Abstract

Individual differences and polysemy have rich literatures in cognitive linguistics, but little is said about the prospect of individual differences in polysemy. This article reports an investigation that sought to establish whether people vary in the senses of a polysemous word that they find meaningful, and to develop a novel methodology to study polysemy. The methodology combined established tools: sentence-sorting tasks, a rarely used statistical model of inter-participant agreement, and network visualisation. Two hundred and five English-speaking participants completed one of twelve sentence-sorting tasks on two occasions, separated by a delay of two months. Participants varied in how similarly they sorted the sentences as compared to other participants, and mean agreement across all 24 tasks did not meet an established threshold of acceptable agreement. Between the two test phases, inter-participant agreement varied to a significant but trivial degree. Networks generated for each dataset varied in the degree to which they captured all participants’ responses. This variation correlated with inter-participant agreement. The data collectively suggest that word senses may be subject to individual differences, as is the case in other linguistic phenomena. The methodology proved replicable and has a promise as a useful tool for studying polysemy.