\input zb-basic
\input zb-ioport
\iteman{io-port 01794690}
\itemau{Pinkal, Manfred}
\itemti{On semantic underspecification.}
\itemso{Bunt, Harry (ed.) et al., Computing meaning. Vol. 1. Dordrecht: Kluwer Academic Publishers. Stud. Linguist. Philos. 73, 33-55 (2001).}
\itemab
This paper, devoted to underspecified semantic representations (USRs) of natural language (NL), provides an overview on the motivations for USR use, explains what USRs may be taken to mean and how they can be used in reasoning. USRs are shown to be efficient not only when faced with the ambiguity met in NL processing, but also in allowing robust processing for incomplete linguistic information, as often happens in NL speech and text understanding. An important discussion is on USR meaning; the author argues that to understand a USR as a partial description of logical formulae, or as a disjunction of the set of unambiguous logical formulae which it describes, should be inadequate. To approach the USR meaning, the paper examines various possibilities for defining the entailment relations for USRs. Manfred Pinkal suggests a distinction between two USR entailment concepts, a `dynamic' one that takes into account the phenomena of discourse parallelism, and a USR `static' entailment, that does not. The conclusion would be that USRs form a layer of information, which may be truth-conditionally irrelevant but which is indispensable for discourse semantics. The investigated concepts and ideas on USR of NL are intended to be applied within the CHORUS project, Saarbr\"ucken University (Germany).
\itemrv{Neculai Curteanu (Ia\c{s}i)}
\itemcc{}
\itemut{underspecified semantic representation; natural language semantics; robust processing; incomplete linguistic information; natural language understanding; discourse theory}
\itemli{}
\end