Case and the Parameterization of Scope Ambiguities

Abstract

This paper argues for a version of Aoun and Li's account of quantifier-scope ambiguities, and for treating the English/Chinese contrast in ambiguity patterns as a difference in how Case is assigned to subjects: under government (Chinese) vs. agreement (English). I argue that the Copy Theory of movement allows the prediction of ambiguity in Someone must have left and Everyone doesn't like squid; English subjects in Spec,IP optionally delete at LF, allowing their VP-internal copies to adjoin below must/n't (giving the subject's narrow scope reading). But subjects with governed Case cannot delete, so the Chinese equivalents are unambiguous. The same distinction underlies the Chinese/English difference for (e.g.) Everyone recommended a book. English subjects, with agreement Case, allow objects to raise over them at LF; but Chinese subjects, with governed Case, block objects from raising, preventing ambiguity. Evidence for this account comes from correct predictions about ambiguity in sentences with double objects or PPs; correlations with ECP phenomena; and the fact that Japanese, Korean, and Persian behave (largely) like Chinese.