This paper distinguishes two types of language, V-to-I type vs. V-to-V type, with a view to deriving two distinct patterns of associating (non-)specific interpretations with subject positions. In the former type, verbs move to a higher functional head, either overtly or covertly, and non-specific numeral indefinites may appear in [Spec, TP] at S-structure. This is the case with English, where cardinal subjects are commonplace. In the latter type, verbs do not move to a higher functional head, and non-specific numeral indefinites cannot appear in [Spec, TP] at S-structure. This is the case with Chinese, where cardinal subjects are relatively rare. Developing the idea that syntax-semantics mapping is cyclic and closely aligned with syntactic predication, we provide a principled account of an asymmetry between Chinese declarative and modal constructions with respect to their subject specificity.