Researching language learning: theories, evidence, claims
Epistemological assumptions and ILL research
Incorporating the social: interpretivist alternatives
Commentators such as these have contributed to ILL research which sees its object as extending beyond the 'factors' which vary across teaching practices, learners and languages, and there is undoubtedly a healthy and growing literature focusing on ILL in its broader social context. Critics of the dominant approach have written of the 'imposition of an orthodox social psychological hegemony on SLA ... [which] gives preeminence to the research practice of coding, quantifying data, and replicating results. ... At best it marginalizes, and at worst ignores, the social and the contextual dimensions of language' (Firth and Wagner 1997: 288). These authors propose that socio cultural contexts should be afforded much greater recognition in ILL research, and they and others have suggested research programs capable of doing this.
We would not want to suggest that there is an absolute divide between what we are presenting, for the purposes of our particular argument, as two very broad approaches, or traditions. The debates in the literature clearly indicate some fundamental points of difference among the protagonists: see, for example, responses to Firth and Wagner by Long (1997), Kasper (1997), Poulisse (1997); or the exchanges among Beretta and Crookes (1993), Gregg (1993), Long (1993), Schumann (1993), Block (1996) and Gregg et 01. (1997). Nevertheless, several of these authors are at pains to identify areas of shared assumptions, and there have been a number of calls for complementary approaches in ILL research, or 'hybrid' research, as Ellis (1990: 202) calls it, commenting on 'the old controversy between those who favor an experimental approach and those who favor ethnographic enquiry or the use of introspective methods' which he sees as 'a sterile and unnecessary debate'.
However, for some of those who reject the dominant tradition, the epistemological differences are too profound to ignore, or to overcome by simply mixing methods. Van Lier (2001: 90), for example, commenting on the implications of conceptualizing language learning settings as complex systems, observes:
It is of the utmost importance to realize how different the job of researching language learning becomes once we decide that the social context is central. To continue looking for operationally defined, discretely measured, statistically manipulated, and causally predictive variables would be to approach one job with tools that belong to another.
As McGroarty (199B: 600) recognizes, '[a]t issue is the matter of description of the appropriate unit or units of analysis'. Discussing whether 'motivation', for example, can legitimately be conceptualized as a quality residing within each individual, as traditional research has tended to assume, she speculates whether groups (classes, companies) could perhaps '... be called learning organizations, capable of demonstrating inquisitiveness or motivation as an aggregate, not only as a collection of individuals' (p.601). Those who are dissatisfied with approaches which use experimentation and variables are usually committed to embracing a much more comprehensive 'unit of analysis'.
Thus in relation to language itself, Wardhaugh (1998), for example, is particularly outspoken about the shortcomings of what he identifies as 'the psycholinguistic approach, which ... ignores most of language and all of society' (p.589). Beaugrande repudiates any conception of 'language by itself' as a basis for investigating communication, including processes of teaching and learning language (s). In relation to learning, '[flrom the sociocultural stance', according to Lantolf and Pavlenko (1995: 110), 'acquiring a second language entails more than simple mastery of the linguistic properties of the L2. It encompasses the dialectic interaction of two ways of creating meaning in the world (interpersonally and intrapersonally)'. Constructivists such as McGroarty (1998) call for a recognition that '. .. the activity of L2 students and teachers is simultaneously linguistic, affective, and social; theory and investigation must acknowledge all three domains to render the picture complete' (p.604), and advocates of qualitative methods such as Davis (1995) see the mas:
an alternative to mainstream SLA studies in viewing acquisition not only as a mental individualistic process, but one that is also embedded in the sociocultural contexts in which it occurs. From this point of view, mental processes are not unimportant, but they are situated in a larger sociocultural context that is equally important. In other words, ethnographers and other qualitative researchers take a holistic perspective in conducting research. (p,432)
ILL research which develops different 'tools for the job' includes studies focusing particularly on learners as active social agents, those exploring the contribution of sociocultural theory to this field, those developing the 'ecological' metaphor, which connotes language learning as an organic, rather than mechanistic, process (Kramsch 2002b), and those whose critique of the dominant tradition is explicitly political. These are by no means discrete sub-fields: there is much cross-fertilization among these perspectives, which- certainly if considered together- appear to be gaining ground in ILL research.
Associated particularly with the first of these four broad trends is Breen, who, as early as 1985, observed, 'We need a metaphor for the classroom through which teacher and earners [sic] can be viewed as thinking social actors and not reduced to generators of input-output nor analyzed as dualities of either conceptual or social beings' (Breen 2001 [1985]: 127). In a recent edited collection of studies about 'learner contributions to language learning', he argues that '... language learning is, in essence, cultural action towards meaning within con text' (2001: 182). One illustration of the way in which such an orientation contrasts with that of the three example studies summarized on page 91 above is provided by Norton (2001). She undertook an extensive ethnography of immigrant language learners in Canada, and in the findings reported here she contrasts the responses of one student from Poland and another from Peru to their ESOL (English for speakers of other languages) course. Norton posits the notion of these learners as positioning themselves in relation to 'imagined communities', a concept which could not readily be accommodated within the 'variables' stance.
