Dilemmas of discourse (analysis)

Jan Blommaert, Discourse: A critical introduction (Key Topics in Sociolinguistics). Cambridge: Cambridge University Press, 2005. Pp. xiii, 299. Pb $32.99.

James Paul Gee, An introduction to discourse analysis: Theory and method. 4th ed., revised. London: Routledge, 2003. Pp. 176. Pb $29.95.

Sara Mills, Discourse. London: Routledge, 2004. Pp. viii, 168. Pb $19.67.

Jan Renkema, Introduction to discourse studies. Amsterdam: Benjamins, 2004. Pp. 363. Pb $42.95.

Henry G. Widdowson, Text, context, pretext: Critical issues in discourse analysis. Oxford: Blackwell, 2004. Pp. 185. Pp $27.95.

UNDERSTANDING THE CONCEPT(S) OF “DISCOURSE”

Reviewing five stimulating books within the scope of one short article is by necessity doomed to fail in several respects: It is impossible to do justice to each book in an appropriate way, apart from briefly summarizing the most important aspects of the contents of each book. I will therefore attempt to point to innovative, sometimes controversial ideas and relevant debates with respect to the field of discourse studies (DS). However, some details will have to be neglected by the very nature of the review article genre. Moreover, the many cross-disciplinary links cannot be adequately addressed. I therefore apologize in advance and leave potential readers to judge on their own. The five books were selected on the basis of their publication date. I am, of course, aware of the fact that many other relevant books on discourse analysis (DA) and DS have appeared in recent years which have informed my own knowledge and assessment but can only be touched upon in passing. (See, for example, Howarth 2000, Johnstone 2002, Phillips & Jørgensen 2002, Schiffrin, Tannen & Hamilton 2001, Titscher et al. 2000, Torfing 1999, van Dijk 1997, Wetherell et al. 2001, Wodak & Meyer 2001.)

Before venturing into more detail, it makes sense briefly to clarify some of the many different meanings of “discourse,” several of which are discussed extensively later on. The term “discourse analysis” stems etymologically from the Greek verb ana-lyein ‘deconstruct’ and the Latin verb discurrere ‘running back and forth’. The compositum “discourse analysis” has in recent decades penetrated many disciplines, such as sociology, philosophy, history, literary studies, cultural studies, anthropology, psychology, and linguistics. In all these disciplines the term carries distinct meanings, ranging from a social science methodology to the label for a whole field, a subdiscipline of linguistics, a critical paradigm, and so forth. Reisigl (in press) lists 23 meanings of “discourse” used by Michel Foucault throughout his famous lecture in the Collège de France on “Orders of discourse,” while Maas 1988 demonstrates that the meaning of “discourse” has shifted from “scholarly deliberation” to “dialogue” in recent years (see also Wodak 1996:20ff.). In his seminal lecture, Foucault formulates a number of crucial axioms about the nature and contexts of discursive events (énoncés):

Although Foucault refers to many definitions of “discourse” in the course of his famous lecture, it is equally important to note what “discourse” is not supposed to mean in Foucault's work – specifically, that it is neither defined thematically nor by a strict system of concepts, and that it is not an object but rather a set of relationships existing between discursive events. These stipulations open the door to a dedicated functional approach, enabling the cultural critic to identify both static and dynamic relationships between discursive events and to address the causes and consequences of historical change (see Wodak 2005a, 2005b).

However, in the tradition of Wittgenstein's “language games” (1967) and thereafter Austin's “speech acts,” (1960) “discourse” is mainly understood as linguistic action, be it written, visual, or oral communication, verbal or nonverbal, undertaken by social actors in a specific setting determined by social rules, norms, and conventions. As early as the 1990s, while distinguishing DA from text linguistics, van Dijk (1990:164) defined “discourse” as “text in context”; the latter concept probably being one of the most complex, vague, and challenging notions for research in DA (Panagl & Wodak 2004, Duranti & Goodwin 1992; and see below).

