One of the most remarkable things about human cognition is the ability to take a finite number of words and combine them to generate an infinite number of new meaningful sentences. As an English speaker who has never heard the sentence or entertained the thought, “There are no bears on Mars”, you nevertheless have no trouble understanding what it means. Not only can you understand it, you can judge that it is very likely true and make conclusions on that basis: if there are no bears on Mars, that means there are no brown bears there, no bear cubs, no bears climbing Martian trees. How is it so easy for us to understand new sentences and thoughts, judge whether they’re true, and reason through to related thoughts? One plausible explanation is that the thinking of complex thoughts is the result of a rule-governed system that combines meaningful component units (e.g. words) in systematic ways.

This system – its structure and development – are the focus of my research. To study it, we must study what properties of a word or concept allow it to combine with others in the right way. Do the mental representations that participate in this system all have something in common, and if so, what distinguishes them from other kinds of representations that are not directly accessible to combinatorial language and thought, like, for example, representations of line orientations in visual cortex, or our implicit biases? If language is the most obvious way we combine meanings into more complex thoughts, does it play a special role in enabling the combination of meanings, or is it just how we communicate what we are able to do in our heads non-linguistically? Does learning a label allow that label’s meaning to productively combine with others, or does it merely attach a sound to a representation that already had all the right combinatorial properties?

Exploring these questions means exploring our shared humanity -- how all of us can think new thoughts so quickly and productively, and how we communicate those thoughts to each other.