Why do humans reason? Arguments for an argumentative theory

More documents

Recommendations

Info

Commentary/Mercier & Sperber: Why do humans reason?by different processes. Smith et al. (2009) suggest that interpersonalargumentation in group decisions changes goal-specificweights and not the order by which evidence is considered andthus its recall success. M&S find similar differences in implicitinference processes when inferences are made by a single individualor by several individuals in a group setting.In summary, preference and inference tasks seem connected,among other things, by a common argumentative nature, whichmay suggest shared cognitive mechanisms (Weber & Johnson2009).Reasoning, robots, and navigation: Dual rolesfor deductive and abductive reasoningdoi:10.1017/S0140525X10002955Janet WilesSchool of Information Technology & Electrical Engineering, University ofQueensland, Brisbane 4072, Australia.wiles@itee.uq.edu.au http://www.itee.uq.edu.au/≏janetw/Abstract: Mercier & Sperber (M&S) argue for their argumentativetheory in terms of communicative abilities. Insights can be gained byextending the discussion beyond human reasoning to rodent and robotnavigation. The selection of arguments and conclusions that aremutually reinforcing can be cast as a form of abductive reasoning that Iargue underlies the construction of cognitive maps in navigation tasks.Mercier and Sperber’s (M&S’s) theory of the adaptive value ofargumentative reasoning is intriguing from a computational perspective,since the search for arguments that support a given conclusionis computationally more difficult (viewed as a reasoningproblem) than logical reasoning. The first logical solvers weredeveloped in the 1950s (Newell & Simon 1956). Argumentativecomputers are yet to be developed.Argumentative reasoning, defined broadly as the discovery ofstatements to support a given conclusion can be cast as a formof adbuctive reasoning, or inferring a precondition from a consequent(following Peirce 1931–35). Such reasoning is logically fallacious,but as M&S’s target article details, it is typical of humanbehaviour to select arguments and conclusions that together aremutually reinforcing.We accept M&S’s arguments for the adaptive value of argumentativereasoning as a communicative skill. However, just asquestions have been raised in other fields about the evolutionof the sophisticated communicative abilities of humans, we canalso ask how an argumentative ability could have evolved.Many evolutionary adaptations are thought to be exaptations;that is, new uses for existing structures. Verbal argumentativereasoning obviously draws on linguistic ability, but it need notpostdate it. We consider the possibility that cognitive abilitiesunderlying argumentative reasoning may predate the evolutionof language by exapting abductive abilities from other domains.Reasoning is not the only domain where adaptive behaviourmay utilise abductive reasoning. A much more ancient evolutionaryability, which humans share with other mammals, birds, reptiles,and even insects, is the ability to navigate. Much is knownabout the navigational systems of mammals, including theneural representations of places (O’Keefe & Dostrovsky 1971)linked into cognitive maps (O’Keefe & Nadel 1978; Tolman1948), grid cells (Moser et al. 2008), and head-direction cells(Taube et al. 1990). Complementing neural studies are computationalmodels and embodied robots, and it is the fully functionalrobotic systems (Arleo & Gerstner 2000; Kuipers 2000; Milford& Wyeth 2003; Thrun 2003) that provide insight for thiscommentary.Two approaches can be contrasted for robotic navigationalsystems: a logically correct approach based on Bayesianreasoning (analogous to deductive reasoning), and one basedon a bio-inspired approach that exploits a form of abductivereasoning to constructive a cognitive map. In mobile robots, akey problem is to maintain an estimate of one’s current locationwhile exploring and mapping a new environment (called simultaneouslocalisation and mapping [SLAM]).Given informationabout localisation (such as a Global Positioning System [GPS]),mapping is a relatively straightforward deductive reasoningproblem, and conversely, given a map, localisation is straightforward.However, when both tasks must be solved simultaneously(in the absence of GPS), the errors in each compound. Manylocations do not have unique landmarks; apparently unique featuresof one environment may turn out to be present only transientlyor to be shared by other locations. Even recognising apreviously visited location at a later time can be challenging. Invision-only SLAM, one of the best-performing systems is theRatSLAM system (Milford 2008), inspired by the hippocampalmapping system of the rodent. Initially developed using placecells and head-direction cells, it was discovered early on thatthe robots also needed something akin to grid cells (althoughwhen the model was first developed in 2003, grid cells themselveswere yet to be discovered). RatSLAM learns the paths that arobot traverses through its environment and links them intomaps. It uses a unique optimisation system that maintains informationthat is locally consistent, while