Approaches to Media Texts

Ruth Wodak & Brigitta Busch. Handbook of Media Studies. Editor: John D H Downing, Denis McQuail, Philip Schlesinger, Ellen Wartella. Sage Publication. 2004.

In our contribution, we focus on qualitative linguistic approaches to media texts—especially on the approaches developed within critical linguistics and critical discourse analysis. There are several important reasons for this choice: In recent decades, there has been a significant increase in international interest in applying qualitative research methods to the study of social and cultural processes. Although the traditional empirically oriented approach to media texts, mainly represented by quantitative content analysis, is still widespread in mass communication research (McQuail, 2000, p. 235), some observers (e.g., Jensen & Jankowski, 1991) speak of a “qualitative turn” in media studies. This shift of paradigm is not a question of preferences for particular methodologies but corresponds to conceptual and theoretical frameworks distinct from the traditional sender-receiver model.

We cannot, however, elaborate on all the important research in conversation analysis (CA) and sociolinguistics, which has been concerned with media analysis, due to the shift of paradigm mentioned above. CA emerged in the 1960s (see Titscher, Wodak, Meyer, & Vetter, 2000, for a summary). It is based on ethnomethodology (Garfinkel, 1967; Sacks, Schegloff, & Jefferson, 1974) as an interpretative approach to sociology, which focuses mainly on the organization of everyday life. Despite the specificity of its name, CA represents a generic approach to the study of social interaction. Much of the media text research in this field focuses on relevant aspects of broadcast news interviews (Greatbach, 1986; Heritage, 1985), talk radio (Hutchby, 1991), and talk shows (Gruber, 1991; Kotthoff, 1997). CA describes the formal structure of conversations (openings, turn takings, closings, topic control, interruptions, etc.) and analyzes how they operate under the institutional constraints of media. The strength of CA is based in detailed linguistic description, focusing on the organization of interaction, without considering the context. Context is defined within the text, dependent on the explicit mentioning of relevant factors by the speakers (see Schegloff, 1998).

In recent approaches to media texts mentioned above, however, the “text” as such has been somewhat “decentralized,” and the focus of interest has shifted to the (social, cultural, political) context and to the “localization” of meaning. A similar change of paradigm in approaches to texts has been occurring in linguistics. Media texts are also frequently being used as data corpora in linguistic analysis. Garrett and Bell (1998, p. 6) point out that more than 40% of the papers published in the leading journal Discourse & Society are based on media texts. In this chapter, we argue that the agendas in both disciplines are obviously converging and that interdisciplinary approaches to media texts can offer deeper insights.

The Concept of the Text

The present trend in approaches to media texts can be characterized by turning away from “text-internal readings, where readers are theorized as decoders of fixed meanings, to more dynamic models, where meanings are negotiated by actively participating readers” (Meinhof, 1994, p. 212). It would be beyond the scope of this contribution to discuss the different strands that have led to a more dynamic view of the text. But we would like to emphasize that some of the works that have influenced the change of paradigms in media studies have been equally influential in critical linguistic approaches, such as aspects of the work of the Bakhtin Circle by the early 20th-century Russian semioticians, Halliday’s (1978) work on social semiotics and pragmatics, the Foucauldian notion of discourse, and argumentation theories. Van Dijk’s sociocognitive approach has also had a considerable impact (see below).

All these approaches endorse an interactive model of communication, which is far more complex than the traditional models in mass communication. Media texts are perceived as dialogic, and the readings depend on the receivers and on the settings. Researchers presume, therefore, that readers/listeners or viewers interact with media (not only by writing letters to the editor but also by interpreting and understanding them in specific subjective ways). Media texts also depend on intertextual relations with many other genres, diachronically or synchronically. Texts relate to other texts, represented by the media, through quotes or indirect references, thus already adding particular meanings or decontextualizing and recontextualizing meanings. Media thus produce and reproduce social meanings.

Barthes (1966/1994), in his essay “Introduction to the Structural Analysis of Narrative,” differentiates between the work and the text. Work refers to the artifact, to the fixed pattern of signifiers on pages, whereas text refers to the process of meaning making, of reading. Fiske (1987/1989) takes up Barthes’s differentiation to distinguish between a program (on television) and a text: “Programmes are produced, distributed, and defined by the industry: texts are the product of their readers. So a programme becomes a text at the moment of reading, that is, when its interaction with one of its many audiences activates some of the meanings/pleasures that it is capable of provoking” (p. 14).

