07.10.2013 Views

Main trends of research in the social and human ... - unesdoc - Unesco

Main trends of research in the social and human ... - unesdoc - Unesco

Main trends of research in the social and human ... - unesdoc - Unesco

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

552 Raymond Boudon<br />

The notion <strong>of</strong> model. The present considerations wil be limited to <strong>the</strong> concept<br />

<strong>of</strong> <strong>the</strong> ma<strong>the</strong>matical model.<br />

A model <strong>in</strong> this sense <strong>of</strong> <strong>the</strong> term is always a hypo<strong>the</strong>tical-deductive <strong>the</strong>ory.<br />

In o<strong>the</strong>r words, it is a set <strong>of</strong> premises or axioms from which a set <strong>of</strong> consequences<br />

is deduced.<br />

In certa<strong>in</strong> cases, <strong>the</strong>se consequences may be checked by observation <strong>and</strong> <strong>in</strong><br />

that case <strong>the</strong> model is verifiable. But this property <strong>of</strong> verifiability, or more pre-<br />

cisely falsifiability (s<strong>in</strong>ce, as Popper has stressed, a universal proposition can be<br />

proved false, but never proved true), is not <strong>in</strong>dispensable, for <strong>the</strong>re are models<br />

which are not falsifiable.<br />

This po<strong>in</strong>t gives us an <strong>in</strong>itial dist<strong>in</strong>ction between <strong>the</strong> notion <strong>of</strong> <strong>the</strong> model <strong>and</strong><br />

that <strong>of</strong> <strong>the</strong> <strong>the</strong>ory (<strong>in</strong> <strong>the</strong> sense <strong>in</strong> which <strong>the</strong> natural sciences use <strong>the</strong> latter word).<br />

In effect, many examples can be cited <strong>of</strong> non-verifiable models such as <strong>the</strong><br />

multi-factorial analysis <strong>in</strong> psychology which always produces a solution, what-<br />

ever <strong>the</strong> data available, <strong>the</strong> solution itself be<strong>in</strong>g rejected or accepted on criteria<br />

o<strong>the</strong>r than that <strong>of</strong> congruence with <strong>the</strong> observed situation, which, by def<strong>in</strong>ition,<br />

is always achieved. Similarly, a multitude <strong>of</strong> models <strong>in</strong> ma<strong>the</strong>matical economics<br />

could be cited where it is impossible to decide on unambiguous criteria whe<strong>the</strong>r<br />

<strong>the</strong> model is true or false. Take for <strong>in</strong>stance <strong>the</strong> models <strong>of</strong> Samuelson, Goodw<strong>in</strong><br />

or Phillips on economic cycle <strong>the</strong>ory. These models show that, for certa<strong>in</strong> hy-<br />

po<strong>the</strong>ses as to <strong>the</strong> immediate or deferred l<strong>in</strong>kages between <strong>the</strong> global economic<br />

quantities (<strong>in</strong>vestments, sav<strong>in</strong>gs, consumption, etc.) <strong>and</strong> for certa<strong>in</strong> classes <strong>of</strong><br />

value <strong>of</strong> <strong>the</strong> models’ parameters, <strong>the</strong> global quantities mentioned manifest<br />

oscillations. Certa<strong>in</strong> <strong>of</strong> <strong>the</strong>se models can <strong>the</strong>n be judged preferable to o<strong>the</strong>rs<br />

to <strong>the</strong> extent that <strong>the</strong>y will reproduce more faithfully <strong>the</strong> general characteristics<br />

<strong>of</strong> <strong>the</strong> real cycles. Thus Goodw<strong>in</strong>’s model, as aga<strong>in</strong>st <strong>the</strong> Samuelson-Hicks<br />

<strong>the</strong>ory, has <strong>the</strong> advantage <strong>of</strong> giv<strong>in</strong>g <strong>the</strong>oretical cycles where <strong>the</strong> ris<strong>in</strong>g phase is<br />

longer than <strong>the</strong> fall<strong>in</strong>g, as <strong>in</strong> <strong>the</strong> real cycles. The first will <strong>the</strong>refore be preferred<br />

to <strong>the</strong> second. But it cannot be said that we are really <strong>in</strong> a position to verify<br />

this type <strong>of</strong> model (show it to be ei<strong>the</strong>r false or not false); <strong>in</strong> short, we can say<br />

that Goodw<strong>in</strong>’s model is more ‘realistic’ than <strong>the</strong> Samuelson-Hicks model, but<br />

we cannot say that it is more veridical.<br />

It could easily be shown that many models are universally recognized to be<br />

useful, although <strong>the</strong>y are not strictly <strong>the</strong>ories <strong>in</strong> <strong>the</strong> usual sense, mean<strong>in</strong>g that no<br />

criterion <strong>of</strong> falsifiability can be used directly on <strong>the</strong>m, or that <strong>the</strong>y cannot be<br />

made <strong>the</strong> subject <strong>of</strong> an experimentum crucis. But this is not to say that <strong>the</strong>y are<br />

<strong>the</strong>reby reduced to <strong>the</strong> status <strong>of</strong> arbitrary constructs. No experimentum crucis<br />

can be imag<strong>in</strong>ed which could <strong>in</strong>validate, for example, game ‘<strong>the</strong>ory’ or phonolog-<br />

ical classification. Similarly, no falsifiability criterion can be worked out from<br />

<strong>the</strong> psychoanalytic ‘<strong>the</strong>ory’ <strong>of</strong> personality. To conclude that such ‘<strong>the</strong>ories’ are<br />

mean<strong>in</strong>gless would run counter to <strong>the</strong> weight <strong>of</strong> op<strong>in</strong>ion. What it does, on <strong>the</strong><br />

contrary, seem important to br<strong>in</strong>g out is that <strong>the</strong>se examples suggest <strong>the</strong> neces-<br />

sity <strong>of</strong> reth<strong>in</strong>k<strong>in</strong>g <strong>the</strong> whole question <strong>of</strong> <strong>in</strong>ductive logic <strong>in</strong> <strong>the</strong> <strong>social</strong> <strong>and</strong> <strong>human</strong><br />

sciences, s<strong>in</strong>ce <strong>the</strong> logical pr<strong>in</strong>ciples <strong>in</strong>herited from <strong>the</strong> natural sciences, as de-<br />

scribed for example by Karl Popper <strong>in</strong> his famous Logik der Forschung, are <strong>of</strong> no<br />

service whatever for enabl<strong>in</strong>g us to underst<strong>and</strong> why unverifiable models or

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!