also estimating a globalmap.If a location is considered a “conclusion” in a mapping task,and features of the environment are considered “arguments tosupport that conclusion,” then systems that are effective at navigationare of necessity abductive reasoners. Maps are constructedby using locations for which there is evidence, and evidence isretained when it is useful for localisation. Maps and their evidenceneed to be mutually reinforcing to be useful. The hippocampushas been linked to many aspects of cognition as well asspatial memory. Argumentative reasoning may well be thelatest of its exapted abilities.Some empirical qualifications to thearguments for an argumentative theorydoi:10.1017/S0140525X10002840Christopher R. WolfeDepartment of Psychology, Miami University, Oxford, OH 45056.WolfeCR@muohio.eduhttp://think.psy.muohio.edu/home/Abstract: The empirical research on the psychology of argumentationsuggests that people are prone to fallacies and suboptimal performancein generating, comprehending, and evaluating arguments. Reasoningand argumentation are interrelated skills that use many of the samecognitive processes. The processes we use to convince others are alsoused to convince ourselves. Argumentation would be ineffective if wecouldn’t reason for ourselves.Mercier and Sperber (M&S) are insightful in proposing a stronglink between reasoning and argumentation. Understanding theargumentative context sheds light on the processes of reasoning.However, empirical research on the psychology of argumentationcontradicts several of their key claims. Contrary to their position,reasoning biases are common even in the context of argumentation,the confirmation bias is not a feature of argumentation andactually weakens arguments, and people cling to claims lessrigidly than is tacitly assumed by the authors.M&S’s review of the literature on the psychology of argumentationis surprisingly sparse. Unfortunately, the data suggest thatpeople are subject to fallacies and suboptimal performance ingenerating, comprehending, and evaluating arguments. Kuhn92 BEHAVIORAL AND BRAIN SCIENCES (2011) 34:2
Commentary/Mercier & Sperber: Why do humans reason?(2001) found that 4-year-old children are able to distinguishbetween evidence and explanation. However, children areoften poor at generating arguments (Means & Voss 1996). In adolescents,national educational assessments find that only about15% of 12th graders are adequately skilled in developingwritten arguments (Beatty 1996; Greenwald et al. 1999). AnneBritt and colleagues conducted a series of experiments on thecomprehension, evaluation, and generation of argumentativetexts by college students. Among their findings, only about 35%of participants were able to identify the main claims andreasons in arguments, 37% failed to reject unsupported arguments,32% failed to reject unwarranted arguments, 48%included other-side information in their arguments, and 65%wrote essays that did not include a single reason to supporttheir claims (Britt & Kurby 2005; Britt & Larson 2003; Brittet al. 2005; Larson et al. 2004). Britt et al. (2008) found that universitystudents have difficulty precisely recalling the main predicateof argument claims, but less difficulty recalling thepredicates of comparable narrative statements. Wolfe et al.(2007) found that highly implausible reasons and warrants (forexample, Paul should walk to the store “because walking is theabsolute best exercise that will ever exist, and exercising canlead to immortality”) yielded higher agreement than the sameclaims without support. Argumentation is a fundamental skillthat permeates human thinking (Voss & Van Dyke 2001).However, M&S have painted an unduly optimistic portrait ofour argumentation abilities.Some of what M&S describe as “confirmation bias” is sometimescalled the “myside bias” (Baron 1995; Perkins et al. 1991;Toplak & Stanovich 2003; Wolfe & Britt 2005; 2008; Wolfeet al. 2009a). Although some authors use the terms interchangeably,confirmation bias typically refers to a biased search for orweighing of evidence, whereas myside bias refers to biases ingenerating reasons or arguments (Wolfe & Britt 2008). M&Sstate that the confirmation bias “is a consequence of the functionof reasoning and hence a feature of reasoning when used for theproduction of arguments” (sect. 3, para. 1, emphasis theirs). Mycolleagues and I have conducted a series of experiments on argumentation,and the evidence differs in key respects from theirassertions (Wolfe & Britt 2005; 2008; Wolfe et al. 2009a).Wolfe & Britt (2008) had participants write argumentativeessays under different conditions. Some were assigned to writeessays for and others against an unpopular proposition. Participantshad access to a number of pro and con online texts, andwe also examined their search behavior. We found that themyside bias was pervasive. However, it was not linked to participant’spersonal opinions. People exhibited the myside bias whenarguing for the side with which they personally disagreed just asoften as for the side with which they agreed. We have replicatedthis finding, yet also find significant correlations between opinionstrength and myside bias on nonargumentative