Introduction
3. Data collection
3.3. Methods of data collection
3.3.1 The triangulation approach
3.3.1.2. Interviews
Interviewing is multifaceted, and scholars in sociolinguistics have developed several different approaches to interviewing (see for example Labov 1972a; b, 1984) in order to gain specific qualitative research data. "Interviewing is both a direct source of information on belief and knowledge systems and a form of vicarious observation to increase case examples of various types of overt behaviour" (Pearsall 1970: 346). One has to question, however, whether behaviour in interviews is indeed 'overt' behaviour, especially if the interviewer is an outsider to the community and has not established a comfortable relationship with their participants. On the basis of this question, I focused on qualitative interviews in this particular study, as most researchers in the social sciences have argued for the benefits of a 'natural' atmosphere that is as similar as possible to the normal and familiar environment of the interviewee.
The initial interviews with learners were held in the three high schools, but I subsequently interviewed anyone who would allow me to do so, i.e. teachers, parents, security guards. I also established a network with other Umlazi residents 155
and started interviewing, for instance, neighbours of the people I was staying with.
The main characteristics that all the interviewees had in common was that they were all isiZulu mother-tongue speakers, residents of Umlazi and between the ages of 15 to 40. Most interviews were recorded and subsequently transcribed. In some cases, however, I only made notes during the interview, as some informants felt uncomfortable about recording their voice on tape.
As the interview framework was qualitative, the role I adopted during the interview was of crucial importance. The idea was to let the interview develop into a conversation in which I was more in the position of an active listener, rather than an interrogator. Hence, it was supposed to be the informants who controlled the interview rather than me, the interviewer, and consequently the recorded data yielded information on a variety of topics. Nevertheless, I did at times prompt the interviewees to focus on the issue of language.
In the theoretical framework chapter of this thesis I discussed the slippery surface we find ourselves on when discussing concepts such as 'ethnic identity'. While the person interviewed may, for instance, be well aware of a particular ethnic identity, it is of little use to directly and explicitly ask him/her about it. In fact, various field researchers choose not to inform their informants at all about where the focus of the investigation lies (McCafferty 2001).112 My aim was to let the interviewee speak freely about the uses and the functions of language(s) in their lives, in particular the mother-tongue isiZulu. As already mentioned, I attempted to take up the role of an interested listener, rather than an interrogator, which, nevertheless included questions and comments from my side.
Initially, I always gave the interviewees the option to conduct the interview in isiZulu, but the majority of the participants, however, preferred to speak English with me. This choice must not be mistaken for a preference to speak English, but
In the interviews I found it necessary in most encounters, and 'ethical' to give the participants of this study at least some general information on what I was interested in. The issue of language offers a less 'loaded' and less complicated platform for discussion when compared to the issue of ethnicity.
156
rather a choice influenced by the awkward presence of an interpreter. Interestingly, as my own isiZulu proficiency improved, I found myself increasingly making use of English-isiZulu code-switching during the interview and conversations.
On several occasions I held group interviews, either with learners in a classroom, or after school in small groups. Given the fact that the relationship of language, identity and ethnicity is deeply intertwined with general attitudes towards sociolinguistic dynamics, I chose controversial topics in order to reveal notions of identity construction among learners. Accordingly, learners and teachers discussed issues such as the eleven-official-language policy, the status of isiZulu in KZN, or the question of whether English in South Africa has been acting as a liberator or as an oppressor. As the focus of this study generally revolves around identity negotiations among isiZulu LI-speakers, which is very much embedded in the social system, it seemed adequate to discuss attitudes and perceptions in a larger group. Hence, I refer to the qualitative method of the 'group discussion' as described by Mayring (1996: 58) whereby sensitive topics and subjective meanings are frequently revealed more easily within a larger group of people. Accordingly, discussions in a group can provide a fruitful platform on which psychological barriers may be overcome with less difficulty. When individuals find themselves on common ground with others, collective attitudes are naturally exposed more readily.
I focused on general issues that would implicitly give information about the attitude and perceptions of individuals and groups towards a particular language, i.e. isiZulu and/or English. Another productive topic of discussion was the question of whether 'African literature' should be written in an indigenous African language, or whether books published by an African in English on an African setting would qualify for 'African literature'. This technique offered the advantage of stirring the learners' different perceptions in the class and gave me the chance to listen and observe. Unsurprisingly, attitudes within the class were naturally split and quite frequently at least two lobbies emerged with different viewpoints. The danger always existed that a few individuals, generally males, would dominate the
157
discussion, and I paid close attention to the possibility that one person may influence others by "inhibiting them from saying what they really think or leading the group into apparent agreement" (Wray, Trott and Bloomer 1998: 182).
Individual interviews were based on the narrative approach, whereby the participants were guided to speak about any language(s)-related matters in their lives (Mayring 1996: 55). Accordingly, there were few structural limitations and a high level of flexibility on my behalf in order to elicit the thoughts, emotions, perceptions and opinions of the individual participant. Interviews with people who were hesitant to talk about their individual perceptions and opinions tended to evolve more into a problem-centred interview (Mayring 1996). In those cases, I generally initiated the interview by asking the individual about his/her feelings about one of the previously mentioned discussion topics. Although interviews, if handled with care and critically explored, certainly provide a useful source of data elicitation, I regard it as insufficient to rely solely on interview findings for the purpose of this study. Researchers in the social sciences have increasingly become reluctant to trust what people say in interviews (see, for instance Vaux and Cooper 1999: 18). The actual behaviour of participants has frequently proven to be at odds with the information given during an interview. It is for this reason that it was imperative for this investigation to employ an ethnographically orientated method to create a holistic and more authentic picture of the sociolinguistic dynamics at work in the township community.