Agreement and Anaphora

Agreement has been seen to be of very great importance in the operation of anaphora at a number of points in the book so far: in Chapter 3, in particular, which I ended by suggesting, with Brame (1983), that the three types of 'strict' nominal anaphora investigated there could be analysed as involving a two-stage agreement process as a result of function-argument interpretation; and in Chapter 5 (cf. 5.4), we saw illustrations of differences in referential perspective upon a human referent that could be achieved via the use of two distinct modes of agreement within a third person pronoun in French, as well as the way in which the marking for gender in an 'antecedentless' third-person pronoun in that language was able, in conjunction with various contextual functions, to make available a basic-level noun-predicate in the language in question which matches its gender value and conventionally names the type of entity of which the actual referent is an exemplar. In the first type of instance (the operation of 'strict' nominal anaphora seen in Chapter 3), the agreement process involved is more or less automatic, the anaphor performing an essentially grammatico-semantic operator-like function in determining both the nature of the derived predicate of which it is an intrinsic part, and, through this, the selection of a potential argument to which that derived predicate can be applied as a function. The choice of controller argument is, however, strictly limited by the grammatical constraints under which this type of anaphor functions. ¹