that it fell to his lot to defend methodologicalindividualism against its critics during the height of the debate in the1950s. Because of the many critical attacks, Watkins was forced to attempt aclarification of the doctrine of methodological individualism and to state it in away less vulnerable to critique. A result of this is that Watkins has provided moreexplicit statements of methodological individualism than any other adherent ofthis principle.Watkins’s first treatment of methodological individualism is in the article‘Ideal Types and Historical Explanation’ (1952a), where he presents Weber’sideal type as a forerunner of the principle of methodological individualism (aswe have seen, Weber not only launched the ideal type, but was a spokesman ofmethodological individualism as well). Watkins’s methodological individualism isderived from the fact thatAll social phenomena are, directly or indirectly, human creations. A lump ofmatter may exist which no one has perceived, but not a price which no onehas charged, or a disciplinary code to which no one refers, or a tool whichno one would dream of using. From this truism I infer the methodologicalprinciple which underlies this paper, namely, that the social scientist cancontinue searching for explanations of social phenomena until he hasreduced it to psychological terms. I am not, of course, denying that suchthings as a long-term price movement will partially determine other events,which will be partially in terms of it. I only assert that it too is, in principle,explicable, and explicable in terms of individual attitudes towards thingsand other people.To sum up my argument so far; An understanding of a complex socialsituation is always derived from a knowledge of the dispositions, beliefs, andrelationships of individuals. Its overt characteristics may be establishedempirically, but they are only explained by being shown to be the resultantsof individual activities.(Watkins, 1952a: 28f)According to Watkins, then, his individualism is a methodological principle,which may be inferred from the truistic ontological thesis that social phenomenaare human creations. But is it really a methodological principle? Watkins saysthat the social scientist ‘can’, ‘in principle’ reduce social phenomena to psychologicalterms. This looks more like an epistemological thesis. A methodologicalprinciple is a rule telling us not what we could, but what we should do. This epistemologicalthesis is about explanation, and (with the possible exception ofMenger) Watkins is first, among the methodological individualists discussed sofar, to formulate it as implying a reduction to psychology. It may also be noted

212 Popperian methodological individualismthat the psychological factors Watkins invokes in order to explain socialphenomena are ‘dispositions’ in the sense of Gilbert Ryle (Watkins, 1952a:35–40). It may be recalled that Gilbert Ryle introduced his notion of ‘disposition’as a way of avoiding mentalistic language and as part of a behaviouristanalysis of ‘mental’ phenomena (Ryle [1949] 1963: 301–11). Presumably,Watkins’s use of the term ‘disposition’ – and despite his frequent use of the term‘attitude’ – is motivated by a similar wish to break with the subjectivism ofAustrian methodological individualism.Watkin’s first statement of the principle of methodological individualism gaverise to questions concerning its exact status. He replies that he had first thoughtthat it was analytic, and for three reasons: (1) because it is entailed by the truismthat ‘social things’ are created by ‘personal attitudes’, (2) because the social scientistand the historian have ‘direct access’ only to concrete individuals and (3)because it appeared ‘invulnerable’. Watkins now gives up his belief in the analyticityof the principle of methodological individualism and changes its status‘from a rule to an aspiration’ (Watkins, 1952b: 186). In a revised and expandedversion of his ‘Ideal Types and Historical Explanation’ (1953), can be found arestatement of methodological individualism, together with a statement of itsopposite; methodological holism.This principle [methodological individualism] states that social processesand events should be explained by being deduced from (a) principlesgoverning the behaviour of participating individuals and (b) descriptions oftheir situations. The contrary principle of methodological holism states thatthe behaviour of individuals should be explained by being deduced from (a)macroscopic laws which are sui generis and which apply to the social systemas a whole, and (b) descriptions of the positions (or functions) of the individualswithin the whole.(Watkins, 1953: 729)Watkins has now settled for a strictly methodological version of methodologicalindividualism. The epistemological thesis that social phenomena can, inprinciple, be explained in terms of individuals, is replaced by the methodologicalprinciple that they ‘should’ be thus explained. This means that Watkins has alsoreverted to his original belief concerning its status: Methodological individualismis a rule, not just an aspiration. Watkins gives two reasons for accepting methodologicalindividualism and rejecting methodological holism.1 Whereas physical things can exist unperceived, social ‘things’ like laws,prices, prime ministers and ration-books, are created by personal attitudes.(Remove the attitudes of food-officials, shop-keepers, housewives,etc., towards ration-books and they shrivel into bits of cardboard.) But

Popperian methodological individualism 213if social objects are formed by individual attitudes, an explanation oftheir formation must be an individualistic explanation.2 The social scientist and the historian have no ‘direct access’ to theoverall structure and behaviour of a system of interacting individuals(in the sense that a chemist does have ‘direct access’ to such overallproperties of a gas as its volume and pressure and temperature, whichhe can measure and relate without any knowledge of gas-molecules).But the social scientist and the historian can often arrive at fairly reliableopinions about the dispositions and situations of individuals. Thesetwo facts suggest that a theoretical understanding of an abstract socialstructure should be derived from more empirical beliefs about concreteindividuals.(Watkins, 1953: 729)Methodological individualism is no longer entailed by any self-evident truthabout society and the invisibility of social structures, but is supported by an ontological(1) and an epistemological (2) thesis, respectively, and it is no longeranalytical. ‘The principle whose status I have been trying to elucidate is amethodological rule which presupposes the factual assertion that human socialsystems are not organisms’. And, while there is no counter-evidence to this assertion,‘one cannot assert a priori that it is true’ (1953: 730f). Methodologicalindividualism, then, has an ontological basis, and this basis, according toWatkins, is ‘the assumption that society is not some kind of organism, but reallyconsists only of people who behave fairly intelligibly and who influence eachother, directly or mediately, in fairly comprehensible ways’ (p. 732). The epistemologicalargument used to support methodological individualism is ofunmistakable empiricist origin. Watkins’s talk about ‘direct access’ very muchresembles Bertrand Russell’s notion of ‘knowledge by acquaintance’, and hequotes, with approval, A.J. Ayer’s statement that ‘the English state … is a logicalconstruction out of individual people’ (p. 730).In response to some critical observations by May Brodbeck (1954), Watkinssets out to clarify his ideas on the principle of methodological individualism, but,unfortunately, achieves the contrary; increasing confusion. His first attempt atclarification is intended to defuse any criticism to the effect that social events arethe result, not only of individual activity, but of the physical environment as well.I will begin by reformulating the principle. It is based on the metaphysicalcommonplace that social events are brought about by people. Speakingloosely, one can say that climate, famine, the location of minerals and otherphysical factors help to determine history, just as one can say that alcoholcauses road accidents. But speaking strictly, one should say that alcoholinduces changes in people who drink it, and that it is the behaviour of someof these affected people, rather than alcohol itself, which results in roadaccidents … Thus the fact that physical causes operate in society does not

214 Popperian methodological individualisminvalidate the assumption on which the principle of methodological individualismrests. For they operate either by altering (and sometimes destroying)people, or through people’s ideas about them. In either case it is people whodetermine history, however people themselves are determined. Now if socialevents like inflation, political revolution, ‘the disappearance of the middleclasses,’ etc., are brought about by people, then they must be explained interms of people; in terms of the situations people confront and the ambitions,fears and ideas which activate them. In short, large-scale socialphenomena must be accounted for by the situations, dispositions and beliefsof individuals. This I call methodological individualism.(Watkins, 1955a: 58)So far, Watkins has only achieved a restatement of his earlier version ofmethodological individualism, but with an ambiguous relation to its ontologicalbasis. The ambiguity consists in the fact that he, on the one hand, argues that itis ‘speaking loosely’ to say that physical factors help to determine history, and, onthe other hand, says that the situation helps to explain social phenomena.Unless, of course, Watkins intends by the ‘situation’ only other individuals. Buteven so, he runs into trouble, since also other individuals exert only an indirectinfluence upon the actions of a particular individual. So, it would seem, Watkinswill either have to admit ‘loose speaking’ in the explanation of socialphenomena, or else, omit any mention of the situation in his statement ofmethodological individualism.Watkins also breeds confusion by seemingly changing methodological individualismfrom a principle concerning the explanation of social phenomena into athesis about the meaning of concepts. When, in his polemic with Brodbeck, hegoes on to discuss the relation of concepts such as ‘group climate’ and ‘cohesiveness’to the principle of methodological individualism, it turns into a thesis aboutthe meaning of holistic concepts.I do not see how such concepts can be meaningful and yet involve no referenceto individuals. If ‘The Jewish race is cohesive’ does not mean that, forinstance, Jews usually marry Jews, live in close communities, share religiousrituals, etc., if it does not refer to Jewish people (whose behaviour can beobserved), then I do not see how it can be tested or have any empiricalcontent.(Watkins, 1955a: 61)Concern with concepts seems to be a temporary aberration, however. Watkins’sfinal word on this matter is decidedly against a preoccupation with the meaningof concepts.I should have realized, when I insisted that all large-scale happenings are, inprinciple, individualistically explainable, that there are many social methodologistswho concern themselves with the uninteresting question of

Popperian methodological individualism 215analysing sociological concepts rather than with the interesting question ofways of explaining what those concepts describe, and that these methodologistswould naturally tend to mis-read me as insisting that all sociologicalconcepts are individualistically analysable.(Watkins, 1959b: 243)As we saw in the previous quotation, this is exactly what Watkins insisted inhis reply to May Brodbeck. The statement in the last quotation is more typical ofWatkins, however, and more in keeping with the ‘spirit’ of Popperian philosophy,which certainly does not recommend any preoccupation with the meaning ofconcepts. Another thing to notice about this quotation, is that Watkins hasreturned to his initial formulation of methodological individualism as an epistemologicalthesis: social phenomena are said to be individualistically explainable,in principle.Watkins’s last extensive treatment of methodological individualism is in‘Historical Explanation in the Social Sciences’ (1957b), where it undergoes stillanother change. Methodological individualism now becomes an ontological, ormetaphysical, thesis about the ultimate constituents of the social world.According to this principle [methodological individualism], the ultimateconstituents of the social world are individual people who act more or lessappropriately in the light of their dispositions and understanding of the situation.Every complex social situation, institution, or event is the result of aparticular configuration of individuals, their dispositions, situations, beliefsand physical resources and environment.(Watkins, 1957b: 105f)The reason why Watkins changes his formulation of methodological individualismfrom a strictly methodological rule to an ontological thesis may be that hehas come to feel that it is too strong a demand to put upon explanations in socialscience and history, that they must all be in terms of individuals. This guessderives some support from Watkins’s recognition of so-called ‘half-way’ explanations.There may be unfinished or half-way explanations of large-scale socialphenomena (say inflation) in terms of other large-scale phenomena (say, fullemployment); but we shall not have arrived at rock-bottom explanations ofsuch large-scale phenomena until we have deduced an account of themfrom statements about the dispositions, beliefs, resources and inter-relationsof individuals.(Watkins, 1957b: 106)The difference may seem unimportant, but there is undoubtedly a weaker claiminvolved after the substitution of ‘rock-bottom explanations’ for merely ‘explanations’.The reason is that this allows for the use of so-called ‘half-way

216 Popperian methodological individualismexplanations’ in social science and history; if only in the interim, while awaitingthe ‘rock-bottom explanation’ of large-scale social phenomenon in terms of individuals,their dispositions, beliefs and interrelations. The recognition of half-wayexplanations illustrates the difference between epistemological and methodologicalindividualism. The epistemological fact that social phenomena areexplainable and/or definable in terms of individuals, does not imply themethodological rule that they should be explained in this way.In this article, Watkins introduces a still weaker version of methodologicalindividualism and contrasts it with sociological holism.If methodological individualism means that human beings are supposed tobe the only moving agents in history, and if sociological holism means thatsome superhuman agents or factors are supposed to be at work in history,then these two alternatives are exhaustive.(Watkins, 1957b: 106)This version is weaker than his earlier versions, because it is possible to agreewith Watkins that human beings are the only moving agents in history, and yet todeny that human beings are the only constituents of social reality, or the onlycauses of social events.Watkins’s weakening of methodological individualism is the result of hisencounter with the anthropologist Leon Goldstein, who accused him of breedingconfusion by using two different principles of methodological individualism; onetruly methodological and one ontological (Goldstein, 1958). Watkins answersthat he has distinguished these two principles from the very beginning, and so,denies being guilty of confusion. What he does not mention, however, is that heoriginally reserved the name ‘methodological individualism’ for the strictlymethodological principle, while later using this name also for the ontologicalthesis. This is unimportant. More important is the fact that Watkins now makes asynthesis of the two, so that methodological individualism comes to compriseboth versions: ‘(1) Human beings (together with their material resources andenvironment) are the only causal factors in history. (2) Explain all social events interms of human factors’ (Watkins, 1959a: 320).Watkins did return to methodological individualism in an article on Hobbesfrom 1976, where he restates and rejects his position in the 1950s. This retrospectivestatement is perhaps the most clearly worded formulation there is of theoriginal, strong version of methodological individualism. In order not to miss, ormisinterpret, anything in this careful and authoritative statement of methodologicalindividualism, I quote it in full:Let S be any (human) society, institution, or social process, and let an I-predicate be a one-term predicate that is predicable of individual people;that is to say, it always makes sense, though it may be false, to predicate anyI-predicate of any individual. I-predicates can designate physical, psychological,and psycho-physical properties of individuals. Thus ‘tall’, ‘ambitious’

Popperian methodological individualism 217and ‘weary’ are all I-predicates. Let an R-predicate be a relational predicatewhich designates a relation either between individuals (‘a is a cousin of b’) orbetween an individual and a thing (‘a owns this house’) or between individualsand things. There is no restriction on the type of relation that anR-predicate can designate – it may be spatial, physical, biological or social –provided that it is a relation into which individual people can enter.Then my old position can be restated thus:1 An adequate description of S will essentially involve predicates – say S-predicates – that are neither I- or R-predicates.2 However, the explanans of an adequate explanation of the formation ofS, or of the subsequent functioning of S or of changes in S, will essentiallyinvolve only I- and R-predicates. If S-predicates still figure in ourexplanans we have an ‘unfinished’ or ‘half-way’ explanation: we couldproceed to a deeper explanans containing no S-predicates.3 Moreover, explanations of the formation of properties designated bythe I-predicates in our explanans for S will in turn essentially involve I-predicates but not S-predicates.(Watkins, 1976a: 710)Despite the apparent clarity of this statement of methodological individualism, itinvolves one important ambiguity. The notion of an ‘R-predicate’ begs a fundamentalquestion in the metaphysical issue of individualism versus holism: Whatis the nature of social relations? Are they external, or are they internal. Only ifthey are best conceived of as external relations is it possible to accept Watkins’statement as an adequate explication of methodological individualism. I willreturn to the matter of internal versus external relations in a sequel to this workon social holism. For the time being I merely notice that Watkins no longer (or atleast in 1976), subscribes to this version of methodological individualism. He stillbelieves, though, ‘that the right way to explain social phenomena is to exhibitthem as the (largely unintended) resultants of the activities of interacting individuals’(p. 712).I suggested in the introduction to this section that Watkins tried to clarify theprinciple of methodological individualism in reply to its critics. I do not think hewas altogether successful in this endeavour. There were too many shifts in hisposition. Paradoxically, however, Watkins has done us a service, by failing to statemethodological individualism unambiguously as one single principle. By failingto do so, he has called our attention to the fact that what people have called‘methodological individualism’ is really a number of distinct, but interrelated,theses and principles.Strictly speaking, methodological individualism is a rule telling us to explainsocial phenomena in terms of individuals, etc. This rule is supported by the epistemologicalthesis that such explanations are possible in principle and the

218 Popperian methodological individualismontological thesis that social phenomena are caused by individuals and nothingelse. Since Watkins is an adherent of the covering-law model of scientific explanation,methodological individualism can also be stated as a principle aboutlaws. Social phenomena can, and should, be explained by being deduced fromindividualist laws and initial conditions. There are no irreducible social lawsgoverning the behaviour of collectives. Methodological individualism has, finally,been stated as a principle concerning the meaning of concepts, but Watkins findsthis version less interesting.Watkins is a bit uncertain about the status of methodological individualism.He first believed that it is an analytical principle, valid a priori, then changed itsstatus from a rule to an aspiration and back again to a rule, justified mainly byphilosophical (epistemological and ontological) arguments. As distinguished fromPopper’s categorical version of methodological individualism, which admits ofno exception, Watkins’s version allows for half-way explanations in terms ofmacro phenomena.Institutional individualismMethodological individualism has also been defended by some other Popperiansand, especially, by Joseph Agassi and Ian C. Jarvie. They have both done a greatdeal to clarify and develop certain ideas in Popper’s philosophy of the socialsciences. They have made explicit what was only implicit and they have developedwhat was embryonic. Above all, they have tried to bring together Popper’smethodological individualism with his institutionalism. But in attempting this,Agassi and Jarvie reached for the impossible. As stated by Popper, methodologicalindividualism and methodological institutionalism are irreconcilabledoctrines. In their hands, therefore, methodological individualism eventuallyturned into something else, which Agassi called ‘institutionalistic individualism’.Joseph AgassiIn his highly suggestive article ‘Methodological Individualism’ (1960), JosephAgassi declares that his aim is ‘to defend institutionalistic individualism’, whichhe considers ‘to be Popper’s great contribution to the philosophy of the socialsciences’ (p. 244). Institutionalistic individualism is a via media between psychologismand holism, which combines the reasonable elements in both doctrines,while rejecting what is untenable (p. 248). Another important part of Popper’ssocial science methodology is situational logic (p. 264), or, what today goes by thename of ‘rational choice’. According to Agassi, the combined use of individualism,institutionalism and situational logic is the characteristic mark of allserious work in social science and history.I agree that Popper’s institutionalism, combined with situational logic, is animportant contribution to the methodology of the social sciences. I also agreethat it represents a via media between psychologistic individualism and theextreme view that wholes have aims and interests of their own. I even agree that

Popperian methodological individualism 219this methodology may be conceived of as a form of individualism. But it is notcompatible with Austrian methodological individualism and it is not compatiblewith Popperian methodological individualism, as stated by Popper, himself, andby Watkins. Institutionalism and methodological individualism are traditionallyopposite views, 8 and, as Agassi eventually realises, they remain opposites also inthe writings of Popper and Watkins. Agassi’s institutionalistic individualism,therefore, is really a new version of methodological individualism. Before realisingthis, however, Agassi gets entangled in the same contradictions as Popper,himself.The main thesis of Agassi’s article is that the controversy between holism andindividualism has been based upon the implicit assumption that ‘if “wholes”exist, then, they have distinct aims and interests of their own’, and that Popperhas developed a via media between the two traditional views, by rejecting thisassumption. According to Agassi, Popper ‘asserts that “wholes” do exist (though,of course, not in the same sense in which people exist), but they have no (distinct)interests’ (1960: 247f). As we have seen, the belief that holism implies that wholeshave aims and interests of their own does, indeed, loom large in the writings ofmethodological individualists. It is not at all that common among holists themselvesand Popper’s combination of institutionalism and situational logic is notthe only via media between psychologism and the view that wholes have aims andinterests of their own. But it is one possible alternative to these doctrines.Thus … not the aims of institutions but rather their existence affects theindividual’s behaviour; the existing institutions constitute a part of the individual’scircumstances which together with his aims determine his behaviour… While according to psychologistic individualism only material conditionsmay be considered as relevant circumstances, according to Popper the existenceof institutions may be considered as relevant circumstances too.(Agassi, 1960: 247)This is a fair characterisation of Popper’s institutionalism, but it differs fromhis methodological individualism, by including institutions in the relevantcircumstances. Indeed, if institutions are ‘wholes’ affecting the behaviour of individuals,the term ‘institutional individualism’ looks very much like a contradictio inadjecto.According to Agassi, then, Popper’s ‘institutional individualism’ differs frompsychologist individualism, by allowing us to take institutions for granted andusing them to explain the behaviour of individuals (1960: 261–3). But it differsalso from Popper’s methodological individualism, and Agassi, like Popper, hasgreat difficulty separating the two. According to Agassi, psychologistic individualismentails the view that ‘all statements about societies and social institutions …should be viewed as shorthand assertions about many individuals’ (p. 246). Butthis does not make psychologist individualism different from Popper’s methodologicalindividualism. According to Popper, methodological individualism isequal to the analysis of sociological models in nominalist terms (see pp. 201f ),

220 Popperian methodological individualismand nominalist definitions are ‘shorthand symbols … introduced to cut a longstory short’ ([1945] 1966: vol. 2, 14).Agassi also says that the individualist’s acceptance of institutional analysisamounts to this: ‘He does not deny that the behaviour of an individual isconstrained and influenced by social factors provided that we can explain suchconstraints and influence as results of choices of other individuals’ (Agassi, 1960:245). On this view, institutionalism ‘accords with the classical individualistic ideathat social phenomena are but the interactions between individuals’ (p. 267).This seems to me a fair account of methodological individualism, but closer topsychologism than to institutionalism. Instead of taking institutions for granted,the individualist is now supposed to explain them as the result of the choice ofother individuals or, in other words, as the result of interaction between individuals.Institutions have disappeared from view. They have dissolved into thetheory of ‘psychologistic individualism’ that ‘society is the sum-total of individuals’interactions’ (p. 264), or in Popper’s words ‘the product of interactingminds’ ([1945] 1966: vol. 2, 90). If Agassi is able to find a difference betweenpsychologism and methodological individualism, this is only because, at times, heconceives of psychology as individual psychology (1960: 246). Methodologicalindividualism ‘does not accord with the classical individualistic-psychologisticidea that this interaction depends on individuals’ aims and material circumstancesalone’ (p. 267). A strange view, this classical psychologist individualism.To see society as the sum-total of individuals’ interactions, and yet, to deny thatindividuals, when interacting with each other, affect each other’s behaviour.Agassi eventually comes to express serious doubts about the correctness of hisearlier interpretation of Popper’s methodological individualism and its relationto psychologism. In reply to J.O. Wisdom’s criticism that Popper’s methodologicalindividualism seems reductionist even in Agassi’s interpretation of it, Agassinow says that he does not know how Popperian it is. He agrees with Wisdom‘that Watkins’s formulation of methodological individualism is reductionist, asHayek’s and Keynes’s (I was not aware that Keynes ever formulated methodologicalindividualism, or even accepted it). ‘Popper himself, I now tend to agree withGellner is unclear about matters and so is free for all’. Also, ‘I do not mean toreaffirm the position I held in 1960, and I do not know whether today I will at allendorse methodological individualism’ (Agassi, 1972: 326).This last statement notwithstanding, Agassi did reaffirm his earlier position.In a new article, now with the title ‘Institutional Individualism’ (1975), Agassirepeats the views of his ‘Methodological Individualism’, but without ascribingthose views to Popper and without designating them ‘methodological individualism’.Still later, Agassi (1977: 290ff) drops the notions of ‘methodological’ and‘institutional individualism’ altogether. He now prefers to discuss the issuesinvolved in terms of ‘psychologism’ versus ‘collectivism’ and of the reducibilityof sociology to psychology. Agassi’s own position is that of anti-reductionism, aposition which he ascribes to Hayek and Popper. If we are to believe Agassi,Hayek has come to realise, with Popper, that reductionism is not a viable option.

Popperian methodological individualism 221‘And so the edge of his psychologism is gone; indeed if we follow Popper we haveto admit that the edge of all reductionism is gone’ (p. 292).In a recent article, written in celebration of Popper’s Open Society, Agassi hasapparently grown weary of the attention paid to the lack of clarity about institutionsin Popper’s methodology, ‘let us not spend time on efforts to clarify a textbeyond a reasonable limit. Investments in clarification of texts can prove themnot acceptably clear; efforts to clarify them should yield some return’ (Agassi,1997: 510). Since it is my aim, in this book to clarify the meaning of ‘methodologicalindividualism’, I hope that I am excused, if I have tried to reach someclarity about Popper’s institutionalism, but I agree with Agassi, that it is reallynot possible: ‘Popper’s theory of the autonomy of sociology renders institutionsreal in some sense. He was unclear about this. So it is better left alone. We havethus reached a limit on reading Popper on individualism in all respects saveresponsibility’ (Agassi, 1997: 510).Agassi (1977: 298, note 62) refers to Ian C. Jarvie’s Concepts and Society (1972)as the clearest statement of the diverse contemporary views regarding individualism.This may be, but, in my opinion, Jarvie does not succeed in his attempt toreconcile Popper’s methodological individualism with his institutionalism and histheory of world 3, without removing the inconsistencies.Ian C. JarvieIan Charles Jarvie entered the intellectual scene as an astute defender ofPopper’s methodological individualism, guaranteeing that all statements aboutsocial classes could be ‘reduced to individualist terms’. This guarantee is basedupon a firm belief in nominalism and, therefore, valid for all social wholes orcollectives. To the question: ‘Why is the soldier more ontologically “real” thanthe Army?’, Jarvie replies: ‘Simply because “Army” is merely a plural of soldierand all statements about the Army can be reduced to statements about theparticular soldiers comprising the Army. The Army has, then, no “being”, thereis no “Ghost in the machine”’ ( Jarvie, 1959: 57). 9 Jarvie’s first explicit statementof methodological individualism is a restatement of Watkins’s final version. Itcomprises ‘the metaphysical theory … that human individuals are the onlycausal factors in society’, and the consequent ‘axiomatic methodologicalprescription: “Explain all social events in terms of human factors”’ ( Jarvie, 1961:12).In his first book, The Revolution in Anthropology (1964b), Jarvie changes sidesfrom Watkins to Agassi. He now argues that institutions must be reckonedamong the determinants of human behaviour and, therefore, enter explanationsof social action.Popper’s great discovery is that aims and ideal types are not enough toexplain action; circumstances, material and institutional, are required in theexplicans. Institutions are, so to speak, a third force in the society and their

222 Popperian methodological individualismexistence and importance is the reason sociology is autonomous with respectto psychology.( Jarvie, 1964b: 113)The immediate impression created by this statement is that Popper’s ‘greatdiscovery’ conflicts with the premiss of methodological individualism, as statedby Jarvie above. It could be, of course, that the material and institutional circumstancesare included in the ‘etc.’, but this is not likely. Institutions do not belongto the same category as human decisions aims and actions. Also, matters of someimportance do not belong in the et cetera. Another conflict, in my view, is thatbetween Jarvie’s view of institutions as a ‘third force’ in society, and his earlierview that individuals are the only causal factors in society. This conflict too, canbe made to disappear; if you admit of forces that are not causes. But this is evenmore unlikely. Recognising institutions as social forces is tantamount tosuggesting that they are causal factors in social life.Another thing to notice about Jarvie’s version of methodological individualism,is that it has turned from a principle about the explanation, or analysis, ofsocial phenomena – which it was always before Agassi – into a principle aboutthe explanation of human behaviour, or action. The premiss of methodologicalindividualism now is ‘that all human behaviour must be explained in terms ofhuman decisions, aims, actions, etc’ (p. 111). An important consequence of thischange becomes apparent in Jarvie’s review of Robert Brown’s Explanation inSocial Science.It is an important metaphysical issue to decide what we are prepared toexplain, and in what terms we are prepared to explain it. Social events can beexplained in terms of other social events, individual decisions can beexplained in terms of social events, and so on. But the key metaphysical questionis whether social action can be explained other-than-individualistically.( Jarvie, 1964a: 70)Jarvie’s ‘methodological individualism’, now, is an extremely permissive one.It does not deny the existence of social wholes, nor does it prohibit explanationsof social events and individual decisions in terms of social wholes. Its restrictionsconcern only the explanation of action, which must be in terms of the aims ofindividuals, since only individuals have aims. But not even social actions can beexplained in terms of aims alone. Material circumstances and social institutionsare part also of the explanation of social action. Compared to the claims ofuniversal applicability, made by earlier adherents of methodological individualism,Jarvie’s statement represents a dramatic reduction of its intended scope.Jarvie’s main complaint against Brown’s book is that it does not treat theimportant holism–individualism issue. To make up for this deficiency, he gives hisown account of this important problem in the metaphysics and methodology ofthe social sciences.

Popperian methodological individualism 223Holism is that view of society which claims that there are autonomous socialwholes with aims; individualism is the position that only human individualshave aims. These two positions clash in a whole series of ways. Holists tendto be interested in the broader picture, historical patterns, wars, revolutions,hydraulic civilizations, and the like. Individualists may treat of these things,but they also look at the minutiae of day-to-day social behaviour, and alwaystheir explanations are in principle reducible to those involving typical individualsacting to realise certain ends in a situation composed of otherindividuals and of institutions and traditions, or to macro-effects of manyindividuals so acting, or to unintended effects of individuals so acting.Holists tend to look for social wholes and to use them as explicans; individualiststend to put social wholes in the category of explicanda. What isexplanation for one is problem for the other.( Jarvie, 1964a: 69)The first two sentences in this quotation reproduce Agassi’s idiosyncraticrendering of metaphysical holism, as implying that wholes have aims of theirown. The last two sentences state the assumed methodological corollary of thisview. I believe it is an important observation that holists use social wholes asexplanans and individualists put them in the explanandum (I use another, morecommon, terminology than Jarvie). I also believe that this observation suggests areasonable explication of methodological individualism as the principle that theexplanans (that which explains), in a social scientific explanation, must refersolely to individuals and their interaction and material circumstances, but not tosocial wholes.The problem with Jarvie’s account of the holism–individualism issue is thatthere is a gulf between the metaphysical and the methodological problemsinvolved. It is true that holists tend to put social wholes in the explanans, but thisdoes not, in the least, imply that wholes have aims of their own. It implies onlythat they exist and make a difference. If it is correct, as Jarvie suggests, thatholists put social wholes in the explanans, then his own version of ‘methodologicalindividualism’ is holistic. I have already quoted Jarvie to the effect that‘institutions are irreducible wholes’, and that they ‘are required in the explicans’.There is a way out of this dilemma, but the costs are high: ‘To try to preventmisunderstanding I should stress that individualism includes institutions as irreduciblewholes. Individualism is not the view that only individuals exist; it is theview that only individual aims exist’ (1964a: 70, note 1). To maintain that ‘individualismincludes social wholes as irreducible wholes’ looks very much like acontradiction in terms.Jarvie is aware that his interpretation of methodological individualism mayappear unorthodox. When he returns to this subject in Concepts and Society (1972),it is with much precaution, since ‘some ideas in this book may be thought to beincompatible with methodological individualism’ ( Jarvie, 1972: xii). One suchidea, no doubt, is that ‘among the cardinal realities facing the individual … arehis social surroundings, especially institutions’, and that ‘these are as concrete

224 Popperian methodological individualismand as real as his physical surroundings’ (p. xiii). What, then, is methodologicalindividualism? Well, it is ‘not a reductionism that would eliminate all but individualsfrom sociological explanation’ (p. 157). In order to clear up this‘misunderstanding’, Jarvie invokes J.O. Wisdom, who has made another attemptto reconcile Popper’s institutionalism with his methodological individualism. 10John O. WisdomWisdom’s point of departure is the problem that Popper’s methodological individualism‘looks like’ a reductionist theory and, therefore, as incompatible withhis institutionalism. According to Wisdom, however, reductionist individualismand institutionalism are but two poles in Popper’s methodology and the contradictionbetween them is merely apparent. ‘In fact, Popper’s view allows a placeto both poles, and (assuming that this is not an inconsistency, and I hold it is not)the first task is to dissect the thesis and present it free of apparent contradiction’(Wisdom, 1970: 272). Wisdom’s solution to the problem of apparent contradictionis something called ‘situational individualism’ (p. 290). According to thisposition, social wholes are ‘partially’ or ‘distributively’, but not ‘globally’, or‘collectively’, reducible to individual activities.The ‘reductionist’ programme aims at dispensing with all institutionalwholes after the ‘reduction’ is carried out. Now Popper is fully aware, Ithink, that this is impossible, for you can dispense with one or even moreinstitutional wholes, but only in an institutional setting … [Popper’s] viewinvolves a partial reduction, in which any one institutional whole is‘reducible’ though not all such wholes at once. That is to say, when anygiven institution is ‘reduced’ to the aims of individuals, this is effected onlyat the cost of introducing some other whole, which in turn can be reducedbut only at a similar price … Thus, with ‘reduction’ in this form, whateverwhole is ‘reduced’, some whole is always left over ‘unreduced’. This I take tobe Popper’s position.(Wisdom, 1970: 274)If this is, indeed, Popper’s position, he certainly does not say so in his ownwritings about methodological individualism. The position Wisdom ascribes toPopper is more of a reconstruction than of an interpretation. As such, it mayhave some merits of its own, but it is uncertain if it qualifies as a version ofPopperian methodological individualism. The problem with Wisdom’s reconstructionis that like Buridan’s ass, it cannot decide which alternative to choose.The result is that it is incompatible both with Popper’s methodological individualismand with his institutionalism.Wisdom’s analysis hinges, in part, on a faulty interpretation of Popper’s use ofthe notion of ‘unintended consequences’. He suggests that, according to Popper,both actions and institutions have unintended consequences (p. 275). 11 This ismost probably wrong. According to Popper, only the actions of individuals can

Popperian methodological individualism 225have unintended consequences. This becomes clear when Popper contrasts hisown analysis of unintended consequences to that of Marx.For Marx is a methodological collectivist. He believes that it is the ‘system ofeconomic relations’ as such which gives rise to the unwanted consequences –a system of institutions which, in turn, may be explicable in terms of ‘meansof production’, but which is not analysable in terms of individuals, theirrelations, and their actions. As opposed to this, I hold that institutions (andtraditions) must be analysed in individualistic terms – that is to say, in termsof the relations of individuals acting in certain situations, and of the unintendedconsequences of their actions.(Popper [1945] 1966: vol. 2, 323f, note 11)I do not believe this is a correct interpretation of Marx’s view, but the importantpoint, for my present purposes, is that Popper denies that institutions, assuch, have unintended consequences. Popper admits that institutions may haveunintended consequences. But not in addition to the unintended consequencesof the actions of individuals. According to Popper’s methodological individualism,institutions are the actions of individuals. Even the unintendedconsequences of the actions of individuals are actions of individuals. They areactions of other individuals (or some natural event). Popper, then, does not holdthat both actions and institutions have unintended consequences. But neitherdoes Wisdom, it would seem:Thus the independent power of every institution lies in the capacity toproduce unforeseen unintended consequences; but these are the results onlyof individual intentions, so that nothing over and above the individual intentionsand their consequences is needed for building up the content ofinstitutions. Thus Popper can maintain a position with two poles, which areapparently incompatible, namely individualism and institutionalism.(Wisdom, 1970: 276)I am unable to see that Wisdom has succeeded in removing the apparent incompatibility.How, if the unforeseen unintended consequences are the results only ofindividual intentions, can it be an independent power of institutions to producethem?Wisdom is correct to ascribe the following thesis to Popper: (1) institutionscannot be reduced to the aims or purposes of individuals. Wisdom’s mistake is tojump from this thesis to the conclusion that Popper adheres also to (2) ‘institutionscan never be expressed wholly in terms of individuals’, and also to (3) ‘theidea of an institution or of society as an independent source of power’. In theearly writings referred to by Wisdom, Popper, far from holding (2), repeatedlyinsists that institutions must be reduced to individuals (see pp. 201, 207f). As to(3), it is not only absent from Popper’s early writings, but it is hard to see that itfollows even from Wisdom’s ‘situational individualism’. If, as it says, any one, that

226 Popperian methodological individualismis every, social whole, though not all at once, can be reduced to the aims of individuals,how can they have an independent power of their own? It is equallydifficult to accept Wisdom’s argument that a collective, but not a distributivereduction, ‘would render social wholes mere epiphenomena of individual activities’(p. 290). Wisdom’s situational individualism seems to be an epistemologicalthesis and, as such, ontologically neutral. The apparent ontological implicationof situational individualism, moreover, seems absurd. According to situationalindividualism, reduction is a relative matter. Every social whole can be reduced,but only at the cost of leaving some other social whole unreduced. If this thesishas any implication for the ontological status of social wholes at all, it seems tobe this: social wholes are sometimes epiphenomena and sometimes not,depending upon what we choose to reduce and what to leave unreduced in eachparticular distributive reduction.By suggesting that social institutions are partially, or distributively, reducible toindividuals, Wisdom’s situational individualism is incompatible also withPopper’s institutionalism, at least as part of his theory of world 3. There is nodoubt, whatsoever, that social institutions, according to this theory, are altogetherirreducible to the thoughts and actions of individual human beings. 12I find it hard to make sense of Wisdom’s situational individualism. It is notclear to me if he is an institutional individualist, like Agassi and Jarvie, or anoriginal methodological individualist, like Watkins, admitting of half-way explanations.His argument that all institutions are reducible, just not all at once, maybe interpreted as a claim about the merely practical impossibility of completereduction.ConclusionPopper did not, himself, make any radical break in his statement of methodologicalindividualism. But this is not to say that he was a mere epigone of hispredecessors. Far from it. Popper did introduce some new elements in his versionof methodological individualism. In particular, he broke with the subjectivism ofthe Austrians. Most important, however, Popper was a methodological institutionalist,in addition to being a methodological individualist, and this turned outto be of considerable importance for the further development of methodologicalindividualism. The simultaneous adoption of two traditional opposites, created atension in his social science methodology, which he never solved, himself. It waseventually solved by his disciple Joseph Agassi, who created a new version ofmethodological individualism out of two originally distinct elements in Popper’smethodology: methodological individualism and institutionalism. He called it‘institutional individualism’ and it differs from the original Austrian version ofmethodological individualism, by assigning an important explanatory role toobjectively existing social institutions in social scientific explanations.Institutional individualism is, I believe, incompatible with methodological individualism,as stated by Popper himself, but it is, I also believe, true to the spirit ofhis methodology of the social sciences, including his institutionalism, his situa-

Popperian methodological individualism 227tional logic and his theory of World 3. For this reason, and also because, it wasdefended by most of his followers, I conceive of institutional individualism as‘Popperian institutional individualism’.Figure 7.1 Popperian institutional individualismAs I have already suggested in the previous chapter, we may call the originalversion of methodological individualism, the strong version and Popperian institutionalindividualism a weak version. The main difference between them is thatthe latter, but not the former, admits of social institutions in the explanans, orantecedent, of social scientific explanations (arrows from S to I in Figure 7.1). Itshould be added, that arrows turned downwards represent, not only a transitionfrom macro to micro, but an explanation of the actions of individuals in terms ofreal social institutions. There is, therefore nothing half-way about such explanations.They are as rock-bottom as those in terms of individuals. 13In my opinion, Popper’s main contribution to social science methodology washis institutionalism and his situational logic, which may be seen as an earlyattempt to launch rational choice as a general approach in the social sciences.The suggested combination of institutionalism and rational choice, has, indeed,developed into one of the most vigorous research programmes in contemporarysocial science (see chapters 9 and 10).

8 Economics: the individualistscienceThe ever-present intellectual puzzle that confronts economic study is how toexplain the social result in terms of individual actions.(Kuznets, 1963: 52)It is usually thought that mainstream economics is the purest exemplar ofmethodological individualism.(Arrow, 1994: 2)Economics is, and always has been, considered the most individualistic of thesocial sciences (Hausman, 1992: 97f; Bicchieri, 1993: 9). 1 It would not becorrect, however, to maintain that all economics is individualistic. More precisely,the claim should be that orthodox, or mainstream, economics is individualistic(Himmelweit, 1977: 22ff). In addition, there are the remnants of some heterodoxresearch programmes, such as Marxism and institutionalist economics, which areclearly and explicitly holistic – except for the recently emerging ‘analyticalMarxism’, which is a rational choice theory based on methodological individualism(see pp. 309–18). There is also a waning ‘neo-Ricardian school ofeconomics’, which is presumably not individualistic. 2 Finally, there is macroeconomics,which is hard to classify along the individualist-holist axis. Theprevailing opinion, however, is that macroeconomics must be provided with solidmicrofoundations in order to qualify as a strictly individualist theory (Janssen,1993). More exactly, then, the claim should be that mainstream microeconomicsis individualistic, or believed to be so (Nelson, 1986; 1989).According to Lawrence A. Boland (1982: 13ff), individualism is one of tworelated, but autonomous, methodological rules on the hidden agenda of neoclassicaleconomics – the other is inductivism. But neoclassical economics is not justindividualistic, it is guided by the rule of psychologistic individualism. This rulesays that an explanation is not to be considered successful until all exogenousvariables of a comparative statics model have been reduced to psychologicalstates and natural constraints. Non-individualist factors, e.g., institutions, are onlyallowed among the endogenous, i.e., explained, variables in the model.Institutional constraints may be considered temporarily fixed, but not exogenous(see also Boland, 1986: 9–11).

Economics: the individualist science 229By saying that neoclassical economics is based on psychologistic individualism,we are saying specifically that neoclassical theories or analyses mustpermit only two types of exogenous variables: natural constraints andpsychological states. Of particular concern is the psychologistic individualistrequirement that no social institution that appears in our explanations mustbe allowed to play the role of an exogenous given.(Boland, 1982: 49)In a later work, Boland (1992: 24f) makes the further distinction betweenthose exogenous variables, which are included in the model and those exogenouslygiven conditions which do, or may, influence the endogenous variables,but which are treated as fixed and included in the ceteris paribus. The differencebetween exogenous variables and exogenous conditions is of course relative to themodel used by the economist, or to the particular specification of that model.Exogenous conditions may be further divided into those, which are considered‘exogenously fixed by a prior maximization process’, e.g., institutions, and thosewhich are exogenously given, e.g., natural constraints. According to the pychologisticindividualism of neoclassical economics, only psychic states of individualsare permitted as exogenous variables. Social institutions may appear as exogenouslyfixed conditions, but it is part of the neoclassical programme to turn theminto endogenous variables in a more inclusive neoclassical model. 3Methodological individualism is closely associated with equilibrium analysisand is most clearly expressed in the theory of general equilibrium (see below),but Boland’s explication is in terms of the comparative static methodology ofMarshall, where the distinction between endogenous and exogenous variables ismost clearly visible. What is static, and treated as exogenously fixed, in the shortrun is variable in the long run. If, for instance, social institutions have to betreated as exogenously fixed in the short run, they may, nevertheless, be seen asendogenous in a long-run general equilibrium. Methodological individualism issaved (Boland, 1986: 43ff).I believe that Boland’s psychologistic individualism provides a good explicationof the individualist aspirations, if not always of the practice, of neoclassicaleconomists. 4 As he suggests, himself: ‘While almost all neoclassical theoristsopenly accept the requirements of methodological individualism, few seem fullyaware of the complexity of the requirement. Moreover, they are unaware of theconflicts between their tools of analysis and their commitment to methodologicalindividualism’ (Boland, 1986: 10). As we shall see, this quote from Boland is afair statement about economists. The recent development of economic theorygives us reason to believe that Boland is correct also to maintain that this versionof methodological individualism is too strong.But, according to Boland, there is another, weak version of methodologicalindividualism, called ‘institutional individualism’ (Boland, 1982: 32–7). As wehave seen in chapter 7, the roots of this version are in the methodology of KarlPopper, but it was his pupil Joseph Agassi, who stated it most clearly. LawrenceBoland, who was once a student of Joseph Agassi, picked it up and brought it to

230 Economics: the individualist sciencebear on economics. The simple idea of institutional individualism is to permitsocial institutions among the exogenous variables of economic models and toaccept them as exogenous variables and conditions, without insisting upon theirendogenisation in some more comprehensive model. It may be argued that thiswas the position really taken by Marshall, but it seems to have few adherentsamong contemporary economists, who tend to prefer psychologistic individualism.I have discussed economic theory until the early 1930s in earlier chapters. Inthis chapter, I am going to discuss modern microeconomics, which, for mypurposes, is economic theory from about 1935 to the present day. But evenwithin modern microeconomics there is a division between two approaches: (1)The neoclassical approach going back to Leon Walras and Alfred Marshall. Thisapproach focuses on economic equilibrium and favours a mathematicalfunctionalmethod. (2) The Austrian approach going back to Carl Menger,Eugen von Böhm-Bawerk and Friedrich von Wieser. This approach denies thatthe market is ever in equilibrium and emphasises instead its character ofcompetitive process. It uses a ‘causal-genetic’ method. Of these approaches theneoclassical is by far the most influential today, but it is above all among Austrianeconomists that one finds the defenders of methodological individualism. Bothapproaches are individualistic, but the Austrian approach is generally considered‘more’ individualistic. It always focuses on the individual decision-maker, especiallythe entrepreneur, in its analysis of economic life. In addition, there aresome more or less distinct schools, such as the Chicago School and the VirginiaSchool, also called Public Choice, which are both individualistic. I have alreadydiscussed Austrian Economics at considerable length in chapter 3, and I amgoing to discuss the Chicago School and Public Choice in the next chapter.Macroeconomics and microfoundationsIn 1936 John Maynard Keynes published The General Theory of Employment, Interestand Money. The result is sometimes described as the ‘Keynesian revolution’(Samuelson, 1946: 187; Klein, 1947). It was the beginning of macroeconomics.Keynes’s book belongs to that category of famous books, which all talk about,but few have actually read. One reason for this is, no doubt, that it is ‘an exceedinglydifficult book’ (Hicks, 1937: 147). 5 I do not claim to fully understandKeynes’s General Theory – if such a thing is possible – but I hope that my briefpresentation is not too wide off the mark.Keynes set out to criticise what he called ‘the classical theory’, which includedboth classical and neoclassical economics, but especially the theory – as distinguishedfrom the methodology – of Alfred Marshall. 6 His own theory waslaunched as a ‘general’ theory, including the classical theory as a special case,‘only realised when the propensity to consume and the inducement to investstand in a particular relationship to one another’ (Keynes [1936] 1973: 28). Themain point, made by Keynes against the classical theory was that it wronglyassumes that full employment is the normal state of affairs, while voluntary

Economics: the individualist science 231unemployment is abnormal and bound to disappear in the long run. Accordingto Keynes, an economy with substantial involuntary unemployment may verywell be in a state of equilibrium and as to the long run, we are all familiar withKeynes’s most famous expression: ‘in the long run we are all dead’. This impliesa denial of Say’s law, according to which supply creates its own demand.It is a common belief that Keynes’s explanation of involuntary unemploymentwas in terms of sticky wages (see Leijonhufvud, 1969: 14f). Forpsychological and institutional reasons, it is difficult to lower money wages. The(classical) way to achieve full employment, therefore, is largely blocked and theresult is unemployment. This was not Keynes’s main point, however. The maindifficulty with the classical view is that it does not reckon with individuals’ incentiveto liquidity, which interferes with their propensity to consume (Keynes, 1937:234f). If people prefer to save their money, instead of spending it on consumptiongoods, there will be a decrease in effective demand, with harmful effects oninvestment and, ultimately, on employment. It is not that individuals actuallyhoard money, but that their inclination to do so affects the rate of interest.Keynes gives several summary statements of his theory, but the most succinctcan be found in his article ‘The General Theory of Employment’ (1937), whichis a reply to some critics:The theory can be summed up by saying that, given the psychology of thepublic, the level of output and employment as a whole depends upon theamount of investment. I put it in this way, not because this is the only factoron which aggregate output depends, but because it is usual in a complexsystem to regard as the causa causans that factor which is most prone tosudden and wide fluctuation. More comprehensively, aggregate outputdepends on the propensity to hoard, on the policy of the monetaryauthority as it affects the quantity of money, on the state of confidenceconcerning the prospective yield of capital assets, on the propensity to spendand on the social factors which influence the level of the money-wage. Butof these several factors it is those which determine the rate of investmentwhich are most unreliable, since it is they which are influenced by our viewsof the future about which we know so little.(Keynes, 1937: 221)Two elements of his system were picked out by Keynes himself (1937: 222f)as most heterodox. (1) He was among the first to insist upon the importance ofindividuals’ expectations about the future for their decisions to consume, save, orinvest. (2) He was also first to highlight the psychological law behind the so-called‘multiplier’; the fact that individuals do not spend all of their income on eitherconsumption goods, or capital goods, but may have a preference for liquidity.The multiplier is the ratio between increases in income and investment and it isimportant because it tells us something about the total effect on economicactivity of an investment. A multiplier close to unity will make the wheels goround.

232 Economics: the individualist scienceAnother, more systematic, mode of summary presentation can be found inthe original version of The General Theory, where Keynes divides the elements ofthe economic system in those ‘we usually take as given’, those ‘which are theindependent variables’ and those ‘which are the dependent variables’ (Keynes[1936] 1973: 245). This Marshallian way of presenting his theory suits mypurposes well, since it casts some light upon the issue of methodological individualism.Figure 8.1 The elements of the economic system according to KeynesSource: Keynes ([1936] 1973: 245)In Figure 8.1, it appears that Keynes was not committed to methodologicalindividualism. The given elements of the economic system include many institutionaland social structural elements and Keynes makes it perfectly clear thatthey do influence the independent variables, even though relegated to the ceterisparibus. Of the independent variables, the first is psychological, the second technicaland the third institutional. It might be added that the institution of money,while not mentioned explicitly by Keynes, plays an important role in his argument.We may agree with Klein (1947: 103) that, on the view of Keynes, ‘thecausal forces are found outside the price system in the psychology, expectations,

Economics: the individualist science 233habits and institutions of the population’. It would seem, therefore, that Keynes’stheory fails to qualify as an exemplar of strict methodological individualism. It isstill possible that it passes as an example of institutional individualism. There issome evidence, however, to suggest that Keynes, himself, saw his own theory aseven more holistic. In the preface to the French edition (1939) of The GeneralTheory of Employment, Interest and Money, Keynes explains why he calls it a ‘general’theory.I mean by this that I am chiefly concerned with the behavior of theeconomic system as a whole – with aggregate incomes, aggregate profits,aggregate output, aggregate employment, aggregate investment, aggregatesaving rather than with the incomes, profits, output, employment, investmentand saving of particular industries, firms or individuals. And I arguethat important mistakes have been made through extending to the system asa whole conclusions which have been correctly arrived at in respect of a partof it taken in isolation.(Keynes [1936] 1973: xxxii)This is a somewhat idiosyncratic use of the term ‘general’, with respect to scientifictheories. On the other hand, it is a perfectly adequate characterisation of a‘macrotheory’, in the most usual sense of that term. But Keynes seems to go onestep further and suggest that his macrotheory is also a holistic theory, or a theoryabout the system as a whole. This interpretation is supported by his critique oftheories that draw conclusions about the system as a whole from knowledgeabout the parts taken in isolation. 7Keynes’s holism is not absolute, however, and did not prevent him frombasing some of his arguments about the behaviour of the economy as a whole,upon assumptions about the psychology of individuals. It may be argued, therefore,that Keynes did provide ‘microfoundations’ for his macroeconomics. Twothings should be noticed about the role of these macrofoundations, however: (1)Keynes’s General Theory depends upon these psychological assumptions onlypartly and loosely, not wholly and precisely. There is no exact logical relationbetween them, such that the former is deducible from the latter. While it ispossible to argue, then, that Keynes provided ‘microfoundations’, in a sense, hewas far from achieving a microreduction, and even denied that it is possible to doso. This, I believe, follows logically from his holism. Thus, it is wrong to maintain,as does J.W.N. Watkins (1952a: 33–6; 1953: 734–7), that Keynes was amethodological individualist. At least he was not a methodological individualistin the strong sense suggested by Watkins, himself. (2) Keynes’s ‘microfoundations’was not neoclassical. On the contrary, Keynes derided the neoclassical view ofeconomic man and assumed more ‘animal spirits’ instead. Keynes’s basicpsychological assumption concerns the propensity to consume, which in its turnis subject to a preference for liquidity, or a propensity to hoard (Keynes [1936]1973: 89ff, 194ff).Keynes was probably more interested in practical than theoretical matters.

234 Economics: the individualist scienceThe main purpose of his General Theory was to find a cure for economic depressions.We do not know today exactly what Keynes, himself, suggested, and whatwas added by his followers, the ‘Keynesians’, but it is certain that he saw animportant role for government, in stimulating consumption and investment. ‘Thecentral controls necessary to ensure full employment will, of course, involve alarge extension of the traditional functions of government’ (Keynes [1936] 1973:379).Keynes’s theory made a revolution, but it also met with much resistance andcritique. Because of the policy implications and political use of Keynes’s ideas,the critique has usually been a mixture of theoretical, methodological and politicalelements. It is not by accident that you find, among the most energeticcritics, economists of a libertarian persuasion; members of the Austrian,Chicago and Virginia School of Economics. 8I am most interested in the methodological critique, because it is to a largeextent a manifestation of methodological individualism. The main gist of thiscritique is that theories about aggregates – whether empirical or theoretical – areincomplete and in need of individualist microfoundations. This type of critiqueis older than that directed at Keynes. It goes back, at least, to Carl Menger, whocriticised Gustav von Schmoller and the other members of the ‘Younger’German Historical School, for attempting to turn their (collectivistic) historicalstatisticalapproach into the only method of economics, when, in fact, it is onlyone of several, equally legitimate, approaches (Menger [1871] 1976: 35ff; 1884;1889). Of these, however, the individualistic theoretical approach is Menger’spreferred alternative.Menger’s view of the use of statistics was adopted and developed by Friedrichvon Hayek (1931: 3ff; 1933a: ch. 1), who used it first against the theory of businesscycles and later against Keynesian macroeconomics and econometrics.Hayek saw all theories of aggregates and averages as essentially incomplete andin need of individualistic underpinning. There can, for instance, be no ‘directcausal connections between the total quantity of money, the general level of allprices and, perhaps, also the total amount of production’ (Keynes [1936] 1973:245).For none of these magnitudes as such ever exerts an influence on the decisionsof individuals; yet it is on the assumption of a knowledge of thedecisions of individuals that the main propositions of non-monetaryeconomic theory are based. It is to this ‘individualistic’ method that we owewhatever understanding of economic phenomena we possess; that themodern ‘subjective’ theory has advanced beyond the classical school in itsconsistent use is probably its main advantage over their teaching.If, therefore, monetary theory still attempts to establish causal relationsbetween aggregates or general averages, this means that monetary theorylags behind the development of economics in general. In fact neither aggregatesnor averages do act upon one another, and it will never be possible to

Economics: the individualist science 235establish necessary connections of cause and effect between them as we canbetween individual phenomena, individual prices, etc.(Hayek, 1931: 4f)This was written before the publication of The General Theory, but it hits thelatter as well and Hayek repeated the argument with explicit address to Keynes.In retrospect, it seemed to Hayek that the chief significance, and fault, of TheGeneral Theory was that it contributed to the rise of macroeconomics and econometricsat the expense of microeconomics. 9A view, closely akin to that of Hayek, was expressed by Tjalling C. Koopmansin his critique of a book by two institutional economists: Arthur F. Burns andWesley C. Mitchell. The title of the book was Measuring Business Cycles (1946) andKoopman’s main point was that it consisted of ‘Measurement without Theory’(1947). With an analogy, much used by economists, Koopmans maintained thatBurns and Mitchell represent the ‘Kepler stage’, but fail to reach the ‘Newtonstage’, of economics. I do not know how well Koopmans’s arguments hit thetarget, but he was certainly correct to maintain theory is needed even for thechoice of relevant measures. More interesting from my point of view, however, isthe fact that Koopmans is certain that the needed theory must be of a certainkind. By ‘economic theory’, Koopmans means, ‘in this context the theoreticalanalysis of the aggregate effects of assumed patterns of economic behavior ofgroups and individuals’ (p. 164). It is clearly implied, by Koopmans, that the‘Newton stage’ of economics is reached when aggregate phenomena areexplained ‘with reference to the underlying economic behavior of individuals’ (p.165).Koopmans’s article gave rise to a controversy between Rutledge Vining andhimself, which moved the focus of attention from empirical research versustheory to holism versus individualism. For some social scientists, these controversiesare coextensive, perhaps even identical, but for most defenders of holism insocial science, including Keynes and Vining, they concern different issues. Thus,for Vining (1949a), it is not necessary that the dynamics of the economic systembe explained in terms of individuals.I believe that in our discussions of trade fluctuations, national and international,we deal with the behavior of an entity that is not a simple aggregateof the economizing units of traditional theoretical economics. I think thatwe need not take for granted that the behavior and functioning of this entitycan be exhaustively explained in terms of the motivated behavior of individualswho are particles within the whole. It is conceivable – and it wouldhardly be doubted in other fields of study – that the aggregate has an existenceapart from its constituent particles and behavior characteristics of theparticles.(Vining, 1949a: 79)

236 Economics: the individualist scienceI believe that much of the statistical regularities that are to be observed inpopulation phenomena involves the behavior of social organisms that aredistinctly more than simple algebraic aggregates of consciously economizingindividuals. I think that in a positive sense the aggregate has an existenceover and above the existence of Koopmans’s individual units and behaviorcharacteristics that may not be deducible from the behavior of these componentparts.(Vining, 1949a: 80f)Vining expressed himself in a way open to obvious objection and Koopmans wasquick to seize the opportunity of mounting a counter-attack:I cannot understand the meaning of the phrase ‘the aggregate has an existenceapart from its constituent particles and behavior characteristics of itsown not deducible from the behavior characteristics of the particles’. If atheory formulates precisely (although possibly in probability terms) thedetermination of the choices and actions of each individual in a group orpopulation, in response to the choices and actions of other individuals or theconsequences thereof (such as prices, quantities, states of expectation), thenthe set of these individual behavior characteristics is logically equivalent tothe behavior characteristics of the group. Such a theory does not have anopening wedge for essentially new group characteristics. Any deus exmachina who should wish to influence the outcome can only do so byaffecting the behavior of individuals.(Koopmans, 1949: 87)Koopmans goes on to affirm the existence of essentially social phenomena,such as fads, fashions, power struggles and price wars, but maintains ‘that suchsocial phenomena are necessarily acted out by individuals as members ofgroups’ (p. 87). This argument, alone, does not establish the necessity or, even,the desirability of building a theory of group behaviour on the foundation ofindividual behaviour, but Koopmans (p. 87) provides three other arguments tothis effect. The third argument is motivated by policy considerations and irrelevantfor my purposes. Two arguments remain: (1) There is no possibility ofestablishing the relevant aggregate behavior separately from the behaviour ofindividuals. This is in distinction from physics, where Boyle’s law was establishedfor the aggregate behaviour of molecules prior to its derivation from the kinetictheory of gases. (2) Unlike physicists, who know atoms and molecules only indirectly,economists have direct knowledge of individual human beings, themselvesincluded.In a footnote Koopmans makes an interesting point which is typical formethodological individualists, especially the Marshallians. Institutions may beaccepted as exogenously given in a particular analysis, but these institutions may,in turn, be explained in terms of rational choice.

Economics: the individualist science 237It is true that the choices of individuals are restrained by a framework ofinstitutional rules enforced or adhered to by the government, the bankingsystem and other institutions. These rules … can to some extent be taken asgiven for the analysis of economic fluctuations. In a deeper analysis, theserules and the changes in them would need to be explained further fromchoices by individuals interacting, in various degrees of association witheach other, through political processes.(Koopmans, 1949: 87)As we will see in the next chapter, what Koopmans refers to as a ‘deeper analysis’is what the new institutionalism in economics, and public choice in particular,has attempted over the last fifty years.The exchange between Vining and Koopmans is revealing in several ways:first, it includes some standard formulations of holism and individualism insocial science. Second, it uses some standard arguments in the defence of eachrespective doctrine. Third, it is typical of the debate between holists and individualistsin that the protagonists tend to express themselves in certain standardidioms and then to engage in unproductive and, often, deliberate misunderstandingsof each others’ views on the matter.Thus, it is standard among social holists to maintain that social wholes havean existence apart from, or over and above, individual human beings. It is no lesscommon for individualists to deny that there is anything over and above individuals.While holists insist that there are other units of analysis than individualhuman beings, such families, firms and economic systems, individualists retortthat our knowledge of such units can only be derived from knowledge aboutindividuals.Comparing the above statements of Vining and Koopmans, there is littledoubt that Koopmans is most in the right. Thus, Koopmans is no doubt correctto insist that group behaviour is equivalent to the behaviour of individuals ingroups, but this is probably not what Vining wanted to deny. In his finalrejoinder to Koopmans, Vining dismisses his first argument as being little morethan a tautology. 10No issue would be involved over the triviality that in moving a house onemust move the constituent bricks; but no one would contend that from theproperties of individual bricks one could infer all the properties of the housebeing constructed. There is more to the developing house than can belearned from the bricks. Whether or not this is the case with regard tohuman beings and the structural and functional characteristics of evolvingsocietal forms is not a matter of logic, but rather a matter of fact.(1949b: p. 92)The heart of the matter, I believe, is the so-called ‘aggregation problem’ (cf.Hartley, 1997: 132–46; 182ff). While Koopmans admits that ‘the mathematicalderivation of aggregate equations from individual behavior equations’ is difficult

238 Economics: the individualist scienceand maybe impossible with present mathematical tools, he does not seem tobelieve that there are any principal problems involved (Koopmans 1949: 87).Vining, on the other hand, maintains that social wholes have structural and functionalproperties that may make the obstacles to aggregation insurmountable or,rather, that there is more to macroanalysis than aggregation (Vining 1949a:79ff). 11Hayek’s and Koopmans’s critique of macroeconomics and econometrics,then, concerns the so-called aggregation problem, or how to derive relations betweenaggregates from the (rational) behaviour of individuals or, at least, to establishsome kind of logical bridge between them. This problem beset Keynesianmacroeconomics from the very beginning and it is, of course, a manifestation ofthe individualistic quest for microfoundations (see Klein, 1946a: 93; Arrow[1951] 1968: 641f). 12The Keynesian theory has been severally attacked by some economistsbecause it is couched in terms of aggregate concepts like total consumption,employment, income, etc. These aggregative concepts, it is argued, get awayfrom the more fundamental economic concepts of the individual and arethus misleading. But the aggregative or macro approach is not only labeledas misleading; it is also labeled as incorrect. For example, economists askhow can there be a stable relationship between total community consumptionand total community income unless the distribution of income withinthe community is taken into account?(Klein, 1947: 56f)For most economists, it was a matter of course that Keynes’s general theoryhad to be provided with neoclassical microfoundations. Indeed, ‘[t]he fact thatone cannot understand macroeconomics without providing microfoundationsthus seems to have become one of the unspoken assumptions held byeconomists’ (Hartley, 1997: 124) and in most cases the choice of foundation wasWalrasian (Solow, 1989: 29). The result was a ‘neoclassical’, or ‘neo-Walrasian’synthesis, represented by economists such as John R. Hicks, Oscar Lange,Lawrence R. Klein, Paul A. Samuelson and Don Patinkin (Weintraub, 1979: ch.4; Feiwel, 1985: 7–16). The origin of this development was an article by Hicks,where he emphatically denied that there is a big difference between Keynes’system and the ‘classical’ theory. According to Hicks, whatever Keynes theory is,‘it is not the General Theory. We may call it if we like, Mr. Keynes’s special theory. TheGeneral Theory is something appreciably more orthodox’ (Hicks, 1937: 152). Itwas Hicks, who introduced the well-known IS-LM diagram, for purposes ofcomparison between ‘classical’ theory and Keynes’s theory. 13 According toHicks, Keynes’s ‘general theory’ is but a special case of a more general theory,which might be represented by two curves: IS (representing the relation betweenIncome and interest, which must be maintained in order to make saving equal toinvestment) and LM (representing the relation between Income and interestcorresponding to a given amount of money). The economic system is in equilib-

Economics: the individualist science 239rium where the two curves intersect (that is, where the investment rate equals themoney rate).Hicks adhered to Keynes’s view that liquidity preference is the main cause ofunemployment. This view was soon to be challenged by Franco Modigliani inanother famous article (Modigliani, 1944). Modigliani followed Hicks in usingthe IS-LM apparatus in his restatement of Keynes’s theory. He also agreed withHicks, that Keynes’s theory is not the General Theory he claimed himself, but aspecial case, called by Modigliani (p. 56), the ‘Keynesian case’. When it comes toan explanation of this case, however, Modigliani takes exception to Hicks’s analysis.The main cause of underemployment equilibrium is rigid wages, notliquidity preference (pp. 65ff). The liquidity-preference theory holds only in alimiting case of the more general theory of rigid wages (p. 76). Modigliani,himself, did not ascribe this general theory to Keynes, but for some reason it hasturned into a common belief that Keynes explained unemployment equilibriumin terms of rigid, or sticky, wages.By the end of the 1960s some economists began to question the mainstreamsynthesis of neoclassical economics and Keynes’s general theory on the groundthat it was based on a misrepresentation of the latter. Robert Clower (1965)described the mainstream synthesis as ‘the Keynesian counterrevolution’ andAxel Leijonhufvud (1967; 1968) argued that it is necessary to make a distinctionbetween ‘Keynesian economics’ and the ‘economics of Keynes’. A carefulreading of Keynes, himself, made it clear to Leijonhufvud that ‘the propositions,prescriptions, and opinions frequently advanced as “Keynesian” bear little relationto Keynes’s views’ (Leijonhufvud, 1969: 9). 14An important issue was whether Keynes’s theory is a special case of neoclassicaleconomics, as Hicks and Modigliani maintained, or if (neo)classical theoryis a special case of Keynes’s general theory, as he suggested himself, and asClower and Leijonhufvud maintained. I tend to believe that neither view isentirely correct. At least, I do not believe that either theory can be seen as aspecial case of the other in the same sense as Newtonian mechanics is said to bea special case of Einstein’s theory of relativity. For this to be the case, it must bepossible to deduce one theory from the other, within certain boundary conditions.But the prevailing opinion, today, is that the neoclassical synthesis hasfailed to achieve the logical integration necessary to build Keynes’s macroeconomicson the foundation of neoclassical microeconomics. 15Despite the critique by Clower and Leijonhufvud, the neoclassical synthesiswas a great scientific and political success until the 1970s, when it suddenlyceased to fit the facts and to work properly as an aid to policy. Two rivalsappeared on the scene, promising to provide a remedy: the monetarist ‘Counter-Revolution’ of Milton Friedman (1970a) and the new classical ‘RationalExpectations Revolution’ of Robert E. Lucas and Thomas J. Sargent (Miller,1994). 16 The main message of both revolutions was that the neoclassicalsynthesis is not neoclassical enough and, in fact, fails to provide adequate microfoundations.Instead of providing microfoundations for Keynesianmacroeconomics, they wanted to drop Keynes altogether and build a new

240 Economics: the individualist sciencemacroeconomics directly on neoclassical foundations. Both monetarism and the‘new classical economics’ are based on the neoclassical assumptions ofmaximising behaviour and equilibrium. The main contention of Friedman isthat equilibrium will always be restored in the long run (Friedman, 1970b:223ff). 17 This is the first article of faith of Chicago economics. Lucas andSargent agree that equilibrium is the telos of the economic system, but theirmain contribution was on the behavioural side, where the assumption of rationalbehaviour is turned into the assumption of rational expectations. In the framework,first suggested by Lucas (1972: 103), ‘all prices are market clearing, allagents behave optimally in light of their objectives and expectations, and expectationsare formed optimally’ (see also Lucas, 1975: 1113).The idea of rational expectations was first introduced by John F. Muth, whomaintained that the expectations of economic agents are based on the sameinformation as those of economists and are, therefore, essentially the same as thepredictions of economic theory (Muth, 1961: 316). Or, in the version of Lucasand Sargent ([1979] 1981: 307): ‘agents are assumed to make the best possibleuse of the limited information they have and to know the pertinent objectiveprobability distributions’. The adoption of this assumption has important policyimplications. Economic agents are supposed to take government policy intoaccount when forming their own expectations about the future and act accordingly.Much that government decides, therefore, turns out to be futile, becauseundone, or counteracted, by the adaptive responses of economic agents, at leastif government policy is based on Keynesian macroeconomics. But not everything!Some measures, on the part of government, are unanticipated, and thesemeasures are the true causes of business cycles. ‘All aggregate output movementsin the models studied above result from movements in a single monetary-fiscalshock to aggregate demand’ (Lucas, 1975: 1139).The most serious fault with Keynesian econometric models, according toLucas and Sargent, is that they base predictions on the future behaviour ofeconomic agents on their past behaviour, even when government interventionleads to a change in the rules of the game (Lucas, 1976; Sargent, 1980). Theeconomy is treated like a hydraulic machine, which can be manipulated bygovernment, without any strategic behaviour at all on the part of the individualscomprising the economic system. On a more principal level, then, the fault withKeynesian macroeconomics is that it lacks microfoundations, in the form of thepostulate that agents act rationally in their own interest (Lucas and Sargent,1979 [1981]: 299–302).A problem with the new classical economics is that, on closer inspection, itlooks suspiciously like the worst kind of holism. It has been common inmacroeconomics to make use of so-called representative agents – first introducedby Marshall – and the new classical economics is no exception. Put simply,a representative agent is a representative of the economy as a whole, or somemarket. The aggregate economy is treated as if it were an individual agent. Notonly is this procedure, dubious in itself, it is flagrant violation of methodologicalindividualism. And yet, it is often used as a way of providing microfoundations

Economics: the individualist science 241for macroeconomics. As such it is, of course, pure humbug (cf, Kirman, 1992;Hartley, 1997; Kinkaid, 1997: 93f). To treat the economy as a whole, as if it werean individual is no less holistic, and no less problematic, than to invoke nationalinterests, folk souls, collective consciousness, social needs and the like.Some twenty years after the revolution of Lucas and Sargent, new classicaleconomics is in the same position as Keynesianism was in the 1970s. Few seemto believe in it any more. There even seems to be, at the present moment, areturn to more Keynesian ideas in macroeconomics. Be that as it may, it hasbeen suggested by leading macroeconomists that what economics needs mostof all is macrofoundations for microeconomics, rather than the reverse (Solow,1989: 32; Colander, 1993). As we shall see in the next section, Solow is notalone.A number of conclusions can be drawn from this treatment of macroeconomics:1 Keynes’s macroeconomics does not lack microfoundations, but according toorthodox neoclassical economists, Keynes’s animal spirits are the wrongmicrofoundations. They appear ad hoc, incomplete, hard to formalise and,therefore, ill suited for the task of deriving macroeconomics from the microfoundation.Keynes macroeconomics does not follow deductively from hispsychological assumptions. The latter only support the former to some degree.2 Even if Keynes’s macroeconomics has its own microfoundations, this theoryis not an example of strong methodological individualism. In addition topsychological assumptions, it relies on assumptions of a more institutional,even holistic nature. It may still be the case that Keynesian macroeconomicsqualifies as an example of institutional individualism.3 The problem of providing Keynesian macroeconomics with Walrasian or,more generally, neoclassical microfoundations has not been solved andcannot be solved, because they include incompatible theoretical elements.They only way to achieve a reconciliation would be by modifying one, orboth, theories. As a matter of fact, most Keynesians did modify Keynes, soas to make it compatible with neoclassical microeconomics.Demonstrating the dependence of all macroeconomics on microeconomicprinciples is essential for the fulfilment of the (methodological) individualistrequirements of neoclassical economics. However – and this is not widelypointed out – this ‘necessity’ presumes that microeconomic theory, in theform of general equilibrium theory, is a successful individualist programme.(Boland, 1982: 80)General equilibrium theoryThe theory of general equilibrium is at the heart of theoretical economics andconsidered by many as its crowning achievement. The first to use the term ‘equilibrium’in economics seems to have been the Scottish social philosopher James

242 Economics: the individualist scienceSteuart in 1769, but the origin of the idea is better traced to his more wellknownfellow-countryman and contemporary Adam Smith (Arrow, 1968: 376;Milgate, 1989). 18 In particular, it is the notion of a ‘natural price’ that foreshadowsthe modern idea of economic equilibrium. More generally, however, itis the famous invisible hand, which suggests the existence of an equilibrium in anindividualistic, or free market, economy (see Arrow and Hahn, 1971: vi–vii, 1f).It was Adam Smith, who first realized the need to explain why this kind ofsocial arrangement does not lead to chaos. Millions of greedy, self-seekingindividuals, in pursuit of their own ends and mainly uncontrolled in thesepursuits by the state, seem to ‘common sense’ a sure recipe for anarchy.Smith not only posed an obviously important question, but also started usoff on the road to answering it. General Equilibrium Theory as classicallystated by Arrow and Debreu (1954) is near the end of that road.(Hahn, 1981: 123)In the first place, then, the economic theory of general equilibrium is a theoryof social order. Arrow (1968: 376) mentions two aspects of the notion of generalequilibrium, as used in economics: (1) determinateness – ‘that is, the relationsthat describe the economic system must form a system sufficiently complete todetermine the values of the variables’ – and (2) balance of forces. In the impressivehistory of the theory of general equilibrium, Bruno Ingrao and GiorgioIsrael (1990), the authors claim (p. 3) that the ‘invariant paradigmatic nucleus’,or ‘core’ of this theory can be identified as ‘the aim to demonstrate the existence,the uniqueness, and the global stability of the equilibrium’ (see also Arrow, 1968).In the second place, Adam Smith also seemed to suggest that this order is asocial optimum or, at least, highly beneficial to the members of society. This is amore controversial statement, whether attributed to Smith, or not. In the thirdplace, the theory of general equilibrium, as first conceived by Adam Smith, isusually considered ‘individualistic’, in both a methodological and a politicalsense. This is also a matter of controversy. Paul Samuelson, for instance, seems todeny that Adam Smith’s invisible hand was entirely individualistic.In summary: these individualistic atoms … are not isolated from otheratoms. Adam Smith, who is almost as well known for his discussion of thedivision of labour and the resulting efficiency purchased at the price ofinterdependence, was well aware of that. What he would have stressed wasthat the contacts between the atoms were organized by the use of marketsand prices.(Samuelson, 1966: 1411)The origin of the idea of economic equilibrium, then, lies in classicaleconomics. A more explicit version of the theory of economic equilibrium didnot appear, however, until the marginalist revolution and the development ofneoclassical economics. There are two forms of this theory: (1) the theory of

Economics: the individualist science 243general equilibrium developed by Leon Walras and (2) the theory of partial equilibriumadvanced by Alfred Marshall. In the former all economic agents and,therefore, all markets are in equilibrium. In the latter, equilibrium betweensupply and demand is limited to a particular branch, or market.There is also a difference of methodology between Walras and Marshall. Theformer starts from a model of the economy, made up of extremely simpleelements, and goes on to construct successively more complex and realisticmodels of the economy. I assume that this methodology is in the Cartesian tradition,as is that of Austrian Economics. Marshall, on the other hand, took hispoint of departure in the real economy and, then, went on to construct modelsof increasing abstraction, using the famous ceteris paribus; the assumption thatthings, not in the model, remain constant. This leaning on empirical reality mayreflect a certain influence of the German Historical School on Marshall’smethodology. In the terminology of Hayek (see p. 117), we might say thatWalras’s method is ‘synthetic’, while Marshall’s method is ‘analytic’.When Joseph Schumpeter coined the term ‘methodological individualism’,and used it to designate the actual procedure of theoretical economics, heintended, first of all, the theoretical economics of Leon Walras. This may seem abit odd, since, of the three makers of the marginalist revolution, Walras was, byfar, the least interested in the human individual. If Walras’s theory is ‘individualistic’– as it is – this is largely by implication, and it seems to have escaped hisown notice. Walras’s theory of general equilibrium is presented as a huge systemof mathematical equations, representing the working of a lifeless mechanism.Human beings can be fathomed as the deus ex machinae, who make the wheelsturn around, but they do not enter the scene. Walras divided economics in topure economics, which is a science, applied economics, which is an art, andsocial economics, which is an ethical theory (Walras [1874] 1984: 58ff). Pureeconomics is the theory of prices and of the allocation of scarce resources, orsocial wealth. Applied economics is a theory of industry and the production ofwealth. Social economics, finally, is the theory of institutions and of the distributionof wealth. According to Walras, only social economics is a moral science,based on the assumption that individual human beings are persons with a freewill and self-consciousness. His own theory of pure economics, however, heconceived of as a natural science of things. Since buyers on the market are pricetakers, ‘any value in exchange, once established, partakes of the character of anatural phenomenon, natural in its origins, natural in its manifestations andnatural in essence’ (p. 69).Walras’s pure economics, then, is a theory based on an extremely thin andatomistic conception of the human individual, but this does not imply thatWalras was a programmatic methodological individualist. Unlike pureeconomics, social economics is a theory of institutions (Walras [1874] 1984: 63)and there is nothing to indicate that Walras conceived of it as an individualistictheory.In this section my concern is with the theory of general equilibrium as developedby Walras and, more recently, by economists such as Kenneth Arrow,

244 Economics: the individualist scienceGerard Debreu, Lionel W. McKenzie and Frank H. Hahn. I hope that noserious injustice is done to economic theory, if I concentrate on theArrow–Debreu model of general equilibrium, and if I rely on Arrow as my mainguide to the subject. As suggested by John Geanakoplos (1989: 57), ‘TheArrow–Debreu model of general equilibrium is relentlessly neoclassical; in fact ithas become the paradigm of the neoclassical approach’. He goes on to suggestthat ‘[t]his stems in part from its individualistic hypothesis’ (p. 57).As we shall see later, the alleged individualism of general equilibrium theoryis somewhat problematic. Even Kenneth Arrow, himself, has come to expressdoubts about the purity of its individualism. At the beginning of his career,however, he had no such doubts and suggested a general (mathematical) modelfor the social sciences, which he described as the ‘individualistic viewpoint’. Ihave suggested in chapter 2 that Adam Smith’s invisible hand of the market isone paradigm of methodological individualism. At least it was turned into oneby the marginalist economists; most explicitly by Carl Menger, but implicitly alsoby Leon Walras. Arrow’s model for the social sciences is the clearest statement ofthis paradigm I have seen. Because of this, I will quote at some length:In most mathematical and, generally, in most deductive studies in the socialsciences, the starting point is the behavior of the individual. Each individualis conceived of as acting in the way determined partly by his psychology andhis physical surroundings and partly by the actions of others. If there are nindividuals, we may denote the actions of individuals i by A i , and the nonsocialdeterminants of his behavior by P i . Then the actions of the firstindividual may be described by a symbolic equation,A 1 = f (P 1 ,A 2,…, A n ).There is one such equation for each individual. Together they constitute nequations in the n variables A 1,…, A n . In general, these may then be solved toexpress the actions of all individuals in terms of the data P 1,…, P n . Therefore,given the reaction of each individuals to his total (social and other) environment,as expressed in relations of type (1), and given the nonsocialenvironmental factors, which we may term exogenous, we can determinethe behavior of society in the sense that we can determine the behavior ofany individual in society.(Arrow [1951] 1968: 640)From this individualistic viewpoint, then, the actions of each individual is seen asthe result of (1) his/her psychology, (2) the physical surrounding and (3) theactions of other individuals. The behaviour of groups is explained by aggregatingthe behaviour of individuals.In this early paper by Arrow, there is no special reference to general equilibriumtheory, but there is little doubt that Arrow has this theory in mind whenexplicating the individualistic viewpoint in social science. This belief isconfirmed by an article written some eight years later and five years after the

Economics: the individualist science 245famous 1954 article on the existence of equilibrium in a competitive economy.With specific reference to this theory, he writes: ‘In this individualistic framework,every relevant variable, except those classified as exogenous for the wholeeconomic system, is the result of a decision on the part of some one individualunit of the economy’ (Arrow, 1959: 42).Figure 8.2 The methodological individualism of general equilibrium theoryThe economic general equilibrium theory is, of course, an extreme simplification.This has always been pointed out by its critics: institutionalists, Marxists,Austrians and Keynesians. But it is readily acknowledged also by its creators (see,e.g. Arrow and Hahn, 1971: vi; McKenzie, 1989: 28f). Social institutions, such asproperty rights and other laws, the state, the firm, the family and money, areignored, or treated as exogenously given. But simplification is not, itself, aproblem. Science proceeds by simplifying complex reality. Nevertheless, there is agrowing feeling, even among economists working in this field, that general equilibriumtheory is, indeed, in trouble and that part of the trouble is that it is toofar removed from reality (Hahn, 1981). According to one critic,General-equilibrium analysis is still remarkably backward in examining howa system works under a given institutional organization, since it is not seriouslyconcerned with observing, formulating and analyzing these customsand rules. The antiquated theory that prices are fixed at the point wheredemand and supply coincide is still being used as hitherto, or else all wehave is a plethora of new models and new theories based on ideas whichassume crude institutional arrangements conceived from highly superficialobservations. Since this is the state of affairs even with the market, otherinstitutions are all disregarded in the mock-up created by the generalequlibriumtheorists in spite of the fact that institutions such as banks, thecentral bank, labour unions, the government and capitalists’ organizationsplay a highly significant role in the actual economy.(Morishima, 1984: 65)I am a dilettante in this field, but the main problem seems to be a failure ofthis theory to account satisfactorily for the three related phenomena of disequilibrium,time and uncertainty. In what follows, I am going to mention some ofthe limitations and problems connected with the economic theory of generalequilibrium, not because of any intention, on my part, to criticise this theory –economists, themselves, do this much better – but because it casts doubts upon

246 Economics: the individualist sciencethe alleged individualism of economics, or, at least, upon one particular versionof methodological individualism.A first crack in the wall of general equilibrium theory was detected by Arrowalready in the 1959 article quoted above. In this article, it is argued that ‘thereexists a logical gap in the usual formulations of the theory of the perfectlycompetitive economy, namely that there is no place for a rational decision withrespect to prices as there is with respect to quantities’ (p. 41). The reason is thatindividuals are assumed to be price takers. Arrow’s first inclination, naturally,was trying to fill the gap. But thirty-five years later, the problem remains andnow it is seen as a devastating blow to the prevailing assumption that standardeconomic analysis conforms to the strictures of methodological individualism.It is a touchstone of accepted economics that all explanations must run interms of the actions and reactions of individuals. Our behavior in judgingeconomic research, in peer review of papers and research, and in promotions,includes the criterion that in principle the behavior we explain and thepolicies we propose are explicable in terms of individuals, not of othersocial categories. I want to argue today that a close examination of even themost standard economic analysis shows that social categories are in fact usedall the time and that they appear to be absolute necessities of the analysis.(Arrow, 1994: 1)Arrows first example of a social category in economics is the prices faced byeconomic agents in the theory of competitive equilibrium. ‘What individual haschosen prices?’ The answer is none. There is no explanation of the formation ofprices in terms of individuals. But, according to Arrow (1994: 4): ‘They aredetermined on (not by) social institutions known as markets, which equate supplyand demand’. Prices thus presuppose the social institution of the market. It is notclear whether Arrow conceives of prices, themselves, as making up a social institution.To do so is not unusual, however (Schumpeter, 1909: 217; cf. Boland,1982: 50f), as evidenced by the following statement in a well-known textbook: ‘Inthe traditional models of microeconomics, prices in an impersonal marketplace constitutethe institutional framework’ (Kreps, 1990a: 5). 19But there are other social institutions that cannot be entirely ignored ineconomic theory. The most obvious, perhaps, is money. In general equilibriumtheory, money plays no role, but in real economies, prices are money prices, andeconomic theory cannot afford to neglect this simple fact, as, of course, it doesnot. But, then, the problem arises of how to give an individualistic explanation ofthe social institution of money. This problem was faced already by the firstconsistent methodological individualist among economists; Carl Menger, whoattempted exactly this: to explain the origin of money in terms of individualsalone (see pp. 90f ).Other social institutions, hard to ignore in economics, are households andfirms. Economic theory is often stated in terms of ‘individual agents’, which aremainly of two sorts: buyers and sellers (Walras [1874] 1984), or consumers and

Economics: the individualist science 247producers (Arrow and Debreu, 1954). Behind those labels, however, you usuallyfind the household and the firm (Arrow, 1968: 377). 20In fact, even in economics, the unit of the theory of production is not reallythe individual but the firm, which is an operating organization of individuals.Similarly, the unit of consumption is really the household, not theindividual consumer.(Arrow [1951] 1968: 640)This simple fact is made an explicit part of the theory of general equilibriumby Arrow (1968: 382) and by Arrrow and Hahn (1971: 18): ‘Demand and supplydecisions are taken by two kinds of agents: households and firms’. 21On closer scrutiny, then, the ‘individual’ agents of economic theory are reallycollective agents. How does this square with methodological individualism? Theapparent answer would seem to be: not at all (Udehn, 1987: 185). 22 There aretwo ways out of this dilemma for the methodological individualist. The first is tostretch the notion of an ‘individual’ so as to include households and firms (seeKhalil, 1997). This way out has been suggested by Marteen C.W. Janssen.it is clear that the claim that economics is based on a form of individualismonly has a chance to survive potential criticism if a broad notion of ‘individual’is taken: as households and firms are treated as basic concepts ineconomics, the position that economic theory is non-individualistic would bea trivial one if the notion of ‘individual’ was taken in a more narrow sense.( Janssen, 1993: 10)I find this solution question-begging and a bit desperate. I agree, of course, thatit is possible and, even, necessary to treat collectivities as unitary actors, or ‘individuals’,for many purposes in social science, but this does not savemethodological individualism (Hindess, 1984; 1986: 115–19; 1988: 44–8).Methodological individualism was always about individual human beings. Thesecond way out is more consistent with original methodological individualism. Itis to reduce all collective action to the actions of individual human beings.According to Armen A. Alchian and William R. Allen, the first postulate ofeconomics says that ‘The Unit of Analysis Is the Individual’. This means thatThe actions of groups, organizations, communities, nations, and societiescan best be understood by focusing attention on the incentives and actionsof members. When we speak of the goals and actions of the USA, we arereally referring to the goals and actions of the individuals in the USA. Abusiness, a union, or a family may be formed to further some commoninterest of the constituents, but group actions are still the results of decisionsof individuals. Therefore, do not ask ‘Why does the U.S. government makethe decisions it does?’ or ‘Why does General Motors do this?’ or ‘Why doesa union do that?’ Ask instead, ‘Why does a decision maker decide as he

248 Economics: the individualist sciencedoes?’ The principles of economics are based on postulates about responsesof individuals to changes in the environment.(Alchian and Allen [1964] 1967: 14)There is a certain ambiguity here. Should we understand the above postulateonly as a reminder that collectives are made up of individuals and their actions,or as an injunction actually to carry out the reduction. I suspect that mosteconomists are satisfied with the first, ‘ontological’ solution, but there are alsoattempts to explain the actions of firms and households as the outcome of theactions of individuals.It is also part of general equilibrium theory that individuals exchange witheach other things called ‘commodities’ or ‘goods’. Rights of property andcontract are not always mentioned, but they are implied by the various statementsof this theory. Arrow and Debreu (1954: 270) write that economic agents‘possess’, or ‘hold’, assets, or commodities. Debreu (1959: ch. 5) says explicitlythat his theory of equilibrium applies to a ‘private ownership economy’. Arrowand Hahn (1971) assume that consumption ‘goods are owned by individuals andhouseholds’ and that goods of higher order are also ‘private commodities’.Certain property rights are, thus, assumed by the theory of general equilibriumand it might be argued that also the state is assumed, since property rights are, inall market societies, instituted and defended by the state. 23So far, I have concentrated on the institutional element in the presuppositionsof general equilibrium theory. (This element turns it into an example of institutionalindividualism). But there is also the problem that, even with the aid of thisinstitutional framework, general equilibrium theory really fails to provide anindividualistic explanation of prices and the allocation of resources (Janssen,1993: ch. 7). Arrow and Debreu proved the possible existence of an equilibrium,but the aim of general equilibrium theory is also to demonstrate uniqueness andstability. To achieve these aims has proven much more of a problem (Kirman,1989: 127; Ingrao and Israel, 1990: 360f). 24Starting with the latter, it is safe to say that no real economy is ever exactly ina state of equilibrium and there is little reason to assume that it is necessarilystable, in the sense that it always converges to some equilibrium. An obvious question,then, is how the economy gets from disequlibrium to equilibrium. ‘Theattainment of equilibrium requires a disequilibrium process’ (Arrow [1987a]1990: 27). It can be argued, therefore, that the theory of general equilibrium isin need of disequilibrium foundations (Fisher, 1983: ch. 1). Leon Walras tried tosolve this problem by introducing an auctioneer, 25 adjusting prices and quantitiesuntil the market is cleared. Walras’s auctioneer, however, is a mythical figure,solving the problem in the typical manner of a deus ex machina (Leijonhufvud,1968: 390). If taken literally, however, an auctioneer is a social institution and, assuch, a Trojan horse in any individualist theory of the economy (see Mirowski1981: 596). There are, of course, more recent attempts to model the adjustmentprocess, but according to Fisher (1989: 37) there is still a ‘lack of a satisfactorytheory about the disequilibrium behaviour of agents’ (see also Arrow, 1987: 203).

