A social realist approach to research in applied linguistics
A multi-strategy approach in applied linguistic research methods
As we have already observed, the key research issue is not resolved by choosing between qualitative and quantitative methods. In various branches of the social sciences, including applied linguistics, approaches are being developed which cannot be easily categorized under either of these headings. The shortcomings of both measurement and interpretation when either is used exclusively are identified by several applied linguists.
One researcher who provides an account of how and why he has modified his research practice is Spolsky (2000), who has used traditional questionnaires as a means of access to speakers' attitudes to learning additional languages. But he has been frustrated by the limitations of mere correlations and by what is not learned by using them. Although not embracing explicitly a stratified ontology or realist philosophy, Spolsky's reflections on his changing attitude to research methodology echo some of the points we - and others we have cited - have been making. As he says:
To attempt to reduce their [his subjects'] orientation, attitude, and motivation to a single scale is clearly distortion. I do not deny the 'truth' behind simpler models of motivational and attitudinal effects on language learning, but they remain the bare skeleton, ... unlikely to account for the complexity of language practices and ideology.
Spolsky has therefore enriched his data collection methods, by not only recognizing the probabilistic qualities of sociolinguistic practices, but also augmenting the traditional correlational study with 'long interviews that give us an opportunity to explore in conversation and through stories and anecdotes the attitudes, identities, and ideologies of our subjects and to gather reports of language use in various domains and with various members of their social networks' (p.162).
Other researchers who feel constrained by the reductionism inherent in translating human behavior into numbers are also drawn to incorporate more vivid descriptions into their accounts of language practices. Thus Coupland (1997), for example, identifies the short - comings of 'autonomous' applied linguistic research, claiming that the decontextualized patterns of language use revealed by statistical methods cannot illuminate 'how age-salience imbues language in use ... [and] is communicatively valorized'. Coupland notes the problems of reconciling the advantages of the qualitative approach he came to develop in his own studies with the loss of 'scientistic, objectivist, expert research practice', and, although his conclusions (in this and subsequent work) are more thoroughly constructionist than we would find convincing, it is worth noting that he feels drawn to a mixture of methods and approaches to analysis, in order to capture the depth of the issues (here, how ageing and language practices are interrelated) with which, as an applied linguist/sociolinguist, he is concerned.
Another factor which leads to the deployment of a range of methodological approaches in applied linguistic studies is the desirability of investigating both the micro and macro dimensions of the social world.
One approach towards achieving this aim is taken by Norton (2000), who explains how the methodological framework she adopted was influenced by researchers seeking to 'investigate the complex relationship between social structure on the one hand, and human agency on the other, without resorting to deterministic or reductionist analyses' (p.21). She provides a powerful description of the experience of immigrant women in Canada, whose progress as learners of English unfolds in the context of 'inequitable relations of power'. She finds the concepts of both 'motivation' - as a fixed personality trait - and 'acculturation' into the target language group - as a process for which the learner is responsible - wholly inadequate to account for the differentiated experiences of her informants, arguing that 'a language learner's affective filter cannot be understood apart from his or her relationship to larger, and frequently inequitable social structures' (p. 120). Her ethnographic method enables her to present detailed portraits of these women's psychobiographies - their personal trajectories in the context of their family relationships and expectations, career opportunities in their home countries and the much more limited opportunities for work afforded them as immigrants to Canada. The 'social settings' of these workplaces valorize different languages differently, so we discover why one of the women, of Vietnamese background, came under pressure to learn Italian rather than English, where the former served as a marker of social solidarity among the employees in the factory where she worked. Norton also refers to material resources, glossed as 'capital goods, real estate and money' (p.7), and identifies a key role for history as integral to explanation. However, her analysis gives much less attention to these, the domain of contextual resources, than to those of situated activity and social settings, while, from her 'poststmeturalist feminist' perspective, 'discourse' once again assumes a role in social relations which we find unpersuasive, and she appears to underestimate the shaping effects of material inequalities on people's choices, opportunities and ambitions.