Furthermore, language-specific meanings exist as well as distinct uses within the Anglo-American academic community on the one hand, and European scholarship on the other. For example, in British research, the term “discourse” is frequently used synonymously with “text,” meaning authentic, everyday linguistic communication. The French “discours,” however, focuses more on the connection between language and thought, for instance meaning “creation and societal maintenance of complex knowledge systems” (Ehlich 2000:162). In German pragmatics “Diskurs” denotes “structured sets of speech acts.” In the analysis of discourse, the meaning of the notion of discourse is therefore closely linked to the particular research context and theoretical approach (see Titscher, Meyer, Wodak & Vetter 2000).

It is not within the scope of this review article to elaborate further on the quasi-inflationary uses of the notion of “discourse.” Nor is it – unfortunately – feasible to discuss the relevant philosophical debates between, for example, Michel Foucault, Jürgen Habermas, Ernesto Laclau, Seyla Benhabib, Hannah Arendt, or Niklas Luhmann here; it is important, however, to acknowledge that discourse analysts should be required to present their theoretical background and consider other approaches beyond the necessarily limited scope of their school, discipline, or academic culture. This implies, especially for linguists who by nature of their discipline should be competent in more than one language, including literature from different research paradigms in different cultures, in languages other than English; unfortunately, this is rarely the case in the Anglo-American world, where references are more often than not restricted to research published in English by authors of British, American, Canadian, or Australian origin, interspersed with a few translations from prominent, often “trendy” scholars.

INTRODUCING DS/DA

“Discourse Studies is the discipline devoted to the investigation of the relationship between form and function in verbal communication” (Renkema 2004:1). This definition is the starting point of an extensive discussion about what the specific features of discourse might be and which functions are realized through distinct choices in form. Furthermore, it raises the question of whether a separate discipline – DS – is necessary at all (1). Throughout this book, Renkema illustrates in an extremely accessible way which textual forms recur as patterns in specific social contexts and which different theoretical and methodological approaches could be used to analyze discursive practices adequately. Although Renkema's book covers a wide range of approaches, he succeeds in presenting these with small illustrative examples of analysis in the context of their respective philosophical traditions.

He starts out in Part I by introducing Bühler's famous Organon model, and proceeds to Speech Act Theory, Relevance Theory, Conversational Maxims, and Politeness Theory. These theories are framed as “Communication as action,” followed by a second chapter in Part I, “Discourse in communication,” in which the author discusses the foundations of linguistic pragmatics, the classical definition of “sign” by Peirce and Morris, and communication models imported from communication theorists such as Shannon and Weaver, finally venturing to the central question, “What makes discourse discourse?” (48). In this short section, Renkema draws on etymology as well as on text linguistics; he focuses on cohesion and coherence, which define the specific syntactic and semantic connections between sentences producing “texts,” in contrast to a sequence of unrelated sentences. The famous seven criteria for establishing textuality proposed by de Beaugrande & Dressler 1981 are introduced at this point. Thus, Renkema situates DS as nodal point between text linguistics, pragmatics, sociolinguistics (by drawing on Hymes's speaking model, which defines the most important dimensions of any social context) and Hallidayan Functional Grammar (Halliday 2002). Such an integrative approach based on the historical roots of DA allows readers to understand the development of a scientific paradigm and its complexity. Moreover, Renkema draws on the relevant scholarly literature across national traditions, mostly in German and English.

Part II is dedicated to linguistic forms of discourses in all dimensions of language, from genre, to discourse semantics, up to syntactic connections, contextual phenomena, and style. I regard this part as the core of the book because it leads readers through the most complex and difficult aspects of DA: How do we define a “topic” of a text? Which are the most relevant cohesive devices? What are the salient characteristics of multimodal discourse in relation to other genres, such as conversations, or written documents, or even electronic media? In sum, this section enables readers to confront the “hard work” of detailed and explicit linguistic analysis in a systematic way.

Of course, such a section cannot be digested in one go. The different linguistic structures and forms need to be identified and applied in relation to specific research questions. Moreover, each section contains about two pages of questions and assignments, as well as annotated bibliographical information that allows readers to practice their acquired knowledge and to use this book independently of any taught course.

Part III encompasses four special modes of communication, which are selected because of the relevance of genres: conversation, informative discourse, narratives, and argumentation and persuasion. Most discourses and texts contain elements of all these modes. When reading through this part, it is striking that the usual polemics are absent: Renkema does not get involved in debates between conversation analysis and more hermeneutic approaches. He focuses on the applicability of approaches and not on the dogmatic issues of “what is right or wrong” that have dominated recent debates. Readers are left to form their own judgments.

