Words and Roots – Polysemy and Allosemy

Robyn Anne Carston (University College London)

Polysemy is cross-categorial, so the linguistic unit ultimately underpinning polysemy is not the word but the categoryless root of related words, e.g. the root √stone common to the three homophonous words (noun, verb and adjective) ‘stone’, whose meanings are semantically-pragmatically interrelated.  In some current root-based approaches to morpho-syntax (Marantz 2013, Harley 2014, Embick 2015), it is proposed that roots themselves are polysemous, albeit with meanings that are realized only in specific grammatical contexts (a phenomenon known as allosemy, parallel to allomorphy on the phonological side). In this talk, taking the polysemy of words as established, I assess the notion of allosemy and conclude that there is very little evidence for its existence, and that roots themselves are (most likely) meaningless, with atomic meanings (or ‘Content’) mapped to categorized linguistic structures of varying complexity (Borer 2013, 2014), that meaning being typically pragmatic in origin (Carston 2022, 2024).