Choice Functions in Semantics

Klaus Heusinger, Ruth Kempson
2004 Research on Language and Computation  
This article argues that definite NPs are interpreted depending on contextual salience, rather than on the uniqueness condition of their descriptive content. The salience structure is semantically reconstructed by a global choice function that assigns to each set one (most salient) element. It is dynamically modified by the context change potential of indefinite and definite NPs. The anaphoric potential of definite NPs can be accounted for by the interaction of the context change potential and
more » ... ontextual salience structure. , Bruce 1 , ii who has been making himself 1 very salient by dashing madly about. Abstract. This paper develops a unified analysis of "functional" anaphora arid wide-scope indefinites. A new operator is added to Jacobson's variable-free semantics of functional readings, which leads to an analysis of theses readings using the general Skolem function interpretation of wide-scope indefinites. This accounts for the distriutional, technical and intuitive similarities between the two phenomena. Moreover, after formally characterizing the class of generalized quantifiers that are treated by the proposed mechanism, it is argued that this class is a good approximation of the quantifiers that empirically support functional readings. Abstract. This paper sets out to give a natural pragmatic explanation of several aspects of the interpretation of singular indefinite noun phrases. We develop a uniform account of characteristic features of their use which have been dealt with only partly in other semantic paradigms (in particular the dynamic, the E-type and the choice function one). We give an intuitive motivation for the familiar discourse dynamic features of the use of these expressions, and, taking due account of the structuring of information in more involved contexts, account for their behaviour in negated, conditional, quantified, and intensional constructions. Abstract. In this paper, I argue that the 'ambiguity' between mention-all and mentionsome readings of questions can be resolved when we relate it to the decision problem of the questioner. By relating questions to decision problems, I (i) show how we can measure the utilities of both mention-all and mention-some readings of questions, and (ii) give a natural explanation under which circumstances the mention-some reading is preferred. Abstract. In Counterfactuals, David Lewis noticed that definite descriptions and conditionals display the same kind of non-monotonic behavior. We take his observation literally and suggest that if-clauses are, quite simply, definite descriptions of possible worlds [related ideas are developed in M. Bittner (2001) Proceedings from SALT XI, CLC, Cornell University, Ithaca, pp. 36-55]. We depart from Lewis's analysis, however, in claiming that if-clauses, like Strawsonian definite descriptions, refer. We develop our analysis by drawing both on Stalnaker's Selection Function theory of conditionals and on von Heusinger's Choice Function theory of definiteness, and by generalizing their analyses to plural Choice/Selection Functions. Finally, we explore some consequences of this referential approach: being definites, if-clauses can be topicalized; the word then can be analyzed as a pronoun that doubles the referential term; the syntactician's Binding Theory constrains possible anaphoric relations between the if-clause and the word then; and general systems of referential classification can be applied to situate the denotation of the descriptive term, yielding a distinction between indicative, subjunctive and 'double subjunctive' conditionals.
doi:10.1007/s11168-004-0903-7 fatcat:ye4rpurgcfhebaxlijllutq54a