This article presents a series of studies focusing on L2 production and processing of adjective-noun collocations (e.g., social services).In Study 1, 810 adjective-noun collocations were extracted from 31 essays written by Russian learners of English. About half of these collocations appeared frequently in the British National Corpus (BNC); one-quarter failed to appear in the BNC at all, while another quarter had a very low BNC frequency. Based on frequency data and mutual information (MI) scores, it was discovered that around 45% of all learner collocations were, in fact, appropriate collocations, that is, frequent and strongly associated English word combinations. When the study data were compared to data from native speakers, very little difference was found between native speakers (NS) and non-native speakers (NNS) in the use of appropriate collocations. Unfortunately, the high percentage of appropriate collocations does not mean that NNSs necessarily develop fully native-like knowledge of collocation.In Study 2, NNSs demonstrated poorer intuition than NS respondents regarding the frequency of collocations. Likewise, Study 3 showed that NNSs were slower than NSs in processing collocations. Overall, the studies reported here suggest that L2 learners are capable of producing a large number of appropriate collocations but that the underlying intuitions and the fluency with collocations of even advanced learners do not seem to match those of native speakers.

One of the choices available in English is between one-word verbs (train at the gym) and their multi-word counterparts (work out at the gym). Multi-word verbs tend to be colloquial in tone and are a particular feature of informal spoken discourse. Previous research suggests that English learners often have problems with multi-word verbs, and may even avoid their use. This study explores this issue further by comparing the likelihood of using multi-word vs. one-word verbs by both native speakers and advanced nonnatives. A questionnaire of 26 multi-word/one-word verb pairs showed that nonnatives were less likely to use multi-word verbs than native speakers in informal spoken contexts. Moreover, the amount of exposure to native-speaking environments did not have an effect on the likelihood of using the multi-word verbs. However, a corpus analysis of the same verb pairs showed that the one-word verbs are often more frequent in both written and spoken discourse.

Copyright © Cambridge University Press 2018. A reading experiment combining online and offline data evaluates the effect on second language learners' reading behaviors and lexical uptake of three gloss types designed to clarify word meaning. These are (a) textual definition, (b) textual definition accompanied by picture, and (c) picture only. We recorded eye movements while intermediate learners of English read a story presented on-screen and containing six glossed pseudowords repeated three times each. Cumulative fixation counts and time spent on the pseudowords predicted posttest performance for form recall and meaning recognition, confirming findings of previous eye-tracking studies of vocabulary acquisition from reading. However, the total visual attention given to pseudowords and glosses was smallest in the condition with picture-only glosses, and yet this condition promoted best retention of word meaning. This suggests that gloss types differentially influence learners' processing of novel words in ways that may elude the quantitative measures of attention captured by eye-tracking.

© 2018 Elsevier Ltd Conformity to the epistemological orientations of academic disciplines is often reflected in the ways in which knowledge is constructed and communicated through certain linguistic features in academic genres. This paper explores instances of such conformity in hard and soft science disciplines, as it is reflected in the use of recurrent word combinations in different rhetorical moves of research article abstracts. A corpus-driven approach was adopted to identify the most frequently occurring word combinations of different length in a corpus of 5910 abstracts from six disciplines. Following this, the identified sequences were classified according to their communicative functions in different moves of the abstract. A mixed-methods approach was then adopted through which the patterns of variation reflected in the use of the identified sequences were examined both quantitatively and qualitatively. It was found that members of different academic domains have different priorities for representing their research in academic abstracts.

The field of vocabulary research is witnessing a growing interest in the use of eye-tracking to investigate topics that have traditionally been examined using offline measures, providing new insights into the processing and learning of vocabulary. During an eye-tracking experiment, participants’ eye movements are recorded while they attend to written or auditory input, resulting in a rich record of online processing behaviour. Because of its many benefits, eye-tracking is becoming a major research technique in vocabulary research. However, before this emerging trend of eye-tracking based vocabulary research continues to proliferate, it is important to step back and reflect on what current studies have shown about the processing and learning of vocabulary, and the ways in which we can use the technique in future research. To this aim, the present paper provides a comprehensive overview of current eye-tracking research findings, both in terms of the processing and learning of single words and formulaic sequences. Current research gaps and potential avenues for future research are also discussed.