Renkema also draws on theories and models that are largely unknown or non-mainstream: models to measure text comprehensibility, psycholinguistic and cognitive approaches to text planning and text production, Toulmin's argumentation theory juxtaposed with Petty and Cacioppo's Elaboration Likelihood Model, and so forth. This enables readers (students and scholars alike) to learn about the range of existing options without having to reinvent the wheel when undertaking their own analysis. Finally, Part IV is dedicated to special interests within three domains (from many other available choices): “Discourse and cognition,” “Discourse and institution,” and “Discourse and culture.” This choice reflects DS very well. Social problems are addressed, and this enables the reader to get acquainted with research on metaphor and discourse processing as well as on gender, racism, politics, law, and health care. In sum, DS is applied to interdisciplinary research. The last part of the book contains the key to the questions, answers to the assignments, an extensive bibliography, and the index. This book can be used as textbook, handbook, reference work, or exercise book; it opens up a field and defines it at the same time – probably for years to come.

An introduction to discourse analysis by James P. Gee is a very different kind of introduction to the field of DS. Gee presents his unique approach to DA through extensive definitions using his own terminology and examples, which are analyzed step by step. This is the third edition and the fourth reprint, which points to the popularity of Gee's approach and to his capacity of touching upon both mainstream and cutting-edge research and a range of unsolved problems in the field. Gee starts off by defining, not unlike Renkema, the meaning of “language” and the main functions of discourse: “If I had to single out a primary function of human language, it would not be one, but the following two: to scaffold the performance of social activities (whether play or work or both) and to scaffold human affiliation within cultures and social groups and institutions” (Gee 2003:1). What does Gee mean by this quite general definition? Basically, he points to the dialectic among institutions, cultures, situations, and agents who perform activities produced and reproduced in these settings. The dialectics consist in the fact that contexts and actors as well as activities influence one another simultaneously and produce and reproduce one another, and are thus also constantly changing (which relates to a definition in Fairclough & Wodak 1997, where we stress a similar position, albeit in a different framework).

Gee continues that “language-in-use is everywhere and always ‘political’” (1). What does “politics” mean in this context? “By ‘politics’ I mean something that Aristotle would have recognized, though not, perhaps, today's ‘Democrats’ and ‘Republicans’. By ‘politics’ I mean anything and anyplace where human interactions and relationships have implications for how ‘social goods’ are or ought to be distributed. By ‘social goods’ I mean anything that a group of people believes to be a source of power, status, or worth” (2). Very explicitly, Gee introduces his critical and evaluative stance, in common sense and everyday language, using many metaphors that certainly appeal to an American readership. Little or no reference is made to the vast literature in the social sciences, where the definitions of “politics” or “values” fill whole volumes over centuries. This is still legitimate, as it allows readers to link such explanations to their everyday knowledge. However, such an approach also implies dangers: Scholars from neighboring disciplines might feel that their research is not presented in an adequate form. An annotated bibliography would be of help here, perhaps an idea to be considered for the next edition.

Gee continues to set out his terms of reference. He notes that no DA theory can be right or wrong, just more or less adequate. I could not agree more. Moreover, he also states that no theory belongs to or has been created by just one person: All of us are connected through many debates and discourses, and thus we consciously or subconsciously draw on one another's work. This is a true statement that all our students should be introduced to right at the beginning of their studies. Scientific endeavor is all about relating the known to new issues and discovering some innovative aspects or applications. None of us works in an isolated way; Gee acknowledges several major influences on his approach throughout this book.

The goals of DA are also defined: to clarify “tools of inquiry” and strategies of applying these to texts of different kinds, used to investigate problems of a social nature. Throughout this book, the metaphor of scaffolding or construction prevails. Gee teaches us how we “build” our world through communicative activities and defines six areas of “reality” (12): aspects of the material world; activities; identities and relationships; politics; connections (used in the sense of intertextuality); and semiotics (using different symbol systems). In analyzing specific activities, all these six areas have to be analyzed. The analysis itself needs “tools of inquiry,” which consist of situated identities (which positions are enacted in a setting), social languages (different styles), Discourses (with a capital D; integrating language with non-language, i.e. material goods), and Conversations (with a capital C; long-running themes or motifs throughout different texts or interactions) (12–13).

