Dissertation Abstract

The Structure of Thought

While the philosophical literature is replete with debates about whether perceptual experiences

are conceptual or nonconceptual, there is a widespread consensus that thoughts are universally

conceptual. My dissertation challenges this consensus, drawing on empirical results from the cognitive

sciences to argue that some thoughts are nonconceptual.

I begin the dissertation by explaining how I understand the notions of thought and conceptuality.

Very roughly, I hold that thoughts are contentful mental states of a subject that causally and inferentially

mediate between perception and action, facilitate learning and are stored in memory. I argue that

thoughts are appropriate objects of scientific investigation, and that they are possessed by animals and

humans alike. Whether thoughts are conceptual, I argue, is largely a question about their structure—in

particular, about whether thoughts are structured like sentences, and thus composed of concepts in much

the way that a sentence is composed of words. I isolate two prominent theories which maintain that

thoughts are conceptual in this sense: one which maintains that thoughts have conceptually structured

contents and the other of which maintains that thoughts have conceptually structured vehicles.

Armed with this understanding of thought and conceptuality, I consider and rebut several

arguments that philosophers have advanced in favor of the thesis that thoughts must be conceptual.

These include the arguments that thoughts must be conceptual to explain: (1) how inference is possible,

(2) how thinkers can produce an infinite number of new thoughts, (3) how thinkers manage to obey

certain closure conditions with respect to the thoughts they can think, and (4) how thinkers can reidentify

particulars over time. I defend the claim that while these arguments tend to show that thoughts must

have some structure or other, they do not establish that thoughts must have conceptual structure.

Having cleared space for the possibility of nonconceptual thoughts, I employ two strategies to

argue for their actuality. First, I isolate properties of some thoughts which are best explained on the

assumption that they are nonconceptual. For example, so-called analog magnitude thoughts, which

represent magnitudes such as number, time, distance and rate, give rise to Weber’s Law, which holds

that the ability to discriminate two magnitudes is a function of their ratio. (For example, it is easier to

discriminate 6 from 12 than 18 from 24.) On the assumption that these thoughts are conceptual, and thus

structured like sentences, there is no particular reason to expect this result. But if we instead assume

that these thoughts exploit an internal mental magnitude that is a direct analog of the magnitude in the

world which the thinker represents, we have a straightforward explanation of why thinkers obey Weber’s

Law: as the ratio of two worldly magnitudes approaches one, the two mental magnitudes become more

similar and are thus harder to discriminate.

Second, I demonstrate that there are essential properties of conceptual thought which are not

instantiated by all thoughts. For example, insofar as thoughts are conceptual, the concepts from which

they are composed should be capable of recombining freely like the words in a sentence. Thus, if you

can think that Amy is funny and that Bob is gracious, you should also be capable of thinking that Amy is

gracious and that Bob is funny. I argue that analog magnitude thoughts engender violations of this

condition (which is sometimes known as the “Generality Constraint”). We thus have reason to deny that

analog magnitude thoughts are conceptual.

The existence of nonconceptual thought sheds light on several further topics in the philosophy of

mind. For instance, philosophers who hold that perceptual experiences must have conceptual content

are often motivated by the claim that perceptual experiences must have the same kind of content as

thought, which they assume to be conceptual. However, if I am right that some thoughts are

nonconceptual, this motivation is undermined. As another example, consider the question whether

animals have thoughts. If we view all thought as conceptual, the minds of animals are bound to look

mysterious: we either over-intellectualize their minds by attributing conceptual thoughts to them or we

under-intellectualize their minds by denying that they have thoughts at all. By contrast, if we allow that

some thoughts are nonconceptual, we give ourselves the resources to attribute thoughts to animals

without ignoring the significant differences between their thoughts and our own.