Frequency and efficiency in Spanish proverbs

Abstract

Zipf's law states that there is an inverse relationship between a word's length and its frequency; the more frequent words tend to be the shortest. Following this premise, I investigate if this universal property in language can extend to domains beyond the word. As such, the present study analyzes the use of proverbs, a specific type of fixed expressions, in a Spanish corpus of news language. The motivation for this study is to determine if more frequent proverbs are more likely to be shortened, relative to their lower-frequency counterparts. The results of this study indicate that there is a positive correlation between a proverb's general frequency in a corpus and its reduction rate. This paper argues that usage-based models of language representations are better equipped to account for the mechanisms used in the production of proverbs, compared to the traditional view of fixed expressions as enlarged single words.