Chapter 13 Concluding Remarks

The promise of generative grammar in the 1960s was that it would give us insight into the nature of mind and brain.
The path we have traveled in the past several hundred pages has, I think, put us in a better position to realize that
promise. A quick review will help see why.

The goal of Part I was to show that the promise should not be relinquished. The most rewarding way to investigate
language is stillfro mthe mentalistpoint of view—focusing especiallyon what ittakes for a child to acquire facility in a
language. The last thirty years have brought phenomenal progress in understanding linguistic structure, language
acquisition, language processing, and language deficits. These years have also brought many important advances in
understanding the brain and the learning capacities of animals, humans, and computationally modeled neural systems.
Everything about language appears richer and more complex than it did thirty years ago, thanks in part to the better
tools we have for looking at it now, but thanks also to decades of patient and impassioned research, exploring more
ground more carefully.

Theupshotofallthiswork,onbalance,is thatwecanconsiderablysharpenthequestionsposed bymentalism.Instead
of simply asserting that language must be lodged in the brainsomehow, and debating whether children must come to
language acquisition withsomeinnate capacities specific to language, it is now possible to articulate the issues more
clearly. Among the questions that have emerged here as critical are these:

• How are discrete structural units (such as phonological features and syntactic categories) instantiated in the
  nervous system?


• How are hierarchical structures, composed of such units, instantiated in the nervous system?


• How are variables instantiated as elements of such structures in the nervous system?


• How does the brain instantiate working memory, such that structures can be built online from units and
  structures stored in long-term memory, and in