Code-Switching as an Interactional Resource in Garrwa/Kriol Talk-in-Interaction

Abstract

This paper is a study of bilingual talk among Garrwa/Kriol speakers in the communities of Borroloola and Robinson River, NT, focussing on choices speakers make between traditional and non-traditional Indigenous languages in ordinary conversations. The analysis presented here supports the view of code-switching, recently summarized by Matras, as an interactional resource—a means by which speakers can structure their talk around the local contingencies of an interaction. Language choice may be symbolic of a particular social stance or ‘social arena’ in a given context, but the fact of language shift (regardless of the direction of the shift), may be equally significant in demarcating conversational activities, and marking shifts in perspective.

Keywords:

• Code-Switching
• Garrwa
• Kriol
• Conversation

Acknowledgements

I would like to thank the audience of the 2009 IPrA panel on ‘Talk-in-Interaction in Aboriginal Communities’ for their stimulating feedback on the first version of this paper. Many thanks also to Patrick McConvell, Felicity Meakins and Rod Gardner for their comments on various written versions leading to this one. I take responsibility for any remaining errors. I am very grateful to the Garrwa people who consented to be recorded and have me pick their language use apart.

Notes

¹Many Yanyuwa people speak good Garrwa, but not vice versa. Anecdotally, both Garrwa and Yanyuwa people say that Garrwa is an ‘easier’ language than Yanyuwa, although it is unclear whether this is just due to grammatical differences. Yanyuwa has a number of grammatical features that are lacking in Garrwa, including an elaborate noun class system. There may also be social reasons for this asymmetry, however.

²This ‘social model’ of code-switching was challenged by Stroud (1992) for assuming ‘Western’ models of personhood and identity which he argued were not applicable in the Melanesian community he examined. As my study is focused on conversational code-switching and not on identity functions, I will not dwell on this challenge, except to add that potential cultural differences in the practice of identity, and attribution of intentions to self and others does not interfere with the analysis of interactional uses presented here. There is growing evidence that there are fundamental principles of sequence organization, turn taking and repair that transcend cultural differences and are therefore likely to be attributable to species-specific social behaviour [e.g. papers in Enfield and Levinson (2006)].

³The data used for McConvell's (1988) study were collected in the 1970s and while there has clearly been changes in the balance of languages used in Kalkaringi, Meakins (2008) reports that these patterns of code-switching continue among older people today. Younger people have largely shifted to Gurindji Kriol.

⁴See Matras (2009: 122) for a recent demonstration of how features of social and conversational structure may work together. In his view, the differences between a ‘social arenas’ model (based on Myers-Scotton's work) and the ‘conversation analysis’ approach are methodological rather than theoretical. Proponents of a social arenas model are more focused on using code-switching to support models of social relations and participant structures, while conversation analysts are more focused on questions of how code-switching is used locally as a resource for structuring talk and other activities.

⁵When children are addressed in Garrwa, they either fail to respond or they respond with laughter.

⁶I have mostly adopted English orthographic conventions in transcribing the Kriol of Garrwa people. This is partly because its status as a variety of Kriol vs a variety of Aboriginal English has not been properly established either within the community or among analysts. The Conversation Analytic transcription allows for some representation of actual pronunciation however.

⁷In the following code-switching examples I have boldfaced the use of AE/Kriol, leaving the Garrwa talk as regular font.

⁸It is possible that these seemingly odd lexical insertions are explainable in terms of brief shifts in the indexation of social arenas, as McConvell found in the Daguragu data. However there is insufficient contextual information concerning the particular relationships of these participants to attempt such an analysis here.

⁹There is no Garrwa equivalent of ‘pub’, but there are many possible Garrwa formulations for identifying the location of the grandson.

¹⁰I have kept the overlapping talk of the girl in the transcription but this is talk from another conversation that was picked up on the recording (there are a number of younger people present at this time who are talking with each other and not with the older women). Daphne's utterance here is ‘in the clear’ as far as her conversation with Kate and Hilda is concerned.

¹¹However, the request is unsuccessful—Kate never gets any chips from Mary.

¹²While the linguist/author is not mentioned overtly here, she is discussed later as the person behind Kate's reason for talking so much. In fact, Kate later complains that she is working her too hard and too long and is tiring her out.

¹³‘Narrative perspective’ is a term adopted in Mushin (2005) and other work to analyse the deictic orientation of the storytelling as representing a narrator viewpoint (e.g. with third person reference, past tense forms, etc.) or a character viewpoint (e.g. through the direct representation of a character's speech, thoughts or perceptions), or some combination of both. Narrative perspective-taking also involves the ways in which storytellers switch from a deictic centre within the narrative storyworld to the here and now deictic centre of the storytelling event. This aspect is closest to McConvell's (1994) analysis of narrative code-switching as a device to mediate between the storytelling proper and metacomments by the storyteller about the story.

¹⁴It has been shown in many studies that direct speech reporting in storytelling is not about the accurate representation of prior utterances (e.g. Clark & Gerrig 1990; Mayes 1990; Mushin 1994; Clift & Holt 2007). In storytelling they serve a range of functions associated with stance taking, or ‘positioning’ in Goffman's terms. Indeed stories are rife with representations of speech events which never occurred at all. It is highly unlikely that Ellen is consciously attempting to accurately represent the language that was spoken at the time, even if it were actually the case.