What is the history of the belief that representation requires an intentional stance? I am a neuroscientist and we regularly use representation in what I believe is a very different sense: something like a 'token realization.' For example, I show you a bar of a particular orientation and a neuron in your cortex fires. Other bars fail to evoke that response. A typical neuroscience paper might say something like: that neuron's activity represents a bar of that orientation. Is there a difference here? I think this concept of representation as a 'token realization' (maybe a bad term) is central to the description of brain function by practicing scientists.

The term "representation" is a very slippery one in philosophy. The U.S. philosopher H. P. Grice ( some info can be found at http://www.artsci.wustl.edu/~philos/MindDict/grice.html ) distinguished two sense of the word "meaning," but his distinction has relevance to contemporary talk about "representation".

Grice asks us to reflect on the difference between sentences like these two.

1 -- Those spots mean measles

2 -- The "occupied" sign means that someone is using the lavatory.

He points out that if sentence 1 is true, then the occurrence of spots entails the existence of measles. Equivalently, if the occurrence of spots doesn't entail the existence of measles, then it's not correct to say that the spots mean measles. If, for example, the same sorts of spots can be produced by an allergic reaction to penicillin, then one should have said, "those spots mean either measles or an allergic reaction to penicillin."

On the other hand, the truth of sentence 2 doesn't entail that if the "occupied" sign is present, then someone is in the lavatory. "Occupied" means occupied, even if some practical joker has managed to trigger the sign on an empty toilet.

Grice calls the first kind of meaning "natural meaning," indicating that it's the sort of relation that exists when there is a natural connection between the sign and the thing signified, when it's a matter of natural law that the appearance of the sign is, as a matter of natural law, dependent on what the sign stands for. Because we have such relations in nature -- because smoke means fire, and bear scat means bear -- we can and do speak of "natural signs".

In the second case, however, the relation between the sign and the thing it stands for is not determined by any natural law. In fact, it was quite up to us human beings which sign -- which representation, which set of symbols -- to use to represent occupied-ness. We could have all used the sign the French chose -- "occupe" (with an accent). That is, the relation between "occupied" and being occupied is conventional -- a matter of social agreement, rather than natural law.

Now a really interesting question (to me, anyway) at the intersection of philosophy of language and philosophy of mind is this: is convention the only way to get non-natural meaning/representation? Because if it is, then we have a problem ab9out thoughts. The problem is this. Thoughts seem to be kinds of representations. When I think about Paris, I am representing the city of Paris in my mind. And when I think that Paris is lovely in the spring, I'm representing the situation of Paris's being lovely in the spring. But these representations have to be non-natural in Grice's sense. Why? Because it's perfectly possible for me to think about Paris even if Paris (alas!) isn't there, before me, in all its splendor. My Paris thought does not indicate the presence of Paris, in the way the spots indicate the presence of Paris.

But on the other hand, it can't be that the relation between my thought about Paris and the city itself is conventional -- it can't be like the relation between the English word, "Paris" and the city. Why not? Because to set up a convention, one needs to have thoughts that already have representational content. Suppose I want to use a certain neurological state as my personal symbol for Paris. To make that policy for myself, I'd have to form the intention to use that state to refer to Paris. But in order to form such an intention, I'd have to already have a way of representing Paris to myself -- and we're back where we started: with a representational state that can't have natural meaning, but that can't be conventional, either. The project of explaining how there could be this sort of in-between case -- a case of non-natural, but non-conventional meaning, or represntation, is one that has absorbed the energies of a number of prominent philosophers of mind, including Jerry Fodor, Fred Dretske, and Ruth Millikan. These philosophers all have theories purporting to show how a relation of non-natural, but non-conventional meaning could arise out of a natural world filled with merely natural signs.

Many philosophers want to reserve the term "representation" for items that have non-natural meaning. Some want to reserve the term "intentionality" (which in its broadest sense just means "aboutness") for the same purpose. Now the way in which the specialized neurons you speak of "represent" some specific orientation of a bar sounds to me like "natural" representation -- the firing of the specialized neuron is a natural sign of the presence of a bar oriented in that particular way. If that's right, then the philosophers alluded to above will balk at calling the neuronal firing a "representation" at all. But you don't have to argue with these philosophers about how to use words. Just ask them this: can non-natural mental representations (like my current thought about a bar oriented plumb to the plane of the earth) be somehow explained in terms of the neurological natural signs with which you are familiar?

Read another response by Louise Antony
Read another response about Mind