Unifying everything: Integrating quantitative effects into formal models of grammar

Abstract

Quantitative effects can be divided into two general groups: syntagmatic effects like collocations and grammaticalization, and paradigmatic effects like family size, entropy, alternations and collostructions. By now we have more than ample evidence that quantitative effects play a role in language processing, acquisition and language change. It is also clear that core grammar is, at least for the most part, independent from quantitative effects, eg. the frequency of a construction says nothing about its grammatical properties. What is still sorely lacking is an interface between grammar and usage that allows the usage module to access grammar in a systematic way.