Studying policy and politics: bringing linguistic ethnography to interpretive policy analysis

Interpretive policy analysis and linguistic ethnography have both added important dimensions to our work. Rather than attributing different aspects of our work to different approaches (which tends to dichotomise rather than synthesise), we conclude by describing how our own dialogue across interpretive policy analysis and linguistic ethnography has enabled us to better understand and analyse the dispersed, microlevel practices that make-up health policy and planning, and the role of think tanks in shaping these.

This dialogue between linguistic ethnography and interpretive policy analysis is best illustrated with a concrete example. The work of Goffman is frequently drawn on by those adopting linguistic ethnography (see Rampton, Maybin and Roberts, this volume). It was an area of work that we engaged with on the Key Concepts course, encouraging us to connect with dramaturgical theory (specifically performance and impression management, Goffman, 1959). Whilst this work is highly relevant to our own research, it tends to focus on specific instances of interaction in single, defined spaces (e.g. classroom, police interview room) rather than multiple and dispersed micro-level policy practices. Engagement with interpretive policy analysis enabled us to foreground policy and planning as the object of our study, theorise them and then, building on Goffman's work (1959, see also Degeling, 1996), connect with relevant sensitising concepts (front-stage and back-stage) to guide our analysis. In this way, we connected the concepts and methods of linguistic ethnography with the policy-oriented approach of interpretive policy analysis.

The social theory allied to interpretative policy analysis has been important in guiding our conceptualisation of policy as a series of dispersed, micro-level practices (Yanow, 1996; Bacchi, 2000; Stone,

2001; Hajer and Wagenaar, 2003; Hajer, 2005; Wagenaar, 2011). This has guided data collection to ensure (as far as possible) that a range of actors, settings, interactions, artefacts and arguments allied to policy and planning were included within our study. We have at times been frustrated by a limited attention to data and close analysis of language within interpretive policy analysis (see, for instance, an article examining the development of alcohol policy, which employs 'rhetorical frame analysis' but then holds back from detailed analysis of language or rhetoric [Hawkins and Holden, 2013]). For us, this is where linguistic ethnography comes to the fore: the dialogue between language and context is one of the qualities of linguistic ethnography that we valued most and, guided by social and political theory, has led us to explore both 'front-stage' and 'back-stage' contexts of think tanks' work, along with the different linguistic resources employed in each; and to undertake close analysis of think tank language and arguments.

There is a clear commitment in linguistic ethnography to making analytic claims accountable to evidence (Rampton, Maybin and Roberts, this volume). We described above how policy is simultaneously distributed via multiple and dispersed policy practices, and at the same time held in common via overarching policy narratives that make sense of events and guide action. This presents particular challenges for us in 'making analytic claims accountable' (ibid.). Whereas many linguistic ethnography studies are able to demonstrate the key analytic point in a small strip of talk or a few exchanges, analysis of health policy and planning narratives often stretches over multiple and extended documents and/or exchanges and cannot easily be captured in a specific segment of data. Whilst this does not detract from close analysis, it does make it difficult to present 'evidence' of analytic claims in the conventional way. This is an irresolvable characteristic of our work. However, the dialogue between linguistic ethnography and interpretive policy analysis has been helpful in simultaneously encouraging us to undertake close analysis of policy practices (in ways that are unusual for linguistic ethnography) and pushing us to evidence our analysis (in ways not traditionally undertaken within interpretive policy analysis).

It has been argued that linguistic ethnography brings 'a formal, abstract discipline and tried and tested, finely-tuned methods for analysing text together with more open, reflexive social orientation of ethnographic methods, which offer analytic purchase on the related social practices and structures' (Tusting and Maybin, 2007: 576). This dual focus encourages rigorous linguistic work that also addresses social practice. However, it can lead to methodological tensions between the more 'closed' focus on linguistic text and a more 'open' sensitivity to context (ibid.). These tensions are characterised by on-going discussions about if and how language ties ethnography down and ethnography opens linguistics up (Rampton, 2007). In our work on think tanks, we regard this process as a balancing act, negotiating ethnography and language throughout the study so as to enable a conversation about text and context and expand understanding of the ways in which think tanks work to shape health policy.

Methodological tensions also raise interesting questions about the analytic gaze and what counts as data. We have drawn attention to the specific focus on policy as the object of our study and described the ways in which different data sources have, together, helped us to understand the work that think tanks do in shaping NHS reforms. It was interpretive policy analysis that provided a focus on policy, guided consideration of text and context, and enabled us to answer the 'what is data' question. Rather than asking 'what is going on here and how do we know it?' as we were encouraged to do on the Key Concepts course (Shaw, Copland and Snell, this volume), interpretive policy analysis encouraged us to ask 'how does a policy mean?' (Yanow, 1996), and 'how do we know it?'. This guided us to collect different sources and genres of data (documents, interviews) that, together, could provide insight into the overarching architecture of health policy and the narratives, coalitions and arguments allied to it.

Our study of think tanks' role in shaping health policy has illustrated the benefits that can be gained from a dialogue between linguistic ethnography and interpretive policy analysis. Linking language, context and social practice has been particularly valuable. This, of course, is characteristic of other work in the social sciences (Hammersley, 2007; Tusting and Maybin, 2007), with researchers having long been interested in the intimate relationship between policy, political language and political acts (see, for instance, Edelman, 1988; Degeling, 1996; Yanow, 1996; Bacchi, 2000; Fischer, 2003). History aside, it seems that linguistic ethnography has provided us with what Rampton (2007) calls 'a site of encounter': an opportunity to engage with and reflect on 'established lines of inquiry' (for instance, sociolinguistics), to introduce new foci for linguistic ethnography research (for instance, policy and planning), blend linguistic ethnography with relevant social theory and methodology (for instance, through the lens of interpretive policy analysis) and actively engage with an interdisciplinary community of like-minded scholars.