Meeting Description: Although words always express units of meaning, languages frequently use different strategies of meaning lexification, depending on a number of psychological and historical factors. Studying polysemy (the phenomenon of one word sharing several meanings) and semantic shifts (change of meaning over time) allows to better understand how individual concepts relate to each other. Thus, in 2015 we showed that for a certain set of terms expressing coarse concepts related to natural substances, topography and astronomy, the languages of the world are surprisingly regular in mapping out the relationships between them. Other work has extended the domain to a much larger set of concepts, though they studied only a few major languages. Additionally, the Evolution of Human Languages program is currently building up a general typology of polysemies and semantic shifts within the so-called basic lexicon to augment the study of language relationships in prehistory.
We would like to use these previous studies and ongoing projects to help us develop a unified theory of the deep structure of lexical semantics in collaboration with computational linguists. We aim to overcome a major shortcoming of the existing work—that, rather than intrinsic structures inherent in the universal lexical semantic network, it uses only words from English, or another European language, as a metalanguage. Collaboration on this subject between linguists and computational scholars will also result in progress on at least three different fronts:; (1) building a proper methodological foundation for lexics-based phylolinguistics (where historical semantics remains one of the more fuzzy and problematic areas); (2) understanding how to deal with the fact that cross-linguistically, words may be correlated not only with words but with phrases as well; and (3) assisting in the construction of universal translators.