The lexical profile of modern American proverbs: Detecting contextually predictable keywords in a database of American English proverbs

Abstract Proverbs (as Time is money) are conventionalized expressions that are “equivalent to a sentence” and “express generalized experiences or value judgements” (Steyer 2015: 209-210). This study aims at describing the lexical structure of the proverb inventory and at identifying ‘proverbial keywords’ which may be assumed to play an important role in storing proverbs in the mind (Luckert 2018). A database of American English proverbs was compiled and content words were tested for their contextual predictability (with COCA as ‘normative corpus’ using a ‘goodness-of-fit’ test). The results suggest that 59.4 % of the word lemmas are significantly over-represented in the proverb corpus (i.e. ‘keywords’). This finding underpins the assumption that a considerable number of words are strongly associated with the proverb as a category. In the experimental part of the project it was tested whether the ‘keywords’ differ in how well they contribute to a strengthened memory representation of proverbs.