Null Subjects in Early Child English and the Theory of Economy of Projection

It is well-known that young children may omit referential subjects regardless of whether they are acquiring a pro-drop language such as Italian or a nonpro-drop language such as English. The classic proposal of Hyams (1986) according to which these early null subjects instantiate pro in both types of languages has recently come under attack from various sides. Bloom (1990, 1993) and Valian (1991) argue that missing subjects in early child English are a non-syntactic performance phenomenon that is due to a production bottleneck which severely limits the utterance length of young children, a view which they support with an inverse correlation between subject-length (i.e. full NP, pronoun, null) and VP-length. Rizzi (1994a,b) and Hyams (1994) maintain that empty subjects in early child English are a syntactic phenomenon but relate them to adult English Diary Drop and German-style Topic Drop instead of Italian-style pro-drop. In particular, they argue that like adult Diary /Topic Drop and unlike adult pro-drop, these missing subjects in early child language are restricted to the first position of non-Wh root clauses. In this paper1, we present new evidence from Adam (CHILDES, Brown 1973 and MacWhinney in press) that suggests that not all missing subjects in early child English can be reduced to performance limitations or Diary/Topic Drop. Between age 2;3 and age 2;11 (files 1-18), Adam produces numerous Wh-questions without an overt subject (e.g. "Where go?") and the VP-length of these examples is not greater than the VP-length in Wh-questions with an overt subject pronoun. Adam's data moreover display a clear-cut distinction between finite (i.e. agreeing) and non-finite (i.e. non-agreeing) Wh-questions. Whereas the number of empty subjects in finite Wh-questions is negligible, there are almost as many non-finite Wh-questions without an overt subject