Once a reader has actually learned all these definitions by heart, the analyses throughout Gee's book become understandable and interesting. However, I personally always find it difficult to learn new meanings of terms that have already become conventionalized throughout a field, such as “conversation” or “discourse.” One has to pose the question of why Gee introduces such new uses of old terms – to define a particular point of view, to introduce a new methodology, or because all previous, other methodologies have proved to be inadequate; or too complex? I could well imagine using “leitmotif” or “theme” instead of “Conversation,” or “registers/varieties/styles” instead of “social languages,” a term that implies so many associations, from Bernstein's “codes” to “sociolects.” On p. 38, in a footnote, Gee relates “Discourse” to “discourse” as used by Foucault, Lave and Wenger, Geertz, Bourdieu, or Wittgenstein. However, no space is dedicated to justifying the new term nor to distinguishing it from or relating it to other, similar notions. Gee only refers to literature published in English; very few prominent authors from non-English/American backgrounds whose work has been translated are integrated (Bakhtin, Wittgenstein, Foucault, Bourdieu, Heidegger, and Propp), although Gee's approach relates well to discourse analysts such as Ehlich, Jäger, or Maas (see Titscher et al. 2000). One hopes that our ever more globalized world will lead to the acknowledgment of research across a wider geographical area and in more languages.

The appendix to the book lays out the most important principles of Functional Systemic Grammar. It is a good idea to put such basic linguistic knowledge into a final chapter, so that readers who are not trained as linguists can look up relevant categories there; experts do not have to read the appendix. However, if Gee draws on Halliday, why not use Hallidayan terminology throughout?

What makes Gee's book so stimulating are the selected topics and problems that he addresses and the interesting range of examples chosen for analysis. Sample texts from Darwin (56), interviews with young Latinas (70), words like “coffee” or “work” related to their cultural frames (81), and long sequences of narratives are analyzed in great detail (a note on p. 98 draws our attention to similar methods of analysis, quoting Fairclough, Kress, Duranti, Gumperz, and van Dijk, unfortunately without further elaboration of these approaches). In my view, the most interesting example is to be found in chap. 7. Here, Gee provides readers with a step-by-step analysis of a story – “the return of the table” – which he deconstructs into “stanzas” (topical sequences), detecting the sub-stories, themes, and identities as well as paraverbal features. This methodology is certainly relevant for many scholars in the social sciences and humanities.

Sara Mills reflects on the multiple uses and meanings of the term “discourse” in the introduction to her book Discourse. She starts off by citing three renowned dictionaries (one of them French), then turns to more general introductory texts (Crystal, Benveniste, Hawthorn, Fowler) and finally comments on Foucault, whose approach informs this piece of research: “In this book, I aim to detail the use Foucault himself made of the term discourse” (Mills 2004:24). Mills investigates these uses first in contrast to an equally inflationary term, “ideology.” She continues throughout the book to reflect on the definition of discourse in feminist thinking and colonial and postcolonial theories, and she finally turns to linguistics, to Critical Discourse Analysis (CDA) and social psychological approaches. The value of this book lies in the huge range of research covered (including work from the French-speaking world) as well as in the critical review of various methods of DA, illustrated by short text examples.

The very detailed discussion of Foucault – in spite of the complexity and even inaccessibility of some of his writings – deconstructs the subtle and underlying meanings in his work. Specifically, the distinction between notions of ideology (such as that of Althusser) and Foucault's “subject-less” concept of discourse points to a much-neglected dimension of DS. Such a notion of discourse opens up an interesting question: Why do so many researchers in DA and CDA refer to Foucault's concepts while emphasizing simultaneously power relations and activities, obviously performed by actors? In my view, inherent contradictions become apparent (see Wodak 1996, Meyer 2001, Weiss & Wodak 2003). Obviously, we need to find ways to integrate Foucault's powerful theory of discourse with theories of communicative action, for example that of Habermas 1981. Chouliaraki & Fairclough 1999 also discuss other approaches, such as Giddens, Laclau and Mouffe, and Bourdieu in this context; an integrative theory, however, still needs to be developed (see above).

