Methods of Critical Discourse Analysis

INTRODUCING QUALITATIVE METHODS provides a series of volumes which introduce qualitative research to the
student and beginning researcher. The approach is interdisciplinary and international. A distinctive feature of these
volumes is the helpful student exercises.
One stream of the series provides texts on the key methodologies used in qualitative research. The other stream contains
books on qualitative research for different disciplines or occupations. Both streams cover the basic literature in a clear
and accessible style, but also cover the cutting edge issues in the area.

SERIES EDITOR
David Silverman (Goldsmiths College)
EDITORIAL BOARD
Michael Bloor (Cardiff University and University of Glasgow)
Barbara Czarniawska (Gteburg University)
Norman Denzin (University of Illinois, Urbana-Champaign)
Uwe Flick (Alice Salomon University of Applied Sciences, Berlin)
Barry Glassner (University of Southern California)
Jaber Gubrium (University of Missouri)
Anne Murcott (University of Nottingham)
Jonathan Potter (Loughborough University)
For a list of all titles in the series, visit www.sagepub.co.uk/iqm
Methods of Critical Discourse Analysis

Second Edition

Edited by
Ruth Wodak and Michael Meyer
Editorial arrangement and Chatper 1 Ruth Wodak and Michael Meyer 2009
Chatper 2 Siegfried Jger and Florentine Maier 2009
Chatper 3 Teun A. van Dijik 2009
Chatper 4 Martin Reisigl and Ruth Wodak 2009
Chatper 5 Gerlinde Mautner 2009
Chatper 6 Theo van Leeuwen 2009
Chatper 7 Norman Fairclough 2009

Thsis second edition first published 2009

First edition published 2001
Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the
Copyright, Designs and Patents Act, 1988, this publication may be reproduced, stored or transmitted in any form, or by
any means, only with the prior permission in writing of the publishers, or in the case of reprographic reproduction, in
accordance with the terms of licences issued by the Copyright Licensing Agency. Enquiries concerning reproduction
outside those terms should be sent to the publishers.

SAGE Publications Ltd

1 Olivers Yard
55 City Road
London EC1Y 1SP
SAGE Publications Inc.
2455 Teller Road
Thousand Oaks, California 91320
SAGE Publications India Pvt Ltd
B 1/I 1 Mohan Cooperative Industrial Area
Mathura Road
New Delhi 110 044
SAGE Publications Asia-Pacific Pte Ltd
33 Pekin Street #02-01
Far East Square
Singapore 048763
Library of Congress Control Number: 2008932562
British Library Cataloguing in Publication data

A catalogue record for this book is available from the British Library

ISBN 978-1-84787-454-2
ISBN 978-1-84787-455-9 (pbk)
Typeset by C&M Digitals (P) Ltd, Chennai, India
Printed in Great Britain by CPI Antony Rowe, Chippenham, Wiltshire
Printed on paper from sustainable resources
Contents

Acknowledgements
Notes on contributors

1 Critical discourse analysis: history, agenda, theory and methodology

Ruth Wodak and Michael Meyer

2 Theoretical and methodological aspects of Foucauldian critical discourse analysis and

dispositive analysis
Siegfried Jger and Florentine Maier

3 Critical discourse studies: a sociocognitive approach

Teun A. van Dijk

4 The discourse-historical approach (DHA)

Martin Reisigl and Ruth Wodak

5 Checks and balances: how corpus linguistics can contribute to CDA

Gerlinde Mautner

6 Discourse as the recontextualization of social practice: a guide

Theo van Leeuwen

7 A dialecticalrelational approach to critical discourse analysis in social research

Norman Fairclough

References

Index
Acknowledgements

This book could not have been written without the help of many people. We wish to
express our gratitude to them all.
Our endeavour to provide readers with a comprehensive book introducing the different
approaches to Critical Discourse Analysis started in the form of a seminar on CDA at the
Department of Linguistics (University of Vienna) in 1999. We appreciate the discussion
and work with the student participants which provided an important stimulus for our plan
to publish the first edition.
We appreciate the collaboration with the authors of the contributions to this book:
Norman Fairclough, Siegfried Jger, Florentine Maier, Gerlinde Mautner, Martin Reisigl,
Teun van Dijk and Theo van Leeuwen. Not only did they write instructive and interesting
chapters, they submitted them on schedule, and were very open to our criticism, comments
and suggestions.
Bryan Jenner has been indispensable in helping to recast some of our more
idiosyncratic pieces of non-native English into an acceptable form. Anahid
Aghamanoukjan, Ulla Ernst and Edith Fazekas contributed to fine-tuning the manuscript in
an essential way. Claire Lipscomb at Sage has been an extremely supportive editor who
has made a very positive contribution to the final results. Last but not least we thank the
six anonymous reviewers who helped us to avoid a number of shortcomings in the first
edition but who are not responsible for those in the second.
Notes on contributors

Norman Fairclough is Professor Emeritus of Language in social Life at Lancaster

University in the UK. He has written extensively in the field of Critical Discourse
Analysis. His main publications include: Language and Power (1989), Discourse and
Social Change (1992), Media Discourse (1995), Critical Discourse Analysis (1995),
Discourse in Late Modernity (1999) (with Lilie Chaouliaraki) and New Labour, New
Language? (2000). His main current interest is in language (discourse) as an element in
the contemporary social changes that are referred to as globalization, neoliberalism,
new capitalism, the knowledge economy and so forth. Over the past three years he has
been working specifically on aspects of transition in Central and Eastern Europe,
especially Romania, from a discourse analytical perspective (see
www.ling.lancs.ac.uk/profiles/263). Email: [email protected]

Siegfried Jger has been a Professor for German Language at the University of
Duisburg/Essen (Germany) since 1972, and since 1987 he has also been head of the
Dusiburg Institute of Linguistic and Social Research. His major research areas are
Discourse Theory and Critical Discourse Analysis based on Michel Foucaults theory.
He has conducted several projects dealing with a broad range of topics, e.g. language
barriers, right-wing extremism, immigration and racism in mass media and everyday talk,
media coverage of crime, Jewish publications in the 19th century, Christian
fundamentalism and antiSemitism in the German and Polish mass media. For his
publications, see diss-duisburg.de. Email: [email protected]

Florentine Maier is a research assistant at the Academic Unit for Nonprofit Management
at the Wirtschaftsuniversitt Wien (WU, Vienna University of Economics and Business).
Her primary fields of research are the spread of managerialism into Nonprofit
Organizations, and the employment of older workers. Before entering the Academic Unit
for Nonprofit Management, she was a Research Assistant at the Institute of Human
Resource Management at the WU. She has also worked as a lecturer in Human Resource
Management at the University of Applied Science at the Vienna Chambers of Commerce,
and as a teacher of Chinese. She gained her doctoral degree in social and economic
sciences from the WU in 2008. Email: [email protected]

Gerlinde Mautner is Professor of English Business Communication at the

Wirtschaftsuniversitt Wien (Vienna University of Economics and Business). A graduate
of the University of Vienna, she has spent a year at each of the linguistics departments of
the universities of Birmingham (UK), Lancaster, and, most recently, Cardiff. Combining
critical discourse analysis with corpus linguistic methodology, her research focuses on
corporate and marketing communications, as well as on the discourse of the public and
nonprofit sectors. Email: [email protected]

Michael Meyer is Professor for Business Administration at the Unit for Nonprofit
Management at the Wirtschaftsuniversitt Wien (WU, Vienna University of Economics
and Business). Since 1989, Michael Meyer has held a range of academic positions, such
as Assistant Professor for Marketing and Cultural Management, and Social Researcher in
various interdisciplinary projects. Currently, he is head of the Research Institute for
Nonprofit Organizations at the WU. He is an experienced lecturer and researcher in the
field of Organizational Behaviour. His areas of specialization and current research
include: Managerialism, Nonprofit Organizations, Civil Society and Civic Participation,
Social Systems Theory, Organizational Theory, and Careers (for his publications, see:
www.wu-wien.ac.at/npo/english/team/mm). Email: [email protected]

Martin Reisigl has a PhD in Applied Linguistics and currently teaches Applied
Linguistics at the University of Vienna. He is also working on his habilitation project
supported by research fellowships of the German Alexander von Humboldt Foundation
and the Austrian Academy of Sciences (APART). From May 2006 until January 2007, he
was a visiting professor at the university La Sapienza in Rome, Italy. From February until
June 2007, he was a visiting fellow at the Institute for Human Sciences/Institut fr die
Wissenschaften von Menschen (IWM) in Vienna. His research interests include (critical)
discourse analysis and discourse theory, text linguistics, academic writing,
sociolinguistics, (political) rhetoric (language and discrimination, nationalism, racism as
well as populism), language and history, linguistics and literature, argumentation analysis
and semiotics. Email: [email protected]

Teun A. van Dijk was Professor of Discourse Studies at the University of Amsterdam
until 2004, and is at present Professor at the Universitat Pompeu Fabra, Barcelona. After
earlier work on generative poetics, text grammar and the psychology of text processing,
his work since 1980 has taken a more critical perspective and has dealt with discursive
racism, news in the press, ideology, knowledge and context. He is the author of several
books in most of these areas, and he has edited The Handbook of Discourse Analysis (4
vols, 1985), the introductory book Discourse Studies (2 vols, 1997) as well as the reader
The Study of Discourse (5 vols, 2007). He has founded six international journals:
Poetics, Text (now Text and Talk ) , Discourse and Society, Discourse Studies,
Discourse and Communication and the internet journal Discurso and Sociedad in
Spanish (www.dissoc.org), the last four of which he still edits. His most recent
monographs in English are Ideology (1998), Racism and Discourse in Spain and Latin
America (2005), Discourse and Context (2008), Society and Discourse (2008), and his
last edited books (with Ruth Wodak), Racism at the Top (2000), and Discourse and
Racism in Latin America (2008). Teun van Dijk, who holds two honorary doctorates, has
lectured widely in many countries, especially also in Latin America. For further
information, see his website: www.discourses.org. Email: [email protected]

Theo van Leeuwen is Professor of Media and Communication and Dean of the Faculty
of Arts and Social Sciences at the University of Technology, Sydney. He has published
widely in the areas of critical discourse analysis, multimodality and visual semiotics. His
books include: Reading Images The Grammar of Visual Design (with Gunther Kress);
Speech, Music, Sound: Multimodal Discourse The Modes and Media of
Contemporary Communication (with Gunther Kress); and Introducing Social Semiotics
and Global Media Discourse (with David Machin). His latest book Discourse and
Practice was published in 2008. He is a founding editor of the journal Visual
Communication. Email: [email protected]

Ruth Wodak is Distinguished Professor of Discourse Studies at Lancaster University. In

addition to various other prizes, she was awarded the Wittgenstein Prize for Elite
Researchers in 1996. Her research interests focus on discourse studies, gender studies,
language and/in politics, prejudice and discrimination, and on ethnographic methods of
linguistic field work. She has held visiting professorships in Uppsala, Stanford
University, University of Minnesota, University of East Anglia and Georgetown
University. In 2008, she was awarded the Kerstin Hesselgren Chair of the Swedish
Parliament (at the University of rebro). Recent book publications include Ist sterreich
ein deutsches Land? (with R. de Cillia, 2006); Qualitative Discourse Analysis in the
Social Sciences (with M. Krzy anowski, 2008); Migration, Identity and Belonging (with
G. Delanty and P. Jones, 2008), The Discursive Construction of History: Remembering
the Wehrmachts War of Annihilation (with H. Heer, W. Manoschek and A. Pollak,
2008) and The Construction of Politics in Action: Politics as Usual? (2009). See
www.ling.lancs.ac.uk/profiles/265 for more information on her ongoing research projects
and recent publications. Email: [email protected]
1
Critical Discourse Analysis: History, Agenda, Theory and Methodology1

Ruth Wodak and Michael Meyer

CDA What is it all about?
A brief history of the CDA Group
The common ground: discourse, critique, power and ideology
Main research agenda and challenges
Differences and similarities beyond the social dimension
Methodology
Summary and criticism

CDA What is it all about?

In this chapter, we first provide a brief story how it all began; then we present an
overview of some important research agendas in CDA and discuss new challenges for
CDA research. Secondly, we discuss the various theoretical and methodological
approaches assembled in this volume from a sociological and epistemological
perspective.2 There, we focus mostly on three central and constitutive concepts: power,
ideology and critique. We also, of course, summarize some of the salient principles
which are constitutive of all approaches in CDA. In addition, we mention some important
criticism which CDA has been confronted with in the past years (see Billig, 2003, 2008;
Chilton, 2007; Chilton and Wodak, 2007; Wodak and de Cillia, 2006 for an extensive
discussion of these issues).
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. Therefore, we will continue to
use CDA exclusively here (see Anthonissen, 2001; Chilton and Wodak, 2007 for an
extensive discussion of these terms and their history). The manifold roots of CDA lie in
Rhetoric, Text linguistics, Anthropology, Philosophy, Socio-Psychology, Cognitive
Science, Literary Studies and Sociolinguistics, as well as in Applied Linguistics and
Pragmatics.
Nowadays, some scholars prefer the term Critical Discourse Studies (CDS). For
example, Teun van Dijk provides us with a broad overview of the field of (C)DS, where
one can identify the following developments: between the mid-1960s and the early 1970s,
new, closely related disciplines emerged in the humanities and the social sciences.
Despite their different disciplinary backgrounds and a great diversity of methods and
objects of investigation, some parts of the new fields/paradigms/linguistic sub-disciplines
of semiotics, pragmatics, psycho-and sociolinguistics, ethnography of speaking,
conversation analysis and discourse studies all deal with discourse and have at least
seven dimensions in common (see Van Dijk, 2007a; Wodak, 2008a):

an interest in the properties of naturally occurring language use by real language

users (instead of a study of abstract language systems and invented examples)
a focus on larger units than isolated words and sentences and, hence, new basic
units of analysis: texts, discourses, conversations, speech acts, or communicative
events
the extension of linguistics beyond sentence grammar towards a study of action and
interaction
the extension to non-verbal (semiotic, multimodal, visual) aspects of interaction
and communication: gestures, images, film, the internet, and multimedia
a focus on dynamic (socio)-cognitive or interactional moves and strategies
the study of the functions of (social, cultural, situative and cognitive) contexts of
language use
an analysis of a vast number of phenomena of text grammar and language use:
coherence, anaphora, topics, macrostructures, speech acts, interactions, turn-taking,
signs, politeness, argumentation, rhetoric, mental models, and many other aspects of
text and discourse.

The significant difference between DS and CDS (or CDA) lies in the constitutive
problem-oriented, interdisciplinary approach of the latter, apart from endorsing all of
the above points. CDA is therefore not interested in investigating a linguistic unit per
se but in studying social phenomena which are necessarily complex and thus require a
multidisciplinary and multimethodical approach. The objects under investigation do not
have to be related to negative or exceptionally serious social or political experiences or
events this is a frequent misunderstanding of the aims and goals of CDA and of the term
critical which, of course, does not mean negative as in commonsense usage (see
below). Any social phenomenon lends itself to critical investigation, to be challenged and
not taken for granted. We will return to this important point and other common
misunderstandings of CDA below.
We would also like to emphasize right at the beginning of this volume that it is obvious
that the notions of text and discourse have been subject to a hugely proliferating number
of usages in the social sciences. Almost no paper or article is to be found which does not
revisit these notions, quoting Michel Foucault, Jrgen Habermas, Chantal Mouffe, Ernesto
Laclau, Niklas Luhmann, or many others. Thus, discourse means anything from a
historical monument, a lieu de mmoire, a policy, a political strategy, narratives in a
restricted or broad sense of the term, text, talk, a speech, topic-related conversations, to
language per se. We find notions such as racist discourse, gendered discourse, discourses
on un/employment, media discourse, populist discourse, discourses of the past, and many
more thus stretching the meaning of discourse from a genre to a register or style, from a
building to a political programme. This causes and must cause confusion which leads to
much criticism and more misunderstandings (Blommaert, 2005; Reisigl, 2007; Wodak,
2008a; Wodak and de Cillia, 2006). This is why each contributor to this volume was
asked to define their use of the term integrated in their specific approach.

A brief history of the CDA Group

CDA as a network of scholars emerged in the early 1990s, following a small symposium
in Amsterdam, in January 1991. Through the support of the University of Amsterdam,
Teun van Dijk, Norman Fairclough, Gunther Kress, Theo van Leeuwen and Ruth Wodak
spent two days together, and had the wonderful opportunity to discuss theories and
methods of Discourse Analysis, specifically CDA. The meeting made it possible to
confront with each other the very distinct and different approaches, which have, of
course, changed significantly since 1991 but remain relevant, in many respects. In this
process of group formation, differences and sameness were laid out: differences with
regard to other theories and methodologies in Discourse Analysis (see Renkema, 2004;
Titscher et al., 2000; Wetherell et al., 2001; Wodak, 2008a) and sameness in a
programmatic way, both of which frame the range of theoretical approaches. In the
meantime, for example, some of the scholars previously aligned with CDA have chosen
other theoretical frameworks and have distanced themselves from CDA (such as Gunther
Kress and Ron Scollon); on the other hand, new approaches have been created which
frequently find innovative ways of integrating the more traditional theories or of
elaborating them (see below).
In general, CDA as a school or paradigm is characterized by a number of principles:
for example, all approaches are problem-oriented, and thus necessarily interdisciplinary
and eclectic (see below). Moreover, CDA is characterized by the common interests in
demystifying ideologies and power through the systematic and retroductable
investigation of semiotic data (written, spoken or visual). CDA researchers also attempt
to make their own positions and interests explicit while retaining their respective
scientific methodologies and while remaining self-reflective of their own research
process.
The start of the CDA network was marked by the launch of Van Dijks journal
Discourse and Society (1990), as well as by several books which were coincidentally
(or because of a Zeitgeist) published simultaneously and led by similar research goals.3
The Amsterdam meeting determined an institutional start, an attempt both to constitute an
exchange programme (ERASMUS for three years)4, as well as joint projects and
collaborations between scholars of different countries, and a special issue of Discourse
and Society (1993), which presented the above-mentioned approaches. Since then, new
journals have been created, multiple overviews have been written, and nowadays CDA is
an established paradigm in linguistics; currently, we encounter Critical Discourse
Studies, The Journal of Language and Politics, Discourse and Communication and
Visual Semiotics, among many other journals; we also find several e-journals which
publish critical research, such as CADAAD. Book series have been launched (such as
Discourse Approaches to Politics, Culture and Society) , regular CDA meetings and
conferences take place, and handbooks are under way. In sum, CDA (CDS) has become
an established discipline, institutionalized across the globe in many departments and
curricula.

The common ground: discourse, critique, power and ideology

When deconstructing the label of this research programme we view CDA basically as a
research programme, the reasons for which we will explain below we necessarily have
to define what CDA means when it employs the terms critical and discourse. Michael
Billig (2003) has clearly pointed to the fact that CDA has become an established
academic discipline with the same rituals and institutional practices as all other academic
disciplines. Ironically, he asks the question whether this might mean that CDA has
become or might become uncritical or if the use of acronyms such as CDA might serve
the same purposes as in other traditional, non-critical disciplines; namely to exclude
outsiders and to mystify the functions and intentions of the research. Most recently, Billig
has reiterated this question under a new frame: do scholars who employ CDA write in the
same way mainly by using nominalizations extensively, like the many texts which they
criticize (Billig, 2008)?

The problem with talking about the unconscious, repression, mental representations, mirror-stages, etc., is that it is
easy to assume that we have solved problems by discovering things. And the more we write about these
things, the more we take their existence for granted. Analysts might have once understood these concepts semi-
metaphorically, but soon they write about them literally. In my view, the cognitive psychology of mental
representations, or the psychoanalysis of the unconscious and repression, makes psychology too easy and too
non-materialistic too prone to accept that non-material entities provide the solution to the puzzles that, in effect,
analysts are avoiding. And that is why I advocate that we should be examining nominalizing (not nominalization),
representing (not representations), repressing (not repression) and so on.

Billig (2008) ends his quite provocative argument by stating:

[t]here is no reason for supposing that for academics, writing their academic articles, the active forms are
psychologically primary. In my article and in this reply, I have struggled to resist the grammatical forms with
which my fingers are so familiar. I have redrafted, often with a struggle, many sentences which spontaneously
 spilled out in the passive form. I have probably used the first person singular here more times than I have done in
all the rest of my publications put together. And so now, I do not want to end by promoting a new label. To adapt
a very famous phrase, the point is not to categorize language, but to change it.

We cannot follow Billigs recommendations in detail or answer his questions in this

chapter extensively. However, we believe that he points to potentially very fruitful and
necessary debates for CDA. More specifically, Billig points to the danger which can
befall any discipline (or school or group) when it becomes established and
institutionalized after having been marginalized for a long time. Once established, he
argues, one might forget the basic desiderata and become corrupted by the system in our
case, the Knowledge-based Economy (KBE) which influences all our lives in so many
ways (Billig, 2008).
At this point, it is important to stress that CDA has never been and has never attempted
to be or to provide one single or specific theory. Neither is one specific methodology
characteristic of research in CDA. Quite the contrary, studies in CDA are multifarious,
derived from quite different theoretical backgrounds, oriented towards different data and
methodologies. Researchers in CDA also rely on a variety of grammatical approaches.
The definitions of the terms discourse, critical, ideology, power and so on are also
manifold. Thus, any criticism of CDA should always specify which research or
researcher they relate to. Hence, we suggest using the notion of a school for CDA, or of
a programme, which many researchers find useful and to which they can relate. This
programme or set of principles has, of course, changed over the years (see (Fairclough
and Wodak, 1997; Wodak, 1996).
Such a heterogeneous school might be confusing for some; on the other hand, it allows
for continuous debates, for changes in the aims and goals, and for innovation. In contrast
to total and closed theories, such as Chomskys Generative Transformational Grammar
or Michael Hallidays Systemic Functional Linguistics, CDA has never had the image of
a sect and does not want to have such an image. This heterogeneity of methodological
and theoretical approaches that can be found in this field would tend to confirm van
Dijks point that CDA and CL are at most a shared perspective on doing linguistic,
semiotic or discourse analysis (Van Dijk, 1993a: 131). Below, we summarize some of
these principles, which are adhered to by most researchers.

The notion of discourse

CDA sees language as social practice (Fairclough and Wodak, 1997), and considers the
context of language use to be crucial. We quote one definition which has become very
popular among CDA researchers:

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 and Wodak, 1997: 258)

Within this definition, the term discourse is of course used very differently by different
researchers and also in different academic cultures (Wodak, 2006a, b). In the German and
Central European context, a distinction is made between text and discourse, relating to
the tradition in text linguistics as well as to rhetoric (see Brnner and Graefen, 1994;
Vass, 1992; Wodak and Koller, 2008 for summaries). In the English speaking world,
discourse is often used both for written and oral texts (see Gee, 2004; 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 Jger and Maier in this volume).
The discourse-historical approach relates to the sociocognitive theory of Teun van
Dijk (1998) and views discourse as structured forms of knowledge, whereas text
refers to concrete oral utterances or written documents (see Reisigl and Wodak in this
volume).

The critical impetus

The shared perspective and programme of CDA emphasize the term critical, which in
the work of some critical linguists can be traced to the influence of the Frankfurt School
and Jrgen Habermas (Anthonissen, 2001; Fay, 1987: 203; Thompson, 1988: 71ff):
Critical Theory in the sense of the Frankfurt School, mainly based on the famous essay
of Max Horkheimer in 1937, indicates that social theory should be oriented towards
critiquing and changing society, in contrast to traditional theory oriented solely to
understanding or explaining it. The core concepts of such an understanding of critical
theory are:

Critical theory should be directed at the totality of society in its historical

specificity.
Critical theory should improve the understanding of society by integrating all the
major social sciences, including economics, sociology, history, political science,
anthropology and psychology.

What is rarely reflected in this understanding of critique is the analysts position itself.
The social embeddedness of research and science, the fact that the research system itself
and thus CDA are also dependent on social structures, and that criticism can by no means
draw on an outside position but is itself well integrated within social fields, has been
emphasized by Pierre Bourdieu (1984). Researchers, scientists and philosophers are not
outside the societal hierarchy of power and status but are subject to this structure. They
have also frequently occupied and still occupy rather superior positions in society.
In language studies, the term critical was first used to characterize an approach that
was called Critical Linguistics (Fowler et al., 1979; Kress and Hodge, 1979). Among
other ideas, those scholars held that the use of language could lead to a mystification of
social events which systematic analysis could elucidate. For example, a missing by-
phrase in English passive constructions might be seen as an ideological means for
concealing or mystifying reference to an agent (Chilton, 2008).
Nowadays, this concept of critique 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 et al., 1973; Titscher et al.,
2000: 88). Hence, critique is essentially making visible the interconnectedness of things
(Fairclough, 1995a: 747; see also Connerton, 1976: 1139). The reference to the
contribution of critical theory to the understanding of CDA and the notions of critical
and ideology are of particular importance (see Anthonissen, 2001, for an extensive
discussion of this issue).
Critical theories, thus also CDA, want to produce and convey critical knowledge that
enables human beings to emancipate themselves from forms of domination through self-
reflection. Thus, they are aimed at producing enlightenment and emancipation. Such
theories seek not only to describe and explain, but also to root out a particular kind of
delusion. Even with differing concepts of ideology, critical theory seeks to create
awareness in agents of their own needs and interests. This was, of course, also taken up
by Pierre Bourdieus concepts of violnce symbolique and mconnaissance
(Bourdieu, 1989).
In agreement with its critical theory predecessors, CDA emphasizes the need for
interdisciplinary work in order to gain a proper understanding of how language functions
in constituting and transmitting knowledge, in organizing social institutions or in
exercising power (see Graham, 2002; Lemke, 2002; Martin and Wodak, 2003). In any
case, CDA researchers have to be aware that their own work is driven by social,
economic and political motives like any other academic work and that they are not in any
privileged position. Naming oneself critical only implies specific ethical standards: an
intention to make their position, research interests and values explicit and their criteria as
transparent as possible, without feeling the need to apologize for the critical stance of
their work (Van Leeuwen, 2006: 293).

Ideology and power a kaleidoscopic view

The critical impetus of CDA and other critical research programmes is certainly the
legacy of enlightenment (Horkheimer and Adorno, 1969/1991). Critique regularly aims at
revealing structures of power and unmasking ideologies. Ideology is then not understood
in a positivistic way, i.e. ideologies cannot be subjected to a process of falsification. Nor
is it the Marxian type of ideology according to the economic base/superstructure
dichotomy that is of interest for CDA.
Political scientists name four central characteristics of ideologies:

1. Power is more important than cognitions.

2. They are capable of guiding individuals evaluations.
3. They provide guidance through action.
4. They must be logically coherent. (Mullins, 1972)

Although the core definition of ideology as a coherent and relatively stable set of beliefs
or values has remained the same in political science over time, the connotations
associated with this concept have undergone many transformations. During the era of
fascism, communism and the Cold War, totalitarian ideology was confronted with
democracy, the evil with the good. If we speak of the ideology of the new capitalism
(see Van Dijk and Fairclough in this volume), ideology once again has a bad
connotation. Obviously, it is not easy to capture ideology as a belief system and
simultaneously to free the concept from negative connotations (Knight, 2006: 625).
It is, however, not that type of ideology on the surface of culture that interests CDA, it
is rather the more hidden and latent type of everyday beliefs, which often appear
disguised as conceptual metaphors and analogies, thus attracting linguists attention: life
is a journey, social organizations are plants, love is war, and so on (Lakoff, 1987; Lakoff
and Johnson, 1980, 1999). In daily discussion, certain ideas arise more commonly than
others. Frequently, people with diverse backgrounds and interests may find themselves
thinking alike in startling ways. Dominant ideologies appear as neutral, holding on to
assumptions that stay largely unchallenged. Organizations that strive for power will try to
influence the ideology of a society to become closer to what they want it to be. When
most people in a society think alike about certain matters, or even forget that there are
alternatives to the status quo, we arrive at the Gramscian concept of hegemony. With
regard to this key concept of ideology, Van Dijk (1998) sees ideologies as the
worldviews that constitute social cognition: schematically organized complexes of
representations and attitudes with regard to certain aspects of the social world, e.g. the
schema [] whites have about blacks (Van Dijk, 1993b: 258).
Furthermore, it is the functioning of ideologies in everyday life that intrigues CDA
researchers. Fairclough has a more Marxist view of ideologies and conceives them as
constructions of practices from particular perspectives:

Ideologies are representations of aspects of the world which contribute to establishing and maintaining relations of
 power, domination and exploitation. They may be enacted in ways of interaction (and therefore in genres) and
inculcated in ways of being identities (and therefore styles). Analysis of texts is an important aspect of
ideological analysis and critique (Fairclough, 2003: 218)

Power is another concept which is central for CDA, as it often analyses the language use
of those in power, who are responsible for the existence of inequalities. Typically, CDA
researchers are interested in the way discourse (re)produces social domination, that is,
the power abuse of one group over others, and how dominated groups may discursively
resist such abuse (e.g. Van Dijk in this volume). This raises the question of how CDA
researchers understand power and what moral standards allow them to differentiate
between power use and abuse a question which has so far had to remain unanswered
(Billig, 2008).
There are as many concepts of power as there are social theories. There is almost no
sociological or socio-psychological theory which does not provide a distinctive notion of
power, with a Weberian definition as the lowest common denominator: power as the
chance that an individual in a social relationship can achieve his or her own will even
against the resistance of others (Weber, 1980: 28).
At least three different approaches to power can be distinguished:

power as a result of specific resources of individual actors (e.g. French and Raven,
1959)
power as a specific attribute of social exchange in each interaction (e.g. Blau, 1964;
Emerson, 1962, 1975)
power as a systemic and constitutive element/characteristic of society (e.g. from
very different angles, Foucault, 1975 and Giddens, 1984).

Michel Foucault focuses on technologies of power: discipline is a complex bundle of

power technologies developed during the 18th and 19th centuries. Power is thus
exercised with intention but it is not individual intention. Foucault focuses on what is
accepted knowledge about how to exercise power. One way of doing this is by
threatening with violence. However, suggesting how happy people will become if they
buy specific consumer products is also an exercise of power; marketing provides us with
a large body of knowledge of powerful techniques. Though Foucault also combines the
notions of power and domination in a Weberian tradition, he focuses primarily on
structure. Foucault recommends an analysis of power with a rather functionalist strategy:
in his historical analysis in Surveiller et Punir (Foucault, 1975), he always asks and
answers questions concerning the social functions and effects of different technologies of
surveillance and punishment. How do things work at the level of ongoing subjugation, at
the level of those continuous and uninterrupted processes which subject our bodies,
govern our gestures and dictate our behaviours?
Within CDA, power is mostly perceived in the third way, not only because Foucault is
one of the theoretical godfathers of CDA, but also because the text in CDA is often
regarded as a manifestation of social action which again is widely determined by social
structure. Besides, CDA researchers rarely work with interactional texts such as
dialogues (Chilton, 2004; Lalouschek et al., 1990; Wodak, 2009 as exceptions).
Consequently, it is not the individual resources and not the specifics of single-exchange
situations that are crucial for CDA analyses, but the overall structural features in social
fields or in overall society. Power is central for understanding the dynamics and specifics
of control (of action) in modern societies, but power remains mostly invisible. Linguistic
manifestations are under investigation in CDA. This relation between social power and
language is a permanent topic not only in CDA (Fairclough, 1989/1991; Wodak, 1989)
but also in sociology (Bourdieu, 1991) and sociolinguistics (e.g. Ng and Bradac, 1993;
Talbot, 2003; Young and Fitzgerald, 2006).
An important perspective in CDA related to the notion of power is that it is very rare
that a text is the work of only one person. In texts, discursive differences are negotiated;
they are governed by differences in power that is in part encoded in and determined by
discourse and by genre. Therefore, texts are often sites of struggle in that they show traces
of differing discourses and ideologies contending and struggling for dominance.
Thus, the defining features of CDA are its concern with power as a central condition in
social life, and its efforts to develop a theory of language that incorporates this as a major
premise. Closely attended to are 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 (Iedema, 1997; Iedema and Wodak, 1999; Muntigl et al., 2000).
Power is about relations of difference, and 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 and
expresses power, and is involved where there is contention over and a challenge to
power. Power does not necessarily 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. Language provides a finely articulated vehicle for establishing differences in power
in hierarchical social structures.
In sum: CDA can be defined as being fundamentally interested in analysing opaque as
well as transparent structural relationships of dominance, discrimination, power and
control as manifested in language. In other words, CDA aims to investigate critically
social inequality as it is expressed, constituted, legitimized, and so on, by language use
(or in discourse). Most critical discourse analysts would thus endorse Habermass claim
that language is also a medium of domination and social force. It serves to legitimize
relations of organized power. Insofar as the legitimizations of power relations are not
articulated language is also ideological (Habermas, 1967: 259).
Main research agenda and challenges
In this section, we summarize some important research agendas which are currently of
interest in CDA. We then also list examples of research linked to these agendas and
challenges. Although we, of course, encounter a vast amount of research and also many
methodological and theoretical approaches, we have decided to restrict ourselves to six
major areas and related challenges:

1. Analysing, understanding and explaining the impact of the Knowledge-based

Economy on various domains of our societies; related to this, the recontextualization
of KBE into other parts of the world and other societies (transition).
2. Integrating approaches from cognitive sciences into CDA; this requires complex
epistemological considerations and the development of new tools. Moreover, we
question in which ways such approaches depend on Western cultural contexts and
how, related to these issues, Eurocentric perspectives could be transcended.
3. Analysing, understanding and explaining new phenomena in Western political
systems, which are due to the impact of (new) media and to transnational, global and
local developments and related institutions. More specifically, phenomena such as
depoliticization and participation need to be investigated in detail.
4. Analysing, understanding and explaining the impact of new media and related genres
which entails developing new multimodal theoretical and methodological
approaches. Our concepts of space and time have changed, and these changes
interact in dialectical ways with new modes and genres of communication.
5. Analysing, understanding and explaining the relationship between complex historical
processes, hegemonic narratives and CDA approaches. Identity politics on all levels
always entails the integration of past experiences, present events and future visions.
The concepts of intertextuality and recontextualization are inherently tied to
interdisciplinary approaches.
6. Avoiding cherry picking (choosing the examples which best fit the assumptions) by
integrating quantitative and qualitative methods and by providing retroductable, self-
reflective presentations of past or current research processes.

Of course, the many issues of Discourse and Society, Journal of Language and Politics,
Visual Semiotics, and Critical Discourse Studies, to name but a few, have published a
huge variety of CDA-oriented research over the past decade which we cannot review in
detail. We therefore necessarily have to refer readers to the many handbooks and journals
in the field.5

Language of the New Capitalism and the Knowledge-based Economy (KBE)

In Jessop et al. (2008), many aspects and dimensions of the impact of KBE on higher
education are explored from sociological, educational and CDA perspectives. KBE has
penetrated most domains of our Western societies and is also colonizing other parts of the
world. Indeed, globalization and competitiveness rhetoric (Muntigl et al., 2000) seem to
be ubiquitous, and the quantification and economization of knowledge serve to rank social
institutions and individuals. Through detailed case studies, the recontextualization of
more global policy strategies can be illustrated on the micro level (Falkner et al., 2005).
This, of course, requires interdisciplinary research as well as new theories on transition
and social change (see Krzy anowski and Wodak, 2008).
Chouliaraki and Fairclough (1999) explain and elaborate 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, of struggle
and also as a site where language is often 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 reveals the fallacy of such assumptions, and illustrates the mediating and
constructing role of the media with a variety of examples. He has also been concerned
with the Language of New Labour (Fairclough, 2000a) which, of course, is part and
parcel of KBE ideology.
In this vein, Jane Mulderrig (2006) develops a methodology and interdisciplinary
theoretical framework for historically analysing the exercise of governmental power in a
specific policy field, thus addressing at least three of the above-mentioned challenges. At
the theoretical level, this work employs an interdisciplinary approach to CDA by
grounding close textual analysis in both educational sociology and neo-Marxist state
theory. It thus contextualizes the linguistic findings in terms of historical developments in
the capitalist state over the last four decades. This research also develops novel ways of
using corpus tools in CDA, and in particular it demonstrates their heuristic value in
directing the analysts gaze in unexpected and often fruitful directions (see Mautner in this
volume; and Baker et al., 2008). More specifically, the themes explored in this research
include the rise of a more personalized form of governmental identity, and the prominent
role of the pronoun we in legitimizing government action and deflecting its public
accountability. A further strand to this research makes a direct link between the verbs
used to perform government actions and the increasing social significance of governing
at a distance or networked governance as described by, among others, the social
theorist Nikolas Rose (1999).
Phil Graham also elaborates the problems of New Capitalism while integrating a
strong historical perspective (Graham, 2002). The historical investigation of hortatory
genres, for example, compares the emergence of, and struggles between, the Church,
divine right royalties and secular forces over legitimate uses of the sermon form in
Western Europe between the 10th and 14th centuries, with contemporary struggles over
genres that are used to motivate people on a mass scale. The main focus of Grahams
research is to explore and explain the relationships between new media, new genres,
institutions and social change at a macro level. The perspective is primarily historical,
politicaleconomic, relational and dynamic. Genres are produced, textured and
transformed within institutional contexts over long periods of time. In turn, institutions
invest years in some cases, millennia in developing, maintaining and adapting generic
forms to changing social conditions in order to maintain or gain power. Graham believes
that at certain times in history, certain genres become very effective in motivating or
manipulating large sections of society. Because genres are developed within institutions,
and thus within the realms of vested interests, they display inherent axiological biases.
The nature of knowledge and its status as a commodity form immediately become
problematic. In the tradition of dialectical argumentation, Graham accepts the claims that
knowledge can become a dominant commodity; that a global economy can be built on
such forms; and that our new media must, in some fundamental way, underpin the
emergence of this new form of political economy. The research problem is therefore
formulated as an historical investigation into the relationship between language, new
media and social perceptions of value.

Focusing cognition

The seminal book by Teun van Dijk and Walter Kintsch, Strategies of Discourse
Comprehension (1983), triggered research in discourse and cognition from
interdisciplinary and critical perspectives. In this book, they considered the relevance of
discourse to the study of language processing. Currently, interest in cognition has grown,
and many scholars attempt a combination of new cognitive theories (on conceptual
metaphors, for example) with CDA (Charteris-Black, 2006; Musolff, 2004). Some of this
research draws on earlier attempts which integrated cognition, sociolinguistics and
discourse analysis (such as Lutz and Wodak, 1987; Wodak, 1986a, 1996; Wodak and
Schulz, 1986) by proposing new approaches (see below).
Much of the focus in this area has been placed on researching social inclusion and
exclusion. Teun Van Dijk, for example, has recently paid special attention to the
discursive reproduction of racism in Spain and Latin America (Van Dijk, 2005). The
study by John Richardson on the (mis)representation of Islam, and research on the
representation of migrants, asylum-seekers and refugees in the British Press, have
elaborated research on racism, antisemitism and xenophobia in intricate ways, by
combining quantitative and qualitative methods, and by focusing on argumentation as well
(see Baker et al., 2008; Delanty et al., 2008; Richardson, 2004; Wodak, 2008b, 2008c).
Moreover, the focus on theorizing context and knowledge is apparent. Van Dijk argues
that whereas (critical and other) discourse studies have paid extensive attention in the last
few decades to the structures of text and talk, they only paid lip-service to the necessity of
developing the relations between text and context (but see Panagl and Wodak, 2004).
Most approaches, also in CDA, define the influence of the social context on language
variation and discourse in terms of objective social variables such as gender, class, race,
ethnicity or age. Van Dijk argues that no such direct influence exists, because social
structures and discourse structures cannot be related directly, and need the mediation of
an interface. He shows that this interface must be cognitive, in the sense that it is not
objective social situations, but the subjective definitions of the relevant properties of
communicative situations that influence text and talk. These definitions are then made
explicit in terms of a special kind of mental model (see Van Dijk, this volume). In sum,
Van Dijk emphasizes that CDA should also not limit itself to a study of the relationship
between discourse and social structure, such as racism and other forms of power abuse,
but that language use always presupposes the intervening mental models, goals and
general social representations (knowledge, attitudes, ideologies, norms, values). In other
words, the study of discourse triangulates between society/culture/situation, cognition,
and discourse/language.
Paul Chilton, who is a cognitive linguist, has never explicitly applied the term CDA to
his own work and has always worked within a cognitive framework, principally on the
discourse of politics and the international relations framework (cf. Chilton, 1994a,
1994b, 1996a, 1996b, 2004, 2005a, 2005b; Chilton and Lakoff, 1995). His most recent
(2004, 2005a) and ongoing work departs from CDAs tendency to allegedly reify social
structures and processes and raises major research questions relating to the relationship
between language and social cognition in the evolution of the human species. More
particularly, he has drawn on cognitive evolutionary psychology to ask whether there
might exist an innate critical instinct. If this was the case, he argues, then what is the
role of critical discourse analysis? Chiltons argument is that the most fundamental issue
is whether societies provide the freedom to enable the critical instinct to operate. This
position is, of course, extremely vulnerable and has been challenged by other CDA
researchers (Fairclough in this volume; Van Dijk, 2007a; Wodak, 2007). Linked to this
approach is a concern with universal aspects of language and the human mind, a concern
that is also reflected in his current collaborative work on comparative discourse analysis
that crosses linguistic, cultural and political boundaries (Chilton, 2007; Chilton et al.,
forthcoming). Comparative discourse analysis, he argues, is the most serious challenge
facing practitioners of CDA, if CDA is to overcome its Eurocentric drift and respond to a
globalized scholarly environment.
The emergent blend of CDA, cognitive linguistics and corpus linguistics has become a
huge priority that is also recognized by many other scholars (see, for example, Koller and
Davidson, 2008). However, what remains unsolved is the apparent contradiction that
CDA starts from a complex social problem or phenomenon; cognitive linguistics,
however, starts from the individual mind, and corpus linguistics from the largely (but not
fully!) decontextualized text. Thus, integration will have to address these epistemological
considerations. Both Chilton and Koller (and many other CDA researchers) state that the
analysis of powerful, even hegemonic discourse(s) which however do not necessarily
discriminate against a particular social group has to be enhanced, for example, by aspects
of corporate self-presentation, as well as the impact of new genres, and so on (Koller,
2008).