Economics: the individualist science 249Ingrao and Israel (1990: ch. 12), go further and argue that attempts to solve theproblem of achieving global stability has come to a dead end. In this situation,‘[t]he only way possible is a thorough re-examination of the theory’s basichypothesis, i.e., a true paradigmatic revolution’ (p. 362).Since I am not an economist, myself, I do not know what such a revolutionmight involve, but according to one economist, Lawrence C. Boland, it mostprobably involves some institutional element incompatible with ‘psychologisticindividualism’. According to the latter doctrine, only psychological states andnatural constraints, but not social institutions, are permitted as exogenous variables.But disequilibrium implies that, at least, one individual is not maximising,and the explanation for this has to be some institutional constraint uponbehaviour. The typical way out for neoclassical economists is to accept disequilibriumand social institutions in the short run, and/or on a local level, but toclaim that equilibrium is achieved in the long run, and/or on a global level. (SeeBoland, 1982: ch. 3; 1986: passim.)The problem of uniqueness is no less serious. In a number of papers from the1970s, Hugo Sonnenschein, R. Mantel and Gerard Debreu showed that in aneconomy where all individuals have well-behaved excess demand functions,almost anything can happen (Hartley, 1997: 190–3). 26 General equilibrium,theory, therefore, is not in such a state as to be able to provide secure microfoundationsfor macroeconomics (Rizvi, 1994). It is probably much worse: Accordingto, at least, one mainstream economist, Alan Kirman (1989), the theory ofgeneral equilibrium is really empty: ‘The Emperor has no Clothes’ (Kirman,1989). As a sociologist, I am not competent to assess the validity of this judgement,but I note that Kirman blames methodological individualism, or the‘individualistic approach’, for the alleged failure. The root of the problem is thatgeneral equilibrium theory starts from isolated individuals acting independentlyof one another. In order to obtain uniqueness and stability, however, we mustassume that the behaviour of individuals is orderly:If we are to progress further we may well be forced to theorise in terms ofgroups who have collectively coherent behaviour … The idea that weshould start at the level of the isolated individual is one which we may wellhave to abandon … It is clear that making assumptions on the distributionof agents characteristics amounts, in some sense, to making assumptionsabout the organization of society.(Kirman, 1989: 138)The main root of trouble is that the real economy is ‘in time’, whereasgeneral equlibrium theory is ‘out of time’, to borrow two expressions from JohnHicks ([1976] 1982). In a real economy individuals have to base their decisionsupon expectations about an irremediably uncertain future. Reaching equilibrium,in this situation, is a bit like trying to hit a moving target without knowingits trajectory. Equilibrium becomes a matter of making the expectations of allindividuals consistent, but this task may very well be impossible (cf. Hollis, 1987:

250 Economics: the individualist science112). Only the future can tell. But in the meantime we have to conclude thatgeneral equilibrium theory is plagued by the problem of indeterminacy, in theform of multiple equilibria (Arrow [1987a] 1990: 33–5; 1987: 211).One way to address the problem of indeterminacy is by a theory of knowledge,or learning. If individuals behave in certain ways, this may be because theyhave learned to behave in these ways. This is the way forward suggested byArrow (1989: 21f). But knowledge is predominantly social knowledge and, therefore,problematic from the individualistic point of view. As Arrow points out(1994: 7), ‘Technical and other knowledge exists in social form: books or universities’.To this extent, at least, exogenously given knowledge is incompatible withmethodological individualism.Arrow refers to Lawrence A. Boland for an illuminating discussion of individualismwith special regard to the implicit problem of knowledge (Arrow, 1994: 3,note 1). It may be added that Boland has also been explicit about the need tointroduce a theory of knowledge, or learning, in economics (Boland, 1982: ch. 4;1986: part III; 1992: part II). Like Arrow, he discusses technical knowledgeembodied in institutions and argues, convincingly I believe, that it cannot bemade compatible with the neoclassical programme of treating all social institutionsas endogenous variables (Boland, 1992: ch. 8).Game theoryThis individualistic viewpoint, as we may term it, is explicit in the maintradition of economic thought and is completely accepted in the vonNeumann–Morgenstern game theory.(Arrow [1951] 1968: 640)It is a common opinion that game theory is a typical example of methodologicalindividualism (Harsanyi, 1968; Elster, 1982). As we shall see below, however,there are also those who doubt that game theory is necessarily individualistic. Asin the case of general equilibrium theory, this divergence of opinion can beexplained as due to the existence of two versions of methodological individualism;one weak, the other strong, which is a major source of confusion.When Arrow and Debreu constructed their model of general equilibrium in1954, they borrowed from the mathematics of game theory, as developed by vonNeumann and Morgenstern ten years before (Arrow and Hahn, 1971). Even so,the impact of game theory on economics has been less and much later than theArrow–Debreu model of general equilibrium. Actually, game theory became animportant tool in political science, before it became prevalent in economics.The classic treatise on game theory, John von Neumann’s and OscarMorgenstern’s Theory of Games and Economic Behavior (1944), was written by amathematician and an economist, but was intended to be of use also in sociology.The economist Oscar Morgenstern was trained in the Austrian Schooland this is sometimes used to explain the individualistic nature of game theory

Economics: the individualist science 251(Leonard, 1992: 55; Mirowski, 1992: 136f). The evidence for this suggestion liesmainly in the nature of game theory itself, but there is also the following statementby von Neumann and Morgenstern ([1944] 1953: 7): ‘We believe that it isnecessary to know as much as possible about the behavior of the individual andabout the simplest forms of exchange’. Indeed, Kenneth Arrow seems to believethat game theory is more individualistic than the economic theory of competitiveequilibrium. ‘The current formulation of methodological individualism isgame theory’ (Arrow, 1994: 4). The reason for this is that, in game theory, ‘pricesnever appear as objective phenomena; they are only subjective, that is expectationsheld in the agents minds’ (p. 5). This is clearly in conformity with theAustrian version of methodological individualism.In another respect, however, game theory is less individualistic than isorthodox economics. As we have seen, the point of departure of marginalisteconomics – Austrian Economics, in particular – was the isolated individualengaged in economising. The main problem of neoclassical theory of the marketis, of course, to explain interaction, in the form of exchange, but this makes nodifference for economic man. In the neoclassical theory, individuals interact onlyanonymously and impersonally on the market, and without considering therational choices of other individuals. Even when acting on the market, the individualacts as an ‘isolated individual’ engaged in a ‘play against nature’. This ismost clearly seen in the case of the equilibrium theory of Walras and Pareto, inwhich ‘participants face fixed conditions and act like a number of RobinsonCrusoes – solely bent on maximizing their individual satisfactions, which underthese conditions are again independent’ (von Neumann and Morgenstern [1944]1953: 15). 27 It was the main business of game theory to challenge this RobinsonCrusoe model and replace it by a social exchange model. In the former, ‘Crusoeis given certain physical data (wants and commodities) and his task is to combineand apply them in such a fashion as to obtain a maximum resulting satisfaction’(von Neumann and Morgenstern [1944] 1953: 9). In the latter, the problem isdifferent, because the participant ‘must enter into relations of exchange withothers’ (p. 10). ‘This would justify assuming rational behaviour on the part ofothers as well’ (p. 31). In this situation, simple utility-maximisation is notpossible, because ‘all maxima are desired at once – by various participants’ (p.11).The difference between Crusoe’s perspective and that of a participant in asocial economy can also be illustrated in this way: Apart from those variableswhich his will controls, Crusoe is given a number of data which are‘dead’; they are the unalterable physical background of the situation … Nota single datum with which he has to deal reflects another person’s will orintention of an economic kind – based on motives of the same nature as hisown. A participant in a social exchange economy, on the other hand, facesdata of this type as well: they are the product of other participants’ actionsand volitions (like prices). His actions will be influenced by his expectations

252 Economics: the individualist scienceof these, and they in turn reflect the other participants’ expectations of hisactions.(von Neumann and Morgenstern [1944] 1953: 11f)The distinction made by von Neumann and Morgenstern, between aRobinson Crusoe model and a social exchange model, has been elaborated bylater game theorists, using a different terminology. John C. Harsanyi, forinstance, makes the distinction between ‘individual decision theory’ and the‘theory of rational behaviour in a social setting’, including game theory. ‘Theproposed basic difference between decision-theoretical situations and game situationslies in the fact that the latter involve mutually interdependent reciprocalexpectations by the players about each other’s behavior; the former do not’(Harsanyi, 1977: 10). In the terminology of Jon Elster, traditional economicsworked with the assumption of ‘parametric rationality’, and game theory withthe assumption of ‘strategic rationality’ (Elster, 1979a: 68f; 1979b: 18f, 117f).It goes without saying that rationality is more problematic in a strategic thanin a parametric environment. Not surprisingly, therefore, game theory is plaguedby the same difficulty as the economic theory of equilibrium: the existence ofmultiple equilibria (see, e.g. Kreps, 1990c: 507ff; 1990b: 95–107). The suggestedsolutions are also the same: determinacy is provided by invoking social institutionsin the form of norms and conventions (Mirowski, 1986; Ferejohn, 1991:298f). ‘In short, game theory offers the lesson that methodological individualismcan only survive by expanding the notion of rational agency. The challenge iswhether there are changes of this sort which will preserve the individualistpremise’ (Hargreaves Heap and Varoufakis, 1995: 33). It is a bit surprising,therefore, that Arrow should have come to the conclusion that game theory is thecurrent formulation of methodological individualism (see above). As Arrownotes, himself, the problem of multiple equilibria is found in game theory no lessthan in the theory of general equilibrium. Indeed, it is arguable that gametheory is the main source of indeterminacy in contemporary microeconomics(Cowen, 1998: 137–44).There is, however, also another possible entry for social institutions in gametheory: the rules of the game (cf. Janssen, 1993: 35). Game theory is a mathematicaltheory, and as such it has no need for social institutions. A game isdefined in terms of players, pay-offs, strategies, moves and so on. But mathematicssays nothing, at all, about social reality. In real life, however, most gamesare social situations created by institutional and structural constraints. 28 VonNeumann and Morgenstern ([1944] 1953: 23–5) hint at the embeddedness ofgames in social organisation, but are not very explicit about it. Later game theoristsare more explicit on this point, however. Harsanyi (1977: 88), for instance,maintains that social situations, analysed as game situations, include thefollowing ‘rules of the game’:1 The social conventions observed by the players (e.g., legal or moral rules).

2 The laws of nature (e.g., the physical, chemical, and biological laws governingthe performance of the human body and of material equipments used).3 The initial distribution of resources (e.g., bodily strength, economic resources,military equipment) among the players – including the initial distribution ofinformation, technological knowledge, and practical skill.Martin Shubik, similarly, makes clear thatEconomics: the individualist science 253my basic approach to economics is through the construction of mathematicalmodels in which ‘the rules of the game’ derive not only from theeconomics and technology of the situation, but from the sociological, political,and legal structure of society as well. Private ownership of land, therights of various public or private groups to tax, and the existence of certainfinancial institutions are examples of legal and social features that mayrequire delineation in a particular model. Similarly, rigid prices, a classstructure that sets money lenders and merchants apart from aristocrats,priests, and peasants, and the redistribution of economic goods by appealsto social justice or time-honored custom are non-economic factors that maybe present to some degree and that should be reflected in any realisticmodel.(Shubik, 1982: 10)Thus, it would seem that many, or most, real game situations are constitutedby social institutions. This seems to be Arrow’s conclusion too. In real-life games,‘the rules of the game are social’. The reason is that ‘individual behavior isalways mediated by social relations’ (Arrow, 1994: 5). 29 Therefore, if generalequilibrium theory fails to satisfy the conditions of methodological individualism,game theory does too.In a recent book on game theory by Shaun P. Hargreaves Heap and YanisVaroufakis (1995), the issue of methodological individualism appears as problematic,but unsettled. They argue that game theory makes a clear separationbetween individual choice and (social) structure – something which methodologicalindividualists like. But it is not without problems, since methodologicalindividualists, would also like to explain social structures as the outcomes ofprevious choices of individuals. According to Hargreaves Heap and Varoufakis(p. 32), it is a characteristic mark of all methodological individualists that theysee social structures, as ‘merely the deposits of previous interactions’ (see alsoField, 1979: 55; 1984: 698ff).the individualist will want to claim that ultimately all social structures springfrom interactions between some set of asocial individuals; this is why it is‘individualist’. These claims are usually grounded in a ‘state of nature’ argument,where the point is to show how particular structures (institutional

254 Economics: the individualist scienceconstraints on action) could have arisen from the interaction between asocialindividuals.(Hargreaves Heap and Varoufakis, 1995: 32)It is obvious that, Hargreaves Heap and Varoufakis, like Arrow, conceive of‘methodological individualism’ in the original, strong sense of that term, prevalentin economics. In contrast to this view, Janssen (1993: 35–40) argues thatgame theory is the main example of methodological individualism, exactlybecause it explains the aggregative outcome of interacting individuals in an institutionalenvironment. But that is, because he conceives of methodologicalindividualism as ‘institutional individualism’, in the sense of Agassi and Boland.ConclusionIn the beginning of this chapter, I quoted Lawrence A. Boland, who suggestedthat neoclassical economics is guided by a methodological rule of psychologisticindividualism. This means that only psychic states and natural constraints areallowed as exogenous variables. If there are any social institutions among theexogenous variables, they must be reduced to psychic states and naturalconstraints. We have seen that this rule does indeed govern the work of neoclassicaleconomists, but we have also seen that it is constantly violated. Socialinstitutions pop up everywhere, if not as explicitly stated exogenous variables, soat least as exogenously given conditions. The new institutional economics, whichis the subject of the next chapter may be seen as an attempt to deal with thisproblem.We have also seen, however, that some eminent economists, such as KennethJ. Arrow and Alan Kirman have begun to doubt the desirability and even thepossibility of methodological individualism. Economic theory has recently beenplagued by fundamental indeterminateness in the form of multiple equilibria.One way out of this impasse may be to introduce determinateness in the form ofsocial institutions. This is the solution to ‘the problem of order’, suggested byTalcott Parsons in The Structure of Social Action ([1937] 1968: ch. 2–3).

9 The new institutionaleconomicsIn the theories of general equilibrium and of games, social institutions aretreated as exogenously given and neglected. The Marshallian methodology ofcomparative statics, used, above all, in the analysis of partial equilibrium is a bitmore complex. In comparative statics, social institutions are often treated asexogenously fixed, for the purposes of short-run analysis. But it is assumed thatthey are the result of an earlier process of constrained maximisation, whichmeans that they may be turned into endogenous variables in a model of longrunequilibrium. According to Lawrence A. Boland (see pp. 228–30), this is notjust an assumption, but a goal, if distant, on the ‘hidden agenda’ of neoclassicaleconomics. The ambition of neoclassical economics is to turn all social institutionsinto endogenous variables, leaving only psychic states and natural givens asexogenous variables in economic analysis. This methodological individualism, Ibelieve, is clearly manifested in some of the contributions to the new institutionaleconomics.But Boland also recognised another, less demanding version of methodologicalindividualism, which, following Joseph Agassi, he called ‘institutionalindividualism’. In this form of methodological individualism, social institutionsare accepted as exogenous variables and conditions in economic models, withoutany requirement that they must be endogenised. As we shall see, this moremoderate position is also common among representatives of the new institutionaleconomics.According to the traditional division of labour in the social sciences, socialinstitutions used to be the province of anthropology, political science and sociology.Today, this division of labour is no longer very clear-cut. One of themost significant developments in recent economics is the emergence of a newinstitutional economics. 1 It is possible to see this development as a response tosome of the problems and limitations of traditional microeconomics (Demsetz,1966: 61). According to the new institutionalists, the main problem withorthodox economics is its lack of realism, and the remedy is to bring back in allthe social institutions left out of analysis by the theories of general equilibriumand of games. But the new institutional economics is also an attempt to usesome of the tools of the economic approach to phenomena outside the traditionalscope of economics. Because of this, the new institutional economics is

256 The new institutional economicsalso a manifestation of, what has been called, ‘economic imperialism’ (Tullock,1972; Stigler, 1984; Udehn, 1992).Before turning to the new institutionalism, I will say a few words about theold institutionalism and, especially, about the differences between the new andthe old institutionalism in economics. I will also say a few words about the differencesbetween the new institutionalism in economics and the newinstitutionalism in sociology and political science.The old institutionalism in economics is in many ways the North Americancounterpart of the Historical School in Germany (Hutchison, 1984: 21f). Likethe latter, it engaged in largely descriptive investigations of social institutions andtheir role in economic life. Its most well-known representatives include,Thorstein Veblen, John R. Commons, Wesley C. Mitchell and Clarence Ayres.The old institutionalism was most vital in the beginning of this century, but isstill alive. More recent adherents are Allan G. Gruchy, Warren J. Samuels andMarc R. Tool.The new institutionalism in economics started in the 1960s – although itsroots are older – and includes names like Ronald H. Coase, Armen A. Alchian,Harold Demsetz, Douglass C. North, Andrew Schotter, and Oliver E.Williamson. New institutionalists are usually dismissive of the contributions ofthe old institutionalists to economics, and tend to emphasise their own allegianceto the main assumptions of mainstream economics: both classical (see, e.g.Coase, 1984) and neoclassical (Eggertsson, 1990: 3–32). It may even be arguedthat, except in political matter, the new institutionalism is closer to Marxism thanto the old institutionalism. At least they share a sense of the importance of propertyrights (Eggertsson, 1990: 33, note 1). 2In a comparison of the old and the new institutionalism, Malcolm Rutherford(1994) suggests that they might be distinguished in terms of the followingdichotomies: ‘formalist versus anti-formalist; individualist versus holist; rationalchoice versus behaviourist; evolutionary or invisible-hand versus collectivist; noninterventionistversus interventionist’ (p. 5). While arguing that these differencesare often exaggerated and that they are sometimes larger within than betweenthe two traditions, Rutherford nevertheless reaches the conclusion that the oldinstitutionalism is largely anti-formalist, holist, behaviourist, collectivist andinterventionist, while the new institutionalism is formalist, individualist, rationalist,evolutionary and non-interventionist (173–6). 3 For my purposes, theimportant points are that the old institutionalism is holistic and behaviouristic.Consequently, it is of no further interest in a book about methodological individualism.According to Rutherford, however, it is not as simple as that. Following Agassiand Boland, he makes the distinction between psychologistic and institutionalindividualism (1994: 36–50). With the help of this more differentiated view ofthe matter, it is possible to argue that the old and new versions of institutionalismmeet on the common ground of institutional individualism. My interest is neverthelesslimited to the new institutionalism, because it originates in anindividualist tradition. I agree with Rutherford that the work of the new institu-

The new institutional economics 257tionalists in economics belongs in both categories; institutional as well as psychologisticindividualism. But while the ambition is usually to stay within the latter,strong version of methodological individualism, most of their work ends up inthe former, weak version of this principle. At least, this is what I will try to showin this chapter.It is often maintained that institutionalism is the traditional approach in politicalscience (see Udehn, 1996: 2f). The roots of this institutionalism can betraced back to Aristotle’s famous comparative analysis of constitutions in ThePolitics. When we turn to the new institutionalism in modern political science,however, it is important to distinguish, at least, two branches: (1) One rationalchoice branch, which has much in common with the new institutionalism ineconomics. This branch will be discussed in the next chapter. (2) Another, sociohistoricalbranch, has more in common with the old institutionalism ineconomics and political science and also with the new institutionalism in sociology.In contrast to the new institutionalism in economics, it sees institutions asirreducible and fairly stable entities, capable of moulding the behaviour of individuals(March and Olsen, 1989: 16ff). As such it is different from individualisticapproaches to politics.Like political science, traditional mainstream sociology used to be institutionalistic.Emile Durkheim, for instance, suggested that ‘sociology … can be definedas the science of institutions, their genesis and functioning’ ([1895] 1982: 45) andTalcott Parsons says much the same thing in The Social System (1951: 36–51).What is usually called ‘the new institutionalism’ in sociology is, in my opinion,not as different from the old institutionalism, as some of its adherents wish tomaintain (Powell and DiMaggio, 1991: 11–33). There is a certain shift of focusfrom structure to culture, but it is not dramatic. Closely connected to this shift isanother one, from a normativist to a cognitivist conception of social institutions.A third difference is that, whereas the old institutionalism was a generalapproach to social phenomena, the new institutionalism is largely confined to thetheory of organisations. Among the things that unite the old and new versions ofinstitutionalism in sociology are a pronounced anti-rationalism and antiindividualism.4 For this reason, also the new institutionalism in sociology is verydifferent from that in economics (Powell and DiMaggio, 1991: 2–11; Sjöstrand;1993).To come back to the new institutionalism in economics. As we have alreadyseen, it is supposed to be a formalist and individualist rational choice approach.Considering its roots in microeconomics, this is not at all surprising. It should benoted, though that the roots of the new institutionalism in economics are not tobe found in the theory of general equilibrium, but in less orthodox approaches.The most orthodox and most influential source of the new institutionalism ineconomics is the Chicago School, which brought property rights and transactioncosts into economic analysis. The Chicago School is largely Marshallian andfocuses on maximising behaviour and partial, rather than general, equilibrium.But there is also the important exception of Armen Alchian, who suggested thatthe assumption of maximising behaviour might be replaced by that of selection.

258 The new institutional economicsAnother important source is the Virginia School of Public Choice, whichextended economics to political institutions and political decision-making. Itsmethodology is based on the three assumptions of self-interest, exchange andmethodological individualism. A third source is Austrian Economics. It differsfrom the Chicago and Virginia Schools in several respects. Although arguablythe most important source of methodological individualism and rational choicein modern social science, it has recently become less dependent on both theseassumptions. At least this is the case with Friedrich von Hayek’s evolutionarytheory of rule systems.Richard N. Langlois (1986a: 1) has identified four different strands in the NewInstitutional Economics: (1) the evolutionary theory of Nelson and Winter (2) themodern Austrian school, as influenced by F.A. Hayek, (3) the transaction-costeconomics of Ronald Coase and Oliver Williamson, and (4) the property rightsliterature of Armen Alchian and Harold Demsetz, which is also inspired byCoase. This list is far from exhaustive, but it includes those theories which, Iguess, most people would recognise as manifestations of the new institutionaleconomics, and, among these, the third and the fourth are probably the mostself-evident candidates of inclusion. Be that as it may, Langlois, himself, is mostfavourable to the evolutionary alternatives and, especially, to the Austrian versionof evolutionary institutionalism (1986a: 4ff; 1986b: 247–53).My interest is not in the new institutionalism, per se, but in its putative individualism.From this point of view, I have found it most fruitful to concentrate onthose strands of the new institutional economics, which I have not discussedbefore. This means that I will make only brief mention of the Austrian versionof evolutionary economics at the end of this chapter. Before getting on with thistask, it might be wise to say something about the concept of ‘institution’. Aninstitution is either defined as a ‘pattern of behaviour’, or as the rule(s), ornorm(s) prescribing this, or that, pattern of behaviour. My view is that neitherdefinition is sufficient, by itself, but both are necessary and sufficient. An institution,then has two aspects: one behavioural and one normative. A rule thatnobody follows is not an institution and a pattern of behaviour, without anormative element, is a bodily movement, or an individual habit, but not a socialinstitution.The most important sources of the new institutionalism in economics are twoseminal papers by the Chicago economist Ronald Coase. In ‘The Nature of theFirm’ (1937), he took issue with the fundamental assumption, underlying mainstreameconomics, that co-ordination of economic activities takes place onlywith the help of the price mechanism. If this is so, why is an organisation, suchas the firm, necessary? ‘Our task is to attempt to discover why a firm emerges atall in a specialised exchange economy’ (p. 390). The answer to this question isthat the price mechanism is not without costs. First, there are costs of discoveringwhat the prices are. Second, there are ‘costs of negotiating and concludinga separate contract for each exchange transaction which takes place on a market[which] must also be taken into account’. Coase calls them ‘contract costs’ (pp.390f). In addition, there is a fundamental uncertainty about the size of these

The new institutional economics 259costs. 5 For all these reasons, it might, for certain purposes, be better to replacethe price mechanism with organisation; the vertical co-ordination, or integration,of economic activities by an entrepreneur.In ‘The Problem of Social Cost’ (1960), Coase set out to challenge the standardeconomic analysis of externalities, by Arthur C. Pigou. According to thelatter, externalities, not included in any calculation of costs, invariably leads tosocial costs, which call for intervention by the state. Against this view, Coase firstargues that, in the absence of any costs for market transactions, negotiationsbetween parties would lead to the maximisation of social welfare, irrespective ofthe initial distribution of rights (pp. 2ff). This is the so-called ‘Coase theorem’.But, since transactions are far from being costless, the Coase theorem is not validin a real economy. In a real economy, therefore, the nature and distribution ofproperty rights is of fundamental importance and do affect social welfare (pp.15ff). Nevertheless, economists are wrong to assume that it is always efficient toset a price on externalities. If, for instance, this policy leads to the closing downof businesses producing externalities, it might very well lead to an inefficientsocial outcome.The important point of Coase’s first article is that there are costs of transaction,the important point of his second article is that property rights matter. Inboth cases, the message is that economics should study real economic systemsand this includes, most importantly, taking social institutions into account(Coase, 1988b: 23ff; 1992). What does this mean, in terms of methodologicalindividualism? 6 Coase does not, as far as I know, discuss methodological individualismexplicitly, but my impression is that he could not be a strongmethodological individualist. It is possible to interpret his theory of the firm asan example of strong methodological individualism, where social institutions (thefirm) is explained in terms of the (maximising) behaviour of individuals, but itwould not be his own interpretation, since transaction costs depend uponanother institution: law. ‘As I came to realize when I wrote “The Problem ofSocial Cost,” all these interrelationships [between the costs of transacting andthe costs of organizing] are affected by the state of the law, which also needs tobe taken into account in the analysis’ (Coase, 1988b: 47). What Coase says aboutinstitutions, leads in the direction of institutional individualism. It is prettyobvious that Coase sees social institutions as causal factors, with a strong directeffect on economic behaviour and an indirect effect on traditional economicvariables such as resource allocation and on wealth and its distribution.Integration in a firm, for instance, ‘creates a different institutional setting, andour problem is to discover what effect this has on economic behavior’ (Coase,1988b: 31; see also Coase, 1978: 208; 1992: 718). This is institutionalism, notindividualism.It should also be pointed out that Coase never made explicit use of theassumption of maximising behaviour and in his later writings explicitly deniesthat it is needed in economic analysis (1984: 231). The only assumption that isneeded is that individuals respond to relative costs and benefits. The emergence ofthe firm, for instance, is not a matter of utility maximisation, but of ‘adaptation to

260 The new institutional economicsthe existence of transaction costs’ (1988a: 7). Coase does not say if he conceivesof adaptation as a result of rationality or selection, but both are common in thenew institutionalist literature on organisation and property rights. A commonconsequence of rejecting the assumption of maximisation, seems to be that theanalysis ends in crude functionalism (cf. Granovetter, 1985: 488f): institutionsexist because they are beneficial – as if a social maximiser had been part of theprocess and settled for the most efficient one (cf. Field, 1984: 694). One exampleis provided by the following statement by Coase: ‘Markets are institutions thatexist to facilitate exchange, that is, they exist in order to reduce the cost ofcarrying out exchange transactions’ (Coase, 1988a: 7).Coase’s ideas are the sources of two important branches of the new institutionaleconomics: the economic theory of organisation, especially the firm, andthe theories of property rights and of law. Both of them are based on the idea oftransaction costs. But Coase was not alone. In ‘The Nature of the Firm’ (1937:398ff), Coase refers to Frank Knight’s discussion of uncertainty, in Risk,Uncertainty and Profit (1921), as the reason for the existence of entrepreneurs.Coase (1988b: 20) denies, however, that Knight was the source of his own ideason the nature and emergence of the firm. 7 In any case, Knight was an importantfigure in Chicago by the time Coase arrived there a second time in the 1950sand the similarity of their ideas may be part of the explanation why Coase wassaved from oblivion. In Chicago, he found likeminded people, who were receptiveto his ideas.A brief look at the social institutions treated by the new institutionaleconomics reveals that they fall largely into two classes: social organisations andgeneral rules of the game. I will divide my presentation of the sub-field of neoinstitutionaleconomics according to this classification.Social organisationsThe economies of modern industrialised society can more appropriately belabelled organisational economies than market economies. Thus, evenmarket-driven capitalist economies need a theory of organisation as muchas they need a theory of markets.(Simon, 1991: 42)For a long time, neoclassical economics was presented as if the actors on themarket were individual human beings and economics, consequently, a theoryabout individuals. In actual fact, however, the market, like the rest of society, ismade up of organisations, and little else (cf. Ahrne, 1990; 1994). Neoclassicaleconomics, therefore, is really about organisations, such as the household, thefirm and, less obviously, the state. As I have already suggested (p. 247), this is aflagrant violation of methodological individualism. Adherents of this principle,therefore, have tried to escape this embarrassing situation in various ways and/ortried to belittle the significance of this violation.

The new institutional economics 261Although in modern economics, collections of individuals are sometimestreated as ‘entities’ for analytical purposes (examples of ‘the household’, ‘thefirm’, and even occasionally ‘the state’ spring to mind) the ultimate unit ofanalysis is always the individual; more aggregative analysis must be regardedas only provisionally legitimate. In other words, the economist is alwayssensitive to the possibility that the holistic treatment of groups of individualsmay mislead greatly, or involve overlooking dimensions of reality that areextremely important.(Brennan and Tullock, 1982: 225)As I have already shown in chapter 4, this individualist perspective has beenadvocated most explicitly by the Austrian School. Later on in this chapter, wewill see that it is a constitutive element also of the Public Choice approach of theVirginia School. Members of the Chicago School are, perhaps, a little bit lessexplicit about it (see, however, Alchian and Allen [1964] 1967: 14), but, I believe,no less committed to the principle of methodological individualism. The ambitionof many Chicago economists is apparently to analyse all social phenomena,as if they were markets, where individuals contract with one another in order tomaximise their utility. Social organisations are typically seen as frozen, or crystallised,markets, where individuals are parties to long-term contracts. I interpretthis approach to social organisations as a manifestation of methodological individualism.According to a recent book on this subject, the new institutional economics isbased on three basic assumptions: methodological individualism, utilitymaximisationand individual rationality. The first is stated thus:1 Methodological individualism: An entirely new interpretation is given to therole of individual decision makers. Methodological individualism emphasizesthat people are different and have different and varied tastes, goals,purposes, and ideas. Hence, the implication is that ‘society,’ ‘the state,’ ‘thefirm,’ ‘political parties,’ and so on are not to be understood as collective entitiesthat behave as though they were individual agents. The organization orcollectivity per se is no longer the main focus. Rather, it is thought that atheory of social phenomena must start with and base its explanations on theviews and behaviors of the individual members whose actions give rise tothe phenomena being studied.(Furubotn and Richter, 1997: 3; see also 1991: 4)In this section I am going to take a broad view of organisations, as including notonly firms, but also households and states (cf. Ahrne, 1990; 1994).The householdThe words ‘economy’ and ‘economics’, derive from the Greek word oikos, whichmeans household; and the first texts on economics were about this topic. I will

262 The new institutional economicsmake a short digression on the Greek theory of the household, as a relief tocontemporary economic theory. The first Greek work on the household is theOeconomicus, written as a Socratic dialogue by Plato’s contemporary Xenophon(426–354 BC). It is a treatise on the art of estate management (1979: 363), butnot very interesting from a theoretical point of view. 8A second work in this genre is the Oeconomica, attributed to Aristotle, but,according to authoritative opinion, not written by him. It is a treatise both onhousecraft and statecraft, but my present interest is in the first. According toOeconomica, the household is older than the state and consists of ‘(1) humanbeings, and (2) goods and cattle’ (1962: 329). The human part of the householdis made up of woman, man and children, and often enough also of slaves. It isthe task of women to look after the house and of men to defend it. With respectto children, ‘[it] is the mother who nurtures, and the father who educates’ (p.333). There is no need for the author to add that it is the task of men to be thehead, or master, of the household. This is self-evident and, therefore, taken forgranted: ‘There are four qualities which the head of household must possess indealing with his property’ (p. 339; my italics). The first duty of women is to obeytheir husbands in everything, except what happens inside the house. This is thewill of the Gods (pp. 402f). The duty of men is to care for their wives, in order tosecure their ‘agreement, loyalty and devotion’ (p. 411), or in other words, theirsubmission. In this way, both husband and wife are ‘guardians of the commoninterest’ (p. 413). The Oeconomica ends by imprinting on us the importance ofunity of mind, invoking the support of Ulysses: ‘For, says he, there is no greaterblessing on earth than when husband and wife rule their home in harmony ofmind and will’ (p. 417).Aristotle also wrote about economic matters and mainly about the household.Unlike the author of Oeconomica, however, Aristotle recognised two types ofeconomic activity: household management and money-making (The Politics, 40ff),the former consisting of production of goods for use (use-value), the latter ofexchange of goods in the form of trade (exchange-value). Money-making in theform of commerce and, especially, the taking of interest was deemed unnaturaland morally unacceptable. This Aristotelian view on this matter was adopted bythe Catholic Church and became the ruling dogma throughout the MiddleAges.In Ancient writings on household management, then, there is absolutely nodoubt that the head of the house is a man. 9 It is the duty of the husband tocommand and the duty of his wife, children and slaves to obey. ‘For the male ismore fitted to rule than the female, unless conditions are quite abnormal, andthe elder and fully grown more fitted than the younger and undeveloped’(Aristotle, Politics, pp. 49f). Since it is assumed as self-evident that men are superiorto women in intelligence and virtue, it is also the duty of the husband toeducate both his wife and his children. ‘For the rule of free over slave, male overfemale, man over boy, are all natural’ (p. 52). Ancient writings on economics,then, are not just about household-management, they are also about patriarchy,and as such paradigmatic examples of ‘ideology’ in a Marxist sense.

The new institutional economics 263The idea of economics as the science of household management survivedwell into the eighteenth century. It was held, for instance by Adam Smith’steacher Francis Hutcheson (see Finley, 1973: 17ff). This is really not surprising.The household remained the most important economic organisation for thousandsof years. If we are to believe Max Weber ([1922] 1978: 375–80; 1927:115ff), it remained so until the rise of modern rational capitalism and itremained a locus of patriarchy, resting on traditional authority.By the time of Adam Smith, production had moved more and more from thehousehold to the capitalist enterprise and economics was turned into a theory ofthe market. The focus of attention was still on production, however, whichmeans that the household disappeared from economic analysis, if not fromhistorical reality. In reality, the household had turned increasingly into a unit ofconsumption and it is as such that it reappears in neoclassical economics. Thehousehold is a unit of consumption, represented by a utility function. UnlikeAristotle and Weber, neoclassical economics did not at all concern itself with theinternal organisation of the household. It was conceived of as a unity and withfew exceptions, treated as such until Gary Becker turned the wheel around andreinstated the household as a unit of production in economic theory.One exception was Paul Samuelson (1956), who reacted to the way socialindifference curves are commonly used in economic analysis. The family, forinstance is not a monolith, but a group of different individuals, with differentpreferences (pp. 8–12).Since most ‘individual’ demand is really ‘family’ demand, the argument canbe made that such family demands have been shown to have none of thenice properties of modern consumption theory. However, if within thefamily there can be assumed to take place an optimal reallocation of incomeso as to keep each member’s dollar expenditure of equal ethical worth, thenthere can be derived for the whole family a set of well-behaved indifferencecontours relating the totals of what it consumes: the family can be said toact as if it maximizes such a group preference function.(Samuelson, 1956: 21)The main attempt to open the black box of the household, has been made byGary S. Becker, as part of his programme of ‘economic imperialism’. Thisprogramme is based on the belief that ‘the economic approach provides a valuableunified framework for understanding all human behavior’ (Becker, 1976:14). The heart of the economic approach to human behaviour is the ‘relentless’and ‘unflinching’ use of the ‘combined assumptions of maximizing behavior,market equilibrium and stable preferences’ (p. 5).Becker’s attempt to use the economic approach to analyse the household hasled, not only to theories of marriage and of the family, but also to a new theoryof consumer behaviour. A first element in this new theory is a theory of the allocationof time (Becker 1976: ch. 5). In order to consume, most people have towork and the time spent at work is inversely related to the time spent on

264 The new institutional economicsconsumption. Hence, the necessity to allocate time between them. If economicsis a theory of the use of scarce resources, as some maintain, time must beincluded. For human beings, at least, nothing is so inexorably scarce as time.Running out of money is one thing, running out of time is another, much moredefinite thing.A second element in Becker’s theory of consumption is altruism (Becker1976: chs 12–13). Members of a household care for one another, which impliesthat their utility functions are interdependent, not separate. In particular, parentscare for their children. This means that they share in their children’s joys andderive pleasure from everything that benefits them. They even derive pleasurefrom helping them out when they get into self-inflicted trouble, a phenomenonwhich follows from the ‘rotten kid’ theorem: ‘if a head [of the family] exists, othermembers also are motivated to maximize family income and consumption, even if their welfaredepends on their own consumption alone’ (p. 270). The ‘head’ of a family is thatmember, man or woman, ‘who transfers general purchasing power to all othermembers because he [sic] cares about their welfare’ (p. 253). An advantage withthis way of looking at the household is that it may be represented by one singleutility-function.A third element in Becker’s theory of consumer behaviour is that heconceives of the household as a unit of production; an organisation on a parwith the firm, which may be described by a production function. This does notmean, however, that the modern household is a unit of production in the samesense as the Ancient household. The modern household does not produce ordinaryconsumer goods, but ‘commodities’, like health, longevity, prestige, style, ordistinction, sexual pleasure, etc.An important ingredient in Becker’s theory of the family is the householdproduction function. It is assumed that households use their resources in a waythat maximises this function and from this assumption a number of conclusionsfollow about, for instance, the division of labour and demand for children in thefamily. I will not go further into the many interesting and sometimes controversialresults of Becker’s analysis, but point to one of its most controversialfeatures. Even if Becker, himself, denies it ([1981] 1991: 14), his analysis is biasedby the common utility-function towards consensus, whereas in real life, family lifeis often a hotbed of conflicts (cf. Nelson, 1996: 60ff). Opening the black box ofthe family does not suffice to achieve a fully individualistic analysis of itThe assumption of one single utility-function for the household has been seenas a violation of methodological individualism, and, of course, it is no less problematicthan the use of a social indifference curve. The typical solution to thisproblem is game theory, which is in this respect, at least, more individualistic. Animportant representative of a more individualistic approach to the household isPierre-André Chiappori.THE USUAL APPROACH of household desicion making derives householdbehavior from the maximization of a unique utility function. This ‘neoclassical’formalization, however, can be critizised on the ground that it

The new institutional economics 265somehow contradicts the basic requirements of methodological individualism.Indeed, since the household consists of several members, its behaviorshould be analyzed as the result of several individually rational decisions.That is to say, each member should be characterized by his (her) own utilityfunction; and the ‘collective’ household decisions should be analysed withina formal framework which would model the interactions between members.(Chiappori, 1988: 63)There is already a large literature around Chiappori’s individualistic model ofcollective labour supply in the household, but this is not my business. I notice,though, that many contributions to this literature make a point of its individualism.What I have not noticed, however, is any reference to power andpatriarchy. On the contrary, Chiappori assumes, like Becker, ‘that internal decisionprocesses are cooperative, in the sense that they systematically lead toPareto-efficient outcomes’ (Chiappori, 1992: 438). 10The firmSo too in private industries, the man in authority – bailiff or manager – whocan make the workers keen, industrious and persevering – he is the manwho gives a lift to business and swells the surplus.(Xenophon, 426–354 BC) 11According to Max Weber, modern rational capitalism is characterised, aboveall, by the ubiquity of the modern profit-making enterprise. 12 This type of enterprisefirst arose in the West, as did rational capitalism (Weber [1922] 1978: 165).Its most typical feature is that it engages in profit-making based on capitalaccounting ([1922] 1978: 71–4, 90f). Weber did not say that enterprisesmaximise profit, but he did observe that those enterprises that fail to make aprofit will go out of business (Weber, 1904–5: 54f). This is an echo of Marx’ssuggestion that capitalists are forced to accumulate in order to stay in business,but it also foreshadows the more recent idea of Alchian that market competitionselects those firms that are most rational in their pursuit of profit.Another important feature of the capitalist enterprise, according to Weber, isthat, like the state, it is organised as a bureaucracy. Unlike a market, a bureaucracyis made up of relations of authority (Weber [1922] 1978: 220ff). Webergoes even further and maintains that the model of the capitalist enterprise ismilitary discipline – a particularly effective form of authority ([1922] 1978:1155f). It should be mentioned here that Weber made the important distinctionbetween economic power and authority, as two forms of domination (Herrschaft)([1922] 1978: 941ff). Economic power is ‘domination by virtue of a constellationof interests’ and authority is ‘power to command and duty to obey’ (p. 943). Thefirst form of domination springs from ownership and control of resources, whichcan be used to affect the incentives of other people, who are excluded from the

266 The new institutional economicsuse of these resources. The second form of domination is a right and, therefore,effective in the long run only if legitimate, i.e., if people in subordinate positionsaccept it. The relation between superior and subordinate in a capitalist enterpriseis a relation of authority, even if it is entered into because of an asymmetryof economic power (944f). What the employer buys from the employee is theright to command the latter during his/her hours of work. I mention Weber’sinfluential view of authority because the phenomenon of authority is somethingof a touchstone for an economic theory of organisation.Much has been written about the capitalist enterprise and its developmentsince Weber. One of the most discussed ideas in this literature, is that capitalismin the twentieth century is different from earlier forms of capitalism, because ofthe separation of ownership and management in the modern corporation(Bearle and Means, 1934). This development casts some doubts on the assumption,made in economic theory, that firms maximise profits. While it isreasonable to assume that owners of firms are motivated primarily by profitmaximisation,it is not equally obvious that managers share this interest. Ifmanagers maximise their own utility, this will usually lead to a deviation fromprofit-maximisation.More recently, the sociologist James Coleman has argued that the moderncorporation is the most characteristic feature of our age: indeed, that themodern corporation pervades society and dominates our life. Over the last twohundred years, or so, there has been a great transformation from primordialtypes of social organisation, like the family, to rational organisations, with anentirely different structure. Whereas the former is made up of persons, theelements of the latter are positions, or offices. Behind this transformation lies acertain development of law, which made it possible. Of pivotal importance is theidea of legal, or juristic, persons, which gave rise to the existence of corporateactors. (See Coleman, 1991b; 1993a.)Our knowledge of the firm derives mainly from sociologists, economic historians,business historians and unorthodox economists. For a long time, orthodoxeconomics did not, at all, contribute to this knowledge. Neoclassical economicsused to take the profit-maximising firm as exogenously given and treated it like ablack box. No attention whatsoever was paid to the inside of the firm; to theindividuals that make it up and do the acting. It is largely due to the rise of thenew institutional economics that the firm has become an important topic ofmodern neoclassical economics.The main origin of the neo-institutional theory of the firm is in the pathbreakingarticle by Ronald Coase about ‘The Nature of the Firm’ (1937). It wasnot until the 1960s, however, that it attracted the attention of economists andturned into a classic. I have already mentioned the key idea of transaction coststhat Coase used to explain the emergence of firms on the market and will notengage in unnecessary repetition. I will turn, instead, to the further developmentof the economic theory of the firm.The closest descendant of Coase’s theory of the firm is the transaction costseconomics of Oliver E. Williamson, who also coined the term ‘new institutional

The new institutional economics 267economics’ (1975: 1). 13 According to Williamson, firms, or hierarchies, arise asattempts to economise on transaction costs. In certain circumstances, it becomeseconomically rational to replace external market transactions with internalauthority relations. More precisely, this is the case when (1) bounded rationalityand uncertainty make it difficult to specify future contingencies in a contract andwhen (2) the pairing of opportunism and small-number relations make transactionsrisky. It may be noted that markets and hierarchies only represent twoextreme types of governance structure, with many hybrid forms in between(Williamson, 1975: 106ff; 1979: 245ff; 1993: 86ff). It may also be noted thatthese governance structures are subject to change and, therefore, call for aprocess analysis (Williamson, 1988: 71f 76–88; 1995: 215ff). But these complexitiesare not my concern here.The explanation Williamson gives for the emergence of the firm and, moregenerally, for the existence of different types of governance structure rests oncertain assumptions about human beings and the environment in which they act.There are two behavioural assumptions: (1) bounded rationality and (2) opportunism.14Following Herbert Simon, Williamson replaces the assumption of maximisingbehaviour with that of bounded rationality. People have a limited capacity forcalculation and are, therefore, unable to maximise in most situations. Instead,they do the best they can, something Simon calls ‘satisficing’. People withbounded rationality are still intendedly rational and they do economise on scarceresources. At times, it would even seem that Williamson suggests that satisficing ismaximising (cf. Elster, 1979a: 73f), as when bounded rationality becomes part ofthe optimisation problem involved in the choice of governance structure(Williamson, 1979: 245f), but this is probably not what he intends. Boundedrationality represents a second level of rationality, between the strong level ofmaximisation and the weak level of process rationality, represented by Alchian,and Nelson and Winter. On this second level, people learn by their mistakes andadapt consciously to new circumstances.The second behavioural assumption is that some people are liable to opportunism,which means self-interest with guile. They lie and try to mislead oneanother in various ways, when such behaviour is in their own interest.Opportunism is an important source of uncertainty in economic transactionsand, therefore a main source of these costs, themselves.In addition to the above behavioural assumptions, there are three assumptionsconcerning the situation in which people act. The basic unit of analysis inWilliamson’s transaction costs economics is not the individual, but the transaction.The dimensions used to describe transactions are (1) uncertainty, (2)frequency and (3) asset specificity. 15 High values on these dimensions, creates asituation with high transaction costs in which firms are likely to arise.Uncertainty should not be mistaken for mere complexity and it is not a result ofbounded rationality. Uncertainty as a property of the environment, means that itis impossible to foresee what is going to happen, even if we were ‘lightning calculator[s]of pleasures and pains’ (Veblen, 1898: 389). There are things we simply

268 The new institutional economicscannot know and which no amount of information can reveal the probability of.Whether this depends upon limited knowledge or indeterminism is not important.It is for God to be omniscient. For human beings knowledge will always belimited. Frequency, of course, is a matter of how often partners transact with oneanother. It goes without saying that only frequent transactions are candidates forbecoming internalised in a firm. Asset specificity, finally, means that transactionpartners invest in technology and knowledge that is tailor-made to fit a specifictransaction partner, but which has few, or no, other uses. In this situation, weneed an economics of ‘idiosyncrasy’ and ‘small numbers’. One implication isthat it is costly to switch to other transaction partners and that partners to suchtransactions, therefore, become dependent upon one another. In the extremecase you get a situation of bilateral monopoly.To summarise: according to Williamson, firms, or hierarchies, are likely tooccur because people are boundedly rational and opportunistic and in situationscharacterised by uncertainty, frequency of transactions and asset specificity.It is pretty obvious that Williamson’s transaction costs economics is not anexample of strong methodological individualism. In Markets and Hierarchies (1975)the development of hierarchies takes place in the setting of the market (p. 20).Later, Williamson becomes more concerned to compare different governancestructures. In one way, this means that they become endogenised, since theyappear as the result of the economising behaviour of individuals, but in anotheranalytical role they turn into exogenous variables, or antecedents. The comparisonbetween different governance structures is in terms of their effects. 16 InWilliamson’s most recent writings, the holistic element is even more pronounced.He now appears as an advocate of ‘discrete structural analysis’ (Williamson,1993: 76ff; 1995: 221ff), which is to all intents and purposes a typically holisticenterprise.Coase’s and Williamson’s transaction cost analysis is probably the most influentialeconomic theory of the firm in contemporary social science. One reason itseems congenial to many non-economists is, probably, that it sees firms as hierarchies,made up of relations of authority. This is a familiar view of organisationsin sociology, business administration and political science. But it is not equallycongenial to orthodox economists. Members of the Chicago School, forinstance, have been unwilling to accept this particular argument of Coase andWilliamson. In Chicago, everything must be explained in terms of utilitymaximisationand market equilibrium. In this perspective, there is no place forauthority. From this point of view, it is also dubious to assume that firms engagein profit-maximisation. Following a suggestion by Gary Becker, Armen Alchian(1965) suggested that the assumption of profit-maximisation should be replacedby that of utility-maximisation, together with appropriate constraints on theopportunity set.A further step in the development of an economic theory of the firm wastaken by Alchian and Demsetz in their well-known article ‘Production,Information Costs, and Economic Organization’ (1972). The main thesis of thisarticle is that a firm is, essentially, a long-term contract between employer and

The new institutional economics 269employee, not some relation of authority, or disciplinary power. The relationbetween an employer and an employee is no different from that of a customerand his/her grocer. Like an employer with her employee, the customer cancommand the grocer to do certain things in exchange for money payment (pp.777f). Firms emerge (1) when it is efficient to use team-production and (2) whenit is economical to estimate the marginal productivity of the various inputs (p.783). Of particular importance, in the approach of Alchian and Demsetz, is theproblem of monitoring employees and of metering their productivity. In thewords of Alchian (1979: 251f) alone: ‘a firm is essentially a contractual arrangementfor reducing the costs of detecting and monitoring (adjusting rewardsappropriately) joint production performance’. 17Alchian and Demsetz do not reject the transaction cost approach of Coaseand Williamsson, but they do find it wanting in precision (Alchian and Demsetz,1972: 783f). Alchian, in particular, is sceptical: ‘A host of activities are encompassedby the rubric “transaction costs.”’ Also, ‘[t]he answer, “high transactioncosts” is merely the name for whatever it is that leads to the “firm”’ (Alchian andWoodward, 1987: 110). This is close to suggesting that explanations in terms oftransaction costs are tautological. This is denied, however, by Steven Cheung(1983: 6–9), who attempts a specification of the notion of ‘transaction costs’. Hedoes not, however, defend the idea that firms are characterised by authority.It is not quite correct to say that a ‘firm’ supersedes ‘the market.’ Rather onetype of contract supersedes another type. Coase’s main concern is a type ofcontract under which an input owner surrenders a delimited set of rights touse his input in exchange for income. He is therefore directed by a visiblehand of a price mechanism. It takes remarkable insight to see that as thistype of contract increases there will be fewer product markets.(Cheung, 1983: 10)Demsetz later turned to the much discussed problem of the separationbetween ownership and control in the modern corporation. The main thrust ofhis argument is that the suggested difference of interest between managers andowners is much exaggerated. In his typical Chicago manner, he ‘view[s] theownership structure of the firm as an endogenous outcome of a maximisingprocess in which more is at stake than just accommodating to the shirkingproblem’ (1983: 377). The general problem, addressed by Demsetz, is, of course,that of the relation between principal and agent, which has generated a literatureof its own within the economic theory of the firm. One of the most citedarticles is Michael C. Jensen’s and William H. Meckling’s ‘Theory of the Firm:Managerial Behavior, Agency Costs and Ownership Structure’ (1976), which isof particular interest, since it is a manifestation of methodological individualism.As a point of departure, Jensen and Meckling (1976: 306) observe that thefirm is traditionally treated as a ‘black box’ in economics. Their own theory ofthe firm is based, largely, on the contractual view of Alchian and Demsetz, butthey are more explicitly individualistic: ‘It is important to recognize that most

270 The new institutional economicsorganizations are simply legal fictions which serve as a nexus for a set of contracting relationshipsamong individuals’ (p. 310). From this point of view, there is really nodifference between things that take place within firms and things that take placebetween firms.Viewing the firm as the nexus of a set of contracting relationships amongindividuals also serves to make it clear that the personalization of the firmimplied by asking questions such as ‘what should be the objective function ofthe firm’, or ‘does the firm have a social responsibility’ is seriouslymisleading. The firm is not an individual. It is a legal fiction which serves as afocus for a complex process in which conflicting objectives of individuals(some of whom may ‘represent’ other organizations) are brought into equilibriumwithin a framework of contractual relations. In this sense the‘behavior’ of the firm is like the behavior of a market; i.e., the outcome of acomplex equilibrium process. We seldom fall into the trap of characterizingthe wheat or stock market as an individual, but we often make this error bythinking about organizations as if they were persons with motivations andintentions.( Jensen and Meckling, 1976: 311)The agency relationship is a (contractual) relation between two (or more)parties, such that the agent is hired, or appointed, to act on behalf of the principal.The so-called agency problem is the problem of making the agent act inthe best interest of the principal. The paradigmatic agency relationship is thatbetween employer and employee, but there are many others. It is possible andcommon, for instance, to see the relation between government and citizens as arelation between principal and agent. In economic theory, the problem ofagency is solved by creating the right kind of incentives for the agent, subject toagency costs, of course. In recent agency theory, therefore, the prevailing viewseems to be that the firm is best viewed as an ‘incentive system’ (Holmström andMilgrom, 1994). In other words: ‘Firms are complex mechanisms for coordinatingand motivating individual’s activities’ (Holmström and Roberts, 1998:75). 18There is, at least, one more economic theory of the firm that demands ourattention: the property rights theory advanced by Oliver Hart and his associates.To some extent, this theory represents a return to the ideas of Coase andWilliamson. Thus, transaction costs and incomplete contracts are consideredimportant for the emergence of firms. Because there are such things as transactioncosts and incomplete contracts, it may be advantageous to gain residualrights of control over the assets of suppliers. According to Hart et al., ownershipmeans control, and control is the ability to exclude others from use of an asset.The firm is seen as a set of property rights. The defining characteristic of a firmis ownership of non-human assets, such as machines, inventories, buildings, cash,patents, copyrights, etc. The reason ownership of non-human assets is consideredmost important is that it leads indirectly to control over human assets. 19

The new institutional economics 271An important idea underlying the analysis is that a key right provided byownership is the ability to exclude people from the use of assets. We haveargued that this authority over assets translates into authority over people:an employee will tend to act in the interest of his boss … In the absence ofany nonhuman assets, it is unclear what authority or control means.(Hart and Moore, 1990: 1150)In the issue between Coase and Williamson, on the one hand, and Alchianand Demsetz, on the other, concerning the importance of authority in firms,Hart takes a middle position. On the one hand, ‘Coase’s view that firms arecharacterised by authority relations does not really stand up’ (Hart, 1989: 157).Also a customer in a grocer’s shop can tell the grocer what to do as party to acontract. On the other hand, he agrees with Coase and Williamson that ‘theemployer has much more leverage over his employee than the customer has overhis grocer’ (Hart, 1989: 164). The reason is that much more is at stake for theemployee than for the grocer’s customer. 20 As a sociologist, my own view of thismatter, is that Hart goes too far in reducing authority to control. Authority is notjust control, it is the right to exercise control. I agree with Alchian and Demsetzthat there is an element of control in all contracts, but this element has to dowith rights and obligation, not simply with control. It is closer to the truth to saythat there is an element of authority also in the ownership of physical assets,than to say that there is no authority in relation between employer andemployee.It is common to see the purpose of an economic theory of the firm as that ofsuggesting an answer to the essentialist question: ‘What is a firm?’ (see, e.g.Grossman and Hart, 1986: 691; Hart and Moore, 1990: 1120). Differentanswers have been provided, but few of them very convincing. To suggest that afirm is a nexus of contracts is not very helpful since it fails to tell us in whatrespect a firm differs from a market. And, indeed, this seems to be the mainpurpose in mind of those advocating this view (Alchian and Demsetz, 1972;Jensen and Meckling, 1976). The firm is reduced to a nexus of contractsbetween individuals and, therefore, no different from a market. Or, if there is adifference, it is only a difference of degree. What is gained by this reduction isthat you make the firm amenable to economic analysis, according to the stricturesof methodological individualism. The loss involved in this analysis is that itconflicts with the common sense view that firms are, indeed, different frommarkets. And this common sense view is true, even if there are many forms ofinteraction in between firms and markets.An important observation, in this connection, is that principal-agent models –contrary to the express purpose of Jensen and Meckling to do exactly this – failto open the black box of firms. A typical feature of principal-agent models isthat they treat both the principal and the firm as black boxes. The owners of thefirm are treated as a homogenous group, represented by the abstract figure of aprincipal, whereas employees are represented by the more concrete personage ofthe manager (see Holmström and Tirole, 1989: 87). Principal-agent theory, then,

272 The new institutional economicsdoes not tell us anything about the internal organisation of the firm (Hart, 1989:1759).Another attempt to cast light on the inside of the firm was made byHolmström and Milgrom (1994), who view it as an incentive system. As a definitionof the firm, however, this view is even less helpful than to see the firm as anexus of contracts. Incentives are everywhere and, in no way specific to firms. Ifyou want to know what a firm is, therefore, you have to look elsewhere for ananswer. This is not to say, of course, that incentives are not important in firms.To provide employees with an efficient system of incentives is vitally importantfor every firm, and more so if the environment is highly competitive.To suggest that a firm is a set of property rights is potentially more illuminating,even if the problem of distinctiveness remains. Not only firms haveproperty rights in physical assets. The idea that ownership of physical assets maylead to authority over human assets seems to be a step in the right direction. Theoverwhelming majority of firms own physical assets and control human assets. Ifwe add that they are engaged in production of goods and services for themarket, we have probably mentioned the most essential characteristics of firms.The stateThe state emerged earlier in history than did the firm, or modern corporation.According to Max Weber ([1922] 1978: 327ff, 336f, 1094ff; 1927: 338ff), onceagain, this is not a mere accident, but a ‘necessary’ chronological order, since arational state administration is a precondition for the development of modern,rational, capitalism. Rational capitalism cannot work without some predictability,and in modern society only the state can guarantee the order that makes forpredictability. Similar thoughts have been advanced, more recently, by DouglasNorth, who is one of a few neoclassical economists, to have developed a theoryof the origin of the state. 21The state never figured prominently in mainstream economics. For a longtime, the state was treated as exogenously given and ignored. One exception wasthe theory of public finance, another welfare economics. In the new institutionaleconomics, however, more attention is given to the state, if not always explicitly.Most interest in the state is shown by Public Choice theory, but it is implicit invarious theories about parts of the state, such as government and bureaucracy.Before getting into the economic theory of the state, it may be useful to makea distinction between (1) theories of the origin of the state, (2) theories of theworking of the state and (3) normative theories of the proper tasks of the state.Normative theories are not my concern, but it is not always easy to separatethe positive and normative in economics. Type 1 and type 2 theories often mergewith type 3 theories. The obvious example of a merger between 1 and 2 is thetheory of the social contract (Hardin, 1997: 33). The merger of 2 and 3 is oftena consequence of the concept of ‘efficiency’, as used in economics.I think it is possible to say something about methodological individualismalready at the outset. My hypothesis is that type 1 and type 3 theories tend to be

The new institutional economics 273individualistic, whereas type 2 theories are almost necessarily institutionalistic.After all, to analyse the working of the state presupposes that its exists as an institution.I am not going to support this hypothesis by any survey of economictheories of the state. I will resort, instead, to the less reliable method of case study.My case is James Buchanan, who is both one of the most explicit methodologicalindividualists and one of the main theorists of the state, among economists.James Buchanan started his career, as an economist, in the field of publicfinance. This is a type 2 theory of the state, which tends to be institutionalistic,with intellectual roots in the organic theory of the state. As I suggested at thebeginning of chapter 3, methodological individualism is largely a response to thistheory, and James Buchanan is no exception. Already in his first writings, he tooka stand, for an individualistic, and against an organismic, theory of the state. Inthe first, ‘the state is represented as the sum of its individual members acting in acollective capacity’ (Buchanan, 1949: 496). 22 ‘The state has no ends other thanthose of its individual members and is not a separate decision-making unit. Statedecisions are, in the final analysis, the collective decision of individuals’ (p. 498).This view of the individual as the basic unit of all decision-making, lies at theheart of Buchanan’s methodological individualism in all his writings. 23 In TheCalculus of Consent, the individualistic theory of the state is stated like this:Having rejected the organic conception of the State and also the idea ofclass domination, we are left with a purely individualist conception of thestate. Collective action is viewed as the action of individuals when theychoose to accomplish purposes collectively rather than individually, and thegovernment is seen as nothing more than the set of processes, the machine,which allows such collective action to take place. This approach makes thestate into something that is constructed by men, an artifact.(Buchanan and Tullock [1962] 1965: 13)Buchanan’s methodological individualism is most clearly visible in his theoryof the social contract, which exhibits the typical ambiguity with respect to thepositive/normative distinction. It is my feeling that the positive element isstrongest in Buchanan’s earlier works, but vanishes altogether in his constitutionaleconomics (see p. 281). In The Limits of Liberty (1975), which is not exactlyan early work, there is an individualistic theory of the origin of the state, whichis a version of the theory of the social contract. ‘Almost by definition, theeconomist who shifts his attention to political process while retaining his methodologicalindividualism must be contractarian’ (1989a: 82).Buchanan is not only a methodological individualist, however, he is also aninstitutionalist and this element is there almost from the beginning. Already inthe early days of the Virginia School, around 1960, ‘[t]here came to be anincreasing awareness of the importance of the institutional setting and of institutionalconstraints for the operation of an economy’ (Buchanan, 1986: 11). Thisinstitutionalism is stated most explicitly and emphatically by Buchanan in thearticles included in What Should Economists Do? (1979b):

274 The new institutional economicsThe essential subject matter for the economist consists of human behaviourin social institutions, not of human behavior in the abstract. The tendencyof economic theorists to overlook this simple fact provoked the reaction ofthe American institutionalists, a reaction which surely was misguided in itsemphasis, but which, nonetheless, pointed to a serious deficiency in theevolution of economic science.(Buchanan, 1979b: 134)Buchanan’s institutionalism, then, differs both from that of mainstream theoreticaleconomics and from the old institutionalists. It differs also ‘from that whichthe habitual political scientist adopts’.As I have tried to indicate, the shift in thinking is a simple one. It involvesonly the shift from the organizational entity as the unit to the individual-inthe-organization.Instead of trying to examine the institutions of politics asorganizations, the whole approach involves trying to examine the interactionsamong individuals as they carry out assigned roles within theseinstitutions.(Buchanan, 1979b: 157)This passage might equally well have been written by a traditional sociologist.‘Role’ is not a concept normally used by economists, nor is the statement that‘Institutions matter’ (Buchanan, 1979b: 282), even if Buchanan invokes AdamSmith to support this claim in economic theory. Also the following statementlooks more like a piece of orthodox sociology than of orthodox economics.The institutions (economic, geological, legal, political, social, technological),which define the sizes of community within which an individual findshimself, impose external bounds on possible behavior. Parallel to theseexternal constraints there are also internal limits or bounds on what we maycall an individual’s moral-ethical community.(Buchanan, 1979b: 225)What finally distinguishes Buchanan from a traditional sociologist is rationalchoice and methodological individualism. In his view, institutionalism is notincompatible with methodological individualism. On the contrary: ‘If we adoptthe methodological imperative that all choice analysis be reduced to inquiry intoindividual behavior, the importance of institutional constraints becomes evident.The feasible choice options open to the individual are, in part, determined bythe institutional setting’ (1989b: 42). Stated even more emphatically: ‘rationalchoice precepts dictate totally different behavior in divergent institutionalsettings’ (p. 43). To repeat, ‘methodological individualism, does not imply orrequire that individual choice behavior is invariant over changes in the institutionalsettings’ (1989b: 38). ‘[T]here is no difficulty at all in analyzing individualchoice behavior under differing institutional settings, and in predicting how these

The new institutional economics 275varying settings will influence the outcomes of the interaction process’ (1988:105).It is pretty obvious, I believe, that Buchanan is an institutional individualist, inthe sense of Joseph Agassi and Lawrence Boland. Like Marshall’s comparativestatics and Popper’s situational analysis, Buchanan’s version of rational choicerelies on ‘situational determinism’ (Latsis, 1972; 1976) in the form of institutionalism.24The most typical social institution, however, is not the organisation of thestate, but the rules it creates, enforces, and, above all, obeys (Brennan andBuchanan, 1985). Actually the two belong very much together. According toNorth (1981: 21), for instance, ‘[o]ne cannot develop a useful analysis of the statedivorced from property right’. The most important rules, of all, are those laiddown in the constitution. For this reason, Buchanan has created a new branch ofeconomics, which he calls ‘constitutional economics’, and which is his most originalcontribution to economic theory. Rules, of various sorts, including those ofconstitutions, are the topic of the next section. Before I turn to this topic,however, a brief note on another type of organisation.Interest groupsInterest groups seem to be fitting objects for an economic theory of groups, ororganisations. Common wisdom seems to suggest that interest groups are voluntaryorganisations created by individuals in order to be able, better to further thecommon interest. This fits well with Ludwig von Mises’s individualistic view that‘society is concerted action, or cooperation’ for a common purpose (Mises [1979]1966: 143). Somewhat paradoxically, it is also a theory that the economistMancur Olson ascribes to sociologists and political scientists, and a theory he setsout to punctuate. According to Olson’s famous argument, rational, self-interested,individuals will not pursue collective interests, unless rewarded, for doingso, or punished, for not doing so (Olson, [1965] 1971).I have treated interest groups, at great length in an earlier book, and littlewould be gained from entering that interesting topic here. Precious little hasbeen said about methodological individualism in connection with interestgroups. 25 Like the rest of the new institutional economics, theories of interestgroups treat social institutions either as endogenous or exogenous variables. Bothpatterns are richly represented in the literature. Economic theories of collectiveaction, including that of Mancur Olson, typically conform to the first pattern.Public choice theories of rent-seeking, and similar theories, such as GaryBecker’s theory of pressure groups (Becker, 1983; 1985), treat interest groups asexogenous variables and conform to the second pattern. 26ConclusionI have looked at the theories of organisation put forward within the new institutionaleconomics. It is pretty obvious, even from a cursory glance, that social

276 The new institutional economicsinstitutions are treated both as endogenous and exogenous variables. It is myimpression, though, that economists differ in their reaction to this fact. For some,exogenous institutions are a problem, since they violate the requirement of strictmethodological individualism. This attitude is common, if not ubiquitous, in theeconomic theories of the family and the firm. For others, such as OliverWilliamson and James Buchanan, exogenous social institutions are not aproblem, but rather a solution to many problems in economics. It might beargued, however, that no theory of social organisation can be entirely free fromexogenous social institutions. If we believe, for instance, that no social organisationis likely to emerge in an environment totally free from social rules, it followsthat all theories of social organisation, will include at least one type of socialinstitution among its exogenous variables, or conditions, namely social rules.Social rulesThere are many kinds of social rules, from the informal codes of etiquette, ormanners, to the formal rules of written law. Economists are, for obvious reasons,most interested in those rules that regulate market transactions, viz., the closelyrelated laws of property and contract. But not exclusively. There is also a moregeneral theory of ‘law and economics’ and a theory of the most fundamental ofall laws: the constitution.Property rightsThe neo-institutional theory of property rights started with Armen A. Alchian’sarticle ‘Some Economics of Property Rights’ (1961), which was an attempt toanalyse and compare different types of property rights. This project wascontinued by Harold Demsetz (1964; 1966; 1969), who followed Coase in criticisingwelfare economics for seeing all negative externalities as a loss of welfarethat should be remedied. The mistake of this ‘nirvana approach’ is to comparethe existing market economy with an ideal that does not exist and could notexist. A more reasonable approach would be to compare different existing institutionalarrangements, with different property rights and different transactioncosts.The dominating concern in these early contributions to the theory of propertyrights is to compare different institutional arrangements with respect to theireffects. This is institutional individualism and, as such, wanting from the perspectiveof the more orthodox methodological individualism, prevailing in Chicagoat that time. Milton Friedman, who was one of the dominating figures inChicago at that time, expresses this view in the following manner: ‘On ourapproach these [institutional conditions] too, are to be regarded as resultants ofan economic equilibrating process, not as physical data’ (Friedman, 1956: 13).In line with this view, there is the early attempt by Demsetz (1967), to explainalso the emergence of property rights. Demsetz’s thesis is that ‘property rightsdevelop to internalize externalities when the gains of internalization become

The new institutional economics 277larger than the costs’ (p. 350). This happens when new technology and newmarkets create a change in economic values. New property rights, then, seem tobe the result of rational choice. But actually Demsetz takes an agnostic attitude:I do not mean to assert or to deny that the adjustments in property rightswhich take place need be the result of a conscious endeavour to cope withexternality problems. These adjustments in property have arisen in Westernsocieties largely as a result of gradual changes in social mores and incommon law precedents. At each step of this adjustment process, it isunlikely that externalities per se were consciously related to the issue beingresolved.(Demsetz, 1967: 350)Another attempt to explain the emergence of property rights was made bySteven N. S. Cheung (1969). His particular example was the choice of contractualarrangement between tenants and landowners in Asian agriculture and,especially in pre-communist China. His main thesis was that the particular formof these contracts is a function of two variables: transaction costs and risks.Parties to such contracts choose arrangements that minimise transaction costsand disperse risks between them. Cheung, admits, however, that ‘some level oflaw enforcement by legal authorities is taken for granted’ (p. 41). This is institutionalindividualism, once again. 27A third attempt to provide an economic explanation of the emergence ofproperty rights was made by Douglass C. North and Robert Paul Thomas. Theirproject was an ambitious one; amounting to nothing less than trying to explainThe Rise of the Western World (1973; see also 1970; 1971). The analytical tools theyused were those developed by the new institutional economics, and especially bySteven Cheung. Their main argument was that efficient property rights, ratherthan new technology, was the main factor behind the rise of the West. Theirstory, although fairly simple, is nevertheless too complex to be told here. Onepart of it is the change in the relative prices of land and labour, which resulted inthe development of private property in land and free labour. Another part is theinvention of new institutions, such as the joint-stock company, forms of insuranceand international financial markets, which made it possible to economiseon transaction costs. This, latter, part was a necessary condition for the emergenceof market society.The explanations suggested by Demsetz, Cheung, and North and Thomasmay be interpreted as neoclassical attempts to endogenise property rights(Pejovich, 1972: 310), as suggested by Milton Friedman above. As such, they mayalso be interpreted as manifestations of methodological individualism. I haveshown, in the cases of Demsetz and Cheung, that things are not quite thatsimple. But what about North and Thomas?In their general model of institutional change, North and Thomas make useof a distinction between two types, or levels, of institutions:

278 The new institutional economicsThere are the fundamental institutions that specify the basic ‘ground rules’such as the underlying ‘constitutional’ basis of property rights and basicdecision rules with respect to political decision-making, and then there arethe secondary institutional arrangements which may be created withoutaltering the basic institutions.(North and Thomas, 1970: 10; see also Davis and North, 1970: 133)It is clear from this quotation that an explanation of secondary institutionalarrangements must always treat the fundamental institutions, either as exogenousvariables, or as exogenous givens. Such explanations, therefore, are examples ofinstitutional individualism, manifest or latent. But what about the explanation offundamental institutions? In North’s and Thomas’s explanation of the growth ofthe Western World, the decisive change in the ground rules was a result ofcumulative changes in the secondary institutions:The essential key to our essay is that cumulative changes in secondary institutionsultimately led to a restructuring of fundamental institutions inWestern Europe over this period, in such a way that individuals in thesociety were now encouraged to strive for greater productivity in theirundertakings.(North and Thomas, 1970: 10)I conclude that also in the case of North and Thomas endogenisation of someinstitutions is achieved only by treating other institutions as exogenous variables.In the early 1970s, the property rights paradigm had become established andit was time to summarise the results of the growing literature. According toAlchian and Demsetz (1972: 17) three questions have been suggested: ‘What isthe structure of property rights in society at some point in time? (2) What consequencesfor social interaction flow from a particular structure of property rights?and, (3) How has this property right structure come into being?’. Of these, thefirst question has no obvious relevance for the issue of methodological individualism,the second question implies institutional individualism and the thirdquestion may be answered either in a strictly individualistic, or in an institutionalistway. It depends on the nature of the exogenous variables; whether theycomprise only facts about psychic states and natural environment, or whetherthey include also facts about institutions other than those explained.LawProperty rights are a species of the genera law. They are extremely important, ofcourse – especially for economic activity – but, nevertheless, are just one type oflaw among others. The economic theory of property rights may be conceived ofas a special case of a more general economic theory of law. Such a theory hasbeen developed by Richard A. Posner, among others. Posner, himself, calls thisbranch of theory ‘Law and Economics’, as distinguished from ‘New Institutional

The new institutional economics 279Economics’ (Posner, 1993). His argument is that there is a considerable overlap,but not an identity, between the two. His main reason for seeing a difference isthat both Coase and Williamson seem to reject the assumption of utilitymaximisation.This may be, but I still believe that both Coase’s and Williamson’stheories belong in the broad category of rational choice theories and, therefore,that the similarities are more important than the differences. Since I also believethat the main branch of the new institutional economics; what Eggertsson hascalled ‘Neoinstitutional Economics’, is best understood as a rational choicetheory of social institutions, I conceive of them as really belonging in the samecategory as Posner’s economic theory of law, which also relies on a ‘redefinitionof economics as rational choice’ (Posner, 1981: 3).Posner’s most systematic study, Economic Analysis of Law (1972), covers a vastground and there is no possibility of going into the details of his analysis. Myway out is to concentrate on the approach, which is the same, irrespective ofsubject matter. Posner’s approach is essentially the economic approach of GaryBecker, but owes a lot also to Milton Friedman. ‘Economics, the science ofhuman choice in a world in which resources are limited in relation to humanwants, explores and tests the implications of assuming that man is a rationalmaximizer of his ends in life, his satisfactions – what we call his “self-interest”’(Posner [1972] 1977: 3). This approach is applied to a large number of legalphenomena, all of which are ‘shown’ to be economically efficient for society as awhole. Posner’s reasoning comes out clearly in his discussion of the morality ofcommon law:The view of the common law as a system for promoting economic law as asystem for promoting economic efficiency will strike many readers as anincomplete – if not severally impoverished – theory of common law, particularlyin its disregard of the moral dimension of law. Surely, it will beargued, the true purpose of law, especially of those fundamental principlesembodied in the common law of England and the USA, is to correct injusticesand thereby vindicate the moral sense.In fact, there appears to be no fundamental inconsistency betweenmorality and efficiency. Moral principles – honesty, truthfulness, trustworthiness(for example, keeping promises), selflessness (for example considerationfor others), charity, neighbourliness, avoidance of negligence and coercion –serve in general to promote efficiency. That such principles have survived …for thousands of years suggests they are traits that by and large enrich ratherthan impoverish the society that cultivates them.(Posner [1972] 1977: 185)What immediately strikes the eye, in this quotation, is that Posner seems tohave jumped from individual utility-maximisation to social efficiency as the mainexplanatory principle. It may be added that the word ‘jump’, in this case, isparticularly fitting, since Posner leaves out the steps that would take us from theformer to the latter. It is not at all clear how individual utility-maximisation leads

280 The new institutional economicsto social efficiency. The only clue is Posner’s suggestion that moral principleshave survived because they are efficient. But, as several commentators haveobserved (Field, 1981: 184f, 196; Granovetter, 1993: 22ff), this is a functionalexplanation and, as such, scientifically dubious, unless supplemented by anaccount of the mechanism that make functional (efficient) principles survive. 28Posner’s Economic Analysis of Law is notorious for its absence of suchmechanisms. 29In his well-known article ‘A Theory of Primitive Society, with SpecialReference to Law’ (1980) – also included as chapters 6 and 7 in his second mostimportant work, The Economics of Justice (1981) – Posner’s functionalism is evenmore explicit, but he also goes some way to meet the criticism directed at thisdoctrine. Posner closes his discussion of primitive law in the book by asking twoquestions: ‘First, if it is true, as I have argued, that the legal and other socialinstitutions of primitive society are economically rational or efficient, whatmechanism drives primitive society to this surprising result?’ (p. 204). No lesssurprising than this result is the note Posner attaches to his question: ‘I emphasizeonce again that, in suggesting that primitive people are economicallyrational, I am not making any statement about their conscious states. Rationalbehavior to an economist is a matter of consequences rather than intentions andin that respect resembles the concept of functionality in traditional anthropology’(p. 204, note 72; cf. 1980: 5). This is a controversial statement, to say theleast, since rational choice is usually conceived of as a form of intentional explanation.If consciousness and intentions are not involved, the most obviousalternative is selection, and this is presumably what Posner really has in mind:efficient institutions survive because they are best adapted to the environment.One reason selection is a possible mechanism is that primitive societies and theirenvironment are so stable. When the rate of change is slow societies have time todevelop efficient adaptations (1980: 53; 1981: 205).The second question, Posner addresses is this: what factors are treated asexogenous (not possible to change) and what factors are treated as endogenous(possible to change)? He admits that in his analysis, both of common law and ofprimitive law, a background of other institutions is treated as exogenous. Theefficiency of the endogenously adapting institutions, therefore, is relative to theexogenously fixed background institutions (Posner, 1981: 205f). More important,for my purposes, is that Posner, thereby, seems to relinquish any pretensions tobeing a strong methodological individualist. The question is, whether he is amethodological individualist at all?For my purposes, the most important point is that Posner’s analysis seems tobe, neither rational choice, nor methodological individualism. Rational choice isa form of intentional action, but Posner renounces the assumption of consciousintention on the part of actors. In its stead, he talks about adaptation, butwithout specifying the mechanism at work. In the absence of such a specification,Posner’s analysis ends in an untenable form of functionalism: socialinstitutions are said to survive, because they are efficient for society as a whole.But absent a mechanism of selection, or some other form of adaptation, such

The new institutional economics 281functional explanations presuppose the existence of a social maximiser and thisis, of course, holism, not methodological individualism.ConstitutionsPosner’s treatment of law in ‘primitive society’ was part of an attempt to explain‘The Origins of Justice’, which is the title of Part Two of The Economics of Justice.A possible motive for such an endeavour would be to make a clean sweep ofexogenous institutions, in order to satisfy the demands of strict methodologicalindividualism. As we have seen, this does not seem to be Posner’s motive, butthere are others.To seek the historic origin, not of this or that institution, but of all social institutions,is one way to achieve a complete endogenisation. Another route is morelogical and consists in the attempt to endogenise the logically and constitutivelymost fundamental social institution; that social institution upon which all otherinstitutions ultimately rest, the ‘ground rules’ of North and Thomas. More oftenthan not, these routes coincide: the most fundamental social institution is alsoassumed to be the first. This is the typical procedure in the theory of the originalsocial contract.Another name for the most fundamental social institution is the constitution ofa society. I will end my treatment of the economic theory of property rights andother laws with a mention of constitutional economics – a branch of economics,which is the creation of James Buchanan.According to Buchanan (1990: 12–15), constitutional economics, like the restof economics, is based on methodological individualism. As I have argued in anearlier work Udehn (1996: 176–84), however, James Buchanan’s constitutionaleconomics is really a normative theory, and the methodological individualism itis based on is really ethical individualism. Sine this interpretation finds supportin Buchanan, himself (1989b: 62), I assume that it is correct, and since this is awork on methodological individualism, I leave Buchanan’s constitutionaleconomics to the political philosophers (cf. Brennan and Hamlin, 1995).Evolutionary economicsIn the early stages of social science, especially anthropology and sociology,evolutionism was among the most important theoretical currents, but holismwas the rule and individualism the exception. Also economics had its brand ofevolutionism; the institutional economics of Thorstein Veblen, which washolistic. But also neoclassical economics had some advocates of a more evolutionaryapproach (Langlois, 1986a). The main example is the AustrianEconomics of Carl Menger, but there was also Joseph Schumpeter and AlfredMarshall. Among the later Austrians, it is mainly Friedrich von Hayek, whohas maintained the evolutionary legacy of Menger. In the 1950s evolutionismalso popped up in the utilitarian Chicago School of Economics. I havealready mentioned Armen Alchian’s famous article ‘Information, Uncertainty