Mills takes up this point while elaborating on the impact of feminist theory on DA and comes to the conclusion:

The final two chapters of Mills's book emphasize the huge impact of Foucauldian analysis on postcolonial studies, and on CDA and other DA approaches. Mills challenges much research in CDA because it seems to have a “simplistic” view of power relations (140). However, although Mills does acknowledge Fairclough's research, it is a pity that she does not consider other work in CDA, such as Teun van Dijk's Socio-Cognitive Model, Scollon and Scollon's Mediated DA, Blommaert's Sociolinguistic DA, or even my own Discourse-Historical Approach. Nevertheless, this book is to be highly recommended to any graduate student interested in the subtleties of DS and its philosophical dimensions and epistemological background.

NEW PERSPECTIVES IN DA?

Two new books have been published recently both of which set out to make an innovative contribution to the field of DS: Discourse by Jan Blommaert, and Text, context, pretext by Henry Widdowson. In promoting their own approaches, they both need to distinguish themselves from other approaches in some way: by reviewing them critically, in pointing to unsolved problems; by integrating some aspects while elaborating others; by polemical debates, thus reducing (or even distorting) relevant dimensions; or by finally not mentioning important research at all. Of course, this often is not intentional but rather occurs owing to the argumentative dynamic inherent in scholarly debates.

In one case, however, the polemical debate has a long history – a trajectory of criticisms and replies, between Widdowson, on the one hand, and Fairclough and other CDA scholars, on the other. (See Seidlhofer 2003 and Titscher et al. 2000 for a summary of these debates.) In this review, I will not refer to this long (and also boring) history of the ongoing debate unless it helps to clarify Widdowson's position.

Jan Blommaert explicitly positions himself in the tradition of ethnography, anthropology, sociolinguistics, and critical research. Instead of analyzing “power alone” (emphasis in original), Blommaert attempts to offer an “analysis of power effects, of what power does to people, groups, and societies, and of how this impact comes about. The deepest effect of power everywhere is inequality, as power differentiates and selects, includes and excludes” (Blommaert 2005:1–2). This, of course, has been an important issue in sociolinguistics since the 1960s. Blommaert combines this goal with DA: “Discourse is language-in-action, and investigating it requires attention both to language and to action” (2). This statement leads us back to Wittgenstein's notion of “language game” and to the first part of Renkema's book (see also “Discourse-Sociolinguistics,” Wodak 1996). Blommaert draws on text linguistics and pragmatics but also widens the field by including “nonlinguistic” conceptions and features (3), quite similarly to Gee. In sum, Blommaert defines his book “as not a linguistic book” (3), but rather an analysis of activities, contextualized in a space-time frame (i.e., in a material context). Blommaert clearly relates to much research by R. Scollon and Van Leeuwen, although the latter is not mentioned at this point. Another relevant concept introduced by Blommaert is “voice,” though not in the Bahktinian sense: “Voice stands for the way in which people manage to make themselves understood or fail to do so. In doing so, they have to draw upon and deploy discursive means which they have at their disposal, and they have to use them in contexts that are specified as to conditions of use” (4–5). This implies that – if the conditions are not met – people lose access or are misunderstood, or “fail” to make themselves understood.

After exploring the important influence of linguistic anthropology and sociolinguistics on his own work, Blommaert lists five principles that guide us throughout this book: In analyzing language, the focus needs to be on what language use means to its users; language operates differently in different contexts; the units of analysis are actual and densely contextualized forms in which language occurs in society; language users have repertoires which they employ owing to contextual needs; and communication is influenced by the structure of the world system. Apart from the last, all these principles were established long ago across the disciplines. They relate to all the books reviewed but also to research in pragmatics, DA, CDA and sociolinguistics. Only the global dimension, which Blommaert quite rightly emphasizes, has been neglected so far.

Surprisingly, however, Blommaert states on p. 16 that little similarity exists between these principles and CDA, which only postulates ethnographic and contextualized research but does not actually perform it (however, see detailed overviews of CDA in Wodak & Meyer 2001, Weiss & Wodak 2003, Wodak & Chilton 2005). On the other hand, one of the many points that has to be accepted as valid criticism is the noticeable absence of fieldwork in non-European countries (Blommaert, 35; but see Lazar 2005).

