Stage 6 of the journey addresses various aspects of how information is conveyed and organized in a sentence, beginning with the stubborn problems that led to the development of dynamic semantics (‘donkey sentences’ and cross-sentential anaphora), through properties of expressions that are used to refer (descriptions, proper names, indexical expressions), to the organization of information (presuppositions and projective content, topic, focus, coherence). As such, it is a step to the next stage that concerns utterance meaning.

Chapters 9–10 explore how the assignment-variable-based treatments of relativization and pronominal anaphora with noun phrases may be extended to capture shifting phenomena in other clausal structures, such as conditionals, correlatives, and questions. This chapter develops a syntax/semantics for ‘if’-clauses as free relatives, interpreted as plural definite descriptions of assignments (possibilities). Clause-internal movement of a (possibly implicit) definite/maximality operator introduces quantification over assignments. The compositional semantics affords a uniform analysis of ‘if’-clauses in diverse types of conditionals—conditionals with ‘if’-clauses in post-nominal, sentence-final, and sentence-initial positions, and conditionals with/without an overt modal or ‘then’. The choice-function-based analyses of anaphoric pronouns as copies of a linguistic antecedent are extended to anaphoric proforms in correlatives. Comparisons between conditional correlatives and correlatives of individuals are briefly considered.

Chapter 4 outlines a procedural relevance-based analysis of the definite determiner the. The definite article, it is argued, signals to the hearer that he should seek out an existing conceptual file on which to resolve reference. The indefinite article, on the other hand, instructs the hearer to open a new conceptual file. As interpretation proceeds, the hearer seeks to align the conceptual content within the nominal of the definite description with that in the target conceptual file. This approach to the contribution that definite descriptions make to speaker meaning is then applied to cases of misdescription and, it is claimed, it also offers fresh perspective on the referential–attributive distinction. Finally, stylistic effects which may arise from the choice and content of definite descriptions are discussed.