282 The new institutional economicsand the Allocation of Resources’ from 1950. The main impetus to a newevolutionary economics, however, was given by Richard R. Nelson and SidneyG. Winter, with the publication of An Evolutionary Theory of Economic Change in1982. 30Is the evolutionary branch of institutional economics individualistic? Not ifmethodological individualism is equated with rational choice. As we have seen(p. 257), the Chicago version of evolutionary economics replaced rational choiceby natural selection. Since I do not accept this equation of methodological individualismwith rational choice (see p. 288), the question remains and it raisesanother question: What does methodological individualism mean in the case ofevolutionary theory?There is an obvious answer to this question: in order to comply with the principleof methodological individualism, the unit of adaptation, or selection, mustbe the human individual. But this answer is not sufficient. It is also necessary, Ibelieve, that the environment is made up of nature and other individuals, but notof social collectives and/or wholes. The second criterion is, necessary, in order toexclude functionalism. If a social institution, develops and survives, because itcontributes to the social system as a whole, we have clearly to do with a case ofsocial holism, not of methodological individualism.If we consider the case of Hayek in the light of these considerations, wereach the conclusion that his evolutionary theory is not in accordance withmethodological individualism. I have argued earlier (Udehn 1987: 207f) thatHayek moved away from methodological individualism, in a more holistic direction,from the 1950s and onwards. The first signs of this change of direction is inan article ‘On Degrees of Explanation’ from 1955 (in Hayek, 1967). In thisarticle, Hayek argues that the social sciences differ from physics, in that they dealwith complex phenomena, which permit only of ‘explanation of the principle’.Examples of theories providing such explanations, according to Hayek, areGeneral Systems Theory and Cybernetics. I do not want to suggest guilt by association,but since these theories are generally considered holistic (Phillips, 1976:ch. 4; Udehn, 1987: 89f), we might at least suspect a move in the direction ofholism.In ‘The Theory of Complex Phenomena’ (1964), Hayek goes further in theholistic direction, by arguing that social science has to deal with persistent structures,made up of systematically connected elements. This makes it necessary torefer to the relative position of these elements, when explaining their behaviour.Hayek’s holism is most clearly expressed in ‘Notes on the Evolution ofSystems of Rules of Conduct’ (1967), and in his last book: The Fatal Conceit(1988), where it is part of a theory of cultural evolution. Hayek makes a distinctionbetween systems of rules which govern the conduct of individuals and ‘theorder or pattern of actions which results from this for the group as a whole’(1967: 66). There is no holism implied in this distinction, but, then, Hayek goeson to suggest that we must reconstruct the overall order before we shall be ableto discover the functions of the rules of individual conduct and also that the

The new institutional economics 283overall order of a group as a whole interacts with the environment (p. 70). This isbeginning to smack of holism. 31The overall order of actions in a group is in two respects more than thetotality of regularities observable in the actions of the individuals andcannot be wholly reduced to them. It is so not only in the trivial sense inwhich a whole is more than the sum of its parts but presupposes also thatthese elements are related to each other in a particular manner. It is morealso because the existence of the whole cannot be accounted for wholly bythe interaction of the parts, but only by the interaction with an outsideworld both of the individual and of the whole.(Hayek, 1967: 70)Hayek also defends a kind of sociologism, in the form of accepting social normsand social orders as exogenously given, against various individualistic attempts toexplain their emergence. The alternative explanation he proposes is typicallyfunctionalist.This implies a sort of inversion of the relation between cause and effect inthe sense that the structures possessing a kind of order will exist because theelements do what is necessary to secure the persistence of that order. The‘final cause’ or ‘purpose’, i.e., the adaptation of the parts to the requirementsof the whole, becomes a necessary part of the explanation whystructures of the kind exist: we are bound to explain the fact that theelements behave in a certain way by the circumstance that this sort ofconduct is most likely to preserve the whole – on the preservation of whichdepends the preservation of the individuals, which would therefore not existif they did not behave in this manner.(Hayek, 1967: 77)This is full-blown functionalism of the most objectionable type, at least fromthe point of view of methodological individualism: to explain the emergenceand continued existence of social institutions in terms of their contribution tothe preservation of the social whole. As the methodological individualist J.W.N.Watkins has pointed out (1976), this type of functionalist explanation is vacuouswithout the specification of some mechanism of selection. Another methodologicalindividualist, Victor Vanberg (1986: 81–9), has pointed out that Hayek’stheory of cultural evolution is based on group selection and, therefore, incompatiblewith methodological individualism (see also Sugden, 1993: 397–403).Jack J. Vromen (1995) has defended Hayek against the charge of holism, byarguing that his unit of selection is social order, rather than the group. His argumentrests on the following presupposition: ‘One order can be replaced byanother in various ways, many of which are in accordance with the individualisticnotion that individuals, and not groups of individuals, are responsible forthe emergence, maintenance and alteration of rules’ (p. 172). Vromen believes

284 The new institutional economicsthat a reinterpretation of Hayek’s theory of cultural evolution along the lines ofthis truism – or tautology? – can make it consistent with his methodological individualism(p. 173), but I do not think so.Vromen goes on to suggest that the problem with Hayek’s evolutionism is notgroup selection, but his organicism (p. 174), as if the two were independent. Healso suggests that Hayek’s organicism derives from that of Menger, which was ofa holistic and functional type. As we have seen, this is, indeed, the case withHayek’s organicist evolutionism, and this is the reason it conflicts with hismethodological individualism, but Menger is much more critical of organicismand he does contrast it with the atomistic approach, which is the source ofHayek’s methodological individualism (Menger [1883] 1963: 122ff). The conclusionis that the conflict between Hayek’s cultural evolutionism and hismethodological individualism remains.It is arguable that the evolutionism of Carl Menger was more individualisticthan that of Hayek. At least, he did more to specify the mechanism at work inhis main example: the evolution of money. This example, I have suggested (pp.90f ), is a paradigm of an individualist (invisible-hand) explanation of the emergenceof a social institution. It is commonly agreed that money is a convention,in the sense of David Hume and David Lewis (1969), that is, a social institution,which is arbitrary, but in the interest of all. Contemporary social scientists havethe advantage, over Menger, of being able to use the tool of game theory toexplicate and explain social conventions, as games of co-ordination. Importantcontributions, to the game theoretic literature on the emergence of social institutionsinclude Ullmann-Margalit (1977), Schotter (1981), Axelrod (1984) andSugden (1986). The evolutionary use of game theory raises the same questionsabout methodological individualism, as does game theory, generally. Since I havediscussed game theory in the previous chapter, I leave the matter with repeatingmy conclusion: the explanatory use of game theory in social science will typicallyimply exogenously given social institutions. ‘For evolutionary game theory, socialstructure comes in the form of the organization or social environment that doesthe selecting of strategies’ (Kinkaid, 1997: 138).What about the Chicago branch of evolutionary economics? Does it obey thestrictures of (strict) methdological individualism? The answer is simple and short:it does not. The unit of selection is the firm, not the individual.In his classical article ‘Uncertainty, Evolution, and Economic Theory’ (1950),Armen A. Alchian argued that economics do not need the dubious assumptionthat firms are engaged in profit-maximisation. In a world of pervasive uncertaintyand incomplete information, it is really impossible to maximise. Instead,firms adapt, imitate and use trial and error, in order to survive. And those firmsthat succeed in making a profit are the survivors. Rationality is replaced by selection.This argument would prove to be of crucial importance for the emergenceof the new institutional economics. First of all, because it was eventually pickedup by its evolutionary branch, but also because it led to rethinking profitmaximisation.Most Chicago economists accepted Alchian’s critique of profit-maximisation

The new institutional economics 285and his emphasis on incomplete information, but suggested various ways toretain the assumption of maximisation (Demsetz, 1996). Milton Friedman (1953:22) absorbed ‘natural selection’ in his own approach of maximisation, with thehelp of instrumentalism. Natural selection makes firms behave as if theymaximised profit. George Stigler (1961) agreed that information is incompleteand costly to acquire, but included these costs in the maximisation problem.Individuals will search information to the point where these costs equal the benefitsderived from this information. Stigler obviously does not reckon withuncertainty, in the more fundamental sense of Frank Knight (see p. 121). GaryBecker, at first, also wanted to do without maximisation (1962), but eventuallydecided that the economic approach is based on the assumptions of maximisingbehaviour, market equilibrium and stable preferences (1976: 5). In the end, mostChicago economists, including Alchian himself (see Alchian and Allen [1964]1967: 14), settled for global utility-maximisation, rather than selection. Even iffirms do not maximise profit, it might still be the case that its employeesmaximise their own utility. The reason for this, I believe, is methodological individualism.Whereas profit-maximisation is associated with firms, utilitymaximisationis associated with individuals.A third important contribution to evolutionary neoinstitutional economics isAn Evolutionary Theory of Economic Change (1982), by R.R. Nelson and S.G. Winter,which draws on both the Austrian and Chicago contributions, but most of all onSchumpeter. In this book, it is neither the group, nor the firm, which is the unitof selection, but habits, or routines. It was argued by Winter already in 1964 thatthe Chicago version of evolutionism lacks a counterpart of the gene; a mechanismthat preserves and reproduces successful behaviour. In Nelson and Winter(1982), routines are launched as the social analogue to the biological gene. As Ihave already mentioned, this is a break with methodological individualism, ifthis principle is equated with rational choice. Another possible break withmethodological individualism is that Nelson and Winter try to build amacrotheory of ‘industry behavior, as contrasted with individual firm behavior’(p. 36). They do recognise the need for microfoundations, however, and this iswhere the routines come in. So, what are routines?Routines, according to Nelson and Winter (1982: 96ff), are capabilities oforganisations and, as such, contrasted with skills, which are capabilities of individuals.Nelson and Winter make a distinction between individuals andorganisations, which is typical of sociological holism. It is rather clear, I believe,that they conceive of routines as social facts, in a sense similar to that ofDurkheim (see pp. 34–6), even if they are not sociological reductionists, like thelatter. ‘To provide a plausible account of the relations between the capabilities ofan organization and the capabilities of individual organization members, givingboth the “reductionist” and the “holistic” viewpoints their due, is a majorconceptual undertaking – and one that orthodoxy has not seriously attempted’(p. 63). I believe that this declaration is a fair description of what Nelson andWinter actually do, except that there is more emphasis on the “holistic” viewpoint.To me this suggests the following conclusion: if the evolutionary theory of

286 The new institutional economicsNelson and Winter is individualistic at all, it represents a very weak form ofmethodological individualism, where the institutional and structural elementspredominate. As they themselves suggest, this is something the (methodologicallyindividualist) orthodoxy has never attempted.More recently, Ulrich Witt (1987) has tried to build an evolutionaryeconomics on methodological individualism, but there are signs indicating thathe has, since his first attempt in this direction, become more sceptical about thepossibility of an individualistic evolutionary economics, (see, e.g. 1991: 92).Another evolutionary economist, Geoffrey Hodgson (1988: ch. 3; 1998: 175–7),claims that the new institutional economics is inspired by methodological individualism,but fails to follow its precept. The attempt to explain all socialinstitutions in terms of individuals, leads to an infinite regress (see, also Udehn,1987: 227–31). This critique, which derives from Karl Popper (see p. 204), hitsonly the strong, psychologistic, version of methodological individualism, not theweak version of institutional individualism.ConclusionThe new institutional economics is Janus-faced. It includes both an attempt toendogenise all social institutions and a practice of treating social institutions asexogenously given to analysis. It is possible to conceive of the new insitutionalismin economics as a return to the type of social theory that existed before the riseof neoclassical economics. The attempt to endogenise all social institutionsusually takes the form of some version of the theory of the social contract, whilethose who accept social institutions as exogenous variables and conditions arecloser to the Scottish Enlightenment and the classical political economy ofAdam Smith (see chapter 2).Figure 9.1 Social institutions as endogenous variablesThe attempt to achieve a complete endogenisation of social institutions,however, seems to be a Sisyphus work. The endogenisation of one institution isachieved only by treating other institutions as exogenously given, and there is noend to it (cf. Wisdom, pp. 224–6). This predicament of those who adopt thepsychologistic version of methodological individualism is avoided by those whoadopt Popper’s institutionalism, or the position called ‘institutional individualism’

The new institutional economics 287by Joseph Agassi and Lawrence Boland. It is my impression that the latter positionis the more common among new institutionalists in economics, but this isjust an impression.Figure 9.2 Social institutions as endogenous and exogenous variables

10 Rational choiceindividualismAmong the more noteworthy developments in recent social science is theupsurge of a theory and/or methodology called ‘rational choice’. This wouldnot be my business, however, were it not for the fact that rational choice islaunched as a form of methodological individualism. Sometimes it is eventreated as the only form of methodological individualism (Janssen, 1993: 26).This view, which is most common among economists (Sproule-Jones, 1984: 169),is not adopted in this book. As I hope is evident from previous chapters of thisbook, there are other forms of methodological individualism than rationalchoice (Van Parijs, 1981: 302–4; Elster, 1989a: 105: Vanberg, 1994: 7).In an earlier treatise on methodological individualism (Udehn, 1987), I evenexpressed doubts concerning the purity of rational choice, or ‘situational logic’,as a form of methodological individualism. It seemed to me that rational choiceis much more permissive concerning the explanatory use of social institutionsand social structure than is the original principle of methodological individualism.My first response to this situation was to deny that rational choice isnecessarily individualistic. It all depends upon the entities used in a rationalchoice explanation. If social institutions, social structure, or some other type ofsocial entity, enters the explanandum, or, better, the antecedent, of a rationalchoice explanation, then it is not compatible with the principle of methodologicalindividualism. I did suggest a second route, however, in the form of adistinction between a strong and a weak version of methodological individualism.Today I prefer this second route and conceive of rational choice theory as oneform of methodological individualism. My reason for this, is that this view of thematter seems to have emerged victorious, at least among political scientists, sociologistsand analytical Marxists. Indeed, rational choice stands out as the mostvital form of explicit methodological individualism in contemporary socialscience. It should be noted at the outset, however, that not all agree that rationalchoice is necessarily a form of methodological individualism (Farmer, 1982; Satzand Ferejohn, 1994), not even all rational choice theorists, themselves do so, aswe saw in chapter 8.The first, and still the main, field of application of rational choice outsideeconomics is in political science. I have dealt with this field of application, tosome extent in the previous chapter and, at great length, in Udehn (1996). Since

Rational choice individualism 289I prefer not to repeat myself more than necessary, I will be extremely brief here.I do not think any injustice is done to political science, however, by brevity sinceit is my impression that the big issue in political science has been rational choice,not methodological individualism. Most political scientists seem to take it prettymuch for granted that rational choice is individualistic, but without going intomuch detail about its exact meaning and intent (cf. Ward, 1995: 79). It may alsobe noticed that rational choice entered political science mainly in the form ofgame theory, which I have already discussed in chapter 8.The first really comprehensive use of game theory in political science was, byWilliam H. Riker and Peter C. Ordeshook, in An Introduction to Positive PoliticalTheory (1973). The basic assumption of their approach was that of rationalityand, as a corollary, that of individualism. ‘What we insist upon … is that collectivities,regularities about people in them, and the common goals and values ofcollectivities can be understood only by understanding the individual personswho make up the collectivities. And in this sense our method is individualistic’ (p.37).In Game Theory and Political Theory (1986), Ordeshook begins with two assumptions:‘methodological individualism and purposeful action. Methodologicalindividualism holds that we can understand social processes and outcomes interms of people’s preferences and choices’ (p. 1). This means that groups do notact, except by means of their members. ‘The assumption of methodologicalindividualism is but a reminder that only people choose, prefer, share goals, learnand so on, and that all explanations and descriptions of group action, if they aretheoretically sound, ultimately must be understandable in terms of individualchoice’ (p. 1). Six years later, Ordeshook (1992) has replaced the assumption ofpurposeful action with the much more narrow assumption of ‘individual actionmotivated by self-interest’ (p. 175). According to Ordeshook’s interpretation ofrational choice, this means that ‘political institutions are the product of the selfinterestof those who establish them’ (p. 177). It does not seem, however, as ifOrdeshook suggests that all institutions must be treated as endogenous, since healso mentions the constraints they impose on human behaviour (pp. 186f, 191f).William H. Riker is much more clear on this point. In a paper delivered at thePublic Choice Society (Riker, 1988), he makes a forceful plea for taking socialinstitutions much more seriously than economists usually do. He is especiallydissatisfied with the Chicago economists mechanical use of utility-maximisation,but also with the work of James Buchanan. To take social institutions seriously,according to Riker, is among other things to see their role in the selection ofequilibria. ‘What particular one [equilibrium] is arrived at … is mainly a matterof the constraint imposed by institutional arrangements, which are, therefore, asimportant for outcomes as the desires, tastes, and values of the participants (p.252).I believe that most political scientists share Riker’s view of the importance ofsocial institutions. The work of Kenneth Shepsle, for instance, is based on theassumption that ‘institutions do matter’ (1989: 131). It is my impression thatmost uses of rational choice in political science are based on this assumption and

290 Rational choice individualismfor a very good reason. Political action typically takes place in an institutionalcontext of well-defined legal rules.On the basis of this extremely brief survey, I draw the unwarranted conclusion,that most rational choice individualists, among political scientists, areinstitutional individualists in the sense of Agassi and Boland; they see thenecessity of treating social institutions as exogenous variables in their models ofpolitical activity. A different view seems to be taken by Robert Grafstein (1991:262ff), who argues that Shepsle’s institutionalism, despite appearances to thecontrary, fails to go beyond the state of nature, typical of social contracttheory.Because methodological individualism is hard to give up, many theorists, aswe have seen, find that the best way to accommodate this effect is to pack allthe institutional details, the regularities of behavior variously called norms,or rules, into the heads of the participants. Acting on this knowledge, participantscontinually reinforce the institution – that is, reproduce thebehavioral patterns associated with it. The institution as such does nothing.(Grafstein, 1991: 266)It is clear that by ‘methodological individualism’ Grafstein means psychologisticindividualism and also that he believes that rational choice institutionalism inpolitical science exemplifies this type of individualism, rather than the institutionalindividualism of Agassi and Boland.This interpretation of methodological individualism can be found also in Satzand Ferejohn (1994), who draw the conclusion that rational choice is not necessarilyindividualistic. Having argued that rational choice derives most of itsexplanatory power from a consideration of the social environment of the actingindividuals, they suggest a marriage between rational choice and structuralism,which is not individualistic, or reductive.A different route has been taken by some sociologists, who share the belief ofSatz and Ferejohn in the fertility of joining rational choice and structuralism, butwho nevertheless conceive of this approach as a version of methodological individualism.Obviously it is very different from the original version ofmethodological individualism, since it includes an important holistic element inthe form of social structure. It has been called structural individualism. 1Rational choice sociologyThere has been much more resistance to rational choice in sociology than inpolitical science. The reason for this is, no doubt, that sociology emerged partlyas a reaction to rationalism in social theory. The roots of sociology are mainly inRomanticism and most of the early sociologists were critical of the theory of thesocial contract, of utilitarianism and of economic theory. Although not all ofthem. Herbert Spencer was an individualist and rational choice plays some partin his sociology, even though he is known to sociologists mainly as a founder of

Rational choice individualism 291functionalism. Some may think of Vilfredo Pareto, who was both an economistand a sociologist, and certainly a promoter of rational choice. But not in sociology.Pareto is actually the source of the common idea that economics is aboutrational action and sociology about irrational action (Samuelson, 1947: 90).The real pioneer of rational choice in sociology was Max Weber. We havealready seen, in chapter 4, that Weber wanted to import methodological individualismfrom economics to sociology. We have also seen (pp. 95–7) that Weberwas something of a pioneer of rational choice in economics. It remains to pointout that Weber intended to assign a central role to rational choice also in sociology.It is true that instrumentally rational action is only one of four types ofaction in Weber’s sociology – the others are value-rational, affectual and traditionalaction – but it is the most important, and for two reasons: first, thehistorical development of the West has been such that instrumentally rationalaction is now the dominant type of action in our culture. Second, as I havealready indicated, instrumentally rational action is methodologically primary.When trying to understand an action, we should always start by assuming thatwe have to do with rational action. Only if we fail to account for an action byassuming that it is instrumentally rational, should we turn to other types ofexplanations: that it is due to emotion, tradition or intrinsic values (Weber [1922]1978: 4ff).Weber’s rational choice sociology was no success story. Talcott Parsons at firstwanted to build sociology on a theory of action, but emphasised values morethan instrumental rationality, and soon turned in the direction of systems theory.Another neglected attempt to create a rational choice sociology was made by theeconomist John C. Harsanyi, who also recognised it as a form of methodologicalindividualism and raised it against Parsons’s functionalist sociology. The road tosuccess, according to Harsanyi, is the use of game theory together with his own,new, concept of ‘rationality’, which yields determinate solutions for all gamesituations and, therefore, for most social situations. Social institutions and practices,for instance, can be explained in terms of the balance of power amongindividuals (Harsanyi, 1962; 1966b).Harsanyi recognised two opposite theoretical approaches in the socialsciences: the relationistic approach, which is individualistic, and the functionalisticapproach, which is collectivistic. Since Harsanyi favoured the former, he setout to show its superiority to the latter. He chose to demonstrate its superioritywith respect to social phenomena, such as social status, social values and socialinstitutions (Harsanyi, 1966a: 357f; 1968: 305f).Social status is, besides economic gain, the most important incentive andmotivating force of social behaviour. A person has high social status in a group ifall or most other members of the group show deferential behaviour towards him.The other members of the group show deferential behaviour towards a certainmember of the group if they attach special importance to his activities withinthe group. Social status, thus, is a form of power relationship; as such, it dependsupon the ability to influence other people through rewards and/or punishments.The ultimate explanation of social status, according to Harsanyi, is in terms of a

292 Rational choice individualismperson’s ability to deal out rewards and punishments (1966a: 359–69; 1968:316–21).Social values play an important role in the functionalist and conformistapproach represented by Talcott Parsons. According to Harsanyi, the mostimportant defect with this approach is that it cannot explain social change. Bycontrast, the rationalistic approach can account for both stability and change.Social values, their stability and change, are explained, in all essentials, in termsof the self-interest of individuals. When it serves the self-interest of a sufficientlylarge number of individuals to change their values, and when the pressure forconformity is not too strong, then there will be a change in social values(Harsanyi, 1969: 513–15, 526–32).Like social values, social institutions are explained by Harsanyi in terms of anincentive for people to change their behaviour ‘[A]s a result of certain changesin the society or in its natural environment or in its relations with its externalsocial environment, some people have decided that their interests would bebetter served by a new type of institutional arrangement’ (Harsanyi, 1969: 532).Accompanying this individualist explanation of social institutions, Harsanyiproposes an individualist concept of ‘social function’. The ‘social function’ of aninstitution is defined as ‘all the benefits that various individual members ofsociety derives from its operation’ (p. 532). The individualist concept of ‘socialfunction’ is contrasted with the collectivist concept, according to which socialinstitutions are explained in terms of their contribution to the maintenance ofthe social system as a whole. 2One particularly clear example of the use of a collectivistic concept of functionis the theory of social stratification suggested by Davis and Moore. In theirwell-known (among sociologists, at least) article ‘Some Principles ofStratification’ (1945), they explained the universal fact of social stratification interms of the functional necessity ‘faced by any society of placing and motivatingindividuals in the social structure. As a functioning mechanism a society mustsomehow distribute its members in social positions and induce them to performthe duties of these positions’ (p. 243).Harsanyi feels that there is something suspect about this explanation of socialstatus in terms of the functional needs of society as a whole (Harsanyi, 1966;1968), and so do I. His own explanation of stratification is unfortunately not animprovement upon that of Davis and Moore. By blurring their importantdistinction between social position and the status attached to it, Harsanyi endsup with a circular explanation of social status in terms of social status. Thereason is probably that Harsanyi as an economist and methodological individualistlacks a concept of social structure. 3 Such a concept was introduced inrational choice sociology by the sociologist James Coleman.James ColemanJames S. Coleman was, until his premature death in 1995, the undisputed leaderof rational choice sociology. At the end of his career, he also proclaimed himself

Rational choice individualism 293a methodological individualist, but he was always careful to point out that hiswas a ‘form’ of methodological individualism, implying that there are otherforms as well. I think this caution was very wise, because Coleman’s methodologicalindividualism is of a peculiar kind, far removed from the original version ofthis doctrine (cf. Heckathorn, 1997: 8f). It is different, not only from the originalstrong version of methodological individualism, but also from Joseph Agassi’sweak version of institutional individualism.When Coleman (1994b) looks back upon his life as a sociologist, it appears –to the reader, if not to himself – as a constant attempt to unite holism and individualism.In the 1950s, he tells us, sociology was split in two: first there wastheoretical sociology, which conceived of society as a system. On this holistic sideof the split was also the tradition of community studies, which used the communityas the unit of analysis. On the other side of the split, there was the surveymethod of empiricist sociology, almost exclusively focusing on the individual asthe unit of analysis. In Coleman’s earliest vision of sociology, the unit of analysiswas the social system but it should be investigated with the quantitative methodsof empiricist sociology, rather than with the largely qualitative methods ofcommunity studies (pp. 30f). At this time Coleman was a holist, conceiving ofhimself as a ‘Durkheimian’ (1990d: 49; 1994b: 33). It was only some years laterthat Coleman detected rational choice and came to see it as the solution to theproblem of analysing social systems in terms of individuals (cf. Swedberg, 1996:315–7). In addition to rational choice, however, there has remained a holist sideto Coleman’s work, in the form of a consistent structuralism. It is my thesis inthis section, that Coleman never argued that the behaviour of systems could beexplained in terms of individuals alone. Individuals are usually implicated insocial structures and sociology can only advance by paying attention to this fact.Coleman’s vision of sociology is really a combination of rational choice andstructuralism.The first clear expression of Coleman’s turn to rational choice individualismis probably in the article on ‘Collective Decisions’ from 1964. Inspired by GeorgeHomans’s individualist theory of exchange, he complains that ‘sociologists havecharacteristically taken as their starting-point a social system in which normsexist, and individuals are largely governed by those norms’ (p. 166). The mainflaw with this procedure, according to Coleman, is that it leaves traditional sociologicaltheory unable to answer what should be its central problem; the problemposed by Thomas Hobbes: Why is there not a war of all against all?In this paper, I will proceed in precisely the opposite fashion to that taken bythe advocates of homo sociologicus. I will make an opposite error, but onewhich may prove more fruitful. I want to begin the development of a theoryof collective decisions, and in so doing I will start with an image of man aswholly free: unsocialized, entirely self-interested, not constrained by normsof a system, but only rationally calculating to further his own interest. Thisis much the image of man held by economists, and with it the economistshave answered one part of Hobbes’s question: how is it that although the

294 Rational choice individualismmen who make it up are wholly self-interested, the economic system canoperate with one man’s actions benefiting others. It was the genius of AdamSmith to pose an answer to this part of Hobbes’s question.(Coleman, 1964: 167)Coleman’s own answer to Hobbes’s question is a theory of exchange,inspired, I believe, not only by Adam Smith and George Homans, but also by thepublic choice theory of James Buchanan and Gordon Tullock. Like the latter, heuses game theory to analyse a legislative game of vote trading. Like the latter, healso insists upon self-interest, rather than rationality, as the essence of economicman. Since then, Coleman remained convinced that assuming self-interest is aproper point of departure for social theory (see, e.g., Coleman, 1990c: 31f). 4Since then, Coleman also used to maintain that the central problem of sociologywas to explain social phenomena in terms of individuals, rather than theother way around. His own efforts to solve this problem led him to develop twodifferent, but related, theories: first, there is his general and individualistic theoryof social systems. Second, there is his individualistic explanation of the emergenceof rights and social norms.The basic elements of Coleman’s original theory of social systems are actorsand resources, or events. Actors have interests in these resources and events andcontrol over some of them. The basic theoretical ideas are (1) that individualsare self-interested and act so as to maximise the utility they derive from resourcesand events and (2) that in order to maximise this utility, they exchange controlover resources and events with each other. 5Coleman’s theory of social systems is obviously a theory of exchange, borrowedfrom economics and public choice and, like the latter, it is individualistic. Indeed, itshould be more individualistic, since economics is usually satisfied to treat firmsand households as acting units. ‘For most purposes in sociology, we cannot assumepurposive acting units at a level above the individual’ (Coleman, 1965: 105). Thefundamental unit in Coleman’s analysis of systems of exchange is the acting individual,or person. Acting units larger than single persons are constructed entitiesand can only be justified if it is possible to show that the individuals composing thelarger unit have unitary goals. The analysis of ‘already-formed units of socialorganization’, according to Coleman, ‘must begin with persons, and move up fromthere, or if, in an application, it begins at a level above persons, it must be ultimatelyanalyzable into relations among persons’ (1975: 85f).In the early versions of Coleman’s theory of exchange, individuals have defacto control over resources and events, but no de jure control, or legal rights. Theymay also have control, or power, over other individuals, but no rights tocommand, or authority. Nor are there any social norms to regulate exchangebetween individuals. As we have seen, Coleman’s basic urge was to rid sociologyof explanations in terms of norms and other social entities. Eventually rightsand norms, and their derivatives authority and social capital, become importantelements in Coleman’s social theory, but he retains the individualist inclination toexplain their emergence and maintenance, rather than accept them as given to

Rational choice individualism 295analysis. ‘The question for rational action theory is why and how does a normarise and how is it maintained’. 6 The same goes for rights and authority relations:‘Just as a theory of rational action cannot take social norms as given, eventhough norms are found in all social systems, a theory of rational action cannottake authority relations as given, even though authority is found in all socialsystems’ (Coleman, 1992a: 142).It is my impression, however, that Coleman no longer believes in the possibilityof making a clean sweep. His explanation of the emergence andmaintenance of norms is not in terms of individuals alone. The emergence ofsocial norms is subject to two conditions, each necessary and both sufficient.‘Both conditions may be described as socio-structural’ (Coleman, 1990c: 241). 7The first condition is the existence of ‘externalities of an action which cannot beovercome by simple transactions that would put control of the action in thehands of those experiencing the externalities’ (Coleman, 1990c: 251). Thesecond condition ‘is that under which the second-order free-rider problem willbe overcome by rational holders of a norm’. ‘This condition depends on theexistence of social relations among beneficiaries’ (p. 273). A similar observationconcerns the modern paradigm of methodological individualism: the economicanalysis of market exchange. According to Coleman (1984: 85), economic transactionsand institutions depend upon the existence of trust between actors andsince ‘trust is a relation between two actors’, it depends ‘not simply on the averagelevel of trust, but on the social organization of trust’. The interesting consequence ofthis, for a rational choice theory of trust, is ‘the combination it involves ofrational action theory and structural theory’ (Coleman, 1992a: 147). 8I have presented the individualist side of Coleman’s work and turn now to hisholist side; his structuralism. As I have already indicated, this theme is there fromthe beginning, but it is most pronounced in two of his books: Power and the Structureof Society (1974) and The Asymmetric Society (1982). 9 In these books, Coleman stressesthe importance, for modern society, of the development of a new type of socialactor: states, corporations, trade unions, parties, etc. The emergence of theseactors onto the scene, has led to a complete transformation of social structure. Ofspecial importance is their status as ‘legal’, or‘juristic’ persons, with rights andduties of their own, which makes it possible to concentrate a previously unseenamount of power in these fictitious persons. A basic theme of Coleman’s books isthat society, or the social system, must be conceived of as being made up of twoelements: natural persons and corporate actors. Thus, we arrive at threedifferent kinds of relations: (1) between natural persons, (2) between corporateactors and (3) between natural persons and corporate actors (Coleman, 1974:87ff). The third type of relation, Coleman calls, asymmetric, since it involvesactors, or persons, of different types and unequal in size and power (Coleman,1982: 19ff). A corollary of the distinction between natural persons and corporateactors is that we must make a distinction between the former and the positionsthey occupy in corporate actors (1974: 36, 49; 1982: 14f).Now, these distinctions were never made by a methodological individualistbefore – only by methodological holists – and they provide the main reason for

296 Rational choice individualismarguing that Coleman’s version of methodological individualism is different fromall earlier versions. 10 The main difference is that Coleman admits that corporateactors have interests of their own, interests which are distinct from the interestsof the natural persons, who occupy positions in the corporate actors. Thus,‘corporate actors are, in their actions, motivated towards purposes of their own –very often purely growth – for which membership benefits are viewed merely asconstraints’ (Coleman, 1974: 29). If this is so, social structure enters rationalchoice explanations, not only as constraints upon action, but directly as a determinantof actors’ interests. It is a matter of social structure, not of personalaptitude, that households maximise utility, firms maximise profit, politiciansvotes and bureaucrats budgets – if this is, indeed, what they do.Coleman’s structuralism was not a temporary aberration from the individualisttrack, but a permanent part of his social theory. The arguments he putforward in Power and the Structure of Society and in The Asymmetric Society, arerestated with equal force in Foundations of Social Theory (1990c: parts III and IV)and in his Presidential Address to the American Sociological Association in 1992(Coleman, 1993a). In the latter (1993a: 7f), 11 Coleman argues that positions, oroffices, occupied by persons, not persons themselves, are the elements of thesocial structure of purposively constructed corporate actors. Therefore, relationswithin corporate actors, are between positions, not persons. This has alwaysbeen the central idea of social holism (see Udehn, 1987: 83–97).The bifurcation of Coleman’s sociology into an individualistic and a structuralisticpart is not my own fabrication. It is how Coleman, himself, describeshis approach from an early date: ‘Thus the general characteristics of thisapproach are two: a model of a purposive actor … and a structural system ofactivities in which the actors are embedded’ (1965: 103). This approach reflectssociety itself:Individuals constitute one set of elements, each a purposive acting entity,each occupying positions or roles. The organization of roles, that is, thepurposive social organization, is another type of element in society, withmany of the same properties as actors that the individual has.(Coleman, 1971: 74)At the end of his career, Coleman still describes his approach in terms of twocomponents: ‘One is rational action … and the other is social structure whichprovides the constraints, incentives, and contexts of action that bring about thetransitions between micro and macro levels’. In order to distinguish his ownapproach from other approaches, he calls it the ‘rational-structural approach’(Coleman, 1992a: 134).Because of his structuralist view of the social system, Coleman has becomemore critical of economic theory than he seemed in his programmatic 1964article on ‘Collective Decisions’. Twenty years later, he is still maintaining that thecentral problem of social science is that of ‘moving from a model of individualbehavior to a theory of the behavior of a system composed of those individuals’

Rational choice individualism 297(Coleman, 1984: 86). But now he also maintains that this problem can only besolved by ‘taking social organization explicitly into account in making this transition,rather than assuming it away’ (p. 86). The problem with economics (andwith much quantitative sociology) is that it ignores social structure. This is a pity,because introducing more social structure is the most promising avenue to scientificprogress in economics. Rational choice sociology borrows some elementsfrom neoclassical economics, but differs from the latter by taking social organizationand social institutions seriously. ‘Neoclassical economics would eitherignore such social structure altogether, or treat it as endogenous, while rationalchoice theory in sociology does not hesitate to regard prior social structure asexogenous in analysis of system functioning’ (Coleman, 1994a: 170).The hallmark of rational choice theory in sociology is the combination ofan assumption of rationality on the part of individuals, but replacement ofthe assumption of a perfect market with social structure, sometimesregarded as endogenous and other times regarded as exogenous, whichcarries individual actions into systemic outcomes.(Coleman, 1994a: 167)According to his friend and college Gary Becker (1996: 378), ‘Jim makes amajor advance’, relative to economics, ‘by incorporating social structure into thetheory’. He also observes, however, that ‘Jim was not content simply to takesocial structure as given. He recognized that it was desirable to try to build upthe structure from the interactions among the choices of individuals and otheractors’.I believe I have shown conclusively that there is a structuralist, as well as anindividualist, side to Coleman’s sociology. I will now turn to his methodologicalindividualism and try to show that it is a new version of this principle, differentfrom all earlier versions, including the institutional individualism of Agassi andBoland.Coleman’s methodological individualism is explicitly stated for the first timein a programmatic article in American Journal of Sociology in 1986. In the abstractof this article, he says that his programme for social research has two essentialelements:The first is the use of a theory of purposive action as a foundation for socialtheory; this entails acceptance of a form of methodological individualismand rejection of holism. The second is a focus in social research and theoryon the movement from the level of individual actions to macrosocial functioning,that is, the level of system behavior.(Coleman, 1986c: 1309)A first thing to notice is that Coleman refers to his methodological individualismas a ‘form’ of this doctrine, implying that there are other forms as well. Asecond thing is that his methodological individualism uses a theory of purposive

298 Rational choice individualismaction as the ‘foundation’ for social theory. Other expressions, used by Coleman,are that his form of methodological individualism ‘grounds social theory in atheory of individual action’ (p. 1309) and that ‘(p)urposive action of individualscan be taken as a starting point by sociologists’ (p. 1312). 12 None of them rulesout the use of institutions and social structure as determinants of action and,therefore, as co-determinants of system behaviour. A third thing worthy of attentionis that Coleman’s foundation is a theory of ‘purposive action’. This is indistinction to methodological individualism based on other theories of action, orbehaviour (cf. Coleman, 1979: 76). A fourth observation is that Coleman claimsthat his form of methodological individualism implies a rejection of holism. Thisis a highly debatable assertion, stated with no argument at all to support it. Imaintain, to the contrary, that it is no more the case that all forms of methodologicalholism excludes reference to purposive action, than it is the case that allforms of methodological individualism excludes reference to institutions andsocial structure. Both positions are about equally contentious. A fifth pointconcerns the second element of Coleman’s programme. I suggest that the focuson the movement from individual behaviour to macro-social functioning at thelevel of system behaviour is fully in line with the original strong version ofmethodological individualism, but it is not common among methodological individualiststo talk about ‘macrosocial functioning’ and ‘system behaviour’.Coleman returns to methodological individualism in Foundations of SocialTheory (1990c) and in an interview with Coleman by Richard Swedberg in 1987,but published the same year as the former work (Coleman, 1990d). In Foundations,Coleman is even more cautious, and his methodological individualism less restrictive,than in the article from 1986. He repeats that he is suggesting a form, or‘variant’, of methodological individualism, but adds that it is ‘a special variant’.No assumption is made that the explanation of systemic behavior consists ofnothing more than individual actions and orientations, taken in aggregate.The interaction among individuals is seen to result in emergent phenomenaat the system level, that is, phenomena that were neither intended norpredicted by the individuals. Furthermore, there is no implication that for agiven purpose an explanation must be taken all the way to the individuallevel to be satisfactory. The criterion is instead pragmatic: the explanation issatisfactory if it is useful for the particular kinds of intervention for which itis intended. This criterion will ordinarily require an explanation that goesbelow the level of the system as a whole, but not necessarily one groundedin individual actions and orientations.(Coleman, 1990c: 5)Essentially the same view of methodological individualism emerges in theinterview. Coleman now takes the somewhat odd view that ‘methodological individualismcan work at more than one level’. Thus, ‘for some purposes, one couldcarry out an analysis of a system of action among firms as actors, without goingdown to the individuals within the firm … for some investigations, one would be,

Rational choice individualism 299I think, justified in taking corporate actors as having all the properties of anactor’. Coleman is aware, of course, that this is not how methodological individualismis usually conceived: ‘True methodological individualism takes naturalpersons – the actions of natural persons – as the only starting point and looks atthe system of action that occurs among these’ (Coleman, 1990d: 50).The most noteworthy element in Coleman’s methodological individualism, asstated in the above texts, is that it does not demand that we go down to the levelof individual action when explaining the functioning of social systems. Anotherstrange element is the idea that ‘interaction among individuals is seen to result inemergent phenomena at the system level’ (Coleman, 1990c: 5). This has alwaysbeen a central tenet in the metaphysics of holism, but alien to all forms ofatomism, including methodological individualism (see Udehn, 1987: 83–97). Itcould be, of course, that Coleman conceives of ‘emergence’ in some unorthodoxway, but I do not think so. He maintains that rights and norms are supraindividualsocial entities, with an existence of their own, and irreducible to thelevel of individuals. 13 This is social holism, pure and simple.I have found two other references to ‘methodological individualism’ inColeman’s work. The first is in an article on ‘The Economic Approach toSociology’ (1992a) and the second in ‘A Rational Choice Perspective onEconomic Sociology’ (1994a). In the first, the economic approach to sociology isdefined by a form of methodological individualism, stated, somewhat vaguely, asbeing characterised by the fact that ‘explanation involves the actions of individualactors’ (p. 133). ‘Its central defining property is that a foundation ofrational action underlies all theoretical work’ (p. 134). This is in sharp contrast tothe methodological holism of functionalism in sociology, which ‘introduces teleologyat the level of social systems, implicitly treating the system as an actoracting purposively’ (p. 125 ; see also Coleman, 1992a: 135). 14In the second article, nothing new is added to his earlier explications, exceptthat he now states methodological individualism explicitly in terms of his wellknowndiagrammatic representation of the micro–macro relation in socialtheory (1994a), also introduced in Coleman (1986a), in his programmatic 1986article in The American Journal of Sociology (1986c) and in Foundations of Social Theory(1990c). 15Figure 10.1 Coleman’s micro–macro schemeSource: Coleman (1986a: 347; 1986c: 1321f; 1990c: 5–10; 1994a: 166f)

300 Rational choice individualismWith the help of this diagram, Coleman identifies four types of explanations,or relations between variables, occurring in social theory. In type 2 relations,psychic states and/or individual behaviour are explained in terms of social structureor some other macro-social entity. In type 1 relations, actions of individualsare explained in terms of some psychic state; a drive, motive, or intention. Intype 3 relations, the behaviour of the social system is explained in terms of theactions of individuals. Type 4 relations, finally, connect one macro-social state, orentity, with another. This relation, according to Coleman, represents methodologicalholism (1986c: 1321). Methodological individualism is represented bythe remaining relations 2, 1 and 3, or macro–micro–macro (1986c: 1322; 1994a:166f).Figure 10.2 Methodological individualism according to ColemanSource: Coleman (1986c: 1322; 1994a: 166f)It is obvious, I believe, that Coleman’s version of methodological individualismis not the psychologistic version. Type 2 relations imply that socialphenomena may be part of the explanans and used as exogenous variables.Coleman’s scheme is, in fact, well suited to make the distinction between twodistinct types of methodological individualism, which I have called the strong andweak versions of this principle (see pp. 195–9, 227).Before Karl Popper’s plea for institutionalism in social science, there was onlythe strong version of methodological individualism. This version is representedby type 1 and type 2 relations in Figure 10.2. According to the strong version ofmethodological individualism, methodological holism is represented by both type2 and type 4 relations. With Karl Popper and his followers, however, came theso-called institutional individualism, which is, no doubt, a weak version ofmethodological individualism. It is significant that Hedström and Swedberg(1996: 297) call their first type a situational mechanism after Popper’s situationalanalysis, even if they do not seem to adopt his institutionalism.In one of his last writings, Coleman (1993b) takes the opportunity to clarifyand somewhat amend his earlier position, as stated most fully in Foundations. In areply to the structuralist sociologist Peter Blau, who criticised his analysis of votetrading, Coleman says that he agrees that social structure provides the incentivesthat leads to vote trading.

Rational choice individualism 301But there is another question: Where does this structure, which provides theincentives that lead to vote-trading, come from? Obviously, from prior interdependentactions of individuals and corporate actors, whose interests andactions were shaped by the incentives and constraints provided by the socialand political structure existing at that time. Thus structure at one time(macro-level) generates the conditions which together with existing interestsshape the actions of actors (micro-level) that jointly procuce [sic] outcomeswhich modify the structure of a later time (macro-level) which generatesconditions that again (through constraints and incentives) shape actions(micro-level) that jointly produce outcomes (macro-level) and so on.(Coleman, 1993b: 63)The novelty in this quotation is that Coleman adds a time dimension to hisearlier structural individualism. In his own words, it ‘illuminates the complementarityof structure and action, as well as the alternation in time between themicro–macro and macro–micro’ (p. 63). Coleman’s original diagram can now beextended to take the following form (cf. Abell, 1996: 261):Figure 10.3 Coleman’s micro–macro scheme extended in timeWhat this diagram shows is a never-ending dialectic of structure and rationalaction, which is similar to some other ‘dualistic’ approaches to the relationbetween individual action and social structure. 16 The question is: What part doinstitutions and social structure play in the explanation of social phenomena? Ifwe look for an answer in the writings of Coleman, we will find quite a number ofdistinct ways in which ‘society’ explains, first, the behaviour of individual and,second, the behaviour of social systems.As a first step, we may distinguish between an internal, or subjective, and anexternal, or objective, effect. Every individual living in society is socialised andinternalises, in varying degrees, the values and beliefs prevailing in the society, orgroup, to which he/she belongs. In addition, individuals almost always confront‘society’ as an external constraint upon their actions at the moment of action.

302 Rational choice individualismSome sociologists, such as Weber, see society, entirely as a subjective reality,whereas others, such as Durkheim, insist that society is objective and external toindividuals. Coleman recognises both aspects of society, and criticises the earlyParsons’s approach to institutions, for being too subjective. ‘By refusing to defineinstitutions in terms of structures, Parsons is not rejecting the structural or objectiveorientation to institutions, but wants, rather, to develop a theory based onthe relation of the individual to social structure’ (Coleman, 1990a: 334).According to Coleman (pp. 337f), however, the important thing to be explainedis the emergence of institutional structures; concrete structures of relationsbetween individuals or, rather, between social positions (see above).These external constraints are conceived of, by the strong version of methodologicalindividualism, as simply the actions and reactions of other individuals,but Coleman does not follow suit. He conceives of social structure as a distinctelement in the explanation of social phenomena. The role of this element isobscured by the diachronic form of his diagram and also by the above quotationfrom Coleman. It might seem as if social structure comes into play only in type 2relations and only the actions of individuals in type 3 relations. This is not thecase. Social structure is at work also in type 3 relations. This is most clearly seenin Coleman’s discussions of economic theory. 17 When ‘Introducing SocialStructure into Economic Analysis’ (1984), Coleman argues for ‘taking socialorganization explicitly into account’ when making the transition ‘from a modelof individual behavior to a theory of the behavior composed of those individuals’(p. 86). This is clearly a type 3 relation. The same argument reappears in ‘ARational Choice Perspective on Economic Sociology’ (1994a). Recapitulating therelations in his well-known diagram, he now maintains that the third relationconsists of ‘the combination of the actions of individuals, in some institutionalstructure, to bring about systemic outcomes’ (Coleman, 1994a: 166).Both the institutions through which the micro-to-macro link takes place, andthose through which the macro-to-micro link takes place, may be taken asexogenous in rational choice theory, in studying the effects of particularinstitutional structures on individual actions or on systemic outcomes(Coleman, 1994a: 171)Coleman’s most explicit statement to this effect is in his reply to some criticalremarks by the structuralist sociologist Peter Blau:the ‘interdependence’ of this third step is something which Blau slides over,while it is essential to the theory. The interdependence consists of the particularinstitutional structure within which the actions must take place, theinstitutional structure that consists of constraints and incentives. It is onekind of institutional structure in an economic market, another in theAmerican Congress, still another in the British parliament, and another inbureaucratic organization.(Coleman, 1993b: 63)

Rational choice individualism 303Thus, there is no doubt that Coleman saw institutional structure as part also oftype 3 relations. A more complete form of the Coleman diagram, therefore,would be the following:Figure 10.4 Social structure as a determinant of individual actionThe reply to Blau is interesting, because Coleman makes some amendmentsto his earlier position. From his 1964 article on ‘Collective Decisions’ to hisFoundations, Coleman consistently maintained that it is more important to explainmicro-to-macro, or type 3, relations than to explain macro-to-micro, or type 2,relations (1990c: 6; 1990b: 35). His main reasons for this is that this is the mostdifficult and the most neglected problem in social science, or, at least, in sociology(Coleman, 1986c: 1320ff; 1988: 174). In his reply to Blau, however, headmits that he made the opposite error of neglecting type 1 relations in TheFoundations of Social Theory. ‘Were I writing the book over again, I would giveconsiderably more attention to this macro-micro relation’ (Coleman, 1993b: 63).In the latest writings of Coleman, it is no longer possible to detect any preferencefor an individualistic rational action theory over structuralism. Since theyare complementary, both are equally necessary and equally important.Concluding a discussion of existing and possible areas of application of ‘TheEconomic Approach to Sociology’, Coleman writes:In all four of these areas of application, the role of rational action theory iscentral. Also central, however, is structural theory. Rational action theorydrives the phenomena, for it constitutes an engine of action for the actor. Itscomplement, structural theory, is necessary to translate the action from themicro level of actors to the macro level of systems of action. The resultingsystem of behavior is a consequence both of the ‘engine of action’ providedby rational actor theory and structures of interdependence of actor’s actionswhich generate system behavior from individuals’ actions.(Coleman, 1992a: 147)An interesting thing about this quotation is that Coleman now makes adistinction between rational action theory and structural theory. I believe thatthis distinction is sound, since there is nothing in a theory of rational action, or

304 Rational choice individualismof rational choice, to suggest the idea of social structure. I also believe that itsupports my suggestion that Coleman’s programme is really a combination ofindividualism and structuralism. Therefore, insofar as it is a manifestationof methodological individualism at all, it is best characterised as structuralindividualism.One important question remains: What is the difference between the institutionalindividualism of Popper’s followers and the structural individualism ofJames Coleman? Before suggesting an answer, I admit that it is possible toconceive of structural individualism, as a form of institutional individualism.This is so, because, in my view, as well as in Coleman’s, social structure is institutionalstructure, a direct or indirect, result of social institutions. What,nevertheless, makes structural individualism distinct is that social structure issomething more than mere institutions. In institutional individualism, individualsact in a situation consisting of other individuals and institutions, which, inGidden’s phrase, both ‘enable and constrain’ their actions. Actors may haveheterogeneous desires and beliefs, but they do not necessarily, or typically, occupypositions and they have no relations that derive from their positions. Individualsinteract and their actions are, therefore, interdependent, but there is no interdependenceof the situations they face prior to interaction. Interdependence is acontingency of interaction. The behaviour of the social system is the aggregatedresult of the actions of individuals, or the resultant of their interaction.In structural individualism, on the other hand, actors are occupants of positions,and they enter relations that depend upon these positions. The situationsthey face are interdependent, or functionally related, prior to any interaction.The result is a structural effect, as distinguished from a mere interaction effect.In addition to natural persons, then, there are social positions and corporateactors made up of social positions. The behaviour of social systems is, at least inpart, determined by the structure of those systems.The difference between institutional individualism and structural individualismcan be illustrated by the examples of modern microeconomics andColeman’s analysis of corporate actors. As we have seen above (pp. 245ff ), thereis a growing awareness that microeconomics exemplifies institutional individualism,rather than the original, strong version of methodological individualism.Individuals are price takers and prices depend upon the institution of money. Inaddition, modern markets are possible only because of the rights of propertyand contract. Within this institutional framework, individuals interact and theresult is new prices and a new allocation of resources. The process by which thistakes place is usually described as ‘aggregation’, but as Coleman (1986a: 347;1986c: 1321; 1987a: 154) observes, this is a misleading term: ‘for the phenomenato be explained involve interdependence of individual’s actions, not merelyaggregated individual behaviour’ (1990c: 22). ‘[S]ocial structure is somethingother than the undifferentiated, fully communicating social structure that isassumed in the perfect market of neoclassical economics’ (1994a: 167). Since it isdifficult to provide a positive explication of structural individualism, I suggest a

Rational choice individualism 305negative one: structure is tantamount to a failure to solve the problem of aggregation,in economics, as elsewhere.When it comes to corporate actors, ‘Rights are held by functional parts of theorganization, and then by positions in each of these parts’ (Coleman, 1994a:173). Thus, according to Coleman, social structure is something more than institutions,in the sense of rules of individual behaviour. It is also more than theinterdependent interaction which results from individuals following rules of individualbehaviour. Social structure consists largely of corporate actors,constructed by some individuals with the help of rights and norms created byother individuals, but existing independent of each one of them, or at least ofmost of them. Corporate actors are made up of positions, which are independentof the individuals filling these positions. The social systems emerging fromtype 3 relations, therefore, cannot all be depicted as resulting simply from theactions of individuals acting in an institutional environment. It is the result ofindividuals acting in a structure of interdependent positions, existing prior to theinteraction of individuals filling these positions (Type 6). According to Coleman(1991: 6) these constructed social structures may have unintended consequences.Thus it is not just that individual actions have unintended consequences, as inthe idea of spontaneous order and in Elster’ supra-intentional causality (seebelow), but that there are systematic unintended consequences of social structures(see Coleman, 1992b: 266). An even more complete form of the Colemandiagram, therefore, would be like this.Figure 10.5 Social structure as positions to be filledA possible example of this Type 6 relation would be some forms of matchingbetween individuals and a set of pre-existing social positions that are somehowfilled by acting individuals.There is one more aspect of Coleman’s individualism that I would like tomention. Among the Austrians, methodological individualism was typically tiedto a particular theory, like theoretical economics, or a particular researchprogramme, such as Weber’s interpretive sociology, but there was no claim thatall social theories must conform to this methodological programme. In contradistinctionto this, the Popperian principle of methodological individualism wasadvanced as obligatory for all historians and social scientists. James Coleman

306 Rational choice individualismseems to me closer to the Austrians. He does not say that all explanations inhistory and social science must conform to the principle of methodological individualism.Type 2 and Type 4 relations in Coleman’s diagram are consideredpermissible, despite being holistic. His argument is that type 2 relations shouldbe supplemented by the neglected type 3 relations and that type 4 relationsshould be provided with microfoundations, consisting of a conjunction of type 1,2 and 3 relations. Coleman’s methodological individualism, then is microfoundational,but not reductionist (see p. 336). He wants to supplement holisticexplanations, not abolish them. Since he conceives of macrophenomena as realsocial phenomena, he does not believe that holistic explanations are half-way, inthe sense of Watkins. They are half-way in the sense of being partial, but not inthe sense of being not yet reduced.Raymond BoudonThe main pioneer of rational choice sociology, besides James Coleman, is theFrench sociologist Raymond Boudon. He differs from Coleman mainly in beingless of a structuralist, even if he wrote his first book on this subject (Boudon,1968), and by being more interested in the subjective, or psychological, side ofrational choice. His main source of inspiration is Max Weber (Boudon, 1984: ch.2). Like the latter, he focuses on the subjective meaning of action, or on beliefs(Boudon, 1989). In the end, he is led, by this focus, to launch a cognitive model,as an alternative to rational choice (Boudon, 1996; 1998). More exactly, Boudonconceives of the cognitive model as more general than rational choice, butincluding rational choice as a special case. One reason the cognitive model ismore general is that it includes action that is not self-interested and not eveninstrumental.Unlike Coleman, Boudon is not prepared to abandon homo sociologicus.Boudon’s homo sociologicus is not the one we are used to, however, but a close relativeto homo economicus, differing from the latter mainly in being less rational(1981: 7, 155–63; 1982: 7–9, 153f). At first, Boudon based his version of rationalchoice on Herbert Simon’s idea of bounded, and subjective rationality; later onhe relied more on Max Weber’s and Georg Simmel’s verstehende sociology (1989;1994; 1996).Also unlike Coleman, Boudon is not particularly attracted by Hobbes’sproblem of order. Like most social scientists, Boudon realises that, for explanatorypurposes, at least, the social order must be accepted as given. According toBoudon, the ‘state of nature’ is but one of several legitimate paradigms in sociology(1982: 161f, 183–5). 18In an early article, Boudon (1975) recognises three basic paradigms in macrosociology:functionalism, neo-Marxism and interaction analysis. His ownpreference is for the third, which he also calls ‘analysis of the aggregation of actions’(p. 398) and his main interest, at this time, was in the unintended, especially theperverse, consequences of interaction (1982). Characteristic for interactionistparadigm is situational logic and methodological individualism (1982: 201–5):

Rational choice individualism 307This principle means (in the broad sense used here) that the sociologist mustemploy a method which considers the individuals, or individual actors,included in the system of interaction as the logical atoms of his analysis. Toexpress the same principle in a negative manner, the sociologist cannot besatisfied by a theory which considers an aggregate (a class, a group, a nation)as the most elementary unit to which it will descend(Boudon, 1981: 36f)There are some oddities about Boudon’s methodological individualism. It‘assumes’, for instance, ‘very strongly that one can state a proposition like“Germany prefers the costs of war to those of submission”, for the subject of theproposition designates a many-headed actor provided with a mechanism ofcollective decision-making – the German government’ (1981: 37). It would seemthat Boudon uses the term ‘methodological individualism’ in a weak sense,indeed. This impression is strengthened by the fact that he finds examples ofmethodological individualism, not only in Tocqueville, Weber and Pareto, butalso in Marx, Durkheim, Parsons and Merton (1981: 38; 1982: 7).Boudon returns to methodological individualism in his article ‘TheIndividualistic Tradition in Sociology’ (1987). This time, methodological individualismis given the following explication:To summarize: suppose M is the phenomenon to be explained. In the individualisticparadigm, to explain means making it the outcome of a set ofactions m. In mathematical symbols, M = M(m); in other words, M is afunction of the actions m. Then the actions m are made understandable, inthe Weberian sense, by relating them to the social environment, the situationS, of the actors: m = m(S). Finally, the situation itself has to be explained asthe outcome of some macrosociological variables, or at least of variableslocated at a level higher than S. Let us call these higher-level variables P, sothat S = S(P). On the whole, M = M(m[S(P)]). In other words, M is theoutcome of actions, which are the outcome of the social environment of theactors, the latter being the outcome of macrosociological variables.(Boudon, 1987: 46; see also 1984: 29ff)As far as I can see, this is in all essentials the same version of methodologicalindividualism as that of James Coleman, but expressed in another language. 19 Intranslation, M = M(m) corresponds to Coleman’s type 3 relation, m = m(S)corresponds to Coleman’s type 1 relation and S = S(P), finally, to his type 2 relation.A difference, which is not visible in the two representations ofmethodological individualism, is that Boudon attaches much more importance tom = m(S) than does Coleman to the type 1 relation. Also, being an explication ofmethodological individualism, Boudon’s symbolism lacks anything correspondingto Coleman’s type 4 relation, between two social, ormacro-phenomena. This does not mean, however, that Boudon fails to payattention to this relation. On the contrary, he recognises several paradigms

308 Rational choice individualismbesides the individualistic one, which engage in seeking macro-relations. Thenomological paradigm, in particular, has, as its main objective, ‘to discovermacrosociological regularities, or lawlike statements’ (p. 46). In most studies,belonging to this paradigm, ‘no effort will be made to relate the statistical relationbetween variables to their real causes; that is, the individual behaviors ofwhich they are the outcome’ (Boudon, 1987: 62). Also the interpretive paradigmassumes that, ‘as soon as we leave the low levels illustrated by organization orsmall group studies and proceed to more complex levels, the individual actorscan advantageously be forgotten’ (p. 47).For Boudon, as for Coleman, methodological individualism is another namefor his advocacy of microfoundations, or ‘micro-explanation’. Like Coleman,Boudon regards statements concerning macro-phenomena as descriptive, andrequiring micro-explanation. 20 Behind his methodological individualism and hisplea for micro-explanation, there are certain metaphysical beliefs about man andsociety.the more a sociologist is convinced that social facts are the product of individualactions, the less he will have an interest in searching lawlikeregularities … individual action and the products of individual actionconstitute the only and ultimate reality which a sociologist has to deal with.A relationship between A and B can be nothing else but the product of individualactions, so that A will produce B or its opposite depending on thecomplex set of characteristics of the system of action.(Boudon, 1983: 14)In my view, these statements raise more questions than they answer, but theyshow the kind of considerations typically to be found behind an adherence tomethodological individualism. Perhaps the ultimate motive behind Boudon’smethodological individualism is a wish to salvage some freedom on the part ofthe individual in the determination of his own actions (1981: 6–13, 163–66; 192:chs 1 and 7). As a matter of fact, I suspect that Boudon is guilty of a conflationof social holism with sociologism, or social determinism, which is typical ofthose who accept a considerable amount of holism, but who still want to identifythemselves as methodological individualists.So what is the holistic element in Boudon’s methodological individualism? Aswe have seen, it assumes that the situation of individuals ‘has to be explained asthe outcome of some macrosociological variables, or at least variables located ata higher level than S’ (Boudon, 1987: 46). What is hidden in this suggestion?One obvious candidate is social institutions: ‘By system I mean the outcome ofthe aggregation of a set of individual actions taken within a given institutionalframework’ (1975: 400). Boudon’s methodological individualism, then, is reallyinstitutional individualism.But this is not all. Like Coleman, he conceives of the social system, or itsstructure, as a set of interrelated positions, independent of each individual,which is one step further removed from the strong version of methodological

Rational choice individualism 309individualism. This becomes clear if we take a look at his explanation of thefailure of short-cycle higher education in France.According to Boudon, the logic of the situation for those opting for highereducation is similar to that of the prisoner’s dilemma. Faced with the choicebetween long- and short-cycle higher education, every student has an interest inchoosing the former alternative. But if all do, some of them will come out worsethan if they had chosen the short-cycle alternative (Boudon, 1982: 96–104). Thisconsequence is the result of assuming, among other things, that students areequal and perceive themselves to be equal. By assuming, instead, that studentsfrom different classes, and their families, are different in important respects, weget a model which explains the unequal access to higher education for studentswith different social backgrounds. Besides differences in values and linguisticability, Boudon draws attention to differences in the estimation of advantages,disadvantages and risks accompanying investment in higher education (Boudon,1981: 135–42). Nevertheless, there has been an increase in the average level ofeducation. A further problem is how to explain why this increase has not led tothe expected increase in inter-generational social mobility. Boudon’s answer is interms of a fixed stock of social positions to be filled. When the queue to the toppositions gets longer, there is an increasing number of people whose claim tothese positions becomes ineffective (Boudon, 1981: 67–75).This analysis indicates that Boudon is not only an institutional individualist,but a structural individualist, as well. And, indeed, there is lots of additionalevidence to back the conclusion: Boudon is a structural individualist, likeColeman (cf. Hechter, 1983: 8). 21Analytical MarxismIn the last twenty years, we have witnessed the birth of a somewhat odd creature.Quite a few, seem to look at it as something of a monster. I am thinking ofAnalytical Marxism, also called Rational Choice Marxism. The most scandalousfact about this approach to Marxism is that it endorses methodological individualism,or the claim that Marxism must be provided with microfoundations. 22How is it possible to turn a critic of methodological individualism (see p. 11) intoa methodological individualist? Lots of Marxists believe that it is not at allpossible and the critical literature is, by now, hard to survey. 23 Fortunately, forme, I do not have to take a stand on this issue here. I am interested in the theorycalled Rational Choice Marxism, irrespective of its compatibility with the socialtheory of Marx. This means I can ignore the huge body of literature discussingthe question whether rational choice and methodological individualism arecompatible with Marxism, or not. My own view, stated without an argument tosupport it, is that it is possible to use rational choice in Marxist analysis, but notto reduce the latter to the former. Methodological individualism is, I believe,even harder to square with the basic assumptions of Marx’s social theory. Theonly version of methodological individualism, which is at all conceivable within

310 Rational choice individualisma Marxist approach, is the structural individualism developed by rational choicesociology.With these provisos, I turn to a brief presentation of the most outspokenmethodological individualists among rational choice Marxists: Jon Elster andJohn Roemer.Jon ElsterThe most influential case made for rational choice in social science is probablythat of the Norwegian historian and philosopher Jon Elster. Like Harsanyi,Elster launches game theory as an alternative to functionalism within socialscience. In the case of Elster, however, it is not the functionalism of Parsons, butthat of Marx and the Marxists, which is deemed deficient and therefore to bereplaced by a more scientific approach.Elster’s main objection to functionalist explanation in social science is that itposits a ‘purpose without a purposive actor’, thus presupposing some kind ofobjective teleology. Institutions or behavioural patterns are explained by theirbeneficial consequences without being intended by anyone. Elster agrees, ofcourse, that institutions may have beneficial consequences for some group insociety without being intended by anyone, but this does not explain the existenceof the institution unless the mechanism responsible for the maintenance ofthese institutions is specified. The defect with most functionalist explanationsof social institutions is that they fail to specify the feedback mechanism wherebysocial institutions with beneficial consequences are maintained. 24Elster provides several examples of the illicit use of functionalist explanationsin social science, especially in Marxist social science, and suggests that functionalismshould be replaced by the rational choice approach, especially in the formof game theory. The rational choice approach, Elster claims, is the best amongavailable approaches to human behaviour, and game theory is invaluable to theMarxist analysis of exploitation, class struggle and revolution. 25Contrary to the belief of many Marxists, Elster also claims that Marx himselfused the rational choice approach. It is part of Marxian economics that capitalistsmove their capital to the industry where the highest profits can be found andthat they innovate in order to survive. This surely presupposes rational action onthe part of capitalists (Elster, 1983a: 165ff). Elster also gives some examples inorder to demonstrate the utility of game theory for the analysis of class struggle(Elster, 1982: 465ff; 1985: 371ff).Like Coleman and Boudon, Elster justifies rational choice in terms of a questfor microfoundations. The ground for this seems to be justificatory, but not ofthe usual empiricist variety. Microfoundation is not a matter of observability, butof finding the causal mechanism at work. According to Elster, ‘To explain is toprovide a mechanism, to open up the black box and show the nuts and bolts, thecogs and the wheels, the desires and beliefs that generate the aggregateoutcomes’ (1985: 5). The need for microfoundations seems to be especiallypressing within Marxism. ‘Without a firm knowledge about the mechanisms that

Rational choice individualism 311operate at the individual level, the grand Marxist claims about macrostructuresand long-term change are condemned to remain at the level of speculation’(1982: 454). This is where game theory comes in.Game theory provides solid microfoundations for any study of social structureand social change … For Marxism, game theory is useful as a tool forunderstanding cases of mixed conflict and cooperation: cooperation inproducing as much as possible, conflict over dividing up the product. Gametheory can help understand the mechanics of solidarity and class struggle,without assuming that workers and capitalists have a common interest andneed for cooperation. They do not.(Elster, 1982: 477f)Although optimistic about the future prospects of the rational choiceapproach, Elster has a keen sense of its limitations (Elster, 1979b: 112ff). In this,he differs favourably from many other defenders of the rational choiceapproach, such as Becker and Harsanyi. First of all, game theory is limited in itsscope to the sphere of the intentional. But there are also fields in social life wherecausality reigns. There is, according to Elster, both a sub-intentional and a supraintentionalcausality, circumscribing the field of intentional action. ‘Subintentionalcausality’ refers to socialisation; the process in which our preferencesand beliefs are shaped, while ‘supra-intentional causality’ refers to the causalinteraction between intentional agents, i.e., when individuals act upon the beliefthat other people display traditional behaviour rather than rational action (Elster,1978: 157–63; 1979b: 20, 83–8). He also admits that not all intentional action isrational, that it is subjected to structural constraints and usually takes place withless than perfect information. Nevertheless, ‘there is a hard core of importantcases where the rational-choice model is indispensable’. This model, furthermore,is ‘logically prior to the alternatives, in the sense that the social scientistshould always be guided by a postulate of rationality, even if he may end byfinding it violated in many particular cases’ (Elster, 1979b: 116). This is Elster’sso-called ‘principle of charity’, which he borrows from the philosopher RonaldDavidson. 26Elster is not only a proponent of rational choice, but of methodological individualismas well, and he seems to share the common presumption that the twoare inseparably linked. 27 By ‘methodological individualism’, Elster means ‘thedoctrine that all social phenomena – their structure and their change – are inprinciple explicable in ways that only involve individuals – their properties, theirgoals, their beliefs and their actions. Methodological individualism thusconceived is a form of reductionism’ (Elster, 1985: 5).Elster’s methodological individualism, as stated in this quotation, is the original,strong version of this doctrine. This is suggested by the fact that socialphenomena appear only as that which is to be explained (explanandum), whereasthat which explains (explanans), includes only individuals – their properties, goals,beliefs and actions. An even more clear statement to this effect is this:

312 Rational choice individualismThe elementary unit of social life is the individual human action. To explainsocial institutions and social change is to show how they arise as the result ofthe action and interaction of individuals. This view, often referred to asmethodological individualism, is in my view trivially true.(Elster: 1989b: 13)The interpretation of Elster’s methodological individualism as a strongversion of this doctrine gains additional support from his claim that it is a formof reductionism. The implication of this claim is that Elster is a psychologisticmethodological individualist. ‘Many argue that sociology cannot possibly, or atleast not today, be reduced to psychology. Since I insist that the individualhuman action is the basic unit of explanation in the social sciences, I amcommitted to this reduction’ (1989b: 74; see also 1993: 7).There is one thing, however, which makes for a weakening of Elster’smethodological individualism: the words ‘in principle’. Like J.W.N. Watkins (seepp. 215f), Elster admits of half-way explanations. Temporarily, at least, the socialscientist can be satisfied with explanations at the macro-level. In some situations,it may even be harmful to make a premature attempt to reach rock-bottom(1983b 116; 1985: 6, 359).Elster’s methodological individualism is based on ontological individualism.This is evident from his attempts to clarify this doctrine. Methodological individualism,according to Elster, is not incompatible with the fact that ‘individualsoften have beliefs about supra-individual entities that are not reducible to beliefsabout individuals’. Nor is it incompatible with the fact that ‘[m]any properties ofindividuals, such as “powerful,” are irreducibly relational, so that accuratedescription of one individual may require reference to other individuals’ (Elster,1982: 453).A second indication of strong methodological individualism is that he agreeswith Weber and Hayek that the fact that people entertain holistic concepts, orbeliefs, about society is not a threat to methodological individualism, since thesebeliefs are the beliefs of individuals. Or, in other words, the fact that methodologicalindividualism does not hold in intensional contexts is not a problem,since it is enough for social science that it holds in extensional contexts (Elster,1985: 6). This view suggests the intersubjectivist theory of society treated, atlength in chapter 5. A third indication of strong methodological individualism isElster’s ‘reduction’ of social relations to relational properties of individuals,which follows from the second one. In his reply to Michael Taylor, Elster says:Taylor’s main point is that my view does not allow structures as causes. By astructure he means a set of relations, e.g. relations of production, defined inabstraction from the specific relata. In his opinion, structures, thus defined,have causal efficacy. I disagree. I think beliefs (and, more generally, attitudes)about structures, thus defined, have causal efficacy, but as I say in the book,methodological individualism does not hold within intensional contexts. In

Rational choice individualism 313extensional contexts, what has causal efficacy is a relation with its relata or,as I put it, individuals together with their relational properties.(Elster, 1986b: 67)This is a typical ingredient in an individualistic ontology, which makes forconceiving of social order as interaction, rather than structure (cf. Taylor, 1986:5). On the basis of these clarifications of the doctrine of methodological individualism,I think it is safe to conclude that Elster is also an ontological individualist.This conclusion is supported by his view of social institutions as ‘essentially …collections of human beings’ (1993: 8). The clearest statement to this effect,however, is in a reply to Charles Taylor:Through the triple interdependence studied by game theory – betweenrewards (through envy and altruism), between choices and rewards (throughgeneral social causality), and between choices (through mutual anticipation)– the individual emerges as a microcosmos which sums up in itself the wholenetwork of social relations. This in my view makes for a plausible explanationof the emergence of the norms of solidarity, though here as elsewheretheir persistence may be better explained through the diffuse socializationprocesses in the family than by explicit considerations of this kind.(Elster, 1980: 218f)Elster, then, is not only a methodological individualist, but an ontologicalindividualist, as well: that much is clear. But he is not committed to the mostextreme version of methodological individualism, as exemplified by the theoriesof the social contract and of general equilibrium. He denies being an atomistand he conceives of human beings as social beings (Elster, 1993: 7). Elster’s ideaof the individual as a microcosmos, suggests that his social ontology is closer toLeibniz’s monadology, which would not be a coincidence, since Elster oncewrote a book on Leibniz (cf. Moggach, 1991).Compared to Coleman and Boudon, Elster is much more of an individualist.While the former are structural individualists, Elster is a fairly clear case ofstrong psychologistic individualism. It makes no explicit room for the treatmentof social institutions and social structure as exogenous variables in social scientificexplanations, at least not in rock-bottom explanations. The question remainsif Elster makes more room for social institutions and social structure in otherparts of his work. I think this is the case. I think it is possible to find at leasttraces of institutional and structural elements in his general model for explanationsof human action.According to Elster, human action may be seen as the end result of twofiltering processes. The first filter consists of structural (physical, technical,economic and institutional) constraints upon action. The set of possible actionsremaining within these constraints is called, by Elster, the ‘feasible set’, by mostothers the ‘opportunity set’. The second filter consists of the human mind as thelocus of choice. More specifically, it consists of the individual’s preferences and it

314 Rational choice individualismfunctions so as to single out one alternative, within the feasible set, as the actionto be taken. 28 Exactly what belongs to the external constraints, limiting the alternatives,and what shapes preferences is hard to tell and a matter of disputeamong adherents of rational choice. 29According to Elster, there are two alternatives to rational choice. The first isstructuralism, which may be seen as a limiting case of rational choice, whenthere is only one, or a few, alternatives in the opportunity set. In this case, theindividual has little or no choice. Elster interprets structuralism as implying anassumption of severe limits to choice, or of a strong social determinism. I do notthink this is a fair interpretation of structuralism, but I leave this matter unsettled.It is certainly not the only one, as Elster admits, himself. 30 For my presentpurposes, it suffices to point out that, according to Elster, there are structuralconstraints upon action, which enter rational choice explanations through theopportunity set.Elster finds another alternative to rational choice far more important. Thisalternative says that the mechanism realising one member of the feasible set iscausal rather than intentional. This may be due to various obstacles to rationality,such as habitual behaviour, tradition, values, norms and roles. Elsteragrees that action is causally influenced in this way, but not directly. Traditions,values and norms influence human action indirectly, through the preferences(Elster, 1979a: 76–8; 1979b: 114–16). I think there is some ground for this viewin the case of values, but less so in the case of habits, traditions, norms and roles.In these cases, we come closer to the truth, if we conceive of them as constraintsupon action, or as alternatives to rational choice. 31 I believe that Elster is herecommitting a mistake which is the obverse of that commonly ascribed toDurkheim. While Durkheim was led, by his correct observation that social factsare external to each one of us, to the conclusion that they are external to all ofus – which is a correct conclusion in one sense of ‘all’ 32 – Elster seems to be led,by the fact that values and norms are internalised, to the conclusion that they donot act as external constraints upon action. But this is an unwarranted conclusion.Even if it is true that values and norms, which we acquire in the process ofsocialisation, become internalised and act directly upon my preferences, theinternalised values and norms of other people confront me as externalconstraints, in the form of expectations, sanctions and overt behaviour. The factthat all, or most, individuals, follow the institutional rules of a society, makesthese rules external to all of them, only not all of them collectively. 33Nevertheless, even if Elster is right to suggest that ‘society’ influences peoples’preferences rather than their opportunity sets, this is still a ‘social influence’,which seems to imply a break with strong methodological individualism, but ‘onlyif the causes of the attitudes and beliefs which cause action are themselvesnothing but actions and properties of individuals’ (Taylor, 1986: 4). Elster is notcrystal clear on this point. According to him, this social influence, suggeststhe idea of a general sociological theory, in which preferences and desires areexplained endogenously as a product of social states to the generation of

Rational choice individualism 315which they also make a contribution … which, needless to say, in the presentstate of the arts appear to be light-years away – would include (i) the explanationof individual action in terms of individual desires and beliefs, (ii) theexplanation of macro-states in terms of individual actions, and (iii) theexplanation of desires and beliefs in terms of macro-states.(Elster, 1983a: 86; see also 1983c: 141ff)As we saw above, Elster recognises two types of causality in human affairs:supra-intentional and sub-intentional causality. The explanation of desires andbeliefs in terms of macro-states, mentioned in the quotation above, is a case ofsub-intentional causality. The different forms of this type of causality can besubsumed under the general heading of ‘socialisation’. As such, it is fullycompatible with the theory of society as subjectively meaningful action (seechapter 5) and does not imply any element of holism. But according to Elster,preferences are also shaped by class position and class interest (1982: 468ff;1983c: 141ff; 1985: 460). This introduces a structural element, which is holistic.Paradoxically, then, sub-intentional causality seems to be more holistic thansupra-intentional causality. The latter type of causality is a matter of the interactionbetween individuals, but there is nothing to indicate that this interaction isdetermined by structure.It is a peculiar feature of Elster’s ‘general sociological theory’ that it does notmake room for structural constraints upon action (see Taylor, 1986: 3–5). 34 It ispossible, however, to read the following quotation as implicit recognition of suchconstraints:Simply to postulate causal relationships between macro-variables will notdo. We may observe an empirical regularity … but we have explainednothing until we can show (i) how the macro-states at time t influence thebehaviour of individuals motivated by certain goals, and (ii) how these individualactions add up to new macrostates at time t1.(Elster, 1983a: 84; see also 1983b: 116)Elster’s general sociological theory seems to have much in common with therational choice models suggested by Coleman and Boudon, except that it ismuch less structuralistic than that of Coleman and less structuralistic even thanthat of Boudon. This is rather surprising, considering the fact that Elsterconceives of himself as a Marxist of sorts.Elster’s attempt to reinterpret Marx as a methodological individualist hasbeen rejected by a majority of Marxists (and some others), who feel that this isgoing too far. The number of critical writings is by now enormous, and quite alot deal with methodological individualism. 35 It is quite understandable if Elsterhas grown tired of defending himself against all this critique. Whether for this,or for some other reason, Elster seems to have decided that methodological individualismis not a subject he wants to discuss any more. His recent contributions

316 Rational choice individualismto methodology are about ‘social mechanisms’, which are now treated as separatefrom methodological individualism (Elster, 1998: 47).I will not make another contribution to the already immense number of criticalwritings on Elster’s attempt to make sense of Marx. First of all, because, I donot believe that I have anything of importance to add. Second, because I believethat Elster has done Marxism and social science a great service, irrespective ofthe ultimate validity of his interpretation, or reconstruction. I do believe,however, that it is possible to find examples in Elster’s work on Marx of interpretations,which are more in line with structural individualism than withpsychologistic individualism. One example may be Elster’s suggestion thatMarxist class theory ‘attempts to explain collective action in terms of the classposition of the individuals engaged in it’ (1985: 336). Elster denies that this is aviolation of methodological individualism. According to him it is only a convenientshorthand. This is a somewhat desperate methodological individualism byfiat. It is, of course, possible to claim, with equal right, that all holistic statementsare shorthand for statements about individuals.John RoemerJohn Roemer is a Marxist economist, who has used the tools of neoclassicaleconomics in order to create an ‘analytical’ Marxism. He is also a moral philosopher,who wants to get rid of exploitation. In both capacities, he has relied onmethodological individualism as the proper method of investigation.In his first book, Analytical Foundations of Marxian Economic Theory (1981),Roemer’s main objective is to turn Marxian economics into a science, with thehelp of mathematics, deduction, equilibrium analysis and the microfoundationsapproach (p. 7). He notes that Marxism is a theory of class, rather than of theindividual, and admits that it is possible to ‘build a model in which classes are theatoms of the system’, but chooses the individualist path: ‘The reason is this: thatindividuals act as members of a class, rather than as individuals should be atheorem in Marxian economics, not a postulate’ (p. 7; see also 1982b: 518).Roemer invokes Marx, himself, in support of this methodological individualism.In his next book, A General Theory of Exploitation and Class (1982a) Roemerfocuses his attention on one particular topic of fundamental importance to aMarxist theory of history: exploitation. Once again he uses a neoclassicalapproach. 36 In the model he uses, ‘both exploitation status and class positionemerge endogenously as a consequence of individual optimization in the face ofa constraint determined by one’s ownership of productive assets’ (p. 15). This istraditional methodological individualism, even if one institution, viz., ownership,remains exogenous.Roemer compressed his theory of exploitation to an article (1982c), where theindividualist ambition appears even more clearly (p. 262f). Of those Marxists,who commented on Roemer’s article, at least one, Adam Przeworski, was enthusiasticabout Roemer’s methodological individualism.

Rational choice individualism 317In describing a society, Roemer begins with noting the set of techniques ofproduction available to it and the distribution of wealth characteristic of it.Then he imputes objectives to individuals. This is what is given about aparticular society and this is all that is given. He then shows that when individualspursue their objectives, under the constraints of the technology andthe structure of property, they will enter necessarily into class relationswithin [sic] one another and will produce a definite distribution of income.(Przeworski, 1982: 306)Przeworski then goes on to argue that the orthodox Marxist idea that individualsenter objective social relations and occupy positions in a social structure which isindependent of particular individuals is a tautology. 37 He ends his comments byjoining in the pleas for a methodological individualism (p. 313).Roemer returns to the issue of methodological individualism in the programmaticanthology on Analytical Marxism (1986), but this time he observes adifference between the ‘hegemonic individualism’ of neoclassical economics andthat of analytical Marxism. Whereas the former assumes fixed preferences,‘Marxism … asserts that people’s preferences are in large part the consequenceof social conditioning’ (p. 193). According to Roemer, it is the task of rationalchoice to develop a theory of endogenous preference formation. But this takesplace in an existing social environment.Thus individuals are formed by society, and these individuals react rationallyto their environments to produce tomorrow’s environment, which in turnproduces individuals who think somewhat differently than before, and reactin their environment to bring about yet another equilibrium.(Roemer, 1986: 196)This is the familiar chicken-and-egg history once again, and we may recallPopper’s argument, that the only way to get rid of the social environment is topush the regression back into the state of nature. This was the argument that ledPopper to adopt institutionalism. Roemer chooses another path. He suggests thatan individual’s preferences might be determined by her meta-preferences, whichin their turn, are determined by the expectations of others, given the place sheoccupies in the social structure (1986: 198f). It seems to me that Roemer reintroducessocial structure as an exogenously given variable in his account ofpreference formation. This impression is confirmed, by Roemer’s article on‘Marxism and Contemporary Social Science’ (1989). This time, he seems morewilling to admit the possibility of a non-reducible social determination of preferences.Methodological individualism, now,does not preclude explaining why individuals conceive of their interests asthey do (that is, why they have the particular utility functions they have);

318 Rational choice individualismindeed, the Marxist explanation is that a person’s conception of his interestsis moulded by his place in the economic structure.(Roemer, 1989: 278)Structural individualismA somewhat paradoxical conclusion of this chapter is that, of the four methodologicalindividualists discussed in this chapter, the two Marxists are leastcommitted to structuralism. This is paradoxical because structuralism is oftenconsidered to be one of the most typical features of Marxism. Anyway, the mainresult of this chapter is that there has emerged a new version of methodologicalindividualism that we may call ‘structural individualism’, with a term borrowedfrom the Dutch sociologists Reinhard Wippler (1978a) and Werner Raub (1982).Structural individualism is the weakest form of individualism, for the simplereason that it is a synthesis of individualism and holism. The term holistic individualism,suggested by Philipp Pettit, for a similar position (Pettit [1993] 1996:passim), is an alternative, but I have settled for structural individualism. 38To have reached structural individualism, we have come a long way from themost radical versions of methodological individualism in the theory of the socialcontract and the theory of general equilibrium, and also from the less extremeversions of methodological individualism advanced by J.S. Mill, Carl Menger,Max Weber, Friedrich von Hayek, George Homans, Karl Popper and J.W.N.Watkins. In their versions of methodological individualism no causal, orexplanatory, power is attributed to social phenomena. In the two-level graphicrepresentation I have used in this book, no arrows lead from the social level tothe individual level. According to structural individualism, however, there areseveral distinct ways in which social phenomena influence individuals. Drawingupon the various representatives of structural individualism, I recognise thefollowing types of influence:Figure 10.6 Structural individualismArrow 1 represents the general influence of culture on individuals’ preferencesand beliefs in terms of social structure. This explanation is holistic only ifthe social environment is conceived of as situated on the social level. While most

Rational choice individualism 319structural individualists do that, many methodological individualists conceive ofthis influence as flowing from some individuals to other individuals. The mostobvious example is the theory of society as subjectively meaningful action. Iguess that also Elster would see it this way. Arrow 2 represents the influence ofsocial structure on our preferences and beliefs. The way our preferences andbeliefs are influenced by our own place in the social structure. This type of influenceis more difficult to reduce to the individual level. This type of influence isacknowledged by Elster, but more clearly by John Roemer and most clearly byMichael Taylor and Adam Przeworski. Arrow 3 stands for another type of structuralinfluence on our preferences and actions: that which has to do with oursocial roles. If I engage in the maximisation of profit, for instance, this is onlybecause I am the manager of a firm. If I preach in a church, this is only becauseI am a priest. This type of influence is clearly recognised by James Buchananand James Coleman, for instance. Arrow 4 represents the institutional, or structuralconstraints on the opportunity set. This type of influence is recognised byvirtually all institutional and structural individualists. Arrow 5, finally, representsthe structural determination of interaction; the way in which the interdependenceof the social situations of a set of individuals produce structural effects. Amongmethodological individualists, this type of influence is most clearly and explicitlyrecognised by James Coleman, but I believe it is part also of Raymond Boudon’sversion of structural individualism and of the structural individualism of theDutch sociologists Reinhard Wippler, Werner Raub and Siegwart Lindenberg.

