Chapter 8 — Meaning, memory, prediction

Where the brain explains the signal

By the end of movement 8, “Hey Dr. Miles!” has been faithfully decomposed by the cochlea, encoded by the auditory nerve, localized by the brainstem, classified by primary auditory cortex, parsed into phonemes by the superior temporal gyrus, and assembled into words by higher temporal areas. The cortical pattern of activation is now, in some informational sense, a representation of the input phrase.

But it is not yet a meaning. A meaning, in the sense I am going to use the word, is what the listener actually understands the sentence to be saying, in context, for them, right now. It includes the recognition that they have been addressed, the inference of what response is expected, the priming of relevant memories about who Dr. Miles is, the social-emotional valence of being greeted by name. The neural substrate of all of this extends well beyond auditory cortex into temporal, frontal, and limbic areas. None of it is in the cochlear output.

This movement is about the gap between representing the input and understanding what it means. The thesis is that the brain bridges that gap by inference — by reasoning about what the input is most likely to be given everything else the brain knows.