Abstract

This contribution presents a study of collocations in a subset of the EPTIC corpus, which encompasses comparable discourse samples in English produced under a variety of communicative constraints (oral and written, original and mediated, native and non-native). Results are interpreted against the backdrop of Lanstyák and Heltai's (2012) "constrained language" hypothesis, according to which translations, on a par with other varieties of English produced in language contact settings, share linguistic features that set them apart from monolingual native production. Evidence of such constraints is observed in this study with respect to collocational patterns defined by Mutual Information values. Findings are discussed in the light of the cognitive constraints characterising the different language production tasks under investigation.