While the visual system is busy creating a feature map of a face, the auditory system is doing the same: detecting sounds that higher-level feature detectors identify as phonemes that, even further upstream, are recognized as words.

But if concepts are definitions, why, according to Rey, do subjects make the sorts of typicality judgments that psychologists cite in support of probabilistic theories of concepts?

Given that abstract objects have no spatial location, it follows that concepts, if abstract objects, are not in the head, because they are not anywhere. Psychologists, on the other hand, tend to characterize concepts as mental representations that are realized in the brain, just like written representations might be on a page, or pictorial representations on a canvas.

Right- and left-handers, for instance, reliably differ in the preferences they express toward the same items, and their attribution of goodness to items appears to be a function of their handedness. Similarly, changes in an individual’s body are positively correlated with the individual’s capacity to apply emotion concepts.

You come across a word, such as “knish,” that you don’t understand. The word, has a meaning, but it is not meaningful for you. In coming to understand the meaning of the word, you are drawing on concepts, for instance, PUCK-SHAPED POTATO-FILLED CRUST, that are themselves meaningful.

This point returns us to my earlier distinction between a theory of linguistic meaning, which will articulate the conditions that a linguistic symbol must satisfy in order to become meaningful, and a theory of language learning, which will seek to discover the psychological processes involved in understanding a language.