Matthew James Green

Polysemy in the mental lexicon: relatedness and frequency affect representational overlap

ABSTRACT Meaning relatedness affects storage of ambiguous words in the mental lexicon: unrelated meanings (homonymy) are stored separately whereas related senses (polysemy) are stored as one large representational entry. We hypothesised that word frequency could have similar effects on storage, with low-frequency words having high representational overlap and high-frequency words having low representational overlap. Participants performed lexical decision or semantic categorisation to high- and low-frequency nouns with few and many senses. Results showed a three-way interaction between frequency, task type, and polysemy. Low-frequency words showed a polysemy advantage with lexical decision but a polysemy disadvantage with semantic categorisation, whereas high-frequency words showed the opposite pattern. These results confirmed our hypothesis that relatedness and word frequency have similar effects on storage of ambiguous words.

An eye-tracking evaluation of some parser complexity metrics

Information theoretic measures of incremental parser load were generated from a phrase structure parser and a dependency parser and then compared with incremental eye movement metrics collected for the same temporarily syntactically ambiguous sentences, focussing on the disambiguating word. The findings show that the surprisal and entropy reduction metrics computed over a phrase structure grammar make good candidates for predictors of text readability for human comprehenders. This leads to a suggestion for the use of such metrics in Natural Language Generation (NLG)