Moreover, Blommaert criticizes the “linguistic bias” in CDA (34–35). This implies that the focus of CDA is on language and stops “as soon as the discourse has been produced” (35). I could not agree more: It is salient to investigate what is absent, what is not said, what happens after something was said, who has access to speak and to speech at all, who stays silent, and so forth (see e.g. Thiesmeyer 2004). I also agree that this involves a great deal of ethnographic fieldwork (which I know from long experience), and that only a few researchers in DA and CDA have been involved in this kind of research. However, generalizations are never productive; they are – as in this case – easily proved to be wrong.

Even more surprising is Blommaert's emphasis on CDA's neglect of context, although van Dijk has published extensively on issues of contextual knowledge (van Dijk 2003); Blommaert exemplifies this gap by pointing (51–52) to one of my own papers on doctor-patient interaction (Wodak 1997). I have to apologize for going into detail at this point, having compared Blommaert's presentation with my original paper and – more important – with our book on doctor-patient communication (published in German in 1990 and not quoted by Blommaert). My 1997 paper explicitly draws on the book and mentions the ethnographic fieldwork on pp. 178–79 and p. 195 and in fn. 3. Nevertheless, Blommaert labels the necessarily brief contextual information in a chapter of an edited volume (in contrast to a book-length study) as “prima-facie ethnographies,” used as “framing devices in analyses” (51), and states that “context is often a mere background” in CDA (53).

Our extended ethnographic research in organizations of the European Union (Muntigl, Weiss & Wodak 2000) is also neither mentioned nor quoted. It is a pity that Blommaert needs to downplay other research in order to promote his own important findings and innovative theoretical concepts. Obviously, one might be tempted to say: the more agreement exists, the more one needs to distinguish oneself – something Sigmund Freud (1930) noted long ago when studying group dynamics and the genesis of stereotypes.

The central part of Blommaert's book is dedicated to his research in Belgium (on asylum seekers), in the Congo (on literacy), and on the televised hearings in the Truth and Reconciliation Commission (TRC) in South Africa (see also the 2006 special issue of the Journal of Language and Politics on “Critical linguistic perspectives on coping with traumatic pasts”). These detailed case studies of oral narratives, written documents, handwritten letters, pictures, and many more genres are fascinating to read, as they illustrate the recontextualization of both verbal and nonverbal factors between texts and contexts (“entextualisation” in Blommaert's terminology). Blommaert's many-layered interpretations of textual histories also draw greatly on historical knowledge and sociopolitical expertise.

At this point, another personal remark should be permitted: On p. 38, the author mentions Martin & Wodak 2003 as an “attempt to move CDA into history.” Many readers will imagine my astonishment in reading this sentence in the “Suggestions for further reading” section because Blommaert does not mention the “Discourse-Historical Approach” – although he quotes research employing this type of CDA, first advocated in 1990 in German to analyze postwar anti-Semitism in Austria, and soon afterward discussed in many papers in English, French, Italian, Russian, and German. CDA has thus been involved in interdisciplinary historical research for more than 15 years (see also Flowerdew's 2002 research on Hong Kong, for example). The theme of “social change” is certainly not new in critical research. One of Fairclough's seminal books, Discourse and social change (1992), which influenced several generations of students, needs to be mentioned at this point.

Because of space restrictions, I can only summarize one of the many in-depth case studies in this book (pp. 142 ff). In November 1996, the author was invited to attend a workshop in the Netherlands on the 1944 Warsaw Uprising (not to be confused with the uprising of the Warsaw Ghetto a little earlier) (see Ensink & Sauer 2003). In a very close step-by-step analysis and by drawing on his own participant observation, the author deconstructs the various explicit and latent functions of this workshop, which only few people were aware of at that point. By integrating much more archival material on the Uprising, on the role of the Soviet Army, and on the postwar relationship between Poland and the Soviet Union, now Russia, the contradictory perspectives provided by the participants and their analyses of political speeches were deconstructed.

Interestingly, the Polish participants seemed to equate the Soviets with Russians after 1989 – quite wrongly, of course, as if the Nazis were to be equated with the German population nowadays. A specific historical narrative was thus constructed throughout this conference that links 1944 straight to 1989, a linear path of Poland from occupation and suffering to freedom, by neglecting all other voices. Such official national narratives are discursively constructed in most nation-states and instrumentalized to provide a hegemonic framework with which many can identify (see Heer et al. 2003).

