Trading off informativeness and length in lexicon and grammar

In this talk, I focus on the role of coding efficiency in the lexical domain, following up on earlier work on coding efficiency in grammar (e.g. Haspelmath 2021, on form-frequency correspondences and predictability). I will start out from Zipf’s Law of Abbreviation, which reflects communicative efficiency: Frequently expressed meanings are predictable and thus not very informative, and therefore languages can use short forms for them. But in addition to causing shortness of coding, communicative efficiency also causes frequently expressed meanings to be more differentiated – I call this “Mańczak’s Law of Differentiation” (Mańczak 1966; 1970). There is a substantial body of research that claims that lexical domains (such as colour terms, kinship terms, and body part terms) are more differentiated because of an efficient trade-off of informativeness and complexity (see Kemp et al. 2018 for a survey), but I will argue that informativeness trades off against length of coding rather than “complexity”.