Source: p. 30 Bottom – p. 31 Top, Thought: A Very Short Introduction (2013) by Tim Bayne.

There are, however, serious problems with this account. For one thing, resemblance is ubiquitous, and any brain state [hereafter BNS] resembles multiple objects. Some of my brain states will resemble some of your brain states, but surely it does not follow that my brain states represent your brain states. In order to have any plausibility the advocate of this account must identify the kind of resemblance relations that underwrite mental content, and that has proven very difficult to do. A second problem is that we can think about all manner of things that do not resemble brain states in any way.

In what sense could the mental symbols [hereafter MLSL] for <beauty>, <truth>, and <justice> possibly resemble the properties of beauty, truth, and justice? [1.] Thirdly, and perhaps most fundamentally, it is not at all clear how mere resemblance might explain why a mental symbol has the meaning that it does. [End of 1.] [2.] Even if a brain state does happen to resemble (in some relevant sense) a certain feature of the world, it is utterly mysterious why this fact should entail that it represents that feature of the world. [End of 2.]

I must be missing or misunderstanding something, because 1 and 2 appear obvious to me.

  1. About 1: If a MLSL resembles a real object, then how can the MLSL not mean what it does (i.e. resemblance of the real object)? E.g., of the spectacle below, I will likely forget everything except the 3 main subjects: the lodge juxtaposed before the mountains and trees, and behind the bridge over water. But (remembering) this deficient MLSL still enables me to remember the location, photograph, and how to retrieve the photograph.

  2. About 2: If my BNS is the aforesaid MLSL, then how is it utterly mysterious that the MLSL would represent the photograph?

enter image description here

  • 2
    The photograph is just a splash of paint on canvas, or in this case pixels on computer screen. Without a reader attuned to conventions, absorbed in childhood, about a relation between 2D and 3D representations, pictures and schematizations, etc., nothing "resembles" anything else. To different people Rorschach blots "resemble" all sorts of things, mostly because no conventions attach to these shapes. And brain function certainly can not depend on social conventions for its relation to objects.
    – Conifold
    Commented Jan 4, 2017 at 4:33
  • In the days of Locke and Hume, an idea was in many cases a mental image (but not necessarily so). The author seems to be trying to use the word in a different sense, identifying it with brain states. Since one area of the cerebral cortex in a given state might represent multiple objects, he's using a sort of a question-begging argument to make Locke's and Hume's positions seem absurd. But if an idea is understood in the traditional sense, the authors objections fall to the ground, and there is no reason why meaning couldn't be established in that way. I'd look for a better book.
    – user3017
    Commented Jan 4, 2017 at 13:33

1 Answer 1


Resemblance theories of representation used to be common in philosophy but arguments by philosopher like Nelson Goodman and Wittgenstein made them unpopular.

The basic problem is hinted at in the first paragraph of your quotation: resemblance is neither sufficient nor necessary for representation. For example Barack Obama does resemble his picture but he does not thereby represent his picture, so resemblance is not sufficient; indeed a picture most accurately resembles itself but it doesn't thereby represent itself, also in many respects other humans resemble Barack more than his picture does, they are 3-dimensional and composed of similar material etc. but other humans don't represent him, and more generally pretty much any two objects resemble each other in some ways without thereby representing each other. Resemblance is not necessary because anything can represent any other thing by conventional stipulation, e.g. the sequence of marks Barack Obama represents him linguistically.

Goodman thinks that resemblance is irrelevant for pictorial representation which is a conventional matter. This conclusion has been criticized by suggesting that it conflates two questions: 1. what does it take for a picture to function as a representation, and 2. what determines the representational content of a pictorial representation. It has been argued that Goodman's argument only shows that resemblance does not answer question 1, but it could still play a role in answering 2. (see Laura Perini's "Convention, Resemblance and Isomorphism: Understanding Scientific Visual Representations.")

As Conifold notes brain processes are not subject to conventions, so Goodman's argument would show that resemblance is not enough to explain how "pictures" in the brain can represent. The neuroscientist Stephen Kosslyn has proposed a theory of mental imagery that posits picture-like analog representations in the brain (see https://plato.stanford.edu/entries/mental-imagery/quasi-pictorial.html), in that link it is mentioned that Fodor criticized Kosslyn's theory based on Goodman's and Wittgenstein's arguments, Fodor's suggestion is that for a picture in the brain to represent (to have intentionality) it has to be accompanied by a mentalese description, but it's also notoriously difficult to explain how a mentalese sentence in the brain gets its intentionality.

You must log in to answer this question.