The Acquisition of Restrictions on Pronominal Reference

Abstract

The business of linguistic theory is to provide an empirically adequate description of the adult language, with the stipulation that it is learnable, given a theory of universal grammar and some account of the ways in which children take advantage of linguistic input. Although the goal of constructing learnable grammars has led linguists in somewhat different directions (compare, for example, Chomsky, 1981 with Culicover and Wexler, 1977, and Wexler and Culicover, 1980), the question of which description is to be preferred, is strictly a factual one. Exactly what mental structures limit the child’s ability to form hypotheses, or in other words, exactly what is contained in universal grammar, is subject to empirical investigation. Once this fact is ascertained, then some of the possible linguistic descriptions consistent with it can also be ruled out as unlearnable, given a reasonable set of assumptions about learning theory, as the work by Culicover and Wexler points out.