Linguistic and Nonlinguistic Methods of Textual Analysis

Titscher et al. (2000) provide an overview of current methods of text analysis that cover a broad and diverse range of methods such as grounded theory, ethnographic approaches, psychoanalytically oriented methods, qualitative heuristic text analysis, narrative semiotics, CA, and critical discourse analysis (CDA). On the basis of the definition of text provided by de Beaugrande and Dressler (1981), the authors suggest that the two dimensions of coherence (the semantic dimension, which is constitutive for the construction of meaning) and cohesion (the syntactic dimension) are constitutive of the text. The main difference between linguistic and nonlinguistic analysis is that nonlinguistic methods focus mainly on the semantic dimension of coherence, whereas linguistic methods are based on a systematic analysis of both dimensions. The aim is to make the interconnection between the cohesion and coherence dimensions apparent (Titscher et al., 2000, pp. 49ff).

Linguistic and sociolinguistic analysis pays attention to the linguistic detail, to the form and “texture” of the text (Fairclough, 1995, p. 21), aiming at illuminating socio-cultural contexts. Garrett and Bell (1998) and Fairclough (1995) provide an overview of different text and discourse-analytical approaches and their application in media studies. Approaches situated within critical linguistics (CL) emphasize the importance of the context, the social and historical situativity of the text, and the intertextual/interdiscursive dimension. Thus, the claim is not to unveil “hidden meanings,” as this would imply a static, reified conception of the text, but to identify and analyze discursive strategies, argumentation schemes (topoi), and means of realization (in verbal as well as in other semiotic modes), as put forward by the discourse-historical approach (see below). Bell (1984), for example, while examining the microlinguistic level developed in audience design, considers consonant groups in word endings. Fowler (1991) applies some tools of functional linguistics (transitivity, use of passives, nominalizations, modality, etc.) in studying the language of news media. This means that media analysis is problem oriented and not dogmatically related to the one or other linguistic theory or methodology. What seems appropriate is a multimethod approach that combines different levels of analysis and thus different tools.

Linguistic methods are time-consuming in their detailed attention to the text, especially when it comes to audio or audiovisual texts, which necessitate accurate transcription. In approaches to media texts, mixed methods are very often employed. Examples of such mixed approaches are the work of the Glasgow Media Group (1976, 1980, 1985) on news programs or van Dijk’s work (1998) comparing news reports in different countries. Both combine content analysis with text-linguistic and discourse-analytical approaches.

As far as media are concerned, linguistic approaches have so far been focusing mainly on the moment of the text, in the sense of Fiske’s “program” or Barthes’s “work” (see above). Although there has been increasing interest in audiences in the past years, studies that link media text and reception are still scarce (e.g., Lutz & Wodak, 1987; Meinhof, 1994; Morley, 1980; Richardson, 1998). Meinhof and Smith (2000) elaborate Kristeva’s concept of intertextuality to frame the collection of papers concerned with this link.

The news genre has been the most prominent research focus so far in linguistic approaches to texts, especially in discourse analysis. The press has received comparatively more attention than television, and outside of conversation analysis, radio has been relatively neglected, except for some studies of news programs (e.g., Lutz & Wodak, 1987).


The terms critical linguistics (CL) and critical discourse analysis (CDA) are often used interchangeably. In fact, recently, the term CDA seems to have been preferred and is being used to denote the theory formerly identified as CL. The roots of CDA lie in classical rhetoric, text linguistics, and sociolinguistics, as well as in applied linguistics and pragmatics. The notions of ideology, power, hierarchy, and gender, together with sociological variables, were all seen as relevant for an interpretation or explanation of text. The subjects under investigation differ for the various departments and scholars who apply CDA. Gender issues, issues of racism, media discourses, political discourses, organizational discourses, or dimensions of identity research have become very prominent.1 Bell and Garrett (1998) and Marris and Thornham (2000) provide excellent overviews on recent media studies and their relationships to CDA.

The methodologies differ greatly in all of these studies, on account of the aims of the research and also the methodologies applied: Small qualitative case studies are to be found, as well as large data corpora, drawn from fieldwork and ethnographic research. CDA takes a particular interest in the relationship between language and power. The term CDA is used nowadays to refer more specifically to the critical linguistic approach of scholars who find the larger discursive unit of text to be the basic unit of communication. This research specifically considers more or less overt relations of social struggle and conflict in all the domains mentioned above.

Deconstructing the label of this research program means that we have to define what CDA means when employing the terms critical and discourse. Most recently, Michael Billig (2002) has clearly shown that CDA has become an established academic discipline with the same rituals and institutional practices as all other academic disciplines. Ironically, he asks whether this might mean that CDA has become “uncritical” or whether the use of acronyms such as CDA might serve the same purposes as in other traditional, noncritical disciplines—namely, to exclude outsiders and to mystify the functions and intentions of the research. We cannot answer Billig’s questions extensively in this chapter. But we do believe that he suggests some interesting and potentially very fruitful and necessary debates for CDA.