Multimodality and new genres

Recognition of the contribution of all the aspects of the communicative context to text
meaning, as well as a growing awareness in media studies in general and in the
importance of non-verbal aspects of texts in particular, has turned attention to semiotic
devices in discourse other than the linguistic ones. In particular, the theory put forward
by Kress and van Leeuwen (1996) provides a useful framework for considering the
communicative potential of visual devices in the media. Currently, van Leeuwen is
focusing on the semiotics of handwriting and typography and the question of colour, as
well as on the constraints imposed by certain software, PowerPoint templates, and so on.
Thus, it is important for social semiotics to provide models of semiotic practice that are
appropriate to the practices they model, and as different semiotic practices are very
differently organized, it is not possible to apply a single model to all. Van Leeuwen
claims that the role and status of semiotic practices in society are currently undergoing
change as a result of the fact that it is increasingly global corporations and semiotic
technologies, rather than national institutions, which regulate semiotic production and
consumption (see also van Leeuwen, this volume).
Jay Lemkes recent work has emphasized multimedia semiotics, multiple timescales
and hypertexts or traversals. This work emphasizes the implicit value systems and their
connections to institutional and personal identity in new multimodal genres. The work on
multiple timescales, for example, is an extension of earlier work on ecologicalsocial
systems as complex dynamical systems with semiotic cultures. It is very important in
considering all aspects of social dynamics to look across multiple timescales, i.e. how
processes and practices which take place at relatively faster rates are organized within
the framework of more slowly changing features of social institutions and cultures (see
the concept of non-simultaneity or Ungleichzeitigkeit) (Lemke, 2001, 2002). Lemkes
work has combined both these themes to develop the idea that although we tell our lives
as narratives, we experience them as hypertexts. Building on research on the semantic
resources of hypertext as a medium, he proposed that postmodern lifestyles are
increasingly liberated from particular institutional roles, and that we tend to move, on
multiple timescales, from involvement in one institution to another, creating new kinds of
meaning, less bound to fixed genres and registers, as we surf across channels, websites
and lived experiences. This is seen as a new historical development, not supplanting
institutions, but building up new sociocultural possibilities on and over them.
Moreover, Lemke while building on CDA and extending the tools from systemic
linguistics has increasingly investigated combinations of language, visual media and
dynamicinteractive effects. The most advanced part of this work concerns computer
games. Here, one has to take into account not only the semiotic and semantic affordances
of an interactive environment, but also the phenomenological experience and particularly
the feelings and emotions of the user. This requires developing a number of new tools to
think about and analyse categories of feelings, and their combination with other sorts of
meanings. In this vein, the semantics of evaluations and judgements is a key link, as is re-
thinking feelings, at least in part, as social and distributed, rather than as totally
individual and internal. Of course, there are also cultural dimensions to feelings; this
point addresses similar issues to those mentioned by Chilton (see above), namely that
CDA has to move away from a Eurocentric focus.
Another branch of this work is the critical analysis of transmedia. This is the term now
used for sets of related media (e.g. a book and film, a website and a game, merchandise,
stories written by fans of the movie or games, etc.) that either form a commercial
franchise (e.g. Harry Potter, Star Wars ) or some more loosely connected intertextual set
(see the remarks on KBE above). In either case, there are economic and material
relationships as well as textual and semantic ones (see also the research by Graham
above). Thus, these cases are particularly revealing for a study of the political economy
of signs. Commercial interests and their ideologies interact with consumer interests and
consumer beliefs and desires. Of special importance is the appearance of collectives of
consumers who also become producers of parallel or counter media: for instance, readers
write their own stories changing the values and practices of famous characters from
popular movies, fiction, games, and so on, and these are widely read and distributed on
the internet, for example in blogs. Something similar is also happening with art, games
and music. Fans edit commercial movies, add a musical soundtrack, and create new
montage effects with completely different ideological and narrative meanings from the
original works, which illustrate creative and subversive processes of the
recontextualization of multiple genres and modes in detail.
Therefore, Theo van Leeuwen (2006: 292) argues that:

[c]ritical discourse analysis has also moved beyond language, taking on board that discourses are often
multimodally realized, not only through text and talk, but also through other modes of communication such as
images Overall, then, critical discourse analysis has moved towards more explicit dialogue between social
theory and practice, richer contextualization, greater interdisciplinarity and greater attention to the multimodality of
discourse.

Political discourse

The study of political discourse after the Second World War was triggered in part by the
investigation of National Socialist (NS) language (Klemperer, 1975/1947); it was
essential to understand and explain the roles and importance of language and
communication in totalitarian regimes and their propaganda. Utz Maas was the first
linguist to subject the everyday linguistic practice of National Socialism to an in-depth
analysis: he used NS texts to exemplify his approach of Lesweisenanalyse (Maas, 1984,
1989a, 1989b).
His historical argumentation analysis, based on the theories of Michel Foucault,
demonstrates how in complex ways discourse is determined by society, i.e. in what may
be termed a social practice. In his analysis of language practices during the National
Socialist regime between 1932 and 1938, he showed how the discursive practices of
society in Germany were impacted by the National Socialist (NS) discourse
characterized by socialrevolutionist undertones. NS discourse had superseded almost all
forms of language (practices), a fact that made it difficult for an individual who did not
want to cherish the tradition of an unworldly Romanticism to use language in a critical
reflective way. Discourse is basically understood as the result of collusion: the
conditions of the political, social and linguistic practice impose themselves practically
behind the back of the subjects, while the actors do not understand the game (see also
Bourdieus violnce symbolique). Discourse analysis identifies the rules which make a
text into a fascist text. In the same way as grammar characterizes the structure of
sentences, discourse rules characterize utterances/texts that are acceptable within a
certain practice. The focus is not on National Socialist language per se, but the aim is
rather to record and analyse the spectrum of linguistic relations based on a number of
texts dealing with various spheres of life. These texts represent a complicated network of
similarities, which overlap and intersect. Therefore, it is also important to do justice to
the polyphony of texts resulting from the fact that societal contradictions are inscribed
into texts (see also the concept of entextualization as employed by Blommaert, 2005).
Texts from diverse social and political contexts (cooking recipes, local municipal
provisions on agriculture, texts by NS politicians, but also by critics of this ideology,
who are ultimately involved in the dominant discourse) are analysed in a representative
sample (see Wodak and de Cillia, 2006, for details and an extensive overview of the
field of language and politics).
The study of political institutions and everyday life and decision-making in
organizations has become a major new focus of CDA. Krzy anowski and Oberhuber
(2007), for example, have analysed the European Convention in much detail. The focus on
discursive dimensions of transnational political organizations also led to the elaboration
of discursively constructed visions/conceptions of social and political order in
Europe/the EU. Wodak (2009) focuses on the everyday lives of MEPs and other
politicians because as she argues depoliticization is linked to the democracy deficit
and the huge dissatisfaction about the strong ritualization of politics and the snapshots
provided by media which condense complex political processes into iconic images
(Triandafyllidou et al., 2009). Such studies allow insight into politics as a profession
and into the complexity of political decision-making. If the media, however, allow us to
venture backstage, this usually happens in the context of the sex and corruption scandals
of politicians. (Hence, in the above-mentioned ethnographic studies, backstage access
opens the door to understanding the doing of politics.)
 Much CDA research in the domain of politics centres on right wing populist rhetoric on
many occasions, as right wing populist rhetoric is becoming more and more hegemonic in
many European countries (see Haiderization Krzy anowski and Wodak, 2008; Pelinka
and Wodak, 2002; Richardson and Wodak, 2008; Rydgren, 2005; Wodak and Pelinka,
2002). This research is triggered by the rising dominance and hegemony of this kind of
rhetoric and its apt use of indirect strategies to address multiple audiences (see Reisigl
and Wodak, this volume). The latter research also develops new methodologies for CDA:
the use of ethnography, focus groups and narrative interviews, combined with more
traditional data sources such as newspapers and political speeches (Wodak and Krzy
anowski, 2008).
Research on politics from a historical perspective also co-triggered CDA from the
very beginning. The study for which the Discourse-Historical Approach (DHA) was
actually developed, for instance, first attempted to trace in detail the constitution of an
antisemitic stereotyped image, or Feindbild, as it emerged in public discourse in the 1986
Austrian presidential campaign of Kurt Waldheim (Gruber, 1991; Mitten, 1992; Wodak et
al., 1990). In order to be able to study the discourse about the Waldheim Affair,
context was unravelled into various dimensions.
The DHA has been further elaborated in a number of more recent studies, for example,
in a study on racist discrimination against immigrants from Romania and in a study on the
discourse about nation and national identity in Austria (Kovcs and Wodak, 2003; Wodak
et al., 1999) and in the European Union (Muntigl et al., 2000; Wodak and Van Dijk,
2000). The 1999 study was concerned with the analysis of the relationships between the
discursive construction of national sameness and the discursive construction of difference
leading to the political and social exclusion of specific outgroups. The findings suggest
that discourses about nations and national identities rely on at least four types of
discursive macro-strategies. These are:

constructive strategies (aiming at the construction of national identities)

justificatory strategies (aiming at the conservation and reproduction of national
identities or narratives of identity)
transformative strategies (aiming at the change of national identities)
destructive strategies (aiming at the dismantling of national identities).

More recently, much research has focused on commemorative events which manifest
hegemonic ways of dealing with traumatic pasts in various societies (Anthonissen and
Blommaert, 2007; Blommaert, 2005; de Cillia and Wodak, 2008; Ensink and Sauer, 2003;
Heer et al., 2008; Le, 2006; Martin and Wodak, 2003; Reisigl, 2007; Wodak and de
Cillia, 2007). In most of these studies, media, school books, speeches at national days
and the like are analysed to illustrate the myths which are constructed to provide new,
sanitized narratives which cover up ruptures, war crimes and conflicts which have
occurred in the past. For example, Heer et al. (2008) describe in detail the huge scandal
and crisis when the two exhibitions on war crimes committed by the German Wehrmacht
during the Second World War were opened to viewers, in 1995 and 2001. A carefully
constructed and protected myth was destroyed by these exhibitions the myth that the
Wehrmacht soldiers had been innocent whereas the SS and other units had been the sole
perpetrators.

Differences and similarities beyond the social dimension

The differences between CDA and other DA, pragmatic and sociolinguistic approaches
may be most clearly established with regard to the general principles of CDA. Firstly, the
nature of the problems with which CDA is concerned differs from all those approaches
which do not explicitly express their research interest. In general, CDA asks different
research questions, and some CDA scholars play an advocatory role for socially
discriminated groups. Looking at the CDA contributions collected in this reader, it also
becomes evident that the line drawn between social scientific research which ought to be
intelligible, and political argumentation sometimes gets blurred.
Specifically, we distinguish between approaches which proceed deductively and such
that choose a more inductive perspective. Linked to this distinction is the choice of
objects under investigation: more deductively oriented theories which also propose a
closed theoretical framework are more likely to illustrate their assumptions with a few
examples which seem to fit their claims (e.g. the dialecticalrelational approach and
sociocognitive approach in this volume). More inductively oriented approaches usually
stay at the meso level and select problems where they attempt to discover new insights
through in-depth case studies and ample data collection (for example, DHA, social actors
approach, corpus linguistics approach, dispositive analysis, in this volume). Of course,
all approaches moreover proceed abductively, i.e. oscillate between theory and data
analysis in retroductive ways. However, on a continuum, we are able to distinguish
obvious priorities in choosing entry points and themes (see Figure 1.1).
Related to the choice of more macro- or meso-topics (such as globalization or
knowledge versus un/employment or right wing populism), we encounter differences
in the evaluation of the chosen topics and objects under investigation. Macro-topics are
relatively non-controversial in the respective national or international academic contexts;
some meso-topics, however, can touch on the very core of the respective national
community to which the researcher belongs. For example, research on concrete
antisemitic, xenophobic and racist occurrences is much more controversial in certain
academic and national contexts and regarded as unpatriotic or hostile than many macro-
themes this explains grave problems which critical scholars have encountered when
venturing into such seemingly tabooed fields (see Heer et al., 2008).
FIGURE 1.1 Overall research strategies and theoretical background

In any case, related to the object of investigation, it remains a fact that CDA follows a
different and critical approach to problems, since it endeavours to reveal power relations
that are frequently obfuscated and hidden, and then to derive results which are also of
practical relevance.
Furthermore, one important characteristic arises from the assumption that all
discourses are historical and can therefore only be understood with reference to their
context. Therefore, these approaches refer to such extralinguistic factors as culture,
society and ideology in intricate ways, depending on their concepts of context and their
research methodologies and ways of data collection. Hence, the notion of context is
crucial for CDA, since this explicitly includes socialpsychological, political and
ideological components and thereby postulates interdisciplinary procedures.
Interdisciplinarity is implemented in many different ways in the CDA approaches
assembled in this volume: in some cases, interdisciplinarity is characteristic of the
theoretical framework (dispositive approach, dialecticalrelational approach,
sociocognitive approach); in other cases, interdisciplinarity also applies to the collection
and analysis of data (social actors approach, DHA). Moreover, CDA uses the concepts of
intertextuality and interdiscursivity; in sum, it may be concluded that CDA is open to a
broad range of factors exerting an influence on texts.
 A further difference between CDA and other DA approaches emerges with regard to
the dialectic relationships between language and society. CDA does not take this
relationship to be simply deterministic but invokes the concept of mediation. The
dialecticalrelational approach draws on Hallidays multifunctional linguistic theory
(Halliday, 1985) and the concept of orders of discourse according to Foucault, while the
discourse-historical approach and the sociocognitive approach make use of theories of
social cognition (e.g. Moscovici, 2000). This reflection on issues of mediation between
language and social structure is absent from many other linguistic approaches, for
example from conversation analysis. This is somewhat related to the level of social
aggregation: though CDA concentrates on notions like ideology or power, scholars focus
on different units of analysis the way in which individuals mentally perceive, or the
way social structures determine discourse (see Figure 1.2). In simplified terms, we can
distinguish between more cognitivesocio-psychological and more macro-sociological
structural approaches although, admittedly, this is a rough distinction.
A further characteristic of CDA is that most researchers integrate linguistic categories
into their analyses but to a different extent and with a different focus and intensity. CDA
does not necessarily include a broad range of linguistic categories in each single analysis;
one might get the impression that only a few linguistic devices are central to CDA studies.
For instance, many CDA scholars consistently use social actor analysis by focusing upon
pronouns, attributes and the verbal mode, time and tense; Hallidayan transitivity analysis
and the analysis of argumentative topoi are also used frequently by other social scientists
because these concepts seem to be quite easy to apply without much linguistic background
knowledge. Exceptions always prove the generalization: Reisigl and Wodak (this
volume) and Van Dijk (this volume) illustrate how a broad range of macro and micro
linguistic, pragmatic and argumentative features can be operationalized and integrated in
the analysis of specific texts (Figure 1.2).
FIGURE 1.2 Linguistic depth of field and level of aggregation

In principle, we may assume that linguistic categories such as deixis and pronouns can
be analysed in any linguistic methodology, but they are salient for CDA. Explicitly or
implicitly, CDA distinguishes between the linguistic surface and some kind of deep
structure; for instance, the dialecticalrelational approach speaks of form and texture at
the textual level, the discourse-historical approach of forms of linguistic realization.
As for the methods and procedures used for interpretation, CDA generally views them
as hermeneutic, although this characteristic is not completely evident in the positioning of
every author. Hermeneutics can be understood as the method of grasping and interpreting
meanings. The hermeneutic circle i.e. the meaning of one part can only be understood
in the context of the whole, but this in turn is only accessible from its components
indicates the problem of the intelligibility of the hermeneutic interpretation. Therefore,
hermeneutic interpretation requires detailed documentation. However, the specifics of the
hermeneutic interpretation process are not made completely transparent by many CDA-
orientated studies.6 If a crude distinction has to be made between text-extending and
text-reducing methods of analysis, then CDA, on account of its concentration of very
clear formal properties and the associated compression of texts during analysis, may be
characterized as text-reducing. These findings contradict the mainly hermeneutic
impetus of most CDA approaches.
Methodology
CDA in all its various forms understands itself to be strongly based in theory. To which
theories do the different approaches refer? Here we encounter a variety of theories,
ranging from theories on society and power in Michel Foucaults tradition and theories of
social cognition and grammar, to individual concepts that draw on larger theoretical
traditions. As a first step, this section aims to systematize these different theoretical
influences (see Figure 1.1).
Moreover, this section is devoted to the problem of the operationalization of
theoretical concepts. The primary issue here is how the various approaches of CDA are
able to translate their theoretical claims into instruments and methods of analysis. In
particular, the emphasis is on mediation between grand theories as applied to larger
society, and concrete instances of social interaction which result in texts. In addition to
what can be described primarily as hermeneutics, one finds interpretative perspectives
with differing emphases, among them even quantitative procedures (see Mautner, this
volume).
Particularly worthy of discussion is the way in which sampling is conducted in CDA.
Most studies analyse typical texts. What is typical in which social situation, and for
which aspect of a social problem, however, frequently remains vague.
The links between theory and discourse in CDA can be described in terms of the model
for theoretical and methodological research procedures illustrated in Figure 1.3 (see
below).

Theoretical grounding and objectives

Among the different positions within CDA presented in this book, there is neither any
guiding theoretical viewpoint that is used coherently within CDA, nor do the CDA
protagonists proceed consistently from the area of theory to the field of discourse and
text, and back to theory (see Figure 1.1).
Within the CDA approaches presented here, all the theoretical levels of sociological
and socio-psychological theory can be discovered (the concept of different theoretical
levels is in the tradition of Merton, 1967: 3972):
FIGURE 1.3 Empirical research as a circular process

Epistemology, i.e. theories which provide models of the conditions, contingencies

and limitations of human perception in general, and scientific perception in
particular.
General social theories, often called grand theories, try to conceptualize the
complex relations between social structure and social action and thus link micro-and
macro-sociological phenomena. Within this level, one can distinguish between the
more structuralist and the more individualist approaches. To put it very simply, the
former provide top-down explanations (structure>action), whereas the latter prefer
bottom-up explanations (action>structure). Many modern theories try to reconcile
these positions and imply some kind of circularity between social action and social
structure.
Middle-range theories focus either upon specific social phenomena (e.g. conflict,
cognition, social networks) or on specific subsystems of society (e.g. economy,
politics, religion).
Micro-sociological theories try to explain social interaction, for example the
resolution of the double contingency problem (Parsons and Shils, 1951: 329) or the
reconstruction of everyday procedures which members of a society use to create
their own social order (i.e. ethnomethodology).
Socio-psychogical theories concentrate upon the social conditions of emotion and
cognition and, compared to microsociology, prefer causal explanations to a
hermeneutic interpretation of meaning.
Discourse theories aim at the conceptualization of discourse as a social
phenomenon and try to explain its genesis and its structure.
Linguistic theories, e.g. theories of argumentation, of grammar, of rhetoric, describe
the patterns specific to language systems and verbal communication.

As all these theoretical levels can be found in CDA, in the following, we present a short
overview of the theoretical positions and methodological objectives of CDA approaches.
Among the approaches assembled in this reader, Dispositive Analysis (DA Siegfried
Jger and Florentine Maier in this volume) is closest to the origin of the notion of
discourse, i.e. to Michel Foucaults structuralist explanations of discursive phenomena.
However, it detects a blind spot in Foucaults theory, namely the mediation between
subject and object, between discursive and nondiscursive practices (activities) on the one
hand and manifestations (objects) on the other. Here, DA strategically inserts Aleksej
Leontevs (e.g. 1982) activity theory. Thus, the social actor becomes the link between
discourse and reality. The epistemological position is based upon Ernesto Laclaus social
constructivism, which denies any societal reality outside of the discursive. In that way,
Dispositive Analysis introduces a dualism of discourse and social reality.
As all CDA approaches accept that discourse, understood as language use, is but one
manifestation of social action (Chilton, 2005b: 20), DA is forced to argue against the
Foucauldian notion of discourse which also includes non-linguistic elements. DA
therefore applies Jrgen Links notion of discourse as an institutionalized way of talking
that regulates and reinforces action and thereby exerts power (Link, 1983: 60).
Furthermore, Foucaults concept of the dispositive frames both discursive and
nondiscursive practices and materializations. DA thus explicitly aims at the analysis of
discourses and dispositives.
The Sociocognitive Approach (SCA Teun van Dijk in this volume) is situated on the
socio-psychological dimension of the CDA field. The approach draws on social
representation theory (e.g. Moscovici, 2000). Discourse is seen as a communicative
event, including conversational interaction and written text, as well as associated
gestures, facework, typographical layout, images and any other semiotic or multimedia
dimension of signification. Van Dijk relies on sociocognitive theory and understands
linguistics in a broad structuralfunctional sense. He argues that CDA should be based
on a sound theory of context. Within this claim, the theory of social representations plays
a major role.
Social actors involved in discourse do not only use their individual experiences and
strategies, they rely upon collective frames of perceptions, i.e. social representations .
These socially shared perceptions form the link between the social system and the
individual cognitive system, and perform the translation, homogenization and
coordination between external requirements and subjective experience. Thus, the
approach refers to the link which Chilton (2005a) detects as missing. This assumption is
not new: already in the first half of the 20th century, mile Durkheim (e.g. 1933) pointed
to the significance of collective ideas in constructing social order. Serge Moscovici
(1982) coined the notion of social representations as a bulk of concepts, opinions,
attitudes, evaluations, images and explanations which result from daily life and are
sustained by communication. Social representations are shared among members of a
social group, which was already stated by mile Durkheim: The ideas of man are not
personal and are not restricted to me; I share them, to a large degree, with all the men
who belong to the same social group that I do. Because they are held in common, concepts
are the supreme instrument of all intellectual exchange (cited in Bellah, 1973: 52). Thus,
they form a core element of the individuals social identity (Wagner, 1994: 132). Social
representations are always bound to specific social groups. They are dynamic constructs
and subject to permanent change. Together, they constitute a hierarchical order of mutual
dependency (Duveen and Lloyd, 1990). SCA introduces the concept of context models,
i.e. mental representations of the structures of the communicative situation that are salient
for a participant. These context models control the pragmatic part of discourse, whereas
event models control the semantic part. Three forms of social representations are
relevant in this context:

1. knowledge (personal, group, cultural)

2. attitudes (not in the social psychology understanding)
3. ideologies.

Discourses take place within society, and can only be understood in the interplay of
social situation, action, actor and societal structures.
The Discourse-Historical Approach (DHA) explicitly tries to establish a theory of
discourse by linking fields of action (Girnth, 1996), genres, discourses and texts.
Although DHA is aligned to critical theory, grand theories play a minor role compared
with the discourse model and the emphasis on historical analysis: context is understood
as mainly historical. Thus far, the DHA consistently agrees with Mouzeliss (1995)
pessimistic diagnosis of social research. DHA consequently follows his
recommendations: not to get lost in theoretical labyrinths, not to invest too much in the
operationalization of inoperationalizable grand theories but rather to develop
conceptual tools adequate for specific social problems. The DHA focuses on the field of
politics, where it develops conceptual frameworks for political discourse. Reisigl and
Wodak try to fit linguistic theories into their model of discourse. They make extensive use
of argumentation theory in the example presented in this volume. That does not
necessarily imply that the concepts resulting from argumentation theory are adequate for
other research questions.
As the Corpus Linguistics Approach (CLA Gerlinde Mautner in this volume) is a
quantitative, linguistic extension of CDA, it provides additional linguistic devices for
thorough analysis and can be applied against the backdrop of CDA approaches.
The Social Actors Approach (SAA Theo van Leeuwen in this volume) refers to a
broad scope of sociological and linguistic theories, especially to those explaining the role
of action to establish social structure: representation is ultimately based on practice, on
that which people do it is the primacy of practice which constitutes the theoretical core
of SAA. Therefore, SAA introduces sources from Malinowski to Parsons, and from
Bernstein to Bourdieu. This idea of individual actors permanently constituting and
reproducing social structure is linked with a Foucauldian notion of discourse, somewhat
similar to DA and DRA.
The DialecticalRelational Approach (DRA Norman Fairclough in this volume)
takes a rather grand-theory-oriented position: Fairclough focuses upon social conflict in
the Marxian tradition and tries to detect its linguistic manifestations in discourses, in
specific elements of dominance, difference and resistance. According to DRA, every
social practice has a semiotic element. Productive activity, the means of production,
social relations, social identities, cultural values, consciousness and semiosis are
dialectically related elements of social practice. He understands CDA as the analysis of
the dialectical relationships between semiosis (including language) and other elements of
social practices. These semiotic aspects of social practice are responsible for the
constitution of genres and styles. The semiotic aspect of social order is called the order
of discourse. His approach to CDA oscillates between a focus on structure and a focus on
action. CDA should pursue emancipatory objectives, and should be focused upon the
problems confronting what can loosely be referred to as the losers within particular
forms of social life. DRA draws upon a specific linguistic theory Systemic Functional
Linguistics (Halliday, 1985) which analyses language as shaped (even in its grammar)
by the social functions it has come to serve.
Gathering data
We concluded above that CDA does not constitute a well-defined empirical methodology
but rather a bulk of approaches with theoretical similarities and research questions of a
specific kind. But there is no CDA way of gathering data, either. Some authors do not
even mention sampling, while other scholars strongly rely on traditions based outside the
sociolinguistic field.7 In any case, similar to Grounded Theory (Glaser and Strauss,
1967), data collection is not considered to be a specific phase that must be completed
before analysis begins: it is a matter of finding indicators for particular concepts,
expanding concepts into categories and, on the basis of these results, collecting further
data (theoretical sampling). In this procedure, data collection is never completely
excluded, and new questions always arise which can only be dealt with if more data are
collected or earlier data are re-examined (Strauss, 1987; Strauss and Corbin, 1990).
Most of the approaches to CDA do not explicitly recommend sampling procedures.
Obviously, CLA specifically refers to large corpora of text. DA, SCA, SAA and DRA
mainly rely on existing texts, such as mass media communication or documents. This is
also true for DHA, though it additionally postulates that studies should incorporate
fieldwork and ethnography if possible to explore the object under investigation as a
precondition for any further analysis and theorizing. The focus on already existing texts
implies specific strengths, in particular it provides non-reactive data (Webb et al., 1966),
and certain weaknesses concerning the research questions which have to be
operationalized.
Operationalization and analysis
CDA places its methodology in the hermeneutic rather than in the analyticaldeductive
tradition. In any case, the linguistic character of CDA becomes evident, because in
contrast to other approaches of text and discourse analysis (e.g. Content Analysis,
Grounded Theory, Conversation Analysis; see Titscher et al., 2000), CDA strongly relies
on linguistic categories. This does not mean, of course, that topics and contents play no
role, but the core operationalizations depend on linguistic concepts such as actors, mode,
time, tense, argumentation and so on. Nevertheless, an exhaustive list of linguistic devices
cannot be given, for their selection depends mainly on the specific research questions.
DA distinguishes between a more content-oriented phase of (1) structure analysis and a
more linguistically oriented phase of (2) fine analysis. Within structure analysis, the
media and the general themes have to be characterized. Within the fine analysis, DA
focuses upon context, text surface and rhetorical means. Exemplary linguistic indicators
are figurativeness, vocabulary and argumentation types. DA takes into account both
qualitative and quantitative aspects of these features. DA analyses:

the kind and form of argumentation

certain argumentation strategies
the intrinsic logic and composition of texts
implicit implicateurs and insinuations
the collective symbolism or figurativeness, symbolism, metaphors and so on, both
in language and in design (statistics, photographs, pictures, caricatures, etc.)
idioms, sayings, clichs, vocabulary and style
actors (persons, pronominal structure)
references, for example, to (the) science(s)
the particulars of the sources of knowledge, etc.

SCA generally argues that a complete discourse analysis of a large corpus of text or talk
is totally impossible. If the focus of research is on the ways in which some speakers or
writers exercise power in or by their discourse, research focuses on those properties that
can vary as a function of social power. SCA therefore suggests the following linguistic
indicators:

stress and intonation

word order
lexical style
coherence
local semantic moves such as disclaimers
topic choice
 speech acts
schematic organization
rhetorical figures
syntactic structures
propositional structures
turn-takings
repairs
hesitation.

SCA assumes that most of these are exemplary forms of interaction which are in principle
susceptible to speaker control, but are in practice mostly not consciously controlled.
Other categories, such as the form of words and many structures of sentences, are
grammatically obligatory and contextually invariant and hence are usually not subject to
speaker control and social power. SCA further suggests six steps of analysis:

1. The analysis of semantic macrostructures: topics and macropropositions.

2. The analysis of local meanings, where the many forms of implicit or indirect
meanings, such as implications, presuppositions, allusions, vagueness, omissions
and polarizations, are especially interesting.
3. The analysis of subtle formal structures : here, most of the linguistic markers
mentioned are analysed.
4. The analysis of global and local discourse forms or formats.
5. The analysis of specific linguistic realizations, e.g. hyperbole, litotes.
6. The analysis of context.

DHA unfolds a four-step strategy of analysis: after (1) having established the specific
contents or topics of a specific discourse (e.g. with racist, antisemitic, nationalist or
ethnicist ingredients), (2) the discursive strategies (including argumentation strategies)
are investigated. Then (3), the linguistic means (as types) and the specific, context-
dependent linguistic realizations (as tokens) of the discriminatory stereotypes are
examined (4).
In these studies, DHA focuses on the following discursive strategies:

Referential strategy or strategy of nomination, where the salient linguistic devices

are membership categorization (Sacks, 1992), metaphors, metonymies and
synecdoches.
Strategies of predication which appear in evaluative attributions of positive or
negative traits and implicit or explicit predicates.
Strategies of argumentation which are reflected in certain topoi (i.e. used to justify
political inclusion or exclusion).
 Strategies of perspectivization, framing or discourse representation use specific
means of reporting, description, narration or the quotation of events and utterances.
Strategies of intensification and mitigation intensify or mitigate the illocutionary
force of utterances (Ng and Bradac, 1993).

This methodology aims to be abductive, because the categories of analysis are first
developed in accordance with the research questions, and a constant movement back and
forth between theory and empirical data is suggested. The historical context is always
analysed and integrated into the interpretation, although there is no clear procedure for
this task.
CLA adds a quantitative methodology to CDA: large corpora are analysed by means of
concordance software, thus providing frequency lists and specific measures of statistical
significance. Therefore, CLA applies a rather deductive methodology in selecting
specific words which are relevant for analysis, but also offers concordance lines as a
basis for further (qualitative) interpretation.
In a very general sense, SAA describes and explains social practices by identifying
seven core elements. Firstly, it tries to identify (1) actions within given texts; then it
analyses (2) performance modes and denotes (3) actors which apply (4) specific
presentation styles of their actions. Social practices which are constituted by these
actions take place in (5) specific times and (6) spaces, and actors require (7) specific
resources. Some of these elements of social practice are eligible, while some are deleted,
i.e. are not represented in a specific discourse. Some are substituted; some reactions and
motives are added to the representation of social practices.
DRA suggests a stepwise procedure in preparation for analysis. Like DHA, it prefers a
pragmatic, problem-oriented approach, where the first step is to identify and describe the
social problem which should be analysed. DRA stages of analysis are as follows:

1. Focus upon a specific social problem which has a semiotic aspect, go outside the
text and describe the problem, and identify its semiotic dimension.
2. Identify the dominant styles, genres and discourses constituting this semiotic
dimension.
3. Consider the range of difference and diversity in styles, genres and discourses
within this dimension.
4. Identify the resistance against the colonialization processes executed by the dominant
styles, genres and discourses.

After these preparatory steps, which also help to select the material, DRA suggests (1) a
structural analysis of the context, and then (2) an interactional analysis, which focuses
on such well-known linguistic features as:
 agents
time
tense
modality
syntax

and finally (3) an analysis of interdiscursivity, which tries to compare the dominant and
resistant strands of discourse.
Above, we have only been able to give a brief overview of the core procedures
applied in the different approaches to CDA. Finally, it should be pointed out that,
although there is no consistent CDA methodology, some features are common to most
CDA approaches: (1) they are problem-oriented and not focused on specific linguistic
items, yet linguistic expertise is obligatory for the selection of the items relevant to
specific research objectives; (2) theory as well as methodology is eclectic, both of which
are integrated to be able to understand the social problems under investigation.
Criteria for assessing quality
It seems to be beyond controversy that qualitative social research also needs concepts
and criteria to assess the quality of its findings. It is also indisputable that the classical
concepts of validity, reliability and objectivity used in quantitative research cannot be
applied in unmodified ways. The real issue is how our research can be both
intellectually challenging and rigorous and critical (Silverman, 1993: 144); Silverman
also provides a detailed discussion of these concepts and a reformulation for qualitative
research. Within CDA, there is little specific discussion about quality criteria.
DA at least mentions the classical criteria of representativeness, reliability and
validity. Beyond it, DA suggests completeness as a criterion suited for CDA: the results
of a study will be complete if new data and the analysis of new linguistic devices
reveal no new findings. SCA suggests accessibility as a criterion which takes into
account the practical targets of CDA: findings should at least be accessible and readable
for the social groups under investigation.
DHA suggests triangulation procedures to ensure validity which is appropriate
whatever ones theoretical orientation or use of quantitative or qualitative data
(Silverman, 1993: 156).8 DHAs triangulatory approach is mainly theoretical and based
on the concept of context which takes into account four levels: (1) the immediate
language-or text-internal co-text; (2) the intertextual and interdiscursive relationship
between utterances, texts, genres and discourses; (3) the extralinguistic (social) level,
which is called the context of situation and is explained by middle-range theories; and
(4) the broader sociopolitical and historical contexts. Permanent switching between these
levels and evaluating the findings from these different perspectives should minimize the
risk of being biased. Beyond it, DHA suggests methodical triangulation by using
multimethodical designs on the basis of a variety of empirical data as well as background
information.
Triangulation among different types of data, participants definition of relevance and
problem-based analysis to establish the significance of the sites of engagement and
mediated actions under investigation are suited to bring the analyses back to participants:
to uncover divergences and contradictions between ones own analysis of the mediated
actions one is studying and those of participants.
Nevertheless, rigorous objectivity cannot be reached by means of discourse analysis,
for each technology of research must itself be examined as potentially embedding the
beliefs and ideologies of the analysts and therefore guiding the analysis towards the
analysts preconceptions.

Summary and criticism

The goal of this chapter was to provide a summary of CDA approaches, their similarities
and differences. One of CDAs volitional characteristics is its diversity. Nevertheless, a
few salient cornerstones can be noticed within this diversity:

Concerning its theoretical background, CDA works eclectically in many aspects.

The whole range between grand theories and linguistic theories is evoked, although
each approach emphasizes different levels.
There is no accepted canon of data collection, but many CDA approaches work with
existing data, i.e. texts not specifically produced for the respective research
projects.
Operationalization and analysis are problem-oriented and imply linguistic expertise.

The most evident similarity is a shared interest in social processes of power, hierarchy-
building, exclusion and subordination. In the tradition of critical theory, CDA aims to
shed light on the discursive aspects of societal disparities and inequalities. CDA
frequently detects the linguistic means used by the privileged to stabilize or even to
intensify inequalities in society. This entails careful systematic analysis, self-reflection at
every point of ones research and distance from the data which are being investigated.
Description and interpretation should be kept apart, thus enabling transparency and
retroduction of the respective analysis. Of course, not all of these recommendations are
consistently followed, and they cannot always be implemented in detail because of time
pressures and similar structural constraints. Therefore, some critics will continue to state
that CDA constantly sits on the fence between social research and political argumentation
(Wodak, 2006a), while others will accuse some CDA studies of being too linguistic or
not linguistic enough. In our view, such criticism keeps a field alive because it
necessarily stimulates more self-reflection and encourages new questions, new responses
and new thoughts.

Notes
1 This chapter is based on long and extensive discussions with friends, colleagues and co-researchers as well as
students. Ruth Wodak would like to thank Teun van Dijk, Paul Chilton, Theo van Leeuwen, John Richardson, Jay
Lemke and Michael Billig, as well as all her co-researchers and (former) students. Finally, we would like to thank
the many colleagues we have not been able to mention here.
2 See Anthonissen, 2001; Blommaert and Bulcaen, 2000; Fairclough and Wodak, 1997; Reisigl and Wodak, 2001;
Titscher et al., 2000; Weiss and Wodak, 2003; Wodak and Meyer, 2001; Wodak and Pelinka, 2002, etc.
3 See Language and Power by Norman Fairclough (1989/1991), Language, Power and Ideology by Ruth
Wodak (1989) and Prejudice in Discourse by Teun van Dijk (1984).
4 The Erasmus network 1995 consisted of a cooperation between Siegfried Jger, Duisburg, Per Linell, Linkping,
Norman Fairclough, Lancaster, Teun van Dijk, Amsterdam, Gunther Kress, London, Theo van Leeuwen, London,
and Ruth Wodak, Vienna.
5 Unfortunately, we have to neglect much research here which could certainly also be categorized as critical, such
as feminist CDA (see Lazar, 2005), critical ethnographic approaches (see Blommaert, 2005), research situated
between sociolinguistics, literacy research and CDA (see Tricento, 2005), etc. We have to refer readers to the
many publications and overview essays in the field.
6 The question of whether it is possible to make hermeneutic processes transparent and intelligible at all remains
undecided, although some authors (Oevermann et al., 1979) have developed a hermeneutically orientated method
with well-defined procedures and rules.
7 A general survey on sampling and the selection of texts is given by Titscher et al. (2000).
8 An early proponent of the method of triangulation is Norman Denzin (1970). For further discussion of criteria for
assessing interpretive validity in qualitative research, see, for example, Altheide and Johnson (1994); Morse et al.
(2002).
2
Theoretical and Methodological Aspects of Foucauldian Critical Discourse
Analysis and Dispositive Analysis

Siegfried Jger and Florentine Maier

Introduction
Foundations of discourse theory
From discourse to dispositive
Methods for analysing discourses and dispositives
Introduction
This article aims to give a concise introduction to the methodology of analysing
discourses and dispositives, building on the theoretical insights of Michel Foucault. The
article is mainly aimed towards novices to this approach.1
Critical discourse analysis (CDA) based on Michel Foucaults discourse theory
centres on the following questions:

What is valid knowledge at a certain place and a certain time?

How does this knowledge arise and how is it passed on?
What functions does it have for constituting subjects?
What consequences does it have for the overall shaping and development of society?

