The interdisciplinarity of critical discourse studies research

Interdisciplinarity has been a core tenet of critical discourse studies—a group of approaches to the analysis of texts in their social contexts—since its inception, in what may be seen as a reaction against the sometimes staid and rigid disciplinary boundaries of linguistics and other disciplines. Interdisciplinarity has also been seen as necessarily accompanying analyses of complex social problems such as racism, sexism or other forms of discrimination and social domination. The concept has been multiply re-examined, challenged and reaffirmed by critical discourse scholars (for instance, in Weiss and Wodak, 2003), and the present article continues this work by mapping out the present-day dimensions of interdisciplinarity in different approaches to critical discourse studies. It also attempts to juxtapose these various disciplinary developments, and consider whether interdisciplinarity in and of itself has come to be taken for granted. Finally, it raises questions about whether the move away from an emphasis on the analysis of social wrongs within some of the newer approaches to critical discourse studies may in time lead to a disciplinary schism, or whether the increasingly disciplinary nature of critical discourse studies itself may have become a problem. This article forms part of an ongoing thematic collection dedicated to the concept of interdisciplinarity.


Introduction
C ritical discourse studies (CDS), as critical discourse analysis has increasingly come to be called in recent years, has been conceived of as an interdisciplinary (or sometimes transdisciplinary) endeavour since its outset (Fairclough and Wodak, 1997;Wodak and Weiss, 2003: 1), in what may be seen as a reaction against the sometimes staid and rigid disciplinary boundaries of linguistics and other disciplines. This claim to interdisciplinarity is evidenced at three different levels: the theoretical origins of CDS, the methodologies typically used by critical discourse scholars, and the research contexts in which CDS is applied (see also Wodak and Meyer, 2015a in press). Theoretically, CDS draws not just on linguistics but on critical theory, sociology, politics, psychology and cognitive science, among others. The different methodologies used by scholars, while not fixed or prescriptive, have typically been drawn from traditions such as text linguistics, social psychology, ethnography, corpus linguistics and many others. In terms of application, CDS has widely been applied to different social, political and cultural phenomena that are not necessarily only to do with language, and indeed this has also been seen as an indispensible orientation when analysing complex social problems (Wodak and Meyer, 2015a). The concept of interdisciplinarity in CDS has been multiply reexamined, challenged and reaffirmed by critical discourse scholars (for instance, in Weiss and Wodak, 2003) since they first coalesced into an identifiable group in the late 1980s. This comment article will attempt to examine some of the different ways of understanding interdisciplinarity and consider whether interdisciplinarity in and of itself has come to be taken for granted in different ways within the field.
It would be unnecessary, not to mention impossible, to attempt to replicate in this brief review article the volume-length treatment of interdisciplinarity in CDS edited by Weiss and Wodak (2003). Instead, while referring to this and other previous work where applicable, I will focus on mapping out the presentday dimensions of interdisciplinarity in different approaches to CDS. There appears to be a tension between two directions of travel inherent in many current debates about CDS and its role in relation to linguistics: one centrifugal, pushing away from linguistics, the other centripetal, pulling back towards the discipline (see, for instance, the comments on Meyer, 2009, in Hart andCap, 2014, and the response by Wodak and Meyer, 2015a). On the one hand, there is the emphasis on interdisciplinarity from the outset, and the attempt to not only acknowledge the theoretically diverse origins of the approach as conceived by the "founders" (generally considered to be in alphabetical order Norman Fairclough, Gunther Kress, Teun van Dijk, Theo van Leeuwen and Ruth Wodak), but also bring in new insights from other disciplines by reading and collaborating widely (Wodak and Meyer, 2015a). On the other hand, there is a concern that CDS has become too far removed from what is sometimes seen as its core analytical concern, namely the analysis of linguistic (or more broadly semiotic) forms (Hart and Cap, 2014). However, the latter concern somewhat sidesteps the issue that linguistics itself is an extremely broad and diverse discipline, and has co-opted theories and methods from numerous other disciplines, for example, psychology and neuroscience (cognitive linguistics), computer science and mathematics (corpus and computational linguistics), or biology and physics (phonetics) to name just a few. Regrettably, there are still some linguists who argue that linguistics should only be concerned with the forms and structures of language at sentence-level and below, and while it is certainly useful for CDS scholars to take account of recent developments in describing and understanding language in terms of, for example, syntax or phonetics, it is hardly worth revisiting this now very dated debate.
