Not the word I wanted? How online English learners' dictionaries deal with misspelled words

Abstract

This study looks at how well the leading monolingual English learners’ dictionaries in their online versions cope with misspelled words as search terms. Six such dictionaries are tested on a corpus of misspellings produced by Polish, Japanese, and Finnish learners of English. The performance of the dictionaries varies widely, but is in general poor. For a large proportion of cases, dictionaries fail to supply the intended word, and when they do, they do not place it at the top of the list of suggested alternatives. We attempt to identify some of the mechanisms behind the failures and make further suggestions that might improve the success rate of dictionary interfaces when identifying and correcting misspellings. To see whether it is possible to do better than the dictionaries tested, we compare the success rates of the dictionaries with that of an experimental context-free spellchecker developed by the second author, and find the latter to be markedly superior.