Sorting out the Most Confusing English Phrasal Verbs

In this paper, we investigate a full-fledged supervised machine learning framework for identifying English phrasal verbs in a given context. We concentrate on those that we define as the most confusing phrasal verbs, in the sense that they are the most commonly used ones whose occurrence may correspond either to a true phrasal verb or an alignment of a simple verb with a preposition. We construct a benchmark dataset1 with 1,348 sentences from BNC, annotated via an Internet crowdsourcing platform. This dataset is further split into two groups, more idiomatic group which consists of those that tend to be used as a true phrasal verb and more compositional group which tends to be used either way. We build a discriminative classifier with easily available lexical and syntactic features and test it over the datasets. The classifier overall achieves 79.4% accuracy, 41.1% error deduction compared to the corpus majority baseline 65%. However, it is even more interesting to discover that the classifier learns more from the more compositional examples than those idiomatic ones.