The museum theory of mind and language (or the picture picture)-

Miller attributes to Locke the belief that "understanding an expression consisted in the possession of some associated idea or image," (Miller, 36) and to Frege the following argument against this view

  1. One crucial role of sense is to explain how linguistic communication is possible:  the success of language in facilitating communication between two speakers is to be accounted for in terms of their grasping the same senses.
  2. Private, inner items have no role to play in explaining the practical success of language in facilitating communication between two speakers.
  3. Understanding and expression-grasping a sense-cannot consist in the possession of some inner, private item.

(A)  One way to balk at this argument is to deny the first premiss by instead holding that the success of language in facilitating communication between two speakers is to be accounted for in terms of their grasping similar senses.  Then one might argue that the inner items could be senses, but that if they were similar enough, then they could still explain linguistic communication.  (B)  Another way to balk at it would be to hang onto premiss one, but to try to tell a story about how private inner items can play a role in facilitating communication.

Both of these possibilities seem to undermine the inverted spectrum problem that Miller raises.  When Miller considers two people with inverted spectra talking to one another, he writes,

Now, if the possibility were to obtain, someone advancing an account of sense along the lines of Locke's would have to say that Smith and Jones were not communicating successfully: after all, they would be attaching different senses to the words "blue" and "yellow" because of the difference in the ideas that these words produce.

Well, this is only a problem if identity of sense is necessary for communication.  Maybe the senses of "blue" for Smith and Jones are equal to their mental images, but these are similiar in the relevant respects enough for communcation to proceed (this tack corresponds to (A) above).  The problem with this is that it is totally unclear in what respects the two senses are similar if you hold on to the view that senses are mental images.  If senses are mental images then it seems that the sense of "blue" for Smith is type-identical to the sense of "yellow" for Jones.  The museum theorist would need to spell out a notion of similarity such that Smith and Jones' "blue" senses are more similar than Smith's "blue" sense is to Jones' "yellow" sense.

Tack (B) seems more plausible to me.  If one hold that the associated images still play an important role in fixing sense, then one would disagree with premiss 2 above.  Say Jones has an inverted spectrum and Smith has a standard one.  Jones' wierd images are still what allow him to call objects in the world by their correct color (even if he sees them in a non-standard way).  While this tack seems to work to undermine premiss 2 above, the problem is that it admits that Frege is right after all.  Someone going this tack has already admitted that the sense is different from the associated image, which is Frege's conclusion.

One can avoid Frege's conclusion here in pursuing tack (B) by "biting the bullett" and saying that Smith and Jones really aren't communicating, since their associated images are different.  Given the pathological nature of the example, this seems at first site to be entirely convincing.  Isn't something not being communicated by them?

Miller is very unclear in discussing this example, and it is not a very good one, because everybody has different intuitions about inverted spectra examples.  Many philosophers of mind argue that inverted spectra aren't really possible, for example.  Most important, Miller's final absurdum is horrifyingly weak.  He gives it as

It is possible that we could do everything we currently do with language-use language for all of the purposes for which we currently use it-and yet not really be communciating, and not really attach the same senses to our words. (Miller)

If it is only possible in the weak sense that inverted spectra are possible, then it is not at all clear why we should worry about this.  Miller writes,

But this is absurd: there is absolutely nothing in our everyday conceptions of communication and grasp of sense that allows for such a possiblity. In particular, the admission of such a possibility effectively renders the notion of sense empty. (Miller)

This strikes me as terrible.  Think about Descartes' skeptical hypothesis that all of his conscious experience is an illusion created by an evil demon.  Would it be O.K. to say to Descartes

But this is absurd: there is absolutely nothing in our everyday conceptions of communication and grasp of sense knowledge that allows for such a possiblity. In particular, the admission of such a possibility effectively renders the notion of knowledge empty. (????)

Rather, one should respond by either showing how we can know that the skeptical hypothesis does not hold (as Descartes did) or by showing how  knowledge doesn't require the kind of certainty Descartes was after (as most contemporary fallibalists do).  Likewise one should respond to the envisioned possibility of global miscommunication by providing a positive argument that it is not possible (analogous to Descartes), or by arguing that its mere possibility doesn't in any way undermine actual communication (asnalogous to contemporary fallibalists).

So if the only problem were inverted spectra type examples, I think there would be no good argument against Locke.  Frege's argument is stronger though.  While Miller talks about the mere possibility of people associating different mental images with words,  Frege points out that people actually do have different associated images with words.  Frege writes,

A painter, a horseman, and a zoologist will probably connect different ideas with the name "Bucephalus".  This constitutes an essential distinction between the idea and the sign's sense, which may be the common property of many people, and so is not a part or a mode of the individual mind.  For one can hardly deny that mankind has a common store of thoughts which is transmitted from one generation to another. (quoted in (Miller, 38).

Here biting the bullett really does seem extreme, as one would end up saying that we never actually do communicate.  Since the other version of tack (B) involved admitting that Frege was right, this leaves the Lockean with tack (A), which involved arguing that sameness of sense is not necessary for communication.  Maybe when we are not worried about pathological cases such as inverted spectra there is a notion of similarity of images, such that similarity of sense (sense being equal to mental image) can explain communication. I don't know.

I think prospects for really equating sense with just a mental "picture" of some sort are exceedingly slim.  But, this should be kept seperate from the general issue of "keeping senses private."  Maybe other facts about our mental life are rich enough to non-circularly characterize a similarity space, such that we could say that senses are mental.  This is worth thinking about, but first we must examine other reasons why the picture picture is a bad picture.

Here I will suggest other arguments against the theory.  To present these well be a little more precise about the Lockean theory.

Museum theory of mind:  A person understands a word P if and only if (she possesses the appropriate image corresponding to P (call this i(P)) and knows that i(P) is the image connected with P).

Since this claim is an if and only if it really is a combination of two claims, both of which it is easy to disprove.

Counterclaim 1:  It is possible for someone to understand a word P without possessing "an appropriate image".

(1)  People blind from birth understand language. (2) For many words which we do understand, we just don't have any kind of sensory image.  Think of terms referring to abstractions, like "universal," or "Platonic form."

Counterclaim 2:  Given any candidate for "an appropriate image" corresponding to a word P  it is possible for one to possess the image, know that it is connected with the word P, and still not understand P.

For the sake of argument, assume that there is an image associated with each word.  Say that the image for "dog" is a labrador retriever.  Now assume someone who had never seen a dog was shown a picture of one and told that this was a picture of a "dog" a fake cartoon creature that appeared in all Microsoft software whenever you accessed Microsoft help files (various talking "dogs" were going to replace the annoying smiling paperclip).  Here the person has the appropriate image and knows the image is connected with the word, but still doesn't understand the word.

The point here concerns the conventionality of signs (a mental sign is still a sign). . .
talk about convention

However, sense might still be private in some other respects. . .