The ambiguous nature of idioms has been a persistent challenge for English language learners and researches alike. Problematic issues include identifying which idioms are most pertinent for study, and the question of how frequently idiom forms found in dictionaries and other canonized resources actually function as idioms in real language use. This study differentiates between idiom forms used idiomatically (idiom-principle) versus literally (open-choice principle), and provides quantitative data to assess this difference. The data was obtained through a corpus analysis of 1,000 randomly-selected idioms in 10,000 randomly-selected contexts (10 contexts per idiom), and revealed that the majority of idiom forms were indeed functioning idiomatically in the contexts analyzed, but there were also notable exceptions. The findings are used to support the general notion in the literature that idioms represent a single lexical choice for language users, and the researcher proposes several extensions of the findings for the teaching and researching of idioms.

The research reported in this thesis explores the degree of authenticity of the formulaic language used by NNSs and the extent to which a learner's L1 interferes in the production of different types of multi-word units, namely non-idiomatic recurrent three and four-word combinations (lexical bundles), phrasal and prepositional verbs and collocation. Drawing on Granger's Contrastive Interlanguage analysis (CIA 1996), the investigation is conducted on two different learner sample corpora and subsequently contrasted with a native sample corpus. The study aims to prove that multi-word units pose a challenge for learners for several reasons. In general terms, learners are assumed to operate predominantly on what Sinclair calls the open-choice principle, that is to say their production will be less idiomatic than that of native speakers'. This assumption is independently tested on different types of phraseological combinations. As regards non-idiomatic recurrent word combinations, learners are expected to be more repetitive in their three- and four-word combinations and use less creativity in their writing. Concerning the phrasal verbs, it is highly likely to observe a small number of phrasal verbs in the non-native writing whereas prepositional verbs are considered problematic for learners due to the...