Blommaert's leitmotif is spelled out at the end of the book: “I believe that the truly intellectual challenge for discourse analysis is precisely that: to reconfigure our own discipline in a permanent process of self-critique based on observations of the way in which our object of study changes” (p. 238). Again, I could not agree more, and I would like to add, that – apart from self-reflection – dialogue inside and across disciplines is essential as well.

Henry Widdowson starts his preface by stating that “this book is, in a sense, a reconceptualized and extended version of one that was unwritten und unpublished thirty years ago” (2004:vii). In this unpublished book, Widdowson continues, he would have already touched upon many issues and aspects of DA that were addressed later on, albeit by other scholars. Unfortunately, we all didn't have the chance to read the first (unwritten) book; the new book, however, is certainly worth reading and touches on many controversial issues.

In contrast to Blommaert, Widdowson does not present a distinctive approach with empirical examples of theory application. The dense challenging of other work requests extensive background knowledge and expertise across a whole range of DS and DA theories. This makes the book difficult to access; it certainly cannot be used in teaching except as an example of controversial (or even polemical) writing.

In promoting his own notions of “text, context, pretext,” all of which are certainly at the center of scholarly debates (Panagl & Wodak 2004), Widdowson attempts to distinguish himself, inter alia, from more sociolinguistic approaches (Stubbs, Chafe, Firth, Hymes), from pragmatics (Grice, Sperber & Wilson), from Functional Systemic Grammar (Halliday, Hasan), from CDA (Fairclough, van Dijk, Wodak), from corpus linguistics (Stubbs), from critical linguistics (Fowler, Kress), and possibly from others. If one looks at the research that he challenges, one can observe that he includes very little recent work (specifically from scholars whom he criticizes; see below), and no research in languages other than English, and that he leaves out much relevant research. For example, text linguistics is not written about; the important and path-breaking ideas of de Beaugrande, Dressler, and van Dijk in the 1970s and 1980s are not picked up; and new research in corpus linguistics is also absent (e.g. Baker & McEnery 2005).

We are all of course aware that theories and methodologies are changing, that we as scholars are involved in many fruitful and stimulating debates with our peers, and that therefore – quite often in response to justified criticism – we also change our perspectives or elaborate previously undeveloped parts of our approaches. Thus, constructive criticism is more than welcome (see Meyer 2001, Chilton 2005, Billig 2003). Unfortunately, these quite normal and trivial phenomena are not integrated in Widdowson's challenge of the whole field, as if science stood still.

The criticism of CDA seems to be on the top of Widdowson's agenda and has been for years (see above). Nevertheless, at several stages in this book he does emphasize how much he still admires the prominent proponents of CDA, although later on he accuses them of being unserious scholars; I must confess that this rhetoric reminds me of Marc Antony's famous speech about Brutus. For instance,

Of course, we can agree to disagree and discuss our different opinions in a rational manner, maybe in the kind of utopian discourse proposed by Habermas 1981 for scholarly debates. However, at the end of Widdowson's book, the mode of criticism shifts:

Widdowson continues that students need to be taught a methodology:

Such a patronizing tone is surprising in a scholarly book that does not consider the vast array of recent books on methodology in CDA, such as Fairclough 2003, Scollon & Scollon 2003, 2004, Chilton 2004, Van Leeuwen 2005, Wodak & Meyer 2001, Reisigl & Wodak 2001, and many handbook articles dedicated to a systematic description of methodology (for example, Wodak 2004, Fairclough 2001, Van Dijk 2001, Fairclough & Wodak 1997, Jäger 2001). Widdowson (139) bases most of his criticism on his reading of Titscher et al. 2000 (German edition, 1998), a necessarily brief overview of 14 qualitative and quantitative methods of text and discourse analysis where we explicitly state that it is impossible to give thorough examples of analysis because of the genre and scope of the book. Because of Widdowson's lack of expertise on the subject, I do not want to continue this debate; I am prepared to take it up again when Widdowson has been able to digest the new developments in the field.

