Here is a very simple question: What is a thought?

A Materialist Answer:

A thought is a kind of brain state.

Materialism: Everything that exists is matter, or made up of matter.

So for the materialist, if thoughts exist at all, they are material. In particular, they might be just certain sequences of neural firings in the brain.

(Note: Many materialists reject the characterization of thoughts in terms of brain states, since this seems to rule out artificial intelligence from the start. But we’ll ignore this issue for now.)

A Dualist Answer:

A thought is a state of an immaterial mind (or soul)

Dualism: There are two kinds of substances which exist: Material substance (what bodies are composed of) and Thinking substance (what minds/souls are composed of).

So for the dualist, a thought is just a state of an immaterial substance, whose primary characteristic is that it has thought, i.e., it thinks.

The vast majority of philosophers today are materialists (of one kind or another). Primarily, this is because dualists face the problem of interaction. Material substance and thinking substance appear to causally interact. For example, my thinking that there’s food in the refrigerator can cause my body to go up to the refrigerator and open it. So a thought can cause my body to move in certain ways. But how could an immaterial thought cause something to happen in my body? Intuitively, X can cause something to happen with Y only if there is some kind of physical contact between X and Y. But there is no physical contact between immaterial thoughts and material bodies, since immaterial thoughts are not the sort of thing with which you can have physical contact. So the causal interaction between thoughts and bodily behavior seems utterly mysterious for the dualist.

For most of this course, we will just assume that the materialist answer is roughly correct. A thought is just a kind of brain state.

A Question about Representation

If thoughts are just states of the brain, then these brain states are very unique. For they have the capacity to represent objects. For example, if my thought that “cows are bigger than dogs” is just a brain state, then that brains state has the strange power to represent animals, viz., cows and dogs, and relates them to each other in a certain way, namely, in the “is bigger than” relation. But how can a state of a material object, i.e., a brain, represent other objects?

(Actually, I think an analogous question can be raised for the dualist.)

Nothing else in nature seems to have this particular capacity to represent objects. For example, states of a chemical compound don’t represent anything else. They just are.

So if thoughts are just brain states, how do those brain states represent things?

Caveat: Other material phenomena seem to be able to represent things. For example, a picture of Winston Churchill represents Winston Churchill, even though the picture is nothing but a series of marks on paper.

But the way in which a picture represents is different from the way in which a thought represents:

Distinguish: Primitive representation from derived representation.

Let us say X primitively represents Y just in case X represents Y independently of Y being represented by anything else. On the other hand, X derivatively represents Y just in case X’s representing Y depends on the existence of some other representation.

So, in the case of the picture, the picture represents Winston Churchill because the picture’s representing Winston Churchill depends on the drawer’s intention that the picture should represent him. If there was no such intention, e.g., if the drawer was a small child merely moving her pencil in a haphazard fashion, then the figure would not represent Winston Churchill.

We should say, then, that the picture represents Winston Churchill derivatively: The picture’s representing Winston Churchill derives from the drawer’s intention that the picture should represent him.

Thoughts, on the other hand, like the drawer’s intentions, are primitive representations. I can have a thought that “cows are bigger than dogs” without the help of any other representation. Similarly, the drawer can have the intentions that she has, independently of the existence of any other representation.

This difference between thoughts and pictures, moreover, undermines the idea that thoughts just are mental pictures, or images in the mind (The “imagist” theory.) This is because, in order to represent a particular object, an image need to be interpreted as representing that object. But since an interpretation is a kind of representation, this just shows that mental images are derived representations. So mental images cannot account for how thoughts primitively represent what they do.

This point is worth going over again, using an example. Suppose having a thought about a triangle consisted in having a mental image of a triangle. But consider such an image:

▲

What would make it the case that this image represents a “triangle in general,” as opposed to an “equilateral triangle in general,” or just a “particular triangle on a particular occasion”? One would need an interpretation of the image, which again, would show the image not to be a primitive representation, but a derived one. So, since thoughts are primitive representations, thoughts cannot just be mental images.

To be clear, one certainly can have mental images. Moreover, one can even have thoughts about mental images—I can think about an image of a triangle that I conjure up in my head. But the point is that thoughts cannot be identical to mental images, for the reasons we have given.

Language may be another source of derived representations. The sentence “the cat is on the mat” represents the cat’s being on the mat, due to the speaker’s intentions in uttering that sentence. If the speaker just happened to utter those words at random (e.g. if she was not an English speaker), then the utterance would not represent anything at all. (I should note that this view of language is somewhat contentious, although most philosophers, myself included, accept it.)

A Question About Representation (Revisited)

Thus, we can reformulate the question: If thoughts are just brain states, how do those brain states primitively represent things?

But note we can raise this question, not just about thoughts, but about other mental phenomena too, e.g., concepts, experiences, and perceptions. What is common to these mental phenomena is that they all can represent things. So we may call this a question about mental representation more generally.

If mental representations, i.e., thoughts, concepts, perceptions, experiences, are just brain states, how do they primitively represent the things that they do?

This is a very hard question.

It is so difficult, in fact, that no one really knows how to answer it. But still, there is a lot to be learned from trying to address the issue.

But here’s a simple question: Why do we care?

Why inquire into the nature of thoughts? There are, to my mind, three motivations behind such an inquiry.

Natural Human Curiosity

Humans, in general, want to understand the world they live in. Science, for example, is the systematic attempt on the part of human beings to understand the phenomena they encounter. But just as we want to understand chemical compounds, planetary orbits, and species of iguana, we might just as well want to understand understanding itself.

What is it to understand something?

Presumably, the understanding is a natural phenomenon in the world, just like anything else. So like those other phenomena, we might have a natural human curiosity to understand the faculty of the understanding.

If Locke is to be believed (Essay, Bk. 1, Ch. 1, section 3), understanding the understanding includes, among other things, understanding those “ideas” or “notions” with which the mind is furnished.

Or, in more contemporary terminology, it requires us to understand thoughts which are composed of concepts, as well as the experiences and perceptions by which we come to understand things.

But: The point about “natural curiosity” only answers the question of why some people do care about an inquiry into mental representation. Yet we might be more interested in the question of why we should care about such an inquiry. Even if we are naturally motivated to pursue these questions, is there any reason to think this is a worthwhile pursuit?

Locke’s Motivation

Locke himself was motivated to inquire into “ideas” and “notions,” so as to better understand “the discerning Faculties of a Man,” in hopes that this may bring us “great Advantage, in directing our Thoughts in search of other things.”

In brief, understanding the faculty of the understanding might help us in using that faculty better.

Perhaps, then, we could resolve many disputes that people get into. Or, failing that, our inquiry might show us that “either there is no such thing as Truth at all; or that Mankind hath no sufficient Means to attain a certain Knowledge of it.”

The Technological Motivation.

Moreover, there may be technological benefits in answering questions about mental representation, particularly in relation to the field of artificial intelligence (AI).

What would it mean to build a machine that thinks? It would mean to build a machine that has thoughts. But what is a thought? And how could a physical piece of machinery have thoughts, i.e. states that primitively represent objects? This feeds us back into the questions we are exploring here.

There are many issues in the philosophy of mental representation which we will not explore here. But we will hit (what I take to be) some of the highlights in the field. We will start with a few articles which explain the issues at hand in more detail. Then we will look at some accounts of what thinking is, and what perception is. The last part of the course will be concerned with semantic externalism, which is currently the source of some very cutting-edge debates in the semantics of mental representation.