Adventures in long-distance moving: The acquisition of complex Wh-questions

Date of Completion

January 1990


Language, Linguistics




This thesis investigates the acquisition of long-distance questions and the principles of Universal Grammar that govern them (Chomsky 1981;1986). The questions under investigation are all ones in which a Wh-phrase is moved from an embedded clause, leaving a Wh-trace at the extraction site. Using an elicited production methodology, it is demonstrated that even three-year-old children readily produce a variety of long-distance questions. The methodology is refined to probe children's knowledge of two principles that constrain production of long-distance questions. One principle governs wanna contraction, prohibiting contraction across a Wh-trace in extracting from infinitival clauses, as in (1). The second principle, the Empty Category Principle (ECP), restricts extraction from the subject of a tensed embedded clause, ruling out questions like (2) with an overt complementizer (that):^ (1) *Who do you wanna help you? (cf. Who do you want to help you?)^ (2) *Who do you think that's in the box? (cf. Who do you think is in the box?)^ Another methodology, the Truth Value Judgment task, supplements the data from elicited production by studying a constraint that restricts the meanings that can be assigned to long-distance questions. This constraint of "Strong Crossover" prohibits Wh-movement over a pronoun with the same index, ruling out a bound interpretation of the pronoun, as indicated in (3):^ (3) *Who$\sb{\rm i}$ does he$\sb{\rm i}$ think t$\sb{\rm i}$ has a hat? (cf. Who$\sb{\rm i}$ thinks he$\sb{\rm i}$ has a hat?)^ The findings were that children adhered to the constraints on wanna contraction and Strong Crossover. In extracting from tensed clauses, however, several children consistently produced ungrammatical questions like (2), in apparent violation of the ECP. Other non-adult questions like (4) appeared. I call these "medial-Wh" questions.^ (4) Who do you think who's in the box?^ To explain the problematic data, I explore several current formulations of the ECP. A version by Rizzi (1990) is seen to provide a unified explanation of children's exceptional questions of both kinds. In this framework, these questions are not interpreted as ECP violations but, rather, as attempts by children to preserve the ECP. In sum, the findings support the view that, by three, children have mastered long-distance movement questions and the constraints that apply to them. ^