In the following, I briefly summarize Widdowson's own approach. The first chapter is dedicated to a discussion of the terms “text” and “discourse.” Widdowson observes quite rightly that these terms are used in many, vague, and contradictory ways. He argues that one needs to distinguish among locutionary, illocutionary, and perlocutionary forces and effects of utterances, hence also between the descriptive analysis and the interpretation of texts; the latter necessarily carries values of the interpreters (the whole debate on hermeneutics is thus evoked). The text comes first: “We read plausible meanings into the text, prompted by the purpose and conditioned by the context. In other words (in my words) you derive a discourse from it and it is that which realizes a text as text” (19).

If I understand Widdowson rightly, this means that a neutral, value-free linguistic analysis of the text is possible, and, as a next step, through adding on contextual information, we then arrive at interpretations (our readings) of the text, which serve to transform a “text into a discourse.” I agree with Widdowson that analysis should be separated from interpretation; I also agree that the analysis needs to be systematic and that the terms need to be spelled out. Moreover, I agree that often this has not been the case. However – and here I do not agree – even a descriptive and close linguistic analysis will always depend on specific research questions; and these are by nature driven by interests and values. Otherwise, we would end up with endless analyses of many features without having defined the goals of a given analysis. Therefore, the hope that text and discourse, analysis and interpretation could stay totally distinct (in Widdowson's terms) becomes a positivistic illusion. Let me quote the German historian Leopold Ranke (1867:103) in this context. Ranke wrote about text interpretation and his futile wish “das Selbst gleichsam auszulöschen” (‘to quasi-delete the Self’) when analyzing historical sources.

Another central notion of this book is “context.” Widdowson suggests a cognitive concept instead of the traditional static notions of context. Context is defined as “schematic construct,” much in line with van Dijk's most recent proposals. These schemata are not static, but related to sociocultural conventions. What Widdowson does not elaborate at this point is the operationalization of such a concept of context. Who develops which schemata, and how do we study cognitive concepts at all? Moreover, if we all always use schemata when reading, writing, or interpreting texts, we most certainly also employ them in analyzing texts as well, because the process of analysis necessarily implies reading and understanding the text.

The debate about context leads us to the chapter on “pretext” (74 ff). What is “pretext”? “The term ‘pretext’ generally refers to an ulterior motive: a pretending to do one thing but intending to do something else” (79). As far as I understand Widdowson in this context, “pretext” accompanies our interpretations of texts, guides them, and could be termed “presuppositions” or “background knowledge.” Pretext allows us to mark what is relevant (owing to our cognitive schemata). However, Widdowson also applies this concept to illustrate that some interpretations of CDA are mainly based on “pretext” and not on systematic analysis. At times this has certainly been the case; if one elaborates a theory, often enough some typical examples are used to illustrate this. If, on the other hand, one chooses an abductive procedure – that is, proposing theoretical claims and at the same time an empirical study that could verify or falsify the hypotheses – examples are not chosen for illustration but used as evidence. Both procedures are legitimate and are to be found in DS and CDA (see Meyer 2001).

Methodology, in my view, should be retroductive; that is, any research procedure and device should be transparent and explicit, and thus allow recipients to trace and understand the operationalization and interpretation of results. Unfortunately, Widdowson does not provide us with an example of systematic methodology. He – only in passing – suggests that reception studies might be able to validate some reformulations of texts (p. 171) but observes that no studies of this kind exist (however, see Wodak & Busch 2004). In a note, Widdowson criticizes our study on reformulating news broadcasts in a more comprehensible way (p. 174), but unfortunately without reading the whole study (published in German; Lutz & Wodak 1987, Pfeiffer et al. 1987), where explicit categories for analysis are spelled out.

DS and DA now belong to the mainstream of linguistic and interdisciplinary research. The five books reviewed in this article illustrate that the field is vast and growing, and that many approaches exist in parallel. These five books argue convincingly that DS and DA also contribute to interdisciplinary research and to relevant applications in linguistics and in the social sciences. Accessible textbooks open up the field for newcomers; innovative approaches elaborate theories and stimulate research. Central concepts, such as discourse, text, and context, necessarily remain vague for many. The “dilemmas of discourse analysis” thus are still unsolved or maybe not to be solved: For some authors, DA is too linguistically oriented, for others not linguistic enough. For some, DS lies at the core of linguistics; for others, outside linguistics. Precisely such controversies make research in DS interesting and worthwhile.