Researchers in CDA rely on a variety of grammatical approaches. The definitions of the terms discourse, critical, ideology, power, and so on are also manifold (see below; Garrett & Bell, 1998; van Dijk, 2002; Wodak, 1996a). Thus, any criticism of CDA should always specify which research or researcher they relate to because CDA as such cannot be viewed as a holistic or closed paradigm.

The Notions of Discourse, Critical, Power, and Ideology

CDA is concerned with “language as social practice” (Fairclough & Wodak, 1997) and considers the context of language use to be crucial (Benke, 2000; Wodak, 2000):

CDA sees discourse—language use in speech and writing—as a form of “social practice.” Describing discourse as social practice implies a dialectical relationship between a particular discursive event and the situation(s), institution(s) and social structure(s), which frame it: the discursive event is shaped by them, but it also shapes them. That is, discourse is socially constitutive as well as socially conditioned—it constitutes situations, objects of knowledge, and the social identities of and relationships between people and groups of people. It is constitutive both in the sense that it helps to sustain and reproduce the social status quo and in the sense that it contributes to transforming it. Since discourse is so socially consequential, it gives rise to important issues of power. Discursive practices may have major ideological effects—that is, they can help produce and reproduce unequal power relations between (for instance) social classes, women and men, and ethnic/cultural majorities and minorities through the ways in which they represent things and position people. (Fairclough & Wodak, 1997, p. 258)

Of course, the term discourse is used very differently by different researchers and also in different academic cultures. In the German and Central European context, a distinction is made between text anddiscourse, relating to the tradition in text linguistics as well as to rhetoric (for summaries, see Brünner & Graefen, 1994; Wodak, 1996a). In the English-speaking world, discourse is often used both for written and oral texts (see Schiffrin, 1994). Other researchers distinguish between different levels of abstractness: Lemke (1995) defines text as the concrete realization of abstract forms of knowledge (discourse), thus adhering to a more Foucauldian approach (see also Jäger, 2001).

In the discourse-historical approach, we elaborate and relate to the sociocognitive theory of Teun van Dijk (1985, 1993, 1998) and view discourse as a form of knowledge and memory, whereas text illustrates concrete oral utterances or written documents (Reisigl & Wodak, 2001). Critical media studies view discourse as interactive, as negotiated between producers and audience, as a process in construction. Text is the (oral, visual, or written) manifestation of this (see Garrett & Bell, 1998).

The shared perspective and program of CDA relate to the term critical, which in the work of some “critical linguists” could be traced to the influence of the Frankfurt school or Jürgen Habermas (Anthonissen, 2001; Fay, 1987, p. 203; Thompson, 1988, pp. 71ff). Nowadays, this concept is conventionally used in a broader sense, denoting, as Krings argues, the practical linking of “social and political engagement” with “a sociologically informed construction of society” (Krings, Baumgartner, & Wildly, 1973, p. 808). At the same time, in Fairclough’s (1995) words, “In human matters, interconnections and chains of cause-and-effect may be distorted out of vision. Hence ‘critique’ is essentially making visible the interconnectedness of things” (p. 747; see also Connerton, 1976/1996, pp. 11-39).

For CDA, language is not powerful on its own—it gains power by the use powerful people make of it, specifically in new public spaces or new genres provided by globalized media (Baudrillard, 2000; Fairclough, 2000a; Habermas, 2000; Hall, 2000a, 2000b). In agreement with its critical theory predecessors, CDA emphasizes the need for interdisciplinary work to gain a proper understanding of how language functions in constituting and transmitting knowledge, in organizing social institutions, or in exercising power.

Not only the notion of struggles for power and control but also the “intertextuality” and “recontextualization” of competing discourses in various public spaces and genres are closely attended to. Power is about relations of difference, particularly about the effects of differences in social structures. The constant unity of language and other social matters ensures that language is entwined in social power in a number of ways: Language indexes power, expresses power, and is involved where there is contention over and a challenge to power. Power does not derive from language, but language can be used to challenge power, to subvert it, to alter distributions of power in the short and the long term.

Theoretical and Methodological Approaches

Kress (1990) concentrates on what he terms the “political economy” of representational media—that is, an attempt to understand how various societies value different modes of representation and how they use these different modes of representation. (This is a different sense of the term political economy from the one Wasko deploys in Chapter 15, this volume.) A central aspect of this work is the attempt to understand the formation of the individual human being as a social individual in response to available “representational resources.” One by-product of this research interest has been Kress’s increasing involvement in overtly political issues, including the politics of culture. Moreover, he has been concerned with multimodality and semiotics. Together with Theo van Leeuwen, Kress has developed a taxonomy that allows the precise description and interpretation of visual data (Kress & van Leeuwen, 1996). This work has influenced research on the new media (see Lemke, 2001; Scollon, 1999).

