tailieunhanh - Báo cáo khoa học: "Bridging the Gap Between Underspecification Formalisms: Hole Semantics as Dominance Constraints"

We define a back-and-forth translation between Hole Semantics and dominance constraints, two formalisms used in underspecified semantics. There are fundamental differences between the two, but we show that they disappear on practically useful descriptions. Our encoding bridges a gap between two underspecification formalisms, and speeds up the processing of Hole Semantics. | Bridging the Gap Between Underspecification Formalisms Hole Semantics as Dominance Constraints Alexander Koller Joachim Niehren Stefan Thater koller@ niehren@ stth@ Saarland University Saarbriicken Germany Abstract We define a back-and-forth translation between Hole Semantics and dominance constraints two formalisms used in underspecified semantics. There are fundamental differences between the two but we show that they disappear on practically useful descriptions. Our encoding bridges a gap between two underspecification formalisms and speeds up the processing of Hole Semantics. 1 Introduction In the past few years there has been considerable activity in the development of formalisms for underspecified semantics Alshawi and Crouch 1992 Reyle 1993 Bos 1996 Copestake et al. 1999 Egg et al. 2001 . These approaches all aim at controlling the combinatorial explosion of readings of sentences with multiple ambiguities. The common idea is to delay the enumeration of all readings for as long as possible. Instead they work with a compact underspecified representation for as long as possible only enumerating readings from this representation by need. At first glance many of these formalisms seem to be very similar to each other. Now the question arises how deep this similarity is - are all underspecification formalisms basically the same This paper answers this question for Hole Semantics and normal dominance constraints two logical formalisms used in scope underspecification by defining a back-and-forth translation between the two. Due to fundamental differences in the way the two formalisms interpret underspecified descriptions this encoding is only correct in a nonstandard sense. However we identify a class of chain-connected underspecified representations for which these differences disappear and the encoding becomes correct. We conjecture that all linguistically useful descriptions are chain-connected. To support this claim we .