Coarse coding and the lexicon
Catherine L. Harris · Lingvisticae investigationes. Supplementa · 1994
and enumerating the diverse senses of polysemous words. The former approach captures our intuitions about what is common across a word's uses, but seldom specifies details of the possible range of uses. The latter approach accounts for the range of uses, but obscures their commonalities. Unfortunately, both these approaches put off to be solved another day the problem of explaining how rules for words' contextual integration act to produce the specific interpretation we obtain on hearing words in context. In this chapter I argue that adopting a memory-based approach to lexical representation will illuminate the tension between abstractness and specificity of representation, as well as helping with the question of contextual integration. There are two key ideas in the memory-based approach. The first is that units larger than words (such as phrases, clauses, conventional collocations and idioms) are the primary storage unit. On this view, words are laid down in memory w