The child, the bioprogram and the input data: a commentary on Cziko

The syntax of any given language seems now to be best regarded as the result of interaction between a small set of principles, invariant for all species members, and the properties of the target lexicon (themselves selected from an inventory of all language-possible lexical properties). Creoles thus owe their similarities to the fact that, for every creole, the same set of principles interacts with a lexicon that has, in each case, been largely stripped of grammatical morphemes and their properties by the process of pidginization. What difference should this make to our view of acquisition? Instead of blindly pursuing an unrolling genetic program until it collides with a wall of resistant environmental fact, the child is perhaps better conceptualized as scanning input for morphemes that will have the properties entailed by syntactic principles. Let us see how this would affect the acquisition of TMA. Among universal principles are those that establish X’-structure. From these the child knows that in order to combine subject (NP) and predicate (VP), the latter must be inserted as specifier and complement of a head (INFL) so as to form the superphrase IP. The child is therefore on the