Standards of acceptability in English as an academic lingua franca: Evidence from a corpus of peer-reviewed working papers by international scholars

This paper probes the issue of standards of acceptability in English as an academic lingua franca (academic ELF) by investigating selected patterns of variation in a corpus of working papers produced by international scholars in a unique multilingual context: the EU-funded European University Institute (Florence, Italy). Analysis of video-recorded oral presentations and participant observation over a one-year period revealed that oral language use in this setting frequently diverges in certain areas of the lexico-grammar and syntax of English from native-speaker norms. Such divergences were rarely commented on as inaccurate, however, by either ‘native’ (NS) or non-native (NNS) members of the Institute. Based on these observations, it was hypothesized that similar 'pluralistic' notions of acceptability (cf. Dewey 2007) may also be operating in the written production of established international scholars. To investigate this hypothesis, rater evaluation and corpus linguistics techniques were used to investigate lexico-grammatical variability in a corpus of peer-reviewed (but not officially language-edited) working papers published in the same setting (henceforth, WP Corpus: 75 papers; 847,234 words; average length per paper 11,296 words). Five areas frequently mentioned in ELF literature and in which deviation had been identified in at least 25% of a sample of 40 oral presentations were investigated: (1) preposition choice in 'fixed' expressions; (2) word order in reported (indirect) questions; (3) positioning of adverbs; (4) use of definite and indefinite articles; and (5) functional range of the present perfect tense. To investigate rater evaluations (and hence possible ‘gate-keeping’ to which said production would be subject if presented to international English-language journals), three raters representing different geographical varieties of L1 English (US, GB, New Zealand) were asked to identify and classify all segments the first 4 pages of each paper which they considered deviant. Variability was also investigated by using concordance tools (Wordsmith Tools 4, AntConc) in order to compare the WP Corpus with a corpus of research articles in similar disciplines published in peer-reviewed, copy-edited international journals (RA corpus: 1,221 articles, for a total of 12,861,501 words; average length per paper: 10,534 words). Divergence from NS norms in the WP corpus (at least one divergent form identified by at least two raters in the paper in question) was found to range from a high of 57% for the use of articles to a low of 5% in for embedded questions. Sections 4.1 to 4.5 of the paper present comparative analyses of the WP and RA corpora in order to illustrate and discuss the principal patterns of variability. Among the tendencies identified were shifts in the functional distribution of certain pairs of prepositions between/among; on/over; in/into; under/below, with the former term used more extensively in the WP corpus in certain colligations. Divergences in the WP corpus from ‘NS’ norms were also present in areas in which both descriptive and pedagogical grammars traditionally contemplate some margin for authorial ‘manoeuvre’, such as patterns involving sentence-initial also. Differences in the functional range of the present perfect were also observed (over-extension). The study contributes to the ELF literature by documenting, for selected areas of the lexico-grammar, ways in which English as an academic lingua franca may be moving towards greater endo-normativity; it also illustrates how data triangulization (combining rater evaluations and corpus linguistics techniques) can provide insights into how variability is assessed by NS language professionals and hence, potentially, into processes of ‘linguistic gatekeeping’ in academic publishing.