Another recent collection brings together studies influenced by the theories of Vygotsky and the sociocultural tradition, associated in ILL research particularly with Lantolf (Lantolf 1996, 2000a; Lantolf and Pavlenko 1995). Echoing some of the claims about learners' contributions discussed in the previous paragraph, Donato (2000) comments, '... sociocultural theory is about language classrooms where agency matters. That is, learners bring to interactions their own personal histories replete with values, assumptions, beliefs, rights, duties, and obligations' (p.46). He continues, '[w]here mentalist theories assume that language learning can be understood without taking into account the active and purposeful agent, sociocultural theory maintains that no amount of experimental or instructional manipulation ... can deflect the overpowering and transformative agency embodied in the learner' (p.47). Donato's research contrasts with that of the dominant tradition in that it does not assume that pedagogical 'tasks' are the stable, invariant 'recipes' which some of the versions of ILL research might suggest. Instead, they are perceived as 'emergent interactions' (Lantolf 2000b: 20), irreducible to their constituents.
A third challenge to successionist, variables-based models, overlapping significantly with the previous two, has been given the label 'ecological perspectives'. This is glossed by the editor of a recent collection as 'a convenient shorthand for the poststructuralist realization that learning is a nonlinear, relational human activity, co constructed between humans and their environment, contingent upon their position in space and history, and a site of struggle for the control of social power and cultural memory' (Kramsch 2002a: 5). Contributors to this development in applied linguistic research are by no means confined to ILL, and we return later to some of the other aspects of their work.
Our illustrative example here is an 'in-depth analysis- on an incremental, almost moment-by-moment basis- of what takes place in the course of a first English as a foreign language (EFL) lesson in a Dutch secondary school, on the first day of the academic year' (van Dam 2002: 237). Like much of the research, this study is presented partly in opposition to the dominant paradigm: 'Since my concern was with the emergence of a classroom culture, not with any specific teacher or student behaviors in terms of a communicative (or other) SLA orthodoxy, I adopted an approach that is multidisciplinary and ecological' (ibid.). One of the questions guiding this research- despite the (metaphorical) use of the measurement term 'degree'- will illustrate how different the approach is from that of the studies whose questions we quoted earlier: 'To what degree does speaking in another language entail a reframing of the self?' (p.252).
Finally in this very brief review of challenges to the dominant tradition in ILL research we move on to the critiques which engage with the politics of the enterprise on a more macro scale. According to Roberts (2001), SLA studies in the past few decades have not ignored social context. However, 'often the references to socio-cultural context give it only a marginal role in the process of language development' (p.108). Unsurprisingly, then, ILL is often represented as relatively 'neutral' politically: teachers have the progress of all their students at heart, and research which can identify effective practices thus has a benign potential. But as Roberts continues, 'there is relatively little concern with the social import of second language development', by which she means '[t]he transformation of many cities in Western and Northern Europe from monolingual to multilingual environments' (p.109), and the effects of this on individuals, relationships and groups.
A 'critical' viewpoint characterizes mainstream ILL research questions as inevitably far from value-free, and raises further questions such as 'who decides what counts as "better" practice and outcomes?'; 'how has the language being taught come to be a "target"?' Critics of dominant approaches in applied linguistics draw attention to the political implications of the prominence of English worldwide (e.g. Auerbach 1995; Phillipson 1992, Phillipson and Skutnabb-Kangas 1999). McGroarty (1998: 612) points out that language learning is a process which is heavily influenced by government policies and people's interpretations of those policies, and that learners' social locations and 'local socioeconomic niches' are bound to frame the value and use of particular varieties in particular contexts. Curriculum decisions, similarly, are inevitably constrained by considerations of financial resources, market priorities (which influence which text books get published and distributed) and so on (Pennycook 2001). The 'critical realist' position adopted by Corson (1997) calls for a recognition of these cultural contexts: '... the historic power differentials maintained in the (non-linguistic) discourses of wider social formations, which provide the real social context, still sit outside the concerns of applied linguistics. And when they are included, they turn up in a rather untheorized way' (p.181).
In the light of these ideas, the research question which asks 'what works?' must also be seen as a political one. Policy-makers, as well as practitioners, may look to research to help in the identification of more successful practice in the language classroom. The question 'what works?' is a legitimate one to address to researchers, and it lies behind the calls for evidence-based practice across a range of publicly-funded areas of social life. (See Mitchell 2000 for a call for applied linguistic researchers to adapt to this changing context.) However, this too is an area where epistemological and political questions are inevitably enmeshed with issues of research method. Research in the natural sciences sometimes comes to public attention for being compromised by the priorities of its funders (such as drug companies, for example). Yet the influence of interest groups other than 'scientists' may compromise social scientific research too, as academics struggle to establish and maintain their career positions by attracting research grants from prestigious sources, and publishing their findings so as to gain maximum ratings in the Research Assessment Exercise or its equivalents. It is thus quite plausible that the identification of knowledge and theoretical understanding will be at least partly at odds with the priorities of policy-makers commissioning ILL research.