11 Why methodologicalindividualism?The simple answer to this question is that most methodological individualistsadopt this doctrine because they believe it is good social science; indeed, that itis the best guarantee that our knowledge of society will grow. But why do theybelieve this? Isn’t it for social science itself to tell which methodology is best?Well, things are not that simple. All scientists have ideas about what is goodscience. These ideas are based, partly on experience and partly on a prioriimages of science. It is obvious, for instance, that methodological individualism isbased on epistemological and ontological views about the nature of knowledgeand of society; views which are not so much the result, as the precondition ofresearch.There is one doctrine, in particular, which seems to be an important elementin the justification of methodological individualism: the doctrine of reductionism.Many social scientists and philosophers of science conceive ofmethodological individualism as a particular case of reductionism, and manymethodological individualists share this view. The reason for this is, most probably,that they believe that reduction is the highroad to scientific progress (cf.Kinkaid, 1997: ch. 1).Behind the adoption of methodological individualism, it is also possible todiscern a complex of moral and political considerations. Even though the termmethodological individualism was coined to make a distinction between methodologicaland political individualism and to prevent a confusion between them,there is little doubt that many methodological individualists adopt this doctrinefor moral and/or political reasons, some of them explicitly so.What I am suggesting, then, is that there is some kind of relation betweenmethodological individualism and certain other doctrines, such as epistemologicalindividualism, ontological individualism, reductionism, ethical individualismand political individualism. Such relations between doctrines may be of severalkinds. 1 An important distinction is between logical and psychological relations.This distinction, although analytically clear, is not always clear-cut when appliedto actual relations between doctrines. A first complication is that such relationsare seldom those of entailment, but usually some form of weak implication. Asecond complication is that psychological association of ideas is often based

Why methodological individualism? 321upon logical relations. Ideas become associated because they are logicallyrelated.In this chapter, I will first of all discuss methodological individualism as aninstance of reductionism and the implication of this for the relation betweenmethodological individualism and psychologism. Second, I will look at the morecontroversial question of the moral and political motives behind methodologicalindividualism. Before I get on to my main tasks, however, I will briefly discuss theepistemological and ontological assumptions behind methodological individualism.Philosophical backgroundIt is a commonplace today that science is based on epistemological and ontologicalassumptions about knowledge and society. A host of concepts developed inthe philosophy of science, such as ‘paradigm’ (Kuhn [1962] 1970), ‘researchprogramme’ (Lakatos, 1968; 1970) and ‘research tradition’ (Laudan, 1977), havebeen advanced, at least partly, in order to make that point.Methodology, being a set of rules, or principles for scientific research is themain vehicle for the impact of epistemological and metaphysical beliefs onscience. As we have seen in previous chapters, methodological individualism isoften backed up by epistemological and ontological beliefs. Indeed, methodologicalindividualism is sometimes stated as a thesis about our knowledge of societyand often as an ontological thesis about the nature of society. I will returnto these two theses in the final chapter of this book. Here I will concentrateon the supportive role of epistemology and ontology, but if we conceive of‘methodological individualism’ as including epistemological and ontological individualism,then, it would be more appropriate to talk about an ‘individualisticresearch programme’, or ‘research tradition’.I suggested in chapter 6 that methodological individualism looks very muchlike a version of the empiricist attack on metaphysics. I also pointed out,however, that few of the most influential methodological individualists have beenempiricists. It is clearly possible, then, to be a methodological individualistwithout being an empiricist. I will now look a little bit closer at the epistemologicalbeliefs of methodological individualists.Closest to being empiricists, among the more influential methodological individualists,are John Stuart Mill, George Homans and the Popperians. None ofthem, were radical empiricists, however. They all rejected inductivism. ThePopperians, in particular, deny being empiricists, or positivists. Karl Popper,himself, takes pride in being a severe critic of logical positivism, who has foughta life-long struggle against the inductivism of empiricism and even allows himselfto ask the rhetorical question: ‘Who killed logical positivism?’, without leavingany doubt about the right answer (Popper, 1976: 87–90). His pupils, Watkins andAgassi, have continued this fight by insisting that there is an inescapable metaphysicalelement in all science.No less hostile toward empiricism are Mises (Mises, 1957: 240–50) and Hayek([1952] 1963: 165–94; 1978: 35–49). Mises has called positivism ‘the most

322 Why methodological individualism?conspicuous failure in the history of metaphysics’ (Mises, 1943–4: 536), andHayek is of a similar opinion. Mises and Hayek, as we have seen, belong to theAustrian School of Economics. A most characteristic mark of this school is itsdefence of a theoretical and deductive economics, and its critique of the inductivistgathering of facts advocated by the German Historical School ofEconomics.Finally, Weber has not expressed any opinion about empiricism that I knowof, but his method of Verstehen has been a constant target for attack from positivistphilosophers and sociologists. Despite the anti-empiricism of the main adherentsof methodological individualism, there is a certain affinity between their ownepistemology and that of empiricism.The philosophical roots of Weber’s methodology, as we have seen, are to befound in Kant and in neo-Kantianism, rather than in empiricism. When Weberholds that social concepts are theoretical constructions of the mind, this shouldbe understood as an expression of a Kantian distrust in ‘things in themselves’.Though different in important respects – empiricism rejects the synthetic a priori,and also takes a more passivist view of the acquisition of knowledge –Kantianism and empiricism have this much in common: both reject conceptualand scientific realism in favour of an instrumentalist or conventionalist view ofscientific concepts.The roots of Mises’s methodology are probably to be found both inKantianism and phenomenology. Human action, according to him is a categoryembedded in the logical structure of mind, a presupposition of perception,apperception and experience. Praxeology, the theory of human action is bothvalid a priori and concerned with reality. This claim is usually interpreted as anexpression of the Kantian synthetic a priori, but is probably based upon the apodicticcertainty of phenomenological eidetic intuition (Wesenschau). Hayek’sphilosophical influences, finally, are probably to be found, above all, inphenomenology, but he also admits a certain influence from logical positivism(Hayek, 1967: 268). These influences might seem incompatible, but alsophenomenology has something in common with empiricism, especially with thephenomenalist or sensationalist brand of empiricism. They are both mainlyconcerned with the experience of individuals (or egos), the difference being thatempiricists regard this as unproblematic, while phenomenologists see it as theirmain task to lay bare its a priori presuppositions (Föllesdal, 1972: 427). Anotherpoint of contact between phenomenology and empiricism is their respectiveattempts to constitute or construct the common world out of these experiences(Kaufmann, 1940: 138–42). When Hayek holds that social objects are ‘constituted’or ‘constructed’ out of the actions of individuals, it is probable that heuses these terms in the phenomenological, rather than in the empiricist, sense.Popper shares the view that social objects are theoretical constructions. Thisview, however, appears to be inconsistent with his otherwise realist position.Methodological individualism also imposes special requirements on socialscience, which have no counterpart in Popper’s natural science methodology,

Why methodological individualism? 323and which do not follow from his general epistemology and methodology(Wisdom, 1970: 294f; Johansson, 1975: 99).Popper’s famous criterion of demarcation between science and metaphysics isfalsifiability, or testability ([1934] 1972: 40–2, 78ff). This criterion says nothingspecific about the kind of entities that are allowed to enter scientific explanations,if only it is possible to deduce some testable conclusion from the theory orhypothesis. When turning to the social sciences, Popper obviously feels that thiscriterion is insufficient, and makes the much stronger demand that only entitiesof a certain kind; individual human beings, are allowed within the antecedent ofits explanations. In Watkin’s terminology, methodological individualism is amaterial, and falsifiability a formal, methodological rule (Watkins, 1957b:104–6). The former does not follow from the latter. Theories about social wholesand collectives may be as much open to falsification as are those about individuals,etc. 2What, then, could be the reason for laying down special and more restrictiverules for social science? Why is the invisibility argument thought to be importantfor the social sciences, but not for the natural sciences? I could think of oneanswer, but it is probably not the right one. In the natural sciences, it is possibleto conceive of the entities occurring in its theories as material even if unobservable.It seems natural to believe, that if only we had a good enough microscope,we would be able to see such things as electrons and atoms, which are now unobservable(Harré, 1972: 90–5, 165ff; Shapere, 1977: 529f). In social science, thingsare different. Social phenomena are not material objects with a definite extensionand shape. They are not ‘things’ in the ordinary sense of that word. Socialphenomena lack the most typical characteristics of things; continuity in spaceand time and tangibility. We do not consider it as a possibility, that even if wehad some such thing as a ‘macroscope’, we would be able to detect socialphenomena except as complexes of individuals.Among the methodological individualists, typically empiricist arguments aremost frequent in the writings of Watkins. Not only does he repeatedly point outthat we only have direct access to individuals, he also argues that social conceptsderive their meaning from reference to individuals, and seemingly subscribes tothe idea that social objects are logical constructions out of individual people.This idea presumably also lies behind Jarvie’s contention that statements aboutcollectives, such as classes and armies, can be reduced to statements about individuals(see p. 221).When asked to defend methodological individualism, adherents of this principleusually reply that, after all, society is made up of individuals and nothingelse. 3 It follows from this, that individuals and their actions are the only causes ofsocial phenomena. This belief is often a matter of common sense, but it is sometimesbased on a more sophisticated metaphysics. I will treat the metaphysicalbasis of methodological individualism in more detail in another book and limitmyself to a mere mention of the most relevant doctrines here.Before the term ‘methodological individualism’ was coined by JosephSchumpeter, the name for this principle was the atomist method. ‘Atom’ means

324 Why methodological individualism?indivisible, and it was maintained by the Greek atomists that everything else isdivisible, and made up of atoms. Applied to society, it is usually assumed that theatom is the human individual. All social phenomena are aggregates, collections,or complexes, of individual human beings. The term ‘atomism’ is often used, insocial science, to denote the idea of an ‘isolated individual’, but this is not theonly use of the term. When Carl Menger talked about an ‘atomist method’ andMax Weber conceived of the individual as the atom of analysis, they did notmean the ‘isolated individual’.In previous chapters we have seen that many methodological individualistsreject ‘realism’ and invoke ‘nominalism’ in defence of methodological individualism.Indeed, in early sociology, methodological individualism was often calledsociological nominalism, as opposed to sociological realism. Nominalism is thedoctrine that only particulars exist. Universals are all in the mind. Applied tosociety, this means that only individuals exist, while social collectives are fabricationsof the mind.A third doctrine, which is often mentioned in discussions about methodologicalindividualism, is monadology. This doctrine derives from the Germanphilosopher G.W. Leibniz and is a German counterpart of atomism. Themonad, like the atom is the smallest part of the universe and indivisible. It differsfrom the atom by being endowed with a mind. This mind reflects, in itself, thewhole of the universe and is, in this respect a microcosmos. It is clear that theintersubjectivist theory of society discussed in chapter 5 is influenced byLeibniz’s monadology. Also Jon Elster, seems to conceive of society as made upof monads (see p. 313).ReductionismIt is common to see methodological individualism as a form of reductionism andmany methodological individualists accept this view. But not all, as we shall see.Very much depends upon the answer to the question: Reduction to what? Anatural answer would be psychology, but some methodological individualists,e.g., Weber and Popper, deny that methodological individualism is identical withpsychologism. In the case of Popper, this denial is based on his institutionalism,which is, indeed incompatible with psychologism. But it is still possible to maintainthat the strong version of methodological individualism is identical withpsychologism.Scientific reduction‘Reduction’, as here understood, means elimination without remainder. Thereduced theory (or science) is eliminated and replaced by, or incorporated in, thereducing theory (or science). This is how the term ‘reduction’ was originally usedby philosophers of science. ‘After a theory has been reduced to another itbecomes, in a sense, expendable. This is the heart of the matter’ (Bergmann,1957: 166). Or, in the words of Kinkaid (1997: 14), the ‘root notion behind

Why methodological individualism? 325reduction’ is ‘that one theory can in principle do all the explanatory work ofanother’. 4 In the further development of the doctrine of reduction, however,suggestions have been made to use this term in a broader sense, to cover alsocases which are not eliminations. 5Two kinds of reduction are commonly recognised: (1) Philosophical or epistemologicalreduction, which proceeds by the definition of the (non-logical) terms ofone language by the terms of another language. The typical requirement ofepistemological reduction is that the terms of the reducing language denote entitiesthat are primary relative to the entities denoted by the terms of the reducedlanguage. (2) Scientific reduction, which proceeds by the deduction or derivationof one theory (or science) from another theory or (science). 6 Epistemologicalreduction has been treated in chapter 6 ( pp. 168–78). In this section our concernis scientific reduction.Kenneth Schaffner distinguishes four approaches to scientific reduction. Theyare: (1) the Nagel-Woodger-Quine paradigm; (2) the Kemeny–Oppenheimparadigm; (3) the Popper–Feyerabend–Kuhn paradigm; and (4) the Suppesparadigm (Schaffner, 1967:138f). Of these, the Suppes paradigm may beomitted, since, according to Schaffner, it can be shown to be a weaker form ofthe Woodger–Nagel–Quine paradigm (the Nagel paradigm, for short), ‘in fact soweak as it stands that it will not do as an adequate reduction paradigm’(Schaffner, 1967: 145).Of the various approaches to reduction, Ernest Nagel’s is undoubtedly themost influential. Virtually all subsequent discussions about reduction take theirpoint of departure in this approach. According to Nagel, reduction is formally arelation of logical entailment between two theories (or sciences). One theory, thesecondary, is said to be reduced if it can be deduced or derived from another,primary, theory (Nagel, 1949: 119; 1961: 352). Nagel distinguishes two types ofreduction. Homogenous reduction, where the primary and secondary theories (orsciences) are about the same type of phenomena, and where the secondarytheory (or science) employs only such descriptive terms as occur also in theprimary theory (or science), and with approximately the same meaning. Typicalfor homogenous reduction is that the primary theory is more general than thesecondary theory and includes the latter as a special case, valid within certainboundary conditions. An example of a homogenous reduction is the explanationof Galileo’s laws for freely falling bodies – which apply to the boundary conditionsobtaining at the surface of the earth – by Newtonian mechanics. Thesecond type is heterogeneous reduction, where the primary and secondary theories(or sciences) are of qualitatively different kinds, and where the secondary theory(or science) employs descriptive terms, not to be found in the primary theory (orscience). The most important case of heterogeneous reduction is microreduction,where the primary theory (or science) refers to entities and processeswhich are parts of the entities and processes referred to by the secondary theory(or science). Nagel’s primary concern is with heterogeneous reduction (1949:102–4; 1961: 338–42).For reduction to be possible, certain formal conditions must be fulfilled. In the

326 Why methodological individualism?case of heterogeneous reduction, where the primary and secondary sciencesemploy different vocabularies, there are two conditions, in particular, which arenecessary for the reduction of the latter to the former. There is, first, the conditionof connectability. A link must be postulated between those terms of the secondaryscience that do not occur in the primary science and the theoretical terms of thelatter. Second, there is the condition of derivability, which says that all the laws ofthe secondary science must be derivable from the laws of the primary scienceand the postulated links between their respective vocabularies. Concerning thestatus of these links, Nagel leaves open, the question whether they are coordinatingdefinitions or factual assumptions, or both (1949: 11ff; 1961: 345–58).These formal conditions are not enough, however, for a reduction to count asa scientific achievement. Above all, they fail to sort out genuine scientific reductionsfrom merely philosophical reductions; that sort of linguistic exercise whichuses definition as its sole means and proceeds by replacing the terminology ofone theory, or language, with that of another. To these formal conditions, therefore,Nagel adds the non-formal requirement of independent empirical supportfor the theoretical assumptions of the primary science (1949: 123ff; 1961:358–66).Kemeny and Oppenheim are critical of Nagel’s approach to reduction. Theynotice that of Nagel’s two formal conditions for reduction, that of derivability isreally superfluous, since definability guarantees derivability. Not in the sense thatthe secondary theory is necessarily derivable from previously accepted laws ofthe primary science, but in the sense that, after definition, we are left with laws,the theoretical terms of which all belong to the primary science. This critique isunjust, since it ignores Nagel’s non-formal condition of reduction. Kemeny andOppenheim meet the problem by advancing the requirement that the reducingtheory must be at least as well systematised as the reduced theory.Another critique of Nagel’s approach is that it ‘ignores the fact that the oldtheory usually holds only within certain limits, and even then only approximatively’(Kemeny and Oppenheim, 1956: 13). Kemeny and Oppenheim,therefore, drop the condition of derivability altogether, and instead centre on thecapacity of the reducing theory to explain the same observational data as thereduced theory. One theory T2 is said to be reduced to another theory T1 whenthree conditions are fulfilled: (1) T2 has primitive terms not in T1; (2) all observationaldata associated with T2 can be explained by T1; (3) T1 is at leastequally well systematised as T2 (p. 13).The Kemeny–Oppenheim approach to reduction must be deemed inadequateon several accounts. First, it is based upon the doubtful distinction betweenan observational and a theoretical language. Second, it is based upon an instrumentalistview of scientific theories that squares badly with the idea of‘reduction’. Implicit in the notion of ‘reduction’ is that there is an asymmetricrelation between primary and secondary theory. In the Kemeny–Oppenheimapproach, asymmetry is afforded primarily by degree of systematisation. Butsince Kemeny and Oppenheim allow of ‘reductions’ where the primary andsecondary theories are equally well systematised, we are left with the possibility

Why methodological individualism? 327that both theories may be reduced to each other, the only difference being theirvocabulary. 7An important point made by Kemeny and Oppenheim, however, is thatactual scientific reductions are seldom strict derivations. This point has recurredin the later critique of Nagel’s approach to reduction, and especially, in thecritique by Paul Feyerabend. According to Feyerabend, most cases taken fromthe history of science and cited as examples of reduction do, in fact, show thatthe ‘reduced’ theory is not derivable from, but inconsistent with, the ‘reducing’theory. It is a matter of replacement rather than of inclusion. What can bederived from the primary theory is, at best, an approximation to the secondarytheory (Feyerabend, 1962: 43–8).Confronted with this critique, Schaffner suggests a new approach to reduction,which retains the basic idea of Nagel while incorporating the critique ofFeyerabend. According to Schaffner, what must be derivable from the primarytheory is not the original secondary theory, but a corrected version of thesecondary theory. The primary theory, furthermore, must be able to indicate inwhat respect the secondary theory is incorrect, and why it worked as well as itdid. Reduction, then, becomes a combination of replacement and inclusion.Concerning the status of the connecting link between the terms of the primaryand secondary theories, Schaffner suggests that they are synthetic identities(Schaffner, 1967: 144).This concludes my presentation of the different approaches to reduction. Butthere are also different types of reduction. Nagel mentioned two, homogeneousand heterogeneous reduction, and among heterogeneous reductions, especiallythose which involve phenomena that are microscopic relative to some other,macroscopic phenomena. Reductions of this type have become known as ‘microreductions’.Micro-reduction has been discussed by, among others, Oppenheimand Putnam (1958). 8 According to them, the essential feature of a microreductionis that the objects dealt with by the primary science are parts of theobjects dealt with by the secondary science. For their general concept of ‘reduction’,Oppenheim and Putnam rely on Kemeny and Oppenheim (see above).One advantage with micro-reduction is that it affords asymmetry to theKemeny–Oppenheim approach. Another advantage, according to Oppenheimand Putnam, is that it helps to bring about the ideal of a unitary science, in thestrong sense of unity of laws (Oppenheim and Putnam, 1958: 6–8).The idea of micro-reduction is part of a more comprehensive view of theuniverse as organised into a hierarchy of reductive levels, where each level is theproper domain of a scientific branch or discipline. The perfect reduction wouldbe accomplished if all sciences could be reduced to the science dealing withobjects at the lowest level, that is to physics. Oppenheim and Putnam entertainthe possibility of realising this ideal as a working hypothesis. A first step in therealisation of this ideal is the reduction of theories about social groups to theoriesabout ‘living things’, of which human beings is a sub-class. In support oftheir working hypothesis, Oppenheim and Putnam cite, not surprisingly, theexamples of Mill, Weber and economic theory, but, more surprisingly, also Marx

328 Why methodological individualism?and Veblen. It may also be noticed that they regard methodological individualismas the special form taken by their working hypothesis when applied tosocial groups. According to Oppenheim and Putnam, then, methodological individualismhas everything in common with the idea of micro-reduction (pp.9–18).One form of micro-reduction has been of special importance to the discussionabout methodological individualism. According to May Brodbeck, thereduction of sociology to psychology is accomplished, and the promise ofmethodological individualism fulfilled, if a composition law can be found whichexplains the behaviour of groups from that of individuals.In addition to elementary laws telling how an individual acts in the presenceof one or a few others, we must also have composition laws stating whathappens under certain conditions, as the number of people he is withincreases. The latter, of course, states how he behaves in a group … Giventhe composition laws, the reduction of sociology to psychology is a purelylogical matter, following as it does from these composition laws jointly withthe definitions.(Brodbeck [1958] 1968: 299)The idea of composition laws goes back to C.D. Broad and, more recently, toGustav Bergmann. A composition law is a general principle that explains thebehaviour of complex systems, or configurations, from the behaviour of theirelements. The obvious example of a composition law is the well-known parallelogramlaw for vector addition, used in mechanics to explain the behaviour ofcomplex systems from the behaviour of two-body systems in isolation. Thenotion of ‘composition law’ has also been used to explicate the concept of‘emergence’. Emergence occurs in the absence of composition laws, or if acomposition law should break down at a certain level of complexity, i.e., whenthe number of elements exceeds a certain limit. Emergence, in this sense, is relativeto our knowledge of composition laws. A theory which is emergent todaymay become reducible tomorrow, as a result of the discovery of a new compositionlaw. 9Reduction with the aid of composition laws has a lot in common with theclaims of methodological individualism. (1) The paradigm case of a compositionlaw is the parallelogram law of mechanics, and, according to Watkins, methodologicalindividualism is the social science analogue of mechanicism (Watkins,1957a: 104f). As we have seen in previous chapters, many methodological individualists,starting with John Stuart Mill, have stated their views of explanationin mechanistic terms. (2) Composition laws are elementaristic in the sense that theyexplain the behaviour of complex wholes from the behaviour of their elements.The elements need not be single objects, but may be systems of two or moreobjects. (3) Composition laws explain the behaviour of several objects in interaction.(4) Composition laws are additive, and so, falsify, where found, theproblematic saying that ‘the whole is more than the sum of its parts’. I suggest

Why methodological individualism? 329that we may conceive of composition laws in social science in terms of aggregation.Composition laws are those ‘laws’ that solve the so-called ‘aggregationproblem’, while the statement that the whole is more, or other, than the sum ofits parts is a claim that the aggregation problem cannot be solved. 10But is explanation with the help of composition laws really reduction? Notaccording to Bergmann. On his view, reduction is a relation of derivationbetween two theories, while explanation with composition laws involves only onetheory (Bergmann, 1944: 67f; 1957: 162ff). According to Brodbeck, however, thereduction of sociology to psychology presupposes both (a) the definition of allsocial concepts by individualist concepts and (b) the explanation of thebehaviour of individuals in groups from that of more elementary individualbehaviour with the help of composition laws (Brodbeck [1958] 1968: 297f). Oneof these conditions seems to be redundant. The translation of all social conceptsto individualist concepts, gives us a (trivial) terminological reduction even in theabsence of composition laws. 11 If, on the other hand, there are composition lawsfor every configuration of individuals, there is no reason to establish a connectionwith previous sociological theory. There is, however, the possibility ofinterpreting this case as ‘reduction’ in the sense of Kemeny and Oppenheim; thereplacement, within a certain domain, of one theory by another theory. In thesocial sciences, where formalisation is rare, and derivation, therefore, difficult toachieve, it seems wise to admit of reductions that are not derivations, but thereplacement of one theory by another, especially when the reducing theory is amicro-theory relative to the reduced theory.To conclude this discussion, I will quote Helmut Spinner on the epistemologicaland ontological assumptions behind reductionism, because I believe that hecatches the ‘faith’ and ‘spirit’ of reductionism in a most telling way.The reductionist pattern of scientific development is a picture … of steadyprogress towards a unique set of explanatory principles or laws of ever-increasing generality,empirical content, systematicity, and comprehensiveness – a pattern of continuous,cumulative growth of knowledge resulting in the last resort in a nomological unity ofscience, physicalist or otherwise.(Spinner, 1973: 45)Also, the predominant logico-epistemological level-picture has been supplementedby, and based on, an ‘ontological’ doctrine. This additionalontological assumption postulates, in parallel with the logico-epistemologicalhierarchy of theories, a corresponding ontological hierarchy of reductive levels,based on a part-whole relation, in the order of the ‘things’ … Each of theselevels constitutes a characteristic proper domain of some specific empiricalscience.(Spinner, 1973: 56)What about methodological individualism, then? Is it a species of reductionism?I believe that it is. As we have seen in previous chapters, methodological

330 Why methodological individualism?individualism has been conceived of as a principle about the explanation ofempirical laws, the definition of collective concepts, and the explanation ofsocial phenomena (see also pp. 349–54). The first version, about the explanationof social laws, corresponds to Nagel’s approach to reduction. The secondversion, about the definition of concepts, corresponds to the scientifically trivialphilosophical reductions, which we met with in chapter 6. The third version,about the explanation of social phenomena, finally corresponds roughly to theKemeny–Oppenheim approach to reduction, especially as individualist theoriesare often held to be both formally and empirically superior to the holistic theoriesthey are supposed to replace. 12 It might be added that the strong version ofmethodological individualism is an instance of strong reductionism, while theweak version of methodological individualism is a case of weak reductionism.Some qualification must be made with respect to this conclusion. As MurrayWebster (1973: 266) has pointed out, methodological individualism denies thepossibility of social theory being anything but individualistic, and so, there seemsto be nothing to reduce. Also George Homans occasionally denies that there areany social laws to reduce (Homans, 1969: 15–17), if not always (Homans, 1970a:324f; 1983: 38–43). Most methodological individualists deny that there aregenuine, causal, social laws, but accept the existence of empirical laws, or generalisations.It is still possible to be a reductionist about empirical social laws andthis is exactly what many methodological individualists are.Another problem with reductionism in social science is that there are no theories,which have the formal properties, necessary for reduction to take the formof a strict derivation of one theory, e.g. sociology, from another theory, e.g.,psychology, and yet many methodological individualists say that they are reductionists:J.W.N. Watkins, George Homans, Hummel and Opp, and Jon Elster.The explanation, must be that they use the term ‘reduction’ in a less demandingsense. Probably they intend something like this: whatever can be truly said aboutsocial phenomena can, in principle, be derived from statements about individualsand their actions. Most methodological individualists avoid this roundaboutway, however, and go directly to social phenomena. For them methodologicalindividualism is not a principle about reduction of social theories, but about theexplanation of social phenomena.With these qualifications in mind, I still believe that methodological individualismis a form of reductionism. But why are some social scientists reductionists?A clue to an answer can probably be found in the first quotation from Spinnerabove. Reduction is, for some social scientists, the highroad of Science. This beliefis clearly expressed by Jon Elster, who also believes that reduction has beenachieved in the natural sciences. 13 This is a surprising claim, however. Theprevalent opinion of virtually all philosophy of science is that reduction, in thestrong sense, has not been achieved in one single case, in the natural sciences, 14and it is reduction in the strong sense that Elster wants to see in the socialsciences (see above).The recent discussions about reduction have been much influenced by theidea of supervenience, propagated most successfully by the philosopher Jaegwon

Why methodological individualism? 331Kim. Supervenience has to do with the relation between levels in the reductivehierarchy, and especially between the brain and the mind, but also between individualand society. Applied to the former relation, supervenience means that iftwo individuals are exactly alike in their brains, they are also exactly alike in theirminds, although the reverse does not hold. There may be multiple realisations ofbrain-states in the mind. Applied to the relation between individual and society,this means that there are many microfoundations of a certain macro-state, orevent. There are many different ways in which to make a revolution, win an election,exert authority, marry, etc. At first Kim had reductionist ambitions with theconcept of supervenience (1978), but pretty soon he was led to embrace holism(1984). The idea of supervenience is obviously relevant to the issue of methodologicalindividualism, and it has been frequently used in arguments for andagainst methodological individualism. Since it is not my aim, here, to argue,either for, or against, methodological individualism, I will not go any further intothe doctrine of supervenience and its uses, except to suggest that it was first usedmainly in defence of methodological individualism, but later against it. 15Those methodological individualists who are reductionists, seem to believethat social science is reducible to psychology. 16 Methodological individualism,then, would be identical with psychologism.PsychologismIf methodological individualism is the principle that social phenomena must beexplained in terms of the motives, beliefs and actions of individuals, then it lookspsychologistic. As we have seen there are also some methodological individualistswho accept this conclusion and many more commentators and critics, who alsotake it more or less for granted. 17 But many methodological individualists denythe identity of methodological individualism and psychologism, and some ofthem emphatically so. How is this possible? I will take a brief look at the argumentsof those methodological individualists, who have denied the identity ofmethodological individualism and psychologism, but before I do, a few wordsabout psychology.First of all, there is psychology as a science, existing or potential. Scientificpsychology may be divided in various ways. There is, for instance, physiologicaland mental psychology, individual and social psychology, experimental and nonexperimentalpsychology, explanatory and descriptive psychology, etc. There arealso a number of different paradigms within each of these branches, which I willnot try to enumerate here. Second, there is common sense, or folk psychology,consisting of the (psychological) beliefs people entertain about the minds andbehaviour of themselves and others. For the social scientist, however, folkpsychology is interesting, not so much as a contribution to psychological theory,but as an object of investigation. Since people act in accordance with theirbeliefs, these beliefs are of primary importance for all human sciences. Third,‘psychology’ may mean this stuff itself; the totality of ‘psychic’, or mental

332 Why methodological individualism?phenomena, including sensations, ideas, emotions, volition, etc., which is thedomain, or possible object of investigation of psychology as a science.I return to the arguments methodological individualists have advancedagainst the view that methodological individualism is psychologistic and startwith Max Weber. It would seem that Max Weber’s insistence upon explanationin terms of motives renders his methodology inescapably psychologistic, and yet,he tries hard to escape this conclusion. 18 How can this be? The main answer, Ibelieve, is that Weber, like his neo-Kantian friend Heinrich Rickert and manyothers at that time in Germany (see chapter 2), conceived of psychology as anexperimental natural science; in the case of Weber, as psycho-physics, orbehaviouristic reflex psychology. Above all in his early writings, Weber frequentlyrefers to psychology as ‘experimental psychology’ or ‘psycho-physics’ (seeSchluchter, 2000). His argument seems to be that the cultural sciences deal withmeaningful phenomena and are, therefore, independent of scientific psychologyas a natural science. The cultural sciences rely, instead, upon common sense or‘vulgar psychology’. 19This does not mean, however, that the cultural sciences are, or will everbecome, reducible to common sense psychology, or any other psychology for thatmatter. Weber’s decisive argument against psychologism seems to be that thecultural sciences have another subject matter and utilise another method thandoes psychology. The cultural sciences proceed by the construction of ideal typesof action, not by an investigation of the psychic life of individuals. Economicsand sociology in particular, make use of the ideal type of perfectly rational, orobjectively correct, action, as distinguished from the concrete, usually irrational,actions of individuals. 20I do not find these arguments entirely compelling. In an earlier work (Udehn,1987: 66f) I have rejected, as untenable, the view that scientific disciplines areconstituted by their subject matter and, by implication, also the view that theyare constituted by their method. I maintained that the ‘essence’ of a scientificdiscipline is its typical explanatory factors, and on this point Weber is perfectlyclear: social phenomena are to be explained in terms of the motives of individuals;indisputably a psychological factor. Weber’s attempt to escape psychologismwith the help of the ideal type is also far from convincing, even on his ownaccount of this much debated methodological tool. There is a certain ambiguityabout the epistemological status of the ideal type. According to Weber, it seemsto be both a heuristic device and a hypothesis subject to verification. 21 This willnot do. Verification, if possible at all, is possible only with respect to hypotheses,but not with respect to heuristic devices. 22 What is more, the verification of idealtypes, according to Weber, takes place with the help of psychological experiments(Weber [1913] 1981: 157; [1922] 1978: 10). The use of ideal types, assuch, therefore, is not incompatible with psychologism. It all depends upon whatkinds of ideal types.The conscious effort to distinguish methodological individualism frompsychologism had much to do with the attempt, among philosophers and socialscientists in Weber’s Germany, to free themselves and their disciplines from any

Why methodological individualism? 333influence from psychology (see chapter 3). In the case of Weber, it was anattempt to free economics and sociology from any dependence upon psychology.This is the background to Weber’s rejection of psychologism and it is a backgroundhe shared with the members of the Austrian School of Economics.By and large, the first generation of Austrian economists were psychologicists,Carl Menger and Friedrich von Wieser, in particular. The second generation,however, shared Weber’s attitude to psychology. They wanted to replacepsychology with rational choice. According to Ludwig von Mises, for instance,theoretical social science takes the form of praxeology, or the general theory ofhuman action. Praxeology is individualistic, but does not rely upon psychology.Mises gives two reasons for praxeology’s independence of psychology. First,psychology deals with the ‘inner’ aspect of action, what goes on in people’sminds, while praxeology deals with action as such, no matter what are themotives behind the action (1933: 3; [1949] 1966: 12). Second, praxeology hasanother method and another epistemological status than psychology. The latteris an ‘historical’ or ‘realistic’ science, while the former is strictly formal, like logicand mathematics (Mises, 1933: 10ff; [1949] 1966: 32ff).Mises’s first argument is easily dismissed as depending for its validity upon amuch too restrictive delimitation of psychology. His second argument is more ofa problem. It raises the highly controversial issue of the nature of praxeology,and of its most important branch: economics. Mises is, perhaps, the most wellknownexponent of the view that economics is a priori; that the propositions ofeconomics are independent of empirical reality both in origin and for theirvalidity (Mises, 1933: 23ff; [1949] 1966: 32ff; 1962: 1ff). If this means that thepropositions of economics are analytic truths, economics is neither individualistic,nor psychologistic. Analytic statements do not explain anything whatsoever.Another possibility is that economics is an example of the Kantian synthetic apriori. If so it may not be empirical science, but its content might be ‘psychological’.Most likely, Mises’s praxeology is an eidetic science, in the sense ofphenomenology. If so, Mises has the support of Husserl when rejecting psychologism,but it is far from obvious that Husserl succeeded in his attempt to freephenomenology from psychologism (see chapter 3).Friedrich von Hayek does not share Mises’s view about the epistemologicalstatus of economics, but he agrees that methodological individualism differsfrom psychologism. The task of psychology, according to Hayek, is to explainconscious action. The social scientist, on the other hand, takes conscious actionas his data. His task is to explain the unintended consequences of the actions of manymen (Hayek, 1955: 39). This argument carries some weight. Psychologists do notseem to be interested in large-scale social phenomena. But I do not believe thatthe occurrence of unintended consequences per se is sufficient ground for denyingthat methodological individualism is psychologistic. As I have already suggested,it is not the subject matter that decides the issue, but the explanations.What about Karl Popper then? As we have already seen, Popper is a critic ofpsychologism and a defender of an autonomous sociology. Unfortunately,Popper’s main argument against psychologism – that social science must take

334 Why methodological individualism?account of institutions – is an argument also against methodological individualism,and cannot serve as an argument for a difference between these twodoctrines. A second argument, that the social sciences use a ‘logical’ or rationalist,rather than a psychological method, was deemed untenable when used byWeber, and does not appear any more convincing in the case of Popper. 23 Athird argument, from the unintended consequences of men’s actions, was alsorejected when advanced by Hayek. There remains a fourth argument againstpsychologism, to the effect that social phenomena cannot be explained in termsof human nature alone, but require mention of the environment or situation andthe interaction between men as well for their explanation. 24 This argument isbased upon the tacit presupposition that psychologism is forced to explain socialphenomena in terms of the human nature of isolated individuals. 25 This is notan acceptable view of psychology, at least not of social psychology. It does notseem to be Popper’s view either. According to him, psychologism is the ‘theorythat sociology must in principle be reducible to social psychology’ (Popper [1945]1966: vol. 2, 88), or the doctrine that society is ‘the product of interacting minds’(p. 90). Popper is equally confused about psychology and psychologism, as aboutmethodological individualism.Among Popper’s followers, views about psychologism are divided. Predictably,all except J.W.N. Watkins, reject the view that methodological individualism isidentical with psychologism, but since they are institutional individualists, this isperfectly understandable and is no threat to the thesis that the strong version ofmethodological individualism is psychologistic. J.W.N. Watkins, on the otherhand, states methodological individualism as a psychologistic principle of explanation.Methodological individualism, he says, is the principle ‘that the socialscientist can continue searching for explanations of social phenomena until hehas reduced it to psychological terms’ (1952a: 29). He also ‘suppose[s] that socialtheories derive sociological conclusions from psychological premisses’, therebystating the formal condition for reduction (p. 34). He goes as far in a psychologistdirection as to suggest that social institutions are certain kinds of psychologicaldispositions and, as such, belong to man’s personality (p. 39). In his later writingson methodological individualism, Watkins retreats from his original position anddenies that methodological individualism is psychologistic, but he is able toperform this manoeuvre only with the help of an absolutely impossible versionof psychologism. It is supposed to be the view ‘that all large-scale social characteristicsare not merely the intended or unintended results of, but a reflection ofindividual characteristics’ (1957b: 112).I conclude that despite arguments to the contrary, the strong version of methodologicalindividualism, or individualist reductionism, is indistinguishable from psychologism.It is true that methodological individualists are usually interested in otherphenomena than are psychologists. Usually, they are interested in large-scale, ormacro-phenomena, which rarely catches the attention of psychologists. It mightstill be that their explanations are the same. If so, methodological individualistsmay be conceived of as large-scale psychologists.It is also true that methodological individualists rarely use psychological

Why methodological individualism? 335assumptions borrowed from the science of psychology. They tend to invoke theirown psychological principles of explanation borrowed from folk psychology, ordeveloped in their own sciences. Economists, for instance, make use of assumptionsof rationality and utility-maximisation and self-interest. These assumptionsare not borrowed from the science of psychology, and so, many economists denythat economics depend on psychology. This is, of course, true, in a sense, buttheir own principles of explanation are also ‘psychological’ in content.A related reason why many social scientists deny being psychological reductionistsis that they rely on very simple psychological assumptions, and sometimeson assumptions that only approximate the behaviour of individuals. But onceagain, this does not make the assumptions other than psychological. We mightsay that most individualistic explanations of social phenomena make use of thinpsychological assumptions, as distinguished from the thick psychology of mostexplanations in the science of psychology (cf. Lindenberg, 1996). There are alsoimportant differences between methodological individualists, themselves, in thisrespect. As we have seen in the previous chapter, James Coleman advocates athin psychology, whereas Raymond Boudon and Jon Elster advocate more thickdescriptions of psychological mechanisms.I have argued that strong methodological individualism is psychologistic, but Ihave remained silent on the distinction between individual and social psychology,and for a very good reason. It is not at all easy to make this distinction.According to Jean Piaget (1973: 29): ‘Social psychology comes into all thegeneral problems to do with psychology (differential psychology, personality, andso on), since man is essentially a social being’ (see also Harré 1986: 94ff andMargolis, 1986: 31ff). Following this suggestion by Piaget, we may conceive ofindividual psychology, as based on the idea of an abstract individual, with aconstant human nature, and social psychology as based on the idea of thehuman individual as a social being. This makes it possible to argue that thoseversions of methodological individualism, which are based on the idea of aconstant human nature, e.g., the theory of the social contract and the theory ofgeneral equilibrium, imply a reduction to individual psychology, whereas thoseversions that assume that individuals are social beings, e.g., Austrian methodologicalindividualism, imply a reduction to social psychology.I have argued that strong methodological individualism is psychologistic, butis psychology necessarily individualistic in the strong sense? I take it for grantedthat individual psychology, if such a thing is possible, is individualistic in thestrong sense, but what about social psychology? This branch of psychology hasits roots both in psychology and sociology, but probably more in the latter(Allport, 1924b: ch. 5; Moscovici, 1972: 49). Partly as a consequence of this, it ispossible to distinguish a sociological social psychology from a psychological socialpsychology (Farr, 1978: 510; Secord, 1986), the former being ‘top-down’, thelatter ‘bottom-up’ (Margolis, 1986: 38–46). The latter is interested in the elementaryprocesses of interaction between individuals, while the former focus on theeffects of social structure upon consciousness and behaviour. With W. PeterArchibald (1976: 115–19), I believe that the split between a psychological and a

336 Why methodological individualism?sociological psychology reflects the opposition between individualism and holism(see also Tajfel, 1972: 86ff and Doise, 1978: 28ff). The demarcation betweenindividualism and holism, then, cuts right through the discipline of socialpsychology. This fact is reflected in most textbooks on social psychology, whichtend to be a mix of psychology and sociology, rather than a clearly demarcatedarea of its own (Moscovici, 1972: 43–8).The distinction between a psychological and a sociological social psychologyshould not be confused with that between small group psychology and masspsychology. Number is not a variable that separates sociology from psychology.There is a sociological small group psychology, as well as a truly psychologicalmass psychology (cf. Sztompka, 1979: 100). The main bulk of contemporarysocial psychology is devoted to the study of small groups, but early socialpsychology was more interested in the psychology of nations, masses, crowds,publics and other collectives (see, e.g., Ginsberg, 1928 and Blumer, 1937: 144–6).MicrofoundationsIn the recent development of methodological individualism, it has often beendiscussed as a case of microfoundation. The reason for this is probably‘economic imperialism’. In economics, the clearest manifestation of methodologicalindividualism has for a long time been the request that macroeconomicsshould be provided with microfoundations. When social scientists from otherdisciplines adopt the economic approach, they also adopt the view of methodologicalindividualism as the demand that macro-theories must be provided withmicrofoundations in the form of rational choice (see chapter 10). But is the questfor microfoundations the same as reduction? It depends.In the case of economics, it is typically assumed that the proper microfoundationsof macroeconomics are Walrasian. This assumption makes the quest formicrofoundations in economics a case of strong reductionism. But it is not necessaryto conceive of all macrofoundations in this way. The failure, so far, toprovide macroeconomics with Walrasian microfoundations, may even make usdoubt the very possibility of strong reduction in the social sciences, as in allsciences. An alternative would be to conceive of microfoundations as partial andmore like a case of weak reduction. I think a strong case could be made for theview that a microfoundation is not the same as a reduction, in the original sense ofelimination. The very word ‘foundation’ has a meaning different from the words‘reduction’ and ‘elimination’. If we conceive of microfoundations in this ‘nonreductive’way, they become fully compatible with institutional and structuralindividualism. 26Normative individualismI suggested above, that methodological individualism is based on certain epistemologicaland ontological assumptions. If we include these assumptions in thedoctrine of methodological individualism, it is better conceived of as an individ-

Why methodological individualism? 337ualistic research programme, in the sense of Imre Lakatos. According to SusanJames (1984), however, there are deeper motives hidden behind the adoption ofmethodological individualism and holism respectively. Behind the methodologicalissue, there is an ontological issue, and this issue has importantethico-political implications. According to James (p. 56), this explains the heat ofthe controversy between individualism and holism, and the irreconcilability ofthe parties to the dispute. 27I agree that a full understanding of the debate between individualists andholists is not possible without an understanding of the moral and political issuesinvolved. I do not agree, however, that the methodological issue is barren, or thatit reduces to the ideological issue, as James (1984: 54ff) seems to suggest. Iassume that the methodological issue is a real issue, open to rational discussion,even though it may never be ‘solved’ once and for all. I also assume that themethodological issue can be discussed independently from the ideological issue.To believe otherwise is to commit the naturalistic fallacy. Nevertheless I find itimpossible to deny that there is some kind of connection between methodologicaland normative individualism.To claim that there is some kind of relation between methodological andnormative individualism is of course problematic, if we believe, as I do, thatthere is a categorical distinction between facts and values. Even so I will make noattempt to clarify this relation, since I know that such an attempt would lead to alengthy argument. I may invoke the support of methodological individualist,J.W.N. Watkins (1957a; 112–4; 1958; 314, 357f; 1957/8: 80f), however, who hasargued that, besides analytical and empirical statements, there are in science aclass of statements which are synthetic, but irrefutable and, therefore, metaphysical.To this class of statements belong, what he calls ‘haunted universedoctrines’. Such metaphysical doctrines play a regulative role in science, and insocial science they often comprise, or suggest, moral and political ideas.Examples of haunted universe doctrines are, in natural science, determinism andmechanism, and, in social science, historicism and individualism.A problem with suggesting that methodological individualism is related tonormative individualism is that the very term ‘methodological individualism’ wascoined, by Schumpeter, to make the distinction between methodological andpolitical individualism. Max Weber also made this distinction and maintainedthat the two doctrines are independent. More recently, Jon Elster has argued thatnothing political follows from methodological individualism (1993: 7). 28Political individualismA first thing to suggest a connection between methodological and political individualismis the fact that the former was originally used by Menger as a weaponin the battle of methods (Methodenstreit), which was perhaps first of all a politicalbattle (see, e.g. Alter, 1990; Cubeddu, 1993: 21ff).The second thing to suggest a connection is that, until quite recently, therewas a very strong statistical correlation between methodological and political

338 Why methodological individualism?individualism (see Buchanan [1962] 1965: 315). Today this correlation is probablynot equally strong, but strong enough to suggest some kind ofinterdependency.Most methodological individualists have been, or are, ‘liberals’ in some senseof that term. John Stuart Mill was not only a forerunner to the methodologicalindividualists, but one of the most important figures in the history of liberalism.Carl Menger was influenced by Mill, and a liberal himself, but with a certainleaning towards conservatism. Weber was also a liberal of sorts; a liberal in theGerman sense of that term, which is something rather different from a ‘liberal’ inthe Anglo-American sense. Mises and Hayek are among the most important andenergetic defenders of a classical or laissez-faire liberalism in our century (see, e.g.Mises [1927] 1978; Hayek [1944] 1962). Popper was, until the latest libertarianvogue, at least, the most well-known representative of neo-liberalism in our times.Popper was not a laissez-faire liberal, however, but saw the need for state interventionand the welfare state, as long as it restricts itself to ‘piecemeal socialengineering’, but refrains from ‘utopian social engineering’ ([1945] 1966; 1957).More recently, methodological individualism has been advocated by membersof the Chicago and Virginia Schools of political economy, all of whom seem tobe political individualists.The most glaring exception to the ‘rule’ that methodological individualists arealso political individualists is Analytical Marxism, but this exception is notenough to destroy the pattern.The question arises: Is the correlation between methodological and politicalindividualism spurious, or is there some kind of connection between methodologicalindividualism and liberalism? Many methodological individualistsobviously believe that it is.According to Mises, liberalism is the application of praxeology; the science ofhuman action, to social life. In some of his early writings Mises maintains thatliberalism, because supported by science, is value-free, but in Human Action heconcedes that ‘As a political doctrine liberalism is not neutral with regard tovalues and the ultimate ends sought by action’ ([1949] 1966: 154; see also1943–4: 541–5). According to an already cited passage from Hayek, true individualismis primarily a theory of society, an attempt to understand the forces whichdetermine the social life of man, but also ‘a set of political maxims derived fromthis view of society’ (Hayek, 1948: 6). The set of political maxims referred to is,of course, liberalism. But a similar relation exists between methodological andpolitical collectivism. ‘The collectivist method … leads … directly to politicalcollectivism; and though logically methodological and political collectivism aredistinct, it is not difficult to see how the former leads to the latter’ (Hayek,[1942–4]; 1955: 91f).What then, more precisely, is the relation between methodological individualismand liberalism? First of all, it is not a relation of logical necessity. You mayaccept one of these doctrines without being logically compelled to accept theother. Let us say there is a certain ‘elective affinity’ between methodological individualismand liberalism. Both doctrines rely on a certain view of society, which

Why methodological individualism? 339is for the former a metaphysical presupposition, and for the latter a politicalideal; a view of society as made up of autonomous agents interacting with eachother (making contracts) for their mutual benefit. 29 The actual society mostclosely approximating this ideal society – ‘ideal’ both in the epistemological andethical sense of that term – is the free market society of early capitalism, if thereever was a free market society (cf. Weber, 1949: 44). It is no coincidence thatorthodox economics is both the theory of that society and the most often citedexample of methodological individualism (cf. Schotter, 1985: 1ff).Methodological individualism and liberalism also have an epistemologicalargument in common. Methodological individualism denies that we can haveknowledge about society as a whole or as a totality. But if we cannot have suchknowledge, then, central planning and ‘utopian social engineering’ become, ifnot impossible, then at least dangerous, since we are left in the dark as to theconsequences of such large-scale political measures. Under these circumstances,a free market economy seems preferable to a centrally planned economy and,according to some methodological individualists, also to state interventionismand the modern welfare state. 30HumanismThe supreme perfection of man is that he acts freely or voluntarily, and it isthis, which makes him deserve praise or blame.(Descartes, Philosophical Writings, vol 1: 205)A basic idea of individualism, and of Western thought in general, is that ofhuman freedom; the idea of man as an autonomous being capable of acting rationallyin the pursuit of her own interests. Closely related to the idea of freedom isthat of the dignity of ‘man’. Man alone among the creatures of the earth is freeand rational, created in the image of God. This godlike nature of man is theultimate source of his/her dignity. From the freedom and rationality of manfollows his/her responsibility. From the dignity of man derives his/her naturaland inalienable rights (Lukes, 1973: 43ff).This conception of man originates in Greek Antiquity and finds its mostgenuine expression in Renaissance humanism. In a secularised version, it is thedominant conception of man also in contemporary Western thought, commonsenseincluded. Considering the privileged place it assigns to humankind, thisview is better characterised as ‘humanism’ than as ‘individualism’. In order todistinguish this form of humanism from other forms, I will call it ‘individualisthumanism’. 31Individualist humanism is not a necessary accompaniment to methodologicalindividualism. It is possible to accept the latter without accepting the former, aswitnessed by the existence of behaviourist methodological individualism (see pp.190–9). We may speak, therefore, of a humanist individualism in addition to anindividualist humanism. But if individualist humanism does not follow from

340 Why methodological individualism?methodological individualism, the reverse appears to be the case.Methodological individualism seems to follow from individualist humanism. Ifman is fully autonomous, subject only to the limitations imposed by nature; hisown, as well as that which surround him, it would seem that whatever happensin society is the result, immediate or mediate, of his wilful actions and, therefore,to be explained in terms of individuals etc., that is, according to the strictures ofmethodological individualism. Individualist humanism, then, seems to entailmethodological individualism. But perhaps entail is too strong. Following Watkins,we may say that individualist humanism suggests methodological individualism.Additional light is thrown upon individualist humanism and its relation tomethodological individualism if the above, positive statement of this doctrine issupplemented by a negative one. To what is individualist humanism opposed?Generally speaking, individualist humanism is opposed to every attempt todethrone humankind from its elevated position as the crown of creation, or inthe secularised version; to every doctrine, which reduces the human individual,by explaining away freedom, reason or mind. The human individual, as a free,conscious and rational being is threatened both from above and from below.From above, she is threatened by divine providence and predestination. Ifdestined to carry out the divine plan, mere marionettes in the hands of God (orthe Gods), individual freedom becomes but an illusion. The first expression ofindividualist humanism, therefore, was in the form of an opposition to all formsof theism.From above, the individual is threatened also by historical and social forces,apparently reducing him to a means for the ends of history, a cog in the socialmachine, or an organ in the social organism. Individualist humanism, therefore,is opposed to the doctrines of historicism, holism and collectivism. This is whereindividualist humanism joins forces with methodological individualism.But the threat may also come from below. Above all, humanism is a defenceof the human individual against the recurrent attempts to reduce her/him to apiece of matter; a machine, a moving body, or an organism emitting responses toexternal stimuli. Individualist humanism, therefore, is also opposed to materialism,physicalism and behaviourism. As we have already seen, there is nonecessary relation between methodological individualism and this aspect ofhumanism. As a matter of fact, however, methodological individualists havetended to also endorse this part of humanism. It has often been pointed out thatmethodological individualists, while being reductionists with respect to thatwhich is above individuals (i.e. social phenomena) are usually emergentists withrespect to the human individuals, or her/his mind (see, e.g., Brodbeck, 1954:155).The relation between methodological individualism and individualisthumanism is nowhere, more clearly visible, than in Isaiah Berlin’s essay‘Historical Inevitability’ (1954). Berlin does not mention methodological individualism,but it is clearly present in his text. It is to Berlin that I owe the idea ofilluminating individualist humanism by looking at its opposites. According tohim, it is opposed to physicalism, scientism, historicism, holism, collectivism and

Why methodological individualism? 341environmentalism. But Berlin takes the analysis one step further and asks whatthe doctrines opposed to individualist humanism have in common. He finds thecommon denominator in determinism.All these theories are, in one sense or another, forms of determinism,whether they be teleological, metaphysical, mechanistic, religious, aesthetic,or scientific. And one common characteristic of all such outlooks is theimplication that the individual’s freedom of choice (at any rate here, below)is ultimately an illusion, that the notion that human beings could havechosen otherwise than they did usually rests upon ignorance of facts; withthe consequence that any assertion that they should have acted thus or thus,might have avoided this or that, and deserve (and not merely elicit orrespond to) praise or blame, approval or condemnation, rests upon thepresupposition that some area, at any rate, of their lives is not totally determinedby laws, whether metaphysical or theological or expressing thegeneralised probabilities of science.(Berlin, 1954: 58)In Berlin’s version, then, methodological individualism-cum-humanism isprimarily a voluntarist thesis advanced against all forms of determinism. Berlindoes not suggest, however, that we should give up determinism on the onlyground that it turns freedom into an illusion. We should give up determinismbecause, for the time being, it is based upon mere speculation, and because theconsequences of adopting it are too high a price to pay for mere speculation. Ifdeterminism should turn out to be true and in the end only science, if anything,can tell – we would have to radically change not only our conception of man,but our way of living. The change would be so dramatic as to be almost inconceivable.This is so, because our everyday life and our common-sense way ofthinking is in a fundamental way based upon the tacit presupposition that ‘man’is a free and responsible being (p. 69ff). 32 Even history and social science are to alarge extent based upon this presupposition. When describing and explainingactions in terms of character, purpose and motive, we automatically also evaluatethem.The reason I have paid some attention to the views of Isaiah Berlin is that Ifind them representative of the views of most methodological individualists. Ofthe doctrines mentioned by Berlin as being opposed to individualist humanism,physical determinism has been criticised by Popper in his essay ‘Of Clouds andClocks’, with the telling subtitle ‘An Approach to the Problem of Rationality andthe Freedom of Man’. Popper calls physical determinism a ‘nightmare’, since itimplies thatEverything that happens in such a world is physically predetermined,including all our movements and therefore all our actions. Thus all ourthoughts, feelings, and efforts can have no practical influence upon what

342 Why methodological individualism?happens in the physical world: they are, if not mere illusions, at best superfluousby-products (‘epiphenomena’) of physical events.(Popper, 1972: 217)Popper’s aim is to show ‘how such nonphysical things as purposes, deliberations,plans, decisions, theories, intentions, and values, can play a part in bringing about physicalchanges in the physical world’ (Popper, 1972: 216). J.W.N. Watkins has eventried to show that indeterminism is the central theme, or unifying idea ofPopper’s philosophy, and also that there is a ‘natural coalition’ between indeterminismand falsificationism, and opposed to this, another natural coalitionbetween determinism and inductivism. While the latter coalition ‘depicts man asan induction machine nudged along by external pressures, and deprived of allinitiative and spontaneity’, the former ‘gives him Spielraum to originate ideas andtry them out’ (Watkins, 1974: 407). According to Watkins, then, the idea of thehuman individual as an autonomous being is part also of Popper’s epistemology.In his reply to Watkins, Popper says that he would prefer to see critical realism asthe unifying idea of his philosophy, but admits that Watkins has made a brilliantcase for his interpretation (Popper, 1974b: 1053).Scientism, the doctrine that the social sciences must use the same method as thenatural sciences, has been criticised among methodological individualists, aboveall by Dilthey, Weber, Mises and Hayek. According to them, social phenomenadiffer from natural phenomena in being ideal (Gestige), subjective and meaningful.Therefore, they must be studied with the special method of interpretiveunderstanding (Verstehen). We have already seen in chapter 4, how, in the case ofWeber, Mises and Hayek, methodological individualism was the direct outcomeof their subjectivist methodology. We might add that this is how methodologicalindividualism has sometimes been interpreted; as a thesis about the subjectivenature of social phenomena and the consequent ‘subjectivism’ of social science(Nagel, 1961: 535–46).In the case of Popperian methodological individualism, things are a little bitdifferent. Popper himself is a defender of methodological monism, and so, doesnot accept any radical cleavage between the natural and social sciences. Aboveall, he rejects the idea of a subjectivist social science. According to Popper,subjectivism is out of place in all science, social as well as natural. But he alsorejects scientism, arguing that this doctrine is based on a misunderstanding ofthe natural sciences. Understanding is not a specific method of social science, butthe aim of all science. Popper admits, however, that there is a sense in which wecan understand human phenomena, such as gestures, speech, actions and theproducts of mind, but not natural phenomena, such as solar systems, moleculesand elementary particles. There is, according to Popper, also the method of ‘situationalanalysis’, peculiar to the social sciences, but without a counterpart in thenatural sciences.Environmentalism, according to Mises, ‘is the doctrine that explains historicalchanges as produced by the environment in which people are living. There aretwo varieties of this doctrine: the doctrine of physical or geographical environ-

Why methodological individualism? 343mentalism and the doctrine of social or cultural environmentalism’ (1957:324). 33 Of the latter variety, two in particular; sociology of knowledge andbehaviourism, have been subjected to the criticism of methodological individualists.The trouble with sociology of knowledge is that it seems to imply that ideasare socially determined to an extent that makes free and rational discussionimpossible or illusory. 34The most radical form of social or cultural environmentalism (and scientism,and determinism) is behaviourism. This doctrine depicts man as a passive orreactive being completely determined by his environment, but withoutconsciousness, free will and reason. Behaviourism, therefore, is in obvious oppositionto individualist humanism, and not only implicitly, but explicitly as well.According to B.F. Skinner (the foremost contemporary representative ofbehaviourism) individualist humanism is the main obstacle to the adoption ofthe scientific outlook, and of behaviourism in particular, in the study and runningof human affairs. But in a future beyond freedom and dignityenvironmental contingencies now take over functions once attributed toautonomous man, and certain questions arise. Is man then abolished?Certainly not as a species or as an individual achiever. It is autonomousinner man who is abolished, and that is a step forward.(Skinner, 1974: 210)The behaviourist image of man must appear repulsive to anyone committedto the beliefs and values of individualist humanism. We should expect, therefore,an attitude of hostility towards behaviourism on the part of the majority ofmethodological individualists. This is also what we find. There is a tendencyamong (non-behaviourist) methodological individualists to dismiss behaviourismas too obviously wrong to be worthy of any serious consideration. 35 More specifically,methodological individualists dissociate themselves from behaviourism onthe following issues: most methodological individualists defend a mind-bodydualism against the monism of behaviourism. 36 Most methodological individualistsdefend an activist epistemology against the inductivism of behaviourism.Weber, Mises and Hayek, in particular, insist that social science is about humanaction, as distinguished from behaviour (see chapter 4). Most important, for mypresent purpose, methodological individualists reject behaviourism because,being deterministic, it explains away the unique in man; his consciousness, freewill and reason, and because determinism is coupled with utopian dreams abouta future society based on human engineering, dreams the behaviourist is allegedto share with the believers in holism, collectivism and historicism.Holism, collectivism and historicism; these are the doctrines methodologicalindividualism was originally invented to combat. According to many methodologicalindividualists, holism, collectivism and historicism are, not onlyerroneous, but dangerous. By suggesting that the race, nation, state or class hasinterests of its own, distinct from the interests of its members, and by suggestingalso that this collective is destined to play a special role in history, this trinity of

344 Why methodological individualism?doctrines has served as the intellectual foundation for fascism, nazism andcommunism, and as a legitimation of their crimes. This is the thesis of themethodological individualists (or some of them), and this is why Popper dedicatesThe Poverty of Historicism to ‘the countless men and women of all creeds andraces who fell victim to the fascist and communist belief in Inexorable Laws ofHistorical Destiny’.Holism, collectivism and historicism are dangerous because they lead to acollectivist and historicist ethic, as opposed to the individualist ethic of ourWestern civilisation. This is the message of Mises’s, Hayek’s and Popper’s socialphilosophy and social science methodology.A collectivist ethic is an ethic which makes the interest of some collective theultimate criterion of morality, of what is right and what is wrong. A collectivistethic reduces the individual to a mere means for the ends of the collective.According to Mises ([1957] 1985: 59), collectivism ‘assign[s] to the concerns ofthe collective precedence over the personal wishes of the individuals’, andaccording to Hayek (1944: 111), ‘Once you admit that the individual is merely ameans to serve the ends of the higher entity called the society or nation, most ofthose features of totalitarian regimes which horrify us follow of necessity’.A historicist ethic is somewhat different from a collectivist ethic, but withmuch the same consequences for the individual. A historicist ethic is based uponthe belief that there is progress in history, that the society of tomorrow will bebetter than the society of today. The morally good, therefore, is that which helpsbring about the society of tomorrow, a society that will come whether we like itor not, but with more or less suffering on the way. This historicist ethic has beenascribed by Popper to Marx. It is supposed to be a form of moral positivism: thedoctrine that what is, is good and right, that, ultimately, might is right (Popper,[1945] 1966, vol. 2: 199–211).The intimate link between methodological individualism and individualisthumanism is exhibited clearly in the writings of Joseph Agassi. In his articles onmethodological and institutional individualism, methodological, metaphysicaland ethical considerations are mixed to an extent that makes even an analyticalseparation of them difficult, if not impossible. According to Agassi, the centralthesis of (methodological) individualism is that ‘only individuals have aims andinterests’, also written as the thesis that ‘only individuals have aims and responsibilities’(Agassi, 1960: 244, 248; 1975: 146, 155).The wish to avoid the seemingly deterministic implications of holism, collectivismand historicism is also an important motive for many recentmethodological individualists. It is most obvious, I believe, in the case ofRaymond Boudon, Jon Elster and Philip Pettit, who all reject sociological structuralism,but defend the freedom, or autonomy of human individuals. 37A particularly clear case of an intimate relation between methodological andnormative individualism is the constitutional economics of James Buchanan.While aware of the distinction between methodological and normativeeconomics, at times he suggests that his normative theory of constitutions isbased on methodological individualism, when it is plain that it is rather based on

Why methodological individualism? 345ethical individualism: a doctrine which assigns supreme value to individualhuman beings (see Udehn, 1996, 174ff).ConclusionAmong the motives, or reasons, for being committed to methodological individualism,three seem to be particularly salient: (1) a belief that reduction is the wayof scientific progress; (2) the belief that only methodological individualism iscompatible with humanism; and (3) the belief that methodological individualismgoes together with political individualism. Among the methodological individualists,treated in this book, it is pretty clear that the positivists and rational choiceindividualists are most influenced by reductionism, while the Austrians, membersof the Chicago and Virginia Schools in economics and, to some extent, thePopperians, are motivated by their political persuasion. Humanism seems to bean important motive for methodological individualists of all political persuasions.It is obviously important for the Popperians, but also for many sociologistsand/or Marxists: Jon Elster and Raymond Boudon are, perhaps, the most clearexamples.

12 Methodologicalindividualism restatedIn this last chapter, I will try to bring some threads together and, above all, tosummarise the results of this investigation. The main result is that methodologicalindividualism exists in a bewildering number of different versions. 1 Thissituation has created a Babylonian confusion of tongues, which makes intellectualexchange extremely difficult, if not impossible. In order to discuss the prosand cons of methodological individualism – something I intend to do in a laterwork – it is necessary to keep at least the most important versions of methodologicalindividualism apart.When Joseph Schumpeter coined the term ‘methodological individualism’, hedid not use this term to denote any thesis about social reality, or the proper wayto explain it. As I have interpreted him, he used it to designate the actual procedureof theoretical economics, which is to start from the wants of human beings.I have called Schumpeter’s individualism procedural methodological individualism,as distinguished from the far more common substantive methodological individualism.It is possible to make a further division of procedural methodologicalindividualism, into one theoretical and one empirical variant, but there is notmuch point in it. Procedural methodological individualism is rare and not veryinteresting. The principally important thing is not where you start, but where youend. The most important version of methodological individualism, therefore, isthe substantive version.Strong and weak methodological individualismIn the last decade some leading economists, such as Kenneth Arrow and AlanKirman, have expressed serious concern about the state of economic scienceand, especially, about its crowning achievement, the theory of general equilibrium.They have also more than hinted at the suggestion that the problem witheconomics might be its methodological individualism. At the same time anumber of leading sociologists, such as James Coleman and Raymond Boudonhave expressed a similar concern about sociology. For them, however, methodologicalindividualism appears not as a problem, but as the solution. How is thisdifference possible?A first answer might be sought in the history of the two disciplines. If

Methodological individualism restated 347economics is the most individualistic of the social sciences, and sociology(together with anthropology) is the most holistic, then it seems reasonable toassume that the solution wfor their respective problems lies in a via media, orgolden mean. The main answer to this puzzle, however, is that economists andsociologists mean entirely different things by ‘methodological individualism’.While both Arrow and Kirman think of the original strong version of methodologicalindividualism, Coleman and Boudon think of a version far removedfrom this original. They defend, what I call, structural individualism.The doctrine of methodological individualism, then, ranges from versionswhich require that social phenomena must be fully explained in terms of individuals,to versions which require only that they are partly explained in terms ofindividuals. How large this individualistic part must be is not stated, and cannot bestated, at least not precisely, but it is possible to conceive of a version of methodologicalindividualism, which assigns virtually all explanatory power to socialinstitutions and social structure, and only a small fraction of it to individuals.There has been in the short history of methodological individualism a developmentfrom strong to weak versions of methodological individualism. Thedevelopment has not been linear, but the tendency has been clear. From theextreme versions of methodological individualism, exemplified by the theory ofthe social contract and Walras’s theory of general equilibrium, to the Austrians,who conceived of human beings as social beings, to Popper, who introducedsocial institutions as objective constraints upon action and, finally, to the sociologists,who introduced social structure, made up of positions related to oneanother independently of their occupants. This is an originally holistic idea,which was banned by earlier methodological individualists.Presented in an order of descending individualism, the main versions ofmethodological individualism are these:1 Thetheory of the social contract, which takes as its point of departure the natural(asocial) individual, living without social institutions in a state of nature.2 The theory of general equilibrium, which takes as its point of departure theisolated individual, without social relations interacting on the market, in theabsence of social institutions and technology.3 Austrian methodological individualism, which also starts with assuming theisolated individual, or Robinson Crusoe, but which conceives of humanindividuals as social or cultural beings, attaching subjective meaning to theirown actions and to human artefacts.4 Popperian methodological individualism, which accepts objectively existing socialinstitutions in the explanans, or as exogenous variables, in the social scientificexplanations.5 Coleman’s methodological individualism, which admits of social wholes, in theform of structures of interrelated positions, which exist independently ofthe particular individuals who happen to occupy these positions.I suggest that we call versions 1 and 2, natural individualism, since nothing socioculturalenters the explanans, or exogenous variables, of its explanations. This

348 Methodological individualism restatedversion is sometimes called ‘atomistic’ methodological individualism, but thisdesignation is problematic, since it was used by the Austrians, who entertainedanother conception of the individual.In order to distinguish Austrian methodological individualism from 1 and 2, Isuggest that we call it social individualism, to acknowledge it’s conception of thehuman individual as a social being, and of society as an intersubjective reality,like the theories discussed in chapter 4.I call Popperian methodological individualism, institutional individualism, assuggested by Joseph Agassi. Institutional individualism is the dominating versionof methodological individualism, both in political science and in the new institutionaleconomics, even if many economists cherish the dream of being able toendogenise all institutions.Borrowing a term from some Dutch sociologists, I call Coleman’s methodologicalindividualism structural individualism. This form of methodologicalindividualism is today common and perhaps dominant, among sociological andMarxist methodological individualists.The main divide goes between strong methodological individualism (1–3) andweak methodological individualism (4–5). It is in the move from 3 to 4 that atleast one arrow, representing the direction of explanation and/or causality,changes direction from bottom-up to top-down. This is clearly the most decisivebreak with methodological individualism, as originally conceived. It is a move,which makes it difficult to continue talking about individualism and holism asopposite doctrines. A good illustration of this difficulty is the holistic individualismof Philip Pettit ([1993] 1996: 165ff).Somewhere in between the weak and strong versions of methodological individualism,falls that of J.W.N. Watkins, which seems to have many adherents,most notably, Jon Elster. According to this epistemological version of ‘methodological’individualism, reduction is possible in principle, if not in practice.Because of the complexity of the social world and/or our limited knowledgeabout the mechanisms at work, it might be permissible and maybe inescapableto rest content with explanations in terms of macrophenomena: but onlytemporarily, while awaiting the growth of knowledge about the mechanismswhich make full reduction possible. In the terminology of Watkins, explanationsin terms of macro-phenomena are only ‘half-way’ and second to fully individualistic‘rock-bottom’ explanations. I believe that this version of methodologicalindividualism, even if it implies a weakening of the strong version is still apermutation of the latter. In the words of Watkins, methodological individualismhas changed from being a rule to becoming an ‘aspiration’. Since half-way explanationsdeviate from the ideal of rock-bottom explanations, and are acceptedonly reluctantly, we may call this version of methodological individualism compromisingindividualism.Another form of ‘pragmatic holism’ is represented by the practice amongmethodological individualists to relax the requirement that the actor must be anindividual human being. Many methodological individualists allow collectiveactors, such as firms, households, classes, nations, states and churches, in addi-

Methodological individualism restated 349tion to individual human beings, in their models and explanations. This form ofcompromising methodological individualism – most common economics andgame theory – is often considered unproblematic (see, e.g. Janssen, 1993: 10).Collectives are composed of individuals, and the decisions and actions of collectivesare, ultimately, the decisions and actions of the individuals, out of whichthey are composed.This form of compromising methodological individualism illustrates a verycommon move on the part of methodological individualists. When confrontedwith arguments against methodological individualism proper, they frequentlyswitch to ontological individualism. After all society is made up of human individuals,or is it not? I call this move the ontological twist. It is typical ofcompromising methodological individualism and it is possible, also, to see someversions of the intersubjective and interactionist theories of society as manifestationsof this twist. If society is only in the minds of individuals and in theinteractions between them, a reduction must be possible in principle.This form of reductive methodological individualism must be clearly distinguishedfrom institutional individualism. In the latter version, there is nothinghalf-way about explanations in terms of social institutions. On the contrary,social institutions are conceived of as the most important exogenous, or explanatory,variables in social scientific explanations. According to institutionalindividualism, explanations in terms of social institutions are rock-bottom. Apossible source of misunderstanding must be removed: institutional individualismdoes not imply a denial of the undeniable fact that social institutions arehuman creations. They are human creations, as much as are machines, houses,works of art, etc., but this does not imply that they are, in principle reducible, inthe sense of strong methodological individualism.Versions of strong methodological individualismIn the previous section, I argued that the most important divide goes betweenthe strong and weak versions of methodological individualism. I have alsoargued that only the original, strong version is obviously opposed to methodologicalholism. If we want to understand the opposition between individualists andholists, therefore, we should take our point of departure in the more controversial,strong version. Looking more closely at this version, however, soon revealsthat it is not one but many. The strong version of methodological individualismitself is made up of quite a number of different versions. I will try in this sectionto reduce the partly linguistic, partly real, variety of different versions of strongmethodological individualism to a minimal set, comprising, at least the mostcommon and most important ones, and then, to attempt an explication of, whatI consider, the single most important version of this doctrine.Methodological individualism has been formulated as a doctrine or principleabout ‘collectives’, ‘collective wholes’, ‘collective phenomena’, ‘social institutions’,‘social objects’, ‘social wholes’, ‘social events’, ‘social processes’, ‘socialstructures’, ‘large-scale phenomena’, macro-phenomena and ‘social phenomena’.

350 Methodological individualism restatedThe last term, ‘social phenomena’, is wide enough to cover them all. In the categoryof social phenomena are included the ‘social actions’ of individuals – anotion that is not without problems 2 – but I take it, as a matter of course, thatmethodological individualism applies to all actions of individuals, not just socialactions. Since I am concerned only with the strong version of methodologicalindividualism, I rule out the possibility of a holistic explanation of the actions ofindividuals (cf. Ruben, 1985: 151f).But methodological individualism has also been formulated as a principleabout ‘collective concepts’ (Weber) and sociological ‘models’ (Hayek andPopper). Granted that scientific models are conceptual, we might suggest thatmethodological individualism is either about social phenomena or socialconcepts. These social phenomena, according to methodological individualism,are to be ‘analysed’, ‘understood’, or ‘explained’ in terms of, or ‘reduced to’,‘constructed from’ or ‘constituted by’, something else. Most of these terms arevague, but can probably be reduced to those of ‘definition’ and ‘explanation’,denoting primary tasks of any science. By combination, we arrive at two versionsof methodological individualism: it is either about the definition of socialconcepts or about the explanation of social phenomena. Of these two versions,the first is more common among the subjectivist methodological individualists,while the latter is considered more important, or all-important, by thePopperians (cf. Scott, 1960: 331–6). 3I have omitted, so far, all those formulations of methodological individualismwhich state that social phenomena ‘are’, are ‘constituted’, ‘created’, ‘produced’,‘brought about’ or ‘caused’ by, or the ‘result’ of, something else. This latter groupof verbs indicates that we have to do with an ontological thesis about the nature ofsocial reality, rather than with any methodological rule about its investigation.We could even distinguish between an ontological thesis about the being, and anontogenetic thesis about the becoming of social phenomena. According to thefirst, social phenomena are something. According to the latter, they are caused bysomething (cf. Sztompka, 1979: 303).It has been pointed out by several commentators and critics, and noticed alsoby Watkins and Agassi, that methodological individualism has been formulatedboth as a strictly methodological principle or rule about the study of socialphenomena, and as an ontological thesis about the nature of social reality. 4 Thedistinction between methodological rules and ontological theses might seemunimportant. According to Popper and his followers, ontology and methodologyare mainly two modes of argument. 5 I think the distinction is worth maintaining,however, since it is possible, both to be a methodological individualist withoutendorsing ontological individualism, and to endorse ontological individualismwithout accepting methodological individualism. It is possible to believe in theexistence of social wholes, but to advocate an individualist social science. 6 Itcould be argued, for instance, as we have seen some methodological individualistsdo, that we only have ‘direct access’ to the actions of individuals, and thatsocial science, therefore, has to be individualistic.It is also possible, and more common, to believe that only individuals exist,

Methodological individualism restated 351and yet, to deny the possibility of an individualist social science. It could beargued, for instance, that large-scale social phenomena are too complex for anindividualist social science to be possible (see Udehn, 1987: 178ff). That we haveto see social concepts as shorthand for individuals and to accept explanationsthat are, in Watkins’s terminology, only ‘half-way’ (Israel, 1971: 149f). ‘In theface of complexity, an in-principle reductionist may be at the same time a pragmaticholist’ (Simon, 1978: 86).To return to the strictly methodological principle of individualism, we are leftwith the most important question: In what terms are social concepts to bedefined and social phenomena to be explained? In terms of individuals, ofcourse, but this answer needs some qualification. The most common answeramong methodological individualists is: in terms of the ‘actions’ of individuals.The term ‘behaviour’ has also been used. I will use the term ‘action’, but not inWeber’s restricted sense. I will use the term ‘action’ to denote behaviour, with orwithout a subjective meaning attached to it, and without entering any philosophicaldiscussion about the relation between actions and bodily movements.There is behind action a whole bunch of psychic factors supposed to determineits course. There are ‘needs’, ‘motives’, ‘ideas’, ‘purposes’, ‘expectations’,‘attitudes’, ‘decisions’, ‘intentions’, ‘beliefs’, ‘dispositions’ and ‘interests’, ‘desires’and ‘preferences’. There has been a controversy among philosophers about theproper way to analyse action and its determinants. It is even controversialwhether we can say anything about the ‘inner’ determinants of action at all.Both ‘dispositions’ and ‘intentions’ have been analysed in such a way as tobecome inseparable from action itself, but in different ways. 7 Methodologicalindividualism, although most of its advocates defend the view of the individualas a free agent, is best considered neutral in this controversy. Its main concern isnot the analysis of action, but the analysis of social phenomena. While beingaware of the difficulties of finding an adequate omnibus term for all the innerdeterminants of action, I will refer to them collectively as ‘psychic states’ of theindividual.While Weber, Mises, Elster and Boudon focus on the inner determinants ofaction, Popper and his followers, Buchanan and Coleman emphasise the importanceof the outer situation. This situation is partly physical, partly social. Thephysical part of the situation is not much discussed by the methodological individualists.Only Watkins and Agassi make explicit mention of the physicalsituation in their formulations of methodological individualism. According toWatkins, the physical situation consists of ‘material resources and environment’.But not all physical factors, influencing the actions of individuals, belong to theouter situation or environment. There are also those physical factors that belongto the individual her/himself, or her/his body, such as her/his biological needsand his health. 8 The physical factors influencing the actions of individuals, then,can be divided into those that belong to their ‘physical environment’ and thosethat belong to their own ‘physical states’. It is to be noticed that reference to thephysical situation is made only in that version of methodological individualism,which states something about the explanation of social phenomena. It is

352 Methodological individualism restateduncertain, therefore, if reference to the physical situation is part also of thatversion, which states something about the definition of social concepts.The components of the social situation are even more of a problem. It mostobviously consists of other individuals. But it has also been suggested that, inaddition to individuals, it consists of institutions. 9 I have already argued that theinclusion of institutions in the situation is incompatible with the original, strongversion of methodological individualism. Since institutions are social wholes,even according to the methodological individualists themselves, to admit of themin the situation is to smuggle in holism through the backdoor. The social situation,according to strong methodological individualism, then, consists of otherindividuals. But these individuals are not mere figures in a landscape.Popper, in particular, has included the terms ‘interaction’ and ‘relations’ in hisformulations of methodological individualism. While the incorporation of ‘interaction’presents no problem, that of ‘relations’ does. Popper seems to use theterms ‘interaction’ and ‘relations’ interchangeably. This indicates that heconsiders social relations reducible to the interaction between individuals. Atleast, this is a plausible interpretation of methodological individualism. I believethat strong methodological individualism implies that social relations arereducible to the interaction between individuals (and their psychic states). Whatseems clear, however, is that strong methodological individualism can only assimilatecertain views about the nature of social relations, but must oppose thoseviews which are inextricably mixed up with a holistic view of society. Toconclude, the terms in which social phenomena are to be explained, are: individuals,their physical and psychic states, actions, interaction, social situation andphysical environment.Two additional points must be made before restating methodological individualism.First, methodological individualism is usually stated in the imperative, asindicated by the use of terms such as ‘should’ and ‘must’. But it has also beenstated in the indicative, as witnessed by the use of the term ‘can’, sometimes withthe addition ‘in principle’. It is only when stated in the imperative that methodologicalindividualism can be regarded as a strictly methodological principle orrule. When stated in the indicative, as a general thesis, and with the importantaddition ‘in principle’, methodological individualism becomes an epistemologicalthesis rather than a methodological rule (see Scott, 1960; Rosenberg, 1988:114f). Methodological individualism, then, has been stated at three differentlevels: (1) the methodological, (2) the epistemological and (3) the ontological (cf.Bunge, 1973: 167). Strictly speaking, there is a methodological versionsupported, but not entailed, by an epistemological and ontological thesis, respectively(Dray, 1972: 55–7). I have used the term ‘methodological individualism’ torefer to all versions of this doctrine, but sometimes, for the sake of clarity, (1) hasbeen referred to as ‘methodological individualism’, (2) as ‘epistemological individualism’10 or ‘individualist reductionism’, (3) as ‘metaphysical’ or ‘ontologicalindividualism’, and all three, but no one in particular, simply as ‘ individualism’in social science and history. I have also referred to all all three versions of

Methodological individualism restated 353methodological individualism together as the ‘individualistic researchprogramme’.Second, methodological individualism, as we have seen, has been stated as aprinciple both about the definition of social concepts and the explanation ofsocial phenomena. And yet, something important seems to be missing. Anyone,the least acquainted with the recent discussion about methodological individualism,knows that it has been largely concerned with the reducibility, orirreducibility, of social laws. This view is typical of positivist philosophers andsocial scientists, who conceive of scientific theories as hypothetico-deductivesystems of law-like statements and who believe in the covering-law model ofexplanations. I suggest that John Stuart Mill, George Homans, Hans J. Hummeland Karl-Dieter Opp belong in this category. Also Watkins accepts the existenceof social or ‘macroscopic’ laws, while denying that they are sui generis. This istantamount to holding that social laws are reducible to psychological laws.Methodological individualism, then, is a principle or thesis about (a) the definitionof social concepts, (b) the explanation of social phenomena and (c) thereduction of social laws. 11We are now in a position to attempt a restatement of methodological individ-Figure 12.1 The individualistic research programme, or versions of strong methodologicalindividualism

354 Methodological individualism restatedualism, or the individualistic research programme, and arrive, by combining(1–3) with (a–c), at the following scheme:Of the three levels, at which methodological individualism has been stated, theontological level is, in a sense, the most ultimate. The ‘basic question, and onethat in a sense underlies all the others … is whether societies exist in their ownright at all’ (Pratt, 1978: 108). But this is also the question less likely ever to besettled by argument. The epistemological level may also be considered morebasic than the methodological one. ‘Must’ or ‘should’ presupposes ‘can in principle’.But it also presupposes can in practice, which, together with some otherconsiderations, motivates a special treatment of methodological individualism.Of the three different versions (a–c) of methodological individualism, (b) and(c) must be considered scientifically more important than (a). The definition ofconcepts is, especially in social science, too much of a conventional matter, for anagreement about the right one to be possible. There is also the problem that definitionsare of different sorts, and made for different purposes. The strongestversion of methodological individualism would be the requirement that all socialconcepts must be reduced and entirely eliminated from social science (andhistory). Although sometimes made, this is an unlikely interpretation of the viewsof the methodological individualists. From Weber to Watkins, no one arguedagainst the use of social concepts for descriptive purposes. But in a nominalistfashion they held that social concepts are shorthand for complexes of individuals.A more likely interpretation of the position of the methodologicalindividualists is that they admit of social concepts for descriptive purposes, butforbid their use for explanatory purposes. 12Of versions (b) and (c), the former about the explanation of socialphenomena is more general than the latter about the reduction of social laws. 13Version (b) and, in particular, (1b) also appears to occupy a privileged position asmost important among the methodological individualists themselves. Thisversion is Methodological Individualism. 14 It is true that, on the deductivenomologicalor ‘covering-law’ model of scientific explanation, there can be noscientific explanation without laws. But while both explanation and reductionrequire the existence of individualist laws, reduction, in a strict sense, requires, inaddition, the existence of social laws. The reduction of social laws has beendiscussed in chapter 11.Explanatory methodological individualismI have suggested that version (1b) in Figure 12.1 is the most important. By ‘mostimportant’, I mean first of all, that it is the principle defended by most methodologicalindividualists, the Popperians, in particular. I will now turn to anexplication of this version of methodological individualism, as stated in Figure12.1: Social phenomena must be explained in terms of individuals, their physical and psychicstates, actions, interactions, social situation and physical environment.The expression ‘explained in terms of ’ requires some clarification. We mayrecall Jarvie’s suggestion that individualists tend to put social wholes in the cate-

Methodological individualism restated 355gory of ‘explicanda’ (explanandum), which presumably implies that they tend toput individuals, etc. in the ‘explicans’ (explanans). 15 While based on a soundintuition, this suggestion will not do. As Danto correctly points out, on thedeductive-nomological model of scientific explanation, nothing can be in theexplanandum which is not already in the explanans. His own suggestion is that,for an explanation to satisfy methodological individualism, the initial conditionsmust be facts about individual human beings (Danto, 1965a: 277). While, nodoubt, an improvement upon Jarvie’s suggestion, even this one is not altogethersatisfactory. The notion of ‘initial conditions’ belongs to that of a ‘closed’ or‘deterministic’ system. But in social science there are no closed systems. Besidesinitial conditions, there are the most important boundary conditions, whichtogether with the initial conditions make up the antecedent conditions of anexplanation (Hempel, 1965: 232, 246–51, 335ff). The following, extremelyFigure 12.2 Deductive–nomological explanationssimplified model of a deductive-nomological explanation may help to clarifymatters.A first interpretation of the phrase ‘explained in terms of ’ is that it refers tothe initial and boundary conditions of a deductive-nomological explanation.Methodological Individualism could then be rewritten: In any deductive-nomologicalexplanation of a social phenomenon, the antecedent must refer only to individuals, etc. 16The most often cited examples of laws in social science are those ofeconomics. It is not usual, however, to refer to the explanations of economics ashypothetico-deductive explanations, conforming to the covering-law model ofscientific explanations. As we have seen in chapter 8 and chapter 9, methodologicalindividualism has been explicated, by Lawrence Boland, in terms ofendogenous and exogenous variables. Following Boland, I suggest the followingalternative explication of strong Methodological Individualism: In any social scientificmodel, the exogenous variables and conditions must refer only to individuals, etc., but not tosocial institutions.

