Innate constraints on language variation: Evidence from child language

Date of Completion

January 2003


Language, Linguistics




Within the Principles and Parameters approach to Universal Grammar (Chomsky 1981), language acquisition is assumed to be the process of setting the values of parameters, which are conceived of as innately-specified points of grammatical variation that have multiple consequences for the surface grammar. Given this view, it is expected that parameter-setting, more accurately the time required to accommodate the data indicating the correct parameter-settings plays an important role to explain why language acquisition is not ‘instantaneous’ and proceeds gradually. Yet, despite this expectation, few pieces of clear evidence have been provided for parameter-setting from child language acquisition. This situation has led to the recent, influential hypothesis by Wexler (1996, 1998), which claims that basic parameters are set correctly at the earliest observable stages (Very Early Parameter-Setting, VEPS). ^ In this thesis, I will present evidence against the “strongest” form of VEPS, the hypothesis that all the parameters are set to the adult value at the earliest stages. The evidence comes from the acquisition of three syntactic properties: preposition stranding, scrambling, and resultatives. A strong acquisitional association has been found (i) between preposition stranding and the prepositional complementizer construction, (ii) between the multiple-nominative construction and Japanese-type scrambling, and (iii) between noun compounding and transitive resultatives. These results not only argue for the existence of the relevant parameters, but also constitute a clear indication that parameter-setting in fact plays a significant role in explaining the non-instantaneous and gradual nature of language acquisition. These findings in turn demonstrate that the time course of child language acquisition is a rich source of evidence concerning the innate constraints on language variation. ^