Knowledge refers to all kinds of contents that make up a human consciousness, or in

other words, all kinds of meanings that people use to interpret and shape their
environment. People derive this knowledge from the discursive surroundings into which
they are born and in which they are enmeshed throughout their lives. Knowledge is
therefore conditional, i.e. its validity depends on peoples location in history, geography,
class relations and so on.
Discourse analysis and its extension, dispositive analysis, aim to identify the
knowledges contained in discourses and dispositives, and how these knowledges are
firmly connected to power relations in power/knowledge complexes. Discourse and
dispositive analysis subject these workings of power/knowledge to critique. All sorts of
knowledge can be subjected to analysis. This, for example, includes everyday knowledge
transmitted through everyday communication, scientific knowledge from the natural as
well as the cultural sciences, knowledge transmitted by the media, by schools and so on.
To provide an overview of the methodology for critically analysing discourses and
dispositives, we first need to establish the foundations of Foucauldian discourse theory
on which critical discourse and dispositive analysis build. A brief outline of this
theoretical background is given in the ensuing section on the foundations of discourse
theory. The third section (From discourse to dispositive) then introduces the concept of
the dispositive. Dispositives can be understood as the synthesis of discursive practices
(i.e. speaking and thinking on the basis of knowledge), nondiscursive practices (i.e. acting
on the basis of knowledge) and materializations (i.e. the material products of acting on the
basis of knowledge). In the final section, practical methodological guidelines for
analysing discourses and dispositives are provided.

Foundations of discourse theory

One of the most fruitful approaches to Foucauldian discourse analysis in the cultural
sciences has been developed by literary and cultural scientist Jrgen Link and his team at
the University of Dortmund. Their work, on which we build here, focuses on the analysis
of current discourses and their power effects, on uncovering the linguistic and
iconographic means in particular, the collective symbolism by which discourses
work, and on the function of discourse to legitimize and secure dominance in bourgeois-
capitalist modern society (see Link, 1982).

The concept of discourse

According to Link, a discourse can be defined as an institutionalized way of talking that

regulates and reinforces action and thereby exerts power (Link, 1983: 60, authors own
translation). This definition can be illustrated by the image of discourse as a flow of
knowledge throughout time. Different discourses are intimately entangled with each other
and together form the giant milling mass of overall societal discourse. This milling mass
of discourse is growing constantly and exuberantly.
Discourses are not only mere expressions of social practice, but also serve particular
ends, namely the exercise of power. The concept of power, in the Foucauldian sense,
refers to a whole series of particular mechanisms, definable and defined, that seem
capable of inducing behaviours or discourses (Foucault, 1996: 394). Discourses
exercise power in a society because they institutionalize and regulate ways of talking,
thinking and acting.

Aims of critical discourse analysis

CDA aims to disentangle the giant milling mass of discourse, to chart what is said and can
be said in a given society at a given time with regard to its qualitative spectrum (What is
said? How is it said?), and to uncover the techniques through which discursive limits are
extended or narrowed down.
Last but not least, and to live up to its name, CDA aims to question and criticize
discourses. CDA does so in two ways.
Firstly, CDA reveals the contradictions within and between discourses, the limits of
what can be said and done, and the means by which discourse makes particular statements
seem rational and beyond all doubt, even though they are only valid at a certain time and
place.
Secondly, the critical discourse analyst needs to be clear about the fact that her critique
is not situated outside discourse as this would contradict the fundamental assumptions
of discourse analysis. The analyst can and has to take a stand. She can invoke values,
norms, the constitution, universal human rights and so on. But when doing so, she must
never forget that these values, norms, laws and rights have been discursively constructed
too. This kind of critique is not ideological, because unlike ideology it does not make
claims to absolute truth. A critical discourse analyst needs to be clear about the fact that
her position is also the result of discursive processes. With this outlook, she can enter
discursive arguments, where she may defend and possibly modify her position.
Foucauldian CDA here clearly differs from orthodox Marxist positions, which stipulate
that social existence determines consciousness (as discussed under the terms of false
consciousness and the fetishism of commodities in Marx, 1992/1867). Foucault
understands this relationship rather the other way around and emphasizes the materiality
of discourse (cf. Jger, 2008; Link, 1996).
The question of how the production of discourses and the temporarily and spatially
contingent knowledge they contain are connected to mechanisms and institutions of power
has been central for Foucault (as he states in the introduction to the German version of
The History of Sexuality see Foucault, 1983: 8). To clarify the connections between
discourses and power, it is necessary to first examine how discourses and reality are
connected to each other. The following section examines this connection between
discourse and reality. We then examine how discourses relate to power.

Discourses and reality

Discourses do not merely reflect reality. Rather, discourses not only shape but even
enable (social) reality. Without discourses, there would be no (social) reality. Discourses
can thus be understood as material reality sui generis. They are not a second-class
material reality, not less material than real reality, not passive media into which
reality is imprinted. Discourses are fully valid material realities among others (Link,
1992). Therefore, discourse cannot be reduced to a notion of false consciousness or a
distorted view of reality, as is done by some orthodox Marxist approaches to ideology
critique. Discourse is a material reality of its own. It is neither much ado about nothing,
nor a distortion, nor a lie about reality. This characterization of discourse as material
reality implies that discourse theory is a materialist theory. Contrary to a common
misconception, probably based on the fact that discourse analysis deals with language,
discourse theory is not an idealist theory. Discourse theory deals with material realities,
not with mere ideas. Discourses may be conceptualized as societal means of
production. Discourses are not mere ideology; they produce subjects and reality.
Discourses determine reality, though of course always via intervening active subjects
in their social contexts as co-producers and co-agents of discourses. The subjects are
able to do this because they are entangled into discourse and therefore have knowledge at
their disposal. Discourse analysis is therefore not only about the retrospective analysis of
allocations of meaning, but also about the analysis of the ongoing production of reality
through discourse, conveyed by active subjects.
Discourses and power

Based on the above outline of the connection between discourses and reality, we can now
examine the connections between discourses and power. On the one hand, there is the
power of discourse. On the other hand, there is also something like the power over
discourse.
The power of discourse lies in the fact that discourses delineate a range of positive
statements, which are sayable. This means that they simultaneously inhibit a range of other
statements, which are not sayable (cf. Link and Link-Heer, 1990). As flows of knowledge
through time, discourses determine the way in which a society interprets reality and
organizes further discursive and nondiscursive practices (i.e. further talking, thinking and
acting). To put it more precisely, we can distinguish two effects of discourse. Firstly,
discourses form individual and mass consciousness and thereby constitute individual and
collective subjects. Secondly, since consciousness determines action, discourses
determine action. This human action creates materializations. Discourses thus guide the
individual and collective creation of reality.
From a discourse-theoretical point of view, it is thus not the subject who makes the
discourses, but the discourses that make the subject (which may be irritating for those
attached to the idea of the uniqueness of the individual). The subject is of interest not as
an actor, but as a product of discourses. As Foucault argues:

One has to dispense with the constituent subject, to get rid of the subject itself, thats to say, to arrive at an
analysis which can account for the constitution of the subject within a historical framework. And this is what I
would call genealogy, that is, a form of history which can account for the constitution of knowledges, discourses,
domains of objects etc. without having to make reference to a subject which is either transcendental in relation to
the field of events or runs in its empty sameness throughout the course of history. (Foucault, 1980a: 117)

Foucauldian discourse theory, though often wrongly accused of so doing, does not deny
the subject. It aims to analyse the constitution of the subject in its historical and social
context from a diachronic (i.e. longitudinal) and synchronic (i.e. cross-sectional)
perspective: who was conceived of as a subject at a particular point in time? How, and
how come? For example, unlike in the past, women and children in Western society
usually have subject status today. In modern management, in contrast to traditional
bureaucracy, employees are depicted as subjects. They are empowered, with all the
responsibilities that involves. The subject status of foetuses and apes is something that is
hotly debated. Asylum seekers and criminals are often denied a subject status, for
example when they are portrayed as maniacs, dogs or viruses (these are collective
symbols see the section on collective symbols). In a nutshell, Foucauldian discourse
theory contests the existence of an autonomous subject, but that does not mean that it is
against the subject. The active individual is fully involved when it comes to realizing
power relations in practice. The individual thinks, plans, constructs, interacts and
fabricates. The individual also faces the problem of having to prevail, to assert himself,
to find his place in society.
When analysing the power effects of discourse, it is important to distinguish between
the effects of a text and the effects of a discourse. A single text has minimal effects, which
are hardly noticeable and almost impossible to prove. In contrast, a discourse, with its
recurring contents, symbols and strategies, leads to the emergence and solidification of
knowledge and therefore has sustained effects. What is important is not the single text,
the single film, the single photograph and so on, but the constant repetition of statements.
The philologist Victor Klemperer recognized this mechanism as early as the 1930s, when
he observed the language of the Nazis. In his analysis of the language of the Third Reich
(Klemperer, 2000/2006), he contends that fascist language works like the continuous
administration of small doses of arsenic, which unfold their poisonous effect only over
the long term.
With regard to power over discourse, different individuals and groups have different
chances of influence. However, none of them can simply defy dominant discourse, and
none of them alone has full control over discourse. Discourses are supra-individual.
Everybody is co-producing discourse, but no single individual or group controls
discourse or has precisely intended its final result. Discourses take on a life of their own
as they evolve. They transport more knowledge than the single subject is aware of. In
Foucaults words, [p]eople know what they do; they frequently know why they do what
they do; but what they dont know is what what [sic] they do does (personal
communication, quoted in Dreyfus and Rabinow, 1982: 187). The power effects of
discourses should therefore not necessarily be interpreted as the conscious and
manipulative intent of some individual or group. There may be a difference between a
speakers reasons for using a particular discourse, and the social consequences of doing
so (Burr, 2003: 61).
It should be kept in mind, however, that in the long run, powerful politicians and other
groups can accomplish changes in discourse. For example, the Basic Constitutional Law
that governs the right of asylum in Germany was tightened after over ten years of intensive
political and media lobbying. In a similar way, the stance towards deployments abroad of
the German armed forces was changed in a long process. Examples like these show that
certain groups and individuals have more power over discourse than others, for example
because they have privileged access to the media or greater financial resources.
Exclusions inherent to the structure of discourse can thus be amplified by institutional
conditions.
This section can therefore be summarized as follows: discourses exert power because
they transport knowledge on which collective and individual consciousness feeds. This
knowledge is the basis for individual and collective, discursive and nondiscursive
action, which in turn shapes reality.
From discourse to dispositive
Since knowledge is the basis for acting, we can analyse not only discursive practices, but
also nondiscursive practices and materializations, as well as their relationships with each
other. Building on Foucault, we call the interplay between discursive practices,
nondiscursive practices and materializations a dispositive. In this section, this interplay is
examined.
As human beings, we assign meanings to reality. This is how we bring reality into
existence. Of course, this does not mean that human beings are the creators of the raw
matter of material reality. But people shape and use these raw materials. The assignment
of meanings includes very tangible physical acts, such as when a tree is sawn into boards,
or when boards are joined into a table. People moreover learn the conventions of
assigned meanings through language, which helps them to interpret reality in the way it
has previously been interpreted by others. For example, people learn that a certain object
is called a table, and by so doing, they simultaneously learn what a table is good for.
(Different cultures have different objects, which makes translations from one language
into another difficult.) Dispositive analysis examines how such assignments of meaning
create reality. (In contrast, some other disciplines such as the natural sciences and
medicine examine material reality as an objective, natural or biological given.)
Consciousness does not passively reflect reality, but actively takes hold of it. This
works through discourses, which provide the knowledge for shaping reality. If a
discourse withdraws from the reality that has been built on it, or to put it more precisely,
if people withdraw from that discourse, this part of reality becomes meaningless in the
truest sense of the word. It returns to a blank state. If the knowledge assigned to a
particular part of reality changes, this part of reality turns into a different thing. For
example, tramps may move into an abandoned bank and turn it into their flop. A steel mill
may be closed down and may be turned into an amusement park. In these examples,
original meanings are withdrawn and new meanings are assigned.
Foucaults views on the relationship between discursive and nondiscursive realities
have been somewhat ambiguous. In the remainder of this section, we will first provide an
overview of Foucaults conception of this relationship. Then we will propose a different
understanding, which builds on Foucault, but at the same time goes beyond his ideas.
Foucault writes in his Archaeology of Knowledge that discourses can be treated as
practices that systematically form the objects of which they speak (Foucault, 2002: 54).
But Foucault also sees that nondiscursive practices play a decisive role in forming
objects. He talks about discursive relations in this regard. He states that discursive
relations are

in a sense, at the limit of discourse: they offer it objects of which it can speak, or rather [], they determine the
group of relations that discourse must establish in order to speak of this or that object, in order to deal with them,
name them, analyse them, classify them, explain them, etc. (Foucault, 2002: 50f.)
With this statement, Foucault circumnavigates the problem of the relationship between
discourse and reality without really solving it. It remains unclear what he actually means
by objects. Presumably, he does not mean materializations but themes, theories,
statements or other purely discursive objects.
This circumnavigating becomes most apparent in his attempt to define the dispositive.
In a conversation with several psychoanalysts in 1977, he defined the dispositive as
follows:

What Im trying to pick out with this term is, firstly, a thoroughly heterogeneous ensemble consisting of discourses,
institutions, architectural forms, regulatory decisions, laws, administrative measures, scientific statements,
philosophical, moral and philanthropic propositions in short, the said as much as the unsaid. Such are the
elements of the [dispositive]. The [dispositive] itself is the system of relations that can be established between
these elements. (Foucault, 1980b: 194)

He further differentiates:

[ B]etween these elements, whether discursive or nondiscursive, there is a sort of interplay of shifts of position
and modifications of function which can also vary very widely. (ibid.: 195)

He states that by the term dispositive, he understands:

[] a sort of shall we say formation which has as its major function at a given historical moment that of
responding to an urgent need [urgence]. The [dispositive] thus has a dominant strategic function. (ibid.)

He mentions the example of the control of madness, of mental illness, and later of
neurosis which served the function of assimilating a floating population that would
otherwise be burdensome for an economy (ibid.: 195).
In the further course of the conversation, the psychoanalysts begin to nit-pick his
distinction between the discursive and the nondiscursive. It seems that Foucault is in a
tight spot here; the psychoanalysts drive him into a corner. It is noticeable that his
interview partners are getting on his nerves; he is becoming impatient, even annoyed. He
answers them:

Yes, if you like, but it doesnt much matter for my notion of the [dispositive] to be able to say that this is discursive
and that isnt. If you take Gabriels architectural plan for the Military School together with the actual construction
of the School, how is one to say what is discursive and what is institutional? That would only interest me if the
building didnt conform with the plan. But I dont think its very important to be able to make that distinction, given
that my problem isnt a linguistic one. (ibid.: 198)

Foucault thereby liberates himself and us from a linguistics that is not based on
thought and consciousness. He subordinates language and also linguistics to thought and
knowledge. Thereby he basically turns linguistics into a sub-discipline of the cultural
sciences, which deal with the conditions and results of meaningful human activities.
Human activities are meaningful because they are based on thought and consciousness.
After his archaeological endeavours that aimed to reconstruct the development of
knowledge from a purely materialistic perspective, Foucault arrived at the conviction that
talk/text/discourse alone is not what makes the world tick. He invented the dispositive to
permit a better analysis of historic and current reality. With regard to the concept of the
dispositive, the question about the relationship between discourse and dispositive, as
well as between discourse and reality, is of fundamental importance.
As is evident from the above quotes, Foucault clearly assumes a coexistence of
discourse and objects. They are related elements of the dispositive. The dispositive as a
whole comprises the net that is spun between these elements and connects them.
However, Foucault is unable to tell us what concrete, or to put it very concisely, what
empirical relationship connects discourses and objects. The reason for this problem is
that while he is interested in the nature of the relationship between these heterogeneous
elements, he assumes a dualism of discourses and material reality. Foucault did not see
discourses and material reality as interrelated and unable to exist on their own.
According to him, the dispositive assembles various elements that are connected to each
other, and these connections are what constitute the dispositive (see also Balke, 1998;
Deleuze, 1988). Foucault apparently understands these connections as follows: an urgent
need emerges and an existing dispositive becomes precarious. The need for action arises,
and society or its hegemonic forces, confronted with the urgent need, gather all the
elements they can get hold of to deal with it. These may be speeches, people, knives,
cannons, organizations and so on. By these means, they mend the leak, the urgent need
that has arisen (see Balke, 1998; Deleuze, 1988). The elements of the dispositive are
connected by the common purpose they serve, namely the purpose of dealing with an
urgent need. No other inner bond between these elements is apparent in Foucaults
understanding of the dispositive.
In what follows, we propose a different understanding of the bond between the
elements of a dispositive. This bond exists in the form of nondiscursive practices (in
other words, human actions), which connect the subject and the object, symbolic reality
and material reality. In Foucaults definition of the dispositive, nondiscursive practices
are not mentioned explicitly. The ensuing arguments draw on the activity theory
developed by Leontjev (Leontev [sic], 1978), which is based on LevVygotskys book
Thought and Language (Vygotsky, 1986). Activity theory is essentially an approach to
the critique of ideology, which for the purpose in hand is here given a discourse-
theoretical turn.
As mentioned above, human beings are able to assign meaning to objects. What is
more, only by being assigned a meaning does the object turn into an object. For example, I
can assign the meaning table to a piece of wood that I find in the woods. I can eat my
bread from it and put my mug on it. An object that is not assigned any meaning is not an
object. It is totally nondescript, invisible, even nonexistent. I dont see it because I
overlook it. For example, I dont see the bird that the forester sees (forester syndrome).
I may see a red spot, and I may think there is a red spot. It will have the meaning of a
red spot for me. It is beyond my knowledge whether it is a bird, a flower or the recently
dyed hair of Lothar Matthus, who is going for a walk because he was injured playing in
the last football match and therefore cannot train today. Of course, a friend can tell me:
Look, that is Lothar Matthus, and he used to be captain of the German national team.
Then I may say: Yes, okay, I know him, or no, that was definitely a bird or a flower.
The point is that all meaningful reality exists for us because we make it meaningful for
us, or because our ancestors and neighbours assigned meaning to it, and this meaning is
still valid for us. It is like King Midas with his gold: everything he touched turned into
gold. Similarly, everything that human beings assign meaning to becomes a particular kind
of reality, according to the meaning it was assigned. Ernesto Laclau expressed this
connection elegantly:

By discursive I do not mean that which refers to text narrowly defined, but to the ensemble of the phenomena
in and through which social production of meaning takes place, an ensemble which constitutes a society as such.
The discursive is not, therefore, being conceived as a level nor even as a dimension of the social, but rather as
being co-extensive with the social as such. This means that the discursive does not constitute a superstructure
(since it is the very condition of all social practice) or, more precisely, that all social practice constitutes itself as
such insofar as it produces meaning. Because there is nothing specifically social which is constituted outside the
discursive, it is clear that the nondiscursive is not opposed to the discursive as if it were a matter of two separate
levels. History and society are an infinite text. (Laclau, 1980: 87)

The question is: why, when, how and under what conditions do I assign meaning to
objects? How is the gap between discourse and reality closed? According to Leontjevs
activity theory, meaning is assigned to an object if I derive a motive from a particular
need and therefore aim to achieve a particular aim, for which I use actions and raw
materials. In other words, meaning is assigned to an object through work. The products
thus created can be articles of daily use, but also new thoughts and plans, which again
may give rise to new activities and products. It seems strange that Foucault, who had a
background in psychology, did not know activity theory, which is based on the
materialistic psychology of the early 1930s. Maybe he rejected it because he found that it
centred too much on the subject. Yet the approach is fruitful because it deals with the
connection of subject and object, society and objective reality, through human activity.
Foucault also overlooks the fact that the materializations of work are part of reality.
People create these through their nondiscursive practices, as they build houses, benches,
banks and so on, which exist only as long as they are embedded into discourses. For
example, a bank, as an element of the dispositive of capital, stops functioning as a bank if
it is no longer supported by discourse. It becomes meaningless, reduced to nothing but
raw matter (which, if called this way, again takes on a certain meaning). Alternatively, it
may be discursified anew into another object. Tramps may sleep in the former bank and
thereby assign another meaning to it, turning it into their flop. Foucault also sees this and
writes:

[I]t is not the objects that remain constant, nor the domain that they form; it is not even their point of emergence
or their mode of characterization; but the relation between the surfaces on which they appear, on which they can
be delimited, on which they can be analysed and specified. (Foucault, 2002: 52)
Or to put it more simply: if the discourse changes, the object does not only change its
meaning, it turns into a different object. It loses its previous identity. This may happen
abruptly or as the result of a long process that impalpably but thoroughly changes
everything.
Foucault does not want to define objects with reference to the ground, the foundation
of things (ibid.: 53), but to define objects by relating them to the body of rules that
enable them to form as objects of a discourse and thus constitute the conditions of their
historical appearance (ibid.). However, he gets stuck at this point because he does not
conceptualize subject and object, society and discourse, as connected by activity, i.e. by
nondiscursive practices. Discursive practices remain verbal for him, strictly separated
from nondiscursive practices. With this separation of intellectual activity and
(unintellectual?) physical activity, he shows himself as a product of his times and his
origins, where the bourgeoisie highly valued mental work and believed manual work to
be completely unintellectual. Foucault knows that the signs do more than designate things,
and he states that [i]t is this more that renders them irreducible to the language (langue)
and to speech (ibid.: 54). He wants to reveal and describe (ibid.) this more, but does
not succeed fully. He is not really able to grasp this more. In this chapter, we argue that
this more is the knowledge that enables the transformation of verbally articulated
knowledge into material objects: knowledge about statics, materials, tools, routines and
the like that enters every kind of physical work. However, it is rarely verbalized, or even
cannot be verbalized (i.e. tacit knowledge). For example, a steelworker at a blast furnace
sees when the steel is ready or what ingredients are still missing. However, it may be
impossible to teach this knowledge to somebody else just through verbalization. One may
have to watch and try for oneself. In a way, the knowledge is in the practices.
We may therefore say that reality is meaningful, that reality exists in the way it does,
only insofar as it is assigned meaning by people, who are themselves entangled into and
constituted by discourses. If people no longer assign the same meaning to an object, the
object changes or loses its meaning. This meaning may then at most be reconstructed as a
former meaning that has mixed with other meanings or has ceased to be valid. Even if we
just watch the night sky and see constellations of stars there, we see them as a result of a
discourse. We see the constellations because we have learnt to see them. To assign
meaning is not a noncommittal, merely symbolic act. To assign meaning is to animate
whatever one comes across, to re-shape and change. For example, from the collective
symbolism used with regard to immigrants, it is apparent that many people have learnt to
assign negative meanings to immigrants, and now actually perceive them as floods that
need to be held back, or even as lice or pigs that should be crushed or slaughtered.
Bernhard Waldenfels (1991) confirms this criticism of Foucault, which at the same
time is inspired by Foucault. He writes that it is:

unclear how [Foucault] draws the line between discursive and nondiscursive practices and how he bridges this
line. It even remains unclear if this line is drawn at all. I think that in a way, Foucault drove himself into a blind
alley when he first conceptualized the formations of the order of history as orders of knowledge (epistemes), and
 then conceptualized them as orders of speech (discourses), instead of assuming an order that is shared by all
behavioural registers of people, i.e. by their speech and actions (!) but also their gazes, their bodily conventions,
their erotic relationships, their technical handlings, their economic and political decisions, their expression in art and
religion, etc. It is not clear why any of these areas should not have the functionality that Foucault developed based
on speech alone. (Waldenfels, 1991: 291, authors own translation)

Waldenfels notes that Foucault himself crossed this border at several points. In the
Archaeology of Knowledge, Foucault (2002) mentions the discourse of the painter, who
speaks without words. Foucault also mentions political discourse, where for example
an unexpected revolution may take place, which cannot be traced back to a revolutionary
situation or a revolutionary conscience. From a superficial point of view, this unexpected
revolution is not based on knowledge, while on closer inspection, it is very much based
on knowledge. As often happened, Foucault preferred to do bricolage2 here (cf.
Waldenfels, 1991: 291).
In this text, we continue the bricolage, pick up Foucaults toolbox of theoretical and
practical instruments, and develop his ideas further. So far, we have done so in two
regards. Firstly, Foucaults notion of discourse, which is too bound to the verbal, is
shifted one step backwards, namely into human thinking and knowledge, or in other
words, into consciousness. This is where the contents of thought (including emotions and
perceptions) are located, which are the basis for shaping reality through work.
Secondly, and by the same token, activity theory is made available for discourse
theory. Activity theory explains how subjects and objects are connected to each other.
Foucault mainly saw discourse as somehow connected to reality. Drawing on Leontjev,
we identify the subject as the missing link that connects discourses to material reality.
Subjects do this in all their activities. The way in which those activities become effective
has not been intended in exactly this way by any individual or single group. Nevertheless,
it is human consciousness and physical strength that shape reality. Everything in human
consciousness is discursive, i.e. constituted by knowledge. Moreover, subjects
continuously draw on tacit knowledge. Tacit knowledge is passed on in nondiscursive
practices and materializations, and the researcher can aim to reconstruct tacit knowledge
from both of them.
To summarize the problems discussed in this section, many of the difficulties in
analysing dispositives stem from an insufficient conceptualization of discursive practices,
nondiscursive practices and materializations (see also Bublitz, 1999: 82115). Many of
these problems can be solved by drawing on Leontjevs and others understanding of
materializations and nondiscursive practices as realizations of discourse, i.e. knowledge
(for further discussion see, for example, Jger, 2001a). The following section provides
some ideas on how discourses and dispositives can be analysed systematically.

Methods for analysing discourses and dispositives

To recapitulate briefly our initial definition, a discourse can be characterized as an
institutionalized way of talking that regulates and reinforces action and thereby exerts
power (Link, 1983: 60, authors own translation), as illustrated by the image of a flow of
knowledge throughout time (see Jger, 2004). Together, all discourses form the giant
milling mass of overall societal discourse.
The theoretical considerations introduced in the early sections of this chapter are the
basis for the development of concepts and methods that facilitate analysis. A range of
these concepts and methods is outlined in this section. The scope of this chapter does not
allow for detailed methodological justifications of each of these concepts and methods,
but such explanations can be found in the book on methods of CDA by Jger (2004).
Methods of CDA also include linguistic concepts (e.g. figurativeness, vocabulary,
pronominal structure, argumentation) which can be used to examine the more subtle
workings of texts. However, these linguistic instruments are not described in detail here,
as good explanations can be found in works on style and grammar. Linguistic concepts fill
only one slot in the toolbox of discourse analysis. Depending on the research question
and subject matter, various other tools can be added to the toolbox. In any case, certain
methods are part of a standard repertoire. These are described in the remainder of this
section, putting special emphasis on discourse theory and activity theory as outlined
above.

The structure of discourse

The following suggestions on terminology aim to provide some help in making the
structure of discourses more transparent and amenable to analysis.
Special discourses and interdiscourse
A basic distinction can be drawn between special discourses and interdiscourse. Special
discourses are discourses in the sciences, while interdiscourse refers to all non-scientific
discourses. Elements of special discourses continuously feed into interdiscourse.
Discourse strands
In general societal discourse, a great variety of topics arise. Flows of discourse that
centre on a common topic are called discourse strands. Each discourse strand comprises
several subtopics, which again can be summarized into groups of subtopics.
The concept of discourse strands is similar to the one of discourses. The difference
is that discourse is the more abstract concept, located at the level of statements
(noncs). Discourse strands, in contrast, are conceived of at the level of concrete
utterances (nonciations) or performances located on the surface of texts (cf. Foucault,
2002).
Every discourse strand has a diachronic and a synchronic dimension. A synchronic
analysis of a discourse strand examines the finite spectrum of what is said and sayable at
a particular point in time. A synchronic analysis can cut through a discourse strand at
various points in time, for example at particular discursive events. By comparing these
synchronic cuts, it provides insights into the changes and continuities of discourse strands
over time.
In a way, a synchronic cut through a discourse strand is always also a diachronic one.
This is because each topic has a genesis, a historical a priori. When analysing a topic, the
analyst has to keep an eye on its history. To identify the knowledge of a society on a
topic, the analyst has to reconstruct the genesis of this topic. Foucault has undertaken
several attempts to do so, not only with regard to the sciences, but also with regard to
everyday life and institutions (e.g. the hospital, the prison).

Discursive limits and techniques for extending them or narrowing them down
Each discourse delineates a range of statements that are sayable and thereby inhibits a
range of other statements, which are not sayable (cf. Link and Link-Heer, 1990). The
borders to what is not sayable are called discursive limits.
Through the use of certain rhetorical strategies, discursive limits can be extended or
narrowed down. Such strategies for example include direct prescriptions, relativizations,
defamations, allusions and implicatures. Discourse analysis examines these strategies in
their own right, and also uses them as analytic clues to identify discursive limits: if
tricks are used, this is an indicator that certain statements cannot be said directly
without risking negative sanctions.
Discourse fragments
Each discourse strand consists of a multitude of elements that are traditionally called
texts. We prefer the term discourse fragment, because one text may touch on various
topics and thus contain various discourse fragments. A discourse fragment therefore
refers to a text or part of a text that deals with a particular topic (for example, the topic of
immigration). Or to put it the other way around, various discourse fragments on the same
topic form a discourse strand.
Entanglements of discourse strands
A text usually refers to various topics and therefore to various discourse strands. In other
words, it usually contains fragments from various discourse strands. These discourse
strands are usually entangled with each other. An entanglement of discourse strands can
take the form of one text addressing various topics to equal degrees, or of one text
addressing mainly one topic and referring to other topics only in passing.
A statement where several discourses are entangled is called a discursive knot. For
example, in the statement integrating immigrants into our society costs a lot of money,
the discourse strand of immigration is entangled with the discourse strand of the economy.
In the statement in [insert any Islamic country here], they still live in a patriarchal
society, the discourse strand of immigration is entangled with the discourse strand of
women.
Two discourse strands can be entangled more or less intensively. For example, in
everyday discourse in Germany, the discourse strand of immigration is intensively
entangled with the discourse strand of women, as sexist attitudes and behaviours are
attributed to immigrants (see Jger, 1996).
Collective symbols
An important means of linking up discourse strands is the use of collective symbols.
Collective symbols are cultural stereotypes, also called topoi, which are handed
down and used collectively (Drews et al., 1985: 265). They are known to all members of
a society. They provide the repertoire of images from which we construct a picture of
reality for ourselves. Through collective symbols we interpret reality, and have reality
interpreted for us, especially by the media.
An important technique for connecting collective symbols is catachreses (also called
image fractures). Catachreses establish connections between statements, link up spheres
of experience, bridge contradictions and increase plausibility. Thereby, catachreses
amplify the power of discourse. An example of a catachresis is the statement the
locomotive of progress can be slowed down by floods of immigrants. Here, the symbols
of the locomotive (meaning progress) and floods (meaning a threat from the outside) are
derived from different sources of images. The first one is taken from traffic and the
second from nature. With a catachresis, the images are connected.
Discourse planes and sectors
Different discourse strands operate on different discourse planes, such as the sciences,
politics, the media, education, everyday life, business, administration and so forth. These
discourse planes can be characterized as social locations from which speaking takes
place.
Discourse planes influence each other and relate to each other. For example, on the
media plane, discourse fragments from scientific specialist discourse or political
discourse are taken up. The media also take up everyday discourse, bundle it, bring it to
the point, or especially in the case of the yellow press spice it up with sensational and
populist claims. In this way, the media regulate everyday thinking and exert a
considerable influence on what is and what can be done in politics and everyday life. For
example, the larger-than-life image of Jrg Haider would hardly have come about without
the help of media reports that normalized right-wing populism.
A discourse plane consists of various sectors. For example, womens magazines, TV
news broadcasts and newspapers are different sectors of the discourse plane of the
media.
A discourse plane is tightly interwoven in itself. For example, on the discourse plane
of the media, leading media may also repeat and build on contents that have already been
brought up in other media. It is therefore all the more justified to talk about the media
discourse plane, which especially with regard to the dominant media in a society can
be considered as integrated in its major aspects.
Discursive events and discursive context
All events are rooted in discourse. However, an event only counts as a discursive event if
it appears on the discourse planes of politics and the media intensively, extensively and
for a prolonged period of time.
A major reason why it is important to identify discursive events is that they influence
the development of discourse. For example, the Three Mile Island nuclear accident near
Harrisburg was comparable to the one in Chernobyl. But while the Three Mile Island
accident was covered up for years, the Chernobyl accident was a major media event and
influenced global politics. Whether an event, such as a nuclear accident, becomes a
discursive event or not depends on the power constellations at work in politics and the
media.
Discourse analysis can examine whether an event becomes a discursive event or not. If
it becomes a discursive event, it influences the further development of discourse: the
Chernobyl disaster contributed to a changing policy towards nuclear power in Germany.
Germany is now, albeit hesitantly, going to phase out nuclear power. Environmentalist
discourse, which had been developing for some time, could hardly have achieved this on
its own. At the same time, a discursive event like the Chernobyl disaster can influence the
whole discourse about new technologies, for example by drawing attention to the need to
develop alternative energy sources. Another example for a discursive event is the success
of the FP (Freedom Party of Austria) in the 1999 Austrian National Parliamentary
Elections, and the FPs ensuing participation in government. While the FPs success in
the elections attracted considerable media coverage, the ensuing participation of the FP
(and indirectly Jrg Haider) in government triggered a far greater worldwide response,
and became a discursive mega event that kept the European and US press in suspense for
months. It had an effect on the discourses of the extreme right in countries in Europe and
beyond.
Another reason why the identification of discursive events is important for the analysis
of discourse strands is that they outline the discursive context that a discourse strand
relates to. For example, a synchronic (i.e. cross-sectional) analysis of a discourse strand
can be enriched with diachronic (i.e. longitudinal) elements by adding a chronicle of the
discursive events belonging to it. Such historic references can be very helpful for
synchronic analyses of discourse strands (as, for example, demonstrated by Caborn,
1999).
Discourse positions
A discourse position describes the ideological position from which subjects, including
individuals, groups and institutions, participate in and evaluate discourse. Also, the
media take up discourse positions, which become evident in their reporting. (As noted
above, subject status is nothing natural and obvious, but something that in itself needs to
be established through discourse.)
Subjects develop a discourse position because they are enmeshed in various
discourses. They are exposed to discourses and work them into a specific ideological
position or worldview in the course of their life. This relationship also works the other
way around. Discursive positions contribute to and reproduce the discursive
enmeshments of subjects (Jger, 1996: 47).
Discourse positions can be identified through discourse analysis. But a rough outline of
discourse positions is also part of peoples everyday knowledge. People know roughly
which politicians and newspapers tend towards the left, the right or the centre. Everyday
self-descriptions of ones discourse position, however, should be taken with a grain of
salt. For example, newspapers often describe themselves as independent and
impartial, while from a discourse-theoretical perspective, this is an impossibility.
The discourse positions of subjects may vary widely. For example, with regard to the
discourse strand of the economy, many subjects take up a neoliberal discourse position
and favour privatization, free trade, low taxes, fiscal policy discipline and so on. Others,
in contrast, reject neoliberalism and take up Keynesianism or something even more
unorthodox as their discourse position.
Discourse positions are homogeneous only in their core and become diffuse with
regard to less central issues. For example, subjects who embrace the hegemonic
discourse position of neoliberalism agree that it is in principle right and important to
reduce the nations budget deficit. They do not question the current economic system.
However, they may have differing views on the best way to reduce the budget deficit.
Within a dominant discourse, discourse positions are fairly homogeneous, which itself
is already an effect of dominant discourse. Dissenting discourse positions often belong to
complete counter-discourses (e.g. a fundamental questioning of the current economic
system may not arise from economic discourse, but from ecology or ethics). However,
these counter-discourses can pick up arguments from dominant discourse and subvert
their meaning. For example, some people interpret the widespread saying that time is
money with an anti-capitalist twist, as when saying If time was money, Id already be
rich.

Overall societal discourse and global discourse

All the entangled discourse strands in a society together form the overall societal
discourse. A society is never totally homogeneous but consists of different subcultures. In
Germany, since reunification in 1989, overall societal discourse has become
ideologically more homogeneous, and it seems unlikely that this is going to change easily
(Teubert, 1999).
The overall societal discourse of a society, in turn, is part of global discourse. Even
though global discourse is very heterogeneous, it is probably safe to say that global
discourse has also become more homogeneous and has switched from a confrontation of
the West against communism to a confrontation of the West against the Islamic World.
Overall societal discourse is a very complex network. Discourse analysis aims to
disentangle this net. The usual procedure is to first identify single discourse strands on
single discourse planes (for example, the discourse strand of immigration on the
discourse plane of the media). Subsequently, analyses of this discourse strand on further
discourse planes, such as politics or everyday communication, can be added. At the end
of such analyses, the question is usually how the different discourse planes of a discourse
strand relate to each other. For example, one may examine whether and how the political
discourse plane is linked to the discourse planes of the media and everyday
communication, whether and how the media discourse plane influences the discourse
plane of everyday communication, and so on.

The history, present, and future of discourse strands

Discourse strands have a history, a present and a future. In order to identify the changes,
ruptures, ebbings and recurrences of a discourse strand, it is necessary to analyse longer
periods of time. To put it into Foucaults words, an archaeology of knowledge or a
genealogy is needed. On the basis of such an analysis, even prognoses about discourse
can be undertaken. These can take the form of scenarios based on different future
discursive events.
Of course, an analysis of the history, present and future of overall societal or even
global discourse is an enormous endeavour and can only be tackled in the form of many
single projects. Such single projects, however, are extremely helpful because they create
reliable knowledge about certain sub-zones of overall societal discourse. This scientific
knowledge can be the basis for a change of everyday, political and media knowledge, and
can change behaviours and policies. Work on the discursive plane of science can thus
influence the further development of a particular discourse strand.

On the completeness of discourse analyses

A discourse analysis fully captures the qualitative range of what can be said and how it is
said in one or more discourse strands. It is complete if further analysis leads to no further
new findings. Social scientists who mainly work with large amounts of quantitative data
will be surprised to learn that in discourse analysis, a relatively small amount of
qualitative data suffices to reach this point. The arguments and contents that can be read
or heard about a particular topic (e.g. immigration) at a particular time in a particular
social location are amazingly limited (often in both senses of the word). With regard to
methodology, this means that the analyst continues to analyse new materials until he
notices that arguments begin to repeat themselves. If this is the case, completeness (in the
sense of theoretical saturation) has been achieved.
While qualitative analysis is the bedrock of discourse analysis, quantitative analyses
can also be interesting. The analyst can examine with what frequency particular
statements occur. In this way, focal issues in discourse strands, or statements that have the
character of slogans and are therefore accompanied by a bulk of judgements and
prejudices, can be identified. If a statement occurs very frequently, it has sustained effects
and strongly solidifies a particular knowledge. In diachronic analysis, frequencies can
also be used to identify trends. However, for the explanatory power of a discourse
analysis, the qualitative aspect is of greater importance than the quantitative.

A little toolbox for discourse analysis

In this section, a brief summary of our toolbox for discourse analysis is presented. As
noted above, within the scope of this chapter, we cannot provide detailed methodological
justifications for each of the tools, but these can be found in the volume by Jger (2004).
In our own research projects, we use short handouts like the following as memory aids or
checklists when first dealing with materials.
This outline deals with the practical procedures for subjecting empirically obtained
materials to discourse analysis. In an actual project, these elements are supplemented
with a clarification of the theoretical foundations and methodology used. The tools
presented in the remainder of this section are discussed using the example of a
hypothetical research project on the discourse strand of stem cell research in newspapers,
or to put it more concisely, in the newspaper sector of the discourse plane of the media.
Choosing a subject matter
The first step in a discourse analysis project is usually to choose a subject matter. In the
project report (usually in the introduction), a rationale for the project and its subject
matter has to be given.
It needs to be kept in mind that the relationship between a phenomenon of interest and
particular discourse strands is often not straightforward because a phenomenon may
permeate many discourse strands. For example, in a research project that aims to examine
how racism permeates the media, the researcher has to decide which discourse strand(s)
to focus on. To make the choice, the researcher has to have an initial concept of racism in
mind. (This concept may be developed further in the course of the analysis.) Theoretical
concepts are always debatable, and the researcher needs to clarify and justify which
concept he is working with.3 Equipped with this concept, the researcher can think about
promising discourse strands where racism may be found. In the case of racism, it is, for
example, the discourse strand of immigration, refugees and asylum-seeking. Of course, the
discourse strand of immigration could also be interpreted in the light of other research
interests. To choose a subject matter means to choose a phenomenon of interest and a
discourse strand which will be examined. This discourse strand delineates the scope of
materials for analysis.

