International Journal of Corpus Linguistics - Volume 22, Issue 2, 2017

Multi-Dimensional Analysis (MDA) has been widely used to explore register variation. This paper reports on a project using MDA to explore the features of an interdisciplinary academic domain. Six dimensions of variation are identified in a corpus of 11,000 journal articles in environmental studies. We then focus on articles in one interdisciplinary journal, Global Environmental Change (GEC). It is expected that these articles will diverge sufficiently to produce differences that are analogous to register differences. Instead of identifying these “registers” on external criteria, we use the dimensional profiles of individual texts to identify ‘constellations’ of texts sharing combinations of features. Six such constellations are derived, consisting of texts with commonalities in their approaches to research: the development of predictive models; quantitative research; discussions of theory and policy; and human-environment studies focusing on individual voices. The identification of these constellations could not have been achieved through an a priori categorisation of texts.

This corpus-based study explored the effects of two factors – genre (i.e. speech event type) and disciplinary variation – on spoken academic ELF, from the perspective of lexical bundles (i.e. recurrent word combinations). The material was drawn from a corpus of transcribed spoken academic lingua franca English (ELFA). The investigation involved a quantitative analysis of the use of four-word bundles, in terms of frequency, form, and function, across a range of genres (academic lectures and seminars) and disciplines (Medicine, Social Sciences, Natural Sciences). A qualitative analysis was also carried out to give an in-depth account of functional variations associated with one particular lexical bundle I don’t know if. The results demonstrated that genre and discipline are two important factors that cannot be ignored in understanding academic ELF communication and idiomaticity, and lexical bundles provide useful glimpses on genre and disciplinary variation that are worth following up.

Forensic authorship attribution is concerned with identifying the writers of anonymous criminal documents. Over the last twenty years, computer scientists have developed a wide range of statistical procedures using a number of different linguistic features to measure similarity between texts. However, much of this work is not of practical use to forensic linguists who need to explain in reports or in court why a particular method of identifying potential authors works. This paper sets out to address this problem using a corpus linguistic approach and the 176-author 2.5 million-word Enron Email Corpus. Drawing on literature positing the idiolectal nature of collocations, phrases and word sequences, this paper tests the accuracy of word n-grams in identifying the authors of anonymised email samples. Moving beyond the statistical analysis, the usage-based concept of entrenchment is offered as a means by which to account for the recurring and distinctive production of idiolectal word n-grams.

While discourse markers (DMs) and (dis)fluency have been extensively studied in the past as separate phenomena, corpus-based research combining large-scale yet fine-grained annotations of both categories has, however, never been carried out before. Integrating these two levels of analysis, while methodologically challenging, is not only innovative but also highly relevant to the investigation of spoken discourse in general and form-meaning patterns in particular. The aim of this paper is to provide corpus-based evidence of the register-sensitivity of DMs and other disfluencies (e.g. pauses, repetitions) and of their tendency to combine in recurrent clusters. These claims are supported by quantitative findings on the variation and combination of DMs with other (dis)fluency devices in DisFrEn, a richly annotated and comparable English-French corpus representative of eight different interaction settings. The analysis uncovers the prominent place of DMs within (dis)fluency and meaningful association patterns between forms and functions, in a usage-based approach to meaning-in-context.

The recent emergence of large parallel corpora has represented a leap ahead for cross-linguistic and translation studies. However, the specificities of these corpora and their influence on the nature of observed linguistic phenomena remain underexplored, especially in the field of contrastive linguistics. In this study, we compare the translation equivalences of four concessive adverbial connectives in English and in French across three corpora varying along three dimensions: register, directionality of the translation and translator expertise. Our results indicate that these dimensions affect the cross-linguistic equivalences observed between connectives. We conclude that, in future work, translation-based claims about cross-linguistic equivalences should be balanced according to the type of data analysed. We also identify a pressing need for more rigorously-documented parallel corpora for the English-French language pair.