The work of Fowler, Kress, Hodge, and Trew (1979) has been cited to demonstrate the early foundations of CL. Later work of Fowler (1991, 1996) shows how tools provided by standard linguistic theories (a 1965 version of Chomskyan grammar and Halliday’s [1985] theory of systemic functional grammar) could be used to uncover linguistic structures of power in texts. Not only in news discourses but also in literary criticism, Fowler illustrates that systematic grammatical devices function in establishing, manipulating, and naturalizing social hierarchies. Fowler concentrated on analyzing news discourses and in providing grammatical tools (transitivity and modality) for such an analysis.

Fairclough (1989) sets out the social theories underpinning CDA, and as in other early critical linguistic work, a variety of textual examples are analyzed to illustrate the field, its aims, and methods of analysis. Later, Chouliaraki and Fairclough (1999) explain and elaborate some advances in CDA, showing not only how the analytical framework for researching language in relation to power and ideology developed but also how CDA is useful in disclosing the discursive nature of much contemporary social and cultural change. Particularly the language of the mass media is scrutinized as a site of power and social struggle, as well as a site where language is often only apparently transparent. Media institutions often purport to be neutral, in that they provide space for public discourse, reflect states of affairs disinterestedly, and give the perceptions and arguments of the newsmakers. Fairclough shows the fallacy of such assumptions and illustrates the mediating and constructing role of the media with a variety of examples.

Van Dijk and Kintsch (1983) early on considered the relevance of discourse to the study of language processing. Their development of a cognitive model of discourse understanding in individuals gradually developed into cognitive models for explaining the construction of meaning at a societal level. Van Dijk turned specifically to media discourse, not only giving his own reflection on communication in the mass media (van Dijk, 1986) but also bringing together the theories and applications of a variety of scholars interested in the production, uses, and functions of media discourses (van Dijk, 1985). In critically analyzing various kinds of discourses that encode prejudice, van Dijk is interested in developing a theoretical model that will explain cognitive discourse processing mechanisms (Wodak & van Dijk, 2000). Most recently, van Dijk has focused on issues of racism and ideology (van Dijk, 1998) and on an elaboration of a theory of context (van Dijk, 2001). The sociocognitive model of van Dijk is based on the assumption that cognition mediates between “society” and “discourse.” Long term and short-term memories and certain mental models shape our perception and comprehension of discursive practices and also imply stereotypes and prejudices, if such mental models become rigid and overgeneralized. The methodology used is eclectic, based primarily on argumentation theory and semantic theories.

In the Vienna school of CDA, the investigation of language use in institutional settings is central (Muntigl, Weiss, & Wodak, 2000; Wodak, 1996a). A new focus on the necessity for a historical perspective is also introduced (the discourse historical approach). A second important research focus of the Vienna school of CDA is the study of racism and anti-Semitism in the media and other public spaces (see Wodak, 1996b; Wodak et al., 1990; Wodak, de Cillia, Reisigl, & Liebhart, 1999; Wodak, Menz, Mitten, & Stern, 1994; Wodak & van Dijk, 2000; and below). Thirdly, and of course related to the latter two issues, is the study of identity constructions and changes of identities at national and transnational levels.

Recognition of the contribution of all aspects of the communicative context to text meaning, as well as a growing awareness in media studies generally of the importance of nonverbal aspects of texts, has turned attention to semiotic devices in discourse other than the linguistic ones. Pioneering work on the interaction between the verbal and visual in texts and discourse, as well as on the meaning of images, has been done by Theo van Leeuwen (Kress & van Leeuwen, 1996). Particularly the theory put forward by Kress and van Leeuwen (1996) should be mentioned here, as this provides a useful framework for considering the communicative potential of visual devices in the media (see Anthonissen, 2001; Scollon, 2001). Van Leeuwen studied film and television production as well as Hallidayan linguistics. His principal publications are concerned with topics such as the intonation of disc jockeys and newsreaders, the language of television interviews and newspaper reporting, and, more recently, the semiotics of visual communication and music. His approach has increasingly led him into the field of education. Van Leeuwen (1993) distinguishes two kinds of relations between discourses and social practices:

discourse itself [as] social practice, discourse as a form of action, as something people do to or for or with each other. And there is discourse in the Foucauldian sense, discourse as a way of representing social practice(s), as a form of knowledge, as the things people say about social practice(s). (p. 193)

“Critical discourse analysis,” according to van Leeuwen, is or should be concerned with both these aspects: “with discourse as the instrument of power and control as well as with discourse as the instrument of the social construction of reality” (van Leeuwen, 1993, p. 193). Van Leeuwen (1993) developed a most influential methodological tool: the actors analysis. This taxonomy allows for the analysis of both written and oral data, related to agency in a very differentiated and validated way. The taxonomy has since then been widely applied in data analysis.