356 Methodological individualism restatedA problem with the first explication above, is that few social scientific explanationssatisfy the criteria of a complete deductive-nomological explanation. Mostexplanations in social science (and history) are incomplete. They are, to borrow aterm from Carl Hempel, mere ‘sketches’ of complete deductive-nomologicalexplanations (Hempel, 1965: 423f; see also Brodbeck [1962] 1968: 381–5).Typical for many social scientific and historical explanations is the absence ofany explicitly stated laws. But even in the absence of laws, it is often possible torecognise an antecedent and a consequent. Many social scientific explanationsconsist in the enumeration of a number of causal factors supposed to explain acertain social phenomenon. In that case, the antecedent is simply the sum totalof these causal factors (cf. Little, 1998: ch. 10).Another problem with the covering-law explication is that many social scientificexplanations are statistical, rather than deductive-nomological in form. Inthe case of statistical relationships between variables, it is customary to make thedistinction between dependent and independent variables. In the case of statisticalrelations, I suggest the following explication of MethodologicalIndividualism: In any statistical explanation of a social phenomenon, the independent variablesmust range over individuals, etc. 17The most serious problem with the covering-law explication ofMethodological Individualism is the scarcity, or absence, of clearly stated andagreed upon laws in social science. There has been, in recent social science andphilosophy, an emerging awareness of this predicament and different suggestionsto meet it. The most interesting suggestion is to replace social laws by socialmechanisms. 18 As I have mentioned in chapter 6, one branch of the socialmechanism approach is explicitly individualistic. 19 There is no particular way tostate strong methodological individualism, in terms of social mechanisms. It maybe stated, as a modification of the original version of MethodologicalIndividualism (1b) in Figure 12.1: Social phenomena must be explained by individualisticmechanisms; that is, in terms of individuals, etc.For many social scientists, ‘social mechanisms’ is another name for rationalchoice. I have treated a considerable number of rational choice theories in chapters8, 9 and 10, and concluded that they tend to transgress the boundaries ofstrong methodological individualism. Most rational choice theories are, explicitlyor implicitly, manifestations of institutional and structural individualism. Inorder to satisfy the strictures of strong methodological individualism, a rationalchoice explanation would have to satisfy the following condition: In any use ofrational choice to then explain a social phenomenon, the description of the situation must be interms of individuals, etc., but not of social institutions, or structures.

Notes1 Introduction1 See Wilhelm Dilthey’s Introduction to the Human Science ([1883b] 1988: 209–18) wherethe opposition between an atomistic/mechanistic and a holistic/organismic metaphysicsis a main theme. See also F. Kaufmann (1958: 158): ‘For more than 2000years the nature of the relation between social groups and institutions – particularlysociety and the state – and the individuals “forming” them has been one of thepivotal issues in social philosophy’.2 See, e.g. Ryan (1975: 173). It would seem, however, that Ryan has changed his mindabout this in Ryan (1991: 42) where he argues (against Jon Elster) that the issuebetween individualists and holists is an empirical matter (see also Tännsjö, 1990).3 See also Runciman (1972: 30).4 I find it interesting that Alan Ryan, who once thought that the battle between individualistsand holists is a sham battle, has taken Elster to task for maintaining thatmethodological individualism follows from the truism that no system can workwithout individuals and no explanation neglect that fact. ‘But to be a holist is not todeny any of that. Rather, it’s to bet on the prospect of explaining social and economicphenomena by approaching them in a “system-down” rather than “individuals-up”perspective’. Thus, ‘there is a genuine parting of ways between non-truistic methodologicalindividualism – that is, the research programme based on orthodox rationalman analysis – and Marxism’ (Ryan, 1991: 42f).5 See, e.g. Sensat (1988: 193ff) and Kinkaid (1997: 144). The reason this position hasbecome common is probably the idea of supervenience, which confers some credibilityon it.6 It remains, of course, to show how the course of history is determined by the actionsof individuals. This Hegel does not do, at least not in a way that would satisfy amethodological individualist (cf. Kinkaid, 1986: 509).7 Examples of such question-begging statements of methodological individualism areM. Lessnoff (1974: 77) and Ryan (1975: 172ff). The most trivial of all individualisticreductions of society, however, is that of Hummel and Opp (1968; 1971). Accordingto Ruben (1985: 152) ‘[i]t seems obvious that if social relations were to be included,the doctrine [methodological individualism] would be incapable of giving any sensewhatever to the idea of the explanatory priority of the non-social over the social’.8 I intend also to write a treatise about collectivism and holism within the next fewyears.9 I am working on a manuscript about the history of the individualist theory of societybefore social science. The title of this work is not yet decided, but the manuscript willbe finished in about a year. I conceive of it as companion volume to the present work.

358 Notes2 Background1 See, for instance, Pribram (1912: ch. 4), Vanberg (1975: ch. 1) and Infantino (1998:41ff).2 On the role of Robinson Crusoe in economics, see Novak (1962).3 Some of the more important contributions to the critique of political economy inEngland, can be found in E. Jay and R. Jay (eds) (1986).4 Ruskin’s main writings on economics are The Political Economy of Art (1857), Unto ThisLast (1862) and Essays on Political Economy (1862), all included in Ruskin (1857–62).The term ‘catallactics’ was introduced by R. Whately in his Lectures on Political Economy(1831), as an alternative to ‘political economy’. The point of the new term was tomake clear that economics is a theory of the market, not of the household, or state.The term ‘catallactics’ has been picked up by many methodological and economicindividualists, to make the same point and in order to avoid the collectivism impliedby a science of household, or state, management of economic life.5 Lorenzo Infantino, who makes the observation that there is no necessary link betweenmethodological and political individualism (1998: 131), has nevertheless written abook in which the distinction between the two is constantly blurred.6 For a more precise statement, see Sowell (1974: 39–41) where Say’s law is stated inthe form of six major propositions. See also Blau (ed.) Jean Bapiste Say (1776–1832)(1991).7 As Sowell (1974: 148) has pointed out, however, ‘no one more strongly defendedabstract deductive reasoning than Karl Marx’. The implication is that the relationbetween abstract theory and individualism is far from necessary.8 Public choice is the economic theory of politics, based on the assumptions of (1) selfinterest,(2) exchange and (3) individualism (see Udehn, 1996: 35).9 Robert E. Prasch (1996) has argued that the origin of the a priori method in classicaleconomics is to be found in Scottish common-sense philosophy, rather than inDescartes, which is commonly believed. I do not dispute the role of Scottishphilososphy, but I think it would be wrong to disregard that of Descartes (andGalileo), as mediated by Thomas Hobbes.10 The German term is ‘Historismus’. See, for instance, Troeltsch (1924) and Mannheim(1924), who both use the term Historismus in the second sense.11 It may be pointed out that Berlin denies that Herder is guilty of metaphysical holism.‘For Herder all groups are ultimately collections of individuals; his use of ‘organic’and ‘organism’ is still wholly metaphorical and not, as in later, more metaphysicalthinkers, only half metaphorical. There is no evidence that he conceived of groups asmetaphysical ‘super-individual entities or values’ (Berlin, 1976: 198).12 The distinction between an ‘Older’ and a ‘Younger Historical School’ comes fromSchumpeter (1954: 807ff), who also mentions the ‘Youngest Historical School’,including Arthur Spiethoff, Werner Sombart and Max Weber (pp. 815ff).13 Droysen’s critique of Buckle can be found in a famous review of the latter’s History ofCivilization in England, published in 1862 and added as an appendix to his Historik in1868 (Droysen [1858] 1977: 386–405).14 See ‘Kunst und Methode’, added as an appendix to the 1868 edition of the Historikand translated as ‘Art and Method’ in F. Stern ([1956] 1970: 137–44).15 It is common to see Henri St Simon and Auguste Comte as the founding fathers ofsociology. Others, such as Small (1907), Sombart (1923), Swingewood (1970) andEriksson (1993) argue that the roots of sociology are rather to be found in the ScottishEnlightenment. I have no reason to take sides in this debate.16 Comte uses the term ‘spontaneous’ frequently to denote unintended socialphenomena. It is also used in the title of chapter V in Book VI of The PositivePhilosophy, which is ‘Social Statics, or Theory of the Spontaneous order of HumanSociety’. As far as I know, Comte was first to use the term ‘spontaneous order’ about

Notes 359social phenomena (Aristotle used the term ‘spontaneous generation’, but notabout social phenomena). Since Friedrich von Hayek did read Comte, it is, at least,possible that he picked up the term from him.17 Friedrich von Hayek, wrote two critical essays, which deal to a large extent withComte: ‘The Counter-Revolution of Science’ (1941) and ‘Comte and Hegel’ (1951),both included in Hayek (1955). Comte is also a principal target of critique in Popper’sThe Poverty of Historicism (1957).18 Comte’s advocacy of social engineering is most manifest in his early works (Comte[1819–26] 1974), written while appointed as secretary to St Simon. The main workfrom this period is ‘Plan of the Scientific Operations Necessary for ReorganizingSociety’ (1822), which contains most of Comte’s later ideas in embryo.19 The best treatment of the role of statistics in the rise of social science that I know ofis Ian Hacking’s The Taming of Chance (1990). It is a common, and probably correct,observation that the main impetus to the rise of systematic empiricism was the needon the part of state administrations for official statistics (Österberg, 1988: 18ff;Desrosières, 1991).20 See, however, John Goldthorpe (2000: 266–74), who does not at all share the opinionthat Durkheim’s Suicide was a masterpiece.21 Spencer’s political views are most clearly expressed in the The Man Versus the State([1884] 1982) and other essays included in the volume with this name.22 ‘I have very emphatically expressed my belief in a subjective science of the mind, bywriting a Principles of Psychology, one half of which is subjective’ (Spencer [1864] 1984:18).23 In his Reasons for Dissenting from the Philosophy of M. Comte, Spencer ([1864] 1984: 15)avoids the organism analogy and uses that of mechanism instead. In his later works,society is once again a social organism, but the analogy between the individual andsocial organism is played down. What they both have in common is that they areorganised entities with mutual dependency among the parts (Spencer [1873] 1961:52–4, 298ff; [1876] 1985, vol. 1: 437ff, 580f).24 See, e.g. Buchanan (1949) and Benn and Peters (1959: 290) on the contrast betweenindividualism and organicism.25 See Taylor (1989; 1992: 137ff) and Gray (1985: 245f; 1996: ch. 13). Other examplesare the American sociologist William Graham Sumner (see Sumner, 1992) and theEnglish jurist Wordsworth Donisthorpe (1889), both of whom were influenced bySpencer. Sumner rarely used the term ‘social organicism’ but there is little doubt thathe was an organicist, and even more of an evolutionist. Donisthorpe, on the otherhand went farther in his organicism than Spencer. According to him, also the state isan organism (p. 2) and, like other, social organisms, it is endowed with a group mind(p. 276). Even so, he was a political individualist.26 If there is a debt to Spencer on the part of Hayek, it is unacknowledged. It has beenargued by Paul (1988), that there is an even greater similarity between Hayek andSumner, but Sumner is not even mentioned by Hayek.27 It should be mentioned that Durkheim was not an enemy of all versions of individualism.On the contrary, he was a strong supporter of Kantian individualism, with itsidea of the autonomy of the individual and he was an equally strong supporter of theidea of human rights derived from Kantian individualism. This fact, casts strongdoubts upon some interpretations of Durkheim, especially that of Philip Pettit([1993] 1996: 112, 127ff), who suggests that Durkheim was a collectivist, who deniedany autonomy to individuals.28 Distrust of psychology, subjectivism and humanism has turned into something of atrademark of that type of French structuralism, which conceives of Durkheim andthe Swiss linguist Ferdinand de Saussure as its founding fathers.29 See, e.g. Durkheim ([1895] 1982: 39f, 45, 128f; [1897] 1951: 319f; [1898] 1974: 31;[1900] 1973: 16ff; [1911] 1974: 90–93).

360 Notes30 For Pettit, the issue of individualism versus collectivism has to do with the questionwhether human beings are autonomous or subject to social forces from ‘above’;whereas, that of atomism versus holism, concerns horizontal relations between humanbeings; to what extent human beings are social beings. I agree with the conclusions thatPettit reaches; that collectivism is untenable and holism justified, as he understandsthese doctrines, but I have no use for this understanding, which is too much at odds withcommon usage of the terms ‘individualism’, ‘atomism’, ‘collectivism’ and ‘holism’.3 Psychologism in early social science1 See, e.g., Pandit (1971), Notturno (1985: ch. 2), Cussins (1987), Mohanty (1989) andKusch (1995: 4ff).2 Other important representatives of associationist psychology are David Hartley(1705–57), Thomas Brown (1778–1820) and Alexander Bain (1818–1903).3 See, e.g., Ryan (1970: 156ff), Hollis (1972: 376f) and Oakley (1994: 185).4 For Mill’s epistemological views, see A System of Logic, and especially An Examination ofSir William Hamilton’s Philosophy (1865), which is to a large extent a defence of Britishempiricism and associationist psychology against the more intuitionist view ofHamilton. It may also be mentioned that John Stuart Mill published a new edition ofhis fathers’ Analysis of the Human Mind in 1878, an act that could be interpreted as asign of appreciation.5 It is common to credit Auguste Comte with the invention of the term ‘social science’.According to P.R. Senn (1958) 1991, however, Mill used this term ten years beforeComte, or in 1836. This claim has been contested by J.H. Burns (1959) 1991, whopoints out that Comte used the term science sociale already in 1822 and by G.G. Iggers(1959) 1991, who points out that it was used by Charles Fourier in 1808 and bySismondi in 1803.6 The idea that the association of ideas can take the form of ‘mental chemistry’ wassuggested by James Mill in Analysis of the Phenomena of the Human Mind ([1829] 1878:321). In the Preface to the second edition of this work, John Stuart Mill says (p. xi),that the idea originated in David Hartley’s treatise Observation on Man (1749), who wasindebted to an otherwise forgotten thinker, Mr. Gray. As Infantino (1998: 72) pointsout, however, it was used by Rousseau in his ‘Geneva manuscript’, where it is associatedwith the holistic idea that the whole is not just an aggregate, but greater than thesum of its parts. The latter idea can be traced first of all to Leibniz, but goes furtherback to Greek Antiquity, where Sextus Empiricus denied, in his Outlines of Pyrrhonism(p. 189), that there is such a thing as a whole, on the ground that if wholes exist theymust be other than their parts, which obviously they are not. Empiricus’ argument isclearly polemical, which means that the idea that wholes are other, or more, thantheir parts must be found somewhere else, probably in Aristotle.7 More recently, Jon Elster (1999: 10–13) has turned to the treasure of folk wisdom, inthe form of proverbs, as a source of knowledge about social mechanisms.8 The term and the idea of axiomata media goes back to Francis Bacon (1561–1626).More recently, it has been used by the sociologist Robert Merton (1967: 56–9) as thesource of his own idea of ‘theories of the middle range’. A closer look at the latter,however, reveals that theories of the middle range have little to do with Mill’s axiomatamedia. For one thing, they are not psychological. The real follower of Mill among sociologistsis the psychologistic methodological individualist George C. Homans.9 Mill’s reaction was not typical for a Briton, however. According to Hacking (1990: ch.15), British philosophers and social scientists, generally, had little problem acceptingthe idea of statistical laws, since it was compatible with Hume’s conception of causeas constant conjunction. It was in Germany that the insight emerged in the middle ofthe nineteenth century that statistical regularities are not laws.

Notes 36110 See, e.g. Mill’s review of de Tocqueville’s Democracy in America (Mill [1840] 1965),which contains many institutional explanations by Tocqueville, which Mill endorses,and also some entirely his own. Wilson (1998) has argued that Mill was a structuralist,rather than a psychologistic methodological individualism. This is farfetched andlacks support in Mill’s methodology.11 It was Wicksteed ([1888] 1970: 45f), who introduced the term ‘marginal utility’, toreplace Jevon’s ‘final utility’, in British marginalism. ‘Marginal utility’ was a translationof the German term ‘Grenz-nutzen’, which was coined by the Austrianeconomist Friedrich von Wieser (1884: 126ff).12 On the term ‘catallactics’, see p. 358, note 4.13 A summary statement of the main points made by Wicksteed in The Common Sense ofPolitical Economy can be found in his article on ‘The Scope and Method of PoliticalEconomy in the Light of the “Marginal” Theory of Value and Distribution’ (1914).In this article Wicksteed mentions two assumptions, which are necessitated by the useof the differential calculus in the mathematical treatment of the theory of price: (1)the tastes, desires and resources of individuals are fixed and (2) that the amount ofsupply is also fixed (p. 784). The first of these assumptions, in particular, have sinceplayed an important role as one of the most distinctive marks of the economicapproach to behaviour (see, e.g., Becker, 1976: 5; and Stigler and Becker, 1977), andin the division of labour between economics on the one hand and psychology andsociology on the other hand (see, e.g., Tullock, 1972).14 On Marshall’s organicist leanings, see, e.g., Whitaker (1975: 107ff).15 Pareto took the idea of indifference curves from Edgeworth (see p. 52), but unlike thelatter he dropped the idea of utility: ‘I consider the indifference curves as given anddeduce from them all that is necessary for the theory of equilibrium, withoutresorting to ophelimity [utility]’ (Pareto [1909] 1972: 119, note 4).16 Völkerpsychologie was the joint creation of M. Lazarus and L. Steinthal, who saw it as adiscipline studying the language, mythology, religion, customs and art of a people.They rejected the romantic concept of ‘folk soul’, but accepted that of ‘folk spirit’,which is simply another phrase for ‘culture’ (cf. Kalmar, 1987).17 Völkerpsychologie is usually translated as ‘folk psychology’, but this is unfortunate, sincethis term is more commonly used to designate the psychological beliefs entertained bypeople, i.e. common sense. Völkerpsychologie, on the other hand, is a discipline thatturns ‘folk psychology’ into an object of investigation. Needless to say, it has much incommon with ethnology, social anthropology and sociology.18 An important contribution to the development of functionalist psychology is JohnDewey’s article ‘The Reflex Arc Concept in Psychology’ (1896), which launches afunctionalist view of the psyche as an alternative to the behaviorist stimulus-responsemodel.19 In a later work on The Group Mind (1920), McDougall has turned into a holist arguingthat the group mind is irreducible to the minds of individuals. This made him one ofthe main targets of attack for methodological individualists.20 James Baldwin was an important figure in the early development of socialpsychology: he launched the idea of imitation as the most important mechanism oflearning, independently of Tarde.21 In a short comment made much later, Allport (1961) reflects on his 1924 critique ofthe group fallacy. He still believes that the fallacious idea of collective, or corporate,actors, in social theory is a serious problem and he does not admit of having changedhis mind on that point. I see an important change of view, however. Whereas in 1924,Allport referred to social entities as aggregates and demanded a reduction of sociologyto psychology, he now insists that groups are social wholes with an internalstructure that explains collective, or corporate, action, and he says nothing at allabout reducing sociology to psychology.

362 Notes22 A difference between Allport and more recent advocates of microfoundations, is thatthe former seeks them in behaviourism, while the latter prefer rational choice.23 Dilthey’s lived experience (Erlebnis) should not be conflated with the experience of thetheory of knowledge. According to Dilthey ([1883] 1989: 50): ‘No real blood flows inthe veins of the knowing subject constructed by Locke, Hume, Kant, but rather thediluted extract of reason as a mere activity of thought’.24 Dilthey’s descriptive psychology was subjected to an effective critique by the experimentalpsychologist Hermann Ebbinghaus (1896), who argued that Dilthey did notknow, or understand, the explanatory psychology he criticised and therefore exaggeratedthe differences between descriptive and explanatory psychology. He also accusedDilthey of misunderstanding the methodology of the sciences, in general and evenhis own descriptive psychology. Ebbinghaus’s critique did much to discredit Dilthey’sproject of a descriptive psychology and to undermine his position in Germanacademic life.25 See, e.g., Rickman (1967: 37ff), Habermas ([1968] 1971: 145–48), Outhwaite (1975:24–37) and Makkreel ([1975] 1992: chs. 7–8). See, however, Rickman (1979: 69), whotries to play down the importance of the shift from psychology to hermeneutics in thework of Dilthey.26 Dilthey’s elementary acts seem to be similar to the ‘basic actions’ of ordinarylanguage philosophy (see Danto, 1965b).27 It has been argued by Backhaus (1998) that Simmel owed a lot also to EdmundHusserl. That, in fact, Simmel’s formal sociology should be interpreted as an ‘eidetic’social science in the phenomenological sense (see p. 81).28 Chapter 2 in Frisby (1992) is an article with the title ‘Georg Simmel and SocialPsychology’, originally published in Journal of the History of the Behavioral Sciences, 20,1984, 107–27, which is a main source on Simmel’s view of social psychology and itsrelation to sociology. According to Frisby, the early Simmel took a strong interest inpsychology, and wrote several articles on this subject in the 1880s.29 Simmel’s main work on this theme is The Problems of the Philosophy of History (1892) and‘On the Nature of Historical Understanding’ (1918). One difference between Diltheyand Simmel is that the latter is more of a Kantian, as manifested in his critique of(Ranke’s) historical realism and his epistemological constructivism ([1892] 1977: 76ff;[1917–18] 1980: 149f).30 It may be pointed out that behind Dilthey’s way of making the distinction betweenthe natural and human sciences was a dissatisfaction with Kant’s view of psychology.In A Descriptive and Analytic Psychology, published in the same year as Windelband’saddress, Dilthey is even more outspoken in his critique of Kant’s theory of knowledgeand, especially, of its claim that epistemology is independent of psychology.According to Dilthey ([1894] 1977: 32): ‘it is evidently impossible to connect the spiritualdata which form the matter of epistemology without relying on some idea orother of the psychic nexus. Absolutely no magical trick of a transcendental methodcan make possible what is in itself impossible. Utterly no legerdemain of the Kantianschool can be of any help here’. I do not know, if Windelband had read this beforehis attack on Dilthey, but if he had, it is not surprising that he should feel provoked.31 It is of some interest, for the purposes of this book, to notice that the methodologicalindividualist Friedrich von Hayek is the author of the Preface to the English edition ofthis book (1962). He complains that Rickert’s dichotomy does not allow for a generalisingsocial science, like economics, which is not altogether correct, since Rickert sawgeneralising and individualising as extremes on a continuum and explicitly recognisedthe generalising element in this discipline (p. 109). The following observation is,however, entirely justified: ‘Rickert was scarcely interested in and perhaps hardlyaware of the existence of the kind of compositive or individualist theory thateconomics and possibly linguistics have developed in order to account for the appearanceof regular structures not the result of deliberate human design’ (p. vi).

Notes 36332 A summary of Husserl’s argument against psychologism can be found in Kusch(1995: 41–60).33 Dilthey referred approvingly to Husserl’s Logical Investigations in his study ‘DerStrukturzusammenhang des Wissens’ (1905), published posthumously in Der Aufbau derGeschichtlichen Welt in den Geisteswissenschaften ([1926] 1958: 39ff). There is also a letterfrom Husserl to a student in 1927, which mentions a visit he paid to Dilthey in thewinter 1905/1906. This was the time when ‘the first synthesis between my strivingsand those of Dilthey occurred in the form of a personal conversation during my visitto his home …. That Dilthey identified my phenomenology with psychology as ahuman science and related it to his life-goal of a philosophical foundation of thehuman sciences made a tremendous impression on me’ (quoted from the Preface toR.A. Makkreel and J. Scanlon (eds) Dilthey and Phenomenology, 1987: ix). See also the‘Introduction’ to Husserl’s lectures on phenomenological psychology in Paris 1925(Husserl, 1977: 24f), where Husserl tells his audience about Dilthey’s enthusiasticreaction to Logical Investigations, but also about his initial suspicion against Dilthey’swork on descriptive psychology, which he did not read because of Ebbinghaus’sdevastating critique. When he, eventually, did read it, he saw ‘that Dilthey was in factright with his judgement which has so greatly astonished me, concerning the innerunity of phenomenology and descriptive-analytic psychology’ (p. 25).34 That Husserl did know about Dilthey’s descriptive psychology is evident from theprevious footnote. His critique of Dilthey’s ‘Weltanschauung philosophy’ led to a correspondencebetween them, where Dilthey first protested against being classed as a‘historicist’ and ‘relativist’ and then paid his respect to Husserl, who answered in aconciliatory tone. The correspondence ended in mutual reverence.35 It may be noted, though, that Husserl mentions a ‘general phenomenology’, whichdeals with objects as they present themselves to consciousness (1981: 12).36 This quest for foundations has recently been much criticised by postmodernism/poststructuralism,especially by Jacques Derrida (1964), who argues that Husserl’stranscendental phenomenology is metaphysical, despite Husserl’s claim to thecontrary.37 Also important is the article on ‘The Origin of Geometry’ (written in 1936 and firstpublished by H. Fink in 1939), where Husserl went far in acknowledging the dependenceof phenomenology on the sciences of man: ‘The ruling dogma of a separationin principle between epistemological elucidation and historical, even humanisticpsychologicalexplanation, between epistemological and genetic origin, isfundamentally mistaken, unless one inadmissibly limits, in the usual way, the conceptof “history”, “historical explanation”, and “genesis”’ (Husserl 1981: 264). See alsoJordan (1968) and Landgrebe (1977). It is this later work of Husserl and, especiallythe problematic concept of ‘Lifeworld’ (Carr, 1977; Gurwitch, [1967] 1978), which isthe source of the sociological phenomenology of Alfred Schutz and MauriceMerleau-Ponty. While the former became acquainted with it through personalcontact with Husserl in the early 1930s, the latter read in the Husserl Archives ofLouvain about ten years later.38 This reference is to the Vienna Lecture, which Husserl gave in 1935 and which isadded as an appendix to The Crisis of European Sciences.39 See also the Vienna Lecture (Husserl [1954] 1970: 296–98) and the Prague Lecture(Husserl, 1965: 188–90).40 It might be added that Jaspers was a friend and admirer of Max Weber. Obviously, hewas not led by his admiration to accept Weber’s neo-Kantian rejection of psychologism.

364 Notes4 Austrian methodological individualism1 Over the last decades, there has been a Renaissance for the Austrian School ofEconomics and the literature about it is by now enormous. On the historical andphilosophical background of the Austrian School, see, e.g., Dolan (1976), Smith(1986), Parsons (1990) and Cubeddu (1993). On the German roots of the AustrianSchool, in general, and on Carl Menger, in particular, see especially Alter (1990), butalso Streissler (1990) and Silverman (1990). According to one interpretation, inspiredby Ludwig Lachmann (1970; 1990), and launched, especially, by Don Lavoie (1986;1990), Austrian Economics may be seen as a hermeneutical enterprise (Selgin, 1988).For a serious critique of this view, see Albert (1988); for a mere dismissal, seeRothbard (1989).2 The psychological basis of Menger’s marginalism was a theory of needs rather thanof utility-maximisation. According to Alter (1990: 23ff) this reveals the roots ofMenger’s economics in German Romanticism. It should be kept in mind, though,that Menger was consistently critical of the more holistic aspects of German thought(see, e.g. Milford, 1990) and was the first in the sphere of German economics to advocatemethodological individualism (Streissler, 1990: 60f).3 In his later writings, Menger dissociates himself from Bacon’s inductivism, because itdoes not lead to exact laws (Menger [1871] 1963: 60).4 On The Battle of Methods, see Schumpeter ([1914] 1954: ch. IV), Ritzel (1950) andBostaph (1976).5 Böhm-Bawerk (1890) seems to agree that the Battle of Methods was between adherentsof a historical-inductive and an abstract-deductive method, but preferred to talk,in the latter case, of an ‘isolating method’ (p. 249). See, however, Menger ([1894]1935: 278–280), who denies that the Battle of Methods was about induction versusdeduction, or even isolation. Atomism was the heart of the matter. According toMenger it was a clash between different views of the goals of research and of thetasks of economics, in particular. ‘The Austrian School defends theory based onscientific analysis – and this means to trace the complex phenomena of the economyback to the endeavours of economising human beings and to their psychologicalcauses in order to achieve a deeper understanding of economic phenomena in theirwider context – while criticising the Historical School as insufficient even as a practicalscience of economics’ (p. 280 [my translation]).6 It is commonly believed that Menger’s idea of economics as an exact science derivesfrom Aristotle (Kauder, 1958; Hutchison, 1973: 17ff; Smith, 1990) and this is mostprobably correct, but it should be added that this influence was equally probablymediated by the Aristotelian philosopher Franz Brentano, who was a contemporaryof Menger in Vienna (see Smith, 1986: 8ff; Fabian and Simons, 1986).7 Cf. von Wieser (1914: 162–6), Watkins (1976a: 711), O’Driscoll (1986) and Vanberg(1994: 146–51).8 According to Hutchison (1988), ‘[t]he ferocity of the quarrel derived … from questionsof priority and prestige. As Schumpeter put it, the issue was one of “Luftraumoder Herrschaft”, that is of empire-building and the capture of chairs and jobs’ (p. 528).9 A morphological, or ‘anatomical’ orientation had been suggested, before Menger, bythe holistic economists Wilhelm Roscher and Gustav Schmoller. More recently amorphogenetic approach has been defended by Margaret S. Archer (1985; 1988;1995). In her work, this approach is best understood as a form of genetic structuralism.10 The American sociologist Albion Small, for instance, characterised the AustrianSchool as ‘the attempt (about 1870) to reconstruct economics by appeal topsychology’ (Small [1924] 1967: ch. 12).

Notes 36511 For those who want to go further in these investigations, it may be pointed out thatthe sources of Dietzel’s ‘economic principle’ are economists belonging to the GermanHistorical School and especially Adolf Wagner.12 See Menger ([1883] 1963: 93, 116f, 139ff, 208f; 1884: 77–79).13 Cf. Harsanyi (1978): ‘When on the common-sense level we are speaking of rationalchoice, we are usually thinking of a choice of the most appropriate means to a givenend … However, it has been an important achievement of classical economic theoryto extend this common-sense concept of rational behavior so that it can cover notonly choices among alternative means to a given end but also choices among alternativeends. Under this more general concept of rationality, our choices amongalternative ends are rational if they are based on clear and consistent priorities andpreferences. Formally this means that our choice behavior will be rational if it satisfiescertain consistency requirements or rationality postulates’.14 See Walther (1926: 5–10), Antoni (1940: ch. IV), Mommsen (1974: ch. 1) andSchluchter (1989: ch. 1).15 See, e.g. Pfister (1928), Tenbruck (1959: 583–9), Cahnman (1964: 108–21), Burger(1976: 6, 140–53).16 On Weber’s debt to Austrian Economics, see Tenbruck (1959: 583–9, 603–6),Therborn (1976: 290–5), Binns (1977: ch. 2), Jones (1977: 29ff), Clarke (1982:197–212), Holton and Turner (1989: ch. 2), Schön (1987: 60ff), Hennis (1991: 29–31).17 Some commentators see in Weber’s methodological essays little more than an unsystematicheap of polemical tracts, written as interventions in contemporary debatesand occasioned by Weber’s wish to rebut faulty ideas on the part of his contemporaries(Tenbruck, 1959; Hennis, 1988: 157f). This view gains support from MarianneWeber’s biography of her husband, where it is suggested that methodology was notMax Weber’s main concern (Weber, 1926: 265, 308f). This is, of course, true, butdoes not imply that his methodology is unimportant. Weber saw methodology as ameans to an end of substantive work – a self-evident view on the part of a historianand social scientist – but this does not imply that he saw it as unimportant (see Weber,1949: 116). It is also true that Weber never systematised his methodology into aunified whole, but this does not imply that it lacks thematic unity. Tenbruck hasrecently changed his view of Weber’s methodology. He now believes that it is unitaryand, what is more, that it is an important key to the understanding of Weber’s sociology(Tenbruck, 1986; 1989; 1994). I agree.18 In the enormous literature on Weber’s methodology, his methodological individualismhas been a relatively neglected subject. See, however, Runciman (1972: 24f),Torrance (1974: 145–5), Weiss (1975: 90–3); Holton and Turner (1989: 38–44),Albrow (1990: 135–40, 251–4) and Ringer (1997: 155–62).19 It has been argued, though, that Weber did not follow Menger in his critique of theidea of a ‘national economy’. According to Hennis (1988: 117–125), Weber saweconomics as a political science and, as such, inseparable from the nation state. Themain evidence, for this view is Weber’s Inaugural Address on ‘The Nation State andEconomic Policy’ ([1895] 1980), delivered when Weber was appointed to the Chair inpolitical economy at Freiburg in 1895. It is true that Weber in this early lectureproclaimed that ‘political economy is a political science’ (p. 16). He also called himselfa historian (p. 20) and a disciple of the German Historical School (p. 19). But he alsosaid: ‘As an explanatory and analytical science, political economy is international’ (p.15). This means that, already in 1895, he sided with classical and neoclassicaleconomics, against members of the Historical School, on the intepretation of theoreticaleconomics.20 See G. Wagner and H. Zipprian (eds) Max Webers Wissenschaftslehre (1994), especiallythe contributions by Rehberg (1994) and Scaff (1994).21 Weber seems to have appreciated the work of Othmar Spann, but it is difficult to seewhy. In his Roscher and Knies ([1903–1906] 1975: 253, note 49) and in an unfinished

366 Notescritical review of Simmel, probably written in 1908 (Weber, 1972: 162, note 5),Weber refers to Spann’s ‘perceptive criticism’ of Georg Simmel’s concepts of ‘society’and of ‘sociology’. Weber’s appreciation is hard to understand, since from his point ofview, one problem with Simmel’s concept of society is that it is not individualistenough (Simmel starts with interaction, not action), whereas for Spann (1905:302–44, 427–60), the problem with Simmel is that he is too individualistic. Spann wasa dedicated collectivist and holist, or in his own terminology, a ‘universalist’. Weber’sappreciation of Spann was not reciprocated. In his review of Economy and Society,Spann (1923) is dismissive of the whole project and ends by predicting that Weber’slife-work will not last. He couldn’t have been more wrong.22 This helps explain Weber’s critical stance towards Simmel’s sociology, which is ascience of society (cf. Lichtblau, 1994), but it does not at all explain his seemingappreciation of Spann’s sociology (see note 2), which is also a science of society(Gesellschaftslehre).23 In his necrology over Max Weber, Schumpeter ([1920] 1991: 221) says that hispersonal acquaintance with the man was too limited for giving a portrait of him.What he offers, instead, is a hagiography, depicting Weber as a figure of almostsuperhuman qualities, both as a person and as a scholar. The picture is not entirelygroundless. Weber’s scholarly achievements speak for themselves and testimonies byhis contemporaries suggest that the person was no less impressive. His adversaryOthmar Spann (1923: 770) describes Weber as ‘a demonic, restless man, with theability of making a personal impact on other people’.24 Schumpeter conceived of sociology as the ‘theory of the reciprocal relations betweenindividuals and groups of individuals in the social whole’ (1915: 556 [my translation]),alternatively as the ‘theory of social institutions and principles of social organization’(Schumpeter, 1954: 103, see also 1954: 20f).25 Schumpeter’s sociological writings have recently been collected by R. Swedberg in J.Schumpeter (1991), The Economics and Sociology of Capitalism.26 The reason Schumpeter gives for this view, has much in common with Durkheim’sargument for the existence of social facts. The individual is born into an alreadyexisting objective class situation. ‘The individual belongs to a given class neither bychoice, nor by any other action, nor by innate qualities – In sum his class membershipis not individual at all. It stems from his membership in a given clan or lineage. Thefamily, not the physical person, is the true unit of class and class theory’ (Schumpeter[1927] 1951: 148).27 ‘Around Christmas, 1903, I read Menger’s Grundsätze der Volkwirtschaftslehre for the firsttime. It was the reading of this book that made an “economist” of me’ (Mises, 1978:33).28 In a later article (Mises, 1961: 125), Mises dismisses Weber’s distinction ‘between“social action” and other action’ as being ‘of little importance’.29 Support for this conclusion can be found especially in Mises ([1957] 1985: PartThree), but also in Mises (1942: 249–52; 1962: 45–51).30 There are other Austrians, such as Murray N. Rothbard (1962: 2, 435, note 6; 1979:15–17, 57–61) and Ludwig Lachmann (1969), who have written about methodologicalindividualism, but their discussions are largely exegetical and do not add anythingof importance to the subject.31 In a note on page 38, Hayek writes: ‘It has long been a subject of wonder to me whythere should, to my knowledge, have been no systematic attempts in sociology toanalyze social relations in terms of correspondence and noncorrespondence, orcompatibility and noncompatibility, of individual aims and desires’. When gametheory developed as a systematic attempt in this direction, no sociologists wereinvolved. But one of the pioneers was the Austrian economist Oscar Morgenstern.32 The Counter-Revolution of Science consists of three parts: (1) ‘Scientism and the Study ofSociety’, first published in Economica, 1942–4, (2) ‘The Counter-Revolution of

Science’, first published in Economica, 1941, and (3) ‘Comte and Hegel’, originallypublished in Measure, 1951.33 In his Discourse on Method, Descartes writes: ‘Finally it is almost always easier to put upwith their [custom’s] imperfections than to change them, just as it is much better tofollow the main roads that wind through mountains, which have gradually becomesmooth and convenient through frequent use, than to try to take a more direct routeby clambering over rocks and descending to the foot of precipices’ (1985: 118). I donot know if this idea originates with Descartes, but it is clearly the idea of pathdependence that he is hinting at.34 It is of some interest, for a sociologist, to know that Frank H. Knight translated MaxWeber’s General Economic History into English in an edition that appeared as early as1927. It was thus the first work, by Weber, to appear in English. Three years beforethe much more famous The Protestant Ethic and the Spirit of Capitalism.35 In Risk, Uncertainty and Profit, Knight does not seem to be aware of the gravity of theproblem of rationality in situations of uncertainty. In a review article on Wicksteed’sThe Common Sense of Political Economy, however, Knight implies that the existence ofrisks is not ‘compatible with rational economic choice’ and observes that risk-taking‘cannot possibly be brought under economic principles unless, again, the risk ismeasurable and hence insurable, and very doubtfully even then; at least, I do not seehow the principles of rational choice through quantitative comparison of incrementscan be applied to the desire to gamble, even in such a mechanical case as the roulettewheel, where the essence of the interest is ignorance as to whether the result will be again or a loss’ (Knight 1956: 111).36 In his autobiographical Against the Stream (1975: 6–16), Myrdal tells us that he startedas a methodological individualist and ended as an institutional economist.5 Society as subjectively meaningful interactionNotes 3671 As a reminder, we may point out that also some economists, belonging to the AustrianSchool, such as Ludwig von Mises and Friedrich von Hayek, adhere to this intersubjectivisttheory of society.2 See Lewis and Smith (1980), Alexander (1987: ch. 12), Joas (1987) and Wiley (1994).Of the four pragmatists, it is common to link Mead to Dewey, partly because theywere friends, but according to Lewis and Smith (p. 6), it is possible to link Peirce andMead, because both of them were conceptual and social realists and to distinguishthem from James and Dewey, who were nominalists. Wiley accepts the view of Lewisand Smith, but goes on to argue that Peirce and Mead are united by a commontheory of a semiotic self.3 The influence of Adam Smith on Chicago sociology, is well known and partly mediatedby Albion Small (1907), the founder of the Chicago School of Sociology, whotraced the origins of sociology to Adam Smith. It is often pointed out (see, e.g.Stryker, 1980: 18; Fine 1990: 138 and Udehn, 1996: 99) that Cooley’s idea of alooking-glass self can be found in Adam Smith’s The Theory of Moral Sentiments ([1759]1976: 206). It seems to be less well known that Mead’s ‘generalised other’ exhibitsstriking similarities to Smith’s ‘impartial spectator’ (Eriksson, 1988: 317f; Udehn,1996: 100). The influence of German Romanticism and of Hegel on Cooley, Deweyand Mead is equally well-known, but more diffuse.4 Joas (1987: 94) warns against reducing the Chicago School to epigons of Germanthought, but does not deny that there was an influence. Thomas and Znaniecki, forinstance, were influenced by German Völkerpsychologie (pp. 96ff), and so was Mead,who took the idea of gesture from Wilhelm Wundt (Mead, 1964: 102; see also Joas[1980] 1985: 94–98). Between 1888 and 1891, Mead studied, first with Wundt andlater with Dilthey and even started to write a dissertation under the latter’s direction

368 Notes(Joas, [1980] 1985: 17–20). The influence of Simmel on symbolic interactionism iswell-attested and needs no further elaboration.5 In the case of Mead, I am certain that he was not a theoretical individualist. In thecase of Cooley and Thomas, I am not that certain. See, for instance, Lewis and Smith(1980: 158–66, 187f), who argue that Cooley and Thomas were psychical interactionsistsand social nominalists.6 The main source of Mead’s theory of the self is Mind, Self and Society (1934), especiallypart III, but see also the supplementary essay ‘The Self and the Process ofReflection’, based on a stenographic copy of Mead’s 1927 course in social psychologyand complemented by notes made by one of his students in 1930. Other sources arethe article on ‘The Social Self ’ (1913) and the texts from 1914, 1917 and 1927 (thefirst and third based on student’s class notes), published for the first time in TheIndividual and The Social Self (1982), The Philosophy of the Present ([1932] 1980: 68–90; theonly book published in Mead’s own lifetime), Movements of Thought in the NineteenthCentury (1936: 360–85) and The Philosophy of the Act (1938: 89–91, 150–3, 374–6,445–7). The concept of ‘generalised other’ does not appear in Mead’s early texts.The first appearance of this concept that I have been able to detect is in the article ‘ABehavioristic Account of the Significant Symbol’ (1922), in Selected Writings (1964:246).7 On Mead’s theory of the ‘I’ and the ‘me’, see 1964: chs 12 and 13 (originallypublished 1912 and 1913); 1982: 46, 56f (lecture 1914) and 1934: 173–222, 273–81.8 Both Faris and Blumer gave courses in social psychology at the department of philosophyat the University of Chicago. While Faris gave an introductory course, however,Blumer replaced Mead as teacher of the main course. Since Blumer was a sociologist,the course in social psychology was eventually moved to the sociological department.The differences between Faris and Blumer are revealed by comparing Faris’s article‘The Social Psychology of George Mead’ (1937/8) with Blumer’s article ‘SociologicalImplications of the Thought of George Herbert Mead’ (1969: ch. 2). Such a comparisonfully supports the view of Lewis and Smith on this matter.9 Blumer introduced the term ‘symbolic interactionism’ in his article ‘SocialPsychology’ (1937: 153) to designate a group of social psychologists who, in contradistinctionto the stimulus-response approach, ‘emphasizes the active nature of the child,the plasticity of this nature, and the importance of the unformed impulse’. In hisSymbolic Interactionism (1969: 1), Blumer says that ‘[t]he term “symbolic interactionism”is a somewhat barbarous neologism that I coined in an offhand way in an article…The term somehow caught on and is now in general use’. By calling Blumer the‘architect of symbolic interactionsim’ I polemise with Huber (1973: 275), whosuggests that Mead was the architect and with Fischer and Strauss (1979: 485) whocall Mead the ‘founding father’ of symbolic interactionism. As I see things, Mead wasthe most important progenitor, but not the architect, or founder, of symbolic interactionism.This is not just a quarrel about words. There are important differencesbetween Mead and Blumer, which become obscured if Mead is tied too closely tosymbolic interactionism.10 The differences between Blumer and Mead have been recognised by many commentatorsin the recent literature on symbolic interactionism. See, for instance, MacPhailand Rexroat (1979), Lewis and Smith (1980: 155, 170ff), Collins (1985: 198–204),Alexander (1987: 215ff) and Joas (1987: 84). See, however, Johnson and Picou (1985:54ff), who reject the analysis of Lewis and Smith, and denies that there is any essentialdifference between Mead and Blumer.11 According to J.G. Meltzer and B.N. Petras (1973: 6ff), the Chicago School stressesverstehen, indeterminacy and process, while the Iowa School is characterised by empiricism,determinacy and structure.12 According to W. Skidmore (1979: 189), ‘it is not altogether clear whether or notKuhn’s “self theory” is indeed symbolic interactionism’.

Notes 36913 Manfred F. Kuhn, the leader of the Iowa School of Symbolic Interactionism, criticisedMead and Cooley for excessive subjectivism, and for failing in their relationbetween self and other (Kuhn, 1967: 171–84).14 In ‘Parsons as a Symbolic Interactionist: A Comparison of Action and InteractionTheory’ (1974), J.H. Turner attempted a synthesis of symbolic interactionism andfunctionalism. In his ‘Comments on Turner’, Blumer declared that he sees ‘the jointact as primarily an “organization” of action’ (p. 59). The methodological implicationis that ‘whereas symbolic interactionists would study the process of symbolic interaction,Parsons would study the products of this interaction’ (p. 61). In his reply toBlumer, Turner (1975) complained that symbolic interactionism has ‘few conceptsand propositions to explain the operation of social processes at other than a microlevel of interpersonal interaction’ (p. 65). Also, ‘it is not clear that interactionism hasany concepts to describe emergent social structures’ (p. 66). Turner is, of course,perfectly correct in these judgements, but it is not necessarily a defect of a socialpsychological theory to lack such concepts, even if Mead thought so, and Turnerthinks so, unless the adherents of such a theory claims that it is a comprehensivetheory of society. Concerning Blumer, himself, it may be pointed out that he seems tohave grown more individualistic over the years. In his early article on ‘SocialDisorganization and Individual Disorganization’, a more holistic language is found.He writes, for instance, about people occupying different social positions.15 On society as a negotiated order, see A.L. Strauss, et al. (1963: 147–69; 1964: 15f,146–58, 293–315, 373–77).16 The dramaturgical approach is used in The Presentation of Self in Everyday Life (1959)and in Frame Analysis (1974: ch. 5). The model of the game is used in the essay ‘Fun inGames’, in Encounters (1961: 17–81) and in Strategic Interaction (1970).17 The statement that the individual may be influenced by social processes, does notalter this conclusion, since social processes have their basis in a plurality of individuals.18 For an introduction to phenomenological social science, in general, and phenomenologicalsociology, in particular, see Natanson (1973) and Wolff (1979). For a morecomprehensive treatment of phenomenological sociology and ethnomethodology, seeRogers (1983).19 On the problematic relation between Schutz and Husserl, see Natanson (1968: 235ff;1973: 23), Tiryakian (1965: 30).20 See, especially, A. Gurwitch (1962: 71f), who argues that there is a close affinitybetween Dilthey’s descriptive psychology and the thought of the later Husserl, andsince Schutz’s ideas are a continuation of those of the later Husserl, Schutz’sphenomenology can be seen as a fulfilment of Dilthey’s project of a descriptivepsychology. See also M. Natanson, who writes (1968: 241): ‘To be sure, it is perfectlyreasonable to take the book [Schutz’s The Phenomenology of the Social World] as a sympatheticcorrective for certain philosophical inadequacies in Weber’s sociology; it isequally permissible to classify the work as an application of Husserl’s ideas to the fieldof the social sciences’.21 It may be added that Husserl read Schutz’s Phenomenology of the Social World in the yearof its appearance (1932) and called Schutz a ‘serious and thorough phenomenologist’and ‘one of the very few persons who managed to penetrate to the deepest meaningof my life work, which, unfortunately, is accessible only under heavy difficulties’(quoted in Wagner, 1983: 46). Husserl also asked Schutz to become his assistant.These facts make it highly unlikely that Schutz was guilty of serious misunderstandingof Husserl’s phenomenology.22 In his later writings, Schutz takes his point of departure in Husserl’s idea of a lifeworld(see, e.g., Schutz, 1940: 178–86; 1945b: 82–5, 92–5; 1959: 93, 95).

370 Notes23 All articles referred to in my presentation of Schutz’s phenomenological sociology,except ‘The Social World and the Theory of Action’, can be found in Alfred Schutz,Collected Papers I: The Problem of Social Reality (1962).24 See Schutz, (1945a: 542–5; 1953: 11–14; 1955: 164–6, 193–207).25 As we saw in chapter 3 (p. 69), the idea of ‘second-order concepts’ goes back toWilhelm Dilthey, who intended roughly the same thing as Schutz by his ‘concepts ofsecond degree’. Schutz does not refer to Dilthey, so it is impossible to say if heborrowed the idea of concepts of second degree from him.26 I think, of course, of the structuralist Marxism of Louis Althusser and his followers.27 Merleau-Ponty read the manuscript to Husserl’s The Crisis of European Sciences in 1939,and saw in it a break with Husserl’s earlier transcendental phenomenology. His ownexistentialist phenomenology is largely based on Husserl’s idea of a life-world. SeeMerleau-Ponty, ‘Phenomenology and the Sciences of Man’ (1947) and ‘ThePhilosopher and Sociology’ (1960) (both in Merleau-Ponty, 1974: 247–79; 95–110).28 See Bourdieu (1990: ch. 1) for information on the intellectual scene in France in the1960s. Bourdieu’s favourable attitude to Merleau-Ponty has probably much to do withthe importance attached to the body in his main work Phenomenology of Perception(1945). As is well known, also Bourdieu attaches much importance to the body in hissociology.29 See also Craib (1976), Douglas and Johnson, eds (1977) and Kotarba and Fontana,eds (1984).30 On the relation between Garfinkel and Parsons, see Heritage (1987: 226–32). See alsoSharrock and Andersson (1986: 23–38), who argue that Garfinkel did not intend tobelittle the contribution of Parsons and never rejected it wholesale. See alsoAlexander (1987: 252ff), who suggests that Garfinkel was driven into a more radicalrejection of Parsons by his more radical followers. It has also been maintained thatethnomethodology is the result of an unsuccessful attempt to create a synthesis of theincompatible theories of Parsons and Schutz (Rogers, 1983: 133–6), but this is probablyto exaggerate the influence of Parsons on ethnomethodology.31 Like Blumer, with regard to the term ‘symbolic interactionism’, Garfinkel refuses totake any responsibility for the use of the term ‘ethnomethodology’. While few woulddeny that Garfinkel is the origin, both of the term and the theoretical movement, it isalso clear that it has developed in different directions and become increasinglyheterogenous (see, e.g. Douglas, 1971: 32–5; Attewell, 1974: 182ff; Zimmerman,1978: 6–8).32 See, e.g., Habermas, ([1970] 1988: 109ff; [1981] 1984: 224–30); Mayerl (1973: 276),Wolff (1979: 532–40), Sharrock and Andersson (1986: ch. 1), and Collin (1997: 28)on the influence of phenomenology. See, however, Zimmerman (1978: 6–8), whowishes to play down the dependence of ethnomethodology on phenomenology.33 My view that ethnomethodology belongs with both phenomenology and symbolicinteractionism, as a version of the interactionist and subjectivist, or interpretive,theory of society, seems to be shared by Douglas (1971: 12ff), Habermas ([1970]1988: 109ff; [1981] 1984: 124–30) and Knorr-Cetina (1981: 1ff).34 See Giddens (1979: 172–8), Lemert (1979) and Bleicher (1982: 126–30).35 On the ethnomethodological view of society as a process, see also Attewell (1974:201–5).36 See Cicourel (1964: 197ff; 1974: 13–28), Manning (1971: 244–51) and Zimmerman(1971).37 See Attewell (1974: 196), Skidmore ([1975] 1979: 236f, 255–58), Collins (1981: 85f)and Bleicher (1982: 135f).38 SeeHabermas([1981]1984:92–117;[1981]1987:124–31)andAlexander(1987:271ff).39 On the role of pre-understanding for understanding in ethnomethodology, seeMehan and Wood (1975: 365–67). Bleicher (1982: 135f), who also notes the similaritiesbetween ethnomethodology and hermeneutics, nevertheless observes that, at a

Notes 371certain point, they part company: ‘In contrast [to ethnomethodology], thehermeneutic paradigm has evidenced features of social existence, variously referredto as tradition, language, consensus, which underlies not only the diffused object ofethnomethodology but also the activities of the researcher’.40 See Cicourel (1964: 178ff) and Lemert (1979: 291). Lemert, who argues for ahomology between ethnomethodology and French structuralism is also aware that thehomology is weak and that ethnomethodology belongs mainly in the subjectivistcamp. ‘Ethnomethodology remains largely a psycho-sociological micro-study’ (p.290). It is also ‘true that there are many themes of ethnomethodology which partakeof the more traditional humanistic idea that the relativity of social order is due to thefreedom of the social self ’ (p. 292).41 Lemert (1979) advances the contrary argument, i.e., that the Zimmerman–Wieder–Pollner group is the least subjectivistic, since it takes the most extreme view ofindexicality. Lemert’s interpretation is wrong, I believe, first of all in the interpretationof indexicality, but also because it fails to appreciate the radical subjectivism ofZimmerman–Pollner–Wieder’s views. See, especially, Zimmerman and Pollner (1971:94f) and Wieder (1974: 169–72). Zimmerman (1978: 8–10) has argued thatethnomethodology is neither individualistic, nor subjectivistic. What the argumentamounts to, however, is that ethnomethodology does not take its point of departure inthe isolated individual, but is intersubjectivistic.42 See also Schlegloff and Sacks (1973: 262) and Turner (1970: 181–3). For an argumentthat comes closer to that of Mandelbaum, see Bittner (1965: 78f) and Wieder (1974:162f).43 See, e.g. Giddens (1979: 173–8), Knorr-Cetina (1981: 18f) and Zimmerman (1978:10f).44 The distinction between linguistic and extra-linguistic rules is, of course, a difficultone to draw, but Zimmerman (1978: 10f) definitely goes too far, when seeminglysuggesting that linguistic rules are those rules that determine ‘the use of naturallanguage expressions in interactive situations’. This view would turn virtually allsocial rules into lingustic rules. See Goffman (1983: 2, 24, 32) for a good analysis ofthe problem.45 For a discussion and critique of Douglas’s analysis of Durkheim’s theory of suicide,see Giddens (1977: ch. 9).46 See also Hilbert (1990), Button (1991), Sharrock and Button (1991: 138ff). See,however, Heritage, who maintains that Garfinkel stands squarely in the sociologicaltradition: ‘Garfinkel’s lifelong theoretical endeavours have been directed at a range ofconceptual issues which have always been central topics of sociology. These issues –the theory of social action, the nature of intersubjectivity and the social constitutionof knowledge – are complex and tightly interwoven’ (1987: 225).47 For an introduction and overview of social constructionism, see Gergen (1994) andBurr (1995). For critical, but nuanced, discussions, see Collin (1997) and Hacking(1999).48 This non-individualistic side of Berger and Luckmann has been developed into antiindividualismin the social constructionist theory of organisation.49 See, e.g., Meyer (1986), the various contributions to Thomas, et al. eds, (1987) andDiMaggio and Powell (1991). In a recently published volume, edited by Mary Brintonand Victor Nee (1998), the attempt is made to launch a more individualistic ‘newinstitutionalism in sociology’, based on rational choice. In the first chapter of thisvolume, Victor Nee concludes that ‘[m]ethodological holism in sociology has been anobstacle to acceptance of the choice-theoretic approach underlying the new institutionalistparadigm’ (p. 11). He goes on to maintain that ‘a form of methodologicalindividualism represents a mainstream of modern empirical sociology’. This is adebatable claim, but it is not totally unlikely to be true. A growing number of sociologistsnow seem to adopt what I call structural individualism (see pp. 318f). In this

372 Notesversion of methodological individualism, individual action and social structure aretwo, equally important, elements in the explanation of social phenomena. If I understandNee correctly, it is this weak version of methodological individualism, that hesupports. At least, he reckons with ‘the reciprocal interactions between purposiveaction and social structure’ (pp. 4f).50 Other important forerunners include Marx, the neo-Kantian sociologists Max Weberand, especially Georg Simmel, and the sociologists of knowledge, Max Scheler andKarl Mannheim.51 Bourdieu’s main debt to phenomenology is to Maurice Merleau-Ponty and his analysisof the body.6 Positivism in philosophy and social science1 For an ‘empiricist’ interpretation of methodological individualism, see Brodbeck(1958), Nagel (1961: 535–40), Morgenbesser (1967: 163f), Lachenmeyer (1970:213–16; 1971: 66ff), Addis (1975: 43f), Quinton (1975–6: 7–11), Papineau (1978: 17f),Hughes (1981: 42–7, 63) and Rosenberg (1988: 117f). See also Hund (1982: 271,277), and Archer (1995: 33f), who argue (incorrectly) that methodological individualismis an outgrowth of British empiricism. See, finally, Danto (1965a: 266f) andLukes (1973: 122f), who argue against the ‘empiricist’ interpretation of methodologicalindividualism.2 ‘Imply’ is too strong, since it is possible for an empiricist to be an instrumentalistdefending methodological holism and collectivism, as a parsimonious way ofanalysing complex social reality.3 It is, thus, typical that empiricist critics of Weber and Hayek, such as Brodbeck (1954:141–54), Rudner (1954: 164–8) and Nagel (1961: 540–6) have been more troubled bythe subjectivism, than by the individualism of their methodological individualism.4 On logical constructions, see also Russell the articles ‘Mysticism and Logic’ (1914)and ‘The Relation of Sense-data to Physics’ (1914) in Russell ([1917] 1963).5 See Urmson (1958: 43f); Pears (1968: 58f), and Ayer (1972: 58).6 See Russell ([1917] 1963: 108–10, 153–5; 1912: 48).7 It was J.C. Smuts who coined the term ‘holism’ in his Holism and Evolution ([1926]1936), a work agreeing with the theory of emergent evolution, as suggested by HenriBergson, C. Lloyd Morgan and others. On Neurath’s Marxist sympathies, see([1931a] 1973: 345–53, 358–71; [1931/2] 1959: 306–15).8 On physicalism and unified science, see Neurath (1930/31: 107, 116ff; [1931a] 1973:324ff, 416f; 1931b: 621–23; [1931/2] 1959: 286ff).9 See Neurath (1930/1: 121f; [1931a] 1973: 409–11; [1931/2] 1959: 303f; 1944:26–8).10 For a similar positivist view of social science by a philosopher see Rudner (1966).11 We are told, however, that ‘[i]n finding the meaning of various social concepts theverbal form “What would happen – ?” is very useful’ (1938a: 137). Bridgman deniesthat operationalism implies quantification and would thus seem not to be responsiblefor the road taken by operationalism in sociology.12 I have adopted the term ‘systematic empiricism’ from Willer and Willer (1973). Analternative term would be ‘instrumental positivism’, as suggested by Bryant (1985).On the origin and history of systematic empiricism, see also Obershall (1972),Bierstedt (1977), Halfpenny (1982) and Hacking (1990). An excellent treatment of theearly stage of statistical sociology can be found in Goldthorpe (2000: ch. 12).13 See, e.g., Mills (1959: chs 3 and 5), Österberg (1988: 18ff), Desrosières (1991).14 Among methodological individualists, this view of systematic empiricism has beenmaintained by, among others, John Stuart Mill, Carl Menger, Gabriel Tarde, FloydAllport, Friedrich von Hayek, Ludwig von Mises, Tjalling Koopmans and James

Notes 373Coleman. Among micro-sociologists, Douglas (1974: 4–12) and Knorr-Cetina (1981:1f) and Cicourel (1981: 51ff) have expressed a similar opinion.15 The first to make this argument clearly and persuasively was probably the Germanstatistician Wilhelm Lexis (1837–1914). See Goldthorpe (2000: 283–6). It may be thatLexis was the source of the views of methodological individualists, like Menger andTarde, on this matter.16 Mises invokes the sociologist Gabriel Tarde in support of his view of statisticalmeasures. According to Tarde ([1890] 1962: 102), the regularities demonstrated bystatistics are due to laws of imitation.17 A further corollary of this view is that macroeconomics is not holistic, merely byvirtue of dealing with relations between aggregates. If macroeconomics is holistic – aswe shall see in chapter 8, a controversial question – the reason for this is not that itdeals with aggregates, but must be sought somewhere else.18 Among critics of systematic empiricism making this argument, see Adorno (1957:70–3), Mills ([1959] 1970: 78–80), Phillips (1973: 38). Among representatives ofsystematic empiricism, making the same argument, see above all, Lazarsfeld (1962:471; 1968: 623; 1970: 14f; 1973: 18). See also Boudon (1968: 29, 48), Coleman(1969a: 87; 1969b: 517), Galtung (1969: 25, 37ff, 148–52), van den Eden andHüttner (1982: 12, 26, 55f) and Bryant (1985: 133, 140f, 150, 154–61).19 A note of caution: It is not a question of epistemological subjectivism. Systematicempiricism, in general, and operationalism, in particular, are of course objectivistic inan epistemological sense, as are all sciences concerned with measurement (see, e.g.Lundberg [1929] 1942: 45–79; 1939: 10–22). What I wish to say is that systematicempiricism has been much preoccupied with objective investigations of subjectivephenomena, such as attitudes, opinions and beliefs.20 See, e.g., Lazarsfeld (1959: 67–73; 1970a: 17–35; 1970b: 310–15), Boudon (1968:29ff), Coleman (1969a: 102–5; 1969b: 517–28), Galtung (1969: 25f).21 On contextual propositions, as an example of ‘structural’ analysis, see Lazarsfeld(1959: 69; 1968: 623f; 1970b: 313–15) and Coleman (1969a: 103f).22 For a survey of the fallacy of the wrong level, see Riley (1963: 703ff) and Galtung(1969: 45–8, 79f). See also Elster (1978: 97–106).23 Cf. Hummel and Opp (1968: 218f; 1971: 65–8) and Coleman (1969a: 102–5).24 The novelty of this reorientation should not be exaggerated. There has always beentalk about mechanisms in social science and even before that, as far back as the rise ofmechanism in natural science. My impression is that, until quite recently, the use ofthe term ‘mechanism’ was most common among holistic social theorists: functionalists,structuralists, institutionalists and system theorists. It has been very common, forinstance to conceive of social institutions and organisations as social mechanisms.See, e.g., Lipset ([1959] 1963: 45) who calls democracy a social mechanism andMerton (1967: 44), who suggests that a role-set is a social mechanism. Interesting, inthis connection is that the systematic empiricist, George A. Lundberg tried to launchthe idea of mechanism, on a large scale, in his Foundations of Sociology (1939: 159ff). Heuses ‘the word mechanism to describe that necessary and sufficient set of relations orcircumstances by means of which any behavior whatsoever takes place, whether it bethe jump of an electric spark, the thoughts of a human being … the milling of amob, or the vote of a deliberative assembly’. Mechanisms, in this sense, include folkways,customs, traditions, mores and institutions.25 Rom Harré differs from the other realists in being more micro-oriented and moreinterested in psychology. He is not a methodological individualist, however, but wantsto ‘steer a middle course between individualism and collectivism’ (1993: 12, 34ff). Seealso Harré (1979: 83ff) and Harré, Clarke and de Carlo (1984: 2, 6, 65ff).26 Theories of social exchange have probably played a more important role in socialanthropology than in sociology (see, e.g. Davis, 1992). My neglect of these theories is

374 Notesnot due only to the fact that I am a sociologist, myself, but is justified by the fact thattheories of exchange in anthropology are not typically individualistic.27 Homans sometimes complains about his theory being called a theory of exchange(Homans, 1980: 20; 1983: 35). Since nothing important seems to be at stake I followthe common habit of calling Homans’s version of behaviourism a ‘theory ofexchange’.28 Homans’s four autobiographical writings are ‘Autobiographical Introduction’ toSentiments and Activities (1962), ‘A Life of Synthesis’ (1968), ‘Steps to a Theory of SocialBehavior. An Autobiographical Account’ (1983) and Coming to My Senses. TheAutobiography of a Sociologist (1984).29 On Homans’s critique of functionalism, see also Homans (1964a: 809ff; 1964b:963–7 and 1967b: 64–70).30 ‘See, e.g., Homans (1964a: 811ff; 1964c: 221–31; 1969: 1–24 and 1975: 313ff).31 The second half of this statement exhibits an interesting similarity to many statementsby ethnomethodologists, but it is one of very few similarities between theseapproaches. For one thing, ethnomethodologists would certainly disagree with thedeterminism of behaviourism. Much more than Parsons, behaviourism treats thehuman being as a ‘cultural dope’.32 See Homans (1964c: 11–31 and 1964d: 113–31), for a restatement of the originalformulation. See Homans (1967a: 30–78) and (1974: ch. 2) for a somewhat modifiedversion.33 See, however, the contributions to Burgess and Bushell (ed.) (1969) and Hamblin andKunkel (ed.) (1977).34 His contribution to the theory of exchange is Exchange and Power in Social Life ([1964]1986). ‘The implicit assumption [of this work] is that macrosociological theory restson the foundation of microsociology. This is the assumption I have come to question’([1964] 1986: vii). The difference between Homans and Blau are clearly visible intheir exchange over the relevance of psychology to the explanation of socialphenomena (Homans, 1975a: 313–28; 1975b: 340–3; Blau, 1975: 329–39). Thedifference between Blau and Emerson emerge clearly in his contribution to Karen S.Cook (ed.), Social Exchange Theory (Blau, 1987), which brings out his doubts about theneed for a microsociological foundation for structural macrosociology.35 The theory of networks, like that of social exchange, has strong roots in anthropology,which I neglect because of ignorance. See Scott (1991: ch. 2) for an accountof the development of social network analysis.36 In his last contribution to a theory of social exchange, Emerson might seem to moveback to a more psychologistic approach (cf. Turner, 1987: 225ff). His main concern,now, is to develop a theory of value, as the foundation of the theory of socialexchange (Emerson, 1987). This appearance is probably deceptive, however, andmerely reflects his wish to develop both the individualist and structuralist elements ofhis theory.37 See Haines (1988) and Mathien (1988) for a discussion of network analysis andmethodological individualism. Both argue that network analysis is a form of methodologicalindividualism, but not a radical or extreme form of this principle. Bothobserve that the focus of network analysis is on social relations, but this is not incompatiblewith methodological individualism. Haines argues that there are importantsimilarities between network analysis and Giddens’s theory of structuration, but this aweak argument. Network analysis works with a dualistic view of individual action andsocial structure, whereas the main point of Giddens’s theory of structuration is toreplace this dualism with a duality, which is something else.38 For a survey, see Wippler (1978c) and Raub (1982). See also Wippler (1978b).39 See, e.g. Lindenberg (1985a; 1985b; 1995), Opp (1978; 1985; 1988) and Wippler andLindenberg (1987).

7 Popperian methodological individualismNotes 3751 This procedure may be seen as part of the early Popper’s strategy of replacing metaphysicalissues with their methodological counterparts. In The Logic of ScientificDiscovery (1934), Popper said that ‘not a few doctrines which are metaphysical, andthus certainly philosophical, could be interpreted as hypostatizations of methodologicalrules’ ([1934] 1972: 55). The ‘principle of causality’, for instance, was excluded‘as “metaphysical” from the sphere of science’ and replaced by the correspondingmethodological rule (p. 61).2 See Popper ([1945] 1966: vol 1: 31–4, vol 2: 9–21; 1957: 26–34; 1962: 103–7).3 See Popper ([1962] 1968: 18–21; 1972: 194–7; 1974a: 12–23; 1976: 17–31). There aretimes, however, when words do matter, even to Popper. Such is the case, for instance,when some members of the Frankfurt School call him a ‘positivist’. ‘A last word on theterm “positivism”. Words do not matter, and I do not really mind if even a thoroughlymisleading and mistaken label is applied to me. But the fact is that throughout my life Ihave combated the positivist epistemology, under the name of “positivism”. I do notdeny, of course, the possibility of stretching the term “positivism” until it can beapplied even to an opponent of positivism such as myself. I only contend that sucha procedure is neither honest nor apt to clarify matters’ (Popper, 1970: 261).4 The separation of persons (or individuals) is not accidental. In his contribution toRoads to Freedom: Essays in Honour of Friedrich von Hayek, Popper says (1969: 199) aboutsituational logic, that ‘[i]t makes room for giving full weight … not only to individualsbut also to institutions’ [my italics].5 Cf. also Popper ([1962] 1968: 350): ‘Democracy as such cannot confer any benefitsupon the citizen and it should not be expected to do so. In fact democracy can donothing – only the citizens of the democracy can act (including, of course, those citizenswho comprise the government)’.6 The incompatibility of Popper’s theory of objective knowledge with methodologicalindividualism, as usually conceived, has been pointed out by several commentators.See, e.g. Ackerman (1976: 166–72), Ingram (1976: 134f), O’Hear (1980: 197ff),Chalmers (1985), Udehn (1987: 204–7) and Stokes (1997: 63).7 On the theory of emergent evolution as a form of holism, see, e.g., Nagel (1961:366ff), Phillips (1976: 32f;), Udehn (1987: 85ff) and Bunge (1996a: 258ff).8 The earlier methodological individualists were in no doubt that individualism andinstitutionalism are opposite methodologies. This opposition was part of the Battle ofMethods and it is confirmed by all Austrians, who lump together the GermanHistorical School, American institutionalism and mainstream sociology, asapproaches committed to a ‘realistic’, empirical investigation of social institutions asopposed to the individualism of theoretical economics. See, e.g., Schumpeter ([1914]1954: ch. 4; 1954: 20f, 26f, 983–95, 864–77) and Mises ([1933] 1976: 4–12; [1949]1966: 4; 1990: 210). The opposition between individualism and institutionalism isalso obvious to the founder of the Chicago School of Economics, Frank H. Knight,who contrasts ‘the “institutional” view of economic activities’ with ‘the individualisticor contractual aspect’ (Knight, 1956: 18; see also 1935/6: 330).9 The expression ‘Ghost in the machine’ is, of course, a reference to Gilbert Ryle’s TheConcept of Mind (1949), where it is argued that the mind, or soul, is a ghost in themachine of the body, that is non-existent.10 In a recent article on Popper’s situational logic, Jarvie (1998) defends Popper’s ‘institutionaland situational individualism’ (p. 374) against the more radical individualismthat he finds in contemporary rational choice. He refers to Jon Elster as one of hisexamples, and this is justified, since Elster is a reductive individualist (see pp. 310–16),but it is not justified as a criticism of contemporary rational choice as a whole. To avery large extent recent rational choice is ‘institutionalistic’, in the sense of Popper,and in some cases even ‘structuralistic’ (see chs 9–10).

376 Notes11 See, however, Wisdom (1981: 10), where he says that ‘The Weber–Popper schemaplaces enormous emphasis upon the unintended consequences of human actions’,but without mentioning institutions.12 A year later, Wisdom has reached this conclusion, himself. ‘Elsewhere I havesuggested the need to allow emergent properties that are in no sense reducible to individuals,their actions and aims, and their foreseeable unintended consequences…Such properties thus constitute a system over and above individuals. Although I put itforward as a development from, though inconsistent with, Popper’s “methodologicalindividualism”, he would seem to have originated essentially the same idea, which heconceives of as the “third world”…, or alternatively shown that the “third world” iscompatible with “methodological individualism”’ (Wisdom, 1971: 143, note 4).13 Mario Bunge, who agrees that Popper’s methodological individualism is differentfrom original methodological individualism suggests that ‘Popper’s social ontologymay therefore be characterized as individholistic rather than as consistently individualistic’(Bunge 1996b: 533; 199: 108).8 Economics: the individualist science1 There is a good reason for this. As we have seen in previous chapters, a majority ofthe founders and defenders of the principle of methodological individualism wereeconomists and philosophers. John Stuart Mill, the most important forerunner of thisprinciple, was both. Carl Menger, the first explicitly to advance methodological individualism,or atomism, was the founder of the Austrian School of Economics andone of the three founders of marginalist economics. Joseph Schumpeter, who introducedthe term ‘methodological individualism’, was also an economist, and he used itto denote the method of theoretical economics. Ludwig von Mises and Friedrich vonHayek, who were the most consistent propagators of this doctrine, stand out as themost prominent representatives of Austrian Economics after Menger. The obviousexception, Max Weber, was a sociologist, but he was also a historian and aneconomist. As a sociologist, however, his aim was to create a sociology after the modelof marginalist economics. Karl Popper, finally, was a philosopher, but he picked upmethodological individualism from the Austrian economists, so there is really nobreak in the link between methodological individualism and economics.2 The parallel existence of classical and neoclassical economics led Hollis and Nell(1975: 13) ‘to divide economic theory into two great mainstreams, one flowing fromthe pens of Marshall and Walras, the other from Ricardo and Marx … Marshall andWalras conceived the economy in the same way as a set of interrelated markets, inwhich households and firms meet as demanders and suppliers… Ricardo and Marxon the other hand, saw the economy as based on methods of production, controlledand operated by one class but worked by another’.3 For a similar view of methodological individualism in economics, see Field (1979;1984) and Mirowski (1981; 1986).4 See, however, Janssen (1993: 14), who claims that ‘contrary to what Boland (1982)argues, economic explanations of individual behaviour (even neoclassical ones) doadopt institutions as primitive terms’. I think Boland is the more correct and, ofcourse, he never denied that neoclassical economists sometimes use institutions as‘primitive terms’. Boland’s argument is that there is ‘a hidden rule on the agenda’ ofneoclassical economics not to accept social institutions as exogenous variables.5 Other prominent economists attesting to the difficulties involved in trying to understandwhat ‘Keynes really meant’ include Frank H. Knight ([1937] 1960: 92), PaulSamuelson (1946: 190f) and Frank H. Hahn (1977: 25).6 It has been maintained by Friedrich von Hayek that Keynes was not very well read ineconomics and knew little more than the theories of Alfred Marshall.

Notes 3777 On Keynes as a methodological holist, see, e.g. Leijonhufvud (1968: 387–401). Seealso Winslow (1989), who gives references to other interpretations of Keynes as anorganicist and methodological holist.8 See, e.g., Hayek (1941: 373–6), Mises ([1952] 1974: chs 4 and 5), Friedman ([1962]1982: ch 5; 1970a), and Buchanan and Wagner (1977). For a collection of criticalviews, see Hazlitt (ed.) (1960). Friedrich von Hayek occupies a special place amongthe libertarian critics of Keynes (because he was also his friend) and blends hiscritique of Keynes, the economist, with an almost hagiographic picture of Keynes,the person and intellectual (see Hayek, 1995).9 See Hayek (1995: 231f, 60f, 241ff, 251f). Keynesian macroeconomics and econometricshave this in common; they both deal with aggregates. The difference betweenthem is that Keynes’s theory is about theoretical aggregates, whereas econometrics isabout statistical aggregates. This difference is clearly visible in the controversybetween Keynes (1939) and Tinbergen (1940). According to Lucas and Sargent(1979: 296): ‘It is the fact that Keynesian theory lent itself so readily to the formulationof explicit econometric models which accounts for the dominant scientificposition it attained by the 1960s’.10 See also Hartley (1997: 176) ‘Of course only people act; of course there is no monsterout there called aggregate output with a life of its own; of course explanations ofaggregate activity may refer to the behavior of people’. Even Keynes ([1936] 1973)motivated his aggregate consumption function with a discussion of the objective andsubjective factors influencing the propensity to consume (chs 8 and 9) and his aggregateinvestment function with a discussion of the expectations of entrepreneurs (ch.12). However, this is a very different thing from an insistence that any model ofaggregate activity must be rigorously derived from an individual’s optimisationproblem.11 The difference between Vining and Koopmans is, I believe, roughly the same as thatbetween Keynes and Tinbergen, as reflected in the well-known debate between them(Keynes, 1939; 1940 and Tinbergen, 1940).12 It may be noted that Klein (1946: 93f) sees no problem in the fact that, in economictheory, individual behaviour is really the behaviour of households and firms, whileArrow ([1951] 1968: 641) seems to admit that, really, it implies a collective basis ofeconomics.13 In Hicks’s original article (Hicks, 1937), the second curve was called LL. It was AlvinHansen, who changed the symbolic representation of the diagram to IS-LM, in his AGuide to Keynes (1953). It may be noted that Hicks originally used the IS-LM diagramto represent the classical theory, of which Keynes’s theory is a special case (cf.Barrère, 1988: xviiif). It may also be noted that he later, when turning from equilibriumto more dynamic analysis, came to assign a rather limited role to IS-LM analysis(Hicks, [1980] 1982).14 Clower and Leijonhufvud were not, of course, first to question the neoclassicalsynthesis. There had been many economists before them objecting to turning Keynesinto a neoclassical economist. Among the more important is Cambridge economistJoan Robinson, who insisted that the Keynesian revolution was a break with Marshalland orthodox neoclassical economics (1962: ch. 4; 1971).15 See Arrow (1974: 4), Weintraub (1979), Janssen (1993) and Hartley (1997: 194).16 Actually there are, at present, at least seven different approaches to macroeconomicsin contemporary economics (see Phelps, 1990): (1) The Macroeconomics of Keynes,(2) The Monetarist Tradition, (3) the New Classical School, (4) the New KeynesianSchool, (5) Supply-side Economics, (6) Neoclassical and Neo-Neoclassical RealBusiness Cycle Theory and (7) Non-Monetary Theories of UnemploymentFluctuations: The Structuralist School.17 The role of maximising behaviour is not made explicit in Friedman’s monetaristtheory, but is discussed in his famous essay on ‘The Methodology of Positive

378 NotesEconomics’ (1953). His conclusion is that it is not really needed for standardeconomic theory to be true and may be replaced by a theory of selection, assuggested by Armen A. Alchian (1950).18 Bruno Ingrao and Giorgio Israel (1990: chs 1–2), trace the roots of Walrasian generalequilibrium theory to France, rather than to Scotland. According to them, A.R.J.Turgot (1727–81) also used the term ‘equilibrium with respect to economicphenomena’ in a letter to David Hume in 1769–70.19 According to Walras ([1874] 1984: 69) prices are natural phenomena, in the sensethat they cannot be affected by individuals exchanging on the market (cf. Mirowski,1981: 595).20 Historically, the household has been both a consumption and a production unit, andthe rise of the firm as the main production unit, is part of the development of capitalism.Thus, it is only in modern market economies that the household is only aconsumption unit and the firm is the main production unit.21 It may be added that Arrow in (1968), saw more problems with the individualism ofgeneral equilibrium than the fact that its atoms are households and firms, not individuals.‘It is perhaps interesting to observe that “atomistic” assumptions concerningindividual households and firms are not sufficient to establish the existence of equilibrium:“global” assumptions … are also needed (though they are surelyunexceptionable). Thus, a limit is set to the tendency implicit in price theory, particularlyin its mathematical versions, to deduce all properties of aggregative behaviorfrom assumptions about individuals’ (p. 382).22 See also Kincaid (1996: 250f; 1998: 93) and Hartley (1997: 184–6).23 On the market as a social institution, see Swedberg (1994) and Burns (1995).24 Serious objections to general equilibrium theory have been advanced from lessorthodox economists, almost from the beginning. An important early critique wasmade by the Austrian economist Oscar Morgenstern (1935), who focussed on theparadoxes resulting from the assumption of perfect foresight. Not only is this anincredibly unrealistic assumption, implying that human beings are some kind ofdemi-gods, but, were it true, it would lead to ‘contradictions’, or paradoxes. Incompetition, mutual knowledge of one another’s plans for the future leads tocomplete indeterminism, since all actors would change their plans on the basis of thisknowledge, indefinitely. Another weighty critique, similar to that of Morgensterncame from G.B. Richardson (1959), who observed that there is nothing in generalequilibrium theory to suggest how the plans of the independent economic actors areco-ordinated. A third important objection was made by the game theorist, MartinShubik (1975), who pointed to three major flaws in general equilibrium theory: ‘(1) Itis independent of the number of competitors. (2) It is a totally static tightly coupledsystem. It is error free…. (3) It implicitly requires symmetric information’ (p. 546).What is new in the recent development of economic theory, is that the problems havebeen recognised by more mainstream economists, some of which have made importantcontributions to general equilibrium theory.25 Walras started his analysis of market exchange, by suggesting that ‘The marketswhich are best organized from the competitive standpoint are those in whichpurchases and sales are made by auction, through the instrumentality of stockbrokers,commercial brokers or criers acting as agents who centralize transactions in such away that the terms of every exchange are openly announced and an opportunity isgiven to sellers to lower their prices and to buyers to raise their bids’ (Walras, [1874]1984: 83f). He goes on to mention other markets, such as grain markets and fishmarkets, which are also highly competitive, but finally decides to use the Paris StockExchange as his preferred example for bringing out the details of his analysis.26 For a summary, see the survey by Shafer and Sonnenschein (1982).27 For a more extended critique of economic theory and of the theory of general equilibrium,in particular, see Morgenstern (1972).

