ABSTRACT
Abstract
The processes of reading and writing present a problem for those who claim that linguistic processing is modular. How is it that the language module, specialized to respond only to speech-like acoustic patterns, can apparently respond also to optical patterns of arbitrary form? It is proposed that cognitive linguistic representations such as those that the module creates can also drive the module, even if these representations are incomplete, and even if they are actually of cognitive origin. Thus, given a convention for transcribing such incomplete representations (an orthography), it is possible to exploit the language module in reading and writing. But what is the biological function of an arrangement in which cognitive linguistic representations and not just semantic ones can be modular inputs as well as modular outputs? Not communication, which could have been managed more straightforwardly without linguistic representations external to the module. The function of this arrangement is rather to bring about language change and diversity and thus, as Nottebohm has proposed, to facilitate subspeciation.