The Duisburg school of CDA (Jäger, 1993, 2001) draws on Foucault’s notion of discourse. According to Jäger (1999, p. 116), discourse is “materiality sui generis,” and discourse theory is a “materialistic cultural theory.” Jäger is also influenced by Alexej N. Leontjev’s “speech activity theory” (Sprechtätigkeitstheorie, Leontjev, 1984) and Jürgen Link’s (1988) “collective symbolism.” As institutionalized and conventionalized speech modes, discourses express societal power relations, which in turn are affected by discourses. This “overall discourse” of society, which could be visualized as a “diskursives Gewimmel” (literally: “discursive swarming”), becomes comprehensible in different discourse strands (composed of discourse fragments from the same subject) at different discourse levels (science, politics, media, etc.). Every discourse is historically embedded and has repercussions on current and future discourse. The uniformity of the hegemonic discourse makes it possible that analysis requires only a “relatively small number of discourse fragments.” Siegfried Jäger and Margret Jäger (1999) offer concrete model analyses dealing with everyday racism, the analysis of the “discourse strand of biopower” in a daily newspaper (S. Jäger), and an analysis of interwoven discourses relating to the “criticism of patriarchy in immigration discourse” (M. Jäger). The discourse of the so-called “new right” in Germany was also analyzed by M. Jäger and S. Jäger (1993), who based their research on different right-wing print media. They identified important common characteristics (e.g., specific symbols, “ethno-pluralism” [apartheid], aggressiveness, antidemocratic attitudes, etc.) as well as significant linguistic and stylistic differences dependent on the different target groups of the newspapers.

Some Research Agendas

With the debate on globalization and on European integration, there is an increasing interest in media and multilingual audiences, cross-cultural and transnational perspectives, and the global-local articulation. Most of the research in these fields focuses on structural or political dimensions and/or on audience research, but there are only a few research projects concerned with media texts. Richardson and Meinhof (1999) contributed to filling the gap with a series of comparative case studies on satellite television programs proposed by news channels addressing a global audience (News Corp’s Sky News and Germany’s, local TV channels in Germany and Britain, and the European TV experience (ARTE, Euro-News), drawing on discourse analysis, applied linguistics, and social semiotics.

Equally interested in addressing multilingual audiences and in multilingual texts is a whole range of sociolinguistic and linguistic research covering a very diverse range of media and genres that vary from multilingual aspects in advertising (Grin, 1996), code switching in popular music (e.g., Bentahlia & Davies, 2002), and different aspects of subtitling and dubbing (Gambier, 1997) to the emergence of “Hinglish” and “Spanglish”—hybrid Hindi-English and Spanish-English spoken codes, respectively—in radio and TV. Nevertheless, researchers working in linguistics and media studies point out that there is a serious lack of systematic research available on language and the media in multilingual settings (Boyd-Barrett, Nootens, & Pugh, 1996; Grin, 1996; Leitner, 1997; Robins, 1997). The new transnational configurations of media landscapes with their particular articulations between the local and the global would thus necessitate deeper insights. Concerning ethnic minorities, the media text-oriented research mainly investigates the representation of the “Other” in (mainstream) media (see below). In the field of minority media, a shift of paradigm has occurred (Busch, 1999b): In the past, questions of access to information and minority rights were a main focus, whereas present work concentrates more on constructions of (multiple) identities. Consequently, audience-centered approaches are now dominant. From the linguistic point of view, the role of media in supporting/reviving minority or less used languages has been a concern throughout. The scarce text-analytical studies of media in minority languages in Europe show that, particularly among smaller language communities, the spectrum of topics covered has considerably narrowed, leading to a focus on questions internal to the group (Busch, 1999a).

The Representation of the “Other”

The representation of the “Other,” the representation of cultural diversity, and the reproduction of racism and xenophobia through media have been key research topics in the past few decades. Such studies have traditionally used a (critical) discourse analysis and cultural studies approach. All these studies focus on the production and reproduction of stereotypes through print media and the internet, as well as through TV. Van Dijk’s sociocognitive approach focuses on the schemata through which minorities are perceived and illustrated, as well as on headlines in the press. Headlines and their syntactic and semantic configuration typically represent “others” as perpetrators and agents, as anonymous and criminal, whereas the police and victims are passivized and presented as suffering.

