In this presentation I will focus on a particular case study -- the antipassive construction of West Greenlandic Eskimo -- which has occupied my attention since 1982. At that time crosslinguistic research in modeltheoretic semantics was still virgin territory. The work therefore had to do start from ground zero -- namely, the development of a field method for investigating semantic intuitions of mostly monolingual speakers with no background in linguistics or logic. This field method revealed that antipassive objects in Eskimo systematically take narrower scope than transitive objects, relative to negation, modals, aspect, etc (Bittner 1987, 1995). Moreover, these exotic scope facts of Eskimo seemed to instantiate a larger crosslinguistic pattern of correlations between case and scope. Subsequent theoretical research sought to explain this pattern in terms of independently motivated universal principles of syntax (Bittner & Hale 1993, 1996a, b) and static compositional semantics (Bittner 1994a, b). However, recent developments in dynamic semantics (esp. Dekker 1993, Stone 1997, 1999) open up the possibility of an alternative explanation. This requires a new wave of fieldwork which, in turn, seems likely to raise some fundamental questions for the existing English-based theories.
Thus, fieldwork and theory seem to relate to each other like two strands of DNA -- though clearly distinct, neither is complete without the other.