A Successful Theory of Mental Representation

A Successful Theory of Mental Representation A successful theory of mental representation must answer two related

questions:

(1) How does representation work?

(2) How is misrepresentation possible?

These questions are related because answering (1) is a requisite for

answering (2): in order to explain how it is is possible that

cognitive systems produce erroneous representations (i.e.

representations that do not correspond to their "proper" meanings), we

must first explain how is it possible that cognitive systems produce

representations at all. Moreover, in the contemporary philosophical

scene, it is an additional requirement that both questions must be

given naturalistic answers, i.e. answers which ultimately account for

intentional phenomena in nonintentional terms. In this essay, I will

briefly expound and contrast the theories advanced, in order to answer

these questions, by three main contemporary philosophers of mind: Fred

Dretske, Jerry Fodor, and Ruth Millikan.

In his essay "Theory of Content I" Jerry Fodor presents the

antecedents of his own theory of representation by referring to the

work of the behaviourist psychologist B.F. Skinner. According to

Skinner's theory, and following standard behaviourist dogma,

semantical facts are to be explained in terms of behavioural

dispositions. Thus the fact that, e.g. the (mental or linguistic)

representation 'dog' expresses the property being a dog can be

reformulated as a subject's production of the representation 'dog' is

under the control of a certain type of discriminative stimu...

... middle of paper ...

... would mean mouse, not mouse-or-shrew, and 'mouse' representations

tokened by shrews would be asymmetrically dependent on 'mouse'

representations tokened by actual mouses. As Robert Cummins has shown,

this can be expressed in the respective truth and falsity of the

following counterfactuals:

(1) If mice didn't cause 'mouse's, then shrews wouldn't either. (True)

(2) If shrews didn't cause 'mouse's, then mouses wouldn't either .

(False)

But, as Cummins argues, if mice and shrews are indistinguishable,

there is no way for statement (1) to be true while (2) is false: the

same state of affairs that makes (1) true makes (2) also true. In this

way, shrew-tokened 'mouse's cannot be regarded as asymmetrically

dependent on mouse-tokened 'mouse's, and thus cannot see 'mouse' as

meaning mouse (and not mouse-or-shrew).