While the aim of language acquisition research is to describe the nature of the child's developing grammar, we face a challenge in that the grammar itself is not visible to us- we must rely on observable behavior. And further, children's behavior is not directly indicative of their grammatical knowledge, in two important ways. First, behavior is inherently a function of both a child's grammatical knowledge and the deployment processes required to implement this knowledge in real time. Second, in certain situations both accurate adult-like grammar and a grammar relying on a non-adult heuristic can yield the same behavior.

I'll present as a test case research investigating the developmental pattern of children's acquisition of Principle C. Principle C is the binding constraint which blocks a reflexive interpretation of sentences like "she likes Katie" (i.e such a sentence can't mean that Katie likes herself). I will show evidence from two lines of study exploring children's understanding of Principle C at 30 months. The first capitalizes on differences in behavior which are indicative of differences in deployment capabilities; from this we can make inferences about the underlying knowledge that children are relying on in terms of whether or not that knowledge is structurally dependent as in adults. The second line of studies aims to contrast behavior predicted by adult-like knowledge of Principle C and that predicted by various possible non-adult heuristics that children could use as alternatives; I will show that in each case adult-like knowledge of Principle C is the only way to account for all of children's behavior at 30 months. Taken together, these results suggest that at the youngest age at which we are (as of yet) able to test children's knowledge of Principle C, children are adult-like not only in their behavior but also in the underlying knowledge which drives it.