The 1986 “Waldheim affair” in Austria, which exposed the former United Nations (UN) general secretary as having lied about his past in the German Wehrmacht, brought latent anti-Semitic prejudices to the fore. “Jews,” “certain circles,” “Jews full of revenge,” “rich Jews,” “Socialist Jews,” and so on were accused of being part of a “world conspiracy” to attack Waldheim all around the world. The political party that launched Waldheim’s candidacy for the Austrian presidency (the Austrian People’s Party), functionalized old and new anti-Semitic stereotypes in this election campaign. Very characteristic of the Austrian variety of such discourses were subtle linguistic features, such as implications, insinuations, and facile categorizations because blatant anti-Semitic slander has been taboo in official contexts in postwar Austria. The election campaign for the right-winger Jörg Haider and the right-wing populist party (Freedom Party) in Vienna 2001 again made use of such stereotypes. This illustrates how, whenever scapegoats are needed to channel anxieties, insecurities, aggressions, or failures, racist and anti-Semitic discourses appear and are reproduced through the media.

Stuart Hall (2000b) has also been able to demonstrate that the British media are biased when writing or talking about minorities and migrants. Specifically, in recent riots and conflicts (such as in Bradford in 2001), the unemployed young people who felt and were in reality excluded from access and participation in many social domains were depicted very negatively and associated with criminality and drug abuse. More and more in Europe, immigrants, especially African men, have come to be blamed for drug-related crimes.

Immigration laws throughout the European Union have become stricter, and hence discrimination and racism became stronger, legitimizing such restrictions. Fear of unemployment was thus highlighted. These totally constructed fears and threats started immediately after the fall of the Iron Curtain in 1989-1990 (see Reisigl & Wodak, 2000, 2001). The media reporting went through three distinct phases: Firstly, a rather paternalistic, condescending tone was applied towards those countries and people “who had no democratic experience.” Secondly, a discourse of “pity” took over, as soon as the living conditions of some population groups became known. And thirdly, as soon as migrants started crossing the borders to the West, racist beliefs and attitudes became loud. Thus, it was possible to study and exemplify the genesis of racism in media reporting (Matouschek, Januschek, & Wodak, 1995).

The terrorist attacks in the United States on September 11, 2001, reinforced anti-Islamic feelings and prejudices. The representations in the media of Muslims and the Islamic religion generalized the fear of terrorism to all people who “look different.” Usama Suleiman (2001) has analyzed the reporting in the Israeli, Palestinian, and U.S. media about a number of important events since the founding of the state of Israel. He was able to show that the representation of Israelis in the Palestinian press, of Palestinians in the Israeli press, and of both conflicting parties in the American press was significantly biased because of the interests of leading political elites. One frequently had the impression that totally different events and people were being written about.

Arab reporting about Israel has become more and more laden with old anti-Semitic stereotypes since the new wars in the Middle East of 2001-2002 (see Wistrich, 2002). Conflicts in that period also led to more anti-Semitic clichés in the European press: analogies to Nazis and to concentration camps were drawn in the French and German media. European Jews—even all Jews—were made responsible for Israeli government policies.

Overall, strategies of generalization, blaming the victims, and victim-perpetrator reversal were increasingly prominent. Stories about one bad experience with “one Jew, Roma, Arab, Turk, and so on” were generalized onto the whole ethnic group. Such “prejudice stories” characterize media reporting as well as everyday racism (Essed, 1993). Disclaimers are another salient feature of such reporting: “Everybody has best Jewish, Turkish … friends, but….” These clauses always introduce massive prejudices. The “denial of racism” (van Dijk, 1988a, 1988b) is another important characteristic. Denying racist, anti-Semitic, or xenophobic attitudes while latently functionalizing them in anti-immigrant reporting is necessary in pluralistic societies that claim to be “open” and “tolerant” (see Martín-Rojo & van Dijk, 1997; Wodak & van Dijk, 2000).