Interdisciplinarity at different stages of the research process To break out of this rather reductive way of thinking about the role of CDS within (or at the margins of) linguistics and other disciplines, I suggest we start by thinking about how interdisciplinarity might apply at the different stages of the research process. Wodak and Meyer's well-known diagram showing the recursive nature of the research process provides a useful template (Fig. 1).
Starting at the top of the circle, as already suggested above, the theories that critical discourse scholars draw on may have their origins within a broad range of disciplines-as Chouliaraki and Fairclough (1999: 16) put it, CDS involves "bringing a variety of theories into dialogue". In applying them to the research problem (moving from theory to operationalization), they may be to a certain extent divorced from their "original" disciplines, while at the same time being adapted and changed to the problem at hand, which will inevitably be influenced by the disciplinary orientation of the researchers. The concept of ideology, for instance, which has been widely discussed in politics, sociology, psychology and philosophy, to name just a few, has been important for most CDS scholars (not least van Dijk's, 1998, book length treatment). The concept has been applied in the analysis of parliamentary speeches and debates (for example, Fairclough and Fairclough, 2013), the discourse on employment and national identity (Richardson and Wodak, 2009) or in my own work on minority language policy (Unger, 2013), to name just a few studies.
The procedures and instruments used, in other words the methodologies developed for a specific CDS project, will again be strongly influenced by disciplinary conventions. Savski (Forthcoming), for instance, draws heavily on interpretive policy analysis to develop his framework for the analysis of policy formation in the Slovenian parliament. At the level of discourse/ text, decisions about specific linguistic phenomena will in turn be dependent on which traditions within linguistics the researchers are familiar and comfortable working with, and which are most relevant to the questions at hand. In fact, there have been calls to combine different forms of linguistic analysis (the term triangulation is often used for this, see, for example, Baker and Levon, 2015, for a recent application). Finally, the interpretation, including the scope and specific focus of what is interpreted, will also be strongly influenced by disciplinary traditions. For instance, ethnographic approaches allow rich description of discursive practices within specific contexts without necessarily generalizing to other contexts, while corpus analysis will allow inferences to be drawn about potentially huge volumes of texts, but perhaps for a more limited range of discursive features.

Theory, method and analysis
The theoretical origins of CDS have been introduced, discussed and deconstructed numerous times (see, for instance, Meyer, 2009, 2015b;Hart and Cap, 2014) so I will not revisit them here. However, it is worth making a brief digression into the current orientations of different approaches within CDS, as recently outlined by Hart and Cap (2014). As Hart and Cap show, there is a complex interplay of different theoretical building blocks and objects of investigation typically tackled by each approach. CDS approaches are not mutually exclusive, and of course there is interaction between and within CDS approaches, as Hart and Cap and others have pointed out, but nevertheless, the idea of identifying different approaches is useful in understanding what unites them and also separates them, where they are in dialogue and where they are in conflict on key theoretical issues. Wodak and Meyer (2015a) arrange the approaches along a cline from inductive to deductive and show "theoretical attractors", while Hart and Cap (2014), apart from adding additional approaches, attempt to show the relationships between different approaches and their "methodological attractors", typical micro-level objects of analysis (for example, argumentation). Furthermore, they suggest two major axes along which approaches can be arranged: cognitive/functional and content/ structure. While this is a useful way of categorizing approaches, as Hart and Cap acknowledge it does not fully do justice to the complexity of approaches-particularly as several approaches suggest recursive, multi-stage analyses, which may for instance involve analysis of both content and structure at different stages, or may include both cognitive and functional dimensions. So for the purposes of this review, in order to illustrate how the approaches coalesce around a common core of CDS, I argue instead that we should categorize CDS approaches according to three broad dimensions based on their main analytical concerns, or "analytical attractors", to continue in the vein of the taxonomies outlined above: textual, social and cognitive/mental (see Fig. 2).
