2009. Journal of South Asian Linguistics 2:5-34.
This paper presents an alternative account of DP distribution that is based on DPs being selected rather than being Case-theoretically licensed. We argue that the fundamental prediction made by Case theory, namely that obliga- torily controlled PRO and overt DPs are in complementary distribution, is not empirically justified. To this end, we provide data from nonfinite clausal adjuncts, complements and gerundivals in Tamil where subject controlled PRO and overt subject DPs seem to alternate in free variation. We further illustrate, with supporting evidence from Malayalam, Sinhala, Latin, Irish, and Middle English as well as the Present-Day English gerundival construc- tion, that this type of problematic alternation is not a language-specific quirk but a widely attested crosslinguistic phenomenon. While standard Case theories are equipped to handle either the occurrence of PRO or that of an overt subject, they are unable to consistently handle the alternation between both types of elements. Our selection analysis is designed to handle the alternations as well as instances where only one DP type is allowed.