This paper concerns a particular problem raised by Mandarin Chinese pronouns, viz. they appear to obey a linear precedence constraint unlike English (e.g., Huang (1982)). This calls into question the nature of UG and how it can account for these crosslinguistic differences.In this paper, experimental analyses of children's (null and lexical) pronoun interpretation in Chinese and English argue for universal 'structure dependence' (including 'command') in the Initial State and against either a universal or a languagespecific role of 'linear precedence' alone. A linear precedence effect is developmentally achieved only in Chinese.The acquisition results provoke a revised theoretical analysis of the grammar of pronouns in Chinese and a strong form of UG. We argue that it is not necessary to propose a language-specific definition of 'command' in Chinese in order to explain the apparent linearity effects on Chinese pronouns, and it is not necessary to propose a linear precedence rule in UG. Rather, consideration of an articulated structure of Chinese NP, which is motivated by the acquisition data, explains essential differences between lexical ta pronominals and null pronominals in Chinese and accounts for linearity effects in the adult language. We propose that ta pronominals are not themselves in argument position and are not N o heads. They are not equivalent to null pronominals.Together the experimental and theoretical results support a 'strong continuity' theory of UG in which universal "principles and parameters" of UG continuously constrain the child's mapping from UG to a specific-language grammar. Language development, and the Chinese precedence effect for pronouns, lies in pragmatic/semantic features connected with the lexical realization of the specifier of a pronominal NP, not in the development of UG. tigation we are led to new insights regarding the grammar of Chinese pronominals and its relation to Universal Grammar (UG). This paper represents one component of a cross-linguistic project in which first language acquisition of English and several other languages are compared in order to investigate the nature of principles and parameters in UG. We assume the theory of UG (as summarized in (1)) to provide a model of the "Initial State" for language development (cf. Lust (to appear, in preparation)).