Another 'critical ethnographer' who disparages the 'ahistorical perspective' of much ethnographic research is Canagarajah (1999), whose four-year study at Jaffna University encompassed the community and classroom settings, and whose data included field notes, formal and informal interviews, recordings of linguistic interaction, textual analysis of written products and questionnaires. Like Norton, Canagarajah describes the different experiences and responses to learning English of various students, but he also provides detailed information about the history of Sri Lanka as a colonized territory. This includes the impact on both official language policy and actual language practice of different interest groups. He explains clearly how local people 'selectively appropriated] the Western culture and values, while benefiting from the economic and social rewards from English education' (p.65), and why the Tamil military had an interest in promoting Tamil and outlawing the use of English. Canagarajah's account also includes evidence of how the 'socially strategic' use of English by the Tamil community acts back on the variety itself, 'modify[ing] the communicative and linguistic rules of English' (p.76), an example of the kind of nonlinear interaction between structure, culture and agency that we have discussed in this text. However, once again we find his theoretical position rather confusing in places, lacking detail about who did what to whom in claims such as 'colonialism and Anglicism had many ways of imposing their reproductive agenda' (p.59), and unhelpful in suggesting that 'language is an active ideological agent that constrains our consciousness' (p.178). In the sense discussed earlier, these claims are conflationary in appearing to endow structural and cultural relations with agency, a power, we have argued, that is possessed only by human beings.
There is a further distinction marking a realist account from those considered above, namely the pursuit of 'a modest form of objectivity', as an aim of social research. Therefore, we find ourselves in disagreement with both Norton's approving reference to the position that no research can claim to be objective (Norton 2000: 21), and with Canagarajah's claim that knowledge is partisan and partial to the communities which construct it' (Canagarajah 1999: 32).
A mix of research methods and of perspectival scale is also evident in a study by Chick (2001). This began as a micro-ethnography of 14-year-old, Zulu-speaking mathematics students in the KwaZulu region of South Africa, where the data at situated activity level was analyzed using an interactional sociolinguistic framework. Chick became aware, however, of the 'limitations of explanations of school failure in terms of culturally-specific styles' (p.234), and decided to investigate the macro context too. He points out that in apartheid South Africa '... the discriminatory legislation tended to make visible what is normally hidden in democratic societies, namely the mechanisms in the wider (macro) society through which groups and individuals exercise power and deny it to others' (p.234). His study describes displays in the classroom of 'safe-talk' (chorused responses to teachers' questions, concealing a poor command of English and an absence of learning taking place), which he suggests are visible traces of 'the unpleasant fact that schooling is structured in such a way as to provide access to opportunities for learning for some students and deny it to others' (p.237). Studies such as his, Chick claims, 'show how features of the macro context, namely the institutional ideologies and bureaucratic structures, constrain what takes place at micro level' (p.238).
The stratified nature of the social world, then, as well as its complex and emergent features, make it impossible to apply a single approach to analyzing it: '... if the world is complex, it is a pointless exercise to try to deal with understanding it on a foundation of abstracted and unreal simplification' (Byrne 2002: 39). There is obviously a risk of seeming to endorse all methodological approaches and none, but the social realist can turn to theoretical principles in the process of discriminating between methods. In making 'judgements about the nature of social life and society', Layder (1993) identifies three main theoretical elements. 'These are, first, the layered nature of social reality; second, the unfolding nature of social activity over time and space, and third, the integrated nature of macro and micro features of social life' (p.108). For Layder, methods that are incapable of grasping the relationship between the macro and micro elements of social life must be ruled out as inadequate.
In sum, then, social realism is not a research method (nor yet a methodology), and studies may well draw eclectically on the extensive expertise developed by social and linguistic researchers during the previous century, including surveys, official statistics, questionnaires, interviews, observation, ethnography and even introspection. We will provide a single worked example of a typical applied linguistic research question, drawn once again from the field of second language teaching and learning, to illustrate the processes a researcher might go through in seeking to investigate it from a social realist perspective.
