This chapter provides a brief summary of preceding chapters, observing that the constructionist approach directly undermines the need for biologically determined knowledge that is specific to language (universal grammar). Generalizations are best described by analysing surface structure instead of positing an underlying level of representation (Chapter 2). The generalizations of language, like generalizations in other cognitive domains, are formed on the basis of instance-based knowledge that is retained (Chapter 3). Children are able to learn certain kinds of generalizations quite quickly, with skewed input like that commonly found in natural language playing a facilitory role (Chapter 4). Generalizations can be constrained by the indirect negative evidence children receive involving statistical preemption of non-occurring patterns (Chapter 5). Generalizations at the level of argument structure are made because they are useful, both in predicting meaning and in on-line production (Chapter 6). Classic island and scope phenomena can be accounted for by recognizing the discourse function of the constructions involved (Chapter 7). Generalizations that appear to be purely syntactic are at least sometimes better analysed in terms of constructions insofar as a patterns' distribution is typically conditioned by its functional role (Chapter 8). Cross-linguistic generalizations can be accounted for by appealing to pragmatic, cognitive, and processing facts that are independently required, without any stipulations that are specific to language (Chapter 9).
Oxford Scholarship Online requires a subscription or purchase to access the full text of books within the service. Public users can however freely search the site and view the abstracts and keywords for each book and chapter.
If you think you should have access to this title, please contact your librarian.