28 See, e.g., Udehn (1987: 173), Willer and Skvoretz (1997: 5f) and Kinkaid (1998:131–5). See also Hollis (1987: ch. 10; 1994: chs 6–8) for an argument to a similareffect.29 A more obvious example would be the famous prisoner’s dilemma, which in the realgame behind the name presupposes social institutions in the form of laws andprisons.9 The new institutional economicsNotes 3791 See Langlois (ed.), (1986), Hodgson (1988), Eggertsson (1990), Mäki, Gustavsson andKnudsen, eds (1993), Rutherford (1994) and Furubotn and Richter, eds (1997).2 For a discussion of the relation between the New Institutional Economics andMarxism, see Heijdra, Lowenberg and Mallick (1988, 1991) and Dorman (1991).3 It is generally agreed that the old institutionalism is holistic and the new institutionalismindividualistic. See for instance, Dugger (1979) and Hodgson (1989: 250–253),who also agrees with Rutherford that the old institutionalism is behaviouristic. Seealso Wilber and Harrison (1978) who agree that the old institutionalism is bothbehaviouristic and holistic.4 The main contributions to the new institutionalism in sociology are collected inMeyer and Scott (1983), Thomas, et al. (1987), Powell and DiMaggio (1991) and Scottand Meyer (1994).5 Coase makes extensive reference to Frank H. Knight on the subject of uncertainty,which has immense importance for the emergence of the new institutionaleconomics.6 According to one commentator at least, it means a lot: ‘What I see as crucial in thenew institutional economics theory is not so much the introduction of transactioncosts as an assumption to be made in the calculation of the decisions, or even thepreoccupation with institutions, for old style institutionalists also raised questionsabout the latter. The core of the theory lies in establishing the connection betweenboth. With this aspect in mind, Coase’s now famous 1937 study extended methodologicalindividualism – which also identified with the neoclassical theorists – to itsultimate consequences’ (Möschel, 1993: 89).7 Frank H. Knight also wrote an article on ‘Some Fallacies in the Interpretation ofSocial Cost’ (1924), which vaguely anticipates some of Coase’s critique of Pigou in‘The Problem of Social Cost’ (Coase, 1960).8 For an interesting analysis of antique economics as a theory of household managementand of male dominance, with special reference to Xenophon, see Foucault([1984] 1985: Part Three, 141–84).9 In Rome the head of the household, or household authority, was called paterfamilias.His authority was divided, by Roman law, into potestas over children and slaves, manusover wife and dominium over his possessions (see Nicholas, 1962, 65ff).10 See, e.g., England and Kilborne (1990) for an approach, which includes power in theanalysis of the family.11 A large part (463–81) of the Oeconomicus is devoted to an amazingly timeless analysisof the ‘agency problem’.12 See Weber ([1922] 1978: 222ff, 1095; 1927: 275ff) and Swedberg (1998: 49).13 Trainn Eggertsson (1990: xii, 7–10) makes a distinction between the ‘new institutionaleconomics’ of Williamson and the ‘neoinstitutional economics’ of Douglass Northand others. The reason for making this distinction is that Williamson relies onHerbert Simon’s idea of ‘satisficing’ behaviour, rather than on maximising behaviour.14 See Williamson (1975: 21–6; 1981: 553f, 571; 1985: 44ff; 1986: 173ff).15 See Williamson (1979: 246–54; 1981: 555; 1985: 52–61; 1986: 179–84; 1988: 69–71;1995: 225).

380 Notes16 See Williamson (1985: 18, 29, 213ff; 1986: 171; 1988: 88; 1993: 75f, 96).17 In a later work, Alchian and Woodward (1987: 113–15) complement the picture bystressing the importance of dependence and the temptation to ‘holdup’ it may create.Holdup becomes possible when a resource is dependent upon a specific use, but losesvalue when separated from this particular use. This is similar to Williamson’s ‘assetspecicificity’ (see p. 268).18 See also Levinthal (1988: 155ff) and Sappington (1991).19 See Grossman and Hart (1986: 692, 716f), Hart (1989: 164) and Hart and Moore(1990: 1121).20 I think it is possible to see a certain similarity to Marxism here. In the Marxianversion of the argument, the fact that in capitalism, workers do not own their meansof production, means that much more is at stake for them than for the capitalists.21 Unlike most other economists, North (1981: ch. 3) does not remain satisfied with thetheory of the social contract. He prefers a ‘predatory or exploitation theory’ (p. 21),which recognises the clash of interest between groups in society and the use ofviolence and ideology in the creation and maintenance of the state.22 Another Public Choice pioneer, Anthony Downs (1957) rejects the individualisttheory of the state. Confronted with Buchanan’s dichotomy between an organismicand an individualistic theory of the state, Downs rejects both. While the organismictheory is false, the individualistic theory is incomplete. The dichotomy between anorganismic and an individualistic theory is ‘false’ (71).23 See, e.g., Buchanan and Tullock ([1962] 1965) and Buchanan (1984: 13; 1989b: 25,37; 1987: 304).24 Also Becker wants determination by constraints, but does not include institutionsamong them. The following statement is probably directed at Becker and Friedman:‘The emphasis on economy-wide rather than on individual constraints has beenpartly responsible for a serious omission in analysis’ (Buchanan: 1989b: 42).25 See, however, Olson ([1965] 1971: 100–102), who accepts the individualist approachand criticises, as metaphysical, the idea that states, races, and classes can have‘“objective” needs and purposes beyond those of the individuals who compose them’.26 Mancur Olson’s theory in The Rise and Decline of Nations (1982) is a more mixed bag,which conforms partly to the first and partly to the second pattern.27 In another article, Cheung (1970) takes his point of departure in Coase’s discussion ofexternalities. Part of his analysis is the presentation of a model of a common fishingground, with implications that ‘point to the possibility of a theory of property rightsformation. Such a theory, however, is not intended here’ (p. 64).28 See Giddens (1977: ch. 2) and Elster (1979b: 28–35; 1983a: 55–68).29 According to Elster (1982: 455) ‘Posner and his school actually tend toward theStrong Functional Paradigm, which most sociologists have abandoned for the moresubtle Main Paradigm’. As support for this claim, he cites the following ‘somewhatgrotesque example’ (Elster, loc.cit), from Posner: ‘The economic case for forbiddingmarital dissolution out of concern for the children of the marriage is weakened if theparents love the child (as we have defined “love”), for then the costs to the child ofdissolution will be weighed by the parents in deciding whether to divorce, and(assuming “full” love) they will divorce only if the gains to them from the divorceexceed the costs to the child – in which event the divorce will be welfare-maximizing.If, as suggested earlier, love is a factor of growing importance in the production ofchildren, this might help to explain why the law is moving toward easier standards fordivorce’ (Posner [1972] 1977: 106).30 For a good overview of the evolutionary branch of the new institutional economics,see J.J. Vromen (1995).31 My quotations are from Hayek (1967), but his view of cultural evolution is the samein Hayek (1982: 17–30) and also, if less explicitly in Hayek (1988: 11–37). An observationof some interest, from the point of view of methodological individualism, is

Notes 381Hayek’s suggestion, in his last works, that the elements of society are individuals andorganisations (1982: 46–8; 1988: 37f).10 Rational choice individualism1 Among political scientists, David Knoke is committed to a methodology which is bothstructuralist and individualistic and similar to the structural-individualism of sociologists(see Knoke, 1990: ch1).2 It is common to attribute the individualist concept of function to the anthropologistBronislaw Malinowski and the collectivist concept to his colleague A.R. Radcliffe-Brown.3 For a more detailed argument to this effect, see Udehn (1987: 171f). For a critique ofHarsanyi’s explanation of social stratification, see also Foldes (1968: 334–7) andBrown (1973: 130–7).4 Coleman is aware that the typical sociologist will object to this procedure, that homosociologicus reenters the game in the form of socialised players, who have already internalisedthe norms of society, and that norms will enter the game also in the form ofthe rules of the game. His reply is that the player’s of games are socialised only in thesense of being rational and prudent agents, capable of estimating the long-term effectof conformity to norms (Coleman, 1990c: 179f).5 Coleman’s theory of social systems, is developed in a large number of articles andbooks from about 1965 onwards. Some sources are Coleman (1966; 1972; 1973: 61ff;1979).6 Coleman (1992a: 137). See also (1987b: 133ff; 1990b; 1990c: 45ff; 241ff; 1990e).Coleman, himself, is of a slightly different opinion: ‘Most social theory not based onmethodological individualism assumes the existence of social norms, and most theorythat is based on methodological individualism disregards their existence altogether’(Coleman, 1986b: 1326).7 See also Coleman (1990b: 49ff) and (1990e: 254ff).8 I believe, for instance, that it is possible to detect a certain change of view betweenthe article on ‘Social Structure and a Theory of Action’ (1975) and ‘Rational Actorsin Macrosociological Analysis’ (1979) and increasingly so in ‘Introducing SocialStructure into Economic Analysis’ (1984) and ‘Psychological Structure and SocialStructure in Economic Models’ (1987c). The change is complete in ‘The EconomicApproch to Sociology’ (1992a). Social structure now appears as an irreduciblephenomenon which cannot be entirely endogenized.9 A very clear and concise statement of Coleman’s structuralism can also be found inan article on ‘Social Systems’ from 1971. Here Coleman makes the typically holisticdistinction between persons and positions and argues that social organizations aremade up of positions, or roles, not of persons. ‘The position has certain attributes,that is actions which its occupant should carry out and actions he can expect frompersons in other roles. But the individual is merely the tenant of the role; the organizationexists independently of him, just as an appartment building existsindependently of its tenants’ (Coleman, 1971: 73).10 Coleman relies for distinction between natural persons and corporate actors on thelegal scholars, Otto von Gierke and F.W. Maitland. Both of these, but the former inparticular, are among the arch-holists in the history of social thought.11 See also Coleman (1991a; 1991b) and (1992c: 117).12 The word ‘starting point’, sometimes used by Coleman (see also 1990c: 31), might betaken as an indication that his methodological individualism is procedural rather thansubstantive. My conjecture is that it is both.13 See Coleman (1987b: 135; 1990b: 35; 1990c: 54, 241, 325; 1992b: 273, 281).

382 Notes14 At other times, he characterises functionalism, not by teleology at the level of thesocial system, but by its failure to account for regularities in system behaviour ‘by themediation of purposes or goals on the part of persons’ (Coleman, 1979: 75).15 Coleman (1986c). See also (1986a; 1987a; 1990c: ch. 1). See also Abell (1996: 256ff)for a presentation of Coleman’s scheme.16 On an abstract level, I think there are some similarities between Coleman’s approachand that of Berger and Luckmann, the networks analysis of Willer, among others, thestructural individualism of Wippler, Raub and Lindenberg, the morphogeneticapproach of Margaret Archer and, perhaps, Mario Bunge’s systemism. At the mostabstract level, I can also see similarities between Coleman and Marx.17 See, however Coleman (1986c: 1321), where also Talcott Parsons is accused of ‘disregardingthe very structural configurations that are essential elements in determiningthe social outcome of a combination of individual actions’. That we have to do witha type 3 relation is clear from the sentence that follows: ‘This micro-to-macroproblem is sometimes called by European sociologists the problem of transformation’.18 According to Boudon, a ‘state of nature’ is characterised by the fact that individualsact without taking into consideration the effects of their actions upon other individuals,but often within an institutional context. This is an unusual view. A state ofnature, as usually conceived, is characterised by the absence of social institutions.What Boudon calls a ‘state of nature’, is better characterised as a ‘Robinson Crusoeapproach’, as exemplified by marginalist economics.19 This is not surprising, since Boudon, himself, observes (1984: 33) that ‘[t]he explanationof M provided by Coleman … consists precisely of a clarification of the terms ofM = M{m[S(M]}. The explanation of M, referred to is that of the dissemination ofnew products by Coleman and his colleagues. It may be noted that P has become Min this formula, but the meaning is the same. M means ‘macrosocial actions’.20 See, e.g., Boudon (1975: 398ff; 1981: 121, 134ff; 1982: 197ff; 1988a; 1988b: 761ff;1998).21 See, e.g., Boudon, (1981: 12–27, 58–85, 145; 1982: 7, 154; 1984: ch. 4.)22 See Daniel Little (1998: ch. 1), for a lucid and concise presentation of this approachto Marxism.23 Among the more noteworthy contributions are Levine, Sober and Wright (1987),Sensat (1988) and Weldes (1989).24 On the critique of functionalism, see Elster (1978: 121f; 1979b: 28–35; 1982: 454–63;1983a: 55–68; 1985: 27–37; 1986b: 202–7).25 See Elster (1982: 463ff; 1983a: 165ff; 1985: 8ff; 1986b: 207ff).26 On the principle of charity, see Elster (1979a: 81f; 1979b: 116f, 154ff; 1983b: 117;1989a: 35f). Elster’s principle of charity is similar to Weber’s proposal that we shouldalways start with the ideal type of rational action and, then, treat concrete actions asdeviations from this ideal type (Weber [1922] 1978: 6).27 See Elster (1982: 453f, 462, 469, 473; 1983b: 111–13). That is to say Elster, seems tobelieve that rational choice is necessarily individualistic, not that rational choice is theonly version of methodological individualism. In later writing, he takes explicit exceptionto the latter belief (see, e.g., Elster, 1989a: 105).28 On the two-filter model of human action, see Elster (1978: 160; 1979a: 65–8; 1979b:77, 113ff; 1982: 463f; 1985: 9).29 According to G.J. Stigler and G.S. Becker (1977: 76f) preferences are stable andcommon to all people. According to Elster (1979a: 75–78; 1979b: 137–47), however,they change and differ between people. My guess would be that, in general, for thoseanxious to defend human autonomy the opportunity set must be large, and actionexplained mainly by preferences.30 Elster quotes the following statement by A.S. MacFarland: ‘The structuralist wouldstudy the fence around the cattle; the behaviourist [read ‘rational choice’, LU] wouldstudy the activity of the cattle within the fence’. I believe that Macfarland’s image of

structuralism is, on the whole, more accurate than Elster’s image of strong determinism(see Elster, 1979b: 113f; 1982: 464; 1983b: 115–8; 1986a: 23).31 Elster would eventually reach the latter conclusion with respect to social norms. See,e.g., Elster (1989a: ch. 3).32 See, e.g. Durkheim ([1897] 1951: 313): ‘The force uniting the conglomerate multitudeof individual cases, scattered over the face of the earth, must necessarily be externalto each of them’ and ‘It is external to each average individual taken singly’ (p. 316;but do not ask me what kind of creature an ‘average individual taken singly’ is). Seealso E. Gellner (1956: 170f): ‘For any individual, the mores, institutions, tacit presuppositions,etc., of his society are an independent and external fact, as much so as thephysical environment and usually more important. And if this is so for each individual,it does follow that it is so for the totality of individuals composing a society’.33 Elster recognises four possible types of interdependence in social life: ‘First the rewardof each depends on the choice of all, through general social causality. Second, the reward ofeach depends on the reward of all, through envy or altruism. Thirdly, the decision of eachdepends on the decision of all: this is the specific contribution of game theory. Lastly, thepreference structure of each depends on the actions of all, through socialization and similarmechanisms’ (1983a: 77; see also 1982: 464; 1985: 10; 1986a: 7). The relevantdistinction for my present purposes is that between the first and third types of interdependencies.According to Elster, the first type is characterised by ‘intentionalinteraction between intentional beings’, whereas the third is ‘purely causal interaction betweenintentional agents’. This takes place when each agent acts upon unjustified assumptionsabout the behaviour of others, e.g., when each agent believes he is the only one whois adjusting to the environment, whereas all others merely follow habit, or tradition’(1983a: 83f; see also 1978: 159; 1979a: 68f). Why unjustified? It seems as if Elsterentertains something like an existentialist belief that social order is due to bad faith,concerning the behaviour of others. But is it not rather the case that the social environmentis to a considerable extent really parametric, rather than strategic?34 See also Taylor (1988: 77–93), for a somewhat extended discussion of structuralism asan alternative to methodological individualism.35 See, for instance, the Symposium on his Logic and Society in Inquiry, 23, 1980: 139–232,the ‘Comments’ on his ‘Marxism, Funcionalism and Game Theory: The Case forMethodological Individualism’ in Theory and Society, 11, 1982: 483–539 andSymposion on his Making Sense of Marx in Inquiry, 29, 1986, 3–77.36 In his reply to Elster’s ‘Marxism, Functionalism and Game Theory’, he makes thesomewhat curious statement that ‘Methodological individualism is the deductivemethod: it attempts to deduce historical observations from basic postulates on individualbehavior that are sufficiently fundamental to be self-evident’ (Roemer, 1982b:514).37 This does not mean that Przeworski rejects structural analysis. See the extendeddiscussion of rational choice and methodological individualism in Przeworski (1990)for a clarification on this point.38 Pettit is also a structuralist, who has produced one of the most interesting casesagainst methodological individualism (Pettit [1993] 1996: ch. 5; see also Jackson andPettit, 1992). Daniel Little, who also combines microfoundations with structuralism,feels that it is too holistic to be called methodological individualism. He prefers theterm “‘methodological localism” in contrast to both holistic social science andmethodological individualism’ (Little, 1998: 198).11 Why methodological individualism?Notes 3831 See Laudan (1977: 54) for a taxonomy of the various cognitive relations, which canexist between two, or more, theories.

384 Notes2 See, e.g. Mandelbaum (1955: 312), Thomas (1979: 32). Cf. Papineau (1978: 17f;46–9). Methodological individualism has even been interpreted by Runciman (1972:24) as a principle about the testability of laws, employing collective concepts, by referenceto individual behaviour. There is no justification for this interpretation in thewritings of leading methodological individualists.3 Very often methodological individualists move from methodological to ontologicalindividualism, without acknowledging the difference between the two, as if the latterentailed the former. I have called it the ‘ontological turn’ or ‘twist’ (p. 125).4 See, also Addis (1975: 38), Radnitzky (1988: xv): ‘Reduction in the strong sense makesthe reduced theory dispensable’, and K. Lennon and D. Charles (1992).5 See, e.g., Sklar (1967), Schaffner (1967), Nickles (1973), Krüger (1976) and Smith(1992). The latter introduces the term ‘strong reduction’ (p. 22) to cover those caseswhich correspond to the ‘common’ meaning of ‘reductionism’.6 On the distinction between epistemological, or philosophical, and scientific reduction,see Popper (1972: 290–5). Cf. Also Nagel (1949: 123; 1961: 358).7 For a critique of Kemeny and Oppenheim, see, e.g., Hempel (1969b: 193), Sklar(1967: 113–7) and Nagel (1970).8 See also T.R. Girill (1976a: 69; 1976b: 387), who maintains that micro-reduction ischaracterised by the occurrence of the parts in the explanans and the whole in theexplanandum. Assuming that individuals are the parts of social wholes, this would bea good explication of methodological individualism.9 See Broad (1937: 44ff) and Bergmann (1944: 212–4; 1957: 131ff).10 See, e.g. Durkheim ([1895] 1982: 102) who says ‘[a] whole is not identical with thesum of its parts. It is something different, and its properties are different from those ofits component parts’ and Smuts ([1926] 1936: 102), who suggests that ‘[i]t is veryimportant to recognise that the whole is not something additional to the parts: it is thesynthesis of parts in a definite structural arrangement...’ Durkheim’s view is echoedby Bergmann (1944: 217; 1957: 158–61), who suggests that the existence of elementaristiccomposition laws is one way of giving meaning to ‘additivity’ and‘emergence’. The most penetrating analysis of the claim that ‘the whole is more thanthe sum of its parts can be found in Nagel (1961: 380–97).11 It is a matter of controversy whether translation of concepts counts as reduction ornot. The simple answer is that it depends upon what you mean by ‘reduction’. If youconceive of sciences as constituted by their vocabulary, translation is reduction, but ifyou believe that they are constituted by their theories, translation is not reduction.12 See, e.g. Homans (1964a: 813–8; 1964c: 967–73; 1967b: 35–43) and Hummel andOpp (1968: 220–3; 1971: 81–6). The same claims on behalf of individualistic theoriesis made by virtually all advocates of microfoundations.13 See Elster (1983: 23f; 1985: 5f; 1989b: 73f; 2000: 24f).14 See, e.g., Friedmann (1982: 20), the contributions to G. Radnitsky (ed.) (1987; 1988)and D. Charles and K. Lennon, eds (1992). See also Little (1991: 192) Kinkaid (1997:ch. 1) and Callender (1999). For a non-reductive view of inter-theoretic relations, seeDarden and Maull (1977) and Maull (1977).15 See, e.g., Macdonald and Pettit (1981: 119ff) and Pettit (1984), who uses the idea ofsupervenience to accept the ‘expressive autonomy of concepts’, but in defence ofontological and methodological individualism. See also Currie (1984), Levine, Soberand Wright (1987), Little (1991: 190–9) and Kinkaid (1986: 507–9; 1997: 70–4) whouse supervenience to argue against methodological individualism.16 See, e.g., Watkins (1952a: 29), Homans (1970a), Hummell and Opp (1968; 1971) andElster (1989b: 74). This does not necessarily imply that social scientists can bereplaced by psychologists. It might still be the case that social scientists differ frompsychologists in terms of subject matter (see, e.g., Brodbeck, [1958] 1968: 297;Gibson, 1968: 106–11; Papineau 1978: 48; Lindenberg, 1996).

Notes 38517 See, e.g. Brodbeck (1954: 148ff; [1958] 1968: 301f), Mandelbaum (1955: 305ff; 1957:222f; 1977: 235ff), Gellner (1956: 174f; 1960: 514–16), Danto (1965a: 269ff), Putnam(1973–4: 135), Addis (1975: 42f), Vanberg (1975: 4), Mellor (1982: 51) and Pettit([1993] 1996: 117ff).18 Weber’s method of verstehen is deemed psychologistic by Runciman (1972: 74f),Outhwaite (1975: 46, 104) and Elster (2000: 26). According to Torrance (1974: 148)Weber was ‘pushed in a psychologist direction by [his] methodological individualism’.Weber’s relation to psychology and his rejection of psychologism, is a recurrent themein Huff (1984).19 See, especially, Weber ([1903–6] 1975: 101–20, 129–52, 172ff; 1908a; 1908b:119–24; [1922] 1978: 7, 13, 19).20 See Weber ([1903–6] 1975: 189f; 1908a: 29f, 32f; [1913] 1981: 154ff; [1922] 1978:19; 1949: 74f, 88f).21 In Weber (1949: 90), it is suggested that the ideal type is a heuristic device, but not ahypothesis, although it offers guidance to the construction of hypotheses. In Weber([1903–6] 1975: 189f) and ([1913] 1981: 157), however, the ideal type has the statusof a hypothesis. In Weber ([1922] 1978: 9–11, 18f), finally, the ideal type is both aheuristic device and a hypothesis.22 See Hempel (1965: 166) and Vanberg (1975: 107f, 121).23 See Popper ([1945] 1966: vol. 2, 97; 1957: 140–3; [1962] 1976: 102f).24 See Popper ([1945] 1966: vol. 2, 90, 97; 1957: 153f; [1962] 1976: 101).25 On this point, see, especially Lichtheim (1965). See also Lessnoff (1974: 106–8) andVanberg (1975: 109ff).26 See, e.g. Little (1991: 195–99; 1998: 10–13; 203) and Kinkaid (1997: 46–8). Cf. alsoSatz and Ferejohn (1994: 85).27 On the connection between methodological and normative individualism, seeBrodbeck (1954: 142), Burman (1979: 357), Blaug (1980: 49f), Heine (1983), Hodgson(1988: 70–2), Madison (1990: 47), Shand (1990: 2, 31f), Kingdom, (1992), Vanberg(1994: 1) and Kinkaid (1997: 2).28 This does not prevent Elster from making the following statement: ‘From Stalin to theRed Guards this outlook [a speculative philosophy of history] has led to a disregardfor individuals that goes far beyond the denial of methodological individualism’ (Elster,1985: 117).29 A good example of this affinity, is the constitutional economics of James Buchanan.This theory is based on the idea of unanimous consent, which is the political counterpartof economic exchange. According to James Buchanan, it is also based onmethodological individualism. But, when used in the context of politics, methodologicalindividualism transforms into political individualism, which, in its turn, is basedon ethical individualism (see Udehn, 1996: 174ff).30 This argument was advanced by Hayek in a number of articles in the 1930s (see,especially, ‘Socialist Calculation’, I–III, ‘Economics and Knowledge’ and ‘The Use ofKnowledge in Society’, all in Hayek, 1948). A similar argument has been used byPopper (1957: 79f).31 There is another kind of humanism, which sees the dignity of human beings in theirachievements rather than in their freedom. Among these achievements are scienceand literature, works of art and social institutions. The latter kind of humanism tendsto be holistic.32 See also Strawson ([1962] 1968: 58–96) and James (1984: 70–5).33 ‘The truth contained in environmentalism is the cognition that every individual live ata definite epoch in a definite geographical space and act under the conditions determinedby his environment. The environment determines the situation but not theresponse. To the same situation different modes of reaction are thinkable and feasible.Which one the actors choose depends upon their individuality’ (Mises, 1957: 326).

386 Notes34 For a critique of sociology of knowledge, from an individualist point of view, Mises(1933: 186–94; [1949] 1966: 4–7, 72–91; [1957] 1985: 122ff), Hayek ([1952] 1963:192f; 1955: 76, 89, 180) and Popper ([1945] 1966, vol. 2: ch. 23).35 According to Watkins (1970: 176f), behaviourism is not only inadequate when appliedto human beings, but ‘unfair to rats’. Agassi (1977: 166) maintains that ‘vulgarbehaviourism, like vulgar Marxism, is but an object of inquiry, not a target for anyserious criticism’.36 See especially Popper, who is the most energetic defender of mind–body dualism, orinteractionism. His most elaborated argument can be found in Popper (1977).37 See Boudon (1981: ch. 8; 1982: chs 1 and 7) and Elster (1978: 160; 1979a: 66–8;1986a: 22f). On the normative element in Elster’s methodological individualism, seeWarren (1988). Also Philip Pettit’s ([1993] 1996) individualism may be interpreted asa defence of human freedom, or autonomy.12 Methodological individualism restated1 It is very common, in the literature on methodological individuals, to make thedistinction between methodological and ontological individualism. At least onecommentator (Scott, 1960) has also made a distinction between methodological andepistemological individualism. Rajeev Bhargava (1992: ch. 1) distinguishes betweenontological, explanatory and semantic individualism, and a variety of different formsof explanatory individualism. What Bhargava calls ‘semantic individualism’ has beendiscussed by Pettit (in Macdonald and Pettit, 1981: ch. 3) as a matter of the ‘expressiveautonomy’ of institutions (see also Tiles, 1984 and Pettit, 1984). I conceive ofsemantic, or expressive individualism as a version of epistemological individualism.Since scientific theories have a linguistic form, I believe that semantics is part of epistemology.2 According to Weber’s well-known definition, ‘[a]ction is “social” insofar as its subjectivemeaning takes account of the behavior of others and is thereby oriented in itscourse’ (Weber [1922] 1978: 4). According to Winch’s also well-known alternative,social action is rule-governed action, presupposing a social setting. (Winch, 1958:116–20).3 This interpretation of methodological individualism – as a principle of explanation –is favoured also by Philip Pettit in Macdonald and Pettit (1981: ch. 3) and by virtuallyall who interpret methodological individualism as a quest for microfoundations, especiallythose who identify microfoundations with rational choice.4 The first to point out that methodological individualism is often stated as an ontologicalthesis was probably Goldstein (1958: 1ff; 1959: 240–1). On the distinctionbetween methodological and ontological individualism, see also Koertge (1974: 198f);Ruben (1982: 1; 1985: 1ff) and Kinkaid (1996: 187–90; 1997: 13f). According to M.Bunge (1979: 15), ‘every theoretical view of society and, for that matter, of anyconcrete object, has two components: an ontological and a methodological. Theformer concerns the nature of society, the latter the way to study it’.5 Empiricists, when faced with metaphysical, or ontological problems have typicallyturned them into epistemological, logical, or methodological questions. This is theonly way to deal with them, when metaphysics is forbidden. The strategy of Popper isto transform ontological (and epistemological) problems to methodological ones.According to Popper ([1934] 1972: 55), ‘not a few doctrines which are metaphysical… could be interpreted as typical hypostatizations of methodological rules’.6 See, e.g., Mises ([1949] 1966: 42ff), Hummell (1973: 138f), Homans (1980: 19) andSztompka (1979: 299).7 See Ryle (1949) for a classic analysis of dispositions. On intentions, see Anscombe(1968: 144–52) and MacIntyre (1972: 48–70).

Notes 3878 According to Watkins (1955: 58) material circumstances influence social phenomenaonly indirectly. Lukes (1968: 124f) has divided facts about individuals in four classes:(i) genetic make-up; brain states; (ii) aggression; gratification; stimulus-response; (iii)co-operation; power, esteem; (iv) cashing cheques; saluting; voting’. See Ruben (1985:128–30) for a comment on Lukes’s classification.9 This is not to deny that institutions are, themselves, made up of actions of individuals.10 Lukes (1973: 107–9) uses the term ‘epistemological individualism’ in a different wayto denote ‘a doctrine about the nature of knowledge, which asserts that the source ofknowledge lies within the individual’.11 On methodological individualism as a thesis about the reducibility of social laws, see,e.g., Brodbeck (1954: 155; [1958] 1968: 286ff), Mandelbaum (1957), Martin (1972:69), Brown (1973: 137), Addis (1975: 55–74) and Mellor (1982).12 See especially Watkins (1957b: 107n; 1959c: 243). See also Danto (1965a: 275),D’Agostino (1979: 28), Flew (1985: 42–47), Perry (1979: 7–15; 1980; 62; 1983: 68).13 The reason is that there might be explanations without laws. Historical explanations,for instance, rarely invoke social laws. In the recent discussion of methodological individualism,there has been strong arguments for explanations in terms of socialmechanisms, which also make no use of laws. Actually, it is hard to find examples ofgenerally accepted laws in the social sciences.14 This is true especially of the Popperians. There are those, who see methodologicalindividualism exclusively as a principle of explanation, e.g., Danto (1965a) Lukes(1968: 122), D’Agostino (1979: 28) and Flew (1985: 46).15 Jarvie’s use of the terms ‘explicans’ and ‘explicanda’, instead of the more commonexplanans and explanandum, may suggest that we have to do with an ‘explication’ inCarnap’s sense of that term (1970: 8). But this is not the case. It is clear from thecontext that Jarvie intends the explanation of social phenomena, not the explicationof social concepts.16 This interpretation of methodological individualism has been suggested by Gellner(1956: 165), Lukes (1968: 129), D’Agostino (1979: 28), Opp (1972: 130; 1979: 9f).17 This interpretation of methodological individualism can be found in Nozick (1977:359), Opp (1979: 49, 65) and Lindenberg (1985a: 107).18 There are different opinions among those advocating social mechanisms, about therelation of mechanisms to laws. While some accept social laws, but want to is to playdown their importance, others deny their existence altogether. My own view, is that itis always possible to formulate statements about social mechanisms as laws, but thereis little or no point in doing so. The deductive-nomological ideal has turned into astraight-jacket on social science, which we had better through away. Laws are alwaysconditional and subject to the ceteris paribus. In other words, they are statements oftendency. In social life, the conditions are so many, that stating general laws seemsrather pointless. It is better to look for causal mechanisms, in the absence of laws.19 See, e.g. Elster (1989b: ch. 1; 1999: ch. I) and Hedström and Swedberg (eds) (1998).

BibliographyAbell, P. (1996) ‘Sociological Theory and Rational Choice Theory’, pp. 252–73 in B.S.Turner (ed.) The Blackwell Companion to Social Theory, Oxford: Blackwell.Addis, L. (1975) The Logic of Society, Minneapolis: University of Minnesota Press.Adorno, T.W. (1957) ‘Sociology and Empirical Research’, pp. 68–86 in T.W. Adorno, et al.The Positivist Dispute in German Sociology, London: Heinemann.Agassi, J. (1960) ‘Methodological Individualism’, The British Journal of Sociology, 11,244–70.—— (1972) ‘I – Listening in the Lull’, Philosophy of the Social Sciences, 2, 319–32.—— (1975) ‘Institutional Individualism’, The British Journal of Sociology, 26, 144–55.—— (1977) Towards a Rational Philosophical Anthropology, The Hague: Martinus Nijhoff.—— (1997) ‘Celebrating The Open Society’, Philosophy of the Social Sciences, 27, 486–525.Ahrne, G. (1990) Agency and Organizatiom: Towards an Organizational Theory of Society,London: Sage Publications.—— (1994) Social Organizations: Interaction Inside, Outside and Between Organizations, London:Sage Publications.Albert, H. (1988) ‘Hermenutics and Economics: A Criticism of Hermeneutical Thinkingin the Social Sciences’, Kyklos, 41, 573–602.Albrow, M. (1990) Max Weber’s Construction of Social Theory, London: Macmillan.Alchian, A.A. (1950) ‘Uncertainty, Evolution and Economic Theory’, Journal of PoliticalEconomy, 58, 211–21.—— (1961) ‘Some Economics of Property Rights’, pp. 127–49 in Economic Forces at Work,Indianapolis: Liberty Press.—— (1965) ‘The Basis of Some Recent Advances in the Theory of Management of theFirm’, Journal of Industrial Economics, 14, 30–41.—— (1979) ‘Some Implications of Recognition of Property Right Transaction Costs’, pp.233–54 in K. Brunner (ed.) Economics and Social Institutions, Boston: Martinus Nijhoff.Alchian, A.A. and Allen, W.R. (1964) University Economics, Belmont, CA: WadsworthPublishing Company, 1967.Alchian, A.A. and Demsetz, H. (1972) ‘Production, Information Costs and EconomicOrganization’, American Economic Review, 62, 777–95.Alchian, A.A. and Woodward, S. (1987) ‘Reflections on the Theory of the Firm’, Journalof Institutional and Theoretical Economics, 143: 110–36.Alexander, J.C. (1985) ‘The “Individualist Dilemma” in Phenomenology and Interactionism’,pp. 25–57 in S.N. Eisenstadt and H.J. Helle (eds) Perspectives on SociologicalTheory, vol. 1: Macrosociological Theory, Beverly Hills: Sage Publications.—— (1987) Sociological Theory Since 1945, London: Hutchinson.

Bibliography 389Allport, F.H. (1924a) ‘The Group Fallacy in Relation to Social Science’, American Journal ofSociology, 29, 688–706.—— (1924b) Social Psychology, Boston: Houghton Mifflin Company.—— (1961) ‘The Contemporary Appraisal of an Old Problem’, Contemporary Psychology, 6,195–6.Alter, M. (1990) Carl Menger and the Origins of Austrian Economics, Boulder: Westview Press.Anscombe, G.E.M. (1968) ‘Intention’, pp. 144–52 in A.R. Whyte (ed.) The Philosophy ofAction, Oxford: Oxford Univerity Press.Antoni, C. (1940) From History to Sociology, London: Merlin Press, 1962.Archer, M.S. (1985) ‘Structuration versus Morphogenesis’, pp. 58–88 in S.N. Eisenstadtand H.J. Helle (eds) Perspectives on Sociological Theory, vol. 1: Macrosociological Theory,Beverly Hills: Sage Publications.—— (1988) Culture and Agency: The Place of Culture in Social Theory, Cambridge: CambridgeUniversity Press.—— (1995) Realist Social Theory: The Morphogenetic Approach, Cambridge: CambridgeUniversity Press.Archibald, W.P. (1976) ‘Psychology, Sociology and Social Psychology: Bad Fences MakeBad Neighbours’, The British Journal of Sociology, 27, 115–29.Aristotle (1962) The Politics, Harmondsworth: Penguin.Arrow, K. J. (1951) ‘Mathematical Models in the Social Sciences’, pp. 635–67 in M. Brodbeck(ed.) Readings in the Philosophy of the Social Sciences, New York: Macmillan, 1968.—— (1959) ‘The Role of Price Adjustment’, pp. 41–51 in M. Abramovitz (ed.) The Allocationof Economic Resources, Stanford: Stanford University Press.—— (1968) ‘Economic Equilibrium’, pp. 376–89 in D.L. Sills (ed.) International Encyclopediaof the Social Sciences, vol. 4, New York: Macmillan.—— (1974) ‘Limited Knowledge and Economic Analysis’, The American Economic Review,64, 1–10.—— (1987a) ‘Economic Theory and the Hypothesis of Rationality’, pp. 25–37 in J.Eatwell, M. Milgate and P. Newman (eds) Utility and Probability (The New Palgrave),New York: W.W. Norton & Company, 1990.—— (1987b) ‘Rationality of Self and Others in an Economic System’, pp. 201–15 inR.M. Hogarth and M.W. Reder (eds) Rational Choice. The Contrast between Economics andPsychology, Chicago: University of Chicago Press.—— (1989) ‘The State of Economic Science’, pp. 11–22 in W. Sichel, (ed.) The State ofEconomic Science.Views of Six Nobel Laureates. Kalamazoo, MI: W.E. Upjohn Institute.—— (1994) ‘Methodological Individualism and Social Knowledge’, The American EconomicReview. Papers and Proceedings, 84, 1–9.Arrow, K.J. and Debreu, G. (1954) ‘The Existence of an Equilibrium for a CompetitiveEconomy’, Econometrica, 22, 265–90.Arrow, K.J. and Hahn, F.H. (1971) General Competitive Analysis, San Fransisco: Holden-Day,Inc.Attewell, P. (1974) ‘Ethnomethodology since Garfinkel’, Theory and Society, 1, 179–210.Axelrod, R. (1984) The Evolution of Cooperation, New York: Basic Books.Ayer, A.J. (1936) Language, Truth and Logic, London: Victor Gollancz.—— (1971) Russell and Moore, London: Macmillan.—— (1972) Russell, London: Fontana.Backhouse, G. (1998) ‘Georg Simmel as an Eidetic Social Scientist’, Sociological Theory, 16,260–81.Bar-Hillel, Y. (1954) ‘Indexical Expressions’, Mind, 63, 359–79.

390 BibliographyBarrère, A. (1988) ‘Preface: The Keynesian Project’, pp. xiii–li, in A. Barrère, (ed.) TheFoundation of Keynesian Analysis, Basingstoke: Macmillan.Barry, B. (1989) Does Society Exist. The Case for Socialism (Fabian Tract 536), London: FabianSociety.Beauvoir, S. de (1949) The Second Sex, London: Picador Classics, 1988.Becker, G.S. (1962) ‘Irrational Behaviour and Economic Theory’, The Journal of PoliticalEconomy, 70, 1–13.—— (1976) The Economic Approach to Human Behavior, Chicago: Chicago University Press.—— (1981) A Treatise of the Family, Harvard, MA: Harvard University Press, 1991.—— (1983) ‘A Theory of Competition Among Pressure Groups for Political Influence’,The Quarterly Journal of Economics, 48, 371–400.—— (1985) ‘Public Policies, Pressure Groups, and Dead Eight Costs’, Journal of PublicEconomics, 28, 329–47.—— (1996) ‘In Memoriam James S. Coleman (1926–1995)’, pp. 377–9 in J. Clark, (ed.)James S. Coleman, London: The Farmer Press.Benn, S.I. and Peters, R.S. (eds) (1959) The Principles of Political Thought. Social Foundations ofthe Democratic State, New York: The Free Press.Berger, P.L. and Luckmann, T. (1966) The Social Construction of Reality, Harmondsworth:Penguin, 1971.Bergmann, G. (1944) ‘Holism, Historicism and Emergence’, Philosophy of Science, 11,209–21.—— (1957) Philosophy of Science, Westport: Greenwood Press.Bergson, H. (1889) Time and Free Will. An Essay on the Immediate Data of Consciousness, NewYork: Harper & Row, 1960.—— (1907) Creative Evolution, London: Macmillan, 1911.—— (1932) The Two Sources of Morality and Religion, Notre Dame, IN: University of NotreDame Press, 1977.Berle, A.A. and Means, G.C. (1934) The Modern Corporation and Private Property, New York:Macmillan.Berlin, I. (1954) ‘Historical Inevitability’, pp. 41–117 in Four Essays on Liberty, Oxford:Oxford University Press, 1969.—— (1976) Vico and Herder, London: Chatto & Windus.Bhargava, R. (1992) Individualism in Social Science. Forms and Limits of a Methodology, Oxford:Clarendon Press.Bhaskar, R. (1979) The Possibility of Naturalism. A Philosophical Critique of the ContemporaryHuman Sciences, Brighton: Harvester Press.Bicchieri, C. (1993) Rationality and Coordination, Cambridge: Cambridge University Press.Bienenstock, E.J. and Bonacich, P. (1997) ‘Network Exchange as a Cooperative Game’,Rationality and Society, 9, 37–65.Bierstedt, R. (1977) ‘A Critique of Empiricism in Sociology’, pp. 43–54 in M. Bulmer(ed.) Sociological Research Methods, London: Macmillan.Binns, D. (1977) Beyond the Sociology of Conflict, London: Macmillan.Bittner, E. (1965) ‘The Concept of Organization’, pp. 69–81 in R. Turner (ed.)Ethnomethodology, Harmondsworth: Penguin, 1974.Blau, P.M. (1960) ‘Structural Effects’, American Sociological Review, 25, 178–93.—— (1964) Exchange and Power in Social Life, New Brunswick: Transaction Books, 1986.—— (1974) ‘Parameters of Social Structure’, American Sociological Review, 39, 615–35.—— (1975) ‘Comment’ on Homans, pp. 329–39 in R. Borger and F. Cioffi (eds) Explanationin the Behavioural Sciences, Cambridge: Cambridge University Press.

Bibliography 391—— (1977) ‘A Macrosociological Theory of Social Structure’, American Journal of Sociology,83, 26–54.—— (1987) ‘Microprocess and Macrostructure’, pp. 83–100, in K.S. Cook (ed.) SocialExchange Theory, Newbury Park: Sage Publications.Blaug, M. (ed.) (1991) Jean-Baptiste Say 1776–1832, Aldershot: Edward Elgar.Bleicher, J. (1982) The Hermeneutic Imagination. Outline of a Positive Critique of Scientism andSociology, London: Routledge & Kegan Paul.Blumer, H. (1937) ‘Social Psychology’, pp. 144–98, in E.P. Smith (ed.) Man and Society. ASubstantive Introduction to the Social Sciences, New York: Prentice-Hall.—— (1969) Symbolic Interactionsim. Perspective and Method, Englewood Cliffs, NJ: PrenticeHall.—— (1975) ‘Exchange on Turner, “Parsons as a Symbolic Interactionist”’, SociologicalInquiry, 45, 59–65.Boden, D. (1990) ‘The World as it Happens: Ethnomethodology and ConversationalAnalysis’, pp. 185–213 in G. Ritzer (ed.) Frontiers of Social Theory. The New Synthesis, NewYork: Columbia University Press.Böhm-Bawerk, E. von (1890) ‘The Historical versus the Deductive Method in PoliticalEconomy’, Annals of the American Academy of Political and Social Science, 1, 244–71.—— (1891) ‘The Austrian Economists’, Annals of the American Academy of Political and SocialScience, 1, 361–84.—— (1896) ‘Unresolved Contradictions in the Marxian Economic System’, pp. 201–302in Shorter Classics of Eugen von Böhm-Bawerk, South Holland, IL: Libertarian Press.Boland, L.W. (1982) The Foundations of Economic Method, London: George Allen & Unwin.—— (1986) Methodology for a New Microeconomics,The Critical Foundations, Boston: Allen &Unwin.—— (1992) The Principles of Economics. Some Lies My Teachers Told Me, London: Routledge.Bostaph, S. (1976) ‘The Methodological Debate Between Carl Menger and the GermanHistoricists’, Atlantic Economic Journal, Vol. 6, 3–16.Boudon, R. (1968) The Uses of Structuralism, London: Heinemann, 1971.—— (1971) The Logic of Sociological Explanation, Harmondsworth: Penguin.—— (1975) ‘The Three Basic Paradigms of Macrosociology: Functionalism, Neo-Marxism and Interaction Analysis, Theory and Decision, 6, 381–406.—— (1977) The Unintended Consequences of Social Action, London: Macmillan Press, 1982.—— (1979a) ‘Generating Models as a Research Strategy’, pp. 51–64 in R.K. Merton, J.S.Coleman and P.H. Rossi (eds), Qualitative and Quantitative Research: Papers in Honor of PaulF. Lazarsfeld, New York: The Free Press.—— (1979b) The Logic of Social Action, London: Routledge & Kegan Paul, 1981.—— (1981) The Logic of Social Action. An Introduction to Sociological Analysis, London: Routledge& Kegan Paul.—— (1982) The Unintended Consequences of Social Action, London: Macmillan.—— (1983) ‘Individual Action and Social Change’, The British Journal of Sociology, 34,1–18.—— (1984) Theories of Social Change, Cambridge: Polity Press, 1986.—— (1986) ‘Individualism and Holism in the Social Sciences’, pp. 33–45 in P. Birnbaumand J. Leca (eds) Individualism: Theories and Methods, Oxford: Clarendon Press, 1990.—— (1987) ‘The Individualistic Tradition in Sociology’, pp 45–70 in J.C. Alexander, et al.(eds) The Micro-Macro Link, Berkeley: University of California Press.

392 Bibliography—— (1988a) ‘Explanation, Interpretation, and Understanding in the Social Sciences’, pp.236–58 in G. Radnitzky (ed.) Centripetal Forces in the Sciences, vol. 2, New York: ParagonHouse.—— (1988b) ‘Will Sociology Ever be a Normal Science?’, Theory and Society, 17, 747–71.—— (1989) ‘Subjective Rationality and the Explanation of Social Behavior’, Rationalityand Society, 173–96.—— (1996) ‘The “Cognitivist Model”’, Rationality and Society, 8, 123–50.—— (1998) ‘Social Mechanisms Without Black Boxes’, pp. 147–71 in P. Hedström andR. Swedberg (eds) Social Mechanisms. An Analytical Approach to Social Theory, Cambridge:Cambridge University Press.Bourdieu, P. (1972) Outline of a Theory of Practice, Cambridge: Cambridge University Press,1977.—— (1980) The Logic of Practice, Stanford: Stanford University Press, 1990.—— (1985) ‘The Genesis of the Concepts of Habitus and of Field ’, Sociocriticism, 2, 11–24.—— (1989) ‘Social Space and Symbolic Power’, Sociological Theory, 7, 14–25.—— (1990) In Other Words. Essays Towards a Reflexive Sociology, Cambridge: Polity Press.Bourdieu, P. and Wacquant, L.J.D. (1992) An Invitation to Reflexive Sociology, Cambridge:Polity Press.Breen, R. and Goldthorpe, J.H. (1997) ‘Explaining Educational Differentials: Towards aFormal Rational Action Theory’, Rationality and Society, 9, 275–305.Brennan, G. and Buchanan, J.M. (1985) The Reason of Rules. Constitutional Political Economy,Cambridge: Cambridge University Press.Brennan, G. and Hamlin, A. (1995) ‘Constitutional Political Economy: The PoliticalPhilosophy of Homo Economicus’, The Journal of Political Philosophy, 3, 280–303.Brennan, G. and Tullock, G. (1982) ‘An Economic Theory of Military Tactics. MethodologicalIndividualism at War’, The Journal of Economic Behavior and Organization, 3,225–42.Brentano, F. (1874) Psychology from an Empirical Standpoint, London: Routledge & KeganPaul, 1973.Bridgman, P.W. (1938a) The Intelligent Individual and Society, New York: Macmillan.—— (1938b) ‘Operational Analysis’, Philosophy of Science, 5, 114–31.Bridgstock, M. and Hyland, M. (1978) ‘The Nature of Individualist Explanation: AFurther Analysis’, The Philosophy of the Social Sciences, 8, 265–9.Broad, C.D. (1937) The Mind and Its Place in Nature, London: Kegan Paul, Trench, Trubner.Brodbeck, M. (1954) ‘On the Philosophy of the Social Sciences’, Philosophy of Science, 21,140–56.—— (1958) ‘Methodological Individualism: Definition and Reduction’, pp. 280–303 inM. Brodbeck (ed.) Readings in the Philosophy of the Social Sciences, New York: Macmillan,1968.—— (1962) ‘Explanation, Prediction and “Imperfect” Knowledge’, pp. 363–98 in M.Brodbeck (ed.) Readings in the Philosophy of the Social Sciences, New York: Macmillan,1968.Brown, R. (1973) Rules and Laws in Sociology, London: Routledge & Kegan Paul.Bryant, C.G.A. (1985) Positivism in Social Theory and Research, London: Macmillan.Buchanan, J.M. (1949) ‘The Pure Theory of Government Finance. A SuggestedApproach’, Journal of Political Economy, 57, 496–505.—— (1962) ‘Marginal Notes on Reading Political Philosophy’, Appendix (pp. 307–22) toJ. Buchanan and G. Tullock, The Calculus of Consent. Logical Foundations of ConstitutionalDemocracy, Ann Arbor: University of Michigan Press, 1965.

Bibliography 393—— (1975) The Limits of Liberty. Between Anarchy and Leviathan, Chicago: Chicago UniversityPress.—— (1979a) ‘Politics Without Romance’, pp. 11–22 in J.M. Buchanan and R.D. Tollison(eds) The Theory of Public Choice – II, Ann Arbor: University of Michigan Press.—— (1979b) What Should Economists Do?, Indianapolis: Liberty Press.—— (1984) ‘Politics Without Romance: A Sketch of Positive Public Choice Theory andIts Normative Implications’, pp. 11–22 in J.M. Buchanan and R.D. Tollison (eds) TheTheory of Public Choice–II, Ann Arbor: University of Michigan Press.—— (1986) Liberty, Market and State. Political Economy in the 1980s, Brighton: WheatsheafBooks.—— (1987) Economics. Between Predictive Science and Moral Philosophy, College Station: TexasA&M University Press.—— (1988) ‘The Constitution of Economic Policy’, pp. 103–14 in J.D. Gwartney andR.E. Wagner (eds) Public Choice and Constitutional Economics, Greenwich, CT: JAI Press.—— (1989a) Essays on the Political Economy, Honolulu: University of Hawaii Press.—— (1989b) Explorations into Constitutional Economics, College Station: Texas A&M UniversityPress.—— (1990) ‘The Domain of Constitutional Economics’, Constitutional Political Economy, 1,1–18.Buchanan, J.M. and Tullock, G. (1962) The Calculus of Consent. Logical Foundations of ConstitutionalDemocracy, Ann Arbor: University of Michigan Press, 1965.Buchanan, J.M. and Wagner, R. (1977) Democracy in Deficit. The Political Legacy of LordKeynes, New York: Academic Press.Buckle, H. (1858) History of Civilization in England, London: John W. Parker and Son.Bunge, M. (1973) Method, Model and Matter, Dordrecht: D. Reidel.—— (1979) ‘A Systems Concept of Society: Beyond Individualism and Holism’, Theoryand Decision, 10, 13–30.—— (1996a) Finding Philosophy in Social Science, New Haven: Yale University Press.—— (1996b) ‘The Seven Pillars of Popper’s Social Philosophy’, Philosophy of the SocialSciences, 26, 528–56.—— (1997) ‘Mechanism and Explanation’, Philosophy of the Social Sciences, 27, 410–65.—— (1998) Social Science under Debate, Toronto: Toronto University Press.—— (1999) The Sociology-Philosophy Connection, New Brunswick: Transaction Publishers.Burge, T. (1986) ‘Individualism and Psychology’, The Philosophical Review, 95, 3–45.Burger, T. (1976) Max Weber’s Theory of Concept Formation. History, Laws, and Ideal Types,Durham, NC: Duke University Press.—— (1978) ‘Droysen and the Idea of Verstehen’, Journal of the History of the BehavioralSciences, 14, 6–19.—— (1994) ‘Deutsche Geschichtstheorie und Webersche Soziologie’, pp. 29–104 in G.Wagner and H. Zipprian (eds) Max Webers Wissenschaftslehre. Interpretation und Kritik,Frankfurt am Main: Suhrkamp.Burgess, R.L. and Bushell, D. (eds) (1969) Behavioral Sociology. The Experimental Study of SocialProcess, New York: Columbia University Press.Burman, P. (1979) ‘Variations on a Dialectical Theme’, The Philosophy of the Social Sciences,9, 357–75.Burns, J.H. (1959) ‘J.S. Mill and the Term “Social Science”’, pp. 151–3 in J.C. Wood (ed.)John Stuart Mill. Critical Assessments, vol. IV, London: Routledge, 1991.

394 BibliographyBurns, T. (1995) ‘Market and Human Agency. Toward a Socio-Economics of MarketOrganization, Performance and Dynamics’, pp. 15–60 in C. Mongardini (ed.) L’Individuoe Il Mercato, Roma: Bulzoni Editore.Burr, V. (1995) An Introduction to Social Constructionism, London: Routledge.Burt, R.S. (1982) Toward a Structural Theory of Action. Network Models of Social Structure, Perceptionand Action, New York: Academic Press.Butler, J. (1990) Gender Trouble. Feminism and the Subversion of Identity, London: Routledge.Button, G. (1991) ‘Introduction: Ethnomethodology and the Foundational Respecificationof the Human Sceinces’, pp. 1–9 in G. Button (ed.) Ethnomethodology and the HumanSciences, Cambridge: Cambridge University Press.Cahnman, Werner J. (1964) ‘Max Weber and the Methodological Controversy in theSocial Sciences’, pp. 103–27 in Werner J. Cahnman and Alvin Boskoff (eds) Sociologyand History, New York: The Free Press of Glencoe.Cairnes, J.E. (1875) The Character and Logical Method of Political Economy, New York:Augustus M. Kelley, 1965.Caldwell, B.J. (1988) ‘Hayek’s “The Trend of Economic Thinking”’, pp. 175–8 in M.N.Rothbard and Walter Block (eds) The Review of Austrian Economics, vol. 2, Lexington,MA: Lexington Books.Callander, C. (1999) ‘Reducing Thermodynamics to Statistical Mechanics: The Case ofEntropy’, The Journal of Philosophy, 96, 348–73.Carlyle, T. (1841) On Heroes, Hero-Worship and the Heroic in History, pp. 239–467 in SartorResartus, London: J.M. Dent, 1975.—— (1843) Past and Present, New York: New York University Press, 1965.Carnap, R. (1928) The Logical Structure of the World, Los Angeles: University of CaliforniaPress, 1969.—— (1931) ‘Die physikalische Sprache als Universalsprache der Wissenschaft’, Erkenntnis,2, 432–65.—— (1932/3) ‘Psychology in Physical Language’, pp. 165–98 in A.J. Ayer (ed.) LogicalPositivism, New York: The Free Press, 1959.—— (1970) Meaning and Necessity, Chicago: University of Chicago Press.Carr, D. (1977) ‘Husserl’s Problematic Concept of Life-World’, pp. 202–12 in F.A. Ellistonand P. McCormick (eds) Husserl. Expositions and Appraisals, Notre Dame: University ofNotre Dame Press.Carroll, J. (1973) ‘Pareto’s Irrationalism’, Sociology, 7, 327–40.Chalmers, A.F. (1985) ‘Methodological Individualism: An Incongruity in Popper’s Philosophy’,pp. 73–87 in G. Currie and A. Musgrave (eds) Popper and the Human Sciences,Dordrecht: Martinus Nijhoff.Charles, D. and Lennon, K. (eds) (1992) Reduction, Explanation and Realism, Oxford:Clarendon.Charon, J.M. (1979) Symbolic Interactionism. An Introduction, an Interpretation, an Integration,Englewood Cliffs, NJ: Prentice-Hall.Cheung, S.N.S. (1969) ‘Transaction Costs, Risk Aversion, and the Choice of ContractualArrangements’, The Journal of Law and Economics, 12, 23–42.—— (1970) ‘The Structure of a Contract and the Theory of a Non-Exclusive Resource’,The Journal of Law and Economics, 13, 49–70.—— (1983) ‘The Contractual Nature of the Firm’, The Journal of Law and Economics, 26,1–21.Chiappori, P.-A. (1988) ‘Rational Household Labor Supply’, Econometrica, 56, 63–89.

Bibliography 395—— (1992) ‘Collective Labor Supply and Welfare’, Journal of Political Economy, 100,437–67.Chisholm, R.M. (1976) ‘Brentano’s Descriptive Psychology’, pp. 91–100 in L.L. McAlister,(ed.) The Philosophy of Brentano, London: Duckworth.—— (1986) ‘Brentano on Preference, Desire and Intrinsic Value’, pp. 182–95 in W.Grassl and B. Smith (eds) Austrian Economics. Historical and Philosophical Background,London: Croom Helm.Cicourel, A.V. (1964) Method and Measurement in Sociology, New York: The Free Press.—— (1974) Cognitive Sociology. Language and Meaning in Social Interaction, New York: The FreePress.—— (1981) ‘Notes on the Integration of Micro- and Macro-levels of Analysis’, pp. 51–80in K. Knorr-Cetina and A.V. Cicourel (eds) Toward an Integration of Micro- and Macro-Sociologies, Boston: Routledge & Kegan Paul.Clarke, S. (1982) Marx, Marginalism and Modern Sociology. From Adam Smith to Max Weber,London: Macmillan.Clower, R. (1965) ‘The Keynesian Counterrevolution: A Theoretical Appraisal’, pp.103–25 in F.H. Hahn and F.P.R. Brechling (eds) The Theory of Interest Rates, London:Macmillan.Coase, R.H. (1937) ‘The Nature of the Firm’, Economica, 4, 386–405.—— (1960) ‘The Problem of Social Cost’, The Journal of Law and Economics, 3, 1–44.—— (1974) ‘The Choice of Institutional Framework: A Comment’, The Journal of Lawand Economics, 17, 493–96.—— (1975) ‘Marshall on Method’, pp. 409–16 in J.C. Wood (ed.) Alfred Marshall. CriticalAssessments, vol. 1, London: Croom Helm, 1982.—— (1978) ‘Economics and Contiguous Disciplines’, The Journal of Legal Studies, 7,201–11.—— (1984) ‘The New Institutional Economics’, Journal of Institutional and TheoreticalEconomics, 140, 229–31.—— (1988a) The Firm, the Market, and the Law, Chicago: Chicago University Press.—— (1988b) ‘The Nature of the Firm: Origin, Meaning, Influence’, The Journal of LawEconomics, and Organization, 4, 3–47.—— (1992) ‘The Institutional Structure of Production’, The American Economic Review, 82,713–19.Cohen, P.S. (1963) ‘The Aims and Interests of Sociology’, The British Journal for the Philosophyof Science, 14, 246–61.Colander, D.C. (1993) ‘The Macrofoundations of Micro’, Eastern Economic Journal, 19,447–57.Coleman, J.S. (1964) ‘Collective Decisions’, Sociological Inquiry, 34, 166–81.—— (1965) ‘The Use of Electronic Computers in the Study of Social Organizations’,Archives Européennes de Sociologie, 6, 89–107.—— (1966) ‘Foundations for a Theory of Collective Decisions’, The American Journal ofSociology, 71, 615–27.—— (1969a) ‘The Methods of Sociology’, pp. 86–114 in R. Bierstedt (ed.) A Design forSociology, Philadelphia: The American Academy of Political and Social Science.—— (1969b) ‘Relational Analysis: The Study of Social Organization with SurveyMethods’, pp. 517–28 in A. Etzioni (ed.) A Sociological Reader on Complex Organizations,New York: Holt, Rinehart and Winston.—— (1971) ‘Social Systems’, pp. 69–79 in P.A. Weiss (ed.) Hierarchically Organized Systems inTheory and Practice, New York: Hafner.

396 Bibliography—— (1972) ‘Systems of Social Exchange’, Journal of Mathematical Sociology, 2, 145–63.—— (1973) The Mathematics of Collective Action, London: Heinemann.—— (1974) Power and Structure of Society, New York: W.W. Norton.—— (1975) ‘Social Structure and a Theory of Action’, pp. 76–93 in P.M. Blau (ed.)Approaches to Social Structure, New York: The Free Press.—— (1979) ‘Rational Actors in Macrosociological Analysis’, pp. 75–91 in R. Harrison(ed.) Rational Action. Studies in Philosophy and Social Science, Cambridge: CambridgeUniversity Press.—— (1982) The Asymmetric Society, New York: Syracuse University Press.—— (1984) ‘Introducing Social Structure into Economic Analysis’, American EconomicReview, 74, 84–88.—— (1986a) ‘Micro Foundations and Macrosocial Theory’, pp. 345–88 in S. Lindenberg,J.S. Coleman and S. Nowak (eds) Approaches to Social Theory, New York: Russell Sage.—— (1986b) ‘Social Structure and the Emergence of Norms Among Rational Actors’,pp. 55–83 in A. Diekmann and P. Mitter (eds) Paradoxical Effects of Social Behavior. Essaysin Honor of Anatol Rapoport, Heidelberg: Physica.—— (1986c) ‘Social Theory, Social Research, and a Theory of Action’, American Journal ofSociology, 91, 1309–35.—— (1987a) ‘Microfoundations and Macrosocial Behavior’, pp. 153–73 in J. Alexander,et al. (eds) The Micro-Macro Link, Berkeley: University of California Press.—— (1987b) ‘Norms as Social Capital’, pp. 133–53 in G. Radnitzky and P. Bernholz (eds)Economic Imperialism. The Economic Method Applied Outside the Field of Economics, New York:Paragon House.—— (1987c) ‘Psychological Structure and Social Structure in Economic Models’, pp.181–5 in R.M. Hogarth and M.W. Reder (eds) Rational Choice. The Contrast betweenEconomics and Psychology, Chicago: Chicago University Press.—— (1988) ‘Actors and Actions in Social History and Social Theory: Reply to Sewell’,American Journal of Sociology, 93, 172–5.—— (1990a) ‘Commentary: Social Institutions and Social Theory’, American SociologicalReview, 55, 333–9.—— (1990b) ‘The Emergence of Norms’, pp. 35–59 in M. Hechter, K.-D. Opp and R.Wippler (eds) Social Institutions. Their Emergence, Maintenance and Effects, New York: Walterde Gruyter.—— (1990c) Foundations of Social Theory, Cambridge, MA: The Belknap Press of theHarvard University Press.—— (1990d) ‘Interview with James S. Coleman’, pp. 47–61 in R. Swedberg (ed.)Economics and Sociology, Princeton, NJ: Princeton University Press.—— (1990e) ‘Norm-Generating Structures’, pp. 250–73 in K.S. Cook and M. Levi (eds)The Limits of Rationality, Chicago: University of Chicago Press.—— (1990f) ‘Rational Organization’, Rationality and Society, 2, 94–105.—— (1991a) ‘Constructed Organization: First Principles’, The Journal of Law, Economics,and Organization, 7, 7–23.—— (1991b) ‘Natural Persons, Corporate Actors, and Constitutions’, Constitutional PoliticalEconomy, 2, 81–106.—— (1992a) ‘The Economic Approach to Sociology’, pp. 133–48 in G. Radnitzky (ed.)Universal Economics. Assessing the Achievements, New York: Paragon House.—— (1992b) ‘The Problematics of Social Theory. Four Reviews of Foundations of SocialTheory’, Theory and Society, 21, 263–83.—— (1992c) ‘The Vision of Foundations of Social Theory’, Analyse und Kritik, 14, 117–28.

Bibliography 397—— (1993a) ‘The Rational Reconstruction of Society’, American Sociological Review, 58,1–15.—— (1993b) ‘Reply to Blau, Tuomela, Diekmann and Baurmann’, Analyse und Kritik, 15,62–69.—— (1994a) ‘A Rational Choice Perspective on Economic Sociology’, pp. 165–80 in N.J.Smelser and R. Swedberg (eds) The Handbook of Economic Sociology, Princeton, NJ:Princeton Univeristy Press.—— (1994b) ‘A Vision of Sociology’, Society, 32, 29–34.Collin, F. (1997) Social Reality, London: Routledge.Collini, S., Winch, D. and Burrow, J. (1983) That Noble Science of Politics. A Study in Nineteenth-CenturyIntellectual History, Cambridge: Cambridge University Press.Collins, R. (1981a) ‘On the Microfoundations of Macro-sociology’, The American Journal ofSociology, 86, 984–1014.—— (1981b) ‘Micro-Translation as a Theory-Building Strategy’, pp. 51–80 in K. Knorr-Cetina and A.V. Cicourel (eds) Advances in Social Theory and Methodology, Boston:Routledge & Kegan Paul.—— (1985) Three Sociological Traditions, New York: Oxford University Press.Comte, A. (1819–26) The Crisis of Industrial Civilization. The Early Essays of Auguste Comte,London: Heinemann, 1974.—— (1836–42) The Positive Philosophy, trans. H. Martineau, 1855; New York: AMS Press,1974.—— (1851–4) Système de Politique Positive, pp. 309–480 in August Comte and Positivism. TheEssential Writings, New York: Harper & Row, 1975.Cook, K.S. (1987) ‘Emerson’s Contribution to Social Exchange Theory’, pp. 209–22 inK.S. Cook (ed.) Social Exchange Theory, Newbury Park: Sage Publications.—— (1991) ‘The Microfoundations of Social Structure: An Exchange Perspective’, pp.29–45 in J. Huber (ed.) Macro-Micro Linkages in Sociology, Newbury Park: Sage Publications.Cook, K.S. and Emerson, R.M. (1978) ‘Power, Equity and Commitment in ExchangeNetworks’, American Sociological Review, 43, 721–39.Cook, K.S., Emerson, R.M. and Gillmore, M.R. (1983) ‘The Distribution of Power inExchange Networks: Theory and Experimental Results’, American Journal of Sociology,89, 275–305.Cook, K.S., O’Brien, J. and Kollock, P. (1990) ‘Exchange Theory: A Blueprint for Structureand Process’, pp. 158–81 in G. Ritzer (ed.) Frontiers of Social Theory: The NewSyntheses, New York: Columbia University Press.Cooley, C.H. (1902) Human Nature and Social Order, New York: Schocken Books, 1964.Coulon, A. (1995) Ethnomethodology, Thousand Oaks: Sage Publications.Cowen, T. (1998) ‘Do Economists Use Social Mechanisms to Explain?’, pp. 125–46 in P.Hedström and R. Swedberg (eds) Social Mechanisms. An Analytical Approach to SocialTheory, Cambridge: Cambridge University Press.Craib, I. (1976) Existentialism and Sociology: A Study of Jean-Paul Sartre, Cambridge:Cambridge University Press.Cubeddu, R. (1993) The Philosophy of the Austrian School, London: Routledge.Currie, G. (1984) ‘Individualism and Global Supervenience’, The British Journal for thePhilosophy of Science, 35, 345–58.Cussins, A. (1987) ‘Varieties of Psychologism’, Synthese, 70, 123–54.D’Agostino, F.B. (1979) Individualism and Collectivism: The Case of Language’, ThePhilosophy of the Social Sciences, 9, 27–47.

398 BibliographyDanto, A. (1965a) Analytical Philosophy of History, Cambridge: Cambridge University Press.—— (1965b) ‘Basic Actions’, pp. 43–58 in A.R. White (ed.) The Philosophy of Action,Oxford: Oxford University Press.Darden, L. and Maull, N. (1977) ‘Interfield Theories’, Philosophy of Science, 44, 43–64.Davis, J. (1992) Exchange, Buckingham: Open University Press.Davis, K. and Moore, W.E. (1945) ‘Some Principles of Stratification’, American SociologicalReview, 10, 242–9.Davis, L. and North, D.C. (1970) ‘Institutional Change and American Economic Growth:A First Step Towards a Theory of Institutional Innovation’, The Journal of EconomicHistory, 30, 131–49.de Boer, T. (1976) ‘The Descriptive Method of Franz Brentano, It’s Two Functions andTheir Significance For Phenomenology’ pp. 101–7 in McAlister, (ed.) The Philosophy ofBrentano, London: Duckworth.Debreu, G. (1959) Theory of Value. An Axiomatic Analysis of Equilibrium Economics, New York:John Wiley & Sons.Demsetz, H. (1964) ‘The Exchange and Enforcement of Property Rights’, The Journal ofLaw and Economics, 3, 11–26.—— (1966) ‘Some Aspects of Property Rights’, The Journal of Law and Economics, 9,61–70.—— (1967) ‘Toward a Theory of Property Rights’, American Economic Review, 57, 347–73.—— (1969) ‘Information and Efficiency: Another Viewpoint’, The Journal of Law andEconomics, 12, 1–22.—— (1983) ‘The Structure of Ownership and the Theory of the Firm’, The Journal ofLaw and Economics, 26, 375–90.—— (1996) ‘Rationality, Evolution, and Acquisitiveness’, Economic Inquiry, 34, 484–95.Denzin, N.K. (1969) ‘Symbolic Interactionism and Ethnomethodology: A ProposedSynthesis’, American Sociological Review, 34, 922–34.Derrida, J. (1964) “‘Genesis and Structure” and Phenomenology’, pp. 154–68 in Writingand Difference, London: Routledge & Kegan Paul, 1978.Descartes, R. (1985), The Philosophical Writings of Descartes, Cambridge: Cambridge UniversityPress.Desrosières, A. (1991) ‘Social Science, Statistics and the State’, pp. 195–218 in P. Wagner,B. Wittrock and R. Whitley (eds) Discources on Society. The Shaping of the Social ScienceDisciplines, Dordrecht: Kluwer Academic Press.Dewey, J. (1896) ‘The Reflex Arc Concept in Psychology’, Psychological Review, 3, 357–70.Dietzel, H. (1882) Ueber das Verhältnis der Volkswirtschaftslehre zur Socialwirtschaftslehre, Berlin:G. Bernstein.—— (1883) ‘Der Ausgangspunkt der Socialwirtschaftslehre und ihr Grundbegriff ’,Zeitschrift für die Gesamte Staatswissenschaft, 39, 1–80.—— (1884) ‘Beiträge zur Methodik der Wirtschaftswissenschaft’, Jahrbücher für Nationalökonomieund Statistik, Neue Folge, 9, 17–44, 193–259.—— (1895) Theoretische Socialökonomik, Leipzig: C.F. Winter‘sche Verlagshandlung.Dilthey, W. (1883a) Introduction to the Human Sciences, Selected Works, vol. 1, Princeton, NJ:Princeton University Press, 1989.—— (1883b) Introduction to the Human Sciences. An Attempt to Lay the Foundation for the Study ofSociety and History, Detroit: Wayne State University Press, 1988.—— (1894) ‘Ideas Concerning a Descriptive and Analytic Psychology’, pp. 23–120 inDescriptive Psychology and Historical Understanding, The Hague: Martinus Nijhoff, 1977.

Bibliography 399—— (1900) ‘The Development of Hermeneutics’, pp. 246–63 in Selected Writings,Cambridge: Cambridge University Press, 1976.—— (1926) Der Aufbau der Geschichtlichen Welt in den Geisteswissenschaften. Gesammelte Schriftenvol. 7, Stuttgart: B.G. Teubner, 1958.—— (1927) ‘The Understanding of Other Persons and Their Life-Expressions’, pp.152–64 in K. Mueller-Vollmer (ed.) The Hermeneutics Reader, Oxford: Basil Blackwell,1986.Dodd, S.C. (1939) ‘A System of Operationally Defined Concepts for Sociology’, AmericanSociological Review, 4, 619–34.—— (1942) Dimensions of Society, New York: Macmillan.Doise, W. (1978) Groups and Individuals. Explanations in Social Psychology, Cambridge:Cambridge University Press.Dolan, E.G. (ed.) (1976) The Foundations of Modern Austrian Economics, Kansas City: Sheedand Ward.Donisthorpe, W. (1889) Individualism. A System of Politics, London: Macmillan.Dorman, P. (1991) ‘Marxism, Methodological Individualism and the New InstitutionalEconomics: Further Considerations’, The Journal of Institutional and Theoretical Economics,147, 364–73.Douglas, J.D. (1971) ‘Understanding Everyday Life’, pp. 1–44 in J.D. Douglas (ed.) UnderstandingEveryday Life. Toward the Reconstruction of Sociological Knowledge, London:Routledge & Kegan Paul.Douglas, J.D. and Johnson, J.M. (eds) (1977) Existential Sociology, Cambridge: CambridgeUniversity Press.Downs, A. (1957) An Economic Theory of Democracy, New York: Harper and Brothers.Dray, W.H. (1972) ‘Holism and Individualism in History and Social Science’, pp. 53–8 inP. Edwards (ed.) The Encyclopedia of Philosophy, New York: Macmillan.Droysen, J.G. (1858) Historik. Vorlesungen über Enzyklopädie und Methodologie der Geschichte,Munich: R. Oldenburg Verlag, 1977.Dugger, W. (1979) ‘Methodological Differences between Institutional and NeoclassicalEconomics’, Journal of Economic Issues, 13, 899–909.Durkheim, E. (1887) Ethics and Sociology of Morals, Buffalo, NY: Prometheus Books, 1993.—— (1888) ‘Course in Sociology: Opening Lecture’, pp. 43–70 in On Institutional Analysis,Chicago: Chicago University Press, 1978.—— (1893) The Division of Labor in Society, New York: The Free Press, 1964.—— (1895) The Rules of Sociological Method and Selected Texts on Sociology and its Method,London: Macmillan, 1982.—— (1897) Suicide, New York: The Free Press, 1951.—— (1898) ‘Individual and Collective Representations’, pp. 1–34 in Sociology and Philosophy,New York: The Free Press, 1974.—— (1900) ‘Sociology in France in the Nineteenth Century’, pp. 3–22, in On Morality andSociety, Chicago: Chicago University Press, 1973.—— (1909) ‘Sociology and the Social Sciences’, pp. 71–90 in On Institutional Analysis,Chicago: Chicago University Press, 1978.—— (1911) ‘Value Judgements and Judgements of Reality’, pp. 80–97 in Sociology andPhilosophy, New York: The Free Press, 1974.Ebbinghaus, H. (1896) ‘Über erklärende und beschreibende Psychologie’, Zeitschrift fürPsychologie und Physiologie der Sinnesorgane, 9, 161–205.Eden, van den P.G. and Hüttner, H.J.M. (1982) ‘Multi-Level Research’, Current Sociology,30, London: Sage Publications.

400 BibliographyEdgeworth, F.Y. (1881) Mathematical Physics. An Essay on the Application of Mathematics to theMoral Sciences, New York: Augustus M. Kelley, 1967.Eggertsson, T. (1990) Economic Behavior and Institutions, Cambridge: Cambridge UniversityPress.Eliaeson, S. (1990) ‘Max Weber and His Critics. Critical Theory’s Reception of Neo-Kantian Methodology’, International Journal of Politics, Culture and Society, 3, 513–37.Elias, N. (1939) The Civilizing Process, vol. 1, The History of Manners, New York: PantheonBooks, 1968.Ellwood, C.A. (1899) ‘Prolegomena to Social Psychology, I–IV’, The American Journal ofSociology, 4–5, 656–65, 807–22, 98–109, 220–7.—— (1925) The Psychology of Human Society, New York: D. Appleton.Elster, J. (1978) Logic and Society: Contradictions and Possible Worlds, London: John Wiley andSons.—— (1979a) ‘Anomalies of Rationality. Some Unresolved Problems in the Theory ofRational Behaviour’, pp. 65–85 in L. Lévy-Garboa (ed.) Sociological Economics, London:Sage Publications.—— (1979b) Ulysses and the Sirens: Studies in Rationality and Irrationality, Cambridge:Cambridge University Press.—— (1980) ‘Reply to Comments’, Inquiry, 23, 213–32.—— (1982) ‘Marxism, Functionalism and Game Theory: The Case for MethodologicalIndividualism’, Theory and Society, 11, 453–82.—— (1983a) Explaining Technical Change, Cambridge: Cambridge University Press.—— (1983b) ‘Reply to Comments’, Theory and Society, 12, 11–20.—— (1983c) Sour Grapes: Studies in the Subversion of Rationality, Cambridge: CambridgeUniversity Press.—— (1985) Making Sense of Marx, Cambridge: Cambridge University Press.—— (1986a) ‘Introduction’ (pp. 1–33) to J. Elster (ed.) Rational Choice, Oxford: Basil Blackwell.—— (1986b) ‘Reply to Comments’, Inquiry, 29, 65–77.—— (1989a) The Cement of Society: A Study of Social Order, Cambridge: Cambridge UniversityPress.—— (1989b) Nuts and Bolts for The Social Sciences, Cambridge: Cambridge University Press.—— (1989c) Solomonic Judgements: Studies in the Limitations of Rationality, Cambridge:Cambridge University Press.—— (1990) ‘When Rationality Fails’, pp. 19–51 in K.S. Cook and M. Levi (eds) TheLimits of Rationality, Chicago: University of Chicago Press.—— (1993) Political Psychology, Cambridge: Cambridge University Press.—— (1998) ‘A Plea for Mechanisms’, pp. 32–44 in P. Hedström and R. Swedberg (eds)Social Mechanisms: An Analytical Approach to Social Theory, Cambridge: Cambridge UniversityPress.—— (1999) Alchemies of the Mind: Rationality and the Emotions, Cambridge: CambridgeUniversity Press.—— (2000) ‘Rationality, Economy, and Society’, pp. 21–41 in S. Turner (ed.) TheCambridge Companion to Weber, Cambridge: Cambridge University Press.Emerson, R.M. (1962) ‘Power-Dependence Relations’, American Sociological Review, 27,31–41.—— (1969) ‘Operant Psychology and Exchange Theory’, pp. 379–405 in R.L. Burgessand D. Bushell (eds) Behavioral Sociology. The Experimental Analysis of Social Process, NewYork: Columbia University Press.

Bibliography 401—— (1972a) ‘Exchange Theory, Part I: A Psychological Basis for Social Exchange’, pp.38–57 in J. Berger, M. Zelditch and B. Andersson (eds) Sociological Theories in Progress,vol. 2, Boston: Houghton Mifflin.—— (1972b) ‘Exchange Theory, Part II: Exchange Relations and Network Structures’,pp. 58–87 in J. Berger, M. Zelditch and B. Andersson (eds) Sociological Theories inProgress, vol. 2, Boston: Houghton Mifflin.—— (1987) ‘Toward a Theory of Value in Social Exchange’, pp. 11–46 in K.S. Cook(ed.) Social Exchange Theory, Newbury Park: Sage Publications.Endres, A.M. (1997) Neoclassical Microeconomic Theory: The Founding Austrian Version, London:Routledge.England, P. and Kilbourne, B.S. (1990) ‘Markets, Marriages, and Other Mates: TheProblem of Power’, pp. 163–88 in R. Friedland and A.F. Robertson (eds) Beyond theMarketplace. Rethinking Economy and Society, New York: Aldine de Gruyter.Eriksson, B. (1988) Samhällsvetenskapens Uppkomst, Uppsala: Hallgren och Fallgren.—— (1993) ‘The First Formulation of Sociology. A Discursive Innovation of the 18 thCentury’, Archives Européenes de Sociologie, 34, 251–76.Fabian, R. and Simons, P.M. (1986) ‘The Second Austrian School of Value Theory’, pp.37–101 in W. Grassl and B. Smith (eds) Austrian Economics: Historical and PhilosophicalBackground, London: Croom Helm.Faris, E. (1937/8) ‘The Social Psychology of George Mead’, The American Journal of Sociology,43, 391–403.Farmer, M.K. (1982) ‘Rational Action in Economics and Social Theory: Some Misunderstandings’,Archives Européennes de Sociologie, 23, 179–97.Farr, R.M. (1978) ‘On the Varieties of Social Psychology: An Essay on the Relationshipbetween Psychology and Other Social Sciences’, Social Science Information, 17, 503–25.Feigl, H. (1970) ‘The “Orthodox” View of Theories: Remarks in Defence as Well asCritique’, pp. 3–16 in Minnesota Studies in the Philosophy of Science, vol. IV, Minneapolis:University of Minnesota Press.Feiwel, G.R. (1985) ‘Quo Vadis Macroeconomics? Issues, Tensions and Challenges’, pp.1–100 in F.R. Feiwel (ed.) Issues in Contemporary Macroeconomics and Distribution, London:Macmillan.Ferejohn, J. (1991) ‘Rationality and Interpretation’, pp. 279–305 in K.R. Monroe (ed.)The Economic Approach to Politics. A Critical Reassessment of the Theory of Rational Action, NewYork: HarperCollins.Feuer, L.S. (1976) ‘John Stuart Mill as a Sociologist: The Unwritten Ethology’, pp.86–110 in J.M. Robson (ed.) James and John Stuart Mill. Papers of The Centenary Conference,Toronto: Toronto University Press.Feyerabend, P.K. (1962) ‘Explanation, Reduction, and Empiricism’, pp. 28–97 in H. Feigland G. Maxwell (eds) Minnesota Studies in the Philosophy of Science, vol. III, Minneapolis:Minnesota University Press.Field, J.A. (1979) ‘On the Explanation of Rules Using Rational Choice Models’, Journal ofEconomic Issues, 13, 49–72.—— (1981) ‘The Problem with Neoclassical Institutional Economics: A Critique withSpecial Reference to the North/Thomas Model of Pre-1500 Europe’, Explorations inEconomic History, 18, 174–98.—— (1984) ‘Microeconomics, Norms, and Rationality’, Economic Development and CulturalChange, 32, 683–711.

402 BibliographyFine, G.A. (1990) ‘Symbolic Interactionism in the Post-Blumerian Age’, pp. 117–57 in G.Ritzer (ed.) Frontiers of Social Theory. The New Syntheses, New York: Columbia UniversityPress.Finley, M.I. (1973) The Ancient Economy, London: Chatto and Windus.Fischer, B.M. and Strauss, A.L. (1979) ‘Interactionism’, pp. 457–98 in T. Bottomore andR. Nisbet (eds) A History of Sociological Analysis, London: Heinemann.Fisher, F.M. (1983) Disequilibrium Foundations of Equilibrium Economics, Cambridge:Cambridge University Press.—— (1989) ‘Adjustment Processes and Stability’, pp. 36–42 in J. Eatwell, M. Milgate andP. Newman (eds) General Equilibrium (The New Palgrave), New York: W.W. Norton.Flew, A. (1985) Thinking About Social Thinking: The Philosophy of the Social Sciences, Oxford:Basil Blackwell.Foldes, L.P. (1968) ‘A Note on Individualistic Explanations’, pp. 322–37 in I. Lakatos andA. Musgrave (eds) Problems in Philosophy of Science, Amsterdam: North-Holland.Föllesdal, D. (1972) ‘An Introduction to Phenomenology for Analytic Philosophers’, pp.417–29 in R.E. Olson and A.M. Paul (eds) Contemporary Philosophy in Scandinavia, Baltimore:Johns Hopkins Press.Forget, E.L. (1999) The Social Economics of Jean-Baptiste Say: Markets and Virtue, London:Routledge.Foster, J. (1993) ‘Economics and the Self-Organisation Approach: Alfred Marshall Revisited’,The Economic Journal, 103, 975–91.Foucault, M. (1984) The History of Sexuality, vol. 2 The Use of Pleasure, New York: RandomHouse, 1985.Friedman, M. (1953) Essays in Positive Economics, Chicago: University of Chicago Press.—— (1956) ‘The Quantity Theory of Money’, pp. 3–21 in M. Friedman (ed.) Studies in theQuantitative Theory of Money, Chicago: University of Chicago Press.—— (1962) Capitalism and Freedom, Chicago: Chicago University Press, 1982.—— (1970a) The Counter-Revolution in Monetary Theory, London: Institute of EconomicAffairs.—— (1970b) ‘A Theoretical Framework for Monetary Analysis’, pp. 1–62 in R.J. Gordon(ed.) Milton Friedman’s Monetary Framework. A Debate with His Critics, Chicago: Universityof Chicago Press.Frisby, D. (1992) Simmel and Since: Essays on Simmel’s Social Theory, London: Routledge.Furubotn, E.G. and Richter, R. (1991) ‘The New Institutional Economics: An Assessment’,pp. 1–34 in The New Institutional Economics, Tübingen: J.B.C. Mohr.—— (1997) Institutions and Economic Theory. The Contribution of the New Institutional Economics,Ann Arbor: University of Michigan Press.Galtung, J. (1969) Theory and Methods of Social Research, Oslo: University Press.Garfinkel, H. (1967) Studies in Ethnomethodology, Englewood Cliffs, NJ: Prentice-Hall.—— (1968) ‘The Origins of the Term Ethnomethodology’, pp. 15–18 in R.Turner (ed.)Ethnomethodology, Harmondsworth: Penguin, 1974.—— (1991) ‘Respecification: evidence for locally produced, naturally accountablephenomena of order, logic, reason, meaning, method, etc. in and as of the essentialhaecceity of immortal ordinary society (I) – an announcement of studies’ pp. 10–19 inG. Button (ed.) Ethnomethodology and the Human Sciences, Cambridge: Cambridge UniversityPress, 1991.Garfinkel, H. and Sacks, H. (1970) ‘On Formal Structures of Practical Action’, pp.337–66 in J.C. McKinney and E.A. Tiryakian (eds) Theoretical Perspectives and Developments,New York: Appleton–Century–Crofts.