ter Wal (2002) has provided an overview of research in racism and cultural diversity in the mass media for the European Union, for the European Monitoring Center on Racism, Xenophobia and Anti-Semitism ( She found that in the period researched (1995-2000), the predominant methodology used was quantitative content analysis, but in many studies, two or more approaches were incorporated. The most common combination was that of content and discourse analysis or of discourse analysis complemented by ethnographic fieldwork and semiotic analysis. Especially in the Scandinavian countries, Spain, Italy, Austria, and the Netherlands, qualitative discourse analysis was well established in the field. Another frequent approach to media texts was the cultural studies approach, which focuses on the mythical elements on which ideological significations have been built. The majority of the research was on the press, some on television, but virtually none on radio. The perspective gradually shifted from an analysis of news production and news content to a more contextualized analysis, taking into consideration the audience perspective and the possibility of negotiating identities. When the 15 member states of the European Union were compared, it could be shown that in all countries, a big difference existed between tabloids and more elite media, which confirms the difference in modes of expression of prejudice between elites and ordinary people (see Wodak & van Dijk, 2000). On the other hand, all countries employed the linguistic features mentioned above in their reporting and news items. Access to the media was also very difficult for minority-ethnic professionals.

Hate Speech and War

After the fall of the Berlin Wall and the outbreak of armed conflicts in Southeastern and Eastern Europe, media developments in the so-called countries of transition became a focus of interest. On the level of text analysis, the questions of hate speech, biased reporting, and representation of minorities have attracted research interest. Some of these works use quantitative content analysis; others combine quantitative aspects with qualitative text or discourse-analytical approaches. The scope ranges from case studies on particular media (e.g., Kuzmanic, 1999, on racism, sexism, and chauvinism in Slovenian print media; Valic, 1997, on war reporting on local radio in Serbia) to studies on the representation of particular groups (e.g., Erjavec, Hrvatin, & Kelbl, 2000, on Roma in Slovenia). Qualitative work mainly refers to the theoretical and methodological approaches developed by CDA (in particular, van Dijk, 1991; Wodak, 1996b).

During and after the war in former Yugoslavia, two major international text analysis projects were initiated and financed by nongovernmental organizations to investigate hate speech: The project “Media and War” (Skopljanac Brunner, Gredelj, Hodzic, & Kristofic, 2000) brought together a large interdisciplinary group of researchers from Croatia and Serbia, a difficult task in a period of complete communication blockade in the region. Its findings were based on a large body of data drawn from the print media—the two major dailies in the respective countries Vjesnik and Politika—and television news programs on the national (state) stations. Similar to van Dijk’s (1991) comparative study of news discourses, the core of the “Media and War” study combined content analysis and discourse analysis methods complemented with a semantic field analysis (Skiljan, 2000) on the word rat (war), as well as background information on the political situation and the role of the media during the period of disintegration of former Yugoslavia and the outbreak of war in Croatia and Bosnia-Herzegovina. The publication (Skopljanac Brunner et al., 2000) that resulted from this study refrains from drawing general conclusions and does not extrapolate the findings to draw more general conclusions about hate speech. It is left to the reader to compare the conclusions drawn by different authors employing different methods of analysis (tacit triangulation). A main focus is the discursive strategies employed in constructing new national identities in which strategies of creating in-groups and out-groups by emphasizing differences between “us” and “them” play a key role, as well as strategies of internal homogenization, such as invoking “national unity and solidarity,” and of victimizing one’s own group while accusing the other of aggression. It was striking how frequently Croatian media dwelt on locating the newly founded state on a map of the imaginary: Croatia was depicted as an integral part of Europe, and Europe, in turn, was depicted as a centuries-old Schicksalsgemeinschaft—a community formed by historical destiny—based on Christian values.

Another recent major international text analysis project in Southeastern Europe was the “Balkan neighbors project,” which involved researchers in Albania, Bulgaria, Greece, Macedonia, Romania, Turkey, and Yugoslavia (Balkan Neighbours Newsletter, 2000, Issue 10). It was also initiated and financed by nongovernmental organizations. Between 1994 and 2000, the project monitored mainstream print media. In each country, a range of several dailies and weekly political magazines with a different global perspective were selected for the monitoring process. Each affiliated research center extracted from the press texts concerning the other countries involved in the project as well as texts concerning national, ethnic, linguistic, and religious minorities within the country. Within such texts, verbal realizations of stereotypes and prejudices were located and analyzed. As the project had a strong emphasis on dissemination, findings were published at 6 monthly intervals in a substantial news bulletin (Balkan Neighbours Newsletter, 1994-2000), which was circulated to opinion leaders in the different countries. The data were also available on the internet ( Each bulletin comprised a short description of the analyzed papers, an overview of the major political events, and a contextualized compilation of the extracted stereotypes. The long monitoring period made it possible to pin down moments of transformation of particular stereotypes in relation to certain events. Transformations occurred not only when armed conflict broke out but also, for example, after a major earthquake in Turkey, when Greece provided rapid help.

