neural substrate. And noone(Ithink) denies theimportance ofunderstandinghowtheneurons managetoaccomplish
language comprehension and production. So what might a result like Fig. 1.1 tell us about neural instantiation?
First, it is important tounderstand exactlywhat claims are made by thenotation.It is obvious that speakers don'thave
a directcounterpart of the symbol NP in theirheads. Rather,what is significant about the symbol is only that it differs
fro mthe other syntactic categories, not how it is labeled. Si milarly, in saying syntactic structure is modeled by a tree
structure, we are not claiming that speakers literally have trees in their heads. In fact, we often replace tree structures
such as (1a) with the“labeled bracket notation”illustrated in (1b). Some people even use a“box”notation like (1c),
and there are still other notations.
All of these are ways of notating the theoretical claims (a) that words belong to syntactic categories; (b) that words are
in linear order; and (c) that words group hierarchically into larger constituents that also belong to syntactic categories.
These aspects of the notation, then, must be reflected somehow in neural instantiation. Beyond these aspects, the
choice of notation is solely a matter of convenience.
With this understanding in mind, we can think of the combinationof the states of all the neurons in the relevant parts
of the brain as defining a“state-space”with a huge number of dimensions. When someone hears or produces the
sentenceThe little star's beside a big star, their brain can be thought of as being at some point in that state-space; it willbe
at another point for each different linguisticexpression. The notationin which Fig. 1.1 is couched encodes hypotheses
about the significant dimensions of the state-space, and each element in the notation encodes a position in one (or
more) of those dimensions.