Choosing a discourse plane and a sector and characterizing them

Typically, it will be necessary, at least initially, to confine the analysis to one discourse
plane (for example, the media). When examining a discourse plane, analysis may cover
one or several sectors of this plane (for example, the sector of newspapers). The choice
of sector needs to be justified. For example, a sector may be exemplary for how an issue
is dealt with in the media, or a sector may not have previously been examined in any
research project. In the latter case, of course, a review of previous research should
summarize the findings from an analysis of other sectors.
In some cases, it may be possible to examine several discourse planes at once. The
analysis of interactions of several discourse planes in the regulation of mass
consciousness is extremely interesting, but also very time-consuming. To achieve this
task, it is necessary to base the analysis on well-justified examples of sectors of these
discourse planes and instances of their interaction. The task becomes even more complex
if entanglements of discourse strands are also considered.
Accessing and preparing the materials
As a next step, the concrete corpus for analysis needs to be delineated. For example,
when analysing newspapers, the particular newspapers and the time periods under
consideration need to be decided on. Again, these choices have to be justified.
As a preparation for analysis, a general characterization of the newspaper(s) in
question needs to be provided: what is their political orientation, who are their readers,
what is their circulation, and so on?
Stem cell research is a topic that continually crops up in newspapers but is often not
treated very extensively. In this case, it makes sense to use different newspapers from the
period of a whole year, because even if the newspapers are read very carefully, the
whole qualitative variety of the discourse strand becomes apparent only in those
materials that cover a longer period of time.
In contrast, a project that examines the portrayal of women in pop songs could
probably rely on a few exemplary songs (though this would have to be demonstrated in
the particular project).
Analysis
The following sequence of steps can be modified and does not have to be adhered to
dogmatically. The steps usually have to be gone through several times. In these cycles of
analysis, connections between different levels of analysis are discovered, interpretations
are developed and weak arguments are discarded.

A structural analysis of the discourse strand

The detailed steps for a structural analysis of a discourse strand are as follows:

1. A list of all articles of relevance for the discourse strand is compiled. This list
should include bibliographical information, notes about topics covered in the article,
the literary genre, any special characteristics and the section in which the article
appears.
2. Structural analysis should roughly capture the characteristics of articles on particular
aspects of interest, such as any illustrations, the layout, the use of collective
symbols, the argumentation, the vocabulary and so on, and identify which forms are
typical for the newspaper. This outline will be needed later to identify typical
articles for the detailed analysis (see the next subsection).
3. A discourse strand encompasses various subtopics. These are first identified and
then summarized into groups. For example, in the case of the discourse strand of
stem cell research, subtopics may be summarized into groups such as the legal
implications of stem cell research, the benefits of stem cell research, the
technical procedures of stem cell research, the ethical problems of stem cell
research, the costs of stem cell research and so on. The development of groups of
subtopics is an iterative process, which should lead to a good compromise between
parsimony and discriminatory power.
4. The next step is to examine with what frequency particular groups of subtopics
appear. Which ones are focused on and which ones are neglected? Are there any
subtopics that are conspicuous by their absence?
5. If the analysis is diachronic, it will also examine how subtopics are distributed over
the course of time. Are some subtopics particularly frequent at particular times?
How does this relate to discursive events?
6. Discursive entanglements are then identified. For example, the discourse strand of
stem cell research is entangled with the discourse strands of ethics, business and
medicine.

The findings from these steps of analysis are combined and interpreted together. Thereby,
a characterization of the newspapers discourse position begins to emerge. For example,
does the newspaper perceive stem cell research positively or negatively?
The structural analysis of a discourse strand can and should already yield ideas for the
ensuing detailed analysis of typical discourse fragments (see the next subsection) and for
the final synoptic analysis (see the subsection on synoptic analysis). These ideas should
be written down immediately and marked accordingly.

Detailed analysis of typical discourse fragments

To identify the fine detail within the newspapers discourse position and to assess the
effects of this discourse on readers, certain discourse fragments are subjected to detailed
analysis. Discourse fragments that are typical of the particular newspaper are selected for
this purpose. Criteria for typicality are, for example, typical illustrations, typical use of
collective symbols, typical argumentation, typical vocabulary and so on. The typical
forms of these aspects have been identified in the preceding structural analysis.
To select typical discourse fragments, the researcher can proceed in several steps and
rate the articles according to defined criteria. To ensure that the selection is
intersubjectively plausible, several researchers can engage in this rating. The articles that
score highest on typicality are then subjected to detailed analysis. If time restraints
require it, detailed analysis may also be confined to one article only.
If structural analysis has shown that the discourse strand is very heterogeneous, and if
no single homogeneous discourse position can be discerned, the researcher can also
address several typicalities, i.e. several kinds of typical articles.
The procedures for selecting typical articles should be systematic and transparent, but
not mechanical. What is an appropriate procedure depends on the concrete research
project and the discourse strand in question.
The detailed analysis of typical discourse fragments should cover the following
aspects:

Context

Why was this article selected? Why is this article typical?

1. Who is the author? What is her position and status within the newspaper? What
are her special areas of coverage, and so on?
What was the occasion for the article?
In which section of the newspaper does the article appear?

Surface of the text

What is the layout like? What kinds of pictures or graphs accompany the text?
What are the headings and subheadings?
2. How is the article structured into units of meaning?
What topics are touched upon in the article? (In other words, what discourse
 strands is the article a fragment of?)
How do these topics relate to each other and overlap (entanglements of discourse
strands)?

Rhetorical means

What kind and form of argumentation does the article follow? What
argumentation strategy is used?
What logic underlies the composition of the article?
What implications and allusions does the article contain?
What collective symbolism is used (linguistic and graphic, involving, for
3.
example, statistics, photographs, pictures, caricatures, etc.)?
What idioms, sayings and clichs are used?
What are the vocabulary and style?
What actors are mentioned, and how are they portrayed (persons, pronouns
used)?
What references are made (e.g. references to science, information about the
sources of knowledge used)?

Content and ideological statements

What concept of humankind does the article presuppose and convey?

4. What concept of society does the article presuppose and convey?
What concept of (for example) technology does the article presuppose and
convey?
What perspective regarding the future does the article give?

5. Other peculiarities of the article

6. Discourse position and overall message of the article.

In analysing each of these aspects, the researcher has to ask herself what this peculiarity
of the article means, what it implies. For example, what does it mean that a particular
image accompanies this text? What effect does this image create? Each of these
interpretations remains open to revisions. At the end of the detailed analysis, the
interpretations of single aspects are combined into a total interpretation of the article.
Usually, the interpretations of the single aspects fit together like the pieces of a jigsaw
puzzle and form a unitary picture. If one aspect stands out, it is often due to special
circumstances, such as when a photo or a headline has not been provided by the author
but by the editor, who had other purposes in mind, such as spicing up the article. Such
discrepancies also provide important insights into the newspapers discourse position.
Together with the findings from structural analysis, the findings from detailed analysis
form the basis for synoptic analysis.
Synoptic analysis
In synoptic analysis, a final assessment of the newspapers discourse position is made.
For this purpose, the findings from structural analysis and detailed analysis are
interpreted in relation and comparison to each other.
In conclusion, it needs to be emphasized that CDA is not a rigid formula that can be
followed mechanically to produce results. Depending on the research question and the
type of materials used, different procedures are appropriate. The researcher has to keep
this in mind when developing her analytical strategy. In doing so, she needs to be flexible
and imaginative. Therefore, in a way, the best way to learn CDA is to do CDA. An article
like the present one only provides initial insights into the wide range of possibilities. We
therefore close our little toolbox with one more quotation from Foucault, talking about
methodology:

If you want an image, think of a network of scaffolding that functions as a point of relay between a project being
concluded and a new one. Thus I dont construct a general method of definitive value for myself or for others.
What I write does not prescribe anything, neither to myself nor to others. At most, its character is instrumental
and visionary or dream-like. (Foucault, 1991: 29)

Initial thoughts on analysing dispositives

Discourses do not exist independently; they are elements of dispositives. A dispositive,

as outlined above, is the constantly evolving synthesis of knowledge that is built into
language, action and materializations. The concept of the dispositive can be visualized as
a triangle, or rather a rotating circle, with three transit points:

discursive practices (language and thought)

nondiscursive practices (action)
materializations (which are created through nondiscursive practices).
FIGURE 2.1 Dispositives

The visibility of materializations is upheld through discursive and nondiscursive

practices. Figure 2.1, which is a graphic simplification of the concept, summarizes this
relationship.
Though a dispositive has certain durability, it is also subject to historical changes and
is constantly influenced by other dispositives. A synchronic analysis serves to identify the
current state of a dispositive. A particular discursive practice, nondiscursive practice or
materialization can be relevant for various dispositives. For example, the dispositive of
traffic encompasses streets, cars, traffic jams, drivers, traffic signs and so on. But
traffic is also an economic problem that creates costs and affects business. Traffic is
therefore embedded into the economic dispositive. The economic dispositive, in turn, is
embedded into the political dispositive. In a society, dispositives overlap and are
entangled with each other. These entanglements may be what unifies a society.
A dispositive analysis has to include the following steps:

1. Reconstructing the knowledge that is built into discursive practices (through

discourse analysis, as described above). This analysis is the basis for the further
steps in a dispositive analysis. It already creates an awareness of important aspects
of the dispositive, such as uncharted territories in discourse, significant
materializations and so on.
2. Reconstructing the knowledge that is built into nondiscursive practices.
3. Reconstructing the knowledge that is built into materialization, and the nondiscursive
practices that have created these materializations.

Reconstructing this knowledge usually results in texts. A dispositive analysis thus

translates nondiscursive and materialized knowledge into discursive knowledge.
A dispositive analysis should also consider the form in which the examined knowledge
occurs: is the knowledge manifestly apparent? Or is it implicit, for example hidden in
implicatures? Into what arguments is the knowledge packed? And so on. It should be
noted again that the concept of knowledge is here a very broad one, including not only
cognitions but also emotions. Knowledge includes all the contents that make up a human
consciousness.
Since the analysis of discursive elements of a dispositive has already been discussed
extensively above, the remainder of this section will focus on reconstructing the
knowledge that is built into nondiscursive practices and materializations.
Knowledge in nondiscursive practices
Nondiscursive practices, i.e. actions, can be observed and described. For a dispositive
analysis, the task is to reconstruct the knowledge that enables and accompanies these
practices.
For example, the analyst can observe a man who crosses a street and walks into a
bakery, where he buys a loaf of bread. The analysts task is now to find out what this man
knows and wants. The man knows that he has to go to a certain place to be able to buy
bread. He knows that for this purpose he has to dress in a certain way (e.g. put on shoes
and a coat). He knows that when crossing a road, he has to pay attention to the traffic and
observe the traffic rules. Moreover, he knows that the bakery is located in a particular
place, or how to look for a bakery. He knows that in a bakery he can buy bread, and that
he needs money for that. The simple act of buying bread is thus already based on a
considerable amount of knowledge, and this analysis only gives a small hint of it.
The following is a more complex example. Suppose the analyst observes a man who
has dug a hole at the side of a road, and is now standing in this hole and manipulating a
large pipe. To reconstruct the knowledge built into this practice, the analyst has to share
this knowledge and understand what the man is doing. Suppose that to a large extent she is
lacking this knowledge.
There are basically three things she can do to gain an understanding of what the man is
doing.
Firstly, the researcher can of course draw on existing texts. For example, she can
consult previous research, but also more mundane documents such as practitioners
literature, instruction sheets, or field manuals.
Secondly, the analyst can ask the man what he is doing. In research methodology, this is
called an ethnographic interview (see, for example, Spradley, 1979). When the
researcher asks the man what he is doing, the man may answer: I am repairing a burst
pipe. With this information, she already understands better what he is doing. Next, she
may ask him: Why are you doing that? He may answer something like: Because the
pipe has burst, or Thats my job, or I need to earn money somehow, and so on. The
knowledge built into his activity is thus fairly complex and can be followed up to the
economic practice of dependent wage labour.
A large part of knowledge is only available to people in their practices (tacit
knowledge), and people cannot easily explicate it in talk. In other words, people will
know more than they can tell. As a third option, the researcher can therefore rely on
participant observation (see, for example, Agar, 2002; Emerson et al., 1995; Hammersley
and Atkinson, 2007; Spradley, 1980) to learn about this implicit knowledge and make it
explicit in her research. In the extreme case, the researcher may herself learn to dig holes
and mend burst pipes (a fascinating example of such work is Wacquants [2004] study of
boxing).
Knowledge in materializations
When an analyst observes an object, such as a house, a church or a bicycle, obviously he
cannot ask this object about its knowledge directly. However, there are indirect ways of
reconstructing the knowledge built into materializations. Methodological guidelines for
doing so can, for example, be found in multimodal discourse analysis (Van Leeuwen,
2005) and artefact analysis (Froschauer, 2002; Lueger, 2004; for an example of empirical
work that combines multimodal discourse analysis and artefact analysis, see Maier,
forthcoming).
To analyse materializations, the researcher has to rely on his own and his fellow
researchers background knowledge. In addition, he should aim to extend this knowledge
by drawing on the pertinent literature, and by questioning users, producers and other
persons who are experts on the materialization in question.
Artefact analysis, as developed by Lueger (2004) and Froschauer (2002), suggests that
one of the first steps in analysing a materialization is to deconstruct the materialization by
dividing it into its constituent parts and transcribing it into a field protocol. The material
object is thus transformed into a text. Here, another problem arises, which incidentally
also applies to the field notes and observation protocols produced in the participant
observation of nondiscursive practices. The field protocols written by the researcher are
not neutral. Like any text, they pursue particular interests, and in the ideal case, this
interest should be to answer the research question.
In some cases, the researcher may even be able to draw on previous research that has
already discursified the materialization in question as, for example, Caborn (1999, 2006)
has done with regard to state architecture in Berlin.
It should be emphasized again that the meaning of a materialization is not fixed. The
knowledge built into a materialization today may be different from the knowledge it
contained in the past. Legends might have formed around it, and meanings may have
changed. Moreover, a materialization may have different meanings for members of
different cultures (as well as cultures understood in the broad sense of the term). A good
case in point is the Aztec crown of feathers, which is exhibited in the Anthropological
Museum in Vienna. In pre-Columbian times, it was a ritual headdress, worn by priests or
even by the Aztec emperor Montezuma. For the Spanish conquistador Hernn Corts, it
was a treasure which he stole. The Habsburg emperors bought it as an exotic curiosity. In
todays Anthropological Museum, it is an exhibit of scientific value. For todays
descendants of the Aztecs, the crown is a symbol of their cultural identity and stands for
the blossoming and ensuing destruction of their culture. They argue that the crown was
stolen from them, and that the Museum should return it. For Austrian politicians and
diplomats, the crown has thus become a cause for political dissonance with Mexico (for
an in-depth analysis of the transformations of exhibits in anthropological museums, see
Dring and Hirschauer, 1997).
As the example of the feather crown shows, each of the meanings assigned to a
materialization is tightly related to power relations (e.g. whether the crown is a ritual
item, a scientific exhibit, or a symbol of collective identity). In the materialization as
such, these power relations are invisible, and the task of the analyst is to bring them out
into the open. The analyst can only do this if he considers historical contexts (this touches
on the issue of historical discourse and dispositive analysis, which is currently
revolutionizing historical theory; see, for example, Sarasin, 2003).
As these initial thoughts on methods for analysing dispositives indicate, the task of a
dispositive analysis is very complex. It encompasses the analysis of knowledge in
discursive and nondiscursive practices as well as in materializations. An example of such
an analysis is Michel Foucaults (1979) book Discipline and Punish. Also, Victor
Klemperers (2001) Diary of the Nazi Years can be read as a dispositive analysis. Both
authors provided no explicit information on their methodology. They applied their
methodology implicitly, or as Foucault calls it, in the form of bricolage. They analysed
discourse, assembled knowledge, consulted statistics, deconstructed them critically, drew
conclusions, added their own opinions, and so on.
The thoughts on dispositive analysis presented in this chapter do not provide a recipe
or schema. However, they do give some ideas on how to approach dispositive analysis.
A central part of dispositive analysis is the discourse analysis of texts. Moreover,
dispositive analysis comprises the analysis of nondiscursive practices, for which methods
developed in ethnography, such as ethnographic interviews and participant observation,
provide important means. A final component is the analysis of materializations, which can
draw on methods such as multimodal discourse analysis and artefact analysis. An explicit
methodology for combining these approaches has yet to be developed. Such an endeavour
can probably only be achieved in connection with concrete research projects. It would
therefore be desirable for such projects to devote some space and time to explicit
reasoning about methodology. This would promote the development of dispositive
analysis, and would contribute to bridging the existing gaps between discourse analysis
and other methods of empirical social research.

Notes
1 We therefore keep the discussion of theoretical background as short as possible and do not present previous
empirical research projects at great length. To readers particularly interested in these aspects, we recommend the
works by Siegfried Jger (2001b, 2004) and by Margarete and Siegfried Jger (2007).
2 This French term literally means tinkering, fiddling, creative do-it-yourself, making use of any resources
that happen to be at hand
3 For example, a definition of racism that is generally justifiable and well-accepted in the sciences encompasses
the following three elements: (1) One or several people are for biological or cultural reasons constructed as an
ethnic group or even a race. (2) This group is evaluated (negatively or positively, e.g. when blacks are assumed to
be superior jazz musicians). (3) The construction and evaluation take place from a position of power (which in
discourse analysis is obvious, since discourse is per se powerful).
3
Critical Discourse Studies: A Sociocognitive Approach1

Teun A. van Dijk

Terminology and definitions
The discoursecognitionsociety triangle
The example: the defense of capitalism
Topics: semantic macrostructures
Local meanings
The relevance of subtle formal structures
Relating text and context: context models
Discourse semantics: event models
Social cognition
Social situations
Societal structures
Appendix

Terminology and definitions

Although critical approaches to discourse are commonly known as Critical Discourse
Analysis (CDA), I prefer to speak of Critical Discourse Studies (CDS). This more
general term suggests that such a critical approach not only involves critical analysis, but
also critical theory, as well as critical applications. The designation CDS may also
avoid the widespread misconception that a critical approach is a method of discourse
analysis.
For the same reason, I favour the term Discourse Studies (DS), rather than Discourse
Analysis, to designate a multidisciplinary field of scholarly activities that are obviously
not limited to the analysis of text and talk. Moreover, as a discipline, DS has many types
and methods of analysis: it is not a method among others within the humanities and the
social sciences.
CDS is not a method, but rather a critical perspective, position or attitude within the
discipline of multidisciplinary Discourse Studies. Critical research makes use of a large
number of methods, both from Discourse Studies itself, as well as from the humanities,
psychology and the social sciences.
The critical approach of CDS characterizes scholars rather than their methods: CDS
scholars are sociopolitically committed to social equality and justice. They also show
this in their scientific research, for instance by the formulation of specific goals, the
selection and construction of theories, the use and development of methods of analysis
and especially in the application of such research in the study of important social
problems and political issues.
CDS scholars are typically interested in the way discourse (re)produces social
domination, that is, the power abuse of one group over others, and how dominated groups
may discursively resist such abuse.
CDS is not just any social or political research, as is the case in all the social and
political sciences, but is premised on the fact that some forms of text and talk may be
unjust. One of the tasks of CDS is to formulate the norms that define such discursive
injustice. CDS aims to expose and help to combat such injustice. It is problem-oriented
rather than discipline-or theory-oriented. Such a research policy presupposes an ethical
assessment, implying that discourse as social interaction may be illegitimate according to
some fundamental norms, for instance those of international human and social rights. At
the same time, critical analysis should be aware of the fact that such norms and rights
change historically, and that some definitions of international may well mean Western.
As a criterion, we thus call any discourse unjust if it violates the internationally
recognized human rights of people and contributes to social inequality. Typical examples
are discourses that ultimately (re)produce inequalities of gender, race or class.
Finally, socially committed research should be carried out in close collaboration and
solidarity with those who need it most, such as various dominated groups in society. This
also means, not least for students, that CDS research and especially its practical
applications should be accessible, and avoid an esoteric style. In that and many other
senses, CDS researchers are profoundly aware of the role of scholarly activities in
society.

What are Critical Discourse Studies (CDS)?

Although it is virtually impossible to briefly and adequately define a type of scholarly investigation, critical studies
of discourse typically have the following properties:

They aim to analyse, and thus to contribute to the understanding and the solution of, serious social problems,
especially those that are caused or exacerbated by public text and talk, such as various forms of social power
abuse (domination) and their resulting social inequality.
This analysis is conducted within a normative perspective, defined in terms of international human rights, that
allows a critical assessment of abusive, discursive practices as well as guidelines for practical intervention and
resistance against illegitimate domination.
The analysis specifically takes into account the interests, the expertise and the resistance of those groups that are
the victims of discursive injustice and its consequences.
The discoursecognitionsociety triangle
It is within this framework that I propose to formulate and illustrate some of the
principles I try to observe when doing CDS. Given my multidisciplinary orientation, the
overall label I sometimes use for my approach is that of sociocognitive discourse
analysis. Although I dislike labels (because they are reductionist and because I have many
times changed my area and perspective of research), I have few quarrels with this one,
especially since it emphasizes that unlike many of my colleagues in CDS and various
interactionist approaches I value the fundamental importance of the study of cognition
(and not only that of society) in the critical analysis of discourse, communication and
interaction.
This means, among other things, that I am also interested in the study of mental
representations and the processes of language users when they produce and comprehend
discourse and participate in verbal interaction, as well as in the knowledge, ideologies
and other beliefs shared by social groups. At the same time, such an approach examines
the ways in which such cognitive phenomena are related to the structures of discourse,
verbal interaction, communicative events and situations, as well as societal structures,
such as those of domination and social inequality, as mentioned above.

What is cognition?
As is the case for other fundamental notions, cognition is a notion that is jointly defined by all the disciplines
currently integrated under the label cognitive science, such as psychology, linguistics, philosophy and logic as
well as the brain sciences. Some typical cognitive notions used here are, for instance:

Mind, defined, for example, as a central function of the human brain.

Cognition as the set of functions of the mind, such as thought, perception and representation.
Memory: Short Term (Working) Memory (STM) and Long Term Memory (LTM).
Episodic (personal, autobiographic) Memory (EM) and Semantic (sociocultural, shared) Memory as part of
LTM.
Semantic Mental Models (represented in EM) as the subjective representations of the events and situations
observed, participated in or referred to by discourse.
Goals as mental models of the situations to be realized by action.
Pragmatic Context Models: specific mental models of subjective representations (definitions) of the relevant
properties of communicative situations, controlling discourse processing and adapting discourse to the social
environment so that it is situationally appropriate.
Knowledge and its organization: shared, sociocultural beliefs that are certified by the (knowledge) criteria or
standards of a (knowledge) community.
Ideology as the shared, fundamental and axiomatic beliefs of specific social groups (socialism, neoliberalism,
 feminism, (anti)racism, pacifism, etc.).
Attitudes as the socially shared, ideologically based opinions (normative beliefs) about specific social issues
having given rise to debate or struggle (abortion, divorce, euthanasia, immigration, etc.).
Cognitive processes such as the production and comprehension of discourse/interaction on the basis of specific
mental models, controlled by context models, and based on knowledge and ideologies.

The label of the sociocognitive approach does not mean that I think that CDS should be
limited to the social and cognitive study of discourse, or to some combination of these
dimensions. It only means that (at present) I am personally most interested in the
fascinating sociocognitive interface of discourse, that is, the relations between mind,
discursive interaction and society. For instance, in my work on racism (Van Dijk, 1984,
1987, 1991, 1993a, 2005, 2007b), and in my research on ideology (Van Dijk, 1998) and
context (Van Dijk, 2008a, 2009), I have shown that these are both mental and social
phenomena. It goes without saying, however, that the complex, real-world problems
CDS deals with also need a historical, cultural, socioeconomic, philosophical, logical or
neurological approach, among others, depending on what one wants to know (see, for
instance, the various approaches represented in Van Dijk, 1997, 2007a).
Given the verbalsymbolic nature of discourse, explicit CDS of course also needs a
solid linguistic basis, where linguistic is understood in a broad sense. Whatever
cognitive and social dimensions of discourse CDS deals with, it always needs to account
for at least some of the detailed structures, strategies and functions of text or talk. These
may include grammatical, pragmatic, interactional, stylistic, rhetorical, semiotic,
narrative, argumentative or similar forms and meanings of the verbal, paraverbal and
multimodal structures of communicative events.
Having emphasized the necessity of a broad, diverse, multidisciplinary and problem-
oriented CDS, I thus limit my own endeavours to the domain defined by the discourse
cognitionsociety triangle.
In a more or less informal way, we may view the combined cognitive and social
dimensions of the triangle as defining the relevant (local and global) context of discourse.
Indeed, the sociopolitical and problem-oriented objectives of CDS especially need
sophisticated theorization of the intricate relationships between text and context. We shall
see that adequate discourse analysis at the same time requires detailed cognitive and
social analysis, and vice versa, and that it is only the integration of these accounts that
may reach descriptive, explanatory and especially critical adequacy in the study of social
problems.
It should be emphasized that context, as I define it, is not simply some kind of social
environment, situation or structure such as the social variables of gender, age or
race in classical sociolinguistics. Rather, a context is a subjective mental
representation, a dynamic online model, of the participants about the forthemnow
relative properties of the communicative situation. I call such a representation a context
model (Van Dijk, 2008a, 2009). It is this mental definition of the situation that controls
the adequate adaptation of discourse production and comprehension to their social
environment. This is just one of the ways in which cognition, society and discourse are
deeply and mutually integrated in interaction.
Within the theoretical framework of the discoursecognitionsociety triangle, context
models mediate between discourse structures and social structures at all levels of
analysis. This means that society is understood here as a complex configuration of
situational structures at the local level (participants and their identities, roles and
relationships engaging in spatiotemporally and institutionally situated, goal-direction
interaction), on the one hand, and societal structures (organizations, groups, classes, etc.
and their properties and e.g. power relations), on the other hand. This side of the
triangle also includes the cultural and historical dimensions of interaction and social
structure, that is, their cultural variation as well as their historical specificity and change.
It is also at this side of the triangle that we locate the consequences of discursive
injustice, for instance in the form of social inequality.
Finally, it should be stressed that the use of the triangle is merely an analytical
metaphor representing the major dimensions of critical analysis. It should not be
interpreted as suggesting that cognition and discourse are outside society. On the
contrary, human beings as language users and as members of groups and communities, as
well as their mental representations and discourses, are obviously an inherent part of
society. It is also within social structure that language users interpret, represent,
reproduce or change social structures such as social inequality and injustice.

What is discourse?
Discourse analysts are often asked to define the concept of discourse. Such a definition would have to consist of
the whole discipline of discourse studies, in the same way as linguistics provides the many dimensions of the
definition of language. In my view, it hardly makes sense to define fundamental notions such as discourse,
language, cognition, interaction, power or society. To understand these notions, we need whole theories or
disciplines of the objects or phenomena we are dealing with. Thus, discourse is a multidimensional social
phenomenon. It is at the same time a linguistic (verbal, grammatical) object (meaningful sequences or words or
sentences), an action (such as an assertion or a threat), a form of social interaction (like a conversation), a social
practice (such as a lecture), a mental representation (a meaning, a mental model, an opinion, knowledge), an
interactional or communicative event or activity (like a parliamentary debate), a cultural product (like a telenovela)
or even an economic commodity that is being sold and bought (like a novel). In other words, a more or less
complete definition of the notion of discourse would involve many dimensions and consists of many other
fundamental notions that need definition, that is, theory, such as meaning, interaction and cognition.

The example: the defense of capitalism

After a discussion of some of the general notions of CDS, I shall now be more specific
about the kind of discourse structures to be attended to in such a critical approach. At the
same time, I shall illustrate and further develop the theoretical notions introduced above
by analysing some properties of a specific example: a text A Petition Against the
Persecution of Microsoft of the Center for the Moral Defense of Capitalism (now called
Center for the Advancement of Capitalism) downloaded from the internet in 2001
(www.moraldefense.com) (text no longer available on the website in 2008). This petition
criticizes the US government for its legal battle against Microsoft, and asks readers to
sign it.
I have chosen this text as an example because of the widely perceived socioeconomic
problem that vast international corporations may abuse their power in order to dominate
the market and hence limit the freedom of choice of consumers.

Context

Of the communicative situation of this, we do not know much more than that the text was
found on the internet in 2000, and that it was produced by the Center for the Moral
Defense of Capitalism, whose very name suggests a neoliberal, conservative think tank.
A new internet search in December 2007 presented the following definition of the goals
of the Center:

The Center for the Advancement of Capitalism is dedicated to advancing individual rights and economic freedom
through Ayn Rands philosophy of Objectivism. (www.moraldefense.com)

This means that readers who know this Center may already have a more detailed context
model of the communicative situation than other readers (such as information about the
ideological background, identity and goals of the Center), and can interpret the text in that
light. Intended recipients are people who use the web and understand English. A further
aspect of the communicative situation is the assumed previous knowledge about legal
actions against Microsoft because of its abuse of market power (a near monopoly,
because of the pervasive use of Windows as an operating system for PCs) to couple other
programs (such as its web browser Explorer) to this operating system. In other words, the
context analysis focuses on Setting (Time, Place), Participants and their properties and
relations, as well as on their Goals, the Knowledge presupposed by the participants, and
the Ideology of the participants.

Topics: semantic macrostructures

I often advocate beginning Critical Analysis (CrA) with an analysis of semantic
macrostructures, that is, with a study of global meanings, topics or themes. These are
what discourses are (globally) about; they are mostly intentional and consciously
controlled by the speaker; they embody the (subjectively) most important information of a
discourse, express the overall content of mental models of events (see below), and
perhaps most importantly, they represent the meaning or information most readers will
memorize best of a discourse. Discursively, topics or themes are characteristically
expressed in titles, abstracts, summaries and announcements.
For contextual reasons, we select topics as a significant structure to study because they
are usually controlled by powerful speakers, because they influence many other structures
of a discourse (such as its global coherence), and because they have the most obvious
effects on the (memory and consequent actions of) recipients and hence on the process of
reproduction that underlies social power and dominance (for details, see Van Dijk and
Kintsch, 1983).

In our sample text, the title, A Petition Against the Persecution of Microsoft expresses not only part of the topic
(the persecution of Microsoft), but also the self-categorization of the text genre (petition). Theoretically and
psychologically, topics or macrostructures are derived from a text by inference through a process of information
reduction that is being practised especially in text summarization. Thus, we may summarize this text by, for
example, the following macropropositions:
M1 The freedom of enterprise is under attack by antitrust laws.
M2 Successful businessmen are being represented as tyrants.
M3 The suit against Microsoft is an example of this M1 and M2.
M4 Government should not limit the freedom of the market.
M5 Microsoft has the right to do what it wants with its products.
M6 Innovators should not be punished.
M7 We call that the case against Microsoft be dismissed.
In a further reduction, one can summarize these macropropositions with the overall macroproposition (topic): The
US government is requested to stop its judicial persecution of the innovator Microsoft.
We see that these various topics/macropropositions indeed represent very high level, sometimes abstract,
principles. In this case, these propositions are more or less a direct expression of some tenets of a classical
capitalist ideology about the freedom of enterprise. In other words, the macropropositions express the general
ideological principles of the freedom of the market, and then apply these to the special case of Microsoft. We shall
see later that this distinction reflects the difference between socially shared representations, on the one hand, and
more personal mental models, on the other.

Local meanings
Next, a CrA may focus on local meanings, such as the meaning of words (a study that also
may be called lexical, depending on ones perspective), the structures and nature of
propositions, and coherence and other relations between propositions, such as
implications, presuppositions, levels of description, degrees of granularity and so on.
The reasons to give priority to semantic analysis in CrA are mostly contextual: local
meanings are a function of the selection made by speakers/writers in their mental models
of events or their more general knowledge and ideologies. At the same time, they are the
kind of information that (under the overall control of global topics) most directly
influences the mental models, and hence the opinions and attitudes of recipients. Together
with the topics, these meanings are best recalled and reproduced by recipients, and hence
may have the most obvious social consequences.
As is the case for many variable structures of discourse, local meanings may also be
controlled by context models. That is, not all local meanings are equally appropriate in
communicative situations, as we know from the general distinction between formal,
casual and popular discourse styles, as well as from social and cultural taboos. Indeed,
most institutional text and talk is contextually constrained by the specific aims and norms
of institutional interaction and organization.
Theoretically, this means that the generation of meanings based on mental models of
events talked about is controlled by the various categories and contents of context models.
One obvious controlling mechanism is that the mutual knowledge as indexed by the
context model requires that speakers ongoingly in principle only assert what the
recipients do not already know, as we also know from the appropriateness conditions of
the speech act of an assertion. Other constraints are defined for specific kinds of
participants and their identities, roles and relationship: there are often limitations of
content or meaning for specific categories of speakers, and this is especially so in
institutional situations.

At this local semantic level, we may for instance examine the choice of the word persecution in the title of our
sample text, a choice that has various implications that express the ideological perspective of the author (The
Center for the Moral Defense of Capitalism): the action of the government is defined in negative terms, implying a
form of morally or legally reprehensible harassment or force, or abuse of power. At the same time, the choice of
this word implies that Microsoft is the victim of this aggression. In more general terms, lexical selection here
shows the familiar form of negative other-presentation, and positive self-presentation as an organization taking the
defence of the victims. As part of the main macroproposition, the choice of the concept of persecution also
contributes to the organization of the local meanings in the rest of the text. In more cognitive terms, this means
that the choice of this word may influence the formation of the macronodes of the mental model of the readers of
this text.
Similarly relevant is the repeated use of the word rights in the first paragraph, typically associated with
individual and freedom, all profoundly ideological concepts related to the constitution and prevailing ideology of
the United States. In order to be able to qualify the legal action of the government in the starkly negative terms of
a persecution, it needs to be shown that the rights of individuals are being violated, and what these rights are.
The emphasis on rights has several other functions, such as associating Us and Our position with something good
and legitimate, and thus preparing the negative evaluation of the US Government when it violates these rights.
Apart from polarizing the mental model being construed here, this paragraph at the same time functions as an
important premise in the overall argumentation of this text.

For CDS, especially interesting in such a local semantic analysis is the study of the many
forms of implicit or indirect meanings, such as implications, presuppositions, allusions,
vagueness and so on. Again, such meanings are related to underlying beliefs, but not
openly, directly or precisely asserted for various contextual reasons, including the well-
known ideological objective to de-emphasize Our bad things and Their good things.
 In our sample text, there are many propositions that are implied or presupposed, but not explicitly asserted. When
the authors say that antitrust legislation comes under the guise of protecting the public, the expression under
the guise and the quotes imply that it is not true (or merely alleged) that antitrust laws protect the public. Note
also that in the second paragraph, as well as throughout the text, many expressions have ideological
presuppositions, such as:

competitors are envious of successful business men

officials are power-hungry
the business world has creative geniuses
business empires are hard-won.

Apart from further emphasizing the polarization between Government and Business, the local meanings of the text
thus create another polarization between envious competitors and brilliant creators in the business. Notice also that
the lexical choice and metaphors further emphasize these polarizations: envious, power-hungry, hard-won, control,
regulators, and breaking to pieces, etc. are the negative concepts associated with Them, the government (and
some business people), whereas We and those we protect are associated with success, creative geniuses and by
litotes with crime and tyrant. Again, such words not only contribute to the overall polarization of the conceptual
structure of the text, but also to the formation of a biased, polarized model of the events, where the Actors are
neatly differentiated between the Good and the Bad.
The first two paragraphs are formulated in general terms, and apply to rights and their violation, as well as to the
antitrust laws. The third paragraph begins with the functional move of Specification or Example: what has been
said so far specifically applies to the case of Microsoft. Theoretically, this means that the first paragraphs are
rather expressions of (general) social representations, such as attitudes and ideologies, whereas the third
paragraph describes the current case, Microsoft, and thus sets up the more specific mental model based on these
general social representations. Given the ideological slant of the first paragraphs, there is little doubt that this
model, as expressed by the Center, is also ideologically biased, and we may expect that the general polarization
constructed before will be applied here, as is indeed the case. Notice also that conceptual polarization is often
implemented in the text by various forms of hyperboles, as we have already seen in the lexical choice of crime,
tyrants and geniuses. Such hyperboles may even come close to outright lies, for instance when it is asserted
that Bill Gates has been deprived of his right to control his own company.
The use of his, businessmen and the men who have made this country great suggests that men, especially or
exclusively and no women, are involved in business and its success. Thus, apart from expressing a starkly
conservative neoliberal ideology, the Center also professes a sexist ideology by verbally excluding women, thus
contributing to a more overall conservative meta-ideology that also controls the nationalist ideology expressed in
the characteristic form of US self-glorification (the greatness of this country).
Finally, among the many other semantic properties of this text, we should also mention the importance of what is
being left out in the text. Thus, it is suggested that the success of Microsoft is based on the principle of better
products for a lower price, but of course not the well-known practice of the forced bundling of products (like
Windows and its internet browser). Nearly trivially then, we may formulate the general rule that the negative
properties of Us (or those we defend) are either omitted or downgraded in the text. Note that, theoretically,
omission is only a relevant property of a discourse when it can be shown that the omitted information is part of the
mental model (the Center no doubt knows about the illegal practices of Microsoft), or of more general, shared
knowledge that is needed or may be used to produce or understand a text. In this case, the mental model of a
critical reader may of course be different from that persuasively expressed by the Center.

We now have a first impression of the theoretically based practical guidelines on which
discourse structures to study among many hundreds. Of course, this is only an example.
The point is that such a choice is twice context-bound: first, by our own (scholarly) aims,
our research problems, and the expectations of our readers, as well as the social
relevance of our research project. Second, by the relevance of specific discourse
structures studied in their own context, such as the aims and beliefs of the speaker or the
recipients, the social roles, positions and relations of participants, institutional
constraints and so on.

The relevance of subtle formal structures

Besides or instead of the semantic structures just mentioned, we may be more interested
in those structures of text or talk that are usually less consciously controlled or
controllable by the speakers, such as intonation, syntactic structures, propositional
structures, and rhetorical figures, as well as the many properties of spontaneous talk, such
as turn-taking, repairs, pauses, hesitation, and so on.
These various forms generally do not directly express the underlying meanings and
hence beliefs, but rather signal the pragmatic properties of a communicative event, such
as the intention, current mood or emotions of speakers, their perspective on the events
talked about, their opinions of co-participants, and interactional concerns especially such
as positive self-presentation and impression formation. Thus, men may well be able to
hide their negative opinions of women, or white people of black people, but indirectly
their evaluations, position or face, and hence their identity, may be signalled by subtle
structural or formal characteristics of talk or the non-verbal properties of communicative
events (gestures, face work, body position, distance, and so on).

At both the global and local levels, our sample text also has several formal properties that enhance the general
underlying topic and argumentation. Indeed, as we have seen, the very argumentative structure of this petition is
one of the global, formal properties that organizes this text: the general premises, expressed in the first two
paragraphs, focus on constitutional rights on the one hand, and the alleged violation of such rights by the antitrust
laws on the other. Both are then applied to the more particular premise that Microsoft is the victim of this
violation. In the same way, several of the meanings in these and other paragraphs have specific argumentative
functions, such as the reference to the Declaration of Independence as an authoritative and hence credible set of
principles by which the governments duties are evaluated.
Similarly, the discourse may enhance its effectiveness by various rhetorical moves, of which hyperboles have
already been mentioned (geniuses, etc.). There is also the use of the opposite of what is being meant, for
instance in irony or litotes (such as the Microsoft crime of being successful).
Indeed, the very structure of polarization of this kind of ideological discourse not only has a semantic property, but
also formal properties, such as the rhetorical contrast expressed in the fourth paragraph: here, the governments
views that the free market imposes force, and that control is freedom, are criticized as an inversion of
reality as the Center sees it, and it does so by a construction of a contrast.
Finally, in the same way that the semantic and rhetorical polarizations of this text express and help construct
biased models of the case against Microsoft, its formal style is a marker of its genre: the official petition. This
formal style begins with the paraphrase of the Declaration of Independence, but is also lexically expressed in the
Centers own petition, as in the repeated We hold that , not by anyones permission, but by absolute right,
etc., signalling something like a Declaration of the Free Market.

