Computational Grammar Induction for Linguists

In general a grammar describes a (possibly infinite) set of sentences with a finite structural description. Computational Grammar Induction (CGI) deals with the creation of computational models for identification of these infinite sets on the basis of a finite set of examples. CGI is a field in its own right, with its own internal research questions, many of which have no direct impact on the study of human language. Yet it is clear that computational models created by the CGI community might be of interest to the linguistic community because human language after all appears to be an infinite set, the description of which is learned efficiently in a relative short time. There are various domains in which learnability of a language might be interesting for linguists: e.g., first language acquisition, second language acquisition, or the automatic extraction of grammars from corpora. In this article we will focus on first language acquisition with some suggestions for extensions to other areas. with some suggestions for extensions to other areas. Up till now there has been surprisingly little cross-fertilization between the field of linguistics and that of Computational Grammar Induction. There are various reasons for this situation, for example, there is a strong interest in the descriptive study of language in linguistics on the one hand and a strong focus on abstract language models in the CGI community on the other hand. A factor that has certainly contributed this lack of cross-fertilization is the famous conjecture of Chomsky that the efficiency of human language acquisition can only be explained on the basis of an innate Universal Grammar (UG). The UG discussion, which is in fact a revival of the ancient philosophical debate of rationalism (Descartes)