January 5, 2006
Annual Meeting of the Linguistic Society of America
Albuquerque, NM

A central issue of debate and inquiry in theoretical linguistics has been the existence and nature of syntactic objects that do not correspond directly to elements in the speech (or gestural) signal. Ellipsis provides a particularly important and fertile empirical domain for addressing these issues, because it involves a mapping between an arbitrarily complex meaning and the complete absence of a phonological signal.

Traditional generative approaches to ellipsis have assumed that this mapping is mediated by unpronounced syntactic representations, for primarily two reasons. The first reason arises from the guiding theoretical intuition that in general, identity of meaning indicates identity of form; the semantic correspondence between an elided constituent and some antecedent is therefore taken to indicate the presence of a syntactic representation which is not pronounced but which forms the input to interpretation. The second reason is specific to ellipsis: the majority of analyses of ellipsis developed over the past forty years have assumed that an elided constituent is identical to some syntactic antecedent; that is, that ellipsis is licensed by a syntactic identity condition.

Recent work challenges both these underlying arguments for unpronounced syntactic structure in ellipsis. On the one hand, a diversity of approaches to the syntax-semantics interface have been developed which reject the hypothesis that identity of meaning entails identity of form, instead deriving identity of meaning from features of the interpretive system. On the other hand, a great deal of evidence has accumulated which shows that the identity condition in ellipsis should be stated over meanings, not syntactic representations. We think it is thus now possible and indeed necessary to reevaluate the arguments for and against representationalist accounts of ellipsis with a more nuanced eye, and to address directly the question of whether the crucial facts can be explained just as well within a purely interpretationalist approaches which eschew syntactic representations in ellipsis.

This symposium will bring together researchers with an interest and expertise in this domain to provide a state-of-the-art reevaluation of these fundamental questions. Specifically, we would like to focus the debate on the empirical and analytical arguments for positing or not positing unpronounced syntactic structures in ellipsis, in an effort to both sharpen our understanding of the mechanisms that handle ellipsis and further develop our understanding of the syntax-semantics interface. A broader purpose of the symposium is to bring various conflicting claims in the literature about the nature of the data under close scrutiny, with the goal of establishing agreement about what facts a theory of ellipsis needs to explain, independent of framework-specific assumptions.