One conclusion of this discussion of the criteria applied in the choice of the discourse
structures studied in CDS projects is that any method or approach that limits itself to
some genre or dimension of discourse only can by definition only provide a very partial
analysis. Trivially, grammarians usually study grammar, conversation analysts
conversations, and narratologists stories and their structures. Now, if some CDS
researcher, for the double contextual reasons explained above, precisely needs to study
some aspects of grammar, conversation or narration, it is obviously in these more specific
areas of research that one looks for relevant structures. But as soon as the critical aims
of the research project require a broader approach, those scholars who limit themselves
to the study of a single genre or types of structure are often unable to fully deploy their
expertise. Hence also my oft-repeated criticism of the exclusive membership of one
school, approach or scholarly sect, and my plea for diversity, flexibility and
multidisciplinarity as general criteria for CDS. This obviously first of all applies to CDS
and critical scholars themselves! Indeed, there are many ways to do interesting and
relevant discourse analysis, and we also need to develop the more general theories,
concepts and methods that may (also) be used in critical analysis. Indeed, few of the
general notions used so far in our analysis were developed within critical studies.

Relating text and context: context models

It has been assumed above that the relation between discourse and society is not direct,
but needs to be mediated by so-called context models (Van Dijk, 2008a, 2009). That is,
social structures organizations, groups, gender, race, etc. are phenomena that cannot
be directly linked to the mental processes of discourse production and understanding, as
was previously the case in traditional sociolinguistics.
Hence, we need a mediating cognitive device that is able to represent the relevant
structures of the social situation, both locally (micro) as well as globally (macro), and
that at the same time is able to control discourse, the mental processes of production and
comprehension and its situated variation.
Context models, defined as specific mental models, represented in episodic memory,
do just that. They make sure that language users adapt their discourse to the social
environment, so that it is socially appropriate.
Given more or less the same content or information as specified in the (semantic)
mental model of things talked about, context models are needed to appropriately
formulate more or less the same content in different communicative situations. This
means that context models define the genre as well as the style of text and talk.
As summarized above, context models are organized by a relatively simple schema
consisting of fundamental categories, such as:

a spatiotemporal setting
participants
 identities, roles, relationships
goals
knowledge
ideologies
the ongoing social action.

Context models are dynamic and are ongoingly adapted to the communicative situation, if
only because the knowledge of the recipients is constantly changing as a result of the very
discourse itself.
Since context models control all the variable aspects of discourse (intonation, syntax,
lexicon, etc.), they need to be kept activated at least in the background of working
memory and hence cannot be too complex or voluminous (and represent dozens or
hundreds of aspects of the current communicative situation).
It goes without saying that context models as a crucial interface between discourse and
society also play a fundamental role in critical discourse studies, which are premised on
the detailed analysis of some of these discoursesociety relationships, such as those of
power and domination.

In our example of a petition, the context defining the communicative event is rather obvious. The overall societal
domain for this text is that of business or the market, and the overall actions are those of advocating the freedom
of enterprise, and protecting business against government interference. The local setting of the communicative
event is the internet. The communicative role of the participant is that of speaker/writer, author and originator, the
interactional role that of a defender of Microsoft and as an opponent of the government, whereas the
socioeconomic role is that of an organization advocating the freedom of the market. The other participant, the
addressee, is explicitly referred to in the beginning of the text as Fellow Americans, thus pragmatically trying to
emphasize the unity of the we group for which this Center claims to be the defender. It is interesting that
although the proposal for the petition is directed at Fellow Americans, the proposed petition itself is addressed to
the relevant final destinataries: the judge, the Senate Judiciary Committee, the Attorney General and the President
of the United States.
The current communicative action is that of publishing a text on the internet persuading readers to sign a petitition.
This action is being performed through the speech acts of accusing the government, and defending Microsoft. The
(complex) mental structures defining the cognitive dimension of the context consist of the various ideologies
analysed above, as well as the more specific attitudes and opinions (about the legal action of the government
against Microsoft) we have found expressed throughout the text. Although expressing group co-membership in
addressing Fellow Americans, the persuasive structure of the text presupposes that not all Americans may have
the same opinion about the practices of Microsoft. Finally, the text is meaningful for its readers only because it
presupposes a vast amount of common ground and commonsense knowledge, such as about the Declaration of
Independence, Microsoft, antitrust laws and so on, as well as specific (model-based) knowledge about the current
court case against Microsoft.
The important point is that throughout, the text adapts to this subjectively construed context model of the current
communicative situation, for example, as follows:

The meanings of the text are all understandable within the broader framework of the three semantic domains of
business, justice and goverment.
The genre and speech act of the petition is one form of implementing the overall defence of the free market,
which is the global aim of the Center.
The action of the government is defined as a violation of Our rights, and hence is a sufficient condition for the
 success of the current genre and speech act of a petition.
The overall topic semantically realizes the reason for the speech act and genre of this specific petition: Microsofts
rights have been violated.
The argumentative structure is organized in such a way as to optimally sustain the communicative function of this
text as a form of persuasion.
The polarization of the opinions at all levels of the text expresses the attitudes and the ideology of the Center, and
tries to influence those of the readers and final destinataries.
Lexical choice is appropriate for the genre of a formal, public petition.
The text presupposes existent general knowledge about what business, laws, governments, etc. are, as well as
specific knowledge about the process against Microsoft. However, it does not express or presuppose knowledge
that debilitates its defence (e.g. about the illegal practices of Microsoft).

For any kind of CDS research that links text with some social situation, it is important to
realize that whatever the broader social or political situation, it may not reach or impact
on discourse simply because a speaker may find it irrelevant and further ignores the
relevant information in the construction of the context model. Also, the changes that
speakers apply in their discourses, for example, because of politeness or other forms of
persuasion, need to be taken into account.
We see that the notion of subjective context models theoretically implies the important
criterion of relevance, namely that only those properties of communicative situation are
construed as forming part of the context, if they are (now) relevant for the participant.
Note that relevance is not absolute or (only) socially determined, but is relative to the
current knowledge, goals, wishes, interests and personal experiences of the current
speaker or recipient at each moment of a communicative event.

Discourse semantics: event models

Crucial to any theory of discourse is its semantics, which provides a theory of discourse
meaning and interpretation. Traditionally, a linguistic semantics is formulated in terms of
abstract meanings: concepts, propositions and their mutual relations. Thus, local
coherence of discourse is defined as constraints on relations between propositions, on the
one hand, and on the relations between the referents (facts of some mental model) of
such propositions, on the other hand. For instance, a sequence of propositions may be
said to be coherent if the facts (states of affairs, events, actions, etc.) referred to by these
propositions are related, for instance by relations of causality or more generally by
relations of conditionality (such as enabling). We may call this referential or extensional
coherence. On the other hand, coherence may also be established by intensional
(meaning) relations among the propositions themselves, for instance when a proposition
Q is a Generalization, Specification, Example, Illustration, etc. of a previous proposition
P. Since Q is defined here in terms of its function relative to P, we may call this
functional coherence.
As suggested, such an account is rather abstract, and appears to have little empirical
relevance: this is not the way language users go about producing and understanding
meaningful discourse. Instead of an abstract theory of meaning, therefore, we may define
meaning and interpretation in cognitive terms, that is, in terms of mental operations and
representations.
Such an approach also solves a major problem of the more formal account, namely the
nature of reference: notions that define referential coherence, such as referents, facts and
causal relations between facts, are not defined in linguistic (grammatical) semantics. A
cognitive approach to discourse meaning not only may solve that problem, but at the same
time accounts for the subjectivity of coherence: discourses are not coherent in the
abstract, but are coherent for language users, and according to their intentions,
interpretations or understandings.
Thus, discourses are not so much coherent because their propositions refer to related
objective facts in some possible world, but rather to the episodes (events and
situations) as interpreted, defined and (seen to be) related by language users. We have
seen above that such subjective interpretations are represented in episodic memory as
mental models of events and situations. We may now simply say that a discourse is
coherent if language users are able to construct a mental model for it. We may call these
models event models in order to account for the fact that they subjectively represent the
events the discourse refers to. Whereas context models, as discussed above, are
pragmatic, event models are semantic.
Since context models are a specific kind of event model (namely a model of
communicative events), event models have more or less the same structure organized by a
schema with categories such as Setting, Participants and Actions/Events and their
respective subcategories and properties.

Thus, in the analysis of our example, we have repeatedly seen how at all levels of the petition text, structures are
geared not only to the adequate expression of the mental model of the (authors in the) Center for the Moral
Defense of Capitalism, but also to the persuasive construction of a preferred model among the addressees. That
is, this intended model features the macro-opinion that the US Government through its antitrust laws in general,
and its case against Microsoft in particular, violates the basic principles of the freedom of the market. That is, the
current mental model of the Microsoft case is a fairly direct instantiation of more general attitudes about antitrust
legislation and their basic ideologies about the freedom of enterprise. The polarization between Us and Them, or
between Business and Government, and its respective Good and Bad qualities, is thus a specification of more
general opinions about ingroups and outgroups as we know them from the study of ideology (Van Dijk, 1998). In
other words, the authors of the text not only try to adequately express their own model of the events, but
formulate the text in such a way that the intended model be accepted by the readers. This is what persuasion is all
about, and it may be obvious that without an account of mental model structures, such a verbal act and its
concomitant verbal structures cannot be adequately described, let alone explained.

The notion of a mental model also explains another fundamental property of discourse
meaning: its incompleteness. Semantically speaking, a discourse is like the tip of an
iceberg: only some of the propositions needed to understand a discourse are actually
expressed; most other propositions remain implicit, and must be inferred from the explicit
propositions (given a body of world knowledge, to which we shall come back below). It
is the model that provides these missing propositions. Implicit or implied propositions
of discourse are thus simply defined as those propositions that are part of the mental
model for that discourse, but not present in its semantic representation. That is, for
pragmatic reasons as defined by the context model of a discourse (including the beliefs
attributed to the recipient by the speaker), only part of the propositions of a model need to
be expressed for instance, because the speaker believes that such information is
irrelevant, because the recipient already knows these propositions, or because it may be
inferred from other propositions. Hence, mental models at the same time provide an
excellent definition of presuppositions, namely as those propositions of event models that
are implied but not asserted by the discourse.
Besides discourse coherence and implications, the notion of event models provides a
framework for many other, hitherto problematic, aspects of discourse and discourse
processing, briefly summarized as follows:

Event models are a crucial cognitive aspect of the constructionist way people view,
understand, interpret and recall reality. In other words, our personal experiences,
as represented in episodic memory, consist of mental constructs: models.
Event models are not only the result of discourse comprehension, but are also the
basis of discourse production. Event models may be part of our planning of
discourse (what we want to say). Thus, stories are contextually appropriate
(relevant, interesting) formulations of underlying event models, for instance of
personal experiences as stored in episodic memory.
Event models are subjective (personal interpretations of events), but have a social
basis, because they instantiate socially shared knowledge and possibly also group
ideologies (see below). That is, context models explain how discourse may be
ideologically biased, namely when based on event models that instantiate
ideological propositions.
Event models account for the fact that different language users, members of different
communities and of different social (e.g. ideological) groups, may have different
interpretations of events, and at the same time, different interpretations of the same
discourse. This implies that the influence of discourse on the minds of recipients
may also be different.

Social cognition
CDS is not primarily interested in the subjective meanings or experiences of individual
language users. Power, power abuse, dominance and their reproduction typically involve
collectivities, such as groups, social movements, organizations and institutions (Van Dijk,
2008b).
Therefore, besides the fundamental interface of personal mental models that account for
specific discourses, a cognitive approach also needs to account for social cognition, that
is, the beliefs or social representations they share with others of their group or
community. Knowledge, attitudes, values, norms and ideologies are different types of
social representations.
These social representations also play a role in the construction of personal models, as
we have seen in some detail in our brief analytical remarks about the Petition text. That
is, socially shared knowledge and opinions may be instantiated in such models. In other
words, models are also the interface of the individual and the social, and explain how
group beliefs may affect personal beliefs and thus be expressed in discourse. Ethnic or
gender prejudice, which are typically defined for social groups, thus also appear as an
(instantiated) property of individual discourses. And conversely, if the personal mental
model of social events of an influential person is shared by others of a group or
community, mental models may be generalized and abstracted from to form social
representations such as knowledge, attitudes and ideologies. This is of course precisely
the aim of the Petition text.
It is one of the aims of CDS research to analyse specific discourses in this broader,
social framework, for instance by trying to infer (sometimes quite indirectly) which
shared social representations are being expressed or presupposed by discourse. Thus,
critical discourse studies of racism, sexism or classism need to relate properties of
discourse with these underlying, socially shared, representations, which group members
use as a resource to talk about (members) of other groups. Outgroup derogation and
ingroup celebration are the socialpsychological strategies typically defining this kind of
chauvinist discourse.

Ideology

Dominance, defined as power abuse, is often based on, and legitimated by ideologies,
that is, by the fundamental social beliefs that organize and control the social
representations of groups and their members. Many forms of CDS research require such
an ideological analysis, especially because ideologies are typically expressed and
reproduced by discourse.
It is important to stress here that the cognitive framework sketched above suggests that
there is no direct link between discourse and ideology. The basic beliefs of an ideology
(for instance, about the equality of women and men in a feminist ideology) organize
specific attitudes, that is, the socially shared opinions of a group (for instance, about
abortion, sexual harassment or equal pay), which in turn may influence specific event
models (about specific participants and actions), which finally may be related to
discourse under the final control of context models. In other words, to read off
ideologies from discourse is not always possible, precisely because ideologies need to
be very general and fairly abstract. Although we still ignore what the general structure of
ideologies are (van Dijk, 1998), it may be assumed that they are organized by a general
schema consisting of the basic categories that organize the self and other representations
of a group and its members, such as:

membership devices (who belongs to us?)

typical acts (what do we do?)
aims (why do we do it?)
relations with other (opponent) groups
resources, including access to public discourse.

Note that many of the features that were traditionally examined in the (critical or other)
analysis of discourse are here accounted for in more explicit, separate theories of
cognition. That is, meaningfulness, interpretation and understanding of text and context are
described here in terms of specific mental representations, such as event models, context
models and social representations. We are thus able not only to abstractly describe text
and talk, but also to explain how real language users go about producing and
understanding discourse, how their personal and socially shared beliefs affect discourse
production and how these are in turn affected by discourse. No critical account of
discourse is theoretically complete without such a cognitive interface.

Social situations
We shall be relatively brief about the third main component of our CDS approach:
society. For obvious reasons, the social dimensions of CDS usually receive more
attention from CDS researchers than its cognitive aspects, as is also shown in the other
chapters in this book. Note though that our sociocognitive theory explains how social
structures may affect (and be affected by) discourse structures via a theory of social
cognition.
An account of the role of social structures in CDS requires an analysis of both micro
(local) and macro (global) structures of society, that is, of individual social actors and
their situated interactions, on the one hand, and of social groups, movements,
organizations and institutions, as well as their relations, such as power and dominance, on
the other hand.
Note that the micromacro distinction is only analytic. In real life, social members may
experience and interpret such structures at the same time: by locally responding to a
question from a student, which may be part of the somewhat more comprehensive social
activity of giving a class, I at the same time may teach a course and reproduce the
organization of this university as well as higher education at increasingly higher,
macro and abstract levels of analysis and (diminishing degrees of) awareness.
Unlike, for instance, therapists, CDS scholars are less interested in the account of
specific discourses, interactions and situations such as, indeed, the example analysed in
this chapter. Rather, they focus on the more general ways specific discourses may be
instances of more general discourse properties and how such discourse may contribute to
social inequality, for instance by the formation of biased models and ultimately by the
formation or confirmation of ideologies.

Micro vs. macro

As is the case for micro sociologists, discourse analysts deal primarily with text and talk, and hence with a typical
micro dimension of society. Indeed, it is generally assumed that society and its structures as well as its
structures of inequality are locally produced by its members. Yet, I additionally assume that such local
production in interaction is possible only if members have shared social representations such as
knowledge and ideologies. In that sense, local social interaction is again enabled by a macro dimension such
as the social cognitions of collectivities. But then again, such a macro dimension is itself constructed cognitively by
the mental representations of groups of individual social actors. We thus see how for CDS, the micro and macro
dimensions of society, and their analysis, are multiply integrated, as is also embodied in the slogan that CDS is
mainly interested in the discursive reproduction of the social structure of inequality. One of the reasons CDS
research (critically) analyses specific discourses is the methodologically important fact that these are more directly
analysable than abstract structures, and that abstract structures can only be inferred (as language users do) by
studying special cases. Although the macrosocial dimensions of society (such as social inequality) will usually be
the main reasons and aims of CDS, we can only observe and analyse such abstract structures in terms of how
they are expressed or enacted locally in social practices in general, and in discourses in particular, that is, in
specific situations.

As part of an analysis of social situations, let me briefly say a few words about two
central categories of social situations: action and actors.

Discursive action as socialpolitical action

In a similar way as the abstract meanings of discourse are intricately related to beliefs
and other cognitive representations, the discursive acts accomplished in or by discourse
can also hardly be separated from the social acts that define social situations: discourse
is inherently part of both cognition and situations. Indeed, discursive acts are by
definition also social acts. However, since not all social acts are discursive, we need
more than just an analysis of speech acts, such as assertions, promises or threats, or
typical discursive interactions such as turn-taking, interruptions, agreeing, or the opening
and closing of a conversation. There are also a large number of social acts that are the
conditions, consequences, or implications of discursive (verbal) interaction.
Thus, holding a speech in parliament may involve a sequence of speech acts such as
assertions, questions, or accusations, as well as conversational moves and strategies such
as responding to critique, agreeing with members of your own party, refusing to be
interrupted, and many more. In this conversational aspect, parliamentary debates are not
fundamentally different from other forms of public dialogue. However, what is typical
and characteristic is that by such conversational moves, participants globally engage in
legislation, government support or opposition, representation of the voters, and other
global political actions. Similarly, by speaking negatively about refugees in order to
persuade parliament to enact tough legislation against immigrants, this at the same time
implies the social act of derogation and discrimination. Even quite local moves, such as
denials in disclaimers (I have nothing against blacks, but ) may be part of a larger
socialpsychological strategy of positive self-presentation (ingroup celebration) and
outgroup derogation which is so typical of contemporary racisms, especially of the elite.
In other words, in CDS, the action-analysis of text and talk is not limited to a study of
typically discursive doings but also examines the ways in which discursive acts and
structures are deployed in the enactment of broader social and political acts, especially
those that are part of systems of dominance (or resistance against dominance).
The same is true for the embedding of discursive interaction in broader social and
political acts, and the study of the social conditions and consequences of discourse. Thus,
an immigration debate in parliament presupposes the global acts of immigration, and may
globally result in keeping refugees out, and even more locally it may be a consequence of
a political act of government, for instance a decision to close the borders for a specific
group of refugees.

Actors

Similar remarks hold for the actors that define situations as participants in various roles,
such as communicative roles (various kinds of producers or recipients of text or talk),
social roles such as friends and enemies, occupational roles such as politicians, or
political roles such as members of parliament or members of a party. Note that in the
same way as an action may be defined at different levels, and thus relate discursive acts
(such as denials) with social acts (such as discrimination), actors may also at the same
time engage in various identities at the same time, although some identities or group
affiliations will be stronger and more salient than others in a particular context. As we
have emphasized before in the model theory of context, a relevant situational analysis of
discourse does not abstractly examine all the possible identities of speakers or recipients,
but only the locally relevant or more prominent ones, and how these affect or are affected
by discourse.
Note that a local actor analysis of discourse situations at the same time involves an
interface with societal structures: speakers act as members of various social groups, and
we thus have an obvious link between the macrostructures of groups and the
microstructures of interaction, namely via the relation of membership. A similar
relationship was established in mental models, namely between personal beliefs and the
socially shared beliefs of groups.

Societal structures
Local situations of interaction enact, manifest, or challenge global societal structures.
Participants speak and listen as women, mothers, lawyers, party members, or company
executives. Their actions, including their discursive actions, realize larger social acts and
processes, such as legislation, education, discrimination and dominance, often within
institutional frameworks such as parliaments, schools, families, or research institutes.
CDS is mainly interested in the role of discourse in the instantiation and reproduction
of power and power abuse (dominance), and hence is particularly interested in the
detailed study of the interface between the local and the global, between the structures of
discourse and the structures of society. We have seen that such links are not direct, but
need a cognitive and an interactional interface: social representations, including attitudes
and ideologies, are often mediated by mental models in order to show up in discourse,
and such discourse has social effects and functions only when it in turn contributes to the
formation or confirmation of social attitudes and ideologies. White group dominance can
only be implemented when white group members actually engage in such derogating
discourse as an instance of discrimination. Racism and sexism are thus not merely
abstract systems of social inequality and dominance, but actually reach down in the
forms of everyday life, namely through the beliefs, actions and discourses of group
members.

Similar remarks have been made in the analysis of our sample text. In order to fully understand and explain (the
structures of) this text, we not only need to spell out its cognitive and contextual conditions and consequences, but
also the broader societal structures on which such cognitions and contexts are ultimately based, and which at the
same time they enable, sustain and reproduce. We have seen how throughout the text and at all levels, the
negative opinion about the US Government in the Microsoft case is linked with the overall neoliberal ideology of a
free market, in which creative businessmen are the heroes and the government (and its justice system) the
enemies, against whose attacks the Center plays its specific role of defender of capitalist values. That is, the
ideology, as implemented in the mental models constructed for the Microsoft case and as more or less directly
expressed in the text, needs to be linked to societal groups, organizations, structures and relationships of power.
Indeed, the current text is in that respect just one of the myriad of (discursive and other) actions of the business
community in its power struggle with the State. It is only at the highest level of societal analysis that we are able
to fundamentally understand this text, its structures and functions.

It is this permanent bottom-up and top-down linkage of discourse and interaction with
societal structures that forms one of the most typical characteristics of CDS. Discourse
analysis is thus at the same time cognitive, social and political analysis, but focuses rather
on the role discourses play, both locally and globally, in society and its structures.
The relevant relationships run both ways. Societal structures such as groups and
institutions, as well as overall relations such as power or global societal acts such as
legislation and education, provide the overall constraints on local actions and discourse.
These constraints may be more or less strong, and run from strict norms and obligations
(for instance, as formulated in law, such as the acts of judges or MPs), to more flexible or
soft norms, such as politeness norms.
These global constraints may affect such diverse discourse properties as interaction
moves, who controls turn-taking or who opens a session, speech acts, topic choice, local
coherence, lexical style or rhetorical figures. And conversely, these discourse structures
may be heard as (interpreted as, count as) actions that are instances or components of
such very global societal or political acts as immigration policy or educational reform.
It is precisely in these macromicro links that we encounter the crux for a critical
discourse analysis. Merely observing and analysing social inequality at high levels of
abstraction is an exercise for the social sciences and a mere study of discourse
grammar, semantics, speech acts or conversational moves, the general task of linguists,
and discourse and conversation analysts. Social and political discourse analysis is
specifically geared towards the detailed explanation of the relationship between the two
along the lines sketched above.

FURTHER READING
Since the topics dealt with in this chapter would require a vast number of references, I have not made specific
references in the text other than where I concretely refer to some of my earlier work, of which this chapter is a
sample. Obviously, many of the notions dealt with in this chapter have been based on, or inspired by, the work of
others. Thus, my ideas on CDS have been influenced by the many publications of the other scholars represented
in this volume, and I may for the sake of space simply refer to their chapters for detailed references. For the other
notions used in this chapter, see the following references and recommended readings:

Cognition, social cognition, memory, mental models, text processing : Augoustinos and Walker (1995);
Fiske and Taylor (1991); Graesser et al. (2003); Johnson-Laird (1983); Kintsch (1998); Moscovici (2000); Tulving
(1983); Van Dijk and Kintsch (1983); Van Oostendorp and Goldman (1999).
Context: Auer (1992); Duranti and Goodwin (1992); Gumperz (1982); Van Dijk (2008a, 2009).
Discourse and conversation structures: Schiffrin et al. (2001); Ten Have (1999); van Dijk (1997; 2007a,
2007b).
Social situation analysis: Argyle et al. (1981); Goffman (1970); Scherer and Giles (1979).
The analysis of social structure and its relations to discourse and cognition: Alexander et al. (2004);
Boden and Zimmerman (1991); Wuthnow (1989).
Discourse and ideology: Van Dijk (1998).
Discourse and power: Van Dijk (2008b).
Corporate and organizational discourse: Grant et al. (2004).
 Neoliberalism: Rapley (2004).

Appendix
A Petition Against the Persecution of Microsoft
Sign the Petition International Version (for non-US residents)

To: Members of Congress, Attorney General Janet Reno and President Bill Clinton.

Fellow Americans:

The Declaration of Independence proclaims that the governments fundamental purpose is

to protect the rights of the individual, and that each individual has an inalienable right to
the pursuit of happiness. Throughout Americas history, this noble idea has protected the
individuals right to pursue his own happiness by applying his energy to productive work,
trading the products of his effort on a free market and rising as far as his abilities carry
him.
Over the past century, however, this freedom has been under attack, and one notorious
avenue of this attack has been the antitrust laws. Under the guise of protecting the
public, these laws have allowed envious competitors and power-hungry officials to
attack successful businessmen for the crime of being successful. It has led to the ugly
spectacle of the creative geniuses of the business world the men who have made this
country great being branded as oppressive tyrants, whose hard-won business empires
must be broken to pieces and subjected to the control of government regulators.
The Justice Departments current suit against Microsoft is the latest example of this
trend. It is based on envy for the productive ability of Microsoft and its founder, Bill
Gates. The result of this suit, if successful, will be to deprive Mr. Gates of his right to
control his own company, and to deprive the company of its ownership and control of its
own products.
The Justice Departments case and indeed the entire edifice of antitrust law is
based on the bizarrely inverted notion that the productive actions of individuals in the free
market can somehow constitute force, while the coercive actions of government
regulators can somehow secure freedom.
The truth is that the only kind of monopoly that can form in a free market is one based
on offering better products at lower prices, since under a free market even monopolies
must obey the law of supply and demand. Harmful, coercive monopolies are the result,
not of the operation of the free market, but of government regulations, subsidies, and
privileges which close off entry to competitors. No business can outlaw its competitors
only the government can.
We hold that Microsoft has a right to its own property; that it has the authority,
therefore, to bundle its properties including Windows 95 and Internet Explorer in
whatever combination it chooses, not by anyones permission, but by absolute right. We
hold that to abridge this right is to attack every innovators right to the products of his
effort, and to overthrow the foundations of a free market and of a free society.
We do not want to live in a country where achievement is resented and attacked, where
every innovator and entrepreneur has to fear persecution from dictatorial regulators and
judges, enforcing undefined laws at the bidding of jealous competitors. We realize that
our lives and well-being depend on the existence of a free market, in which innovators
and entrepreneurs are free to rise as far as their ability can carry them, without being held
down by arbitrary and unjust government regulations.
As concerned citizens, we ask that the Justice Departments case against Microsoft be
dismissed. We call for a national debate over the arbitrary and unjust provisions of the
antitrust laws and for an end to the practice of persecuting businessmen for their success.

Note
1 This text is a new, shorter and partly rewritten version of the authors contribution to Wodak, R. and Meyer, M.
(eds) (2002) Methods of Critical Discourse Analysis. London: Sage, pp. 95-120.
4
The Discourse-Historical Approach (DHA)

Martin Reisigl and Ruth Wodak

Introducing key concepts and terms
Some tools of analysis and principles of DHA
Approaching the analysis of discourses about climate change
Conclusions

Introducing key concepts and terms

We start our chapter by introducing the notions of critique, ideology and power.
These three concepts are constitutive for every approach in CDA, albeit frequently
employed with different meanings. Therefore, it is important to clarify how they are
conceptualized in the DHA. We then proceed with the delineation of other terms
significant for our purposes, such as discourse, genre, text, recontextualization,
intertextuality and interdiscursivity.
The second section summarizes some analytical tools and general principles of the
DHA, while in the third section, we illustrate our methodology step by step by focusing
on discourses about climate change. In the final section, we refer to the strengths and
limitations of the DHA and point to future challenges for the field.

Critique, ideology and power

Three concepts figure indispensably in all variants of CDA: critique, power, and
ideology.
Critique carries many different meanings: some adhere to the Frankfurt School, others
to a notion of literary criticism, some to traditional Marxist notions. Adhering to a
critical stance should be understood as gaining distance from the data (despite the fact
that critique is mostly situated critique), embedding the data in the social context,
clarifying the political positioning of discourse participants, and having a focus on
continuous self-reflection while undertaking research. Moreover, the application of
results is aspired to, be it in practical seminars for teachers, doctors and bureaucrats, in
the writing of expert opinions or in the production of school books.
The DHA adheres to the socio-philosophical orientation of critical theory. 1 As such, it
follows a concept of critique which integrates three related aspects (see Reisigl, 2003:
7882; Reisigl and Wodak, 2001: 3235 for extended discussions):

1. Text or discourse-immanent critique aims at discovering inconsistencies, self-

contradictions, paradoxes and dilemmas in the text-internal or discourse-internal
structures.
2. Socio-diagnostic critique is concerned with demystifying the manifest or latent
persuasive or manipulative character of discursive practices. Here, we make use
of our contextual knowledge and draw on social theories as well as other theoretical
models from various disciplines to interpret the discursive events.
3. Future-related prospective critique seeks to contribute to the improvement of
communication (for example, by elaborating guidelines against sexist language use
or by reducing language barriers in hospitals, schools and so forth).

It follows from our understanding of critique that the DHA should make the object under
investigation and the analysts own position transparent and justify theoretically why
certain interpretations and readings of discursive events seem more valid than others.
Thompson (1990) discusses the concept of ideology and its relationships to other
concepts and especially to aspects of mass communication thoroughly. He points out that
the notion of ideology has been given a range of functions and meanings since it first
appeared in the late 18th century in France. For Thompson, ideology refers to social
forms and processes within which, and by means of which, hegemonic symbolic forms
circulate in the social world.
Ideology, for the DHA, is seen as an (often) one-sided perspective or world view
composed of related mental representations, convictions, opinions, attitudes and
evaluations, which is shared by members of a specific social group. Ideologies serve as
an important means of establishing and maintaining unequal power relations through
discourse: for example, by establishing hegemonic identity narratives, or by controlling
the access to specific discourses or public spheres (gate-keeping). In addition,
ideologies also function as a means of transforming power relations more or less
radically. Thus, we take a particular interest in the ways in which linguistic and other
semiotic practices mediate and reproduce ideology in a variety of social institutions. One
of the aims of the DHA is to demystify the hegemony of specific discourses by
deciphering the ideologies that establish, perpetuate or fight dominance.
For the DHA, language is not powerful on its own it is a means to gain and maintain
power by the use powerful people make of it. This explains why the DHA critically
analyses the language use of those in power who have the means and opportunities to
improve conditions.
Power relates to an asymmetric relationship among social actors who assume
different social positions or belong to different social groups. Following Weber (1980:
28), we regard power as the possibility of having ones own will within a social
relationship against the will or interests of others. Some of the ways in which power is
implemented are actional power (physical force and violence), the control of people
through threats or promises, an attachment to authority (the exertion of authority and
submission to authority) and technical control through objects, such as means of
production, means of transportation, weapons, and so on (see Popitz, 1992).
Power is legitimized or de-legitimized in discourses. Texts are often sites of social
struggle in that they manifest traces of differing ideological fights for dominance and
hegemony. Thus, we focus on the ways in which linguistic forms are used in various
expressions and manipulations of power. Power is discursively exerted not only by
grammatical forms, but also by a persons control of the social occasion by means of the
genre of a text, or by the regulation of access to certain public spheres.

Discourse, text, context

By employing the DHA, we investigate multifaceted phenomena in our societies. This

implies that the study of (oral, written, visual) language necessarily remains only a part of
the whole enterprise hence, our research must be interdisciplinary. Moreover, in order
to analyse, understand and explain the complexity of the objects under investigation, we
consider many different and accessible sources of data (in respect of external constraints
such as time, funding, etc.) from various analytical perspectives. Thus, we follow the
principle of triangulation, which implies taking a whole range of empirical
observations, theories and methods as well as background information into account (see,
for example, Heer et al., 2008; Wodak, 2007; Wodak et al., 1999). The specific choices
depend on the specific problem in this chapter, on controversies on climate change.
We consider discourse to be:

a cluster of context-dependent semiotic practices that are situated within specific

fields of social action
socially constituted and socially constitutive
related to a macro-topic
linked to the argumentation about validity claims such as truth and normative validity
involving several social actors who have different points of view.

Thus, we regard (a) macro-topic-relatedness, (b) pluri-perspectivity and (c)

argumentativity as constitutive elements of a discourse.2
The question of delimiting the borders of a discourse and of differentiating it from
other discourses is intricate: the boundaries of a discourse, such as a discourse on
global warming or climate change, are partly fluid. As an analytical construct, a
discourse always depends on the discourse analysts perspective. As an object of
investigation, a discourse is not a closed unit, but a dynamic semiotic entity that is open to
reinterpretation and continuation.
Furthermore, we distinguish between discourse and text: texts are parts of
discourses. They make speech acts durable over time and thus bridge two dilated speech
situations, i.e. the situation of speech production and the situation of speech reception. In
other words, texts be they visualized and written or oral objectify linguistic actions
(Ehlich, 1983).
Texts can be assigned to genres. A genre may be characterized as a socially ratified
way of using language in connection with a particular type of social activity (Fairclough,
1995a: 14). Consequently, a manifesto on combating global warming proposes certain
rules and expectations according to social conventions, and has specific social purposes.
A discourse on climate change is realized through a range of genres and texts, for example
TV debates on the politics of a particular government on climate change, guidelines to
reduce energy consumption, speeches or lectures by climatologists.
The DHA considers intertextual and interdiscursive relationships between
utterances, texts, genres and discourses, as well as extralinguistic social/sociological
variables, the history of an organization or institution, and situational frames. While
focusing on all these relationships, we explore how discourses, genres and texts change in
relation to sociopolitical change.
Intertextuality means that texts are linked to other texts, both in the past and in the
present. Such connections are established in different ways: through explicit reference to
a topic or main actor; through references to the same events; by allusions or evocations;
by the transfer of main arguments from one text to the next, and so on. The process of
transferring given elements to new contexts is labelled recontextualization: if an element
is taken out of a specific context, we observe the process of decontextualization; if the
respective element is then inserted into a new context, we witness the process of
recontextualization. The element (partly) acquires a new meaning, since meanings are
formed in use (see Wittgenstein, 1989). Recontextualization can, for instance, be
observed when contrasting a political speech with the selective reporting of the speech in
various newspapers. A journalist will select specific quotes which best fit the general
purpose of the article (e.g. commentary). The quotations are thus de-and recontextualized,
i.e. newly framed. They can partly acquire new meanings in the specific context of press
coverage.
Interdiscursivity signifies that discourses are linked to each other in various ways. If
we conceive of discourse as primarily topic-related (as discourse on x), we will
observe that a discourse on climate change frequently refers to topics or subtopics of
other discourses, such as finances or health. Discourses are open and often hybrid; new
subtopics can be created at many points.
Field of action (Girnth, 1996) indicates a segment of social reality which constitutes
a (partial) frame of a discourse. Different fields of action are defined by different
functions of discursive practices. For example, in the arena of political action, we
differentiate among eight political functions as eight different fields (see Figure 4.1). A
discourse about a specific topic can find its starting point within one field of action and
proceed through another one. Discourses then spread to different fields and relate to or
overlap with other discourses.
We represent the relationship between fields of action, genres and macro-topics in the
area of political action as follows (Figure 4.1):3

FIGURE 4.1 Fields of political action, political genres and discourse topics (see Reisigl, 2007: 3435)

Figure 4.2 further illustrates the interdiscursive and intertextual relationships between
discourses, discourse topics, genres and texts.
FIGURE 4.2 Interdiscursive and intertextual relationships between discourses, discourse topics, genres and texts

In this diagram, interdiscursivity is indicated by the two big overlapping ellipses.

Intertextual relationships are represented by simple bold arrows. The assignment of texts
to genres is signalled by simple arrows. The topics to which a text refers are indicated by
small ellipses with simple dotted arrows; the topical intersection of different texts is
indicated by the overlapping small ellipses. Finally, the specific intertextual relationship
of the thematic reference of one text to another is indicated by simple broken arrows.
Several of these relationships can be illustrated with respect to our case study, if we
choose three texts of the Czech president Vclav Klaus, a fervid opponent of
environmentalist warnings on the dangers related to global warming. On 16 March 2007,
Klaus gave a speech entitled Innsbruck Speech: Three Threats of our Freedom on the
occasion of receiving an honorary doctorate by the University of Innsbruck (= text 1 that
belongs to the epideictic genre of speeches of gratitude). This speech is first and
foremost a part of the discourse about freedom and liberalism (= discourse A), but to a
certain extent also of the discourse about climate change (= discourse B), because
climate change and environmentalism are important topics of the speech.
Three days later, on 19 March 2007, Klaus answered five questions from Republican
politicians of the House of Representatives of the US Congress, in the Committee on
Energy and Commerce, on issues of mankinds contribution to global warming and
climate change. This rather formalized pseudo-dialogic text (= text 2) was realized as a
written letter (dating from 19 March 2007) that was designated to answer the questions
which were formulated in a letter (dating from 6 March 2007) signed by the Ranking
Member of the Committee on Energy and Commerce, Joe Barton, and the Ranking
Member of the Subcommittee on Energy and Air Quality, J. Dennis Hastert. This letter
shares several topics with text 1, relating to climate change, environmentalism and the
threat of freedom, and is also part of various other discourses, for example the discourse
on energy economy. Text 2, thus, is a somewhat hybrid genre, as it adopts features of both
the (sub)genre of a written expert interview and the (sub)genre of a written letter which
resembles the (sub)genre of parliamentary questions and answers.
On 14 June 2007, Vclav Klaus published an article in the Financial Times entitled
Freedom, not climate, is at risk. This text (= text 3) is again part of both discourse A
and discourse B. It picks up many of the topics of text 2 but, of course, manifests
characteristics of another genre the newspaper commentary. This text elicited many
reactions and questions by readers, some of which were published in the newspaper on
21 June 2007, and also answered by Klaus (see www.klaus.cz/klaus2/asp/-default.
asp?lang=EN&CatID=YJrRHRsP).
Our triangulatory approach is based on a concept of context which takes into account
four levels:

1. the immediate, language or text-internal co-text and co-discourse

2. the intertextual and interdiscursive relationship between utterances, texts, genres and
discourses
3. the extralinguistic social variables and institutional frames of a specific context of
situation
4. the broader sociopolitical and historical context, which discursive practices are
embedded in and related to.

In our analysis, we orient ourselves to all four dimensions of context in a recursive

manner (see also Wodak, 2007, 2008a).

Some tools of analysis and principles of DHA

The DHA is three-dimensional: after (1) having identified the specific contents or topics
of a specific discourse, (2) discursive strategies are investigated. Then (3), linguistic
means (as types) and the specific, context-dependent linguistic realizations (as tokens)
are examined.
There are several strategies which deserve special attention when analysing a specific
discourse and related texts (see Step 5 below). Heuristically, we orientate ourselves to
five questions:

1. How are persons, objects, phenomena/events, processes and actions named and
referred to linguistically?
2. What characteristics, qualities and features are attributed to social actors, objects,
phenomena/events and processes?
3. What arguments are employed in the discourse in question?
4. From what perspective are these nominations, attributions and arguments expressed?
 5. Are the respective utterances articulated overtly; are they intensified or mitigated?

