My research centers on pre-linguistic mental representations from faculties such as perception, the number sense, and nonhuman cognition. I mine philosophy and the cognitive sciences to illuminate their format, content, and conscious properties.


Perception is the most primitive form of mental representation, and my research elucidates its nature. Drawing on psychophysics, I have argued that perception has an analog format (like a mercury thermometer). I have also appealed to empirical research on perceptual noise to develop a new version of the argument from hallucination against relational theories of perception. And I have argued that perception (e.g., seeing red) is distinguished from cognition (e.g., imagining or thinking about red) by virtue of being stimulus-dependent.

A glaring fact about perception is that it is often conscious. I have argued that the conscious properties of perception cannot be reduced to its representational properties because attention can alter salience, a conscious property of perception, without altering what is perceptually represented. I have also argued that probabilities or confidences are not reflected in perceptual experience.

Analog Magnitudes and the Number Sense

Empirical evidence suggests that humans and many other animals have an a primitive ability to represent a range of magnitudes, including duration, distance, and number. I have argued that these representations have an analog format and nonconceptual content. I have also defended the thesis that they represent numbers against critics who allege that they only represent non-numerical confounds such as area and density, or unfamiliar substitutes for number such as “numerosities.” One upshot is that there is a phylogenetically widespread mode of pre-linguistic cognition that supports abstract contents but differs from conceptual thought.

Animal Cognition

The existence of pre-linguistic mental capacities such as perception and the number sense raises an intriguing prospect. Can we fully account for nonhuman cognition by appealing solely to nonconceptual/nonlinguistic mental representations? To address this question, I have explored various means of determining whether nonhuman animals engage in conceptual thought or have a language of thought.

One source of skepticism about animal thought derives from our inability to precisely specify its contents—to say just what animals are thinking. But I have argued that such difficulties are to be expected if animal thought has a nonlinguistic format. That’s because translations across formats are always imprecise. Trying to characterize animal thought is like trying to describe the Mona Lisa. Approximations are possible, but precision is not.

Conceptual Thought

While most of my research centers on pre-linguistic mental representation, I have also studied conceptual thought, which underlies linguistic understanding. I have shown how a Fregean account of concepts can make room for modes of presentation that are not reducible to symbols in a language of thought. And I have built on Susan Carey’s account of concept learning to explain how thinkers can learn concepts that are genuinely new.