Animacy, Argument Structure and Unaccusatives in Child English

We offer evidence that children correctly represent unaccusative verbs as taking a single internal argument, rather than an external argument, and they raise this argument to subject position via A-movement, contra Babyonyshev et al. (2001). Our primary evidence comes from the distribution of animate vs. inanimate subjects with intransitive verbs. Perlmutter (1978) argues that inanimate subjects are generally compatible with unaccusative but not unergative verbs. Our data show that children correctly use both animate and inanimate subjects with unaccusatives, and only animate subjects with unergatives. In addition, children correctly limit use of resultatives to unaccusative verbs, and errors involving postverbal subjects occur disproportionally with unaccusatives. If children misanalyzed unaccusatives as unergatives, as argued by Babyonyshev et al. (2001), we would expect similar distributions of these verbs in all of these respects. Thus, we argue that children by age 3 have an adultlike representation of unaccusatives and therefore both respect UTAH (Baker 1988) and represent A-chains.