The formal syntax of alignment change

要旨

Previous accounts claim that the change from accusative to ergative syntax involves reanalysis of a construction, such as passive. This chapter focuses instead on the specific parameter settings that determine ergative, active, or accusative syntax and what leads language learners to change them. The crucial parameter determining non-accusative syntax requires that v assign inherent case to the external argument in its specifier. The chapter argues that the well known case of tense/aspect-based split ergativity in Indo-Iranian does not result from reanalysis of a passive, but from reanalysis of a 'quirky' possessor construction which already had this basic parameter setting, following an influential proposal of Benveniste. It shows how the change from split active to accusative in Old and Early Middle Japanese results from the attrition of inherent case marked subjects in situ, largely through changes in the pronominal system. Alignment shift emerges clearly as a representative example of parametric change.

主題