reasoning tasks(Wolfe & Boone, under review; Wolfe & Britt 2008). Moreover,participants exhibiting the myside bias in their arguments werenot biased in their search. They sought out both pro-side andcon-side texts.The myside bias is indeed a bias – even in the context of argumentation.To illustrate, a content analysis of published authenticarguments found that most writers included other-side informationin their arguments, commonly for the purpose of rebuttal(see Wolfe & Britt 2008). In laboratory experiments, presentingand rebutting other-side information consistently leads tobetter ratings of agreement, argument quality, and impressionsof authors than does excluding other-side information (Wolfe &Boone, under review; Wolfe et al. 2009a). The myside biasweakens arguments measurably.The factors predicting the myside bias in written essays areindividual differences in beliefs about argumentation. Evidencestems from successful tutorials that significantly reduce themyside bias in generating and evaluating arguments (Wolfeet al. 2009a; 2009b), answers to the open-ended question “whatmakes a good argument?” (Wolfe & Britt 2008), and reliable individualdifference measures (Wolfe & Boone, under review;Wolfe & Britt 2009). The context of argumentation changes thenature of the myside bias, but one-sided argumentation is problematicand not an inherent feature of argumentation.A tacit assumption in M&S’s account is that people havestrong, stable preferences and unwavering commitments toclaims. Argumentation is seen as a form of rationalization usedto convince others of claims derived from intuitive processesabout which people are only dimly aware. Yet, starting withearly research on informal reasoning (Perkins et al. 1983), wehave learned that positions are often fluid and tentative. Asreasoning progresses, those positions undergo changes. We typicallyargue about matters that are “debatable,” where reasonablepeople arrive at different conclusions and are open to persuasion.I believe that reasoning and argumentation are interrelated skillsdrawing on many of the same cognitive processes – two sides ofthe same coin. Dual-process theories suggest that people lackaccess to many of our own cognitive processes. Thus, when wehave tentative intuitions that are not well understood, we usemany of the same processes to convince ourselves that in othercontexts we use to convince other people. However flawedthese processes may be, argumentation would be ineffective ifwe were not also able to reason for ourselves.Deliberative democracy and epistemichumilitydoi:10.1017/S0140525X10002888Kevin Chien-Chang WuNational Taiwan University College of Medicine, Taipei, Taiwan.ccwu88@ntu.edu.twAbstract: Deliberative democracy is one of the best designs that couldfacilitate good public policy decision making and bring about epistemicgood based on Mercier and Sperber’s (M&S’s) theory of reasoning.However, three conditions are necessary: (1) an ethic of individualepistemic humility, (2) a pragmatic deflationist definition of truth, and(3) a microscopic framing power analysis during group reasoning.In recent decades, we have witnessed many public policy scandalsand controversies in which, compared with laypersons,experts are not necessarily better in either epistemic or moralaspects (Godlee 2010; Jasanoff 2003; Wynne 1996). Followingthe loss of public trust in expert decision making, new discoursesof deliberative democracy for public policy making haveappeared (Lengwiler 2008). Examples in the new trend ofemphasizing public participation are consensus conferences, citizens’juries, citizens’ panels, and the like (Abelson et al. 2003).Usually, there are two meanings of deliberation: The first is “aparticular sort of discussion” that “involves the careful andserious weighing of reasons for and against some proposition”(Fearon 1998, p. 63). The second is “an interior process” ofreason weighing by an individual (Fearon 1998, p. 63). Deliberativedemocracy adopts the first meaning, so deliberative democracyfits in Mercier and Sperber’s (M&S’s) group-reasoningsituation.Many theoretical arguments support deliberative democracy,but here I focus on the epistemic aspect of deliberative democracy.According to John Dewey’s experimentalist account ofdemocracy, the epistemic power of democracy is inseparablefrom citizens’ collective deliberation. In a democratic regime,citizens are engaged to deliberate collectively over the foreseenconsequences of public policies and then choose, test, evaluate,and revise the policies implemented. Accordingly, the discussionsamong citizens with diverse backgrounds and local knowledgecould facilitate the fair recognition of public interestBEHAVIORAL AND BRAIN SCIENCES (2011) 34:2 93
Page 4: Mercier & Sperber: Why do humans re
Page 7 and 8: Mercier & Sperber: Why do humans re
Page 19 and 20: Commentary/Mercier & Sperber: Why d
Page 35: Commentary/Mercier & Sperber: Why d
Page 39 and 40: Response/Mercier & Sperber: Why do
Page 45 and 46: References/Mercier & Sperber: Why d
Page 55: References/Mercier & Sperber: Why d

Why do humans reason? Arguments for an argumentative theory

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?