A social realist approach to research in applied linguistics
A relational view of applied linguistic issues
We noted that '[social phenomena are what they are by virtue of the internal relations they have to other phenomena' (Danermark et al 1997: 97). Researchers whose practice is consistent with the approach we have been developing should be able to distinguish between those empirical outcomes which are contingent and those which are a necessary product of generative relations. There are echoes here, of course, of the notion of' confounding variables', but, as we hope we have made clear, a relational approach is not hoping to find which variables make a difference, since the world of social interaction is an emergent, complex, densely symbolic world, where relations themselves, not isolable variables, are constitutive of 'what works'. Applied researchers may feel much better equipped to address the policy-makers' concern with 'what works', if the question can be reformulated as 'what works for whom in what circumstances?'.
Two social researchers who have operationalized this approach provide us with an example which, while it does have some relevance to education, is outside the direct concerns of applied linguistics. We shall draw out the ways in which we think principles embodied in this study can be built on within applied linguistics.
For Pawson and Tilley (1997: 46), studies based on successionist, variationist models of social research cannot tell us 'what works for whom in what circumstances', precisely because they work:
with a logic which prioritizes a certain set of observational categories and sequences. This framework tends to overlook the real engine for change in social programs which is the process of differently resourced subjects making constrained choices amongst the range of opportunities provided.
For Pawson and Tilley, research must identify how people are 'differently resourced'; what choices are open to them - and how these are 'constrained'; and what opportunities are afforded them by virtue of their social location. A recognition of the importance of all these elements (that is, resources, constraints and affordances) is the methodological starting point for designing a research program, and one that Pawson has incorporated into a long-term study of the impact of the Simon Fraser University Prison Education Program on the rehabilitation of inmates serving in British Columbian penitentiaries.
The first point that Pawson notes about the program is that it does not 'work' in reducing recidivism in some undifferentiated way. Rather the researcher starts by forming hypotheses about potential generative relations and mechanisms which might underlie educational rehabilitation. Briefly, the point 'is to acknowledge that outcomes of complex programs [such as the Simon Fraser University Prison Education Program] bear the mark of countless contexts and mechanisms' (p.107). Thus in forming theories about what might work for whom in what contexts, the researcher needed to consult both the academic literature and also those who had an 'insider's knowledge' of the program itself, namely those who had taught on it and those who had been taught in it. This mixture of theorizing and qualitative investigation prepares the way for the main body of the investigation, which is an empirical study of outcomes.
In this particular study this meant building up a quantitative picture of which types of inmates and which types of course experience were associated with lower recidivism. Salient differences between individuals were identified from large numbers of cases, often by means of interviews, and grouped to provide an empirical description of hundreds of sub-groups. Data analysis then involved constructing a 'mosaic of outcome patterns', somewhat as we advocated in our discussion of the sociolinguistic research. In this research process, there can be no unqualified finding as to 'what works' to make prison education programs effective in preventing reoffending. Instead, the process of tracing different outcomes back through similar pathways taken by those who were studied leads to a richer and deeper understanding of the particular mechanisms that have operated in specified contexts. This study is innovative, then, in its view of causality and in the methodological means chosen to operationalize it, namely the retroductive tracing of similar outcomes in order to identify which contexts trigger which 'causal mechanisms'.
Another way of approaching the multi-layered phenomena which interest the social researcher is to focus, as Byrne does, on states rather than outcomes. For Byrne, realist accounts should be concerned less with effects and outcomes, which still smack of linearity and fixity, and more with the state of the system as a whole. Such 'states' are inherently unstable, complex and emergent. The research problem we have drawn from Byrne's work involves health and social class, and it exemplifies an attempt to take account of the stratified character of the social world. To account for the differential incidence of TB amongst urban populations, Byrne (in press) illustrates a methodology which he suggests is more capable than many traditional approaches of grasping the emergent nature of a stratified social reality. In accounting for who gets TB and in what contexts, Byrne argues, we must first acknowledge a biological aetiology (people must be exposed to the bacillus) and a genetic component (some people will have a natural resistance to the bacillus). However, whether particular individuals contract the disease or not will be socially contingent, since it will depend on the interaction between these features and other levels of the social world. Byrne identifies four such levels - the individual, the household, the community and the nation-state - existing in a nested hierarchy.
Interaction between these levels is both outwards and back inwards. Thus, for many working people in the 1930s, seeing family members die of TB impelled them towards a radical political critique of inequality. (This critique is an example of what we have called a Cultural Emergent Property.) Such radicalization was a factor in the Labor victory in the General Election of 1945 and in the development of a national health service. This, together with local programs of housing construction and urban improvement, effectively succeeded in wiping out TB in the UK during the 1950s and 1960s and thereby changed markedly the answer to the question of who gets TB and in what contexts. Byrne's example provides a clear illustration of how the social and the individual intersect, and of how the nature of that intersection is best understood in terms of unstable, nested systems of social relations, where 'feedback' can be in both directions.