Bibliography 403Geanakoplos, J. (1989) ‘Arrow-Debreu Model of General Equilibrium’, pp. 43–61 in J.Eatwell, M. Milgate and P. Newman (eds) General Equilibrium (The New Palgrave), NewYork: W.W. Norton.Gellner, E. (1956) ‘Explanations in History’, Aristotelian Society, supplementary vol. 30,157–76.—— (1960) ‘Reply to Mr Watkins’, pp. 514–15 in P. Gardiner (ed.) Theories of History,Glencoe: The Free Press.—— (1975) ‘Ethnomethodology: The Re-Enchantment Industry or the Californian Wayof Subjectivity’, The Philosophy of the Social Sciences, 5, 431–50.Gergen, K.J. (1994) Realities and Relationships. Soundings in Social Construction, Cambridge,MA: Harvard University Press.Gibson, Q. (1968) The Logic of Social Enquiry, London: Routledge & Kegan Paul.Giddens, A. (1977) Studies in Social and Political Theory, London: Hutchinson.—— (1979) Central Problems in Social Theory: Action, Structure and Contradiction in Social Analysis,London: Macmillan.—— (1984) The Constitution of Society, Cambridge: Polity Press.Ginsberg, M. (1928) The Psychology of Society, London: Methuen.—— (1956) Essays in Sociology and Social Philosophy, Harmondsworth: Penguin.Girill, T.R. (1976a) ‘Criteria for the Part-Whole Relation in Micro-Reductions’,Philosophia, 6, 69–79.—— (1976b) ‘Evaluating Micro-Explanations’, Erkenntnis, 10, 387–405.Gisbert, P. (1959) ‘Social Facts in Durkheim’s System’, Anthropos, 54, 354–69.Goffman, E. (1959) The Presentation of Self in Everyday Life, New York: Doubleday.—— (1961) Encounters: Two Studies in the Sociology of Interaction, Indianapolis: Bobbs-Merrill.—— (1967) Interaction Ritual, New York: Anchor Books.—— (1969) Strategic Interaction, Oxford: Basil Blackwell, 1970.—— (1971) Relations in Public: Microstudies of the Public Order, Harmondsworth: Penguin.—— (1974) Frame Analysis, New York: Harper & Row.—— (1983) ‘Felicity’s Condition’, American Journal of Sociology, 89, 1–53.Goldstein, L.J. (1956) ‘The Inadequacy of the Principle of Methodological Individualism’,The Journal of Philosophy, 53, 801–13.—— (1958) ‘The Two Theses of Methodological Individualism’, The British Journal for thePhilosophy of Science, 9, 1–11.—— (1959) ‘Mr Watkins on the Two Theses’, The British Journal for the Philosophy of Science,10, 240–1.Goldthorpe, J. (1996) ‘The Quantitative Analysis of Large-Scale Data-Sets and RationalAction Theory: For a Sociological Alliance’, European Sociological Review, 12, 109–26.—— (1998) ‘Rational Action Theory for Sociology’, The British Journal of Sociology, 49,167–92.—— (2000) On Sociology: Numbers, Narratives and the Integration of Research and Theory, Oxford:Oxford University Press.Gordon, C. and Gergen, K.J. (1968) ‘The Nature and Dimensions of the Self ’, pp. 33–9in C. Gordon and K.J. Gergen (eds) The Self in Social Interaction, New York: John Wileyand Sons.Gordon, D. (1993) The Philosophical Origins of Austrian Economics, Auburn, Alabama: Ludwigvon Mises Institute.Grafstein R. (1991) ‘Rational Choice: Theory and Institutions’, pp. 259–78 in K.R.Monroe (ed.) The Economic Approach to Politics. A Critical Reassessment of the Theory ofRational Action, New York: HarperCollins.

404 BibliographyGranovetter, M. (1973) ‘The Strength of Weak Ties’, American Journal of Sociology, 78,1360–80.—— (1978) ‘Threshold Models of Collective Behavior’, American Journal of Sociology, 83,1420–43.—— (1985) ‘Economic Action and Social Structure: The Problem of Embeddedness’,American Journal of Sociology, 91, 481–510.—— (1993) ‘The Nature of Economic Relationships’, pp. 3–41 in R. Swedberg (ed.)Explorations in Economic Sociology, New York: Russell Sage Foundation.Grassl, W. and Smith, B. (eds) (1986) Austrian Economics: Historical and Philosophical Background,London: Croom Helm.Gray, T.S. (1985) ‘Herbert Spencer: Individualist or Organicist?’, Political Studies, 33,236–253.—— (1996) The Political Philosophy of Herbert Spencer, Aldershot: Avebury.Greaves, B.B. (1996) ‘Introduction’ (pp. 1–10) to B.B. Greaves (ed.) Austrian Economics. AnAnthology, Irvington-on-Hudson, NY: Foundation for Economic Education.Grossman, S.J. and Hart, O.D. (1986) ‘The Costs and Benefits of Ownership’: A Theoryof Vertical and Lateral Integration’, The Journal of Political Economy, 94, 691–719.Gurwitch, A. (1962) ‘The Common-Sense World as Social Reality’, Social Research, 29,50–72.—— (1967) ‘Galilean Physics in the Light of Husserl’s Phenomenology’, pp. 71–89 in T.Luckmann (ed.) Phenomenology and Sociology, Harmondsworth: Penguin, 1978.Haberler, G. (1961) ‘Mises’ Private Seminar’, pp. 190–192 in Ludwig von Mises, Planningfor Freedom, South Holland, IL: Libertarian Press, 1974.Habermas, J. (1968) Knowledge and Human Interests, Boston: Beacon Press, 1971.—— (1970) On the Logic of the Social, Sciences, Cambridge: Polity Press, 1988.—— (1981) The Theory of Communicative Action vol. 1 Reason and the Rationalization of Society,Boston: Beacon Press, 1984.—— (1981) The Theory of Communicative Action vol. 2 Lifeworld and System: A Critique of FunctionalistReason, Boston: Beacon Press, 1987.Hacking, I.C. (1990) The Taming of Chance, Cambridge: Cambridge University Press.—— (1999) The Social Construction of What? Cambridge, MA: Harvard University Press.Hahn, F.H. (1977) ‘Keynesian Economics and General Equilibrium Theory: Reflectionson Some Current Debates’, pp. 25–40 in G.C. Harcourt (ed.) The Microeconomic Foundationsof Macroeconomics, London: Macmillan.—— (1981) ‘General Equilibrium Theory’, pp. 123–138 in D. Bell and I. Kristol (eds) TheCrisis in Economic Theory, New York: Basic Books.Haines, V.A. (1988) ‘Social Network Analysis, Structuration Theory and the Holism-Individualism Debate’, Social Networks, 10, 157–82.Halfpenny, P. (1982) Positivism and Sociology: Explaining Social Life, London: George Allen &Unwin.Hall, R.T. (1993) ‘Introduction’ (pp. 11–53) to E. Durkheim, Ethics and the Sociology ofMorals, Buffalo, NY: Prometheus Books.Hamblin, R.L. and Kunkel, J.H. (eds) (1977) Behavioral Theory in Sociology. Essays in Honor ofGeorge C. Homans, New Brunswick, NJ: Transaction Books.Hardin, R. (1997) ‘Economic Theories of the State’, pp. 21–34 in Dennis C. Mueller,(ed.) Perspectives on Public Choice. A Handbook, Cambridge: Cambridge University Press.Hargreaves Heap, S.P. and Varoufakis, Y. (1995) Game Theory. A Critical Introduction,London: Routledge.Harré, R. (1972) The Philosophies of Science, Oxford: Oxford University Press.

Bibliography 405—— (1979) Social Being: A Theory of Social Psychology, Oxford: Basil Blackwell.—— (1986) ‘Social Sources of Mental Content and Order’, pp. 91–127 in J. Margolis,P.T. Manicas, R. Harré and P.F. Secord, Psychology. Designing a Discipline, Oxford: BasilBlackwell.—— (1993) Social Being, Oxford: Blackwell, 2nd revised edn.Harré, R., Clarke, D. and De Carlo, N. (1985) Motives and Mechanisms: An Introduction to thePsychology of Action, London: Methuen.Harsanyi, J.C. (1962) ‘Models for the Balance of Power in Society’, pp. 442–62 in E.Nagel, P. Suppes and A. Tarski (eds) Logic, Methodology and Philosophy of Science, Stanford:Stanford University Press.—— (1966a) ‘A Bargaining Model for Social Status’, Behavioral Science, 11, 357–69.—— (1966b) ‘A General Theory of Rational Behavior in Game Situations’, Econometrica,34, 613–34.—— (1968) ‘Individualistic and Functionalistic Explanation in the Light of GameTheory: The Example of Social Status’, pp. 305–21 in I. Lakatos and A. Musgrave(eds) Problems in the Philosophy of Science, Amsterdam: North-Holland.—— (1969) ‘Rational-Choice Models of Political Behavior versus Functionalist andComformist Theories’, World Politics, 21, 513–38.—— (1977) Rational Behavior and Bargaining Equilibrium in Games and Social Situations,Cambridge: Cambridge University Press.Hart, O. (1989) ‘An Economist’s Perspective on the Theory of the Firm’, Columbia LawReview, 89, 1757–74.Hart, O. and Moore, J. (1990) ‘Property Rights and the Nature of the Firm’, The Journal ofPolitical Economy, 98, 1119–58.Hartley, J.E. (1997) The Representative Agent in Macroeconomics, London: Routledge.Hausman, D.M. (1992) The Inexact and Separate Science of Economics, Cambridge: CambridgeUniversity Press.Hayek, F.A. (1931) Prices and Production, London: George Routledge and Sons.—— (1933a) Monetary Theory and the Trade Cycle, London: Jonathan Cape.—— (1933b) ‘The Trend of Economic Thinking’, Economica, 13, 121–37.—— (1934) ‘Carl Menger’, Economica, New Series, 1, 393–420.—— (1941) The Pure Theory of Capital, London: Macmillan.—— (1942–4) ‘Scientism and the Study of Society, I–III’, Economica, 9–11, 267–91,34–63, 27–39.—— (1944) The Road to Serfdom, London: Routledge & Kegan Paul.—— (1948) Individualism and Economic Order, Chicago: Henry Regnery, 1972.—— (1952) The Sensory Order, Chicago: Chicago University Press, 1963.—— (1955) The Counter-Revolution of Science. Studies on the Use and Abuse of Reason, New York:The Free Press of Glencoe.—— (1967) Studies in Philosophy, Politics and Economics, New York: Simon and Schuster.—— (1973) ‘The Place of Menger’s Grundsätze in the History of Economic Thought’,pp. 1–14 in J.R. Hicks and W. Weber (eds) Carl Menger and the Austrian School ofEconomics, Oxford: Clarendon Press.—— (1978) New Studies in Philosophy, Politics, Economics and the History of Ideas, London:Routledge & Kegan Paul.—— (1982) Law, Legislation and Liberty: A New Statement of the Liberal Principles of Justice andPolitical Economy, London: Routledge.—— (1988) The Fatal Conceit: The Errors of Socialism, London: Routledge.

406 Bibliography—— (1995) Contra Keynes and Cambridge: Essays, Correspondence, vol. 9 The Collected Works ofF.A. Hayek, ed. B. Caldwell, London: Routledge.Hazlitt, H. (ed.) (1960) The Critics of Keynesian Economics, Princeton, NJ: D. Van NostrandCompany.Hechter, M. (1983) ‘Introduction’ (pp. 3–15) to M. Hechter (ed.) The Microfoundations ofMacrosociology, Philadelphia: Temple University Press.—— (1989) ‘Rational Choice Foundations of Social Order’, pp. 60–81 in J.C. Alexanderand J.H. Turner (eds) Theory Building in Sociology, Newbury Park: Sage Publications.Heckathorn, D.D. (1997) ‘Overview: The Paradoxical Relationship Between Sociologyand Rational Choice’, The American Sociologist, 28, 6–15.Hedström, P. and Swedberg, R. (1996a) ‘Social Mechanisms’, Acta Sociologica, 39,281–308.—— (1996b) ‘Rational Choice, Empirical Research, and the Sociological Tradition’,European Sociological Review, 12, 127–46.—— (1998) ‘Introduction’, pp. 1–31 in P. Hedström and R. Swedberg (eds) Social Mechanisms.An Analytical Approach to Social Theory, Cambridge: Cambridge University Press.Hedström, P., Swedberg, R. and Udehn, L. (1998) ‘Popper’s Situational Analysis andContemporary Sociology’, Philosophy of the Social Sciences, 28, 339–64.Heine, W. (1983) Methodologischer Individualismus. Zur Geschichtsphilosophischen Begründung einessozialwissenschaftlichen Konzeptes, Würzburg: Königshausen & Neumann.Hegel, G.F.W. (1821) Philosophy of Right, Oxford: Oxford University Press, 1967.—— (1837) Reason in History, Indianapolis: The Bobbs-Merril Company, 1953.Heijdra, B.J., Lowenberg, A.D. and Mallick, R.J. (1988) ‘Marxism, Methodological Individualism,and the New Institutional Economics’, The Journal of Institutional andTheoretical Economics, 144, 296–317.—— (1991) ‘Marxism, Methodological Individualism, and the New InstitutionalEconomics: Reply to Dorman’, The Journal of Institutional and Theoretical Economics, 147,375–78.Hempel, G. (1965) Aspects of Scientific Explanation And Other Essays in the Philosophy of Science,New York: The Free Press.—— (1969a) ‘Logical Positivism and the Social Sciences’, pp. 163–94 in P. Achinstein andS.F. Barker (eds) The Legacy of Logical Positivism, Baltimore: Johns Hopkins Press.—— (1969b) ‘Reduction: Ontological and Linguistic Facets’, pp. 179–99 in S.Morgenbesser, P. Suppes and M. White (eds) Philosophy, Science and Method. Essays inHonor of Ernest Nagel, New York: Macmillan.—— (1970) ‘On the “Standard Conception” of Scientific Theories’, pp. 142–63 inMinnesota Studies in the Philosophy of Science, vol. IV, Minneapolis: University ofMinnesota Press.Hennis, W. (1988) Max Weber: Essays in Reconstruction, London: Allen & Unwin.—— (1991) ‘The Pitiless “Sobriety of Judgement”: Max Weber Between Carl Mengerand Gustav von Schmoller – the Academic Politics of Value Freedom’, History of theHuman Sciences, 4, 27–59.Herder, J.G. (1772) ‘Essay on the Origin of Language’, pp. 87–166 in J.J. Rousseau andJ.G. Herder, On the Origin of Language, Chicago: Chicago University Press, 1966.—— (1774) Auch eine Philosophie der Geschichte der Menschheit zur Bildung der Menschheit, Frankfurtam Main: Suhrkamp, 1967.—— (1784–91) Reflections on the Philosophy of the History of Mankind, Chicago: ChicagoUniversity Press, 1968.

Bibliography 407Heritage, J. (1987) ‘Ethnomethodology’, pp. 224–72 in A. Giddens and J. Turner (eds)Social Theory Today, Cambridge: Polity Press.Hicks, J. (1937) ‘Mr. Keynes and the “Classics”: A Suggested Interpretation’, Econometrica,5, 147–59.—— (1976) ‘Time in Economics’, pp. 282–300 in J. Hicks Money, Interest and Wages,Collected Essays on Economic Theory, vol. II, Oxford: Basil Blackwell, 1982.—— (1980) ‘IS-LM – An Explanation’, pp. 318–331 in J. Hicks Money, Interest and Wages,Collected Essays on Economic Theory, vol. II, Oxford: Basil Blackwell, 1982.Hicks, J. and Allen, R.G.D. (1934) ‘A Reconsideration of the Theory of Value’, Economica,New Series, 1, 52–76, 196–219.Himmelweit, S. (1977) ‘The Individual as Basic Unit of Analysis’, pp. 21–35 in F. Greenand P. Nore (eds) Economics: An Anti-Text, London: Macmillan.Hinchman, L.P. (1990) ‘The Idea of Individuality: Origins, Meaning, and Political Significance’,Journal of Politics, 52, 759–81.Hindess, B. (1973a) ‘The Phenomenological Sociology of Alfred Schutz’, Economy andSociety, 2, 1–27.—— (1973b) The Use of Official Statistics in Sociology. A Critique of Positivism and Ethnomethodology,London: Macmillan.—— (1977) Philosophy and Methodology in the Social Sciences, Hassocks, Sussex: HarvesterPress.—— (1984) ‘Rational Choice Theory and the Analysis of Political Action’, Economy andSociety, 13, 255–77.—— (1986) ‘Actors and Social Relations’, pp. 113–26 in M.L. Wardell and S.P. Turner(eds) Sociological Theory in Transition, Boston: Allen & Unwin.—— (1988) Choice, Rationality and Social Theory, London: Unwin Hyman.Hobbes, T. (1651) Leviathan, Harmondsworth: Penguin, 1968.Hobhouse, (1918) The Metaphysical Theory of the State: A Criticism, London: George Allen &Unwin.Hodgson, G.M. (1988) Economics and Institutions, Cambridge: Polity Press.—— (1989) ‘Institutional Economic Theory: The Old Versus the New’, Review of PoliticalEconomy, 249–69.—— (1991) ‘Hayek’s Theory of Cultural Evolution: An Evaluation in the Light ofVanberg’s Critique’, Economics and Philosophy, 7, 67–82.—— (1993) ‘Evolution and Institutional Change. On the Nature of Selection in Biologyand Economics’, pp. 222–41 in U. Mäki, B. Gustafsson and C. Knudsen (eds) Rationality,Institutions and Economic Methodology, London: Routledge.—— (1994) ‘Hayek, Evolution and Spontaneous Order’, pp. 408–47 in P. Mirowski (ed.)Natural Images in Economic Thought, Cambridge: Cambridge University Press.Hollander, S. (1985) The Economics of John Stuart Mill, vol. 1: Theory and Method, Oxford:Basil Blackwell.Hollis, M. (1972) ‘J.S. Mill’s Political Philosophy of Mind’, pp. 374–86 in J.C. Wood (ed.)John Stuart Mill. Critical Assessments, vol. 1, London: Routledge.—— (1987) The Cunning of Reason, Cambridge: Cambridge University Press.—— (1994) The Philosophy of Social Science. An Introduction, Cambridge: Cambridge UniversityPress.Hollis, M. and Nell, E.J. (1975) Rational Economic Man, Cambridge: Cambridge UniversityPress.Holmström, B. and Milgrom, P. (1994) ‘The Firm as an Incentive System’, The AmericanEconomic Review, 84, 972–91.

408 BibliographyHolmström, B. and Roberts, J. (1998) ‘The Boundaries of the Firm Revisited’, The Journalof Economic Perspectives, 12, 73–94.Holmström, B. and Tirole, J. (1989) ‘The Theory of the Firm’, ch. 2, pp. 61–133 in R.Schmalensee and R.D. Willig (eds) Handbook of Industrial Organization, Amsterdam:North-Holland.Holton, R.J. and Turner, B.S. (1989) Max Weber on Economy and Society, London: Routledge.Homans, G.C. (1951) The Human Group, London: Routledge & Kegan Paul.—— (1961) Social Behaviour, London: Routledge & Kegan Paul.—— (1962) Sentiments and Activities, London: Routledge & Kegan Paul.—— (1964a) ‘Bringing Men Back In’, American Sociological Review, 29, 809–18.—— (1964b) ‘Commentary’, Sociological Inquiry, 34, 221–31.—— (1964c) ‘Contemporary Theory in Sociology’, pp. 951–77 in R.E.L. Faris (ed.)Handbook of Modern Sociology, Chicago: Rand-McNally.—— (1964d) ‘A Theory of Social Interaction’, pp. 113–31 in Transactions of the Fifth WorldCongress of Sociology, Louvain: International Association of Sociology.—— (1967a) ‘Fundamental Social Processes’, pp. 27–78 in N.J. Smelser (ed.) Sociology: AnIntroduction, New York: Wiley & Sons.—— (1967b) The Nature of Social Science, New York: Harcourt, Brace and World.—— (1968) ‘A Life of Synthesis’, American Behavioral Scientist, vol. 12, 2–8.—— (1969) ‘The Sociological Relevance of Behaviorism’, pp. 1–24 in R.L. Burgess andD. Bushell (eds) Behavioral Sociology, New York: Columbia University Press.—— (1970a) ‘The Relevance of Psychology to the Explanation of Social Phenomena’,pp. 313–29 in R. Borger and F. Cioffi (eds) Explanation in the Behavioural Sciences,Cambridge: Cambridge University Press.—— (1970b) ‘Reply’ to Blau, pp. 340–3 in R. Borger and F. Cioffi (eds) Explanation in theBehavioural Sciences, Cambridge: Cambridge University Press.—— (1974) Social Behavior. Its Elementary Forms, New York: Harcourt Brace Jovanovich,revised edn.—— (1975) ‘What Do We Mean by Social “Structure”’, pp. 53–75 in P.M. Blau, (ed.)Approaches to the Study of Social Structure, New York: The Free Press.—— (1980) ‘Discovery and the Discovered in Social Theory’, pp. 17–22 in H.M. Blalock(ed.) Sociological Theory and Research, New York: The Free Press.—— (1983) ‘Steps to a Theory of Social Behavior’, Theory and Society, 12, 1–45.—— (1984) Coming to My Senses. The Autobiography of a Sociologist, New Brunswick, NJ:Transaction Books.Huber, J. (1973) ‘Symbolic Interaction as a Pragmatic Perspective: The Bias of EmergentTheory’, American Sociological Review, 38, 274–84.Huff, T.E. (1984) Max Weber and the Methodology of the Social Sciences, New Brunswick, NJ:Transaction Books.Hughes, J. (1981) The Philosophy of Social Research, London: Longman.Humboldt, W. von (1791–92) The Limits of State Action, Cambridge: Cambridge UniversityPress, 1969.—— ([1821] 1967) ‘On the Historian’s Task’, History and Theory, 6, 57–71.Hume, D. (1741/2) Essays. Moral, Political and Literary, Oxford: Oxford University Press,1963.Hummell, H.J. (1973a) ‘Methodologischer Individualismus, Struktureffekte undSystemkonsequensen’, pp. 61–134 in K.-D. Opp and H.J. Hummell, Probleme derErklärung sozialer Prozesse vol. II: Soziales Verhalten und Soziale Systeme, Frankfurt:Athenäum Verlag.

Bibliography 409—— (1973b) ‘Für eine Struktursoziologie auf individualistischer Grundlage’, pp. 135–83in K.-D. Opp and H.J. Hummell, Probleme der Erklärung sozialer Prozesse vol. II: SozialesVerhalten und Soziale Systeme, Frankfurt: Athenäum.Hummell, H.J. and Opp, K.-D. (1968) ‘Sociology Without Sociology’, Inquiry, 11, 205–26.—— (1971) Die Reduzierbarkeit von Soziologie auf Psychologie, Braunschweig: Friedrich Vieweg& Sohn.Hund, J. (1982) ‘Are Social Facts Real?’, The British Journal of Sociology, 33, 270–78.Husserl, E. (1900) Logical Investigations, London: Routledge & Kegan Paul, 1970.—— (1913) Ideas Pertaining to a Pure Phenomenology and to a Phenomenological Philosophy (FirstBook), The Hague: Martin Nijhoff, 1983.—— (1929) Cartesian Meditations. An Introduction to Phenomenology, The Hague: MartinusNijhoff, 1973.—— (1950) The Idea of Phenomenology, The Hague: Martinus Nijhoff, 1973.—— (1954) The Crisis of European Sciences and Transcendental Phenomenology, Evanston: NorthwesternUniversity Press, 1970.—— (1965) Phenomenology and the Crisis of Philosophy, New York: Harper & Row.—— (1977) Phenomenological Psychology: lectures, summer semester, 1925, The Hague:Nijhoff.—— (1981) Shorter Works, Notre Dame, Indiana: University of Notre Dame Press.Hutchison, T.W. (1973) ‘Some Themes from Investigations into Method ’, pp. 15–37 in J.R.Hicks and W. Weber (eds) Carl Menger and the Austrian School of Economics, Oxford:Clarendon Press.—— (1978) On Revolutions and Progress in Economic Knowledge, Cambridge: CambridgeUniversity Press.—— (1984) ‘Institutionalist Economics Old and New’, The Journal of Institutional and TheoreticalEconomics, 140, 20–29.—— (1988) ‘Gustav Schmoller and the Problems of Today’, Journal of Institutional andTheoretical Economics, 144, 527–31.—— (1994) The Uses and Abuses of Economics, London: Routledge.Hyland, M, and Bridgstock, M. (1974) ‘Reductionism: Comments on Some RecentWork’, The Philosophy of the Social Sciences, 4, 197–200.Iggers, G.G. (1959) ‘Further Remarks About Early Use of the term “Social Science”’, pp.154–7 in J.C. Wood (ed.) John Stuart Mill: Critical Assessments, vol. IV, London: Routledge,1991.—— (1968) The German Conception of History. The National Tradition of Historical Thought fromHerder to the Present, Middletown, Conn.: Wesleyan University Press, 1983.—— (1995) ‘Historicism: The History and Meaning of the Term’, Journal of the History ofIdeas, 56, 129–52.Infantino, L. (1998) Individualism in Modern Thought: From Adam Smith to Hayek, London:Routledge.Ingrao, B. and Israel, G. (1990) The Invisible Hand: Economic Equilibrium in the History ofScience, Boston: MIT Press.Israel, J. (1971) ‘The Principle of Methodological Individualism and Marxian Epistemology’,Acta Sociologica, 14, 145–50.Jackson, F. and Pettit, P. (1992) ‘Structural Explanation in Social Theory’, pp. 97–132 inD.Charles and K. Lennon (eds) Reduction, Explanation, and Realism, Oxford: Clarendon.Jacobs, S. (1990) ‘Popper, Weber and the Rationalist Approach to Social Explanation’,The British Journal of Sociology, 41, 559–70.James, S. (1984) The Content of Social Explanation, Cambridge: Cambridge University Press.

410 BibliographyJames, W. (1890) The Principles of Psychology, New York: Dover, 1950.Janssen, C.W. (1993) Microfoundations: A Critical Inquiry, London: Routledge.Jarvie, I.C. (1959) ‘Comments on Charles Taylor’, Universities and Left Review, 5, 57.—— (1961) ‘Nadel on the Aims and Methods of Social Anthropology’, The British Journalfor the Philosophy of Science, 12, 1–24.—— (1964a) ‘Review of Robert Brown’s Explanation in Social Science’, The BritishJournal for the Philosophy of Science, 15, 62–72.—— (1964b) The Revolution in Anthropology, London: Routledge & Kegan Paul.—— (1972) Concepts and Society, London: Routledge & Kegan Paul.—— (1998) ‘Situational Logic and Its Reception’, The Philosophy of the Social Sciences, 28,365–80.Jaspers, K. (1919) Psychologie der Weltanschauungen, Berlin: Julius Springer, 1922.Jay, E. and Jay, R. (1986) Critics of Capitalism: Victorian Reactions to Political Economy,Cambridge: Cambridge University Press.Jensen, M.C. and Meckling, W.H. (1976) ‘Theory of the Firm: Managerial Behavior,Agency Costs and Ownership Structure’, The Journal of Financial Economics, 3, 305–60.Jevons, W.S. (1871) The Theory of Political Economy, Harmondsworth: Penguin, 1970.Joas, H. (1980) G.H. Mead. A Contemporary Re-examination of his Thought, Cambridge: PolityPress, 1985.—— (1987) ‘Symbolic Interactionism’, pp. 82–115 in A. Giddens and J. Turner (eds)Social Theory Today, Cambridge: Polity Press.Johansson, I. (1975) A Critique of Karl Popper’s Methodology, Stockholm: ScandinavianUniversity Books.Johnson, G.D. and Picou, J.S. (1985) ‘The Foundations of Symbolic InteractionismReconsidered’, pp. 54–70 in H.J. Helle and S.N. Eisenstadt (eds) Perspectives on SociologicalTheory, vol. 2: Micro-Sociological Theory, Beverly Hills: Sage Publications.Jones, B. (1977) ‘Economic Action and Rational Organisation in the Sociology of Weber’,pp. 28–65 in B. Hindess (ed.) Sociological Theories of the Economy, New York: Holmes &Meier.Jordan, R.W. (1968) ‘Husserl’s Phenomenology as an “Historical” Science’, Social Research,35, 245–59.Kalmar, I. (1987) ‘The Völkerpsychologie of Lazarus and Steinthal and the ModernConcept of Culture’, Journal of the History of Ideas, 48, 671–90.Kant, I. (1781/7) Critique of Pure Reason, New York: St Martin’s Press, 1965.—— (1786) The Metaphysical Foundations of Natural Science, Indianapolis: The Bobbs-MerrillCompany, 1970.Kauder, E. (1958) ‘Intellectual and Political Roots of the Older Austrian School’, Zeitschriftfür Nationalökonomie, 17, 411–25.Kaufmann, F. (1940) ‘Phenomenology and Logical Empiricism’, pp. 124–42 in M. Farber(ed.) Philosophical Essays in Memory of Edmund Husserl, Cambridge, MA: Harvard UniversityPress.—— (1958) Methodology of the Social Sciences, New York: Humanities Press.Kemeny, J.G. and Oppenheim, P. (1956) ‘On Reduction’, Philosophical Studies, 7, 6–19.Kessler, S.J. and McKenna, W. (1978) Gender. An Ethnomethodological Approach, Chicago:University of Chicago Press.Keynes, J.M. (1936) The General Theory of Employment, Interest and Money, vol. 7: The CollectedWritings of John Maynard Keynes, London: Macmillan, 1973.—— (1937) ‘The General Theory of Employment’, Quarterly Journal of Economics, 51,209–23.

Bibliography 411—— (1939) ‘Professor Tinbergen’s Method’, The Economic Journal, 49, 558–606.—— (1940) ‘Comment’ on J. Tinbergen, ‘On A Method of Statistical Research’, TheEconomic Journal, 50, 154–56.Khalil, E.L. (1997) ‘Is the Firm an Individual?’, Cambridge Journal of Economics, 21, 519–44.Kim, J. (1978) ‘Supervenience and Nomological Incommensurables’, American PhilosophicalQuarterly, 15, 149–56.—— (1984) ‘Concepts of Supervenience’, Philosophy and Phenomenological Research, 45,153–76.Kingdom, J. (1992) No Such Thing as Society? Individualism and Community, Buckingham:Open University Press.Kinkaid, H. (1986) ‘Reduction, Explanation and Individualism’, pp. 497–515 in M.Martin and L.C. McIntyre (eds) Readings in the Philosophy of Social Science, Cambridge,MA: MIT Press.—— (1996) Philosophical Foundations of the Social Sciences. Analyzing Controversies in SocialResearch, Cambridge: Cambridge University Press.—— (1997) Individualism and the Unity of Science. Essays on Reduction, Explanation, and theSpecial Sciences, Lanham: Rowman and Littlefield.Kirman, A. (1989) ‘The Intrinsic Limits of Modern Economic Theory: The EmperorHas No Clothes’, The Economic Journal, 99 (Conference 1989), 126–39.—— (1992) ‘Whom or What Does the Representative Individual Represent?’, Journal ofEconomic Perspectives, 6, 117–36.Klein, L.R. (1946a) ‘Macroeconomics and the Theory of Rational Behavior’, Econometrica,14, 93–108.—— (1946b) ‘Remarks on the Theory of Aggregation’, Econometrica, 14, 303–12.—— (1947) The Keynesian Revolution, New York: The Macmillan Company.Kluckhohn, C. (1963) ‘Parts and Wholes in Cultural Analysis’, pp. 111–33 in D. Lerner(ed.) Parts and Wholes, New York: The Free Press.Knight, F.H. (1921) Risk, Uncertainty and Profit, Boston: Houghton Mifflin.—— (1924) ‘Some Fallacies in the Interpretation of Social Costs’, The Quarterly Journal ofEconomics, 38, 582–606.—— (1933) ‘Social Economic Organization’, pp. 3–19 in W. Breit and H.M. Hochman(eds) Readings in Microeconomics, New York: Holt, Rinehart and Wilson, 1971.—— (1935/6) The Ethics of Competition and Other Essays, London: George Allen & Unwin.—— (1937) ‘Unemployment: And Mr Keynes’s Revolution in Economic Theory’, pp.66–95 in H. Hazlitt (ed.) The Critics of Keynesian Economics, Princeton, NJ: D. VanNostrand Company, 1960.—— (1956) On the History and Method of Economics, Chicago: University of Chicago Press.—— (1961) ‘Methodology in Economics – Part I and II’, The Southern Economic Journal, 22,185–93, 273–82.Knoke, D. (1990) Political Networks. The Structural Perspective, Cambridge: CambridgeUniversity Press.Knorr-Cetina, K. (1981) ‘The Micro-Sociological Challenge of Macro-Sociology:Towards a Reconstruction of Social Theory and Methodology’, introduction (pp.1–47) to K. Knorr-Cetina and A.V. Cicourel (eds) Advances in Social Theory. Toward anIntegration of Micro- and Macro-Sociologies, Boston: Routledge & Kegan Paul.Koertge, N. (1974) ‘On Popper’s Philosophy of Social Science’, pp. 195–208 in K.F.Schaffner and R.S. Cohen (eds) Boston Studies in the Philosophy of Science, vol. 20,Dordrecht: Reidel.Köhler, F. (1922) Wesen und Bedeutung des Individualismus, Munich: Rösl and Cie.

412 BibliographyKoopmans, T.J. (1947) ‘Measurement Without Theory’, The Review of Economics and Statistics,29, 161–72.—— (1949) ‘A Reply’ to R. Vining, ‘Koopmans on the Choice of Variables to be Studiedand of Methods of Measurement’, The Review of Economics and Statistics, 31, 86–91.Kotarba, J.A. and Fontana, A., (eds) (1984) The Existential Self in Society, Chicago: Universityof Chicago Press.Kreps, D.M. (1990a) A Course in Microeconomic Theory, New York: Harvester Wheatsheaf.—— (1990b) Game Theory and Economic Modelling, Oxford: Clarendon Press.Kruks, S. (1990) Situation and Human Existence: Freedom, Subjectivity and Society, London:Unwin Hyman.Krüger, L. (1976) ‘Reduction versus Elimination of Theories’, Erkenntnis, 10, 295–309.Kuhn, M.F. (1967) “The Reference Group Reconsidered’, pp. 171–84 in J.G. Manis andB.H. Meltzer (eds) Symbolic Interaction. A Reader in Social Psychology, Boston: Allyn andBacon.Kuhn, T.S. (1962) The Structure of Scientific Revolutions, Chicago: Chicago University Press,1970.Kunkel, J.H. (1970) Society and Economic Growth, New York: Oxford University Press.Kusch, M. (1995) Psychologism: A Case Study in the Sociology of Philosophical Knowledge, London:Routledge.Kuznets, S. (1963) ‘Parts and Wholes in Economics’, pp. 41–58 in D. Lerner (ed.) Parts andWholes, New York: The Free Press of Glencoe.Lachenmeyer, C.W. (1970) ‘Reduction in Sociology: A Pseudo Problem’, Pacific SociologicalReview, 13, 211–17.—— (1971) The Language of Sociology, New York: Columbia University Press.Lachmann, L.M. (1969) ‘Methodological Individualism and the Market Economy’, pp.89–103 in E. Streissler (ed.) Roads to Freedom, London: Routledge & Kegan Paul.—— (1970) The Legacy of Max Weber, London: Heinemann.—— (1990) ‘Austrian Economics: A Hermeneutic Approach’, pp. 134–45 in D. Lavoie(ed.), Economics and Hermeneutics, London: Routledge.Lakatos, I. (1968) ‘Criticism and the Methodology of Research Programmes’, Proceedings ofthe Aristotelian Society, New Series, 69, 149–86.—— (1970) ‘Falsification and the Methodology of Scientific Research Programmes’, pp.91–196 in I. Lakatos and A. Musgrave (eds) Criticism and the Growth of Knowledge,Cambridge: Cambridge University Press.Landgrebe, L. (1977) ‘Phenomenology as Transcendental Theory of History’, pp. 101–13in F.A. Elliston and P. McCormick (eds) Husserl. Expositions and Appraisal, Notre Dame:University of Notre Dame Press.Landheer, B. (1952) Mind and Society. Epistemological Essays on Sociology, The Hague: MartinusNijhoff.Lange-von Kulessa, J. (1997) ‘Searching for a Methodological Synthesis – Hayek’s Individualismin the Light of Recent Holistic Criticism’, The Journal of EconomicMethodology, 4, 267–87.Langlois, R.N. (1986a) ‘The New Institutional Economics: An Introductory Essay’, pp.1–26 in R.N. Langlois (ed.) (1986).—— (1986b) ‘Rationality, Institutions, and Explanation’, pp. 225–55 in R.N. Langlois(ed.) (1986).—— (ed.) (1986) Economics as a Process. Essays in the New Institutional Economics, Cambridge:Cambridge Universitity Press.

Bibliography 413Latsis, S. (1972) ‘Situational Determinism in Economics’ The British Journal for the Philosophyof Science, 23, 207–45.—— (1978) ‘A Research Programme in Economics’, pp. 1–42 in S. Latsis (ed.) Method andAppraisal in Economics, Cambridge: Cambridge University Press.Laudan, L. (1977) Progress and its Problems: Towards a Theory of Scientific Growth, London:Routledge & Kegan Paul.Lavoie, D. (1986) ‘Euclideanism versus Hermeneutics: A Reinterpretation of Misean APriorism’, pp. 192–210 in I. Kirzner (ed.) Subjectivism, Intelligibility and Economic Understanding.Essays in Honor of Ludwig Lachmann on his Eightieth Birthday, New York: New YorkUniversity Press.—— (1990) ‘Understanding Differently: Hermeneutics and the Spontaneous Order ofCommunicative Processes’, pp. 359–77 in B.J. Caldwell (ed.) Carl Menger and his Legacyin Economics, Durham: Duke University Press.Lazarsfeld, P.F. (1959) ‘Problems in Methodology’, pp. 39–78 in R.K. Merton, L. Broomand L.S. Cottrell, Jr (eds) Sociology Today, vol. 1: Problems and Prospects, New York:Harper & Row, 1965.—— (1962) ‘Philosophy of Science and Empirical Social Research’, pp. 463–73 in E.Nagel, P. Suppes and A. Tarski, (eds) Logic, Methodology and Philosophy of Science, Stanford:Stanford University Press.—— (1968) ‘Evidence and Inference in Social Research’, pp. 608–34 in M. Brodbeck(ed.) Readings in the Philosophy of the Social Sciences, New York: Macmillan.—— (1970a) Main Trends in Sociology, New York: Harper & Row, 1973.—— (1970b) ‘The Place of Empirical Research in the Map of Contemporary Sociology’,pp. 301–18 in J.C. McKinney and E.A. Tiryakian (eds) Theoretical Sociology. Perspectivesand Developments, New York: Appleton-Century-Crofts.Lazarsfeld, P.F. and Menzel, H. (1969) ‘On the Relation between Individual and CollectiveProperties’, pp. 419–516 in A. Etzioni (ed.) A Sociological Reader on ComplexOrganizations, New York: Rinehart and Winston.Leary, D.E. (1978) ‘The Philosophical Development of the Conception of Psychology inGermany, 1780–1850’, Journal of the History of the Behavioral Sceinces, 14, 113–21.—— (1980) ‘German Idealism and the Development of Psychology in the NineteenthCentury’, Journal of the History of Philosophy, 18, 299–317.Leijonhufvud, A. (1967) ‘Keynes and the Keynesians. A Suggested Interpretation’, AmericanEconomic Review, 57, 401–10.—— (1968) On Keynesian Economics and the Economics of Keynes. A Study in Monetary Theory,New York: Oxford University Press.—— (1969) Keynes and the Classics: Two Lectures on Keynes’ Contribution to Economic Theory,London: Institute of Economic Affairs.Lemert, C.C. (1979) ‘De-centred Analysis: Ethnomethodology and Structuralism’, Theoryand Society, 7, 273–88.Lennon, K. and Charles, D. (1992) ‘Introduction’ (pp. 1–18) to D. Charles and K. Lennon(eds) Reduction, Explanation and Realism, Oxford: Clarendon.Leonard, R.J. (1992) ‘Creating a Context for Game Theory’, pp. 29–76 in E.R. Weintraub(ed.) Toward a History of Game Theory, annual supplement to vol. 24 of History ofPolitical Economy, Durham: Duke University Press.Lessnoff, M. (1974) The Structure of Social Science, London: George Allen and Unwin.Levine, A., Sober, E. and Wright, E.O. (1987) ‘Marxism and Methodological Individualism’,New Left Review, 162, 67–84.

414 BibliographyLevinthal, D. (1988) ‘A Survey of Agency Models of Organization’, The Journal ofEconomic Behavior and Organization, 9, 153–85.Lewis, D. (1969) Convention. A Philosophical Study, Cambridge, MA: Harvard UniversityPress.Lewis, J.D. and Smith, R.L. (1980) American Sociology and Pragmatism: Mead, Chicago Sociology,and Symbolic Interactionism, Chicago: Chicago University Press.Lichtblau, K. (1994) ‘Kausalität oder Wechselwirkung? Max Weber und Georg Simmelim Vergleich’, pp. 527–62 in G. Wagner and H. Zipprian (eds) Max Webers Wissenschaftslehre,Frankfurt am Main: Suhrkamp.Lichtheim, R. (1965) ‘Karl Popper’s Defence of the Autonomy of Sociology’, SocialResearch, 32, 1–25.Lindenberg, S. (1977) ‘Individuelle Effekte, kollektive Phänomene und das Problem derTransformation’, pp. 47–84 in K. Eichner and W. Habermehl (eds) Probleme derErklärung sozialen Verhaltens, Meisenheim am Glan: Anton Hain.—— (1985a) ‘An Assessment of the New Political Economy: Its Potential for the SocialSciences and for Sociology in Particular’, Sociological Theory, 3, 99–113.—— (1985b) ‘Rational Choice and Sociological Theory: New Pressures on Economics asa Social Science’, Journal of Institutional and Theoretical Economics, 41 , 244–55.—— (1990) ‘Homo Socio-Oeconomicus: The Emergence of a General Model of Man inthe Social Sciences’, Journal of Institutional and Theoretical Economics, 146, 727–48.—— (1992) ‘The Method of Decreasing Abstraction’, pp. 3–20 in J.S. Coleman and T.J.Fararo (eds) Rational Choice Theory. Advocacy and Critique, Newbury Park: Sage Publications.—— (1996) ‘Choice-Centred Versus Subject-Centred Theories in the Social Sciences:The Influence of Simplification on Explananda’, European Sociological Review, 12, 147–57.Lipset, S.M. (1959) Political Man, London: Mercury Books, 1963.Little, D. (1991) Varieties of Social Explanation. An Introduction to the Philosophy of Science,Boulder: Westview Press.—— (1998) Microfoundation, Method, and Causation, New Brunswick, NJ: TransactionPublishers.Locke, J. (1690) Two Treatises of Civil Government, Cambridge: Cambridge University Press,1960.Lucas, R.E. (1972) ‘Expectations and the Neutrality of Money’, Journal of Economic Theory,4, 103–24.—— (1975) ‘An Equilibrium Model of the Business Cycle’, Journal of Political Economy, 83,1113–44.—— (1976) ‘Econometric Policy Evaluation: A Critique’, pp. 19–45 in K. Brunner andA.H. Meltzer (eds) The Phillips Curve and Labor Markets, Amsterdam: North-Holland.Lucas, R.E. and Sargent, T.J. (1979) ‘After Keynesian Economics’, pp. 295–319 in R.E.Lucas and T.J. Sargent (eds) Rational Expectations and Econometric Practice, Minneapolis:University of Minnesota Press, 1981.Lukes, S. (1968) ‘Methodological Individualism Reconsidered’, The British Journal of Sociology,19, 119–29.—— (1973) Individualism, New York: Harper & Row.Lundberg, G.A. (1929) Social Research. A Study in Methods of Gathering Data, New York:Longman, Green and Co, 1942.—— (1939) Foundations of Sociology, New York: The Macmillan Company.

Bibliography 415Macaulay, T.B. (1829) ‘Mill’s Essay on Government: Utilitarian Logic and Politics’, pp.97–127 in J. Lively and J. Rees (eds) Utilitarian Logic and Politics, Oxford: ClarendonPress, 1978.Macdonald, G. and Pettit, P. (1981) Semantics and Social Science, London: Routledge &Kegan Paul.Macdonald, R. (1965) ‘Schumpeter and Max Weber – Central Visions and Social Theories’,Quarterly Journal of Economics, 79, 373–96.McDougall, W. (1908) Social Psychology, London: Methuen, 1924.—— (1920) The Group Mind, Cambridge: Cambridge University Press.Machlup, F. (1951) ‘Joseph Schumpeter’s Economic Methodology’, pp. 461–74 in Methodologyof Economics and Other Social Sciences, New York: Academic Press, 1978.MacIntyre, A. (1972) ‘A Mistake About Causality in the Social Sciences’, pp. 48–70 in P.Laslett and W.G. Runciman (eds) Philosophy, Politics and Society, Second Series, Oxford:Basil Blackwell.MacIver, R.M. (1911) ‘Society and State’, The Philosophical Review, 20, 30–45.—— (1913) ‘What is Social Psychology’, The Sociological Review, 6, 147–60.—— (1914) ‘Society and the “Individual”’, The Sociological Review, 7, 58–64.—— (1915) ‘Personality and the Suprapersonal’, The Philosophical Review, 24, 501–25.McKenzie, L.W. (1989) ‘General Equilibrium’, pp. 1–35 in J. Eatwell, M. Milgate and P.Newman (eds) General Equilibrium (The New Palgrave), New York: W.W. Norton.MacLean, M.J. (1982) ‘Johann Gustav Droysen and the Development of HistoricalHermeneutics’, History and Theory, 21, 347–65.MacPhail, C. and Rextroat, C. (1979) ‘Mead vs. Blumer: The Divergent MethodologicalPerspectives of Social Behaviorism and Symbolic Interactionism’, American SociologicalReview, 44, 449–67.Madison, G.B. (1990) ‘How Individualistic is Methodological Individualism?’, CriticalReview, 4, 41–60.Mäki, U., Gustavsson B. and Knudsen, C. (eds) (1993) Rationality, Institutions and EconomicMethodology, London: Routledge.Makkreel, R.A. (1975) Dilthey. Philosopher of the Human Studies, Princeton, NJ: PrincetonUniversity Press, 1992.Makkreel, R. A. and Scanlon, J. (1987) Dilthey and Phenomenology, Washington: UniversityPress of America.Malewski, A. (1964) Verhalten und Interaktion, Tübingen: J.C.B. Mohr, 1977.Malthus, T. (1798) An Essay on the Principle of Population and A Summary View of the Principle ofPopulation, Harmondsworth: Penguin, 1970.—— (1820) Principles of Political Economy, vol. 1, Cambridge: Cambridge University Press,1989.—— (1827) Definitions in Political Economy, New York: Augustus M. Kelley, 1971.—— (1830) ‘A Summary View of the Principle of Population’, pp. 219–72 in Malthus(1798).Mandelbaum, M. (1955) ‘Societal Facts’, The British Journal of Sociology, 6, 305–17.—— (1957) ‘Societal Laws’, The British Journal for the Philosophy of Science, 8, 211–24.—— (1977) ‘Psychology and Societal Facts’ pp. 235–53 in R.G. Colodny (ed.) Logic, Laws,and Life, Pittsburgh: University of Pittsburgh Press.Mandeville, B. (1714/29/32) The Fable of the Bees, or Private Vices, Public Benefits, Indianapolis:Liberty Fund, 1988.Manis, J.G. and Meltzer, B.N. (1967) ‘Preface’ to Manis and Meltzer (eds) Symbolic Interaction.A Reader in Social Psychology, Boston: Allyn and Bacon.

416 BibliographyMannheim, K. (1924) ‘Historismus’, Archiv für Sozialwissenschaft und Sozialpolitik, 52, 1–60.Manning, P.K. (1971) ‘Talking and Becoming: A View of Organizational Socialization’,pp. 239–56 in J.D. Douglas (ed.) Understanding Everyday Life, London: Routledge &Kegan Paul.March, J.G. and Olsen, J.P. (1989) Rediscovering Institutions. The Organizational Basis of Politics,New York: The Free Press.Marchionatti, R. and Gambino, E. (1997) ‘Pareto and Political Economy as a Science:Methodological Revolution and Analytical Advances in Economic Theory in the1890s’, Journal of Political Economy, 105, 1323–47.Margolis, J. (1986) ‘Psychology and Its Methodological Options’, pp. 12–52 in J. Margolis,P.T. Manicas, R. Harré and P.F. Secord, Psychology: Designing a Discipline, Oxford: BasilBlackwell.Markovsky, B. (1997) ‘Network Games’, Rationality and Society, 9, 67–90.Markovsky, B., Willer, D. and Patton, T. (1988) ‘Power Relations in Exchange Networks’,American Sociological Review, 53, 220–36.Marshall, A. (1890) Principles of Economics, London: Macmillan, 1920.Martin, M. (1972) ‘Explanation in Social Science: Some Recent Work’, The Philosophy ofthe Social Sceinces, 2, 61–81.Marx, K. (1857–8) Grundrisse: Foundations of the Critique of Political Economy,Harmondsworth: Penguin, 1973.—— (1867) Capital: A Critique of Political Economy, vol. 1, Harmondsworth: Penguin, 1976.Marx, K. and Engels, F. (1848) The Communist Manifesto, Harmondsworth: Penguin, 1967.Mathien, T. (1988) ‘Network Analysis and Methodological Individualism’, The Philosophy ofthe Social Sciences, 18, 1–20.Maull, N.L. (1977) ‘Unifying Science Without Reduction’, Studies in History and Philosophy ofScience, 8, 143–62.Mayerl, W.M. (1973) ‘Ethnomethodology: Sociology without Society?’, pp. 262–79 in F.R.Dallmayer and T.A. McCarthy (eds) Understanding and Social Inquiry, Notre Dame:University of Notre Dame Press, 1977.Mead, G.H. (1932) The Philosophy of the Present, Chicago: Chicago University Press, 1980.—— (1934) Mind, Self and Society, Chicago: Chicago University Press, 1962.—— (1936) Movements of Thought in the Nineteenth Century, Chicago: Chicago UniversityPress.—— (1938) The Philosophy of the Act, Chicago: Chicago University Press.—— (1964) Selected Writings, Chicago: Chicago University Press.—— (1982) The Individual and the Social Self, Chicago: Chicago University Press.Mehan, H. and Wood, H. (1975) ‘An Image of Man for Ethnomethodology’, The Philosophyof the Social Sciences, 5, 365–76.Meinecke, F. (1959) Historism. The Rise of the Historical Outlook, London: Routledge &Kegan Paul, 1972.Mellor, D.H. (1982) ‘The Reduction of Society’, Philosophy, 57, 51–75.Meltzer, B.N. (1967) ‘Mead’s Social Psychology’, pp. 5–24 in J.G. Manis and B.N. Meltzer,(eds) Symbolic Interaction. A Reader in Social Psychology, Boston: Allyn and Bacon.Meltzer, B.N. and Petras, J.W. (1973) ‘The Chicago and Iowa Schools of Symbolic Interactionism’,pp. 3–17 in T. Shibutani (ed.) Human Nature and Collective Behavior. Papers inHonor of Herbert Blumer, New Brunswick, NJ: Transaction Books.Menger, C. (1871) Principles of Economics, NewYork: New York University Press, 1976.—— (1883) Problems of Economics and Sociology, Urbana: University of Illinois Press, 1963.—— (1884) Irrthümer des Historismus in der deutschen Nationalökonomie, Wien: Alfred Hölder.

Bibliography 417—— (1887) ‘Zur Kritik der Politischen Oekonomie’, pp. 99–132 in Kleinere Schriften zurMethode und Geschichte der Volkswirtschaftslehre, The Collected Works of Carl Menger, vol. III,London: London School of Economics and Political Science, 1935.—— (1889) ‘Grundzüge einer klassischen Nationalökonomie und die moderneWirtschaftspolitik’, pp. 185–218 in C. Menger, Kleinere Schriften zur Methode und Geschichteder Volkswirtschaftslehre, The Collected Works of Carl Menger, vol. III, London: LondonSchool of Economics and Political Science, 1935.—— (1892) ‘On the Origin of Money’, The Economic Journal, 2, 239–55.—— (1894) ‘Wilhelm Roscher’, pp. 273–81 in Kleinere Schriften zur Methode und Geschichte derVolkswirtschaftslehre, The Collected Works of Carl Menger, vol. III, London: The LondonSchool of Economics and Political Science, 1935.Merleau-Ponty, M. (1945) Phenomenology of Perception, London: Routledge, 1962.—— (1974) Phenomenology, Language and Sociology, London: Heinemann.Merton, R.M. (1967) On Theoretical Sociology. Five Essays, Old and New, New York: The FreePress.Meyer, J.W. and Scott, W.R. (1983) Organizational Environments. Ritual and Rationality, BeverlyHills: Sage Publications.Milford, K. (1990) ‘ Menger’s Methodology’, pp. 215–39 in B.J. Caldwell (ed.) Carl Mengerand his Legacy in Economics, Durham: Duke University Press.Milgate, M. (1989) ‘Equilibrium: Development of the Concept’, pp. 105–13 in J. Eatwell,M. Milgate and P. Newman (eds) The Invisible Hand (The New Palgrave), New York:W.W. Norton.Mill, J. (1820) ‘Essay on Government’, pp. 53–95 in J. Lively and J. Rees (eds) UtilitarianLogic and Politics, Oxford: Clarendon Press, 1978.—— (1820) The Elements of Political Economy, pp. 203–366 in Selected Economic Writings,Chicago: Chicago University Press, 1826.—— (1829) The Analysis of the Phenomena of the Human Mind, 2 vols, London: Longmans,1878.Mill, J.S. (1836) ‘On the Definition of Political Economy and the Method of InvestigationProper to It’, pp. 407–40 in Philosophy of Scientific Method, New York: Hafner, 1950.—— (1840a) ‘Coleridge’, pp. 117–63 in Essays on Ethics, Religion and Society, Collected Works,vol. X, Toronto: Toronto University Press, 1969.—— (1840b) ‘M. de Tocqueville on Democracy in America’, pp. 105–58 in Mill’s EthicalWritings, New York: Collier Books, 1965.—— (1843–72) A System of Logic: Ratiocinative and Inductive, Collected Works, vols VII–VIII,Toronto: University of Toronto Press, 1974.—— (1848) Principles of Political Economy, Fairfield: Augustus M. Kelley, 1976.—— (1859) On Liberty, pp. 1–141 in Three Essays, Oxford: Oxford University Press, 1975.—— (1865) Auguste Comte and Positivism, Ann Arbor: University of Michigan Press.—— (1873) Autobiography, pp. 1–182 in Essential Works of John Stuart Mill, New York:Bantam Books, 1961.—— (1865) An Examination of Sir William Hamilton’s Philosophy, London: Longmans, Green,Reader and Dyer, 1978.Miller, R.W. (1978) ‘Methodological Individualism and Social Explanation’, Philosophy ofScience, 45, 387–414.—— (1987) Fact and Method. Explanation, Confirmation and Reality in the Natural and the SocialSciences, Princeton, NJ: Princeton University Press.—— (ed.) (1994) The Rational Expectations Revolution. Readings from the Front Line, Cambridge,MA: MIT Press.

418 BibliographyMills, C.W. (1959) The Sociological Imagination, Harmondsworth: Penguin,1970.Mirowski, P. (1981) ‘Is There a Mathematical Neoinstitutional Economics?’, Journal ofEconomic Issues, 15, 593–613.—— (1986) ‘Institutions as a Solution Concept in a Game Theory Context’, pp. 243–64in L. Samuelson (ed.) Microeconomic Theory, Boston: Kluwer-Nijhoff Publishing.—— (1992) ‘What Were von Neumann and Morgenstern Trying to Accomplish?’, pp.113–47 in E.R. Weintraub (ed.) Toward a History of Game Theory, annual supplement tovol. 24 of History of Political Economy, Durham: Duke University Press.Mises, L. von (1922) Socialism. An Economic and Sociological Analysis, London: Jonathan Cape,1951.—— (1927) Liberalism. A Socio-Economic Exposition, Kansas City: Sheed Andrews andMcMeel, Inc.—— (1933) Epistemological Problems of Economics, New York: New York University Press,1976.—— (1942) ‘Social Science and Natural Science’, Journal of Social Philosophy and Jurisprudence,7, 240–53.—— (1943/4) ‘The Treatment of Irrationality in the Social Sciences’, Philosophy andPhenomenological Research, 4, 527–45.—— (1949) Human Action. A Treatise on Economics, Chicago: Contemporary Books, 1966.—— (1952) Planning for Freedom, South Holland, IL: Libertarian Press, 1974.—— (1957) Theory and History. An Interpretation of Social and Economic Evolution, AuburnUniversity, Alabama, and Washington, DC: Ludwig von Mises Institute, 1985.—— (1961) ‘Epistemological Relativism in the Sciences of Human Action’, pp. 117–34 inH. Schoeck and J. W. Wiggins (eds) Relativism and the Study of Man, Princeton, NJ: VanNostrand.—— (1962) The Ultimate Foundation of Economic Science. An Essay on Method, Princeton, NJ:Van Nostrand.—— (1969) The Historical Setting of the Austrian School of Economics, New Rochelle, NY:Arlington House.—— (1978) Notes and Recollections, South Holland, IL: Libertarian Press.—— (1990) Money, Method, and the Market Process, Dordrecht: Kluwer Academic Publishers.Modigliani, F. (1944) ‘Liquidity Preference and the Theory of Interest and Money’, Econometrica,12, 45–88.Mohanty, J.N. (1989) ‘Psychologism’, pp. 1–10 in M.A. Notturno (ed.) Perspectives onPsychologism, Leiden: E.J. Brill.Mommsen, W. (1974) The Age of Bureaucracy. Perspectives on the Political Sociology of Max Weber,New York: Harper & Row, 1977.Morgan, C.L. (1923) Emergent Evolution, London: Williams and Northgate.Morgenbesser, S. (1967) ‘Psychologism and Methodological Individualism’, pp. 160–74 inS. Morgenbesser (ed.) Philosophy of Science Today, New York: Basic Books.Morgenstern, O. (1935) ‘Perfect Foresight and Economic Equilibrium’, pp. 169–83 in A.Schotter (ed.) Selected Writings of Oscar Morgenstern, New York: New York UniversityPress, 1976.—— (1972) ‘Thirteen Critical Points in Contemporary Economic Theory: An Interpretation’,Journal of Economic Literature, 10, 1163–89.Morishima, M. (1984) ‘The Good and Bad Uses of Mathematics’, pp. 51–73 in P. Wilesand G. Routh (eds) Economics in Disarray, Oxford: Basil Blackwell.Möschel, W. (1993) ‘The New Institutional Economics Meets Law and Economics’,Journal of Institutional and Theoretical Economics, 149, 88–91.

Bibliography 419Moscovici, S. (1972) ‘Society and Theory in Social Psychology’, pp. 17–68 in J. Israel andH. Tajfel (eds) The Context of Social Psychology: A Critical Assessment, London: AcademicPress.—— (1988) The Invention of Society, Cambridge: Polity Press, 1993Muth, J.F. (1961) ‘Rational Expectations and the Theory of Price Movements’, Econometrica,29, 315–35.Myers, M.L. (1983) The Soul of Modern Economic Man. Ideas of Self-Interest. Thomas Hobbes toAdam Smith, Chicago: Chicago University Press.Myrdal, G. (1929) The Political Element in the Development of Economic Theory, London: Routledge& Kegan Paul, 1953.—— (1975) Against the Stream. Critical Essays on Economics, New York: Vintage Books.Nagel, E. (1949) ‘The Meaning of Reduction in the Natural Sciences’, pp. 99–135 inR.C. Stauffer (ed.) Science and Civilization, Madison: University of Wisconsin Press.—— (1961) The Structure of Science: Problems in the Logic of Scientific Explanation, London:Routledge & Kegan Paul.—— (1970) ‘Issues in the Logic of Reductive Explanations’, pp. 117–37 in H.E. Keiferand K. Munitz (eds) Mind, Science and History, Albany: State University of New YorkPress.Natanson, M. (1968) ‘Alfred Schutz on Social Reality and Social Science’, Social Research,35, 217–44.—— (1973) ‘Phenomenology and the Social Sciences’, pp. 3–44 in M. Natanson (ed.)Phenomenology and the Social Sciences, vol. 1, Evanston: Northwestern University Press.Nee, V. (1998) ‘Sources of the New Institutionalism’, pp. 1–16 in M.C. Brinton and V.Nee (eds) The New Institutionalism in Sociology, New York: Russell Sage Foundation.Nelson, A. (1986) ‘New Individualistic Foundations for Economics’, Nous, 20, 469–90.—— (1989) ‘Average Explanations’, Erkenntnis, 30, 23–42.Nelson, J.A. (1996) Feminism, Objectivity and Economics, London: Routledge.Nelson, R.R. and Winter, S.G. (1982) An Evolutionary Theory of Economic Change,Cambridge, MA: Belknap Press of Harvard University Press.Neumann, J. von and Morgenstern, O. (1944) Theory of Games and Economic Behavior,Princeton: Princeton University Press, 1953.Neurath, O. (1930/1) ‘Wege der wissenschaftlichen Weltauffassung’, Erkenntnis, 1, 106–25.—— (1931/2) ‘Sociology and Physicalism’, pp. 282–317 in A.J. Ayer (ed.) LogicalPositivism, New York: The Free Press, 1959.—— (1931a) ‘Empirical Sociology’, pp. 318–420 in O. Neurath, Empiricism and Sociology,Dordrecht-Holland: D. Reidel Publishing Company, 1973.—— (1931b) ‘Physicalism’, The Monist, 41, pp. 618–23—— (1932/3) ‘Protocol Sentences’, pp. 199–208 in A.J. Ayer (ed.) Logical Positivism, NewYork: The Free Press, 1959.—— (1944) Foundations of the Social Sciences, Chicago: Chicago University Press.Neurath, O., Hahn, H. and Carnap, R. (1929) ‘The Scientific Conception of the World:The Vienna Circle’, pp. 299–318 in O. Neurath, Empiricism and Sociology, Dordrecht-Holland: D. Reidel Publishing Company, 1973.Nicholas, B. (1962) An Introduction to Roman Law, Oxford: Clarendon Press.Nickles, T. (1973) ‘Two Concepts of Theoretic Reduction’, The Journal of Philosophy, 70,181–201.North, D.C. (1981) Structure and Change in Economic History, New York: W.W. Norton.North, D.C. and Thomas, R.P. (1970) ‘An Economic Theory of the Growth of theWestern World’, The Economic History Review, Second Series, 28, 1–17.

420 Bibliography—— (1971) ‘The Rise and Fall of the Manorial System’, The Journal of Economic History,31, 777–803.—— (1973) The Rise of the Western World. A New Economic History, Cambridge: CambridgeUniversity Press.Notturno, M.A. (1985) Objectivity, Rationality and the Third Realm: Justification and the Grounds ofPsychologism. A Study of Frege and Popper, Dordrecht: Martinus Nijhoff.Novak, M.E. (1962) Economics and the Fiction of Daniel Defoe, Berkeley: University of CaliforniaPress.Nozick, R. (1977) ‘On Austrian Methodology’, Synthese, 36, 353–92.Oakes, G. (1980) ‘Introduction’ (pp. 3–94) to G. Simmel, Essays on Interpretation in SocialScience, Totowa, NJ: Rowman and Littlefield.Oakley, A. (1994) Classical Economic Man: Human Agency and Methodology in the PoliticalEconomy of Adam Smith and J.S. Mill, Aldershot: Edward Elgar.Obershall, A. (ed.) (1972) The Establishment of Empirical Sociology: Studies in Continuity, Discontinuity,and Institutionalization, New York: Harper & Row.O’Driscoll, G.P. (1986) ‘Money: Menger’s Evolutionary Theory’, History of PoliticalEconomy, 18, 601–16.Olson, M. (1965) The Logic of Collective Action. Public Goods and the Theory of Groups, Harvard,MA: Harvard University Press, 1971.—— (1982) The Rise and Decline of Nations. Economic Growth, Stagflation and Social Rigidities,New Haven: Yale University Press.Opp, K.-D. (1972) Verhaltenstheoretische Soziologie, Hamburg: Rowohlt.—— (1973) ‘Die Verhaltenstheoretische Soziologie: ihr Forschungsprogram und einigeihrer Probleme’, pp. 39–60 in K.-D. Opp and H.J. Hummell, Probleme der ErklärungSozialer Prozesse vol II: Soziales Verhalten und Soziale Systeme, Frankfurt: Athenäum.—— (1977a) ‘Soziologie ohne Psychologie?’, pp. 69–89 in G. Eberlein and H.-J. Kondratowitz(eds) Psychologie statt Soziologie? Zur Reduzierbarkeit sozialer Strukturen auf Verhalten,Frankfurt/New York: Campus Verlag.—— (1977b) ‘Die Verhaltenstheoretische Soziologie als Sozialwissenschaftliches“Paradigma”’, pp. 121–56 in H. Lenk (ed.) Handlungstheorien Interdisciplinär, Vol. IV,Munich: Fink.—— (1978) ‘Das “Ökonomische Program” in der Soziologie’, Soziale Welt, 29, 129–54.—— (1979) Individualistische Sozialwissenschaft, Stuttgart: Ferdinand Enke Verlag.—— (1985) ‘Sociology and Economic Man’, Journal of Institutional and Theoretical Economics,141, 213–43.—— (1988) ‘The Individualistic Research Program in Sociology’, pp. 208–24 in G.Radnitzky (ed.) Centripetal Forces in the Sciences, vol. 2, New York: Paragon House.Oppenheim, P. and Putnam, H. (1958) ‘Unity of Science as a Working Hypothesis’, pp.3–36 in H. Feigl, M. Scriven and G. Maxwell (eds) Minnesota Studies in the Philosophy ofScience, vol. II, Minneapolis: Minnesota University Press.Ordeshook, P.C. (1986) Game Theory and Political Theory. An Introduction, Cambridge:Cambridge University Press.—— (1992) ‘The Reintegration of Political Science and Economics’, pp. 171–200 in G.Radnitzky (ed.) Universal Economics. Assessing the Achievements of the Economic Approach, NewYork: Paragon House.Österberg, D. (1988) Metasociology. An Inquiry into the Origins and Validity of Social Thought,Oslo: Norwegian University Press.

Bibliography 421Osterhammel, J. (1987) ‘Varieties of Social Economics: Joseph A. Schumpeter and MaxWeber’, pp. 106–20 in W. Mommsen and J. Osterhammel (eds) Max Weber and HisContemporaries, London: Allen & Unwin.Outhwaite, W. (1975) Understanding Social Life. The Method called Verstehen, London: GeorgeAllen & Unwin.Pandit, G.L. (1971) ‘Two Concepts of Psychologism’, Philosophical Studies, 22, 85–91.Papineau, D. (1978) For Science in Social Science, London: Macmillan.Pareto, V. (1896) Cours d’Economie Politique, pp. 97–122 (selections) in Sociological Writings,Oxford: Basil Blackwell, 1976.—— (1909) Manual of Political Economy, London: Macmillan, 1972.—— (1916) The Mind and Society. A Treatise on General Sociology, 2 vols. New York: DoverPublications, 1963.Parsons, S. (1990) ‘The Philosophical Roots of Modern Austrian Economics: Past Problemsand Future Prospects’, History of Political Economy, 22, 295–319.Parsons, T. (1937) The Structure of Social Action, vol. I: Marshall, Pareto Durkheim, vol. 2:Weber, New York: The Free Press, 1968.—— (1951) The Social System, London: Routledge & Kegan Paul, 1970.Paul, E.F. (1988) ‘Liberalism, Unintended Orders and Evolutionism’, Political Studies, 36,251–272.—— (1989) ‘Herbert Spencer: Second Thoughts. A Reply to Michael Taylor’, PoliticalStudies, 37, 443–448.Pawson, R. (1989) A Measure for Measures: A Manifesto for Empirical Science, London: Routledge.Peacock, J.L. (1986) The Anthropological Lens. Harsch Light, Soft Focus, Cambridge: CambridgeUniversity Press.Pears, D. (1968) Bertrand Russell and the British Tradition in Philosophy, London: Fontana.Pejovich, S. (1972) ‘Towards an Economic Theory of the Creation and Specification ofProperty Rights’, Review of Social Economy, 30, 309–25.Perry, C. (1979) ‘Individualism and Causal Explanation’, Agora, 4, 1–15.—— (1980) ‘Popper, Winch, and Individualism’, International Philosophical Quarterly, 20,59–71.—— (1983) ‘The Explanatory Efficacy of Individualism’, The Philosophy of the SocialSciences, 13, 65–8.Pettit, P. (1984) ‘In Defence of a “New Methodological Individualism”: Reply to J.E.Tiles’, Ratio, 26, 81–7.—— (1993) The Common Mind. An Essay on Psychology, Society, and Politics, New York: OxfordUniversity Press, 1996.Pfister, B. (1928) Die Entwicklung zum Idealtypus, Tübingen: J.C.B. Mohr.Phelps, E.S. (1990) Seven Schools of Macroeconomic Thought, Oxford: Clarendon Press.Philippovich, E. von (1886) Aufgabe und Methode der Politischen Ökonomie, Freiburg: J.C.B.Mohr.Phillips, D.C. (1976) Holistic Thought in Social Science, Stanford: Stanford University Press.Phillips, D.L. (1973) Abandoning Method: Sociological Studies in Methodology, London: Jossey-Bass Publishers.Piaget, J. (1973) Main Trends in Psychology, London: George Allen & Unwin.Pizzorno, A. (1991) ‘On the Individualistic Theory of Social Order’, pp. 209–34 in P.Bourdieu and J.S. Coleman (eds) Social Theory for a Changing Society, Boulder: WestviewPress.

422 BibliographyPollner, M. (1974) ‘Sociological and Common-Sense Models of the Labelling Process’,pp. 27–40 in R. Turner (ed.) Ethnomethodology, Harmondsworth: Penguin.Popper, K.R. (1934) The Logic of Scientific Discovery, London: Hutchinson, 1972.—— (1944–5) ‘The Poverty of Historicism, I–III’, Economica, 11, 86–103, 119–37; 12,69–89.—— (1945) The Open Society and Its Enemies, vol. 1: Plato, vol. 2: Hegel and Marx, London:Routledge & Kegan Paul, 1966.—— (1957) The Poverty of Historicism, London: Routledge & Kegan Paul.—— (1962) Conjectures and Refutations. The Growth of Scientific Knowledge, New York: Harper& Row, 1968.—— (1962) ‘The Logic of the Social Sciences’, pp. 87–104 in T. Adorno, et al., The PositivistDispute in German Sociology, London: Heinemann, 1976.—— (1963–4) ‘Models, Instruments, and Truth. The Status of the Rationality Principlein the Social Sciences’, pp. 154–83 in The Myth of the Framework: In Defence of Science andRationality, London: Routledge, 1994.—— (1969) ‘A Pluralist Approach to the Philosophy of History’, pp. 181–200 in E.Streissler, (ed.) Roads to Freedom. Essays in Honour of Friedrich A. von Hayek, London: Routledge& Kegan Paul.—— (1970) ‘Reason or Revolution?’, Archives Européennes de Sociologie,11, 252–62.—— (1972) Objective Knowledge. An Evolutionary Approach, Oxford: Clarendon Press.—— (1974a) ‘Intellectual Autobiography’, pp. 3–181 in P.A. Schilpp (ed.) The Philosophy ofKarl Popper, Part I, La Salle, IL: Open Court.—— (1974b) ‘Replies to My Critics’, pp. 961–1196 in P.A. Schilpp (ed.) The Philosophy ofKarl Popper, Part II, La Salle, IL: Open Court.—— (1976) Unended Quest: An Intellectual Autobiography, Glasgow: Fontana/Collins.—— (1977) ‘Part I’ (pp. 3–223) in K.R. Popper and J.C. Eccles, The Self and Its Brain. AnArgument for Interactionism, Berlin: Springer.Posner, R.A. (1972) Economic Analysis of Law, Boston: Little, Brown and Company, 1977.—— (1980) ‘The Theory of Primitive Society, with Special Reference to Primitive Law’,The Journal of Law and Economics, 23, 1–54.—— (1981) The Economics of Justice, Cambridge, MA: Harvard University Press.—— (1993) ‘The New Institutional Economics Meets Law and Economics’, Journal ofInstitutional and Theoretical Economics, 149, 73–87.Powell, W.W. and DiMaggio, P.J. (eds) (1991) The New Institutionalism in Organizational Analysis,Chicago: Chicago University Press.Prasch, R.E. (1996) ‘The Origins of the Priori Method in Classical Political Economy: AReinterpretation’, Journal of Economic Issues, 30, 1105–25.Pratt, V. (1978) The Philosophy of the Social Sciences, London: Methuen.Prendergast, C. (1986) ‘Alfred Schutz and the Austrian School of Economics’, The AmericanJournal of Sociology, 92, 1–26.Pribram, K. (1912) Die Enstehung der individualistischen Sozialphilosophie, Leipzig: C.L.Hirschfeld.Przeworski, A. (1900) ‘The Ethical Materialism of John Roemer’, Politics and Society, 11,289–313.—— (1986) ‘Marxism and Rational Choice’, pp. 62–92 in P. Birnbaum and J. Leca (eds)Individualism: Theories and Methods, Oxford: Clarendon Press.Putnam, H. (1973–4) ‘Reductionism and the Nature of Psychology’, Cognition, 2, 131–46.Quetelet, L.A.J. (1835) A Treatise on Man and the Development of his Faculties, Edinburgh:William and Robert Chambers, 1842.

Bibliography 423Quine, W. O. (1953) From a Logical Point of View, New York: Harper & Row, 1961.Quinton, A. (1975–6) ‘Social Objects’, Proceedings of the Aristotelian Society, New Series, 76,1–27.Radnitzky, G. (ed.) (1987) Centripetal Forces in the Sciences, vol. 1, New York: Paragon House.—— (1988) ‘Introduction’ (pp. xiii–xlii) to G. Radnitsky (ed.) Centripetal Forces in the Sciencesvol. 2, New York: Paragon House.—— (ed.) (1988) Centripetal Forces in the Sciences, vol. 2, New York: Paragon House.Ranke, L. von (1836) ‘A Dialogue on Politics’, pp. 102–30 in The Theory and Practice ofHistory, Indianapolis: The Bobbs-Merrill Company, 1973.Raub, W. (1982) ‘The Structural-Individualistic Approach towards an Explanatory Sociology’,pp. 3–40 in W. Raub (ed.) Theoretical Models and Empirical Analyses, Utrecht: E.S.Publications.Raub, W. and Voss, T. (1981) Individuelles Handeln und gesellschaftliche Folgen. Das individualistischeProgramm in den Sozialwissenschaften, Darmstadt: Herman Luchterhand.Rehberg, K.-S. (1994) ‘Kulturwissenschaft und Handlungsbegrifflichkeit. AnthropologischeÜberlegungen zum Zusammenhang von Handlung und Ordnung in derSoziologie Max Webers’, pp. 602–61 in G. Wagner and H. Zipprian (eds) Max WebersWissenschaftslehre. Interpretation und Kritik, Frankfurt am Main: Suhrkamp.Ricardo, D. (1817) The Principles of Political Economy and Taxation, London: J.M. Dent &Sons, 1973.Richardson, G.B. (1959) ‘Equilibrium, Expectations and Information’, The EconomicJournal, 69, 223–37.Rickert, H. (1902) Science and History. A Critique of Positivist Epistemology, Princeton. NJ: D.van Nostrand Company, 1962.—— (1902) The Limits of Concept Formation in Natural Science. A Logical Introduction to theHistorical Sciences, Cambridge: Cambridge University Press, 1986.Rickman, H.P. (1962) ‘General Introduction’ (pp. 11–63) to W. Dilthey, Pattern and Meaningin History, New York: Harper & Row.—— (1967) Understanding and Human Studies, London: Heinemann.—— (1976) ‘Introduction’ (pp. 1–31) to W. Dilthey, Selected Writings, Cambridge:Cambridge University Press.—— (1979) Wilhelm Dilthey. Pioneer of the Human Studies, Berkeley: University of CaliforniaPress.Riker, W.H. (1988) ‘The Place of Political Science in Public Choice’, Public Choice, 57,247–57.Riker, W.H. and Ordeshook, P.C. (1973) An Introduction to Positive Political Theory, EnglewoodCliffs, NJ: Prentice-Hall.Riley, M.W. (1963) Sociological Research I. A Case Approach, New York: Harcourt, Brace &World.Ringer, F. (1997) Max Weber’s Methodology. The Unification of the Cultural and SocialSciences, Cambridge, MA: Harvard University Press.Ritzel, G. (1950) Schmoller versus Menger. Eine Analyse des Methodenstreits im Hinblick auf denHistorismus in der Nationaloekonomie. Frankfurt am Main: Dissertation.Rizvi, A.T. (1994) ‘The Microfoundations Project in General Equilibrium Theory’,Cambridge Journal of Economics, 18, 357–77.Robbins, L. (1932) An Essay on the Nature and Significance of Economic Science, London:Macmillan, 1935.Robinson, J. (1962) Economic Philosophy, Harmondsworth: Penguin, 1994.

424 Bibliography—— (1972) ‘The Second Crisis of Economic Theory’, American Economic Review, 62,supplement 1–10.Robinson, W.S. (1950) ‘Ecological Correlations and the Behavior of Individuals’, AmericanSociological Review, 15, 351–57.Roemer, J. (1981) Analytical Foundations of Marxian Economic Theory, Cambridge: CambridgeUniversity Press.—— (1982a) A General Theory of Exploitation and Class, Cambridge, MA: Harvard UniversityPress.—— (1982b) ‘Methodological Individualism and Deductive Marxism’, Theory and Society,11, 513–20.—— (1982c) ‘New Directions in the Marxian Theory of Exploitation and Class’, Politicsand Society, 11, 253–87.—— (1986) “‘Rational Choice” Marxism: Some Issues of Method and Substance’, pp.191–202 in J. Roemer (ed.) Analytical Marxism, Cambridge: Cambridge UniversityPress.—— (1989) ‘Marxism and Contemporary Social Science’, Review of Social Economy, 47,377–91.Rogers, M.F. (1983) Sociology, Ethnomethodology, and Experience, Cambridge: CambridgeUniversity Press.Roscher, W. (1843) Grundriss zu Vorlesungen über die Staatswirtschaft. Nach geschichtlicher Methode,Göttingen: Dieterichschen Buchhandlung.Rose, A.M. (1972) ‘A Systematic Summary of Symbolic Interaction Theory’, pp. 3–19 inA.M. Rose (ed.) Human Behavior and Social Processes, London: Routledge & Kegan Paul.Rosenberg, A. (1988) Philosophy of Social Science, Oxford: Clarendon Press.Ross, E.A. (1908) Social Psychology, New York: Macmillan.Roth, G. (1976) ‘History and Sociology in the Work of Max Weber’, The British Journal ofSociology, 27, 306–18.Rothbard, M.N. (1962) Man, Economy, and State. A Treatise on Economic Principles, LosAngeles: Nash Publishing.—— (1979) Individualism and the Philosophy of the Social Sciences, San Fransisco: Cato Institute.—— (1989) ‘The Hermeneutical Invasion of Philosophy and Economics’, The Review ofAustrian Economics, 3, 45–59.Ruben, D.-H. (1982) ‘The Existence of Social Entities’, The Philosophical Quarterly, 32,295–310.—— (1985) The Metaphysics of the Social World, London: Routledge and Kegan Paul.Rudner, R.S. (1954) ‘Philosophy and Social Science’, Philosophy of Science, 21, 164–8.—— (1966) Philosophy of Social Science, Englewood Cliffs, NJ: Prentice-Hall.Runciman, W.G. (1972) A Critique of Max Weber’s Philosophy of Social Science, Cambridge:Cambridge University Press.Ruskin, J. (1857–62) Unto This Last, London: J.M. Dent, 1968.Russell, B. (1905) ‘On Denoting’, Mind, 56, 479–93.—— (1912) The Problems of Philosophy, London: Oxford Univeristy Press, 1951.—— (1914) Our Knowledge of the External World, London: George Allen & Unwin, 1926.—— (1917) Mysticism and Logic, London: Unwin Books, 1963.—— (1918) ‘The Philosophy of Logical Atomism’, pp. 31–142 in D. Pears (ed.) Russell’sLogical Atomism, London: Fontana/Collins, 1972.—— (1924) ‘Logical Atomism’, pp. 143–65 in D. Pears (ed.) Russell’s Logical Atomism,London: Fontana/Collins.