According to these five questions, we elaborate five types of discursive strategies. By

strategy, we generally mean a more or less intentional plan of practices (including
discursive practices) adopted to achieve a particular social, political, psychological or
linguistic goal. Discursive strategies are located at different levels of linguistic
organization and complexity.4
TABLE 4.1 A selection of discursive strategies

The first study for which the DHA was developed analysed the constitution of antisemitic
stereotyped images, as they emerged in public discourses in the 1986 Austrian
presidential campaign of former UN general secretary Kurt Waldheim, who, for a long
time, had kept secret his National Socialist past (Wodak et al., 1990). 5 Four salient
characteristics of the DHA emerged in this research project: (1) interdisciplinary and
particularly historical aims and interests; (2) team work; (3) triangulation as a
methodological principle; and (4) an orientation towards application.
This interdisciplinary study combined linguistic analysis with historical and
sociological approaches. Moreover, the researchers prepared and presented an exhibition
about Postwar antisemitism at the University of Vienna.
The DHA was further elaborated in a number of studies of, for example, racist
discrimination against migrants from Romania and the discourse about nation and national
identity in Austria (Matouschek et al., 1995; Reisigl, 2007; Wodak et al., 1999). The
research centre Discourse, Politics, Identity (DPI) in Vienna, established by the second
author of this article (thanks to the Wittgenstein Prize awarded to her in 1996; see
www.wittgenstein-club.at), allowed for a shift to comparative interdisciplinary and
transnational projects relating to research on European identities and the European
politics of memory (Heer et al., 2008; Kovcs and Wodak, 2003; Muntigl et al., 2000).
Various principles characterizing the approach have evolved over time since the study
on Austrian postwar antisemitism. Here, we summarize ten of the most important
principles:

1. The approach is interdisciplinary. Interdisciplinarity involves theory, methods,

methodology, research practice and practical application.
2. The approach is problem-oriented.
3. Various theories and methods are combined, wherever integration leads to an
adequate understanding and explanation of the research object.
4. The research incorporates fieldwork and ethnography (study from inside), where
required for a thorough analysis and theorizing of the object under investigation.
5. The research necessarily moves recursively between theory and empirical data.
6. Numerous genres and public spaces as well as intertextual and interdiscursive
relationships are studied.
7. The historical context is taken into account in interpreting texts and discourses. The
historical orientation permits the reconstruction of how recontextualization functions
as an important process linking texts and discourses intertextually and
interdiscursively over time.
8. Categories and tools are not fixed once and for all. They must be elaborated for each
analysis according to the specific problem under investigation.
9. Grand theories often serve as a foundation. In the specific analyses, however,
middle-range theories frequently supply a better theoretical basis.
10. The application of results is an important target. Results should be made available to
and applied by experts and be communicated to the public.
Approaching the analysis of discourses about climate change

The DHA in eight steps

A thorough discourse-historical analysis ideally follows an eight-stage programme.

Typically, the eight steps are implemented recursively:

1. Activation and consultation of preceding theoretical knowledge (i.e.

recollection, reading and discussion of previous research).
2. Systematic collection of data and context information (depending on the research
question, various discourses and discursive events, social fields as well as actors,
semiotic media, genres and texts are focused on).
3. Selection and preparation of data for specific analyses (selection and downsizing
of data according to relevant criteria, transcription of tape recordings, etc.).
4. Specification of the research question and formulation of assumptions (on the
basis of a literature review and a first skimming of the data).
5. Qualitative pilot analysis (allows testing categories and first assumptions as well
as the further specification of assumptions).
6. Detailed case studies (of a whole range of data, primarily qualitative, but in part
also quantitative).
7. Formulation of critique (interpretation of results, taking into account the relevant
context knowledge and referring to the three dimensions of critique).
8. Application of the detailed analytical results (if possible, the results might be
applied or proposed for application).

This ideal-typical list is best realized in a big interdisciplinary project with enough
resources of time, personnel and money. Depending on the funding, time and other
constraints, smaller studies are, of course, useful and legitimate. Nevertheless, we
believe that it makes sense to be aware of the overall research design, and thus to make
explicit choices when devising ones own project such as a PhD thesis. In the latter case,
one can certainly conduct only a few case studies and must restrict the range of the data
collection (to very few genres). Sometimes, a pilot study can be extended to more
comprehensive case studies, and, occasionally, case studies planned at the very beginning
must be left for a follow-up project.
Because of space restrictions, we only elaborate on a few of the research stages (1, 2,
4 and especially 5) in this chapter. We have decided to focus on argumentation analysis
in our pilot study, since other strategies such as nomination and predication strategies
(which we also take into consideration) are subordinated under the persuasive aims of the
text we want to analyse.
Analysing discourses on climate change and global warming

Step 1: Activation and consultation of preceding theoretical knowledge

The overarching research question on global warming can be approached in various
ways:

(a) What does climate change mean according to the existing (scientific) literature?
What does the relevant literature convey about the relationship between climate
(b) change and modern societies, i.e. the influence of human beings on the global
climate?

A first consultation of the relevant literature supplies us with the following answers:
(a) Climate change in ordinary language use predominantly means global warming,
although other meanings can also be detected: climate change sometimes denotes
global cooling towards a new ice age and sometimes relates to a natural climatic
variation which temporarily leads to a warming or cooling. In scientific terms, climate
change refers to the change of the medial annual temperature, but also to various climatic
alterations including precipitation change, sea-level rise, the increase of extreme weather
events, ozone depletion and so on. A historical semantic reconstruction further reveals
that the scientific and political meaning of the phrase has been extended more recently:
whereas now [i]t refers to any change in climate over time, whether due to natural
variability or as a result of human activity, in the United Nations Framework Convention
on Climate Change (UNFCCC) of 1992, climate change had exclusively been related to
a change of climate that is attributed directly or indirectly to human activity [] in
addition to natural climate variability (Intergovernmental Panel on Climate Change
(IPCC), 2007a, p. 1, downloaded from www.ipcc.ch/pdf/assessment-
report/ar4/syr/ar4_syr_topic1.pdf on 9 February 2008).
(b) Most scientists consider the relationship between climate change and modern
societies to be a causal one in the sense that nature becomes more and more dependent on
human civilization, global warming being the anthropogenic consequence of the
greenhouse effect caused by the worldwide increase in the output of carbon dioxide and
other greenhouse gases.6
After this first orientation, we are now able to formulate a more general discourse-
related research question: what does climate change mean in the specific public
discourse we focus on and how are human influences on climate represented and
discussed in this discourse?
Assumptions related to this question are that the discourse will comprise different and
maybe contradicting interpretations of the nature of climate change (of its existence,
origins and consequences), of the relationship between climate and civilization, and of
possible measures against climate change. If this should be the case, we assume that such
differing discursive representations and positions make it difficult to achieve a political
compromise as a basis for political decisions. Viewed from a historical perspective, we
assume that the discourse (or some facets of the discourse) will have changed over time,
depending on a range of factors to be identified in our analyses.

Step 2: Systematic collection of data and context information

Depending on what data are accessible (by observation, audio-visual recording,
interview, research in archives) and on how much data can be analysed within the
respective research project, a range of empirical data could be collected, considering the
following criteria:

specific political units (e.g. region, nation state, international union) or language
communities
specific periods of time relating to important discursive events, which are
connected with the issue in question, for instance, climate summits or publications of
reports issued by the IPCC and their discussion in public
specific social and especially political and scientific actors (individual and
collective actors or organizations, for example, politicians with different party-
political affiliations, environmentalists, climatologists, national and international
councils on climate change, oil companies, car companies and so forth)
specific discourses in our case, discourses about climate change and particularly
about global warming
specific fields of political action, especially the formation of public attitudes,
opinions and will (e.g. relating to media coverage), the management of international
relations (e.g. relating to international summits and agreements), the fields of
political control (e.g. relating to environmentalist actions), political advertising (e.g.
relating to the promotion of the energy business), the inter-party formation of
attitudes, opinions and will (e.g. relating to the inter-party coordination of
environmental policy), the law-making procedure (e.g. relating to tax laws on carbon
emissions), and specific policy fields, such as environmental policy, energy policy,
economic policy, health policy or migration policy
specific semiotic media and genres related to environmental policy (expert reports,
election programmes, political debates inside and outside parliament, press articles,
TV interviews and TV discussion, leaflets, car advertisements and popular scientific
texts).

In the present case, we focus on one single discourse fragment: Vclav Klauss
Answers to questions from the House of Representatives of the US Congress, Committee
on Energy and Commerce, on the issue of mankinds contribution to global warming and
climate change (19 March 2007; www.klaus.cz/klaus2/asp/default.-asp?
lang=EN&CatID=YJrRHRsP). We chose this text because it is rather brief, easily
accessible on the internet, has been published in several languages (thus guaranteeing a
remarkable communicative scope), and because it relates both to Europe and the United
States. In Step 5, we will have a closer look at this text.

Step 3: Selection and preparation of data for specific analyses

When preparing the corpus for analysis, the collected data are downsized according to
specific criteria such as frequency, representativity, (proto)typicality, intertextual or
interdiscursive scope/influence, salience, uniqueness and redundancy. If it proves
necessary, oral data have to be transcribed according to the conventions determined by
the research question. As we now focus on a single text example, there is no need to
continue the discussion of this step.

Step 4: Specification of the research question and formulation of assumptions

The research question could now be specified with regard to the question of whether:

global warming is perceived as being undisputed by the discourse participants or not

climate change is seen as a natural process or as co-caused by human beings.

Moreover, the research question has to consider opposing political accusations of abuse
and manipulation, and alternative appeals for action (see, for instance, Al Gore, 2007, p.
268 ff. in contrast to Klaus, 2007: 79, 95, 97 ff.). Hence, a possible point of departure for
the further elaboration of our research question could be the analysis of controversial
positions. As critical discourse analysts, we describe and assess such contradictory
positions and their persuasive character on the basis of principles of rational
argumentation and with regard to underlying manipulative strategies.
A second point of departure could be the analysis of media coverage and of the
relationship between the reporting of scientific statements about global warming and the
media recipients knowledge. Allan Bell (1994) has already focused on such issues two
decades ago. In his case study of the discourse about climate change in New Zealand, he
analysed the relationship between the media coverage of scientific explanations and
laypersons understanding of this coverage. Bells research, which led to the insight that
the knowledge about climate change is greater among the socially advantaged than among
socially disadvantaged media users (Bell, 1994: 59), could be compared with todays
situation. The DHA pays special attention to such diachronic comparisons.
On the basis of all these concerns, our research question could be divided into the
following detailed questions:

What social (political, scientific, environmentalist, media) actors participate in the

 specific discourse on climate change? What scientific, political, environmentalist
and other positions are adopted in the different fields of political action relating to
the discourse? (We assume that different actors pursue different and often
conflicting interests.)
What role do scientists play in the triangle of science, policy/politics and the mass-
mediated public sphere? How do they translate their expert knowledge for
laypeople? How reliable are statements of scientists as epistemic and deontic
authorities? How are scientists controlled in modern democratic societies? On the
basis of what criteria do laypeople judge scientific statements? (Here, we assume
that scientists play an important role as experts both in processes of political
decision-making and in the formation of public attitudes, opinions and will.)
What role do the mass media play in the translation of expert knowledge for
laypeople? How are the media controlled in democratic societies? Is it difficult to
understand mass-media texts on climate change? (We assume that the media play a
crucial role in the formation of public attitudes, opinions and will as well as in
the field of political control.)
What validity claims of truth and normative rightness are explicitly made or
presupposed in the discourse in question? How are these claims related to party-
political and in a wider sense ideological alignments? (We assume that
different discourse participants will make different and often contradicting,
ideology-dependent claims on climate change.)
What are the main topics of the discourse? How is the influence of human beings on
the global climate represented and discussed? In more linguistic terms: what
descriptions, explanations, argumentations and narratives about the genesis,
diagnosis, prognosis and avoidance of climate change are presented or reported in
order to support the validity claims in question? What semiotic (especially linguistic
and visual) means are employed to persuade recipients? What contradictions are
constructed in the discourse? (Here, a basic assumption is that we will not find a
single and homogeneous depiction, but a wide range of differing representations
and argumentations.)
What aspects of the discourse change over time? What are the reasons for change?
(This question presupposes that there is a diachronic change involving partial
continuities and new developments.)
What other discourses does the discourse intersect with and relate to? (We assume
that the discourse in question has interdiscursive links with other discourses.)

Of course, we cannot respond to any of these questions exhaustively in this chapter. Only
large-scale interdisciplinary research projects would be able to investigate the above-
mentioned complexities adequately. Smaller projects and pilot studies will necessarily
only focus on some of these aspects.
Step 5: Qualitative pilot analysis
Our pilot study seeks to improve and differentiate the analytical instruments and to
elaborate the assumptions mentioned above. Accordingly, we restrict this qualitative pilot
investigation to one single text (which resembles the parliamentary genre of question
time) from a prominent Czech politician (President Vclav Klaus) who participates very
actively in the American, Czech and German debates on climate change and global
warming from a strictly (neo-)liberal, anti-communist as well as anti-environmentalist
point of view. The text is a hybrid mix of a written expert interview and the written
version of parliamentary questions and answers. It is constructed as a formalized, quasi-
dialogic questionanswer sequence composed of five questions and the respective
answers to representatives of the US Congress in the Committee on Energy and
Commerce. The text was originally published in English, but was reproduced in Czech
and German in Klauss book: Modr, nikoli zelen planeta (Czech) and Blauer Planet
in grnen Fesseln! Was ist bedroht: Klima oder Freiheit? (German)7.
The questioners to whom Klaus is replying remain anonymous in the books. However,
a closer look at the political and historical context shows that Klauss text, from 19
March 2007, was submitted on 21 March 2007 by the Republican J. Dennis Hastert,
Ranking Member of the Subcommittee on Energy and Air Quality, in the Hearing of the
US House of Representatives on Perspectives on Climate before the Subcommittee on
Energy and Air Quality of the Committee on Energy and Commerce and the Subcommittee
on Energy and the Environment of the Committee on Science and Technology. In the
official protocol of the Hearing (p. 137, downloaded from
frwebgate.access.gpo.gov/cgibin/
getdoc.cgi?dbname=110_house_hearings&docid=f:37579.pdf on 10 April 2008), we
read that Klauss text has the form of a letter. On pages 141146 of the protocol, we
learn that Klauss letter answers questions posed in another letter addressed to Klaus on
6 March 2007. This first letter was signed by the two Republicans Joe Barton and J.
Dennis Hastert, who seem to be responsible for the five questions. The two conservatives
invited the Czech president as well the political scientist and statistician Bjrn Lomborg
from the Copenhagen Business School, as a counterpart to the Democrat and
environmentalist Al Gore. Both Gore and Lomborg the latter is often referred to by anti-
environmentalists, because he criticizes Gore for exaggerating the possible consequences
of global warming were personally present and questioned at the Hearing on 21 March
2007, whereas Klauss letter was only submitted and accepted, without any objection. In
their letter, the two Republicans state the following reasons for inviting Klaus to respond
to their questions:

Over the past several decades, as an economist and political leader, you have developed an important perspective
on the forces that effect individual freedom and economic progress and abundance, especially as you have helped
to lead the Czech Republic out of the deadly stagnation of the former Soviet regime to become one of the fastest
growing vibrant economies in Europe. You have also taken public positions regarding the climate change debate.
We believe your perspective on the political, economic, and moral aspects of the climate change debate can be
 useful as we seek to assess the potential impacts of proposed US climate-related regulations on the economic
well-being of its citizens and their ability to contribute to future economic vitality and innovation here and abroad.

The Republican Barton is well known as a vehement sceptic of the anthropogenic thesis
on global warming. During the hearing, Barton attacked Gore for being totally wrong (p.
24) with regard to the depiction of the causal relationship of the change of CO2 levels and
the increase in temperature. Thus, it is understandable why Barton invited the sceptic
Klaus to tell the US Congress his viewpoint. In Table 4.2, the right column next to the text
lists the themes (T stands for theme) contained in the five questions and Klauss first
answer (we list the themes only for the first question; readers may attempt to generate
respective lists for the remaining questions). The third column points to plausible
argumentation schemes (i.e. topoi) and fallacious argumentation schemes (i.e. fallacies).
TABLE 4.2 Klauss answers to questions from the House of Representatives of the US Congress (text,
macrostructure/topics and argumentation)
 Within argumentation theory, topoi can be described as parts of argumentation which
belong to the required premises. They are the formal or content-related warrants or
conclusion rules which connect the argument(s) with the conclusion, the claim. As such,
they justify the transition from the argument(s) to the conclusion (Kienpointner, 1992:
194). Topoi are not always expressed explicitly, but can always be made explicit as
conditional or causal paraphrases such as if x, then y or y, because x (for more details,
see Reisigl and Wodak, 2001: 6980).
Argumentation schemes are reasonable or fallacious. If the latter is the case, we label
them fallacies. There are rules for rational disputes and constructive arguing which allow
discerning reasonable topoi from fallacies (see the pragmadialectical approach of van
Eeemeren and Grootendorst, 1992). These rules include the freedom of arguing, the
obligation to give reasons, the correct reference to the previous discourse by the
antagonist, the obligation to matter-of-factness, the correct reference to implicit
premises, the respect of shared starting points, the use of plausible arguments and
schemes of argumentation, logical validity, the acceptance of the discussions results, and
the clarity of expression and correct interpretation. If these rules are flouted, fallacies
occur. However, we must admit, it is not always easy to distinguish precisely without
context knowledge whether an argumentation scheme has been employed as reasonable
topos or as fallacy.
We analyse this text (see pp. 102109) by focusing on three aspects according to the
three dimensions of the DHA and the five strategies presented in Table 4.1:

1. First, we identify the main discourse topics of the text, extrapolating them from the
themes (listed in the second column).
2. Then, we focus on the main nomination and predication strategies to be found in
Klauss answers.
3. Third, we focus on the argumentation and more specifically on the principal claims
as well as on topoi and fallacies employed to justify these claims (listed in the third
column).

We provide an overview of the basic analytical tools for the specific analysis of
discourses about climate change by adapting the heuristic questions and strategies
presented above in Table 4.3 on pp. 112113 (the right column contains some examples
of the text from Table 4.2).
Identifying the main discourse topics is based on generalizing the established list of
themes from Table 4.2. Figure 4.3 presents the main discourse topics and the three fields
of political action in which our text is primarily located.
The diagram represents the complex topical intersections in the text. It allows for a
first impression of the fact that neoliberal and policy-related topics dominate, whereas
scientific topics are backgrounded by Klaus.
FIGURE 4.3 Selected discourse topics in the questions to, and answers of, Vclav Klaus
TABLE 4.3 Important categories to analyse discourses about climate change
Only a few aspects of nomination and predication can be addressed in this pilot study: the
six most important social actors who are discursively constructed in this text are I,
we, policymakers, environmentalists, developing countries and (the) people. The
most salient predications relating to these actors are listed in Table 4.4.
As Table 4.4 illustrates, Klaus constructs environmentalists only by means of negative
predications. Policymakers, on the one hand, appear as dependent agents wherever they
adopt environmentalist claims and, on the other, as social actors who are requested to
resist environmentalist recommendations in order to protect and foster (neo-)liberal
principles under all circumstances. The I propagates (neo)liberal beliefs and
convictions as well. The we-group, which does not play an important role in the text,
oscillates between a we of politicians (our citizens), a vague addressee-inclusive we,
a we of perceivers, a we of civilization, and a we of all terrestrials. The
developing countries appear as completely dependent on the developed countries and
as potential victims of environmentalist regulations. The people, finally, are represented
as being both endowed with liberal rights and in danger of being deprived of these rights
by environmentalist policies, and, furthermore, as potential beneficiaries of global
warming.
In addition to the (neo-)liberal patterns,8 it is worth looking at who is absent in the text,
i.e. not represented by nomination. Klaus does not name scientists as social actors. They
are only represented indirectly through the adjective scientific that is attributed to
debate. Thus, scientists are backgrounded.
The most important phenomenon in the text is climate change. It is primarily qualified
with predications such as being uncertain, slow, natural, permanent, and probably
unstoppable [], caused not by human behaviour but by various exogenous and
endogenous natural processes (such as fluctuating solar activity).9 As we will see, this
representation of climate change as a possibly permanent natural process is salient in
Klauss argumentation, since it forms the basis on which the fallacies of nature and
uncertainty are grounded, with which Klaus attempts to justify his rejection of the Kyoto
Protocol and other similar initiatives.
Various predications and nominations are relevant elements of the texts argumentation
structure. Klauss answers are highly persuasive. They contain many argumentative
devices; on the meta-linguistic level, words such as argument, argumentation,
debate, justify, conclusion and disagree explicitly indicate the persuasive
character of the text. Hence, we recommend a focus on argumentation and particularly on
content-related argumentation schemes (topoi and fallacies) for the analysis of this text, in
addition to the analysis of nominations and predications (which are linked to and form the
basis for the argumentation schemes).
The analysis of typical content-related topoi and fallacies depends on the macro-topics
of a discourse. There is an impressive amount of literature dealing with field-and content-
related argumentation schemes in various discourses (see, for example, Kienpointner,
1996; Kienpointner and Kindt, 1997; Kindt, 1992; Reeves, 1989; Wengeler, 2003). In the
present context, we refer to several topoi which are mentioned in the literature, but we
also coin new names for topoi and fallacies which occur in our specific data.
TABLE 4.4 Main social actors and predications
Topoi and fallacies in Klauss text are listed in the right column of Table 4.2. They
possess the function of justifying Klauss main claims. These claims most of which are
normative proposals of how policymakers and people in developed countries should act
are formulated from a strictly (neo-)liberal perspective:

Claim 1: Anthropogenic climate change is a dangerous argument (claim of truth).

Claim 2: If policymakers address climate change, they must and should always stick
to the principles of free society (claim of normative rightness, realized as topos or
fallacy of freedom).
Claim 3: The Kyoto Protocol or similar initiatives should be rejected (claim of
normative rightness).
Claim 4: Regulations based on the so-called precautionary principle should be
rejected (claim of normative rightness).
Claim 5: Instead of the Kyoto Protocol or similar initiatives, many small
ecologically harmful things should be avoided; in all cases, freedom and the free
market have to be protected (claim of normative rightness).
Claim 6: If policymakers accept the (maximalist) ecological demands, this will have
 negative consequences for national economies (claim of truth; topos or fallacy of
negative consequence).
Claim 7: If we have to choose between free market and cap-and-trade policies,
capand-trade policies are less important in the context of an irrational fight against
global warming (claim of truth; fallacy of secondary importance as fallacy of the
priority of free market over cap-and-trade policies).
Claim 8: Since developed countries have moral obligations towards developing
countries, they should not introduce large-emission reduction schemes (claim of
normative rightness; topos or fallacy of moral obligation).

These claims are mostly justified by fallacies. The overall structure of Klauss answers is
dominated by two fallacies:

1. The fallacy of uncertainty10 assumes that since science is uncertain in respect of the
existence, causes, consequences and avoidance of climate change, environmentalist
recommendations are not convincing and, thus, it does not make sense to follow
these recommendations.
2. The second recurrent fallacy in Klauss answers is the fallacy of nature: since
climate change is natural, ecological regulations concerning greenhouse gases are
not reasonable, but irrational and wasteful.

Both fallacies can be discredited by a topos of numbers that refers to the vast majority of
climatologists who agree that an anthropogenic climate change does exist with a very high
degree of certainty. 11 In addition, the first fallacy can be countered by the topos of risk
minimization (as a specific topos of priority): if different alternatives carry various risks,
we have to minimize the risks by choosing the alternative with minimal risks. If we
consider this argumentation scheme, Klauss refusal of the precautionary principle
appears unjustified. Furthermore, the topos of risk minimization can also be directed
against the fallacy of secondary importance (claim 7).
We must interrupt our pilot analysis here. In sum, we are able to conclude that Klauss
argumentation is highly fallacious and that there are plausible reasons to reject Klauss
neoliberal position and to accept a just limitation of human freedom for specific
ecological reasons as concerning the well-being of the human species. A more detailed
case study, which would analyse the whole Hearing, Klauss entire book, and the
political role of Klaus as supporter of the US governments negative position with respect
to international greenhouse-gas emission regulations, would differentiate this first pilot
analysis and would gain insight into the broader political and historical contexts of the
specific discourse on climate change.
Such an analysis would also focus on strategies of perspectivization, mitigation and
intensification.12 This would help to recognize the underlying ideological positioning in
this and other discourse fragments produced by Klaus. It would allow reconstructing how
the strict neoliberal perspective co-determines Klauss choice of various rhetorical,
pragmatic and argumentative devices and how frequent Klauss dogmatic anti-communist
stance leads to fallacious intensification strategies such as the one which dominates the
argumentation in the above text: the quasi-equation of environmentalism with communism
that aims to derogate many ecological positions (for more examples of perspectivization,
mitigation and intensification strategies, see Table 4.3, point 5). Moreover, a broader
study would focus on the various ways environmentalists are discursively constructed and
represented as exceptionally powerful and dangerous (topos of threat). In the text, we
analysed, for example, Klaus presupposes that ecological groups dominate and
manipulate politicians and bureaucrats. Such negative other-presentations prepare the
ground for shifting the blame and scapegoating strategies, which could eventually be
used to legitimize and explain political mistakes. Our analysis illustrates that Klaus
organizes most of his strategic discursive manoeuvres with a strict neoliberal, anti-
environmentalist and anti-communist ideological positioning in mind.

Step 6: Detailed case studies

This step consists of detailed case studies on the macro-, meso-and micro-levels of
linguistic analysis, as well as on the level of context. This step, which cannot be
illustrated in this chapter because of space restrictions, interprets the different results
within the social, historical and political contexts of the discourse(s) under
consideration.13
In the present case, this step would lead to general descriptions of the discourse on
climate change in respect of:

social actors and fields of political action

communication obstacles and misunderstandings
contradictory validity claims imbued by political or ideological orientation
salient topics and discursive features
aspects of historical change
interdiscursive relationships, particularly overlaps with other discourses (such as
discourses about globalization, migration or freedom/liberalism).

The overall interpretation would, for example, consider the question of whether the mass-
mediated discourse(s) on climate change and global warming in European states resemble
the discourses in the USA, where company lobbying frequently leads to a balance as
bias in the media coverage: the prevailing scientific consensus on the anthropogenic
influence on global warming is not represented adequately in the media. In contrast,
media coverage seems to suggest that scientists do not agree on this issue quite so
strongly (see Boykoff and Boykoff, 2004; Oreskes, 2004 [both quoted in Rahmstorf and
Schellnhuber, 2007: 83]).
The overall interpretation could further refer to Viehvers stimulating research on
various discourses on climate change (Viehver, 2003). Viehver investigated the media
coverage from 1974 to 1995. On the basis of his comprehensive case study, he
distinguished between six problem narratives about global climate change and its
definition, causes, (moral) consequences and possible reactions to it. According to
Viehver, these narratives gained different salience at different times. He observes that
currently the predominant narrative seems to be the global warming story. Stories
competing with this narrative were and still are, according to the German sociologist,
the global cooling story, the story of the climatic paradise, the story of the cyclical
sunspots, the story of the climatic change as scientific and media fiction and the story
of the nuclear winter (see Viehver, 2003: 268 ff. for more details).

Step 7: Formulation of critique

Our critique is based on ethical principles such as democratic norms, human rights and
criteria of rational argumentation. It points to intended biases in representations
(especially media coverage) and to contradictory and manipulative relationships between
discourses and power structures.
In a theoretical sense, the critique based on our empirical analysis and a theory of
discursive/deliberative democracy offers analytical parameters that evaluate the
quality of public political discourses in which collective learning and decision-
making are at stake.
In a practical sense, the critique might influence current discourses on global warming
and raise the awareness of involved social actors about the problem, more responsibility
and fallacious argumentations.

Step 8: Application of the detailed analytical results

The application of the analytical results stems from the critique. The application should
not only consist of the scholarly publication of the results. In addition, our insights should
also be made accessible to the general public (e.g. by recommendations, newspaper
commentaries, training seminars, further education courses, radio transmissions and
political advising). Such a knowledge transfer requires the recontextualization of
theory, methodology, methods and empirical results into other genres and communicative
practices. This is, of course, a challenging task.

Conclusions
The strengths of the discourse-historical approach include the following:
 its interdisciplinary orientation, which allows avoiding disciplinary restrictions
the principle of triangulation, which implies a quasi-kaleidoscopic move towards
the research object and enables the grasp of many different facets of the object under
investigation
the historical analysis, which allows transcending static spotlights and focusing on
the diachronic reconstruction and explanation of discursive change
practical applications of the results for emancipatory and democratic purposes.

The DHA relates to other CDA approaches in many aspects. However, the DHA like
any inter-or multidisciplinary enterprise should avoid the combination of theoretically
incompatible scientific (re)sources. This caveat remains one of the main theoretical
challenges. Furthermore, many new discourse-related social phenomena (such as the one
discussed in the present chapter) need to be investigated in systematic and detailed ways
from the perspective of our approach.

FURTHER READING
Muntigl, P., Weiss, G., and Wodak, R. (2000) European Union Discourses on Un/Employment: An
Interdisciplinary Approach to Employment Policymaking and Organisational Change. Amsterdam:
Benjamins.
This book presents an interdisciplinary study of EU organizations which involved fieldwork, ethnography,
interviews and the analysis of written and oral data. This study cuts across CDA, sociology, political science and
European studies.

Reisigl, M. and Wodak, R. (2001) Discourse and Discrimination: Rhetorics of Racism and Antisemitism.
London, New York: Routledge.
A comprehensive presentation of the DHA, with case studies on racist, xenophobic and antisemitic rhetoric in the
postwar Austrian context.

Wodak, R. (1996) Disorders of Discourses. London: Longman.

A collection of case studies on organizational communication, framed in an integrated investigation of hospitals,
media and bureaucracies.

Wodak, R. (2009) Politics as Usual The Construction and Representation of Politics in Action.
Basingstoke: Palgrave.
A monograph which illustrates the kind of inter-and post-disciplinary research proposed in the DHA. The
backstage of politics is juxtaposed with the analysis of media soaps about politics and politicians.

Wodak, R. and Krzy anowski, M. (eds) (2008) Qualitative Discourse Analysis in the Social Sciences.
Basingstoke: Palgrave.
An introductory text for social scientists which focuses primarily on the analysis of diverse genres (press articles,
new media, documentaries, focus groups, interviews and broadcasts). The volume is particularly useful for non-
linguists.
Notes
1 See Habermas, 1996; Horkheimer and Adorno, 1969/1991 [1944].
2 Other approaches to CDA do not explicitly link discourse with a macro-topic and more than one perspective
(see Reisigl, 2003: 91 ff.).
3 In three of the eight fields, we distinguish between attitudes, opinions and will. This distinction emphasizes the
difference in the emotional, cognitive and volitional dimensions.
4 Many of these strategies are illustrated in Reisigl and Wodak (2001). In this chapter, we will focus primarily on
nomination, predication and argumentation strategies.
5 The very first critical study which inspired the project on postwar antisemitism in Austria was Wodak et al.
(1985).
6 See IPCC, 2007a, p. 1, downloaded from www.ipcc.ch/pdf/assessment-report/ar4/syr/ar4_syr_topic1.pdf on 9
February 2008; IPCC, 2007b, p.6, downloaded from www.ipcc.ch/pdf/assessment-
report/ar4/syr/ar4_syr_topic2.pdf on 9 February 2008. See also Mller et al., 2007; Rahmstorf and Schellnhuber,
2007.
7 The German translation deviates from the English version on several points. The English translation of the
German title is: Blue planet in green bonds. What is endangered: climate or freedom?. The book has also been
translated into other languages.
8 The frequency of high-value words (miranda) such as freedom, wealth, prosperity and economic growth
fits very well into Klauss (neo-)liberal ideology.
9 Here, the assertion that climate change cannot be influenced is mitigated by probably.
10 Klaus associates this fallacy with the fallacy of superficiality: if scientists dont work seriously, but superficially,
their results are insignificant.
11 Klaus attempts to disparage this topos as a fallacy of numbers, i.e. a myth of scientific consent (Klaus, 2007:
79).
12 See Reisigl (2003: 214235); Reisigl and Wodak (2001: 8185).
13 See, for example, Muntigl et al. (2000); Reisigl and Wodak (2001);Wodak et al. (1999) for such
comprehensive studies.
5
Checks and Balances: How Corpus Linguistics can Contribute to CDA

Gerlinde Mautner
Introduction
Key concepts and a worked example
Using a reference corpus to support interpretation: a second worked example
Summary and critique
Introduction
This chapter focuses on the role that corpus linguistics can play in CDA projects. It will
introduce readers to previous work in this area, explain basic concepts and techniques,
present two worked examples and encourage critical engagement with the methodology.
Those with previous experience of corpus linguistics will be aware that it is a
methodology that uses computer support in particular, software called concordance
programs to analyse authentic, and usually very large, volumes of textual data. Its
potential usefulness for CDA, rather than for lexicography and grammar, may be less
familiar, though. Reflections on the potential of combining corpus linguistics and CDA go
back quite a long way now (e.g. Hardt Mautner, 1995), and in the 1997 edited volume on
discourse studies (Van Dijk, 1997), de Beaugrande argued that [l]arge corpuses offer
valuable support for the project of discourse analysis to return to authentic data (de
Beaugrande, 1997: 42). Still, none of the other contributors to that edition actually used
the method. Awareness of its potential does seem to be growing, however, and there has
been a spate of more recent CDA work using corpus linguistics (e.g. Baker and McEnery,
2005; Baker et al., 2007; Baker et al., 2008; Cotterill, 2001; Fairclough, 2000a; Mautner,
2007; Nelson, 2005; Orpin, 2005). Even so, it seems fair to say that the techniques of
corpus linguistics are not yet generally regarded as being at the core of CDAs
methodological canon. That the present (second) edition of the volume includes this
chapter could thus be said to reflect a change in trend.
What, then, can one expect corpus linguistics to contribute to CDA? In a nutshell, the
potential of this methodology rests on three factors:

Corpus linguistics allows critical discourse analysts to work with much larger data
volumes than they can when using purely manual techniques.
In enabling critical discourse analysts to significantly broaden their empirical base,
corpus linguistics can help reduce researcher bias, thus coping with a problem to
which CDA is hardly more prone than other social sciences but for which it has
come in for harsh and persistent criticism (e.g. Widdowson, 1995, 2004).1
Corpus linguistics software offers both quantitative and qualitative perspectives on
textual data, computing frequencies and measures of statistical significance, as well
as presenting data extracts in such a way that the researcher can assess individual
occurrences of search words, qualitatively examine their collocational
environments, describe salient semantic patterns and identify discourse functions.

This chapter cannot offer detailed step-by-step guidance on project design and execution.
For that, there are other, and arguably more suitable, sources that readers may want to
turn to, notably Baker (2006) and McEnery et al. (2006). However, a few basics will be
covered in the following section, using original sample analyses as well as cross-
referencing existing work in this area. Throughout, the emphasis will be less on technical
detail than on enabling readers to make their own informed judgements on whether the
method is right for them. There are two worked examples: the first shows how a large
reference corpus can be mined for socially relevant information, establishing a
collocational profile of a key expression from the lexis of work, namely unemployed. The
second takes a single newspaper article as its starting point, and uses large-corpus data as
an aid in interpreting what appears to be a particularly loaded expression from the
article the adjective hard-working.
Both case studies are based on the assumption that language and the social are
inextricably, and dialectally, linked. In other words, the way in which labels, in this case
unemployed and hard-working, are used reflects social attitudes, perspectives and
categorizations. And the labels, in turn, shape the way in which social structures and
relationships are perceived. By referring to a person or group as unemployed, one cannot
help implying that being employed is the desired default, just as hard-working comes
with a host of positive connotations directly related to an essentially capitalist work
ethos.
Different approaches to discourse, and concomitant definitions of the term, exist in
abundance, as do various notions of what it means to carry out critical analysis (see
Wodak, 2004: 198199 and Wodak, 2006b for comprehensive overviews). The
perspective adopted in this chapter is functional and constructivist (as well as
unabashedly simple). Discourse is taken to refer to authentic texts used in multi-layered
environments to perform social functions. Analysing discourse is understood as the
systematic attempt to identify patterns in text, link them to patterns in the context, and vice
versa. Doing so critically means unveiling and challenging taken-for-granted assumptions
about language and the social, as well as recognizing discourse as a potentially powerful
agent in social change.
It will not have escaped readers attention that the title of this chapter contains two
hedging devices, one modal (can) and another lexical (contribute). These correspond to
two caveats which are worth spelling out right at the beginning. One is that the usefulness
or otherwise of this method, as of any other, depends crucially on recognizing what kinds
of research questions it is suitable for tackling. With corpus linguistics, the key limiting
factors are the capabilities of the software, as well as the features mainly in terms of
composition and annotation of the electronically held corpora that are used. At the
current state of play, and considering the limitations of those tools that are sufficiently
widely available, there is a very strong bias in favour of the individual lexical item and
clusters thereof. Put simply, the word is the peg that everything else is hung on. It
follows that if the linguistic phenomenon you are interested in is in fact tied to, or at least
crystallizes around, discrete lexical items, then you are likely to find this method a boon
both as a practical and efficient time-saver, and as a powerful heuristic tool helping to
clear pathways to discovery. If, on the other hand, the phenomenon to be focused on is
one that is played out on a larger textual stage, and with varying and unpredictable lexical
realizations, then corpus linguistic methods will be of little or no help. However, at some
point or other, as soon as questions of micro-level linguistic realization are addressed,
even projects located very much at the macro end of the CDA spectrum will have
occasion to benefit from a corpus linguistic approach.
The second caveat, related to the contributing role of corpus linguistics, is that we
need to recall one of the principal tenets of what might be termed mainstream CDA
(broadly, the traditions shaped by Fairclough, Wodak and Van Dijk 2), namely that the
analyst must, precisely, look beyond the text proper in order to unearth socially
meaningful interpretations that can then be enlisted to do socially transformative work.
We need our much-famed context, history, and as firm a grasp as possible of the
politics, in the widest sense, that have a bearing on the production and reception of the
text. This social hinterland and the textual evidence before us are intricately linked, but
rarely in a fully transparent, one-to-one type of relationship; hence the idea of making
corpus linguistics contribute to CDA rather than it doing CDA of its own accord. All
the same, at an Oscar night of methods, my vote would be on corpus linguistics as Best
Supporting Actor, and the present chapter sets out to make the case for that award.
In addition, there is the added benefit that if you decide to include corpus linguistic
methods in your CDA project design, you need not in fact discard, unlearn or in any
other way throw overboard whatever more traditional methods you have grown
accustomed to using. As an ancillary method, corpus linguistics is flexible and
unobtrusive, and if handled appropriately, will enrich but not prejudice the rest of the
research design or the interpretation of the results.

Key concepts and a worked example

As with any method, researchers will want to know, first and foremost, what it can do,
what kind of data and research questions it is suitable for, and what obstacles may be
encountered when applying it. These are the concerns of the present section.

