Native and non-native models in elt: Advantages, disadvantages, and the implications of accent parallelism

Abstract

As a response to the diversification of Englishes into a great many native and non-native varieties, scholars such as Kirkpatrick (2007: 195) have proposed that "context and learner needs" should decide whether native or non-native models are adopted in ELT. However, an analysis of some of the arguments used in Kirkpatrick (2007) and Jenkins (2000, 2007) suggests that they have overemphasised the importance of non-native models, without giving due consideration to those Expanding Circle contexts where native models would be equally, or perhaps even more, appropriate. It will be suggested that a thorough examination of learners' needs and aspirations in different local contexts reveals that these too are actually subject to diversification. In particular, it is important for those learners who prefer to continue to refer to native-speaker models (without necessarily attempting to attain native-speaker targets) to be aware of their interlocutors' attitudes to stigmatised accent features - especially in those cases where parallels exist between native and non-native realisations (e.g. the absence of dental fricatives). Research by Van den Doel (2006) indicates that attitudes to these stigmatised accent features are difficult to predict, and that it would therefore be undesirable to recommend their general use in any pronunciation models.