This chapter has four parts. In the first part I expand the notions of uniqueness introduced in chapter I. There, the nature of topics was determined and it was concluded that they must be specific, i.e., that a card must be defined upon which the entry defined by the predication can be made. In this chapter I will investigate in more detail how new cards are opened and defined and what makes them specific or unique. I will show that this depends on whether the sentence is interpreted as predicated of a stage topic or whether it is interpreted as predicated of an individual topic. I treat the ambiguity which occurs in intensional contexts as the outcome of the two different f-structures; one in which there is a subordinate f-structure involving an individual giving rise to the specific reading, the other in which the subordinate f-structure involves an “intensional” stage giving rise to the nonspecific or transparent readings.

In the second part I discuss the function of focus structure in extended discourse. I distinguish three types of discourse chains; those in which the topic remains constant, those in which the topic changes according to the focus of the previous sentence and those in which a topic set is introduced allowing for any member of this set to be a topic in the following unit of discourse.

In the third part I offer an account of pragmatic anaphora along the lines proposed in Reinhart (1983, 1986). An f-structure-theoretical account for the distribution of coreferent pronouns is provided by the assumption that pronouns must be topics, i.e., they must access cards from the top of the file.