Aspects of the Constructional Nature of Agreement
This contribution is intended to show that agreement is a schematic construction (Culicover & Jackendoff 2005: 192; Steels 2011: 27; Hoffmann 2013: 310), of the core kind (like Goldberg´s argument structure constructions), and that this construction shares with the rest of the constructicon the very same cognitive principles that structure it: categorization, metaphorical extension, metonymy, Gestalt formation and prototypicality effects, usage-based self-structuring, and portions of idiosyncrasy which nevertheless show signs of resting on independently identifiable general cognitive principles. Like all constructions, agreement has a symbolic nature, with a form side (formal co-variance) and a meaning side (‘Unify’). Besides those general dimensions, two particular facts about the grammar of agreement need to be accounted for: the first is the Agreement Hierarchy (Corbett 1979; 2006); the second is the fact that, given vast cross-linguistic differences in the size of the morphological component, which agreement capitalizes on, we should be able to see how languages exhibit different ‘ecologies’, that is, different reflexes of the size (and the frequency) of their agreement systems in the rest of the grammar. Although this work has no serious typological pretensions, agreement operations will be illustrated using English and Spanish, as these two languages show two rather extreme positions in their morphological repertoire. This approach thus assumes that agreement (like transitivity or caused-motion) may be seen as a cross-linguistically valid construction with particular instantiations, and that these instantiations depend to a very large extent on the size of the morphology in a way that will be made precise. Finally, experimental evidence will be used to argue that agreement is very strongly usage-based.