Conclusion

Abstract

Imagine a pair of florists discussing an arrangement of roses and violets. In the course of the conversation, they might have occasion to use any of the following three noun phrases: the flowers, the roses and the violets, and the arrangement. What is the relation between the referents of these noun phrases? Does the language treat them as referring to objects with the same part-whole structure? Are they coreferent? These kinds of questions have been the focus of the preceding pages. In trying to answer them, we appealed for the most part to four linguistic phenomena. The first was semantic selection, in particular the sortal restrictions imposed by predicates on their noun phrase arguments. The second was distributivity, whereby a speaker names a plurality and attributes a property to parts of that plurality. The third phenomenon was reciprocity whereby a speaker names a plurality and claims that a relation holds between parts of that plurality. The fourth phenomenon had to do with quantification. In natural language, quantification involves a quantifier and a domain over which that quantifier quantifies. We studied cases in which a noun phrase is used to name the domain of quantification. An important theme running through much of the discussion was the difference between an answer to one of the above questions that lies in the realm of pragmatics versus one that is semantic. An example of the former might involve the claim that the part-whole structure enters in as the result of a negotiation between speakers in a conversation, formally represented as a free variable in a semantic representation. An example of the latter would take the part-whole structure to be inherent in the entity referred to.