While each approach is represented by a single point in the figure to give its approximate location in relation to three dimensions, in fact they should instead be amorphous overlapping blobs with sometimes fuzzy, sometimes sharp edges, more complex even than Hart and Cap's (2014: 7) already nuanced representation. Each of the approaches takes account of phenomena that can be explained within each of these dimensions to a greater or lesser extent, and each dimension of course has intersections with different disciplinary traditions both within and outwith linguistics. The precise placement of each approach could be the subject of endless discussion and revision, and hopefully will lead to some discussion about the precise nature of our shared endeavour. However, it is my hope that most CDS scholars would agree that there are several identifiable clusters, with two being most salient: the more socially oriented approaches and more cognitively oriented approaches. Within each cluster some approaches take more or and some less account of micro-level textual features, broader social issues and cognitive aspects of discourse production, reception and interpretation. The clusters and, more broadly, the analytical dimensions given in the figure also give some indication of the different levels of attention paid by proponents of each approach to theories and concepts from linguistics and other disciplines-each vertex of the triangle suggests a tendency towards research focussed on textual, social and cognitive/mental approaches respectively within linguistics, and on cognate research in other disciplines. It should be added that all the dimensions interrelate, and almost all the approaches take account of all three dimensions to a certain extent-what the figure attempts to show is that within different approaches there is a different emphasis on analysing and explaining phenomena that are respectively textual, social or cognitive/mental. The corpus linguistics approach is something of an outlier here. In and of itself it lacks the more detailed analysis of social and/or cognitive/mental phenomena that some of the other approaches tend to employ, although it allows the analysis of huge volumes of text. It is also often combined with other approaches to allow more socially oriented interpretations. It should be added that the older, more established approaches tend to be oriented either along the social-textual axis (DHA, DRA, CritL, DP), or the cognitive/mental-social axis (SCA, SAM), while a number of the newer approaches Hart and Cap identify are moving more towards the cognitive/mental-textual axis (CCP, CogLA, CMA). This raises interesting questions about the extent to which CDS, as viewed by these newer approaches can remain critical in the sense of socially critical. With some of these approaches no longer foregrounding the social problem (or social wrong as Fairclough puts it in more recent work), this may represent the beginnings of a disciplinary schism within CDS.

The context of interdisciplinarity
The final form of interdisciplinarity I would like to deal with in this article relates to the contexts from which data for CDS studies is drawn. This again involves two different perspectives which relate to the centripetal and centrifugal directions of travel discussed above. On the one hand, there are scholars who consider themselves to be critical discourse analysts, perhaps having received their initial training in linguistics, or discovering CDS later in their careers. They need to study something in the "real world" outside academia, and thus seek to understand the context of their research. On the other hand, scholars who come from a wide variety of traditions may come to understand CDS as a suitable approach to make sense of the objects of investigation they have previously approached by other means. Whichever direction scholars come from, there is much to be gained by seeking a thorough understanding of different disciplinary approaches to, for instance, institutions, politics, or media. As Chouliaraki andFairclough (1999: 16, cited in Wodak andWeiss, 2003: 6-7) put it, in CDS "the logic of one discipline (for example sociology) can be 'put to work' in the development of another (for example, linguistics)". One example of this drawn from my own work (in collaboration with others) relates to digital media. Because CDS scholars have thus far not paid a great deal of attention to digital media (see KhosraviNik and Unger, 2015, in press for a fuller discussion), my collaborators and I (see, for example, Potts et al., 2014;KhosraviNik and Unger, 2015) found it necessary to draw on a variety of theories and methods more aligned with media and communication studies than linguistics. However, we also benefited from "looking sideways", by attempting to understand how sociologists, political scientists, information scientists, or cultural studies scholars, among many others, were grappling with exactly the same problem, in other words, how to integrate theories from different disciplines into their own work to understand new or changing social phenomena, discursive practices, and research contexts.
Turning the critical eye inwards A final consideration is whether CDS has developed to an extent where it can be considered as a discipline in its own right, which in turn depends on the very nature of disciplines (see Billig, 2003, for a fuller discussion). While there are at the time of writing no departments of CDS at universities, there are certainly journals, conferences and organizations dedicated solely to CDS. There are also postgraduate degrees with a core focus on CDS. At some institutions there may be many more scholars working within CDS than there are in some more traditional disciplines-for instance, in my own institution there are around a dozen active and retired scholars (and many more if you include past and current PhD students) who have published in CDS-related journals, while the Department of Music was closed some years ago. Of course there are other Music departments elsewhere, and some research on music is carried out in other departments, but the point is that the existence of a named entity does not a discipline make. Regardless of what label scholars use for CDSdiscipline, school, programme, approach, field, orientation or any other-we can acknowledge the widespread success and longevity of the ideas it has fostered and developed. But in turn, this should remind us that we should turn our critical eye inward and continually review whether we have become too rigid and staid in our thinking and in rejecting new ideas that do not conform to our established traditions, as Billig (2003) warns might happen. Billig also describes the emergence of critical discourse analysis as a "brand"-the use of the label CDS then, may be seen as a "rebranding exercise". We might ask, in the same self-critical vein, however, whether either of these labels fully captures the complex interplay between different theories, methodologies and contexts described above. Disciplinary modes of thinking may be either a foe or an ally in this self-critique: while we can draw on well-established traditions in other disciplines to open up CDS (following the centrifugal direction), we can also be centripetally trapped in our own disciplinary silos.