ComCogMean

Languages show certain universal tendencies in the way they choose to lexicalize meanings. These tendencies have often been argued to reflect the fact that, across cultural evolution, languages tend to be optimized for efficient communication. However, they can also be the result of various cognitive pressures that exist independently of communication, or as following from how word meaning is learnt, quite independently of communicative efficiency. In this project, we will focus on the role played by meaning in shaping the form of the lexicon (such as word length or the lexicalization of certain logical words). We will study the respective roles of non-communicative cognitive constraints and communicative efficiency in explaining word-meaning mappings, by means of three case-studies P1. Meaning competition, which may explain that languages show limited synonymy, and that there is a ‘mutual exclusivity’ effect in word learning (new words are assigned to novel meanings) P2. Meaning frequency. ZLA relates word frequencies and word length.. We argue that frequencies in the world cannot be ignored in this relation. P3. Meaning complexity. Some words never occur across word languages. Focusing on logical words such as connectives and quantifiers, we explore how the complexity of a concept interacts with communicative pressures to let it occur or not in a lexicon. We will do so by developing explicit formal models, using the tools of probabilistic pragmatics (esp. the Rational Speech Act model) and Bayesian modeling, whose predictions will be tested against data from corpora and experimental data.