The effect of defining model on the accuracy of syntactic class identification in when-definitions in digital dictionaries

Abstract

The aim of the present contribution is to identify the best variant of the single-clause when definition in online monolingual English dictionary entries. The study uses manipulated online dictionary entries and a balanced experimental design on a large sample of advanced learners of English to assess the success of syntactic class identification in three sub-types of the single-clause when-definition. Syntactic class identification was measured by examining the syntactic class of the Polish translations of the target headwords supplied by participants in the experimental task. Significant differences are found between all three defining models, with the most effective being a model using the personal pronoun (you), the least effective being definitions using the indefinite pronoun (someone), and a defining model using a noun phrase (a person) resulting in intermediate scores. The article concludes with practical recommendations for digital lexicography for learners of English that follow from our findings.