Feminist Research

Feminist readings of media texts have both an academic and a political focus (see Kuhn, 2000, 62ff). On one hand, many studies have compared representations of men and women in magazines as well as in newspapers or on TV talk shows (see Eggins & Iedema, 1997; Kotthoff, 1997; Lalouschek, 2002; Wetschanow, 2003; Winship, 1986). On the other hand, feminist criticism focuses on a different deconstruction of texts, on “women’s genres” and the construction of femininity (see Kuhn, 1984). The aim of such studies is to question dichotomies and traditional distinctions, such as the public-private, the knowledge-pleasure, and the masculine-feminine splits (see also Marris & Thornham, 2000, pp. 330ff). The slogan that stems from the women’s liberation movement—“the private is political”—has become very important for these analyses, which also include soap operas, infotainment, talk shows, crime stories, thrillers, and so on.

These first relevant studies were complemented recently by an attempt to include race, class, and ethnicity into such analyses (Squire, 2000). It was agreed that feminist analysis needs to be placed in context and that discourse analysis of media about gender roles needs to be context dependent (see Kotthoff & Wodak, 1997). Much work has also followed postmodern thinking (Brunsdon, 1991) by introducing new genres, the concept of “fragmentation,” and an emphasis on deconstructing subjectivity and on the collapse of boundaries—including those of gender.

The study of Wetschanow (2003) analyzes media reporting on violence against women and on the reporting of rape cases in Austrian print media as well as TV. Although she concentrates on one central European country, the results could quite easily be generalized. By combining quantitative content analysis with qualitative critical discourse analysis of media texts, she illustrates convincingly how strategies of categorization are significantly different for the victims and perpetrators, as well as for men and women. The strategy of victim-perpetrator reversal is applied frequently, and the men who are accused of raping a woman are switched into the role of passive victims who were seduced and could not defend themselves against their sexual drives. Women are depicted as seductresses, as initiators, and thus possibly guilty of the harm done to them. The adjectives employed as attributes mark these differences and characteristics.

The same pattern holds true for violence. Patricia O’Connor (2002) has investigated and also worked with victims of violence (men and women), as well as with perpetrators in prisons. O’Connor (2002) and McElhinny (1997) were able to demonstrate empirically, through their discourse analysis of interactions and media texts, that officials, police officers, and bureaucrats were always represented benevolently, whereas the victims, often enough women, were blamed for being weak, not remembering accurately, or being noncompliant. They are thus doubly harassed: first in the terrible situation and context of violence and, secondly, through the representation in the media and at court or through other bureaucracies that define them.

Eggins and Iedema (1997) studied not only the language of women’s magazines in Australia but also the visual images of women by employing important features of Hallidayan functional systemic grammar and visual grammar, developed by Kress and van Leeuwen (1996) (see above). Eggins and Iedema compared two Australian magazines, New Woman and She. Although both magazines express similar topical dimensions (orientation to appearance, to hetero-sexuality, to women and men in isolation [without other variables], etc.), they address different audiences: New Woman calls for women’s empowerment through individual change and thus neutralizes the possibility of real emancipation as a political process. She, in contrast, provides simple dichotomic answers to and evaluations of complex problems and constructs rigid boundaries between “women and men,” between the “good and bad,” the “beautiful and the ugly,” and so on. The authors conclude that the magazines offer “difference without diversity” and thus—they claim—stabilize the status quo.


In our ever more globalizing world, media have gained more power. The impact of media on political developments and decision making still has to be fully explored. Moreover, the influence of media on the production and reproduction of beliefs, opinions, stereotypes, prejudices, and ideologies also has to be thoroughly investigated and compared throughout different countries worldwide. Qualitative in-depth studies on audiences, reception, and perceptions of readers, viewers, or listeners are also missing.

The cultural influence of the U.S. media on other media (e.g., in Europe) has slowly started to be perceived. This influence is apparent in the construction of new genres, new public spaces, new modes of advertisement, and so on. The impact of transnational media (such as CNN or ARTE) on identity construction has yet to be investigated.

Access to media is another relevant factor. Who are the decision makers; who are the journalists, producers, and investigators; and who is represented and how? And who watches, listens, and reads what? The problem of media literacy and of the com-prehensibility of media poses big questions for participation in democratic societies.

Lastly, research is needed on new genres, which Lemke (2001) labels as “traversals,” as encompassing time and space, such as the internet or channel surfing. Time-space distanciation and time-space compression (Giddens, 2000) have to be considered in their impact on our access to information.

We do not believe that we have enumerated all the relevant phenomena, which require more interdisciplinary and transdisci-plinary research. We hope to have made clear that media research from a critical discourse-analytical point of view, in combination with many other theoretical approaches, might provide some answers to all these important questions.