Bibliography 425—— (1940) An Inquiry into Meaning and Truth, Harmondsworth: Penguin, 1973.—— (1948) Human Knowledge: Its Scope and Limit, London: George Allen & Unwin.—— (1959) My Philosophical Development, London: George Allen & Unwin.Rutherford, M. (1994) Institutions in Economics. The Old and the New Institutionalism,Cambridge: Cambridge University Press.Ryan, A. (1970) The Philosophy of John Stuart Mill, London: Macmillan.—— (1974) J.S. Mill, London: Routledge and Kegan Paul.—— (1975) The Philosophy of the Social Sciences, London: Macmillan.—— (1991) ‘Exploitation, Justice, and the Rational Man’, pp. 29–49 in J.G.T. Meeks (ed.)Thoughtful Economic Man. Essays on Rationality, Moral Rules and Benevolence, Cambridge:Cambridge University Press.Ryle, G. (1949) The Concept of Mind, Harmondsworth: Penguin, 1963.Sacks, H. (1972) ‘On the Analysability of Stories by Children’, pp. 216–32 in R. Turner(ed.) Ethnomethodology, Harmondsworth: Penguin, 1974.Samuelson, P.A. (1946) ‘Lord Keynes and the General Theory’, Econometrica, 14, 187–200.—— (1947) Foundations of Economic Analysis, Boston: Harvard University Press.—— (1956) ‘Social Indifference Curves’, The Quarterly Journal of Economics, 70, 1–22.—— (1966) ‘Modern Economic Realities and Individualism’, pp. 1407–18 in The CollectedScientific Papers of Paul A. Samuelson, vol. II, ed. J.E. Stiglitz, Cambridge, MA: MITPress.Sappington, D.E.M. (1991) ‘Incentives in Principal-Agent Relationships’, The Journal ofEconomic Perspectives, 5, 45–66.Sarbin, T.R. and Kitsuse, J.I. (eds) (1994) Constructing the Social, London: Sage Publications.Sargent, T.J. (1980) ‘Rational Expectations and the Reconstruction of Macroeconomics’,pp. 31–39 in P.J. Miller (ed.) The Rational Expectations Revolution, Readings from the FrontLine, Cambridge, MA: MIT Press, 1994.Sartre, J.-P. (1943) Being and Nothingness, New York: Philosophical Library, 1956.—— (1947) ‘The Humanism of Existentialism’, pp.31–62 in Essays in Existentialism,Syracuse, NJ: The Citadel Press, 1965.—— (1960) Critique of Dialectical Reason, London: Verso, 1976.Satz, D. and Ferejohn, J. (1994) ‘Rational Choice and Social Theory’, The Journal ofPhilosophy, 91, 71–87.Savigny, F.K. von (1814) Of the Vocation of Our Age for Legislation and Jurisprudence, New York:Arno Press, 1975.Sax, E. (1884) Das Wesen und die Aufgaben der Nationalökonomie, Vienna: Alfred Hölder.Say, J.-B. (1803) A Treatise on Political Economy or The Production, Distribution and Consumption ofWealth, New York: Augustus M. Kelley, 1971.Sayer, A. (1984) Method in Social Science: A Realist Approach, London: Hutchinson.Scaff, L.A. (1994) ‘Max Weber’s Begriff der Kultur’, pp. 678–99 in G. Wagner and H.Zipprian (eds) Max Webers Wissenschaftslehre: Interpretation und Kritik, Frankfurt am Main:Suhrkamp.Schaffner, K.F. (1967) ‘Approaches to Reduction’, Philosophy of Science, 34, 137–47.Scheler, M. (1915) ‘Versuche einer Philosophie des Lebens. Nietzsche – Dilthey –Bergson’, pp. 293–339 in Vom Umsturtz der Werte, Bern: Franke, 1955.Schelling, T.C. (1978) Micromotives and Macrobehavior, New York: W.W. Norton & Company.Schelting, A. von (1922) ‘Die logische Theorie der historischen Kulturwissenschaften vonMax Weber und im besonderen sein Begriff des Idealtypen’, Archiv für Sozialwissenschaftund Sozialpolitik, 49, 623–752.

426 BibliographySchlegloff, E. and Sacks. H. (1973) ‘Openings Up Closings’, pp. 233–64 in R. Turner (ed.)Ethnomethodology, Harmondsworth: Penguin, 1974.Schleiermacher, F. (1809–10) ‘General “Hermenenutics”, pp. 225–68 in Hermeneutics andCriticism and Other Writings, Cambridge: Cambridge University Press, 1998.—— (1819–33) Hermeneutics and Criticism and Other Writings, Cambridge: CambridgeUniversity Press, 1998.Schluchter, W. (1989) Rationalism, Religion, and Domination: A Weberian Perspective, Berkeley:University of California Press.—— (2000) ‘Psychophysics and Culture’, pp. 59–80 in S. Turner (ed.) The CambridgeCompanion to Weber, Cambridge: Cambridge University Press.Schmoller, G. von (1883) ‘Review of Carl Menger, Untersuchungen über die Methode der Sozialwissenschaften’,Jahrbuch für Gesetzgebung, Verwaltung und Volkswirtschaft, 7, 239–58.—— (1884) ‘Carl Menger, Die Irrthümer des Historismus’, Jahrbuch für Gesetzgebung, Verwaltungund Volkswirtschaft, 8, 677.Schön, M. (1987) ‘Gustav Schmoller and Max Weber., in W. Mommsen and J. Osterhammel(eds) Max Weber and His Contemporaries, London: Allen & Unwin.Schotter, A. (1981) The Economic Theory of Social Institutions, Cambridge: CambridgeUniversity Press.—— (1985) Free Market Economics. A Critical Appraisal, New York: St Martin’s Press.Schroeter, G. (1985) ‘Ideal Types: Beacons or Blinders?’, History of Sociology, 6, 161–66.Schumpeter, J.A. (1908) Das Wesen und Hauptinhalt der theoretischen Nationalökonomie, Leipzig:Duncker & Humbolt.—— (1909) ‘On the Concept of Social Value’, Quarterly Journal of Economics, 23, 213–32.—— (1914) Economic Doctrine and Method, London: George Allen & Unwin, 1954.—— (1915) ‘Wie studiert man Sozialwissenschaft?’, pp. 555–65 in Aufsätze zur öknomischenTheorie, Tübingen: J.C.B. Mohr.—— (1920) ‘Max Webers Werk’, pp. 220–9 in The Economics and Sociology of Capitalism,Princeton: Princeton University Press, 1991.—— (1927) ‘Social Classes In An Ethnically Homogenous Environment’, pp. 133–221 inImperialism and Social Classes, New York: Augustus M. Kelley, 1951.—— (1954) A History of Economic Analysis, London: George Allen & Unwin.Schumpeter, J.A. (1991) The Economics and Sociology of Capitalism, Princeton, NJ: PrincetonUniversity Press.Schutz, A. (1932) The Phenomenology of the Social World, London: Heinemann, 1972.—— (1940) ‘Phenomenology and the Social Sciences, pp. 164–86 in M. Farber (ed.)Philosophical Essays in Memory of Edmund Husserl, Cambridge, MA: Harvard UniversityPress.—— (1945a) ‘On Multiple Realities’, Philosophy and Phenomenological Research, 5, 533–75.—— (1945b) ‘Some Leading Concepts of Phenomenology’, Social Research, 12, 77–97.—— (1953) ‘Common-Sense and Scientific Interpretation of Human Action’, Philosophyand Phenomenological Research, 1–37.—— (1954) ‘Concept and Theory Formation in the Social Sciences’, The Journal of Philosophy,51, 257–73.—— (1955) ‘Symbol, Reality and Society’, pp. 135–203 in L. Bryson, L. Finkelstein, H.Hoagland and R.M. Maciver (eds) Symbols and Society (Fourteenth Symposium of theConference on Science, Philosophy and Religion). New York: Harper and Brothers.—— (1957) ‘The Problem of Transcendental Intersubjectivity in Husserl’, pp. 51–91 inCollected Papers vol. III: Studies in Phenomenological Philosophy, The Hague: MartinusNijhoff, 1970.

Bibliography 427—— (1959) ‘Husserl’s Importance for the Social Sciences’, Phenomenologica, 4, 86–98.—— (1960) ‘The Social World and the Theory of Social Action’, Social Research, 27,203–21.—— (1962) Collected Papers I: The Problem of Social Reality, The Hague: Martinus Nijhoff.Schutz, A. and Luckmann, T. (1973) The Structures of the Life-World, Evanston: NorthwesternUniversity Press.Schutz, A. and Parsons, T. ([1940–1] 1978) The Theory of Social Action. The Correspondence ofAlfred Schutz and Talcott Parsons, Bloomington: Indiana University Press.Scott, J. (1991) Social Network Analysis: A Handbook, London: Sage Publications.Scott, K.J. (1960) ‘Methodological and Epistemological Individualism’, The British Journalfor the Philosophy of Science, 11, 331–6.Scott, R.R. and Meyer, J.W. (1994) Institutional Environments and Organizations. StructuralComplexity and Individualism, London: Sage.Secord, P.F. (1986) ‘Social Psychology as a Science’, pp. 128–64 in J. Margolis, P.T.Manicas, R. Harré and P.F. Secord, Psychology. Designing a Discipline, Oxford: BasilBlackwell.Selgin, G.A. (1988) ‘Praxeology and Understanding: An Analysis of the Controversy inAustrian Economics’, pp. 19–58 in M.N. Rothbard and W. Block (eds) The Review ofAustrain Economics, vol. 2, Lexington, MA: Lexington.Senior, N.W. (1836) An Outline of the Science of Political Economy, New York: Augustus M.Kelley, 1965.Senn, P.R. (1958) ‘The Earliest Use of the Term “Social Science”’, pp. 84–6 in J.C. Wood(ed.) John Stuart Mill. Critical Assessments, vol. IV, London: Routledge, 1991.Sensat, J. (1988) ‘Methodological Individualism and Marxism’, Economics and Philosophy, 4,189–219.Sextus Empiricus (1996) Outlines of Pyrrhonism, in B. Mates, The Sceptic Way. Sextus Empiricus’sOutlines of Pyrrhonism, New York: Oxford University Press.Shafer, W. and Sonnenshein, H. (1982) ‘Market Demand and Excess Demand Functions,pp. 672–93 in Arrow, K.J. and Intriligator, M.D. (eds) Handbook of MathematicalEconomics, vol. 2, Amsterdam: North-Holland.Shand, A.H. (1990) Free Market Morality. The Political Economy of the Austrian School, London:Routledge.Shapere, D. (1977) ‘Scientific Theories and Their Domains’, pp. 518–65 in F. Suppe (ed.)The Structure of Scientific Theories, Urbana, IL: University of Illinois Press.Sharrock, W. and Andersson, B. (1986) The Etnomethodologists, London: Tavistock Publications.Sharrock, W. and Button, G. (1991) ‘The Social Actor: Social Action in Real Time’, pp.137–75 in G. Button (ed.) Ethnomethodology and the Human Sciences, Cambridge:Cambridge University Press.Shepsle, K. (1989) ‘Studying Institutions. Some Lessons from the Rational ChoiceApproach’, Journal of Theoretical Politics, 1, 131–47.Shibutani, T. (1961) Society and Personality, Englewood Cliffs, NJ: Prentice-Hall.—— (1962) ‘Reference Groups and Social Control’, pp. 128–47, in A.M. Rose (ed.)Human Behavior and Social Processes. An Interactionist Approach, London: Routledge andKegan Paul.Shubik, M. (1964) ‘Game Theory and the Study of Social Behavior: An IntroductoryExposition’, pp. 1–77 in M. Shubik (ed.) Game Theory and Related Approaches to SocialBehavior, New York: John Wiley.

428 Bibliography—— (1975) ‘The General Equilibrium Model is Incomplete and not Adequate for theReconciliation of Micro and Macroeconomic Theory’, Kyklos, 28, 545–73.—— (1982) Game Theory in the Social Sciences. Concepts and Solutions, Cambridge, MA: MITPress.Sidgwick, H. (1874) The Methods of Ethics, Indianapolis: Hackett, 1981.—— (1883) The Principles of Political Economy, London: Macmillan, 1901.—— (1885) The Scope and Method of Economic Science, London: Macmillan.Silverman, P. (1990) ‘The Cameralist Roots of Menger’s Achievement’, pp. 69–92 in B.J.Caldwell (ed.) Carl Menger and his Legacy in Economics, Durham: Duke University Press.Simmel, G. (1895) ‘The Problem of Sociology’, Annals of the American Academy of Politicaland Social Science, 6, 52–63.—— (1892) The Problems of the Philosophy of History, New York: The Free Press, 1977.—— (1900) The Philosophy of Money, London: Routledge and Kegan Paul, 1978.—— (1908) ‘How is Society Possible’, pp. 6–22 in Georg Simmel on Individuality and SocialForms, Chicago: Chicago University Press, 1971.—— ([1908] 1909) ‘The Problem of Sociology’, The American Journal of Sociology, 15,289–320.—— (1908c) ‘Ueber das Wesen der Sozial-Psychologie’, Archiv für Sozialwissenschaft undSozialpolitik, 26, 285–91.—— (1908) The Sociology of Georg Simmel, New York: The Free Press, 1950.—— (1917) ‘Individual and Society in Eighteenth- and Nineteenth-Century Views ofLife (An Example of Philosophical Sociology)’, pp. 58–84 in The Sociology of GeorgSimmel, New York: The Free Press, 1950.—— (1917–18) ‘The Constitutive Concepts of History’, pp. 145–97 in Essays on Interpretationin Social Science, Totowa, NJ: Rowman and Littlefield, 1980.—— (1918) ‘On the Nature of Historical Understanding’, pp. 97–127 in Essays on Interpretationin Social Science, Totowa, NJ: Rowman and Littlefield, 1980.Simon, H.A. (1978) The Sciences of the Artificial, Cambridge, MA: MIT Press.—— (1991) ‘Organizations and Markets’, Journal of Economic Perspectives, 5, 25–44.Sjöstrand, S.-E. (1993) ‘On Institutional Thought in the Social and Economic Sciences’,pp. 3–34 in S.-E. Sjöstrand (ed.) Institutional Change. Theory and Empirical Findings,Armonk, NY: M.E. Sharp.Skidmore, W. (1979) Theoretical Thinking in Sociology, Cambridge: Cambridge UniversityPress, 1979.Skinner, B.F. (1965) Science and Human Behavior, New York: The Free Press.—— (1969) Contingencies of Reinforcement, Englewwod Cliffs, NJ: Prentice-Hall.——(1974) Beyond Freedom and Dignity, Harmondsworth: Penguin.Sklar, L. (1967) ‘Types of Inter-Theoretic Reduction’, The British Journal for the Philosophy ofScience, 18, 109–24.Small, A. (1907) Adam Smith and Modern Sociology, Clifton: Augustus M. Kelley, 1972.—— (1924) Origins of Sociology, New York: Russell and Russell, 1967.Smith, A. (1759) The Theory of Moral Sentiments, Indianapolis: Liberty Press, 1976.—— (1776) The Wealth of Nations, New York: The Modern Library, 1937.Smith, B. (1986) ‘Austrian Economics and Austrian Philosophy’, pp. 1–36 in W. Grassland B. Smith (eds) Austrian Economics, London: Croom Helm.—— (1990) ‘Aristotle, Menger, Mises: An Essay in the Metaphysics of Economics’, pp.263–88 in B.J. Caldwell (ed.) Carl Menger and his Legacy in Economics, Durham: DukeUniversity Press.

Bibliography 429Smith, P. (1992) ‘Modest Reductions and the Unity of Science’, pp. 19–44 in D. Charlesand K. Lennon (eds) Reduction, Explanation and Realism, Oxford: Clarendon.Smuts, J.C. (1926) Holism and Evolution, London: Macmillan, 1936.Solow, R. (1989) ‘The State of Economic Science’, pp. 25–39 in W. Sichel (ed.) The State ofEconomic Science. Views of Six Nobel Laureates, Kalamazoo, Michigan: W.E. Upjohn Institute.Sombart, W. (1923) ‘Die Anfänge der Soziologie’, pp. 3–19 in M. Palyi (ed.) Hauptproblemeder Soziologie. Erinnerungsgabe für Max Weber, Munich: Duncker & Humbolt.Song, H.-H. (1995) ‘Adam Smith as an Early Pioneer of Institutional Individualism’,History of Political Economy, 27, 425–48.Sowell, T. (1974) Classical Economics Reconsidered, Princeton, NJ: Princeton University Press.Spann, O. (1905) ‘Untersuchungen über den Gesellschaftsbegriff zur Einleitung in dieSoziologie’, Zeitschrift für Staatswissenschaft, 59, 302–44, 427–60.—— (1923) ‘Bemerkungen zu Max Webers Soziologie’, Zeitschrift für Volkswirtschaft undSozialpolitik, Neue Folge, 3, 761–70.Spencer, H. (1860) ‘The Social Organism’, pp.383–434 in H. Spencer, The Man Versus theState, Indianapolis: Liberty Classics, 1982.—— (1864) Reasons for Dissenting From the Philosophy of M. Comte, London: Williams &Northgate, 1984.—— (1873) The Study of Sociology, Ann Arbor: University of Michigan Press, 1961.—— (1876) Principles of Sociology, London: Williams & Northgate, 1985.—— (1884) The Man Versus the State. With Six Essays on Government, Society, and Freedom, Indianapolis:Liberty Press, 1982.Spinner, H. (1973) ‘Science Without Reduction. A Criticism of Reductionism withSpecial Reference to Hummel’s and Opp’s “Sociology Without Sociology”’, Inquiry,16, 16–94.Sproule-Jones, M. (1984) ‘Methodological Individualism. Challenge and Response’, AmericanBehavioral Scientist, 28, 167–83.Stern, F. (1956) Varieties of History. From Voltaire to the Present, London: Macmillan, 1970.Stigler, G.J. (1950) ‘The Development of Utility Theory. I-II’, The Journal of PoliticalEconomy, 58, 307–25, 373–96.—— (1961) ‘The Economics of Information’, The Journal of Political Economy, 69, 213–25.—— (1984) ‘Economics – The Imperial Science’, The Scandinavian Journal of Economics, 86,301–13.Stigler, G.J. and Becker, G.S. (1977) ‘De Gustibus Non Est Disputandum’, AmericanEconomic Review, 67, 76–90.Strauss, A.L. (1956) ‘Introduction’ to G.H. Mead, On Social Psychology, Chicago: ChicagoUniversity Press, 1964.—— (1969) Mirrors and Masks. The Search for Identity, London: Martin Robertson, 1977.Strauss, A.L., Schatzman, L., Ehrlich, D., Bucher, R. and Sabshin, M. (1963) ‘TheHospital and Its Negotiated Order’, pp. 147–69 in E. Friedson (ed.) The Hospital inModern Society, New York: The Free Press.—— (1964) Psychiatric Ideologies and Institutions, New York: The Free Press.Strawson, P.F. (1962) ‘Freedom and Resentment’, pp. 71–96 in P.F. Strawson (ed.) Studies inthe Philosophy of Thought and Action, Oxford: Oxford University Press, 1968.—— (1970) Meaning and Truth, Oxford: Oxford University Press.Streissler E. (ed.) (1969) Roads to Freedom: Essays in Honour of Friedrich von Hayek, London:Routledge & Kegan Paul.

430 BibliographyStreissler, E.W. (1990) ‘The Influence of German Economics on the Work of Menger andMarshall’, pp. 31–68 in B.J. Caldwell (ed.) Carl Menger and his Legacy in Economics,Durham: Duke University Press.Stryker, S. (1980) Symbolic Interactionism. A Social Structural Version, Menlo Park, CA:Benjamin/Cummings.Sugden, R. (1986) The Economics of Rights, Co-operation and Welfare, Oxford: Blackwell.—— (1991) ‘Rational Choice: A Survey of Contributions from Economics and Philosophy’,The Economic Journal, 101, 751–85.—— (1993) ‘Normative Judgements and Spontaneous Order: The ContractarianElement in Hayek’s Thought’, Constitutional Political Economy, 4, 393–424.Sumner, W.G. (1992) On Liberty, Society, and Politics, Indianapolis: Liberty Fund.Suppe, F. (1973) ‘The Search for Philosophical Understanding of Scientific Theories’, pp.3–241 in F. Suppe (ed.) The Structure of Scientific Theories, Urbana: University of IllinoisPress; 2nd edn, 1977.Swedberg, R. (1991) Joseph A. Schumpeter. His Life and Work, Cambridge: Polity Press.—— (1994) ‘Markets as Social Structures’, pp. 255–82 in N.J. Smelser and R. Swedberg(eds) The Handbook of Economic Sociology, Princeton: Princeton University Press.—— (1996) ‘Analysing the Economy’: On the Contribution of James S. Coleman’, pp.313–28 in J. Clark (ed.) James Coleman, London: Falmer Press.—— (1998) Max Weber and the Idea of Economic Sociology, Princeton, NJ: Princeton UniversityPress.Swingewood, A. (1970) ‘Origins of Sociology: the Case of the Scottish Enlightenment’,The British Journal of Sociology, 21, 164–80.Sztompka, H. (1979) Sociological Dilemmas, New York: Academic Press.Tajfel, H. (1972) ‘Experiments in a Vacuum’, pp. 69–119 in J. Israel and H. Tajfel (eds)The Context of Social Psychology. A Critical Assessment, London: Academic Press.Tännsjö, T. (1990) ‘Methodological Individualism’, Inquiry, 33, 69–80.Tarde, G. (1890) Laws of Imitation, Gloucester, MA: Peter Smith, 1962.Tarde, G. (1894) ‘Sociology, Social Psychology and Sociologism’, pp. 112–35 in G. Tarde,On Communication and Social Influence, Chicago: Chicago University Press, 1969.—— (1898) Etude de Psychologie Sociale, pp. 73–105 (selections) in G. Tarde, On Communicationand Social Influence, Chicago: Chicago University Press, 1969.—— (1904) ‘ A Debate with Emile Durkheim’, pp. 136–41 in G. Tarde, On Communicationand Social Influence, Chicago: Chicago University Press, 1969.Taylor, M. (1986) ‘Elster’s Marx’, Inquiry, 29, 3–10.—— (1988) ‘Rationality and Revolutionary Collective Action’, pp. 63–97 in M. Taylor(ed.) Rationality and Revolution, Cambridge: Cambridge University Press.Taylor, M.W. (1989) ‘The Errors of an Evolutionist: A Reply to Ellen Frankel Paul’, PoliticalStudies, 37, 436–442.—— (1992) Man versus the State, Oxford: Clarendon Press.Tenbruck, F.H. von (1959) ‘Die Genesis der Methodologie Max Webers’, Kölner Zeitschriftfür Soziologie und Sozialpsychologie, 11, 573–628.—— (1986) ‘Das Werk Max Webers: Methodologie und Sozialwissenschaften’, KölnerZeitschrift für Soziologie und Sozialpsychologie, 38, 13–31.—— (1989) ‘Abschied von der “Wissenschaftslehre”?’, pp. 90–115 in J. Weiss (ed.) MaxWeber heute, Frankfurt am Main: Suhrkamp.—— (1994) ‘Die Wissenschaftslehre Max Webers. Voraussetzungen zu ihremVerständnis’, pp. 367–89 in G. Wagner and H. Zipprian (eds) Max Webers Wissenschaftslehre.Interpretation und Kritik, Frankfurt am Main: Suhrkamp.

Bibliography 431Therborn, G. (1976) Science, Class and Society: On the Formation of Sociology and Historical Materialism,London: Verso, 1980.Thomas, D. (1979) Naturalism in Social Science, Cambridge: Cambridge University Press.Thomas, G.M., Meyer, J.W., Ramirez, F.O. and Boli, J. (1987) Institutional Structure. ConstitutingState, Society, and the Individual, Newbury Park: Sage Publications.Thomas, W.I. (1923) The Unadjusted Girl, New York: Harper & Row, 1967.—— (1927) ‘Situational Analysis: The Behavior Pattern and the Situation’, pp. 154–67 inOn Social Organization and Social Psychology, Chicago: Chicago University Press, 1966.Thomas, W.I. and Thomas, D.S. (1929) The Child in America, New York: A. Knopf.Thomas, W.I. and Znaniecki, F. (1918) The Polish Peasant in Europe and America, 2 vols., NewYork: Alfred Knopf, 1927.Tiles, J.E. (1984) ‘On a New Methodological Individualism’, Ratio, 36, 71–79.Tinbergen, J. (1940) ‘On a Method of Statistical Business-Cycle Research: A Reply’, TheEconomic Journal, 1, 141–54.Tiryakian, E.A. (1965) ‘Existential Phenomenology and the Sociological Tradition’, AmericanSociological Review, 30, 674–88.—— (1973) Sociology and Existential Phenomenology’, pp. 187–222 in M. Natanson(ed.) Phenomenology and the Social Sciences, vol. 1, Evanston: Northwestern UniversityPress.Titchener, E.B. (1921) ‘Brentano and Wundt: Empirical and Experimental Psychology’,American Journal of Psychology, 32, 108–20.Torrance, J. (1974) ‘Max Weber: Methods and the Man’, Archives Européennes de Sociologie,15, 127–65.Tribe, K. (1988) Governing Economy. The Reformation of German Economic Discourse 1750–1840,Cambridge: Cambridge University Press.Troeltsch (1924) Der Historismus und seine Überwindung, Berlin: Pan Verlag Rolf Heise.Tullock, G. (1972) ‘Economic Imperialism’, pp. 317–29 in J.M. Buchanan and R.D.Tollison (eds) Theory of Public Choice. Political Applications of Economics, Ann Arbor:University of Michigan Press.Turner, J.H. (1974) ‘Parsons as a Symbolic Interactionist: A Comparison of Action andInteraction Theory’, Sociological Inquiry, 44, 283–94.—— (1975) ‘Action and Interaction Theory: Some Questions of Theory Building andStrategy – A Reply’, Sociological Inquiry, 45, 65–8.—— (1987) ‘Social Exchange Theory. Future Directions’, pp. 223–38 in K.S. Cook (ed.)Social Exchange Theory, Newbury Park: Sage Publications.Turner, R. (1970) ‘Words, Utterances, and Activities’, pp. 197–215 in R. Turner (ed.)Ethnomethodology, Harmondsworth: Penguin, 1974.Turner, R.H. (1956) ‘Role-Taking, Role Standpoint, and Reference-Group Behavior’, TheAmerican Journal of Sociology, 61, 316–28.—— (1962) ‘Role-Taking: Process versus Conformity’, pp. 20–40 in A.R. Rose (ed.)Human Behavior and Social Process, London: Routledge & Kegan Paul.—— (1978) ‘The Role and the Person’, The American Journal of Sociology, 84, 1–23.Tyrell, H. (1994) ‘Max Webers Soziologie – eine Soziologie ohne “Gesellschaft”’, pp.390–414 in G. Wagner und H. Zipprian (eds) Max Webers Wissenschaftslehre. Interpretationund Kritik, Frankfurt am Main: Suhrkamp.Udehn, L. (1987) Methodological Individualism – A Critical Appraisal, Uppsala: Department ofSociology, Dissertation.—— (1991) ‘The Conflict between Methodology and Rationalization Thesis in the Workof Max Weber’, Acta Sociologica, 24, 131–47.

432 Bibliography—— (1992) ‘The Limits of Economic Imperialism’, pp. 239–80 in U. Himmelstrand (ed.)Interfaces in Economic and Social Analysis, London: Routledge.—— (1996) The Limits of Public Choice. A Sociological Critique of the Economic Theory of Politics,London: Routledge.Ullman-Margalit, E. (1977) The Emergence of Norms, Oxford: Clarendon.Urmson, J.O. (1958) Philosophical Analysis. Its Development Between the Two World Wars,Oxford: Oxford University Press.Vanberg, V. (1975) Die Zwei Soziologien, Tübingen: J.B.C. Mohr.—— (1986) ‘Spontaneous Market Order and Social Rules. A Critical Examination ofF.A. Hayek’s Theory of Cultural Evolution’, Economics and Philosophy, 2, 75–100. (Alsoin Vanberg, 1994: ch 5).—— (1994) Rules and Choice in Economics, London: Routledge.Van Parijs, P. (1981) ‘Sociology as General Economics’, Archives Européennes de Sociologie, 22,299–324.Veblen, T. (1898) ‘Why Is Economics Not an Evolutionary Science?’, The Quarterly Journalof Economics, 12, 373–97.Verhoeven, J. (1985) ‘Goffman’s Frame Analysis and Modern Micro-sociologicalParadigms’, pp. 71–100 in H.J. Helle and S.N. Eisenstadt (eds) Microsociological Theory.Perspectives on Sociological Theory, vol. 2, London: Sage Publications.Vining, R. (1949a) ‘Methodological Issues in Quantitative Economics’, The Review ofEconomics and Statistics, 31, 77–86.—— (1949b) ‘A Rejoinder’ to T.C. Koopmans, ‘A Reply’, The Review of Economics andStatistics, 31, 91–94.Vromen, J.J. (1995) Economic Evolution. An Enquiry into the Foundations of New InstitutionalEconomics, London: Routledge.Wagner, H.R. (1983) Alfred Schutz. An Intellectual Biography, Chicago: University of ChicagoPress.Walras, L. (1874) Elements of Pure Economics, or the Theory of Social Wealth, Philadelphia:Orion Editions, 1984.Walther, A. (1926) ‘Max Weber als Soziologe’, Jahrbuch für Soziologie, Zweiter Band, 1–65.Ward, H. (1995) ‘Rational Choice Theory’, pp. 76–93 in D. Marsh and G. Stoker (eds)Theory and Methods in Political Science, London: Macmillan.Wärneryd, K. (1990) Economic Conventions. Essays in Institutional Evolution, Stockholm:Economic Research Institute.Warren, M. (1988) ‘Marx and Methodological Individualism’, The Philosophy of the SocialSciences, 18, 447–76.Watkins, J.W.N. (1952a) ‘Ideal Types and Historical Explanation’, The British Journal for thePhilosophy of Science, 3, 22–43.—— (1952b) ‘The Principle of Methodological Individualism’, The British Journal for thePhilosophy of Science, 3, 186–89.—— (1953) ‘Ideal Types and Historical Explanation’, pp. 723–43 in H. Feigl and M.Brodbeck (eds) Readings in the Philosophy of Science, New York: Appleton-Century Crofts.—— (1955a) ‘Methodological Individualism: A Reply’, Philosophy of Science, 22, 58–62.—— (1955b) ‘Philosophy and Politics in Hobbes’, The Philosophical Quarterly, 5, 125–46.—— (1957a) ‘Between Analytic and Empirical’, Philosophy, 32, 112–31.—— (1957b) ‘Historical Explanation in the Social Sciences’, The British Journal for thePhilosophy of Science, 8, 104–17.—— (1957/8) ‘Epistemology and Politics’, Proceedings of the Aritotelian Society, New Series, 58,79–102.

Bibliography 433——(1958) ‘Confirmable and Influential Metaphysics’, Mind, 67, 344–65.—— (1959a) ‘The Alleged Inadequacy of Methodological Individualism’, The Journal ofPhilosophy, 55, 392–5.—— (1959b) ‘The Two Theses of Methodological Individualism’, The British Journal for thePhilosophy of Science, 9, 319–20.—— (1959c) ‘Third Reply to Mr Goldstein’, The British Journal for the Philosophy of Science,10, 242–44.—— (1965) Hobbes’s System of Ideas, London: Hutchinson, 1973.—— (1970) ‘Imperfect Rationality’ in R. Borger and F. Cioffi (eds) Explanation in theBehavioural Sciences, Cambridge: Cambridge University Press.—— (1974) ‘The Unity of Popper’s Thought’, pp. 371–412 in P.A. Schilpp (ed.) ThePhilosophy of Karl Popper, 2 vols, La Salle: Open Court.—— (1976a) ‘The Human Condition. Two Criticisms of Hobbes’, pp. 691–716 in R.S.Cohen (ed.) Essays in Memory of Imre Lakatos, Dordrecht: D. Reidel.—— (1976b) ‘Review of F.A. Hayek’s Law, Legislation and Liberty. Volume 1: Rules andOrder’, The Philosophy of the Social Sciences, 6, 369–72.Watson, J.B. (1913) ‘Psychology as the Behaviorist Views It’, Psychological Review, 20,158–77.—— (1924) Behaviorism, New York: W.W. Norton & Company, 1930.Weber, Marianne (1926) Max Weber: A Biography, New Brunswick: Transaction Books,1988.Weber, M. (1895) ‘The Nation State and Economic Policy [Freiburg adress]’, Economy andSociety, 9, 1980, 420–49.—— (1898) Grundriss zu den Vorlesungen über Allgemeine (‘theoretische’) Nationalökonomie,Tübingen: J.C.B. Mohr, 1990.—— (1903–6) Roscher and Knies. The Logical Problems of Historical Economics, New York: TheFree Press, 1975.—— (1904–5) The Protestant Ethic and the Spirit of Capitalism, London: Unwin, 1930.—— (1908a) ‘Marginal Utility and “The Fundamental Law of Psychophysics”’, SocialScience Quarterly, 56, 21–36, 1975.—— (1908b) ‘A Research Strategy for the Study of Occupational Careers and MobilityPatterns’, pp. 103–55 in J.E.T. Eldridge (ed.) Max Weber, London: Nelson, 1972.—— ([1910] 1971) ‘Max Weber on Race and Society, Social Research, 38, 31–41.—— ([1913] 1981) ‘Essay on Some Categories of Interpretive Sociology’, The SociologicalQuarterly, 22, 145–80.—— (1922) Economy and Society, 2 vols, Berkeley: University of California Press.1978.—— (1927) General Economic History, New Brunswick: Transaction Books, 1981.—— (1949) The Methodology of the Social Sciences, New York: The Free Press.—— (1972) ‘Georg Simmel as a Sociologist’, Social Research, 39, 155–63.Webster, M. (1973) ‘Psychological Reductionism, Methodological Individualism, andLarge-Scale Problems’, The American Sociological Review, 258–73.Weintraub, E.R. (1979) Microfoundations. The Compatibility of Microeconomics and Macroeconomics,Cambridge: Cambridge University Press.Weiss, J. (1975) Max Webers Grundlegung der Soziologie, Munich: Verlag Dokumentation.Weldes, J. (1989) ‘Marxism and Methodological Individualism. A Critique’, Theory andSociety, 18, 353–86.Wellman, B. and Berkowitz, S.D. (eds) (1988) Social Structures: A Network Approach,Cambridge: Cambridge University Press.

434 BibliographyWhitaker, J.K. (1975) ‘The Evolution of Alfred Marshall’s Economic Thought and Writingsover the the Years 1867–90’, Introduction to J.K. Whitaker (ed.) The Early Writingsof Alfred Marshall, 1867–1890, London: Macmillan Press.White, H. (1992) Identity and Control. A Structural Theory of Social Action, Princeton, NJ:Princeton Univeristy Press.White, H., Boorman, S.A. and Breiger, R.L. (1976a) ‘Social Structure from MultipleNetworks. I. Blockmodels of Roles and Positions’, American Journal of Sociology, 81,730–80.—— (1976b) ‘Social Structure from Multiple Networks. II. Role Structures’, AmericanJournal of Sociology, 81, 1384–1446.Wicksteed, P.H. (1888) The Alphabet of Economic Science, New York: Augustus M. Kelley,1970.—— (1910) The Common Sense of Political Economy, London: Routledge & Kegan Paul,1933.—— (1914) ‘The Scope and Method of Political Economy in the Light of the “Marginal”Theory of Value and Distribution’, pp. 772–96 in The Common Sense of Political Economyand Selected Papers and Reviews on Economic Theory, London: Routledge & Kegan Paul,1933.Wieser, F. von (1884) Ursprung und Hauptgesetze des wirtschaftlischen Wertes, Vienna: AlfredHölder.—— (1893) Natural Value, New York: Augustus M. Kelley, 1971.—— (1914) Social Economics, New York: Augustus M. Kelley, 1967.Wilber, C.K. and Harrison, R.S. (1978) ‘The Methodological Basis of InstitutionalEconomics: Pattern Model, Story Telling, and Holism’, Journal of Economic Issues, 12,61–89.Wiley, N. (1994) The Semiotic Self, Cambridge: Polity Press.Willer, D. (1967) Scientific Sociology. Theory and Method, Englewood Cliffs, NJ: Prentice-Hall.—— (1992) ‘The Principle of Rational Choice and the Problem of a SatisfactoryTheory’, pp. 49–78 in J.S. Coleman and T.J. Fararo (eds) Rational Choice Theory. Advocacyand Critique, Newbury Park: Sage Publications.Willer, D. and Andersson, B. (1981) ‘Introduction’, pp. 1–21 in D. Willer and B. Andersson(eds) Networks, Exchange and Coercion. The Elementary Theory and its Application, NewYork: Elsevier.Willer, D. and Skvoretz, J. (1997) ‘Games and Structures’, Rationality and Society, 9, 5–35.Willer, D. and Willer, J. (1973) Systematic Empiricism. Critique of a Pseudoscience, EnglewoodCliffs, NJ: Prentice-Hall.Willey, T.E. (1978) Back to Kant. The Revival of Kantianism in German Social and HistoricalThought, 1860–1914, Detroit: Wayne State University Press.Williamson, O.E. (1975) Markets and Hierarchies: Analysis and Antitrust Implications, New York:The Free Press.—— (1979) ‘Transaction-Cost Economics: The Governance of Contractual Relations’,The Journal of Law and Economics, 22, 233–261.—— (1981) ‘The Economics of Organization: The Transaction Cost Approach’, TheAmerican Journal of Sociology, 87, 548–77.—— (1985) The Economic Institutions of Capitalism, New York: The Free Press.—— (1986) ‘The Economics of Governance: Framework and Implications, pp. 171–202in R.N. Langlois (ed.) Economics as a Process. Essays in the New Institutional Economics,Cambridge: Cambridge University Press.

Bibliography 435—— (1988) ‘The Logic of Economic Organization’, The Journal of Law, Economics, andOrganization, 4, 65–94.—— (1993) ‘Comparative Economic Organization: The Analysis of Discrete StructuralAlternatives’, pp. 75–108 in S.-E. Sjöstrand (ed.) Institutional Change. Theory and EmpiricalFindings, Armonk, NY: M.E. Sharpe.—— (1995) ‘Transaction Cost Economics and Organization Theory’, pp. 207–56 in O.E.Williamson (ed.) Organization Theory. From Chester Barnard to the Present and Beyond, NewYork: Oxford University Press.Wilson, F. (1998) ‘Mill on Psychology and the Moral Sciences’, pp. 203–54 in J. Scorupski(ed.) The Cambridge Companion to Mill, Cambridge: Cambridge University Press.Wilson, T.P. (1971) ‘Normative and Interpretive Paradigms in Sociology’, pp. 57–79 inJ.D. Douglas (ed.) Understanding Everyday Life, London: Routledge & Kegan Paul.Winch, P. (1958) The Idea of a Social Science, London: Routledge & Kegan Paul.Windelband, W. ([1894] 1980) ‘History and Natural Science’, History and Theory, 19,165–85.Winslow, E.G. (1989) ‘Organic Interdependence, Uncertainty and Economic Analysis’,The Economic Journal, 99, 1173–82.Wippler, R. (1978a) ‘The Structural-Individualistic Approach in Dutch Sociology’, TheNetherlands Journal of Sociology, 14, 135–55.—— (1978b) ‘Nicht-intendierte soziale Folgen individueller Handlungen’, Soziale Welt, 29,155–79.—— (1978c) ‘The Structural-Individualistic Approach in Dutch Sociology’, The NetherlandsJournal of Sociology, 4, 135–55.—— (1985) ‘Explanatory Sociology: The Devlopment of a Theoretically OrientedResearch Programme’, The Netherlands’ Journal of Sociology, 21, 63–74.Wippler, R. and Lindenberg, S. (1987) ‘Collective Phenomena and Rational Choice’, pp.135–52 in J.C. Alexander, B. Giesen, R. Münch and N.J. Smelser, (eds) The Micro-Macro Link, Berkeley: California University Press.Wisdom, J.O. (1970) ‘Situational Individualism and the Emergent Group-Properties’, pp.271–96 in R. Borger and F. Cioffi (eds) Explanation in the Behavioural Sciences, Cambridge:Cambridge University Press.—— (1971) ‘Science versus the Scientific Revolution’, The Philosophy of the Social Sciences, 1,123–44.—— (1973) ‘The Phenomenological Approach to the Sociology of Knowledge’, ThePhilosophy of the Social Sciences, 3, 257–66.—— (1980) ‘Schemata in Social Science. Part One: Structural and Operational’, Inquiry,23, 445–64.—— (1981) ‘Schemata in Social Science. Part Two: Metatheoretical’, Inquiry, 24, 3–19.Wisdom, J.T. (1931) ‘Logical Constructions (I.)’, Mind, 40, 188–216.—— (1933a) ‘Logical Constructions (IV.) and (V.)’, Mind, 42, 43–66, 186–202.—— (1933b) ‘Ostentation’, Psyche, 13, 164–77.—— (1934) ‘Is Analysis a Useful Method in Philosophy’ II, The Aristotelian Society, supplementaryvol., 13, 65–89.Witt, U. (1987) Individualistische Grundlagen der evolutorishen Ökonomie, Tübingen: J.B.C. Mohr.—— (1991) ‘Reflections on the Present State of Evolutionary Economic Theory’, pp.83–102 in G.M. Hodgson and E. Screpanti (eds) Rethinking Economics. Markets, Technologyand Economic Evolution, Aldershot: Edward Elgar.Wolff, K.H. (1979) ‘Phenomenology and Sociology’, pp. 499–556 in T. Bottomore and R.Nisbet (eds) A History of Sociological Analysis, London: Heinemann.

436 BibliographyWright, G.H. von (1971) Explanation and Understanding, London: Routledge & Kegan Paul.Wrong, D. (1961) ‘The Oversocialized Conception of Man in Modern Sociology’, AmericanSociological Review, 26, 183–93.—— (1970) ‘Introduction’ (pp. 1–76) to D. Wrong (ed.) Max Weber, Englewood Cliffs, NJ:Prentice-Hall.Wundt, W. (1874) Grundzüge der physiologischen Psychologie, Leipzig: Wilhelm Engelmann.—— (1883) Logik. Eine Untersuchung der Principien der Erkenntniss, vol. 2: Methodenlehre,Stuttgart: Ferdinand Enke.—— (1886) Ethik. Eine Untersuchung der Thatsachen und Gesetze des sittlichen Lebens, Stuttgart:Ferdinand Enke.—— (1895) Logik. Eine Untersuchung der Principien der Erkenntniss und der MethodenWissenschaftlicher Forschung, Stuttgart: Ferdinand Enke, 2nd edn.—— (1921) Logik, Stuttgart: Ferdinand Enke, 4th edn.Xenophon (1979) Oeconomicus, pp. 363–525 in Loeb Classical Library: Xenophon IV,Memorabilia, Oeconomicus, Symposium, Apology, Cambridge, MA: Harvard UniversityPress and London: William Heinemann Ltd.Zetterberg, H. (1963) On Theory and Verification in Sociology, New Jersey: Bedminster Press.Zimmerman, D.H. (1971) ‘The Practicalities of Rule Use’, pp. 221–38 in J.K. Douglas(ed.) Understanding Everyday Life, London: Routledge & Kegan Paul.—— (1978) ‘Ethnomethodology’, The American Sociologist, 13, 6–15.Zimmerman, D.H. and Pollner, M. (1971) ‘The Everyday World as a Phenomenon’, pp.80–103 in J.D. Douglas (ed.) Understanding Everyday Life, London: Routledge & KeganPaul.Zimmerman, D.H. and Wieder, D.L. (1971) ‘Ethnomethodology and the Problem ofOrder’, pp. 285–98 in J.K. Douglas (ed.) Understanding Everyday Life, London: Routledge& Kegan Paul.

Index of author namesAddis, L. 372, 384, 385, 387Adorno, T.W. 186, 373Ahrne, G. 260, 261Albrow, M. 365Alchian, A.A. 378; and Demsetz, H. 271;and Woodward, S. 380Alexander, J.C. 367Allport, F.H. 335Alter, M. 337, 364Anscombe, G.E.M. 387Antoni, C. 365Archer, M.S. 5, 364, 372Arrow, K.J. 228, 238Attewell, P. 370Bergmann, G. 324Bierstedt, R. 372Binns, D. 365Bittner, E. 371Blau, P.M. 184, 358Bleicher, J. 370Blumer, H. 336Bostaph, S. 364Boudon, R. 373Brennan, G. 275; and Tullock, G. 261Broad, C.D. 384Brodbeck, M. 340, 372, 385, 387Brown, R. 381, 387Bryant, C.G.A. 185, 372, 373Buchanan, J.M. 338, 380; and Tullock, G.380Bunge, M. 352, 375, 386Burger, T. 365Burgess, R.L. and Bushell, D. 374Burman, P. 385Button, G. 371Cahnman, W.J. 365Carr, D. 363Carroll, J. 58Chalmers, A.F. 375Charles, D. and Lennon, K. 384Chisholm, R.M. 63, 107Cicourel, Aaron V. 370, 371, 373Clarke, S. 365Coleman, J.S. 180, 373Collin, F. 37Collini, S. et al. 46Collins, R. 126, 158, 370Cook, K.S. 374; and Emerson, R.M. 195;et al. 195Craib, I. 370Cubeddu, R. 337Currie, G. 384Cussins, A. 360D’Agostino, F.B. 387Danto, A. 385, 387Darden, L. and Maull, N. 384Davis, J. 373Demsetz, H. 255Denzin, N.K. 150Desrosières, A. 372Doise, W. 336Dolan, E.G. 364Dorman, P. 379Douglas, J.D. 155, 158, 370, 373; andJohnson, J.M. 370Downs, A. 380Dray, W.H. 352Dugger, W. 379Eden, P.G. van den and Hüttner, H.J.M.184, 185Eggertsson, T. 256, 379Elster, J. 267, 288Emerson, R.M. 374England, P. and Kilborne, B.S. 379

438 Index of author namesEriksson, B. 367Fabian, R. and Simons, P.M. 364Farmer, M.K. 288Farr, R.M. 335Feigl, H. 176Feuer, L.S. 46Field, J.A. 376Fine, G.A. 135, 367Foldes, L.P. 381Föllesdal, D. 322Friedman, M. 384Frisby, D. 73Furubotn, E.G. and Richter, R. 261Galtung, J. 373Gellner, E. 2, 385Gergen, K.J. 37Gibson, Q. 385Ginsberg, M. 336Girill, T.R. 384Gisbert, P. 35Goldthorpe, J. 373Gordon, C. 63; and Gergen, K.J. 137Grafstein, R. 381Greaves, B.B. 87Grossman, S.J. and Hart, O.D. 271, 380Gurwitch, A. 363Haberler, G. 107Hacking, I.C. 30, 37, 372Hahn, F.H. 242Haines, V.A. 374Halfpenny, P. 180, 372Hall, R.T. 35Hamblin, R.L. and Kunkel, J.H. 374Hardin, R. 272Harré, R. 323, 335Hart, O. 272, 380; et al. 270; and Moore, J.271, 380Hartley, J.E. 238, 377, 378Hechter, M. 8Hedström, P.: et al. 206; and Swedberg, R.387Heijdra, B.J. et al. 379Heine, W. 385Hempel, G. 176, 384, 385Hennis, W. 104, 365Heritage, J. 370Hicks, J.R. 230; and Allen, R.G.D. 57Hilbert, 371Hindess, B. 138, 158Hodgson, G.M. 281, 379Hollander, S. 15Hollis, M. 249, 360, 379; and Nell, E.J.376Holmström, B.: and Milgrom, P. 270, 272;and Tirole, J. 271Holton, R.J. and Turner, B.S. 365Huber, J. 368Huff, T.E. 385Hughes, J. 372Hummel, H.J. and Opp, K.-D. 357, 373,384Hund, J. 372Hutchison, T.W. 256, 364Iggers, G.G. 23Infantino, L. 358Ingram, 375James, S. 385Janssen, C.W. 228, 288, 376Jarvie, I.C. 375Jensen, M.C. and Meckling, W.H. 271Joas, H. 165, 367Johansson, I. 323Johnson, G.D. and Picou, J.S. 368Jones, B. 365Jordan, R.W. 363Kauder, E. 364Kinkaid, H. 2, 3, 284, 320, 357, 378, 379,384, 385Klein, L.R. 230, 232, 238Knight, F.H. 367Knorr-Cetina, K. 126, 154, 155, 156, 371,373Koertge, N. 386Kuhn, T.S. 321Kusch, M. 85Kuznets, S. 228Lachenmeyer, C.W. 372Lachman, L. 364Lakatos, I. 321Langlois, R.N. 281, 379Latsis, S. 275Laudan, L. 321, 384Lavoic, D. 364Lazarsfeld, P.F. 373Leary, D.E. 59Leijonhufvud, A. 231Lennon, K. and Charles, D. 384Lessnoff, M. 385Levine, A. et al. 382, 384

Levinthal, D. 380Lewis, J.D. and Smith, R.L. 367Lichtblau, K. 366Lichtheim, R. 385Lindenberg, S. 335, 374, 385Lipset, S.M. 373Little, D. 181, 356, 382, 384Lukes, S. 5, 183, 339, 387Lundberg, G.A. 373Macdonald, G. and Pettit, P. 384, 386Macdonald, R. 104Machlup, F. 104MacIntyre, A. 387McKenzie, L.W. 245MacLean, M.J. 26, 27Mäki, U. et al. 379Makkreel, R.A. 362; and Scanlon, J. 363Mandelbaum, M. 384, 385, 387Manis, J.G. and Meltzer, B.N. 131Mannheim, K. 358Manning, P.K. 370Margolis, J. 335Martin, M. 387Mathien, T. 374Mayerl, W.M. 370Mehan, H. and Wood, H. 370Meinecke, F. 20Mellor, D.H. 385, 387Meltzer, J.G. and Petras, B.N. 368Meyer, J.W. 37, 379; and Scott, W.R. 379Miller, R.W. 2Mills, C.W. 372, 373Mirowski, P. 376, 378Moggach, 313Mohanty, J.N. 360Mommsen, W. 365Morgenbesser, S. 372Moscovici, S. 335, 336Myers, M.L. 42Nagel, E. 342, 372, 375, 384Natanson, M. 369Nelson, A. 228, 264North, D.C. 275, 380Notturno, M.A. 360Novak, M.E. 358Nozick, R. 387Oakley, A. 15, 360O’Driscoll, G.P. 364O’Hear, 375Opp, K.-D. 374, 387Ordeshook, P.C. 289Osterberg, D. 372Osterhammel, J. 104Outhwaite, W. 362, 385Pandit, G.L. 360Papineau, D. 372, 384, 385Parsons, S. 107Paul, E.F. 359Peacock, J.L. 8Pettit, P. 35, 384, 385Phelps, E.S. 377Phillips, D.C. 375Phillips, D.L. 373Pizzorno, A. 8Powell, W.W. and DiMaggio, P.J. 257, 379Pratt, V. 35, 354Pribram, K. 358Przeworski, A. 383Putnam, H. 385Quinton, A. 372Index of author names 439Radnitzky, G. 384Raub, W. 374Rehlberg, K.-S. 365Rickman, H.P. 362Ritzel, G. 364Roemer, J. 383Roger, M.F. 150Rogers, M.F. 142, 159, 369, 370Roth, G. 98Rothbard, M.N. 107Ruben, D.-H. 350, 357, 386Rudner, R.S. 372Runciman, W.G. 103, 365, 384, 385Ryan, A. 48, 360Ryle, G. 387Samuelson, P.A. 230, 291Sappington, D.E.M. 380Sarbin, T.R. and Kitsuse, J.I. 162Satz, D. and Ferejohn, J. 288, 385Scaff, L.A. 365Schlegloff, E. and Sacks, H. 371Schluchter, W. 365Schön, M. 365Schotter, A. 284, 339Schumpeter, J. 15Scott, J. 374Scott, K.J. 350, 352Scott, K.S. 374Scott, R.R. and Meyer, J.W. 379

440 Index of author namesSelgin, G.A. 107Sensat, J. 357Shafer, W. and Sonnenschein, H. 378Shapere, D. 323Sharrock, Wes and Andersson, Bob 159,370; and Button, G. 371Silverman, P. 364Simon, H.A. 351Sjöstrand, S.-E. 257Skidmore, W. 368Sklar, L. 384Small, A. 358Smith, B. 107, 364Solow, R. 238Sombart, W. 358Song, H.-H. 10Sowell, T. 358Sproule-Jones, M. 288Stern, F. 358Stigler, G.J. and Becker, G.S. 382Stokes, 375Strauss, A.L. et al. 369Strawson, P.F. 155, 385Streissler, E. 364Stryker, S. 367Suppe, F. 176Swingewood, A. 358Sztompka, H. 350, 386Tajifel, H. 336Taylor, M.W. 31, 32, 383Tenbruck, F.H. von 365Therborn, G. 365Thomas, D. 384Thomas, G.M. et al. 37, 379Tiryakian, E.A. 369Torrance, J. 365Troeltsch, E. 358Tullock, G. 256Turner, J.H. 374Turner, R. 136, 371Tyrell, H. 103Udehn, L. 32, 48, 95, 104, 123, 154, 183,256, 257, 281, 288, 367, 375, 379, 381,385Urmson, J.O. 372Van Parijs, P. 288Vanberg, V. 288, 358, 364, 385Vromen, J.J. 380Wagner, G. and Zipprian, H. 365Wagner, H.R. 151Walther, A. 365Ward, H. 289Warren, M. 386Watkins, J.W.N. 2, 323, 364, 384, 386Weintraub, E.R. 238Weiss, J. 365Wieder, D.L. 371Wieser, F. von 364Wilber, C.K. and Harrison, R.S. 379Willer, D. 186; and Skvoretz, J. 379; andWiller, J. 372Willey, T.E. 67Williamson, O.E. 379, 380Winslow, E.G. 377Wippler, R. 374; and Lindenberg, S. 374Wisdom, J.O. 286, 323Wolff, K.H. 369Wrong, D. 150Zimmerman, D.H. 370, 371; and Pollner,M. 152; and Wieder, D.L. 153, 155

Indexacting individual 73action 298–9, 313–14, 351, 386; human seepraxeology; purposive 298; rational291, 299Acton, Lord 119Agassi, Joseph 10, 43, 162, 197, 210,218–21, 229, 275, 287, 297, 344agency theory 270aggregation 234–6, 237–8, 304–5Alchian, Armen A. 121, 247, 256, 257,258, 267, 268–9, 271, 276, 281, 284–5Alexander, Jeffrey 126, 130, 159, 166, 367,370Allen, Roy G.D. 57Allen, William R. 247Allport, Floyd H. 65, 66–7, 361Alter, Max 20–1Analytical Marxism 309–18, 338analytical properties 184Andersson, Bob 159anthropology 191Archer, Margaret S. 187Archibald, W. Peter 335Arensberg, Conrad 190Aristotle 62, 257, 262, 263, 364Arrow, Kenneth J. 12, 243, 244–7, 248,250, 251, 254associationism 70atomism 88–94, 99, 114, 118, 124, 284,299, 323–4Austrian School 20–1, 24, 49, 56, 63, 73,85, 87, 142–3, 230, 234, 250, 251, 258,281, 306, 333, 335, 364; free frompsychologism 93–5; psychologistic87–93Average Man 29–30Axelrod, Robert 284axiomata media 46, 360Ayer, Alfred J. 170–1, 172–3, 213Ayres, Clarence 256Bacon, Francis 46–7, 179, 360Bacon, Roger 168Baldwin, James 65, 361Bar-Hillel, Yehoshua 155–6Battle of Methods (Methodenstreit) 88–92,364Beauvoir, Simone de 144, 147–8Becker, Gary S. 121, 131, 263–4, 268, 275,279, 297, 380behaviourism 191–3, 196–9, 267, 343being-in-the-world 145Beneke, Friedrich Eduard 59Bentham, Jeremy 15–16, 44, 50, 119Benton, Ted 187Berger, Peter L. 77, 137, 144, 147, 161–2,186Bergmann, Gustav 173, 328, 329Bergson, Henri 64, 86, 139, 145Berkeley, George 44, 168Berlin, Isaiah 22, 340–1, 358Bhaskar, Roy 77, 161, 186–7Blau, Peter M. 164, 184, 194, 302–3Blumer, Herbert 127, 130–3, 153, 368,369, 370Böhm-Bawerk, Eugen 92, 93, 230Boland, Lawrence W. 57, 228–9, 249, 250,254, 275, 287, 297, 355, 376Bosanquet, Bernard 65Boudon, Raymond 187, 189, 306–9, 319,335, 344, 345, 382bounded rationality 267Bourdieu, Pierre 77, 85, 149, 163–4, 370Braithwaite, Richard B. 173Brentano, Franz 62–3, 80, 82, 84Brentano, Lujo 23Bridgman, Percy W. 177, 190Brinton, Mary 371

442 IndexBroad, Charlie D. 328Brodbeck, May 213, 215, 328, 329Brown, Robert 222Buchanan, James 273–5, 276, 281, 294,319, 344, 385Buckle, Henry Thomas 26, 29, 30, 47, 180Bunge, Mario 186Burns, Arthur F. 235Burr, Vivien 162Butler, Judith 162Cairnes, John Elliot 18–19Carlyle, Thomas 11–12, 25, 48Carnap, Rudolf 173–5, 387Cassel, Gustav 57category-bound activities 156–7causality 313–14, 315Chapple, Eliot 190charity, principle of 96Charon, Joel M. 136Cheung, Steven N.S. 269, 277, 380Chiappori, Pierre-André 264–5Chicago School 56, 121, 127, 131, 230,234, 240, 257, 260, 268, 269, 276, 281,284, 338, 345, 367, 368choice 121–2, 123, 124Chomsky, Noam 155Cicourel, Aaron V. 153, 154, 155Clower, Robert 239, 377Coase, Ronald H. 256, 258–60, 266, 271,276Coase theorem 259cognitive ethnomethodology 155Coleman, James S. 34, 187, 188, 194, 195,266, 319, 335, 381–2; individualism of292–5; methodological individualism of297–300, 302; micro-macro scheme299–303; rational choice sociology292–306; structural-individualism295–7, 301–6, 381Coleridge, Samuel Taylor 44collectivism 38–40, 343–4; see also holismCommons, John R. 256composition laws 328–9Comte, Auguste 26, 27–9, 31, 34, 44, 45,47, 49, 62, 67, 68, 358–9, 360concepts 214–15; collective 98–100, 101,103, 110–12, 118, 120, 142, 198, 323;formation 78condition of connectability 326condition of derivability 326consciousness 68, 86, 280; collective 142,146constitutions 281Cooley, Charles Horton 65, 66, 127–8cultural objects 173–4Danto, Arthur 5, 34, 355Davidson, Donald 96Davis, Kingsley 292Debreu, Gerard 12, 244, 248, 249democracy 104Demsetz, Harold 121, 256, 258, 268–9,271, 276–7Derrida, Jacques 363Descartes, René 82, 119, 147, 367descriptions, theory of 169–71, 369Dewey, John 64, 65, 127, 361, 367Dietzel, Heinrich 93, 94–5, 365Dilthey, Wilhelm 49, 63, 67–73, 79, 81, 84,86, 97, 126, 127, 209, 342, 357, 362–3,367, 370dispositions 212Dodd, Stuart C. 183Donisthorpe, Wordsworth 359Droysen, Johann Gustav 25, 26–7dualisms 164, 343, 386Durkheim, Emile 38, 39, 61, 106, 307,383; collective consciousness 142, 146,148, 149, 186; critics of 35–7;description of economics 33–4;external factors 314; holism of 35, 45,160; sociology of 32–5, 128, 159–60,164, 208, 257, 302; on suicide 30, 158,181–2, 359, 371Ebbinghaus, Hermann 77, 362ecological fallacy 185economic imperialism 256economics 28, 32–4, 376; abstractdeductive13, 14–15, 18, 61; agents246–7; aggregates 234–6, 237–8, 240;classical 11–19, 230–1, 358, 376;constitutional 281; consumption 263–4;endogenous/exogenous factors 228–9,280; ends/means 122–3; as historical89; individualistic 12–13, 14, 228–30;inductive/holistic approach 13–14;invisible hand 10, 90; IS-LM diagram238–9; Keynesian 231–4, 239–40, 241;monetarist 234–5, 246, 284, 377;neoclassical 239–40, 255, 266, 297,376; political 11–12, 15–19; practicalscience 89; and prices 246; andproperty rights 248; rational 95–6, 240;as science of action 55; theoretical 89;

as theory of choice 121–2, 123, 124;tri-partition of 89, 92; utilitarian 49–58Edgeworth, Francis Y. 51–2egoistic principle 52Ellwood, Charles 65–6Elster, Jon 2, 96, 187, 252, 305, 319, 324,330, 335, 344, 345, 357, 380, 382–3,386; rational choice 310–16emergence 45, 299, 328emergent evolution 38–9Emerson, Richard M. 194–5empiricism 23, 44, 47, 107, 322, 338, 386;British 168–73; systematic 179–89,372–3Engels, Friedrich 12Enlightenment 168environmentalism 342–3, 385–6epistemology 213, 217–18, 325, 351–3,354, 386equilibrium 114essentialism 201, 202ethnomethodology 85, 126, 137, 165,370–1, 374; common understandings151–2; described 150; indexicality 151,155–8; interpretative/normativeparadigm 153; rules 152–3; society asprocess 150–5; subjectivist 159–60ethology 44, 46–7evolutionary economics, background281–2; firm 284–5; group 282–4;habits/routines 285–6exchange theory 189–90, 294–5, 373–4,378; Homan’s 190–6, 374existentialism 126, 144–9, 167explanation 164, 282, 299–300, 311–12,313, 332, 335, 354–6, 357, 386, 387;half-way 215–16, 312expressive individualism 25–7face-to-face interaction 134–5fallacies 184–5false individualism 15family see householdFaris, Ellsworth 130, 368Fechner, Gustav 60Feigl, Herbert 173feminism 162Ferguson, Adam 10Feyerabend, Paul 327Fichte, Johann Gottlieb 60firm, the, as capitalist enterprise 265–6;contractual view 268–70; evolutionary284–5; principal-agent 271–2; profitmaximisation284–5; property rights270–1; routines 285; transaction costs258–60, 266–8, 269Fischer, Irving 56Fisher, Ronald A. 179folk souls 38, 40, 361Frankfurt School 85freedom 145–6freedom of will 95–6Frege, Gottlob 68Friedman, Milton 121, 239, 276, 277, 279,377–8Fries, Jacob Friedrich 59functionalism 101–2, 116, 190, 280, 283,291, 292, 306, 382Galton, Francis 179game theory 91, 250–4, 264, 284, 289,311, 313, 366, 383Garfinkel, Harold 155–6, 157–8, 159, 370Geanakoplos, John 244Geisteswissenschaften 138Gellner, Ernest 383general equilibrium theory 241–50, 257;indeterminacy 250; institutionalelements 241–8; objections to 378;stability 248–9; uniqueness 249–50generalised other 129Gergen, Kenneth J. 137, 162German Historical School 23, 94, 97, 98,243, 322German historicism see historicismGerman Sociological Association 99, 207Giddens, Anthony 76, 85, 147, 163,164–5, 187Giddings, Franklin H. 65Gierke, Otto von 381Gödel, Kurt 173Goethe, Johann Wolfgang von 25Goffman, Erving 134–5, 153Goldstein, Leon 216Goldthorpe, John 189, 359Gordon, Chad 137Granovetter, Mark 195Gray, Tim S. 31Gruchy, Allan G. 256Gurwitch, Aron 137, 369Habermas, Jürgen 85, 141Hahn, Frank H. 244Hargreaves Heap, Shaun P. 253–4Harré, Rom 186, 373Harsanyi, John C. 252, 291–2, 365Index 443

444 IndexHart, Oliver 270–1Hartley, David 360Hayek, Friedrich von 10, 22, 32, 45, 94,107, 140, 160, 163, 181, 243, 258;aggregation problem 238; collectiveentities 118; defender of laissez-faire338; equilibrium analysis 114;evolutionism of 281, 282–4; footpathexample 117–18; helps Popper 200–1;holism of 282–4; influence of 322; onmethodological individualism 114–21,318, 342, 380–1; natural sciences 117;psychologism of 333; on socialinstitutions 204; and statistics 234–5;subjectivism of 115–17Hedström, Peter 188–9, 300Hegel, Georg Wilhelm Friedrich 22–3, 24,60, 86, 127Heidegger, Martin 85, 137, 144, 145, 146Helmholtz, Hermann von 60Hempel, Carl Gustav 173, 191, 355Henderson, L.J. 190Herbart, Friedrich 59Herder, Johann Gottfried 21, 25, 358hermeneutics 71–2, 154Hicks, John R. 57, 238–9, 249, 377Hildebrand, Bruno 23Hindess, Barry 138historical individualism 25, 78–9historicism 68, 343–4; development of22–4; holistic ideas of 24–7; meaningof 19–20; roots of 21–2Hobbes, Thomas 8, 16, 47, 91, 146, 168,293, 294, 306Hobhouse, Leonard T. 65, 66Hodgson, Geoffrey M. 286holism 37, 38–40, 45, 68, 70–1, 86, 130,145, 155, 162, 175, 184, 194, 214,218–26, 233, 237, 343–4, 357; ofColeman 293, 299, 300; evolutionary282–4; and the firm 268; Hayek’s282–4; and society 312Hollander, Samuel 15Holmström, Bengt 272Homans, George C. 43, 44, 49, 293, 294,318, 321, 330; theory of exchange190–6, 374household: Ancient writings on 261–2;production function 264–5, 378; as unitof consumption 263–4, 328Hughes, Everett C. 131humanism 147, 339–45Humboldt, Wilhelm von 25, 26, 44Hume, David 10, 42, 44, 91, 168, 284,360, 378Hummel, Hans J. 196–9, 330, 357Husserl, Edmund 63, 68, 80–5, 126, 137,139, 143, 145, 333, 363, 369Hutcheson, Francis 263ideal types 95, 96, 107, 211idealism 23, 175Iggers, George G. 25indexicality 155–8individualism 100, 135, 360; aristocratic25; background 7; characteristics of120; and classical economics 11–19;expressive 25–7; false 15; and Germanhistoricism 19–27; holism/collectivism38–40; institutional 17, 19; andinvisible hand 10; objective 24–5;ontological 2, 76; and positivistsociology 27–38;qualitative/quantitative 24–5; andsocial contract 7–9Infantino, Lorenzo 43institutionalism 10, 17, 19, 49, 56, 202–6,210, 230, 256, 304–5; Agassi’s 218–21,255; holism–individualism issue218–26; Jarvie’s 221–4; old 256, 257;Wisdom’s 224–6institutions 121, 122, 124–5, 156, 236;defined 258instrumentalism 201intentionalism 63, 69interaction analysis 306–7interest groups 275interpretative paradigm 153intersubjectivism 131–2, 134, 136, 140–3,148–9, 153, 165, 312, 367inverse deduction 47Iowa School 127Isaac, Jeffrey C. 187James, S. 337James, William 63–4, 65, 86, 127, 141Janssen, Marteen C.W. 247, 376Jarvie, Ian Charles 207, 210, 221–4, 323,354, 387Jaspers, Karl 85, 86, 137, 144, 363Jensen, Michael C. 269–70Jevons, W. Stanley 49, 50–1, 53, 55, 87Kant, Immanuel 59–60, 63, 67, 126, 147,362Kaufmann, Felix 107, 357

Keat, Russel 187Kemeny, John G. 327Kemeny–Oppenheim paradigm 325,326–30Kessler, Suzanne J. 162Keynes, John Maynard 54, 230–4, 239,376–7Kierkegaard, Sören 86, 144Kim, Jaegwon 330–1Kinkaid, Harold 324–5, 357Kirman, Alan 249, 254Klein, Lawrence R. 232, 238Knapp, G.F. 23Knies, Karl 23, 95Knight, Frank H. 121–2, 123, 260, 285Knorr-Cetina, Karin 156knowledge 57–8, 79, 121, 168, 169–70Koopmans, Tjalling C. 235–8, 377Kruks, Sonia 146, 148Kuhn, Manfred H. 131, 368, 369Kunkel, John H. 194Kusch, Martin 59, 85, 86Lakatos, Imre 337Landheer, Bert 137Lange, Oscar 238Langlois, Richard N. 258language 21, 108, 154Latsis, Spiro J. 207Lausanne School 56law 278–81, 387Layder, Derek 187Lazarsfeld, Paul F. 184Lazarus, Moritz 68, 73Leibniz, Gottfried Wilhelm von 69, 83,313, 324, 360Leijonhufvud, Axel 239, 377Lemert, Charles C. 371Lewis, David 284Lewis, J. David 129–30Lexis, Wilhelm 373liberalism 338–9Liefmann, Robert 97life-world 84–5, 141, 143, 148Lindenberg, Siegwart 319linguistic turn 154Little, Daniel 383Locke, John 8–9, 168logical constructions 168–73logical positivism 173–8, 321looking-glass self 127–8Lotze, Hermann 60Lucas, Robert E. 239, 240–1Index 445Luckmann, Thomas 77, 137, 138, 143–4,147, 161–2, 186Lukes, Steven 2, 8, 387Lundberg, George A. 183Macaulay, Thomas B. 16–17, 44McDougall, William 64–5, 65, 361MacFarland, Andrew S. 383Machlup, Fritz 107MacIver, Robert M. 65, 66McKenna, Wendy 162McKenzie, Lionel W. 244macroeconomics 228, 230–41, 377Maitland, Frederic William 381Malewski, Andrzej 196Malinowski, Bronislaw 190Malthus, Thomas 13, 14, 15Manchester liberalism 105Mandelbaum, Maurice 156, 157Mandeville, Bernard 10, 22Mantel, Rolf 249Marcel, Gabriel 145marginal utility 96, 361Markovsky, Barry 195Marshall, Alfred 53, 54–6, 121, 230, 232,240, 243, 257, 275Marx, Karl 11, 12, 22, 148–9, 265, 307,315, 358Marxism 112, 146, 147, 175, 187, 209,228, 256, 380; analytical 309–18, 338materialism 112Mead, George Herbert 127, 128–30, 132,367, 368Meckling, William H. 269–70Meltzer, Bernard N. 136Menger, Carl 22, 49, 63, 75, 87–91, 97,107, 114, 118, 119, 180, 230, 234, 246,281, 284, 318, 324, 333, 338, 364Menzel, Herbert 184Merleau-Ponty, Maurice 85, 86, 137, 144,148–9, 363, 370Merton, Robert 307, 360metaphysics 386methodological individualism: background41; debate on 2–3; different versions ofstrong 349–54; emergence of 20–1,376; epistemological 351–3, 354;explanatory 354–6, 386; graphicrepresentation of 5; meaning of 4;normative 336–45; ontological 350–1,352, 354; philosophical background321–4; as procedural 107; reasons for

446 Index320–1, 384; reductionism 324–36;strong/weak 346–9methodological individualism-in-use 4–5,9, 13, 91methodological pluralism 106methodological principle 211–13micro-macro explanation 307–8micro-reductions 327microeconomics 304microfoundations 49, 185, 233, 240–1,306, 308, 310–11, 336, 362, 383microinteractionism 126–7microsociology 126Milgrom, Paul 272Mill, James 15, 16, 17, 44, 45, 50, 61, 119,360Mill, John Stuart 12, 16, 17–18, 19, 25, 35,36, 42, 43–9, 61, 62, 68, 82, 107, 119,168, 179, 180, 191, 318, 321, 328, 338,360mind 63–4Mises, Ludwig von 63, 107–13, 114, 118,120, 140, 163, 181, 275, 321–2, 333,338, 342Mitchell, Wesley C. 235, 256Modigliani, Franco 239monadology 69, 143, 145, 313Moore, Wilbert E. 292Morgenstern, Oscar 250–4multi-level research 184–5multiplier effect 231Muth, John F. 240Myrdal, Gunnar 121, 123–4Nagel, Ernest 173, 191, 325–6, 330Nagel–Woodger–Quine paradigm 325Natanson, Maurice 137nationalism 21natural science 77–8, 342, 362Nee, Victor 371needs 87–93, 364Nelson, Richard R. 258, 267, 282, 285–6neo-Kantianism 67–8, 73, 77, 85, 97, 141,167, 322neo-Marxism 306neo-Walrasian 238network analysis 195–6, 374Neumann, John von 250–4Neurath, Otto 173, 175–6New Austrian School of Economics 107new institutionalism 286–7, 371, 379;background 255–6; compared with oldinstitutionalism 257; evolutionarytheory 258, 281–6; property rights 258;rational choice branch 257; roots256–8; social organisations 260–76;social rules 276–81; socio-historical257; theories of 258; transaction costs258–60Nietzsche, Friedrich 25, 60, 86, 144nominalism 110, 118, 153–4, 201–2,219–20non-tuism 52, 54normative individualism 336–7; humanist339–45; political 337–9normative paradigm 153norms 294–5, 299North, Douglas C. 256, 272, 277–8, 281,380Nozick, Robert 124Oakley, Allen 15objective idealism 24–5objective mind 210objectivism 75, 143–4, 147, 157–8, 209Ockham’s razor 168Older Historical School 23, 358Olson, Mancur 275, 380O’Neill, John 137ontology 2, 113, 125, 145, 350–1, 352,354, 386operationalism 191Opp, Karl-Dieter 196–9, 330, 357Oppenheim, Paul 327organicism 24, 28–9, 31–2, 90, 101, 284organisation 103, 109, 120, 122, 162, 257organisms 109, 120Pareto, Vilfredo 37–8, 56, 57, 121, 190,291, 307, 361Park, Robert E. 127Parsons, Talcott 141, 160, 257, 291, 302,307, 370part-whole distinction 39, 80, 118–19, 168,237partial equilibrium theory 243particulars 171path-dependence 91, 367Patinkin, Don 238Patton, Travis 195Paul, Ellen Frankel 32Pawson, Ray 187Pearson, Karl 179Pettit, Philip 39, 318, 344, 359, 360, 386phenomenology 63, 68, 80–5, 126,137–44, 165, 167, 322, 363, 369, 372

Philippovich, Eugen von 92philosophy 59, 68, 78physical determinism 341–2physicalism 175Piaget, Jean 335Pigou, Arthur C. 259Ploetz, Dr 99political individualism 25, 104–5, 337–9,385political science 288–90Popper, Karl 46, 86, 163, 173, 350, 375,386; alternative views to 218–26; asdefender of methodological monism342; difference with Austrians 210;essentialism 201, 210; and explanation164; falsifiability 323; andhistoricism/historism distinction 20;institutionalism 49, 202–6, 286, 300;institutions/collectives distinction205–6; on logical positivism 321; onmethodological individualism 38, 43,48, 200–1, 226–7, 229, 318, 375, 387;mind–body dualism 343, 386; naivetheories of society 205; as neo-liberalist338; nominalism 201–2, 210; onphysical determinism 341–2; andpsychologism 333–4; rejection ofdefinition 202; and situational logic161–2, 202, 206–7, 209, 375; on socialobjects 322; world objective 208–9Popper–Feyerabend–Kuhn paradigm 325positivism 78, 167–8, 321–2, 372; andempiricism 168–73; logical 173–8positivist social science, defined 178; andpositivist social theory 189–99; andsystematic empiricism 179–89positivist sociology, and doctrine ofemergence 37–8; and economics 28,32–4; and idealism 35; meaning of 27;organicism of 28–9, 31–2; andpsychology 28, 31–2, 35–6; andstatistics 29–30Posner, Richard A. 121, 278–81, 380pragmatism 27, 126praxeology 107–8, 109–12, 113, 115, 119,141, 322, 338Prendergast, Christopher 142principal-agent models 271–2probability 102–3proper names 169–71property rights 248, 258, 270–1, 276–8Przeworski, Adam 316–17, 319psychological knowledge 57–8psychologism 20, 28, 31–2, 35–6, 202,Index 447203–4, 218–21, 331–6; as approach tophilosophy 42–3; British 43–58;doctrine of 43; folk 42; German 58–77;meaning of 41–2; methodological 43;ontological 42; scientific 42; on trial77–86psychology 332; associationist 63–4;descriptive 68, 70–1, 73; explanatory(synthetic) 70, 72–3; rise of 58–67;social 65–7, 335–6; in USA 63Public Choice theory 230, 237, 258,272–5, 358Pufendorf, Samuel 8, 9Putnam, Hilary 327, 328Quetelet, Lambert A.J. 29–30, 35, 47, 51,180Quine, Willard van Orman 176R-predicates 217Radcliffe-Brown, Alfred R. 190Ranke, Leopold von 23, 26rational choice 8, 195, 228, 236, 257,274–5, 282, 356, 367; analyticalMarxism 309–18; background 288;game theory 289; in political science288–90; sociology 290–309; structuralindividualism 318–19Rational Choice Marxism 309–18rationality 88, 93–7, 107–8, 121, 207Raub, Werner 318, 319realism 78, 110, 118, 201reductionism 41–2, 196–9, 220, 306, 312,324, 387; microfoundations 335;psychologism 331–6; scientific 324–31Reichenbach, Hans 173Ricardian Vice 15Ricardo, David 13, 14, 15, 49Rickert, Heinrich 77–80, 97, 332, 362Rickman, Hans P. 72risk 121Robbins, Lionel 93, 121, 122–3Robinson Crusoe model 53, 88, 122,251–2, 358, 382Robinson, Joan 377Roemer, John 316–18, 319role 136, 153, 198–9, 274Romanticism 21, 23, 24, 25, 98, 127, 290,364, 367Roscher, Wilhelm 23Rose, Arnold M. 135–6Ross, Edward A. 64–5, 65Rousseau, Jean-Jacques 8, 9, 119, 360

448 IndexRuben, David-Hillel 3rules 258, 275, 371; constitutions 281; law278–81; property rights 276–8Ruskin, John 12, 358Russell, Bertrand 168–70, 171, 173, 213Rutherford, Malcolm 256Ryan, Alan 45, 357Ryle, Gilbert 212Sacks, Harvey 156–8St Simon, Henri 27, 358Samuels, Warren J. 256Samuelson, Paul A. 238, 242, 263Sargent, Thomas J. 239, 240–1Sartre, Jean-Paul 77, 85, 86, 137, 144,145–7Saussure, Ferdinand de 359Savigny, Friedrich Karl von 22, 91Sax, Emile 92, 93Say, Jean-Baptiste 13–14Sayer, Andrew 187Schäffle, Albert 93, 101Schaffner, Kenneth 325, 327Scheler, Max 85–6, 86, 137Schelling, Friedrich Wilhelm von 60Schelling, Thomas 187Schiller, Friedrich von 25Schleiermacher, Friedrich 26Schlick, Moritz 173Schmoller, Gustav von 23, 33, 88–9, 97,106, 234Schopenhauer, Arthur 60Schotter, Andrew 256, 284Schroeter, Gerd 106Schumpeter, Joseph 97, 104–7, 110, 122,185, 243, 281, 285, 323, 366, 376Schutz, Alfred 64, 85, 107, 113, 114, 137,138–44, 149, 363, 369science 114–15, 176, 342; social 45–8, 360scientific method 45Searle, John R. 155second-order concepts 69–70self 64, 127–31, 162, 368Senior, Nassau 14–15Senn, Peter S. 360sense-data 168, 169–70, 171Sharrock, Wes 159Shepsle, Kenneth 289, 381Shibutani, Tamotsu 131, 133, 134, 153Shubik, Martin 253Sidgwick, Henry 52–3Simmel, Georg 24, 65, 73–7, 76, 77, 126,127, 362, 366, 368Simon, Herbert 267situational analysis 161–2, 202, 206–7,209, 300, 342Skinner, Burrhus Frederic 191, 192–3, 198,343Slutsky, Eugen 57Small, Albion 364, 367Smith, Adam 10, 12, 13, 91, 127, 128,242, 263, 274, 294, 367Smith, Richard L. 129–30Smuts, J.C. 38, 175social action 100–1, 102, 131–2, 371social concepts 322, 372social constructionism 160–5social contract 7–9, 42, 91, 108, 146, 273social function 292social institutions 207–8, 246–7, 255, 292social mechanisms 188–9, 316, 360, 373,387social organisations 260–1, 275–6; the firm265–72; the household 261–5; interestgroups 275; the state 272–5social phenomena 28–9, 342, 349–51, 354social relationships 99, 102–3social science 360; laws of formation ofcharacter 45–6; laws of mind 45; lawsof society 46–8social situation 352social status 291–2social stratification 292social structures 115–17, 135, 198–9,301–5, 317social systems 294–5, 304, 305social values 292sociation 76society, bottom-up/top-down 163, 186; ascommunicative process 133; influenceof 314–15; internal/external effect301–2; naive theories 205; asnegotiated order 133–4; as spontaneousorder 32–3sociology 65–7, 105, 257; as auxiliary tohistory 119; formal 73, 74–7;interpretative 73, 99; rational choice in290–309; Weberian 99–103solipsism 145, 147Sonnenschein, Hugo 249Spann, Othmar 102, 106, 365Spencer, Herbert 31–2, 109, 290, 359Spinner, Helmut 329Spinoza, Benedict 8spontaneous order 28, 358–9state, the 177–8, 206; background 272;individualistic 273–4; normative

theories 272–3; organic conception273; origin 272; rational choice 274–5;working of 272statistics 181–2, 234, 356, 373Steuart, James 241–2Stigler, George J. 121, 285Strauss, Anselm 131, 133–4, 153structural individualism 76, 77, 197–9,318–19, 371–2structuralism 144, 154–5, 162, 383;Coleman’s 295–7, 301–6structuration 187Stryker, Sheldon 136subjectivism 115–17, 120–1, 146–7, 155,159–60, 342, 359Sugden, Robert 284suicide 181–2, 371Sumner, William Graham 359Suppes paradigm 325Swedberg, Richard 188–9, 300symbolic interactionism 64, 126, 153, 165,368; Chicago School 131; described127; face-to-face 134–5; intersubjective131–3, 134, 136; Iowa School 131;looking-glass self 127–8; psychologicaltheory of self 128–31; as socialcontract 133–4; structuralist 135,136–7; Thomas’s theorem 128systematic empiricism 179–89, 372–3Tarde, Gabriel 35–7, 64, 65, 126, 180Taylor, Michael W. 31, 312, 313, 319Thomas, Robert Paul 277–8, 281Thomas, William I. 65, 66, 127, 128Tiryakian, Edward A. 137, 149Titchener, Edward B. 63Tocqueville, Alexis de 119, 307Tool, Marc R. 256transaction costs, asset specificity 268;frequency 268; uncertainty 267–8transcendental ego 143transcendentalism 143transformation functions 197, 199Tullock, Gordon 294Turgot, A.R.J. 378Turner, Jonathan H. 369Turner, Ralph 136, 153Ullmann-Margalit, Edna 284uncertainty 121universalism 102, 112Urry, John 187utilitarianism 44, 49–58, 123–4Index 449Vanberg, Victor 283Varoufakis, Yanis 253–4Veblen, Thorstein 256, 281verbalism 178verstehen 26–7, 114, 189Verthaltenstheorie 196–9Vico, Giambattista 21–2, 25Vienna Circle 173–8Vining, Rutledge 235–8, 377Virginia School 230, 234, 258, 273, 338,345Völkerpsychologie 46, 61–2, 68, 126, 127Vromen, Jack J. 283–4Wagner, Adolf 33Waismann, Friedrich 173Walras, Leon 12, 49, 53, 57, 87, 238, 243,244, 376, 377, 378Watkins, John W.N. 2, 207, 211–18, 233,312, 318, 321, 323, 330, 334, 337, 342Weber, Ernst Heinrich 60Weber, Max 25, 106, 189, 207, 263, 307,318, 324, 342; atomism 99; because/inorder to motives 139; collectiveconcepts 98–100, 99, 101, 103;concept/reality conflation 98;economics 96–7; emergence of thestate 272; on empiricism 322;explanatory understanding 72–3;individualism 99–100, 101–2, 103, 120,126, 163; methodological issues 97,97–8, 114, 138; political economy 365;power/authority distinction 265–6;psychologism of 332–3; and rationalchoice 291; rationality in economics 94,95–7; social action 100–1, 102, 140;sociological analysis 80, 99–103;subjective society 302Weber–Fechner law of psychophysics 60whole-part distinction see part-wholedistinctionWicksteed, Philip 52, 53–4, 121, 122, 361Wieder, Derek L. 153Wieser, Friedrich von 92, 230, 333Willer, David 195Willey, Thomas E. 130William of Ockham 168Williamson, Oliver E. 256, 258, 266–8,271, 276Wilson, Thomas P. 153Winch, Peter 155Windelband, Wilhelm 77, 78Winter, Sidney G. 258, 267, 282, 285–6

450 IndexWippler, Reinhard 318, 319Wisdom, John O. 161–2, 171–2, 207,224–6, 376Witt, Ulrich 286Wittgenstein, Ludwig 155Wolff, Kurt H. 137world objective 208–9Wundt, Max 62Wundt, Wilhelm 35, 60–2, 63, 367Xenophon 262Younger Historical School 23, 88, 234, 358Zetterberg, Hans 176–7Zimmerman, Don H. 153Znaniecki, Florian 128