Concordancing software

Programs known as concordancers do not, by themselves, produce analyses, but

perform operations on text that make it easier for humans to analyse it. Some of the
information that concordancers provide is quantitative, such as absolute and relative
word frequencies. Programs also compute measures that indicate the relative statistical
significance of the co-occurrence of items. Examples here are t-scores, which capture
certainty of collocation, and MI (mutual information) scores, which tell us about the
strength of the bond between two items, that is, whether there is a higher-than-random
probability of the two items occurring together (Church and Hanks, 1990; Clear, 1993:
281; Hunston, 2002: 73; McEnery and Wilson, 2001: 86).3
To see how this works in practice, let us have a look at the first worked example of
this chapter: building a collocational profile of the adjective unemployed. Although this
serves as a stand-alone case study here, with a focus on method rather than content, it
could also form the nucleus of a substantive contribution to the existing body of
discourse-analytic research on unemployment (see Muntigl et al., 2000 and the 2002
special issue of Text4). As we shall see, the corpus linguistic approach allows the
researcher to work with enormous amounts of data and yet get a close-up on linguistic
detail: a best-of-both-worlds scenario hardly achievable through the use of purely
qualitative CDA, pragmatics, ethnography or systemic functional analysis.
Given the key role that the mass media play in constructing social reality, a corpus of
newspaper articles would appear to be a suitable starting point. Wordbanks Online, 5 a
multi-genre corpus of more than 500 million words of mainly British and American text,
includes a nearly 60-million-word corpus of articles from the British daily newspaper
The Times. This is the subcorpus we will turn to first. The search reveals that The Times
corpus contains 567 instances of unemployed. Table 5.1 gives the ten collocates with the
highest t-scores and MI scores respectively.6
The t-score part of the table is headed, as is invariably the case, by high-frequency
grammatical items. Here, these are an and are, with who, and and for not far behind.
Such function words, devoid of separate meaning as they are, tend not to be as
interesting to discourse analysts as to grammarians, and it is generally safe in a CDA
setting to ignore them and indeed the whole t-score rank scale. Somewhat unusually
though, the top ten by t-score here includes five content words (people, term, long,
million and workers), with the presence of both long and term probably being due to the
phrase long-term unemployed. Comparing this with the collocation list of two other,
randomly picked adjectives, happy and sad,7 we can see how unusual it is for content
words to appear so high up in the t-score rank scale: the group of top ten in the t-score list
for happy, for example, contains no content word at all, and the one for sad includes only
one (the intensifier very). For lexical items to be in the same t-score league as
grammatical ones points to a high degree of what one might call patterned bonding; that is,
of a lexical connection so formulaic that the degree of certainty with which it occurs
equals that of patterns involving grammatical items. Translated into CDA terms, the
phrase-ness of a noun group referring to people could point to the solidified discursive
construction of a social group (a necessary first step towards stereotyping).
TABLE 5.1 t-scores and MI scores for collocates of unemployed in The Times corpus of Wordbanks Online
 Turning now to the right half of Table 5.1, which lists the top ten collocates according
to Mutual Information scores, we can see which social attributes being unemployed is
associated with: unemployable, disadvantaged and homeless. Frequent nominal
collocates include several labels for marginalized, dependent and economically inactive
social groups: househusband, housewives, youths, pensioners, claimants.8 In a full-
blown study, rather than one done for demo purposes only, each of these high-frequency
collocates would be interesting entry points to the corpus. For example, it would be
worth looking at what kind of activities unemployed youths are seen to engage in (by
running a search of unemployed youths followed by a verb form), how the usage of
househusband and housewife compares, how unemployed is linked syntactically to the
negative adjectives it frequently collocates with, how quantification (cf. the collocate 4m)
contributes to establishing the unemployed as a problem group, and so on.
These sorts of questions lead us to another feature of concordance programs, one that a
discourse analyst with a predominantly qualitative mindset might get more mileage out of
than frequencies and statistics: their eponymous capacity to produce concordances. These
are extracts from the corpus, displayed in such a way that the search word or phrase (also
referred to as the node) appears in the middle of a line. The text that the extract comes
from can be accessed at all times and with a simple operation such as a double mouse
click or selecting an option from a menu bar. Accessible co-texts vary from just over 500
characters (e.g. with Wordbanks Online) to full texts (e.g. with Wordsmith Tools). Lines
can be sorted alphabetically: for example, according to the word immediately preceding
or following the search word. When sorted like this, the collocational environment of the
search word can be assessed rapidly, with frequent patterns standing out clearly. As
shown in Table 5.2, for example, the concordance of unemployed followed by and and
another adjective shows up a preponderance of items (six out of a total of nine) with a
negative semantic load: desperate, disadvantaged, divorced, homeless and
unemployable (which occurs twice).
TABLE 5.2 Occurrences of unemployed, followed by and and another adjective, in The Times corpus of
Wordbanks Online
 To make sure that this semantic pattern is not just confined to newspaper discourse, it
is worth checking the corresponding data from the British spoken corpus. The picture is in
fact very similar (Table 5.3).
TABLE 5.3 Occurrences of unemployed, followed by and and another adjective, in the British spoken corpus of
Wordbanks Online

These results are further confirmed if we extend the search to the complete 500+
million word corpus of Wordbanks Online. The adjectives joined to unemployed include
more occurrences of the items already identified in the two subcorpora (see Tables 5.2
and 5.3) as well as new and similarly negative collocates such as angry, demoralized,
destitute, disabled, dreary, drunk, excluded, poor, struggling and underprivileged. By
examining concordances, therefore, we can see more or less at a glance that the search
word, unemployed, has a so-called negative semantic aura, or semantic prosody
(Hunston, 2004: 157; Louw, 1993; Partington, 2004). Alternatively, readers of this
volume will be interested to note, this concept has been referred to as discourse
prosody, for example by Stubbs (2001: 65), in order to emphasize its role in expressing
attitudes and in establishing coherence.
Detractors of corpus-based methods could argue, of course, that one hardly needs a
huge database of text and sophisticated software to prove that being unemployed is not a
pleasant thing. On the other hand, we should not forget, first, that a fair proportion of any
empirical work is devoted, precisely, to finding evidence for the intuitively obvious.
Second, some insights may appear obvious after having emerged from data but were
nothing of the kind before. Supposedly neutral words such as cause or provide, to use one
of Stubbss examples, can in fact be heavily skewed in terms of their evaluative content
when concordance evidence is examined. Patterns are revealed that are not easily
accessible even to native speakers intuition: cause, it turns out, collocates predominantly
with unpleasant events, such as damage, death, disease or trouble, whereas provide
occurs with desirable things, such as care, help, money or service (Stubbs, 2001: 65).
Thus, if a speaker or writer uses provide, that choice in itself implies that what is
provided is being presented as good rather than bad. Third, the concordancer does more
than highlight the evaluative polarity, be it good or bad, of an items collocational
environment. Collocates may also turn out to belong to a class of words that share a
semantic feature, that is, the search word may have a particular semantic preference
(Stubbs, 2001: 88). For example, Baker (2006: 79, 87) concludes from corpus evidence
that refugees has a semantic preference for quantification, collocating frequently with
numbers and phrases such as more and more (see also Baker and McEnery, 2005). In a
study using a similar approach, Mautner (2007) shows that elderly often co-occurs with
items from the domains of care, disability and vulnerability. By the same token, if we
return to our concordance output related to unemployed and its coordinated adjectives,
we can see that, broadly speaking, these denote either social states (e.g. available for
work, excluded, immigrant, nomadic, unemployable, unpaid), negative emotions (e.g.
angry, bored, depressed ), or indeed a condition at the interface of both (unloved). The
collocational profile thus points to the twin nature of unemployment as a social
phenomenon with a manifestly psychological impact on individuals.
Taken together, then, semantic preference and discourse prosody show us what kinds of
social issues a particular lexical item is bound up in, and what attitudes are commonly
associated with it. Importantly, collocational patterns are not merely instantiated in text,
but also cling to the lexical items themselves. Words which are co-selected, Tognini-
Bonelli (2001: 111) reminds us, do not maintain their independence. If a word is
regularly used in contexts of good news or bad news or judgement, for example, it carries
this kind of meaning around with it.
Finally, some software packages, such as Wordsmith Tools, allow the analyst to
compare word lists compiled from different corpora, determining which words are
statistically more frequent keywords in a corpus (Baker, 2006: 125; Baker et al., 2008:
278; Mulderrig, 2006: 123). Faircloughs (2000a) study, for example, focuses on the
keywords of New Labour words, that is, that are more frequent in New Labour material
than in earlier Labour texts, and more frequent, too, than in general corpora (Fairclough,
2000a: 17).
To sum up, concordancing software offers the above features which are useful for CDA
applications (see Table 5.4).
TABLE 5.4 Tools and types of linguistic evidence provided by concordance software

Quantitative
Frequency lists
evidence
Comparisons of wordlists, giving information on relative frequency
(keyness)

Measures of statistical significance:

t-score
 Mutual Information (MI) score

Concordance lines sorted alphabetically, enabling the researcher to

identify:
Qualitative
evidence semantic preference
semantic prosody

Corpus design issues

These days, when linguists talk about a corpus, they generally refer to a collection of (1)
machine-readable (2) authentic texts [] which is (3) sampled to be (4) representative
of a particular language or language variety (McEnery et al., 2006: 5, original italics).
Let us look at each of these four characteristics in turn, with an eye to whatever specific
implications they may have for applications in CDA.
Machine readability is the obvious prerequisite for analysing language with the
concordancing software described in the previous section. This sounds straightforward
enough, but when it is coupled with the second feature authenticity issues of data
quality arise that critical discourse analysts will want to address. Standard concordancers
need plain text files, stripped of formatting, layout and accompanying visuals. While
traditional lexico-syntactic research does not see this as a loss, critical discourse analysts
will (or should). After all, it is one of the foundational assumptions of discourse analysis,
whether of the critical persuasion or not, that meaning-making works simultaneously on
several levels, including the non-verbal. Elements of textual design, including typography,
colour and textimage relationships, are not merely embellishments, but play an integral
role in making text function as socially situated discourse (van Leeuwen, in this volume).
The semiotic reduction that concordancing inevitably entails (Koller and Mautner, 2004)
need not jeopardize the validity of ones analyses, but there ought to be adequate
safeguards to ensure that whatever is lost along the way can be salvaged at a later stage.
In mundanely practical terms, this means collecting and storing hard-copy or scanned
originals for future reference, to be drawn upon should multimodality become an issue.
Likewise, audio or video recordings of spoken data ought to be preserved, so that
contextual clues lost through transcription and conversion into machine-readable format
can be retrieved if and when necessary.
As far as criteria (3) and (4) are concerned sampling and representativeness the
requirements for ensuring methodological rigour are basically no different here than they
are for other approaches (see Mautner, 2008). The first step is to identify the universe of
possible texts (Titscher et al., 2000: 33), while the second involves sampling. This can
be random (i.e. done by first numbering the texts in the universe and then selecting those
with the numbers that a random number generator has picked out). Alternatively, it may be
guided by criteria that are applied systematically and, in a top-down selection process,
narrow down the corpus to a manageable size (e.g. take one article about Topic A from
newspapers B and C published each week between dates X and Y). There is a third
sampling method, common in qualitative research but unlikely to be suitable for corpus
linguistic work, which uses a cyclical process, building a small and homogeneous corpus,
then analysing it and adding to it on the basis of the first results (Bauer and Aarts, 2000:
31). The process is repeated until saturation has been reached a situation, that is,
where adding new data does not yield any new representations (Bauer and Aarts, 2000:
34). The problem with this procedure from a corpus linguists point of view is the flip
side of what makes it appealing to the purely qualitative researcher: that you stop
collecting data as soon as what you find is simply more of the same. In corpus linguistics,
the frequency of an item or structure is taken to be a key indicator of its significance. If
you stop adding text to your corpus as soon as repetition becomes apparent, you are
effectively closing off any frequency-based line of inquiry. This may well be an
acceptable decision to take in a particular project; after all, the qualitative analysis of
concordance lines is as important and valuable as the quantitative inquiry that
concordancing software allows. But it would have to be a decision taken with full
awareness of the loss involved. Certainly, care must be taken not to indulge in hasty
judgements about what can be excluded from the corpus on the grounds that it is similar
to what is already there. Such rashness can easily defeat the whole purpose of the corpus-
building exercise, the point of which is, in a sense, to outwit the analyst who may be
tempted to know the data before rather than after the analysis.
Summing up, corpus design involves the following issues (Figure 5.1):

FIGURE 5.1 Corpus characteristics and issues of corpus design

Types of corpora and data capture

Corpora come in many shapes and sizes. There are huge, multi-million-word corpora
such as The British National Corpus9 (BNC) and Wordbanks Online, from which the
unemployed example in the previous section was taken. These are ready-made,
commercially available, and each comes with bespoke software unique to it (usually a
source of frustration if you want to use both corpora simultaneously). Both were the result
of large-scale projects spanning many years and involving teams of linguists and
computer experts. In CDA, such corpora are ideal for painting on a very large canvas,
investigating how broader social issues are reflected in the genres and discourses
represented in the corpus (such as fiction, newspapers and spoken dialogue). This
approach is used in studies such as those of Krishnamurthy (1996) on racism and Mautner
(2007) on ageism.
At the other end of the spectrum, there are much smaller, do-it-yourself (DIY)
corpora (McEnery et al., 2006: 71), purpose-built by individual researchers or small
teams to investigate specific research questions. Issues of size apart, corpora may also be
classified according to whether they are synchronic, reflecting a language variety at any
given point in time, or diachronic, reflecting historical development. They can be general,
including a wide variety of genres and media, or specialized, focusing on a particular
genre (e.g. corporate mission statements), a particular medium and topic (e.g. web-based
texts on disability rights), a particular genre and topic (e.g. parliamentary speeches on
global warming), or a particular topic in various genres and media (e.g. the Evolution
vs. Creationism debate in sermons, newspaper articles and web logs). Corpora may also
differ in terms of which meta-linguistic information is encoded with the text. Critical
discourse analysts will be particularly keen to insert codes describing extratextual
information (a procedure known as corpus mark-up), such as text type, speakers or
writers sociolinguistic characteristics, or indeed any feature that is relevant for a
particular research question. Mark-up plays a key role in allowing the analyst to relate the
examples shown up by corpus searches back to their original contextual environments
(McEnery et al., 2006: 2223). In addition, corpora may have undergone what is called
annotation, a process of inserting, for example, parts-of-speech tags, prosodic or
semantic information (Baker, 2006: 3842). Table 5.5 summarizes the main types of
corpora.
TABLE 5.5 Main types of corpora
 In building corpora of any size, the World Wide Web has emerged as a key resource.
With the exception of spontaneous spoken language (which, admittedly, is a very
significant exception), the web offers a huge variety of text and sheer unlimited amounts
of it. Cut-and-paste procedures have greatly facilitated the process referred to as data
capture (Baker, 2006: 3138; McEnery et al., 2006: 73). As a result, corpora running to
hundreds of thousands of words can be assembled within the space of a few weeks. (On
the copyright issues involved, see McEnery et al., 2006: 7779.) More importantly still,
this can be done by individual scholars without the help of battalions of research staff,
access to whom is normally dependent on massive research grants and/or a position high
up in the institutional pecking order. Seen from this angle, enlisting corpus methods also
has a democratizing effect on critical research.
Thus, in terms of sheer corpus availability, most critical discourse analysts interests
will be catered for by the web (which makes it all the more surprising that until very
recently, only comparatively few CDA projects were actually based on online material
see Mautner, 2005). Yet in spite of this new abundance, many of the old questions remain,
which brings us back to the design issues raised in the previous section. Modern
technology may have reduced the manual drudgery of corpus-building, but not the need for
brainpower to make the right choices. To recap briefly, the key issues involved all of
them well known from traditional corpus-gathering are: first, ensuring an adequate fit
between the corpus and the research question(s) to be tackled, and, second, the triangle of
issues concerning corpus size, homogeneity and representativeness.

Using a reference corpus to support interpretation: a second worked

example
It is true, as the previous section explained, that corpus linguistic methods can be applied
either to smaller, purpose-built corpora or to large reference corpora. Nonetheless,
critical discourse analysts are likely to get most mileage out of a combination of the two
approaches. Arguably, the smaller your own corpus, the more important it becomes to
validate interpretations by checking them against evidence from larger corpora. We might
feel, for example, that a certain word or phrase is used in a particular text because it
carries a particular evaluative load, but is this borne out by data from a reference corpus
reflecting general language use? What kind of collocational baggage do words
occurring in the text carry in wider universes of discourse? It is through comparative
evidence from large corpora that suspicions can be confirmed or rejected, thus
safeguarding against overinterpretation and underinterpretation (OHalloran and Coffin,
2004). It has to be conceded, of course, that making sense of this comparative evidence
still involves acts of interpretation on the part of the analyst. Neither the quantitative nor
the qualitative evidence that large corpora lay before us speaks for itself and to claim that
it does so would be seriously misguided or, at the very least, naive. On the other hand,
surely, any improvement in CDAs empirical credentials ought to be welcome, tempered
though it may be by the sobering realization that completely mechanized discourse
analysis is impossible. Or, were it possible, it would cease to be critical.
The idea of comparative evidence is illustrated by the second worked example. It
relates to a column in the British popular daily the Sun (25 July 2007), published at a
time when major floods were devastating large areas in central England. The headline
reads Its time to turn off the spongers money tap (with spongers, or scroungers,
referring to people who receive money but do nothing in return). The article is built
around the contrast between the lives of hard-working people whose homes have been
destroyed by the flood in Gloucestershire, and a family with 12 children from Berkshire
(the scrounging Gillespies) who have been given a new council house paid for by the
hard-working taxpayer. The father of the family is quoted as saying, if it was
economical for me to work then I would do. A benefits system producing such attitudes,
the article demands, ought to be changed; also, foreign aid should be put on hold until
our national crisis is sorted out (original italics). There are two large pictures, one of a
flooded house in Gloucester, and a second of the Gillespies 500,000 council home,
which has a small inset showing the couple surrounded by their 12 children. Addressing
politicians from both government and opposition, the columnist pleads: The hard-
working people of Britain should come first.
Like many opinion pieces in the popular press, this column develops a stark black-and-
white dichotomy, pitting us against them. Us refers to the grafting taxpayer (with
grafting being a colloquial British expression meaning hard-working), whose hard-
earned home has been flooded, whereas they are the blatantly idle Gillespies of this
world and, less prominently, recipients of development aid in Africa. In establishing the
we group, the adjective hard-working clearly plays a central role. In an article of
around 1300 words, it occurs five times. In four of these instances, hard-working is part
of a noun group which includes at least one other ingroup marker (people of Britain,
British taxpayers, our own people, the communities). The occurrences, quoted below,
are spread fairly evenly throughout the text, contributing to its cohesion:
 1. theres only one basic rule to remember: the hard-working people of Britain
should come first.
2. Yet frankly, when the hard-working British taxpayers need them most, our
politicians look as washed up as the million-plus doormats floating around the
streets.
3. not to mention the indirect 8 billion of our taxes the PM [Prime Minister] has
recently pledged to Africa. But when our own hard-working people are in trouble,
there should be an instant amnesty on all other benevolent activity until our national
crisis is sorted out [original italics].
4. Theres more rain on the way, so I suggest Brown [= the Prime Minister], Cameron
[= the Leader of the Opposition] et al. get their galoshes on and show they care
about the blighted, hard-working communities () [original italics].
5. the skewed thinking that the hard-working taxpayer is a cash cow only to be
milked and never fed.

On this evidence, it would be fair to claim that in this text, hard-working does
ideological work, establishing a we group and attributing a positive quality to it. But
then, what is it about hard-working, exactly, that makes it such a powerfully positive
label? And is this evaluative load true of general language use, or specific to certain
discourses? In other words, is there a specifically tabloid attitude towards hard-
working people?
Let us see what the Wordbanks reference corpus has to say. First of all, the collocation
list for hard-working from the total 500+ million-word corpus, ordered by mutual
information score, reveals a long list of other positive adjectives. Picking out those that
also have double-digit joint frequencies (that is, not only exhibit a strong collocational
bond with hard-working, evident through an MI score of five or above, but also occur at
least ten times), we arrive at the following list (see Table 5.6).
TABLE 5.6 Collocates of hard-working in the total 500+ million-word Wordbanks corpus, with joint frequencies
of at least 10 and MI scores of at least 5
 Some of these are fairly closely related to the idea of working hard or in a particularly
focused manner (industrious, conscientious, dedicated, disciplined, ambitious, skilled),
but others refer to very general attributes that are quite independent of the domain of
employment: honest, loyal, sincere, decent and caring. There is a semantic preference
for character traits, and the semantic aura is unequivocally positive. To find that there is a
statistically significant collocational bond between these adjectives and hard-working
means that when someone is described as hard-working, there is a higher than random
possibility that one of these other, non-work-related qualities will appear in close
proximity. For each of these collocates, of course, we need to check what the syntactic
relationship with hard-working is, because being close could also mean close and
linked through but, in which case the other adjective would be expressing a contrast, not
a confirmation, of the virtues implied in hard work (cf. the hypothetical, not attested,
phrase hard-working but caring). However, it can be established easily by examining the
relevant concordances such as the one for decent, which is given in Table 5.7 that the
virtuous adjectives are linked to hard-working through and or a comma and do in fact
refer to the same individuals or groups of people. In addition to showing how hard-
working and decent (highlighted in bold capitals) are related syntactically and
semantically when they appear together, the following concordance also reveals a number
of other positive attributes in close proximity (highlighted in bold), some with distinctly
moralizing overtones (e.g. honest family man, genuine and Christian, self-sacrificing).
Such instances of collocation repeated, statistically significant and attested across a
multi-million-word corpus provide objective, empirical evidence for evaluative
meanings, and these meanings are not merely personal and idiosyncratic, but widely
shared in a discourse community (Stubbs, 2001: 215). Effectively, the frequent
collocates of a word become part of its meaning. Thus, by drawing on corpus-based
collocational information, a discourse analyst can replace his or her individual, intuitive
judgement on evaluative meaning with shared assumptions and judgements.
TABLE 5.7 Co-occurrence of hard-working and decent in the 500+ million-word Wordbanks Online corpus
 In further attempting to put the use of hard-working in this Sun article into perspective,
another angle worth looking at is to see how it is used in a newspaper catering for a
different readership. The relevant subcorpus that Wordbanks Online offers is the nearly
60 million words from the British daily The Times. Whereas more than 60 per cent of the
Suns readers are in the C2, D and E social grades, 89 per cent of The Times readers
belong to the A/B/C1 socioeconomic group. 10 The results (Table 5.8) show that hard-
working occurs more than twice as often in the Sun (in relative terms, that is, per million
words) than it does in The Times.
TABLE 5.8 Frequency of hard-working in the Sun and Times subcorpora of Wordbanks Online
 Furthermore, although the lists of high-frequency collocates appear to be rather similar
in the two subcorpora, containing many of the items that showed up when we examined
the whole 500-million-word corpus (see Table 5.6), two differences between the Suns
and The Times collocation lists do stand out. One is that honest and decent, though
present in both lists, are relatively more frequent in the Sun corpus.11 The other is that the
collocation list for hard-working in the Sun includes a collocate the one with the
highest MI score, in fact that is not present in The Times list at all: abiding. Switching
from the collocation list to concordance mode, we can see that all six occurrences of
abiding are due to law-abiding being one of the positive attributes closely associated
with hard-working (Table 5.9).
TABLE 5.9 Co-occurrence of hard-working and law-abiding in the Sun corpus of Wordbanks Online

The social construction involved in this collocational link, rather than any objective
semantic association, could hardly be more obvious. After all, it is perfectly possible to
be lazy and abide by the law, or work a very busy 70-hour week dedicated to breaking
it. Incidentally, taken by itself, law-abiding is also considerably more frequent in the Sun
(145 instances, or 3.2 occurrences per million words) than in The Times (111 instances,
or 1.86 per million).
Summing up, and relating the evidence back to the article that made us turn to the
reference corpus for support, we can draw the following conclusions:

Hard-working is much more than a descriptive label. Its semantic preference and
prosody, evident in the concordance lines, indicate that it is part and parcel of a
moralizing discourse, linking hard work with positive attributes such as decency,
honesty, loyalty, family values and the like.
These patterns are relatively more prominent in the popular, working-class tabloid
the Sun than they are in The Times, which caters for a predominantly middle-class
readership.
In a critical discourse analysis of a text using hard-working, a good case can
therefore be made for arguing that the contribution of hard-working to the overall
meaning of the text is based partly on the ideological baggage that the word carries,
and that this, in turn, is derived from attested patterns of usage in larger universes of
discourse.

Large-corpus evidence thus provides checks and balances by opening a window on

values and attitudes present throughout a discourse community rather than held only by
individual researchers.

Summary and critique

Corpus linguistics has a lot to offer to CDA. It helps researchers cope with large amounts
of textual data, thus bolstering CDAs empirical foundations, reducing researchers bias
and enhancing the credibility of analyses. On the other hand, critical discourse analysts
ought to be self-confident enough to assert that, conversely, corpus linguistics is enriched
by being applied to research questions inspired by social concerns, such as power,
inequality and change. Ultimately, through their theoretical and methodological cross-
pollination (Baker et al., 2008: 297), both CDA and corpus linguistics ought to benefit.
Combining the two approaches typically involves the following steps:

compiling an electronically held corpus that allows the investigation of research

questions arising from social issues
running the corpus through concordancing software that compiles frequency lists,
identifies keywords and reveals statistically significant collocations
analysing concordances qualitatively in order to establish the dominant semantic
preferences and prosodies of lexical items relevant to the social issues under
investigation
putting the results from the purpose-built corpus into perspective by comparing them
with evidence gleaned from large reference corpora.

Alternatively, a multi-million-word reference corpus may itself serve as the starting

point, allowing researchers to build collocational profiles of socially contested lexical
items across a wide range of genres, media and geographical areas.
In spite of the clear benefits involved, there are some areas of potential concern, which
I will deal with in turn under five headings (Figure 5.2).

1. The skills gap and lack of standardization

This is a practical rather than a substantive issue, and may well disappear over time. At
the time of writing, though, it still looms rather large. To anyone advocating the
integration of corpus linguistics into mainstream CDA, it is quite tempting to downplay
the effort involved and make reassuring noises along the lines of its not rocket science.
Indeed it isnt, but there is no denying the fact that becoming a confident user takes time
and effort. Not much, perhaps, for the mundane task of learning to master the tools; but
certainly a significant amount in order to develop the type of mindset that can appreciate
the potential of the method, recognize its limitations, hone your analytical skills and refine
your discovery procedures, so that ultimately you are able to fashion your research
designs accordingly.

FIGURE 5.2 Using corpus linguistics in CDA: key areas of concern

The continuing reluctance of many discourse analysts to become involved may well be
due in part to the deplorable lack of standardization within corpus linguistics. The British
National Corpus and Wordbanks Online, to use just two examples of multi-million-word
corpora, do not use the same software. The same is true of the various concordancing
packages available for analysing DIY corpora (such as Wordsmith Tools or Monoconc
Pro). Search commands differ, screens differ, analytical tools differ: not a happy state of
affairs if all you want to do is get on with the job.

2. Institutional barriers

The second point is related to the first, but located on the institutional rather than the
individual level. Critical discourse analysts and computer linguists do not necessarily
work in the same departments and, if they do, may not communicate well with each other.
They often go to different conferences and publish in different journals. As a junior
researcher, you are likely to be socialized into either the one methodology or the other,
but rarely into both. As most linguists know, but not all care to admit, it is often early
exposure to a particular methodology, rather than any inherent merits this may have, that
tends to bias ones methodological choices for a long time.
At the risk of launching into after-dinner-speech mode, this is the moment to call for
more communication between critical discourse analysts and computer linguists. This
should not, I hasten to add, stop at CDA people begging for IT support, realistic though
this image may be, but should also lead to corpus linguists picking their CDA colleagues
brains on how best to sharpen their computing tools so that they deliver the optimum
value for applications in socially relevant, applied discourse studies. Existing reference
corpora, too, could profit from some overhauling in that respect. In Wordbanks, for
example, source referencing such a key factor in determining context is notoriously
deficient.

3. Resisting temptation in data collection

Whereas the first and second issues related to potential hurdles encountered by those new
to the method, the third centres on the need to curb the enthusiasm of the newly converted.
We saw earlier that the World Wide Web and electronic processing have made for
temptingly laden data tables. And indeed, being able to assemble and analyse large
corpora is a key element in defusing the cherry-picking charges frequently levelled at
CDA. Generally speaking, corpus size undoubtedly boosts representativeness, and this, in
turn, enhances the validity of analysts claims. On the other hand, as is so often the case, a
technological advance comes with strings attached. Somewhat paradoxically, the ease
with which corpora can be assembled can prove to be at once overwhelming and
tempting for the analyst, novice and seasoned researcher alike. They may well react like a
glutton at an all-you-can-eat diner, guzzling data food indiscriminately without due
regard for the principles of discerning composition, be it of a menu or of a corpus. In our
case, these will revolve, as ever, around questions such as: what kinds of texts are most
likely to allow me to answer my research questions? Is the selection of texts which make
up my corpus reasonably representative of the universe of discourse that is out there?
None of these questions, and the principles underlying them, have ceased to be relevant. If
anything, they have become more pressing, precisely because of the embarras de
richesses surrounding the analyst engaged in corpus-building. Amid the bewildering
surplus of easily storable text, it has become easier to lose sight of the need for constant
reflexivity, even in the early stages of a project, and particularly with regard to what
should go into the corpus. This is not a plea for allowing too much biased selectivity too
soon; if it were, it would amount to reverting to the very cherry-picking procedure that a
corpus-based approach wishes to counteract (and which is why cyclical corpus-building
up to saturation was rejected in our discussion of sampling). The point is, rather, that
critical discourse analysts putting corpora together should, quite simply, not get carried
away.

4. Decontextualized data

The fourth area of concern, mentioned earlier but worth restating here, relates to the fact
that both the input to and output from concordancing software is decontextualized,
semiotically reduced language. Although programs allow instant access to wider co-texts
or even the full texts that the concordance lines come from, a considerable amount of non-
verbal information is lost when text is transferred to machine-readable form. Corpus
mark-up can help compensate up to a point but, with the current state of technological
development and commercial availability, it is impossible to run concordancing software
while preserving the full textual integrity of the original. This is an area, therefore, where
the idea of checks and balances needs to work the other way round, with the analyst
having to make sure that whatever information concordancers cannot deal with, such as
typography and pictures, remains accessible somewhere and is not entirely and
irretrievably lost to the analysis.
In this context, we ought also to remind ourselves that concordancing software is
biased towards the discrete lexical unit. Larger-scale discursive phenomena, such as
argumentative patterns, may be captured through corpus linguistic techniques, but only if
they crystallize systematically around certain words, phrases or lexico-semantic patterns.
Finally, and precisely because corpus linguistics has this fantastic potential for
focusing on linguistic detail, there is a need to guard against becoming so engrossed in
building collocational profiles of ideologically loaded individual words that the bigger
picture is lost. There is a fine line between an eye for detail and myopia. Returning
briefly to the hard-working example in the previous section: were this part of a full-
blown study, it would of course be insufficient merely to look at hard-working (central
though it is to this newspaper article). In addition, one would not only have to explore the
full range of synonyms and related expressions that the article uses, but would also have
to delve deeper into the history, politics and social psychology of wage labour and the
work ethic.

5. Language innovation

The fifth issue to be borne in mind is that large and static reference corpora, such as the
BNC and Wordbanks Online, are useless for investigating developments at the sharp end
of language. Where social change is at its fastest and arguably of keenest interest to
CDA these corpora fall silent. Youth culture, advertising and code-switching varieties
emerging among new immigrant populations would be cases in point. For such
applications, building ad-hoc DIY corpora is the only solution.
Essentially, the last four areas of concern all relate to the same issue: the need for a
realistic assessment of a methods potential. Our metaphorical tools for setting to work on
text are subject to very much the same limitations as tools in a literal sense. It makes as
much, or as little, sense to criticize corpus linguistic methods for not permitting more
contextually embedded analysis, or a static, ten-year-old corpus for being silent on the
latest neologisms, as it does to criticize a screwdriver for being no good at hammering in
nails.
What is more, it can be tempting to expect computer-based methods to work miracles
and to obscure or even compensate for flaws elsewhere in the research design. Put more
bluntly, if your sampling technique is faulty and your sample skewed, it will remain so
even when computerized, concordanced and subjected to every statistical procedure
under the sun. Similarly, if your choice of statistical techniques is such that the data are,
as Baker (2006: 179) graphically puts it, subtly massaged in order to produce the
desired results; if the analyst reports results selectively, or ignores inconvenient
concordance lines, then the fault lies not with their methodology but with their integrity.
Whatever the limitations of corpus linguistics, the complexity of discourse is such that
any change in perspective and any insight not otherwise available ought surely to be
welcome as additions to the methodological toolbox. On the other hand, we should not
forget that, in choosing methods, there is a rather thin dividing line between, on the one
hand, eclecticism that is imaginative and productive, and, on the other, aimless
patchworking, which is neither. Whether your research design ends up on the right side of
this divide depends crucially on (1) a clear statement of the aims of your project, (2) a
rigorous assessment of what each method can and cannot do, and (3) robust theoretical
foundations capturing core assumptions about language and the social. If deployed wisely,
corpus linguistics provides an enriching complement to qualitative CDA, aiding
discovery and adding analytical rigour. To return to the metaphor introduced earlier: even
an Oscar-winning supporting actor cannot rescue a bad film, but they can make a good
film great.

FURTHER READING
Baker, P. (2006) Using Corpora in Discourse Analysis. London and New York: Continuum.
This book is ideally suited for critical discourse analysts who are first-time users of corpus linguistic methods. It
combines theoretical background with hands-on advice and several worked examples.

McEnery, T., Xiao, R. and Yukio Tono, Y. (2006) Corpus-based Language Studies: An Advanced Resource
Book. London and New York: Routledge.
This book caters for both novice and more experienced researchers, proceeding from the basics to a section with
key readings from corpus-based language studies. In a third section, six extended case studies are presented,
covering areas as diverse as pedagogical lexicography, L2 acquisition, sociolinguistics, and contrastive and
translation studies.

Stubbs, M. (1996) Text and Corpus Analysis: Computer-assisted Studies of Language and Culture. Oxford
and Cambridge, MA: Blackwell.
Stubbs, M. (2001) Words and Phrases: Corpus Studies of Lexical Semantics. Oxford and Cambridge, MA:
Blackwell.
Both volumes are seminal classics. Although the author positions them, as their respective subtitles reveal, as
Computer-assisted Studies of Language and Culture and Corpus Studies of Lexical Semantics, rather than
CDA, they make essential reading for critical discourse analysts, and especially those who may be challenged in
terms of empirical rigour.
Acknowledgement
Material from the Bank of English reproduced with the kind permission of Harper
Collins Publishers Ltd.

Notes
1 This criticism, in turn, has been refuted strongly by, for example, Fairclough (1996) and Wodak (2006a: 606
609).
2 See Fairclough (1992a); Fairclough (1995a); Fairclough and Wodak (1997); Toolan (2002); van Dijk (2007a);
Wodak (2006b); Wodak and Chilton (2005).
3 The MI score relates the observed frequency of a given co-occurring item within a certain collocational span to
the left and right of the search word to the expected frequency of the co-occurring item in that span (McEnery et
al., 2006: 56). For details of the statistical computation involved, see Matsumoto (2003: 398399).
4 The issue in question is Text 22(3). See, in particular, Graham and Paulsen (2002); Muntigl (2002a); Wodak and
van Leeuwen (2002).
5 See www.collins.co.uk/books.aspx?group=154
6 The cut-off point above which measures are considered to indicate statistical significance is 2 for t-scores and 3
for MI scores (Hunston, 2002: 7172).
7 On the basis of the same number of occurrences, 567, as in the unemployed example.
8 In this particular corpus, steelmen is not as promising a collocate to follow up as it looks, because all nine
occurrences refer to the film The Full Monty.
9 See www.natcorp.ox.ac.uk/
10 According to figures from the National Readership Survey, available at www.nrs.co.uk, accessed 9 August
2007.
11 In a million words, there are 0.3 occurrences of honest in the Sun (15 instances) and 0.1 in The Times (seven
instances). The figures for decent are 0.26 per million for the Sun (12 instances) and 0.06 for the Times (four
instances).
6
Discourse as the Recontextualization of Social Practice: A Guide

Theo van Leeuwen

Introduction
Theoretical background
Discourse and social practice
Social action
Introduction
The term discourse is often used to mean an extended stretch of connected speech or
writing a text. Discourse analysis then means the analysis of a text, or type of text.
Here, I use it in a different sense, building on the work of Michel Foucault (e.g. 1977) and
defining discourses (note the plural) as socially constructed ways of knowing some
aspect of reality which can be drawn upon when that aspect of reality has to be
represented, or, to put it another way, context-specific frameworks for making sense of
things. In this chapter, I will use discourses of leadership as an example. Clearly, there
are different conceptions of what a leader is and does. The discourse of the opinion
leader, for instance, which was developed in the context of US public communication
research (Katz and Lazarsfeld, 1955) is one in which leaders are first among equals,
role models who provide examples for others to follow. Such leaders may never be
formally recognized as leaders, but they will influence what their peers think, say and do.
In other discourses, leaders are constructed as fundamentally different from ordinary
mortals, and said to rule by divine right. In his book on Nazi language, Klemperer
(2000/2006 [1957]: 111 ff.) documents how the Nazi ideology again and again
underlined [Hitlers] uniquely close relationship with the Godhead, his special status as
the chosen one, his special sonship, his religious mission (Klemperer, 2000/2006
[1957]: 111), but in recent times, too, we have seen leaders such as Bush and Blair claim
divine approval for their actions. Clearly, discourses of leadership have a long history.
The New Testament introduced the discourse of the leader as shep-herd, Platos
Republic the discourse of the philosopher king, with the leader as expert. And ever
since then, political thinkers, philosophers and others have struggled to balance peoples
need for freedom and persuasion with societys need for guidance and some form of
coercion.
In this chapter, I argue that discourses are ultimately modeled on social practices, so
that, for instance, knowledge of what leadership is, is ultimately based on what leaders
do. However, discourses will transform these doings, for instance by leaving out some of
the less palatable things leaders may do, or by representing, not just what leaders do, but
also why they do it, and, therefore, why their actions are to be seen as legitimate (cf. Van
Leeuwen, 2007). And different discourses, different ways of making sense of the same
aspect of reality, will do all this in different ways, including and excluding different
things, and doing so in the service of different interests.
As I have said, in this chapter, the term discourse will not be synonymous with the
term text. Yet, evidence for the existence of discourses will inevitably have to come
from texts, from what is said or written about leadership, for instance. More
specifically, it will have to come from the similarity between what is said and written
about a given aspect of reality in different texts that circulate in the same context. It is on
the basis of such similar statements, repeated or paraphrased in different texts, and
dispersed among these texts in different ways, that we can put the puzzle back together
and reconstruct the discourses texts draw on. This chapter introduces methods for just this
kind of reconstruction. After discussing the theoretical background of my approach to
discourse, I will show (1) how to use text analysis to piece together a discourse, and to
connect it to the practice from which it ultimately derives its meaning, and (2) how to
analyse the processes of transformation, or recontextualization (Bernstein, 1981, 1986),
that occur as practices are turned into discourses.
I will use a single example an online leadership questionnaire for assessing the
performance of managers and executives. Called Voices, it was designed by the US
company Lominger-International, and is used by companies and other organizations the
world over. The person to be assessed, referred to by Lominger as the learner,
nominates five people to assess him or her in each of four categories: direct reports (i.e.
people he or she is supervising), peers, customers and others. In addition, the
learner is assessed by his or her boss and by the learner him-or herself. The selected
raters complete the tool online, rating the learner in terms of 30 competencies,
such as hiring and staffing, interpersonal savvy, sizing up people, problem-
solving, confronting direct reports, drive for results, customer focus, integrity and
trust, etc., and also in terms of 30 career stoppers and stallers which are formulated in
terms of overuse of these same competencies. Each competency and each stopper and
staller is glossed by a paragraph of description, for example:

Conflict Management Steps up to conflicts, seeing them as opportunities; reads situations quickly; good at
focused listening; can hammer out tough agreements and settle disputes equitably; can find common ground and
get cooperation with minimum noise. This contrasts then to the following stopper and staller:
Overuse of Conflict Management May be seen as overly aggressive and assertive; may get in the middle of
everyone elses problems; may drive for a solution before others are ready; may have a chilling effect on open
debate; may spend too much time with obstinate people and unsolvable problems.

Both competencies and their overuse are then rated on a five-point scale.
Competencies are rated as (a) a towering strength, (b) talented, (c) skilled/OK, (d)
weakness or (e) serious issue. Overuse is rated as happening (a) constantly, (b)
much of the time, (c) some of the time, (d) every so often or (e) not at all.
Competencies are also rated in terms of their perceived importance for the job of the
learner: (a) mission critical, (b) very important, (c) useful/nice to have, (d) less
important and (e) not important.
As I have said, a single text does not provide enough evidence for reconstructing a
discourse, although it can of course be used for methodological demonstration, as I do in
this chapter. I would nevertheless argue for the special importance of texts such as
Voices. Voices is used very widely and is therefore not only a discourse about
leadership, but also constitutive of actual leadership practices and actual ways of talking
about these practices. When introduced in universities, as it was in the university where I
work, Voices plays a key part in the move from the old elected first among equals
style of leadership to new, corporate leadership discourses and practices (cf. Fairclough,
1993). And as learners are obliged to discuss their weaknesses with an executive
coach, they will more or less be forced to introduce the discourse of Voices into their
thinking and talking about their own role and identity as leaders. For all these reasons, it
is important to critically analyse texts such as Voices, so as to reveal how they construct
leadership.

