Where should English Dictionaries place multi-word expressions? A study of user consultation behaviour

Abstract

Multi-word expressions (MWEs) are a known challenge to lexicographers and language learners alike. Dictionary articles are traditionally organized around single words, so lexicographers are forced to choose the headword under which to include a given MWE. When users encounter a problematic MWE in text, they also face the choice of the component word under which to look it up. The present study attempts to find out what factors can predict such lexical choices so that lexicographers can position the information in the very entry that dictionary users will choose to consult. Based on previous findings, four factors are selected as potentially predictive: lexical frequency, part of speech, word position, and word length. A list of thirty-seven English sentences with MWE’s was presented to 155 Polish students at high school and university levels, asking them to indicate which word they would choose to look up in a dictionary. Our analysis suggests that these choices are best explained in terms of lexical frequency, with a preference for less frequent items. Empirical results are then compared with actual lexicographic practice for these very MWE’s in five major English learners’ dictionaries, indicating there is much room for improvement, and that such improvement is easily achieved using our findings. Finally, it is argued that the issue of MWE lemmatization is relevant for print and digital dictionaries alike.