Corpus-Based Approaches to World Englishes

Abstract

The chapter provides an overview of the developments in synchronic and diachronic corpuslinguistic research into World Englishes (WE), detailing methodological concerns such as sampling frames, representativeness, corpus size, and statistical modelling on the one hand and the broadening scope of corpus-based research from ENL to ESL and EFL varieties on the other hand. It also surveys areas in which corpus-evidence has been applied in the study of WE (e.g. as a testing bed for models of WEs, for the study of language contact, typology and change or as a source of evidence for sociolinguistic and pragmatic variation), providing ample illustration from seminal research papers and recent studies in the field.

Irish English as a World English

“…The general challenge for all models of world Englishes is their demonstration and vindication of hypotheses on the basis of empirical corpus data. Whereas there have been few attempts hitherto, the review of them by Hundt (2020) plausibly concludes that the gap would appear unbridgeable. As the present article goes on to show, there is, however, no shortage of data available-ranging from the ever-growing suite of ICE-corpora to the eWAVE database and the GloWbE corpus as resources of both non-standardized as well as standardized features, for comparison, and for statistical evaluation.…”

“…Such studies are facilitated by the ICEonline software which at a stroke can compare up to 15 national components. A recent article by Hundt (2020) reviews the research which uses ICE-data to substantiate the claims of various WE models as outlined here. Section §22.4.1 of Hundt ( 2020) is entitled "Corpora as a Testing Bed for Models of world Englishes"-specifically Kachru's Three Circles Model, Schneider's Dynamic Model, and the Epicenter Hypothesis.…”

“…The general challenge for all models of world Englishes is their demonstration and vindication of hypotheses on the basis of empirical corpus data. The few attempts that have been undertaken are reviewed by Hundt (2020), whose findings will be presented and whose conclusion that the gap would appear unbridgeable will be acknowledged.…”

The be‐ versus get‐passive alternation in world Englishes

“…Our data come from the ICE, which provides a set of standard, one‐million‐word corpora compiled along the same sampling principles (Greenbaum, 1996). Seeing that the get ‐passive is still increasing and diachronic change might also affect the predictors for the choice in the passive alternation (see Fehringer, 2022), we limit our analysis to the first‐generation ICE components and exclude any of the ICE‐Age 2 components such as ICE‐Nigeria or ICE‐Sri Lanka (see Hundt, 2020). Moreover, not all currently available ICE components have been syntactically annotated, which would make retrieval of be ‐passives problematic.…”

Epicentral influence via agent‐based modelling

“…The only diachronic study so far is Peters (2009), who uses evidence from historical dictionaries – notably first attestations of regional vocabulary – to test whether Australian English served as a regional epicentre for New Zealand English 4 . For epicentral influence on ESL varieties, historical evidence is rarely available, and the few diachronic corpora that are being compiled tend to be too small to study lexico‐grammatical variation of the kind that we are interested in (see Hundt, 2020, p. 522 for references).…”