Theoretical background
Anthropologists and sociologists have always realized that representation is ultimately
based on practice, on what people do. The primacy of practice runs like a thread through
the classics of European as well as American sociology. It is true that sociologists
sometimes derive concrete actions from abstract concepts and processes from systems
Durkheims collective consciousness, Bourdieus habitus, Talcott Parsons systems
theory (1977) and Levi-Strausss structuralist anthropology (1964) are examples. Yet the
primacy of practice also keeps asserting itself in the work of these writers, sometimes
against the grain of their methodology, at other times as a fundamental cornerstone of their
theory (e.g. Berger and Luckmann, 1966). Bourdieu elaborated the primacy of practice
and the fundamental difference between participant knowledge and outsider knowledge
in his Outline of a Theory of Practice (1977) and elsewhere. Talcott Parsons, even in his
systems theory, can still say that the subject of social interaction is in a fundamental
sense logically prior to that of social system (1977: 145), and even Levi-Strauss (1964)
at times derives the meaning of myths from social practices rather than from abstract
schemata. Durkheim leaves no doubt about it, especially in The Elementary Forms of
Religious Life (1976) and Primitive Classification (Durkheim and Mauss, 1963): myths
are modelled on rites, conceptual life on social life, representations of the world on
social organization. And Malinowski (1923, 1935) shows how representation originates
in action and in uses of language that are inextricably interwoven with action, and how
action is then twice recontextualized, first as representation, in narrative speech, and
then in the construction of new realities, in the language of ritual and magic, as
Malinowski calls it. Later, Bernsteins theory of recontextualization applied a similar
idea to educational practices, describing how knowledge is actively produced in the
upper reaches of the education system (1986: 5) and then embedded into a pedagogic
content in the lower reaches, where it is objectified and made to serve the contextually
defined purpose of a discourse of order, a form of moral education, in the
Durkheimian sense. In the approach to critical discourse analysis I present in this chapter,
I connect this idea to the term discourse, used in Foucaults sense (e.g. 1977). This
definition of discourse has also been introduced into critical discourse analysis by
Fairclough (e.g. 2000a) and the emphasis on discourse as social cognition has been
inspired by the work of Van Dijk (e.g. 1998).
 Linguists have generally differed from sociologists in deriving processes (syntagms)
from systems (grammars, paradigms), rather than processes (practices) from systems
(institutions and objectified forms of knowledge). But when linguists began to study texts,
in the 1970s, many found it hard to conceptualize the production and interpretation of texts
without recourse to experience, to world knowledge (e.g. Schank and Abelson, 1977)
or background knowledge (e.g. Brown and Yule, 1983; Levinson, 1983). Martin (1984,
1992) reintroduced the field of discourse, using lexical cohesion analysis to construct
activity sequences sequences of represented activities. Together with the work of
Gleason (1973) and Grimes (1975) who paid attention, not just to represented activities,
but also to represented roles, settings, etc., this work has had a profound influence on
the ideas I present in this chapter, and the main difference is that I have extended it
beyond procedural and narrative texts, in which there tends to be a close relation between
the represented and representing activity sequences, and applied it also to other kinds of
text, in which there is a greater difference between the structure of the text, which may be
some rhetorical, argumentative structure, and the underlying discourse, i.e. the
representation/transformation of a practice together with the purposes, legitimations and
evaluations of that practice.
Finally, the study of the way discourses transform social practices, which in this
chapter is represented especially by my theory of social action, derives to a large degree
from the work of Halliday (1978, 1985), whose theory of transitivity made it possible to
interpret differently worded representations of the same reality as different social
constructions of that reality, and from the work of Kress, Hodge, Fowler, Trew and others
(Fowler et al., 1979; Kress and Hodge, 1979) who demonstrated how Hallidays work
can be used and extended for the purpose of critical discourse analysis, or, as they said,
quoting Whorf (1956), how linguistics can become an instrument of discovery,
clarification and insight for the analysis of the social world (Kress and Hodge, 1979:
14).

Discourse and social practice

As I have said, the approach to discourse I introduce in this chapter is based on the idea
that discourses are recontextualizations of social practices. To bring this out, I start with a
simple schema of the crucial elements of social practices. Actual social practices will
always contain all these elements. Specific discourses about social practices will select
from them, transform them and add further elements.

Actions

The core of a social practice is formed by a set of actions, which may or may not have to
be performed in a specific order. The conflict management text above, for instance,
contains the following actions (I ignore for the moment that they have been transformed in
different ways, for instance by being generalized or represented in a relatively abstract
way):

stepping up to conflict
reading situations
listening
hammering out agreements
settling disputes
finding common ground
getting cooperation.

Performance modes

These actions may have to be performed in specific ways. In the conflict management
text, listening has to be focused, agreements have to be tough, and cooperation has to be
achieved with minimum noise. Clearly, it is not just important what leaders do, but also
how they do it or how they should not do it, as can be seen in the Overuse of Conflict
Management text, where overly aggressive and assertive actions are disapproved of.

Actors

Social actors participate in practices in one of a number of roles as agents (doers of

action), patients (participants to whom actions are done) or beneficiaries (participants
who benefit from an action, whether in a positive or negative sense). In the conflict
management text, the key participant is the learner whose performance is being
assessed. Although the actions in the text would in reality require further participants (the
people who are in conflict, the people who are being listened to, etc.), they have been
deleted in this particular discourse. Only the actions of the learner seem to matter as
behaviour, rather than as what they really are, interactions, actions undertaken for and
with other people. In the Overuse of Conflict Management text, on the other hand, other
people are mentioned, though for the most part vaguely (everyone else, others,
obstinate people).

Presentation styles

The way in which actors present themselves (their dress, grooming, etc.) is an important
aspect of all social practices, even if it may be taken for granted in some representations,
as is the case in Voices. In my experience, senior managers constantly evaluate each
others presentation style, but this is done informally, rather than as part of formal
performance-assessment procedures.

Times

Social practices (or parts of them) will take place at more or less specific times.
Focused listening, for instance, will happen in regular, scheduled face-to-face meetings
with direct reports. I have italicized some examples in the text below:

Timely Decision Making Makes decisions in a timely manner, sometimes with incomplete information and
under tight deadlines and pressure; able to make a quick decision.
Confronting Direct Reports Deals with problem direct reports firmly and in a timely manner; doesnt allow
problems to fester; regularly reviews performance and holds timely discussions; can make negative decisions
when all efforts fail; deals effectively with troublemakers.

Spaces

Social practices (or parts of them) also take place in specific spaces, chosen or arranged
as a suitable environment for the practice. The discourse of Voices, however, steers
away from such concrete specifics, perhaps because it is designed to apply to many
different institutional contexts.

Resources

Social practices also require specific resources, specific tools and materials. Providing
information, for instance, may require computers, an intranet and so on. But these too
have been left out of the Voices text as somehow not relevant to leadership practices.

Eligibility

Specific qualities of the concrete elements of social practices (the actors, settings and
resources) make them eligible to function as actors, settings or resources in those
practices. In fact, the whole of the Voices leadership questionnaire can be seen as a
discourse focusing specifically on eligibility, on the characteristics an actor needs to
have to be eligible to play the role of leader in an organization. I will return to this point
below.
As I have already said, discourses are transformations, or recontextualizations of
social practices. Three types of transformation are particularly important.
Deletion

Some elements of a social practice may not be represented in a particular discourse. As

we have seen, in the conflict management text, all actors other than the learner are
deleted, and so are times, spaces and resources. Such deletions happen for context-
specific reasons. First of all, Voices centres on assessing learners, focusing on their
behavioural patterns, often in abstraction from the specific situations in which these
behaviours occur. Voices also has to be applicable to many different contexts and
therefore tends towards decontextualization, towards leaving out concrete specifics.

Substitution

The key transformation is of course the transformation from an actual element of an actual
social practice into an element of discourse, and this can be done in many different ways.
Actors, for instance, can be represented as specific individuals or as types of people,
they can be referred to in abstract or specific terms, and so on. In the next section, I will
deal in detail with the transformation of social action in discourse. Van Leeuwen (2008)
provides an account of the ways in which actors, times and spaces can be transformed in
discourse.

Addition

Discourses can also add reactions and motives to the representation of social practices.
Reactions are the mental processes which, according to a given discourse, will
accompany specific actions of specific actors, for instance the way the actors feel about
specific actions, or the way they interpret specific actions. Needless to say, in different
discourses, different reactions may accompany the same actions of the same actors. In the
Voices text, as in many other discourses, the focus is on the reactions of the patients,
the people on the receiving end of the leaders actions. In the example below, for
instance, I interpret being direct as a performance mode, and being uncomfortable
and being off-guard as reactions:

Overuse of Integrity and Trust May be too direct at times, which may catch people off guard and make them
uncomfortable; may push openness and honesty to the point of being disruptive

The most important motives are purposes and legitimations. Different discourses may
ascribe different purposes to the same actions. In the following example, the purpose of
getting first-hand customer information is to improve products and services, because
the relevant competency is customer focus. If the competency had been profit focus,
the same action might be given another purpose, for instance meeting demand or
increasing sales:
 Customer Focus Is dedicated to meeting the expectations and requirements of internal and external
customers; gets first-hand customer information and uses it for improvements in products and services

Legitimations provide reasons for why practices (or parts of practices) are performed,
or for why they are performed the way they are. These reasons may be spelled out in
explicit detail or be communicated through what, elsewhere (Van Leeuwen, 2007), I have
called moral evaluation. Moral evaluation is an abstract way of referring to specific
actions which serves to highlight qualities of that action that carry positive connotations
(or, in the case of delegitimation, negative connotations). In the conflict management
text, for instance, a particular action is referred to as getting cooperation. This
formulation does not reveal much about what the leader is actually, concretely, doing
here. Persuading? Giving directives? Bribing? But it does legitimize the action, because it
suggests that it is based on voluntary cooperation, rather than on compliance with top-
down orders, and it also reveals the purpose of the action (getting cooperation).
To reconstruct a discourse (or rather, that part of it that is realized in a specific text), I
enter a texts representations of the concrete elements of the social practice (actors,
actions, times, places and so on), as well as the reactions and motives that have been
added, in the different columns of a table, using the following principles:

1. Actions which are referred to several times in different wordings are combined, but
where an alternative wording adds a purpose or legitimation, it is entered separately
in the relevant column.
2. Where possible, actions are ordered chronologically, e.g.

Where actions are simultaneous rather than sequential, I use a sign. Where there
is a choice of two (or more) possible actions, I use a flowchart notation. Where it is
not possible to decide on any chronological connection, I use a + sign, e.g.

3. Elements other than actions are horizontally aligned with the actions to which they
pertain, e.g. timings with the actions they are timings of, spaces with the actions that
take place in that space, legitimations with the actions they legitimize, and so on.

The example below analyses the conflict management text. For the sake of convenience,
I first show this text again.
 Conflict Management Steps up to conflicts, seeing them as opportunities; reads situations quickly; good at
focused listening; can hammer out tough agreements and settle disputes equitably; can find common ground and
get cooperation with minimum noise.

The text contains two overall labels for this particular episode of leadership practice,
a more neutral, technical one (conflict management) and a more connotative one
(stepping up to conflict) which adds a hint of decisiveness or boldness, hence a
performance mode that pertains to the whole of the practice. I have bolded and
italicized such overall labels in the analysis.
Even this brief example reveals some aspects of the way the Voices discourse
defines leadership. Two motives intermingle: decisiveness and toughness on the one
hand, and fairness and attentiveness on the other hand. It is also clear that the
discourse focuses entirely on the leader and the way in which s/he performs the actions
that define his/her leadership. The other participants are, so to speak, kept out of the
frame.
Three further aspects of discourse need to be mentioned at this point.

1. Some discourses provide discursive resources for other discourses

In Van Leeuwen (2008), I look at how the practice of the first day at school is
recontextualized in different discourses and note that parent-and teacher-oriented first
day discourses rely a great deal on a lay version of child psychology for legitimation.
Parents are advised to take their time because children dont like to be rushed, and to
establish the same routine going to and from school because that will make your child
feel secure. Psychologists are quoted to lend authority to such pronouncements. Clearly,
certain discourses elaborate forms of knowledge (for instance about what children are
like) or systems of moral values (e.g. religions), that can be used in a wide range of
discourses to legitimize a wide range of different practices (cf. Berger and Luckmann,
1966). In my view, Voices is such an expert discourse, supplying legitimate eligibility
criteria for leaders and leadership that can be used in a wide range of settings, as borne
out by this quote from the website of Lominger-International:
TABLE 6.1 Analysis of the conflict management text
 Lomingers Leadership Architect Competencies drive our research-based, experience-tested and integrated talent
management solutions. Flexible, useable and customizable, the suites can be deployed together as a fully integrated
system or individually to meet your immediate business needs. No matter where you start, the Leadership
Architect Competency Library allows you to maintain a common language for leadership

2. Discourses may be modalized

Voices uses a great deal of ability modality. The leader can hammer out tough
agreements, can find common ground and so on. In other words, the Voices discourse
is not so much about what leaders actually do as about what they are able to do, and this
is probably true for all discourses that construct eligibility criteria for specific roles, and
for all discursive practices that test and certify what learners can do.
Ability modality is not the only way in which discourses can be modalized. Practices
may be recontextualized as past or future practices, for instance as actual or possible
practices, or as right or wrong practices. Such modalizations should be indicated in the
analysis, for instance by using superscripts:

3. Discourses combine in specific ways

A given text may include several discourses. Nevertheless, one discourse is usually
central and other, secondary discourses relate to it in specific ways. Texts may, for
instance, contrast past and present discourses about a given practice, usually to position
the present discourse as an improvement on the past, or right and wrong discourses, as
in the case of Voices juxtaposition of competencies and overuses of those
competencies. Other discourses may play the role of preparatory discourses, dealing,
for instance, with practices or arranging spaces for the central practice, or with acquiring
the qualifications needed to participate in the central practice. Secondary discourses
may also be legitimatory. In an interview with the CEO of a large company, conducted as
part of a research project on leadership in which I am currently involved, the CEO
interrupted his description of his companys corporate responsibility practices with a
lengthy account of how he built his own house and included a rainwater tank and low
energy lightbulbs. He did this to show that he was personally engaged in sustainability
practices, and really believed in them, so positioning himself as a charismatic leader who
leads by example and inspires rather than imposes.
Needless to say, what in one context is a central discourse may in another be
secondary, and vice versa.

Social action
In this section, I deal with the ways social actions can be, and are, transformed in
discourse, again using Voices as my main example. I will begin by explaining the main
ways in which social actions can be transformed.

Actions and reactions

I have already touched on the way discourses may infuse a version of a social practice
with representations of the actors reactions to the actions that constitute the practice. As
Berger has said (Berger and Luckmann, 1966: 113), social practices involve not only a
regulatory pattern for externally visible actions, but also the emotions and attitudes that
belong to these actions, and these may be differently construed in different discourses
about these practices.
Reactions can be formulated in a number of ways. They can be unspecified (through
verbs like react, respond, etc.) or specified as cognitive (e.g. grasp), perceptive
(e.g. has a nose for) or affective (e.g. feel):

responds timely to problems with direct reports

quickly grasps the essence and underlying structure of anything
has a nose for talent
makes people feel his or her work is important.
In many discourses, different types of reactions are attributed to different social actors.
Some may be represented as engaging cognitively with social actions, for instance, with
others as reacting at a more emotional level.

Material and semiotic action

Actions can be interpreted as material, as doings (e.g. act), or as semiotic, as

meanings (e.g articulate). Most of a leaders actions are in fact speech acts, hence
semiotic actions, but they may be represented as material actions, to make them seem
more active and dynamic, as in the last two examples below:

acts with customers in mind

c a n articulate the strengths and limitations of people inside and outside the
organization
provides the information people need to know to do their job
creates mileposts and symbols to rally support behind the vision.

Material actions can be transactive, involving two participants, so that the action is
represented as actually having an effect on people or things (e.g. hire and assemble),
o r non-transactive, involving only one participant, which represents the action as
display, as behaviour that does not affect anyone or anything other than the actor him-
or herself (e.g. perform):

hires the best people available

assembles talented staff
tends to trust people to perform.

Transactive material actions are interactive if they are realized by a verb that can only
take a human object (e.g. be candid with ) and instrumental if they can (also) take a
non-human object (e.g. deal with, manage). If the latter dominates in a given
discourse, there is clearly more of a sense of the actor using people instrumentally, to
achieve goals, rather than thinking of them as people with their own goals and interests:

manages all kinds and classes of people equitably

deals effectively with trouble makers
can be candid with peers.

The same distinction applies to semiotic actions. They can be behaviouralized, in which
case the meanings conveyed by the speech acts are not included in the representation, for
example:

communicates effectively
can motivate many kinds of direct reports.

Or they can include those meanings, whether through quotation, rendition (reported
speech) or by specifying the nature of the signified (topic specification), or the signifier
(form specification) as in:

talks about possibilities

can accurately restate the opinions of others
can easily pose future scenarios.

Objectivation and descriptivization

Actions and reactions can be activated, that is represented dynamically, as actions (e.g.
analyse) or de-activated, represented in a static way, as though they are entities or
qualities rather than actions (e.g. analysis):

analyses both successes and failures for clues to improvement

is excellent at honest analysis.

Deactivated representations of social actions may be objectivated or descriptivized. In

the case of objectivation, actions or reactions are represented as though they were things,
for instance by means of nominalization or process nouns, or by various metonyms, such
as substituting the time of the action (temporalization) or the space of the action
(spatialization) for the action itself:

can diffuse even high tension situations comfortably

fosters open dialogue
may have a chilling effect on getting everyones input.

In the case of descriptivization, actions and reactions are represented as more or less
permanent qualities of social actors, that is, usually, either as epithets or as attributes:

can be candid with peers

 may be overly optimistic about how much people can grow.

De-agentialization

Actions and reactions can be agentialized, represented as brought about by human

agency, or de-agentialized, represented as brought about in ways that are impermeable to
human agency through natural forces, unconscious processes and so on.
To mention three types of de-agentialization, in the case of eventuation, the action or
reaction is represented as an event, as something that just happens, without anybody
doing it. This may be done in a number of ways, for instance through verbs of involuntary
action, through verbs like happen or occur, etc. I have not found an example of this in
the Voices text.
In the case of existentialization, an action or reaction is represented as something that
simply exists, usually through an existential process, as in high tension situations.
In the case of naturalization, an action or reaction is represented as a natural process
through abstract verbs such as vary, expand, develop, etc., which link actions and
reactions to specific interpretations of material processes to discourses of rise and fall
or ebb and flood; of birth and death; of growth and decay; of change and development and
evolution; of fusion and disintegration; of expansion and contraction and so on. This will
often have a legitimatory function. Examples in Voices include terms like growth,
development, workflow, turnaround, breakthrough and so on.

Generalization and abstraction

Actions and reactions may be generalized to different degrees. Generalizations abstract

away from the specific actions that make up a practice or some episode that forms part of
it, and label the practice or episode as a whole, as in our earlier example of conflict
management, and as also in the following examples:

distributes the workload appropriately

regularly reviews performance
provides challenging and stretching tasks and assignment.

Actions and reactions may also be represented abstractly, in which case a quality, often
apparently peripheral, is used to name the whole. This I refer to as distillation a quality
is distilled from the whole that has particular relevance in the given context, usually for
purposes of legitimation, as in our earlier example of getting cooperation, and also in:
 builds appropriate rapport
uses diplomacy and tact
eliminates roadblocks.

Overdetermination

I use the term overdetermination to refer to two specific types of representation where a
given social practice stands for more than itself. I have not found an example of this in the
Voices text.
In the case of symbolization, an often fictional social practice stands for a number
of social practices. This is the case in myths, and it is the reason why myths have such an
important function in society: Myths are a model of social action based on a mythical
interpretation of the past (Wright, 1975: 188). The killing of the dragon in the myth can
stand for passing the entrance examination, winning the election, in short for any trial in
which the hero overcomes an obstacle towards achieving his or her goal. This accounts
for its enduring quality as a model for many different practices.
In the case of inversion, one or more elements of the social practice are changed into
its opposite. In comic strips such as The Flintstones and Hagar the Horrible, the
characters, settings and objects are set in the past, but the way they interact and live are
more or less contemporary the ways of contemporary suburban families. This gives a
kind of universality to these practices that helps legitimate them as natural and
unavoidable rather than culturally and historically specific.

The system network in Table 6.2 provides an overview of the categories I have discussed
and also indicates that transformations can and cannot co-occur. Square brackets are
either-or choices and curly brackets simultaneous choices. It is therefore possible, for
instance, for an action to be semiotic (behavioural) and objectivated and de-agentialized
and generalized (but not overdetermined), e.g. workplace communication in a sentence
like good workplace communication improves productivity. But it is not possible for an
action to be both interactive and instrumental.
A text analysis using this network will allow two kinds of question to be asked:

1. What kinds of actions are attributed to what kinds of participants?

2. What kinds of actions tend to be objectivated, de-agentialized and so on?

Here are some observations resulting from my analysis of the Voices text:

Most actions and reactions in the text are attributed to the leader (199), rather than to
employees (28) or customers (4).
 About one third (9 out of 25) of the reactions attributed to employees, and all
reactions attributed to customers, are affective reactions such as like, respect, trust,
feel, etc., for instance. Only two of the reactions attributed to employees are
cognitive:
(employees) like working for (the leader)
(employees) feel their work is important
(employees) need further development
(the leader) gains (customers) trust and respect.
On the other hand, out of a total of 64 reactions of the leader, only seven are
affective and 42 cognitive. The leader is aware of, assesses, plans, judges,
projects, learns, reads the situation and so on.

TABLE 6.2 Social action network

Most of the leaders actions are material (101 out of 199) and semiotic actions (34)
tend to be behaviouralized or topic-specified, as, for example, in:
(leader) represents his interests
(leader) diffuses high tension situations.
Of the leaders 199 actions, only 32 impact on other humans (and only half of these
 are interactive). The rest are non-transactive or instrumental. As mentioned before,
the way in which most of the leaders actions are represented leaves employees and
customers out of the frame, for example:
(leader) practises listening
(leader) holds development discussions.
We might ask: who is he listening to? And: who is he holding development
discussions with?
The leaders actions are overwhelmingly generalized and often represented in terms
of distillations. The most common themes that emerge from these distillations are:
motivation and inspiration (inspire and motivate, create a climate in which people
want to do their best, etc.), future vision (look beyond today, project into the
future, etc.), relationship-building (build rapport, relate to), instrumental methods
(use diplomacy and tact, use rigorous logic, etc.), and finally honesty (present
unvarnished truths etc.).
All the leaders actions are agentialized.
Twenty-nine of the leaders actions are descriptivized (is fair to, has a nose for, is
a teamplayer, is cooperative, is good at figuring out, etc.), whereas this is the case
for only two of the employees actions. Only eight of the leaders actions are
objectivated, while the employees actions, on the other hand, are always
objectivated (career goal, trust, input, work) only their reactions are activated.

What can we conclude from this? First of all, that in this discourse, leaders are
constructed as knowledgeable doers, who act at a broad, general level and whose actions
are imbued with specific purposes and motives. They are at once tough, decisive, quick to
act and patient, understanding and fair, at once practical and down to earth and visionary.
That sounds good, but it is as if they act all this out in front of a mirror. Mirror, mirror on
the wall, who is the toughest of them all? And the employees for and with whom they do
their work are under-represented, while their work is conveyed by nouns and
nominalizations which deprive it of its dynamic and productive character. The customers,
finally, are almost entirely absent. Not to put too fine a point on it, this is not a form of
leadership which is focused on doing, on service, facilitation, enablement, but a self-
obsessed form of leadership which is focused on being a leader, and on glorifying the
characteristic attributes of leaders, rather than their deeds.

FURTHER READING
Bernstein, B. (1990) The Structuring of Pedagogic Discourse. London: Routledge.
Chatper 5 presents the theory of recontextualization which has especially inspired the earlier part of this chapter.
It is difficult to read, but rewarding.

Fairclough, N. (1995a) Critical Discourse Analysis. London: Longman.

Norman Fairclough was the first critical discourse analyst to turn his attention to corporate discourses, as
exemplified by Chapters 5 and 6 in this volume.

Malinowski, B. (1923) The problem of meaning in primitive languages, in C.K. Ogden and I.A. Richards, The
Meaning of Meaning. London: Routledge and Kegan Paul.
Written in the 1920s, this is the classic account of how practice gets to be transformed, or recontextualized, into
discourse.

Van Leeuwen, T. (2006) Critical Discourse Analysis, in K. Brown (ed.), Encyclopedia of Language and
Linguistics. 2nd edn. Vol. 3. Oxford: Elsevier, pp. 290294.
In this encyclopedia article, I have attempted a general overview of critical discourse analysis.

Van Leeuwen, T. (2008) Discourse and Practice New Tools for Critical Discourse Analysis. New York:
Oxford University Press.
Chapters 1 and 3 provide a fuller account of the theories and methods presented in this chapter.

Van Leeuwen, T. and Wodak, R. (1999) Legitimizing immigration control: A discourse-historical analysis,
Discourse Studies 1(1): 83119.
This was the first published version of the theory of discourse as recontextualized social practice.
7
A DialecticalRelational Approach to Critical Discourse Analysis in Social
Research1

Norman Fairclough
Theory and concepts
Fields of application
Methodology
An example: political discourse analysis
An illustration: analysing political texts
Discussion
Appendix 1
Appendix 2

In this Chapter, I introduce and illustrate a methodology for using a dialecticalrelational

version of CDA in transdisciplinary social research (Chouliaraki and Fairclough, 1999;
Fairclough, 2003, 2006). I begin with a theoretical section explaining the dialectical
relational approach, including my view of discourse, of critical analysis, and of
transdisciplinary research. In the second section, I briefly discuss the fields of application
of this approach, and in the third section, I explain the methodology, presenting it as a
series of stages and steps, and identify a number of core analytical categories. In the
fourth section, I present an example, showing the application of this methodology in
researching a political topic, and illustrate the approach to political analysis in the fifth
section with respect to particular texts. The sixth section summarizes what can be
achieved with this methodology and discusses possible limitations.

Theory and concepts

First, a terminological point. Discourse is commonly used in various senses including (a)
meaning-making as an element of the social process, (b) the language associated with a
particular social field or practice (e.g. political discourse), and (c) a way of construing
aspects of the world associated with a particular social perspective (e.g. a neoliberal
discourse of globalization). It is easy to confuse them, so to at least partially reduce the
scope for confusion, I prefer to use semiosis for the first, most abstract and general sense
(Fairclough et al., 2004), which has the further advantage of suggesting that discourse
analysis is concerned with various semiotic modalities of which language is only one
(others are visual images and body language).
Semiosis is viewed here as an element of the social process which is dialectically
related to others hence a dialecticalrelational approach. Relations between elements
are dialectical in the sense of being different but not discrete, i.e. not fully separate. We
might say that each internalizes the others without being reducible to them (Harvey,
1996) e.g. social relations, power, institutions, beliefs and cultural values are in part
semiotic; they internalize semiosis without being reducible to it. For example, although
we should analyse political institutions or business organizations as partly semiotic
objects, it would be a mistake to treat them as purely semiotic, because then we couldnt
ask the key question: what is the relationship between semiotic and other elements? CDA
focuses not just upon semiosis as such, but on the relations between semiotic and other
social elements. The nature of this relationship varies between institutions and
organizations, and according to time and place, and it needs to be established through
analysis.
This requires CDA to be integrated within frameworks for transdisciplinary research,
such as the framework I have used in recent publications cultural political economy
which combines elements from three disciplines: a form of economic analysis, a theory of
the state and a form of CDA (Fairclough, 2006; Jessop, 2004). Transdisciplinary research
is a particular form of interdisciplinary research (Fairclough, 2005b). What distinguishes
it is that in bringing disciplines and theories together to address research issues, it sees
dialogue between them as a source for the theoretical and methodological development
of each of them. For example, recontextualization was introduced as a concept and
category within CDA through a dialogue with Basil Bernsteins sociology of pedagogy,
where it originated (Chouliaraki and Fairclough, 1999).
In what sense is CDA critical? Critical social research aims to contribute to
addressing the social wrongs of the day (in a broad sense injustice, inequality, lack of
freedom, etc.) by analysing their sources and causes, resistance to them and possibilities
of overcoming them. We can say that it has both a negative and a positive character.
On the one hand, it analyses and seeks to explain dialectical relations between semiosis
and other social elements to clarify how semiosis figures in the establishment,
reproduction and change of unequal power relations (domination, marginalization,
exclusion of some people by others) and in ideological processes, and how in more
general terms it bears upon human well-being. These relations require analysis because
there are no societies whose logic and dynamic, including how semiosis figures within
them, are fully transparent to all: the forms in which they appear to people are often
partial and in part misleading. On the other hand, critique is oriented to analysing and
explaining, with a focus on these dialectical relations, the many ways in which the
dominant logic and dynamic are tested, challenged and disrupted by people, and to
identifying possibilities which these suggest for overcoming obstacles to addressing
wrongs and improving well-being.
The social process can be seen as the interplay between three levels of social reality:
social structures, practices and events (Chouliaraki and Fairclough, 1999). Social
practices mediate the relationship between general and abstract social structures and
particular and concrete social events; social fields, institutions and organizations are
constituted as networks of social practices (see Bourdieu on social practices and fields
Bourdieu and Wacquant, 1992). In this approach to CDA, analysis is focused on two
dialectical relations: between structure (especially social practices as an intermediate
level of structuring) and events (or structure and action, structure and strategy) and, within
each, between semiotic and other elements. There are three major ways in which
semiosis relates to other elements of social practices and of social events as a facet of
action; in the construal (representation) of aspects of the world; and in the constitution of
identities. And there are three semiotic (or discourse-analytical) categories
corresponding to these: genre, discourse and style.
Genres are semiotic ways of acting and interacting, such as news or job interviews,
reports or editorials in newspapers, or advertisements on TV or the internet. Part of doing
a job, or running a country, is interacting semiotically or communicatively in certain
ways, and such activities have distinctive sets of genres associated with them.
Discourses are semiotic ways of construing aspects of the world (physical, social or
mental) which can generally be identified with different positions or perspectives of
different groups of social actors. For instance, the lives of poor people are not only
construed through different discourses associated with different social practices (in
politics, medicine, social welfare, academic sociology) but through different discourses
in each which correspond to differences of position and perspective. I use construe in
preference to represent to emphasize an active and often difficult process of grasping
the world from a particular perspective (Fairclough, forthcoming).
Styles are identities, or ways of being, in their semiotic aspect for instance, being a
manager in the currently fashionable way in business or in universities is partly a matter
of developing the right semiotic style.
The semiotic dimension of (networks of) social practices which constitute social
fields, institutions, organizations, etc. is orders of discourse (Fairclough, 1992a); the
semiotic dimension of events is texts. Orders of discourse are particular configurations of
different genres, different discourses and different styles. An order of discourse is a
social structuring of semiotic difference, a particular social ordering of relationships
between different ways of meaning-making different genres, discourses and styles. So,
for example, the network of social practices which constitutes the field of education, or a
particular educational organization such as a university, is constituted semiotically as an
order of discourse. Texts are to be understood in an inclusive sense, not only written texts
but also conversations and interviews, as well as the multimodal texts (mixing language
and visual images) of television and the internet. Some events consist almost entirely of
texts (e.g. a lecture or an interview), while in others, texts have a relatively small part
(e.g. a game of chess).
Discourses which originate in some particular social field or institution (e.g. to
anticipate the example, neoliberal economic discourse, which originated within academic
economics and business) may be recontextualized in others (e.g. in the political field or
the wider educational field). Recontextualization has an ambivalent character
(Chouliaraki and Fairclough, 1999): it can be seen as the colonization of one field or
institution by another, but also as the appropriation of external discourses, often the
incorporation of discourses into strategies pursued by particular groups of social agents
within the recontextualizing field. For example, the transition to a market economy and
Western-style democratic government in the formerly socialist countries of Europe (e.g.
Poland, Romania) has involved a colonizing recontextualization of discourses (e.g.
discourses of privatization) which were however incorporated differently into the
strategies of new entrepreneurs, government officials, managers of state industries, etc.
(Fairclough, 2006).
Discourses may under certain conditions be operationalized or put into practice,
which is a dialectical process with three aspects: they may be enacted as new ways of
(inter)acting, they may be inculcated as new ways of being (identities), and they may be
physically materialized, e.g. as new ways of organizing space, for example in
architecture. Enactment and inculcation may themselves take semiotic forms: a new
management discourse (e.g. the discourse of marketized new public management which
has invaded public sector fields like education and health) may be enacted as management
procedures which include new genres of interaction between managers and workers, or it
may be inculcated as identities which semiotically include the styles of the new type of
managers.
CDA oscillates, as I have indicated, between a focus on structures (especially the
intermediate level of the structuring of social practices) and a focus on the strategies of
social agents, i.e. the ways in which they try to achieve outcomes or objectives within
existing structures and practices, or to change them in particular ways. This includes a
focus on shifts in the structuring of semiotic difference (i.e shifts in orders of discourse)
which constitute a part of social change, and on how social agents pursue their strategies
semiotically in texts. In both perspectives, a central concern is shifting relations between
genres, between discourses and between styles: change in the social structuring of
relations between them which achieves relative permanence and stability in orders of
discourse, and the ongoing working and re-working of relations between them which is
regarded in this approach to CDA as a normal feature of texts.
The term interdiscursivity is reserved for the latter: the interdiscursivity of a text is a
part of its intertextuality (Fairclough, 1992a), a question of which genres, discourses and
styles it draws upon, and how it works them into particular articulations. Textual analysis
also includes linguistic analysis, and analysis where appropriate of visual images and
body language, and these features of texts can be seen as realizing its interdiscursive
features.
Fields of application
The dialecticalrelational approach addresses the general question: what is the particular
significance of semiosis, and of dialectical relations between semiosis and other social
elements, in the social processes (issues, problems, changes, etc.) which are under
investigation? This question is of interest right across the social sciences and humanities,
and I would not want to foreclose the range of potentially fruitful fields of application of
this approach, nor the range of genres or texts it might be applied to. It is true that certain
types of texts would seem to pose particular problems literary texts, for example but
that is a different matter. In general, I would oppose any view of method which seeks to
neatly match methods (methodologies) to fields or text types, or cultivates the view that
researchers need to seek the right method for their data and research questions. In short,
I would not want to limit in advance the fields of application of the dialecticalrelational
approach.
The relationship between approach and applications is not a simple one. The
dialecticalrelational approach in its current form has changed through the process of
being applied in various fields. The beginnings of this approach can be seen in my work
on discourse and social change in the early 1990s (see especially Fairclough, 1992a,
1995a), which itself arose out of earlier work on relations between language, ideology
and power (Fairclough, 1989, 1991). Early applications of that version of CDA included
the marketization of higher education and the enterprise culture project launched by
the Thatcher government, as well as various aspects of political and media discourse
(Fairclough, 1995b) and critical language awareness in education (Fairclough, 1992b).
Important theoretical developments arising out of this work were the conceptualization of
orders of discourse (a concept used already in Fairclough, 1989, 1991, 1992a) as the
semiotic dimension of networks of social practices, and the development of
recontextualization as a CDA category in Chouliarakis research on classroom
discourse (Chouliaraki, 1995; Chouliaraki and Fairclough, 1999), and the foregrounding
of the dialectics of discourse. One application at this stage was to the political discourse
of New Labour (Fairclough, 2000a). Further theoretical developments arose through
exploring neglected semiotic issues in critical realism (Fairclough et al., 2004), and the
incorporation of the dialecticalrelational approach within cultural political economy
(Jessop, 2004), which I addressed specifically from a CDA perspective in my research
on globalization and transition in Central and Eastern Europe2 (Fairclough, 2006).

Methodology
I have referred to a methodology for using a dialecticalrelational version of CDA in
transdisciplinary social research rather than a method, because I also see the process as
a theoretical one in which methods are selected according to how the object of research
(Bourdieu and Wacquant, 1992) is theoretically constructed. So it is not just a matter of
applying methods in the usual sense we cannot so sharply separate theory and method.
This version of CDA is associated with a general method, which I discuss below, but the
specific methods used for a particular piece of research arise from the theoretical process
of constructing its object.
We can identify steps or stages in the methodology only on condition that these are
not interpreted in a mechanical way: these are essential parts of the methodology (a
matter of its theoretical order), and while it does make partial sense to proceed from
one to the next (a matter of the procedural order), the relationship between them in
doing research is not simply that of sequential order. For instance, the step I refer to
below of constructing the object of research does need to precede subsequent steps, but
it also makes sense to loop back to it in the light of subsequent steps, seeing the
formulation of the object of research as a preoccupation throughout. It is also helpful to
distinguish the theoretical and procedural from the presentational order one chooses
to follow in, for instance, writing a paper other generally rhetorical factors will affect
the order in which one presents ones analysis.
The methodology can be seen as a variant of Bhaskars explanatory critique
(Bhaskar, 1986; Chouliaraki and Fairclough, 1999), which can be formulated in four
stages and which can be further elaborated as steps:

Stage 1: Focus upon a social wrong, in its semiotic aspect.

Stage 2: Identify obstacles to addressing the social wrong.
Stage 3: Consider whether the social order needs the social wrong.
Stage 4: Identify possible ways past the obstacles.

Stage 1: Focus upon a social wrong, in its semiotic aspect

CDA is a form of critical social science geared to a better understanding of the nature and
sources of social wrongs, the obstacles to addressing them and possible ways of
overcoming those obstacles. Social wrongs can be understood in broad terms as aspects
of social systems, forms or orders which are detrimental to human well-being, and which
could in principle be ameliorated if not eliminated, though perhaps only through major
changes in these systems, forms or orders. Examples might be poverty, forms of
inequality, lack of freedom or racism. Of course, what constitutes a social wrong is a
controversial matter, and CDA is inevitably involved in debates and arguments about this
which go on all the time.3
We can elaborate Stage 1 in two steps:

Step 1: Select a research topic which relates to or points up a social wrong and which
can productively be approached in a transdisciplinary way with a particular focus on
dialectical relations between semiotic and other moments.
We might, for instance, conclude that such an approach is potentially productive
because there are significant semiotic features of the topic which have not been
sufficiently attended to in existing social research. A topic might attract our interest
because it has been prominent in the relevant academic literature, or is a focus of
practical attention in the domain or field at issue (in political debate or debates over
questions of management or leadership, in media commentary and so forth). Topics are
often given, and they sometimes virtually select themselves who could doubt for
instance that immigration, terrorism, globalization or security are important
contemporary topics, with significant implications for human well-being, which
researchers should attend to? Selecting such topics has the advantage of ensuring that
research is relevant to the issues, problems and wrongs of the day, but also the danger that
their very obviousness can lead us to take them too much at face value. We cannot assume
that such topics are coherent research objects; to translate topics into objects, we need
to theorize them.

Step 2: Construct objects of research for initially identified research topics by theorizing
them in a transdisciplinary way.
Anticipating the example I shall discuss below, let us assume that the selected research
topic is the relationship between national strategies and policies and the global
economy: strategies and policies which are developed for the global economy, or the
adaptation of national strategies and policies for the global economy. We might pin this
down by focusing, for instance, on strategies and policies to enhance competitiveness in
particular countries (the example I discuss relates to competitiveness policies in the UK).
As a topic for critical research, this seems plausible enough: a preoccupation of
contemporary governments is indeed adapting to the global economy, and this process
does indeed have implications for human well-being (it is widely presented as a way
towards greater prosperity and opportunity, but as entailing suffering and insecurity for
some people). One controversial formulation of the social wrong in this case might be
that the well-being (material prosperity, security, political freedom, etc.) of some people
arguably the majority is being unfairly or unjustly sacrificed for the interests of others.
I shall focus below on one particular, political aspect of the social wrong: the
suppression of political differences in favour of a national consensus on strategies and
policies.
Constructing an object of research for this topic involves drawing upon relevant bodies
of theory in various disciplines to go beyond and beneath the obviousness of the topic,
and since the focus is on a specifically semiotic point of entry into researching it, these
should include theories of semiosis and discourse. There are no right answers to the
question of which theoretical perspectives to draw upon: it is a matter of researchers
judgements about which perspectives can provide a rich theorization as a basis for
defining coherent objects for critical research which can deepen understanding of the
processes at issue, their implications for human well-being and the possibilities for
improving well-being. One must work in a transdisciplinary way, either in research teams
which bring together specialists in relevant disciplines, or by engaging with literature in
such disciplines.
What theoretical perspectives might be drawn upon in this case? These might include
(political) economic theories which theorize and analyse the global economy and may
take positions on whether and how it constitutes a realm of necessity, a fact of life; State
and political theory which probe the character and functioning of the State and of national
and international politics in the era of globalization; theories of global ethnography
which address how local groups and individuals seek to adapt to but also sometimes test
and challenge the global economy as a realm of necessity. The importance of discourse
theory is indicated by this implicit questioning of the global economy: a central issue in
both the academic literature and practical responses to the global economy in politics,
workplaces and everyday life is the relationship between reality and discourse the
reality and the discourses of the global economy and of its impact, implications and
ramifications. We can initially identify an analysis of the complex relationship between
reality and discourse as a general formulation of the object of research for a semiotic
point of entry into this topic, but I shall suggest a more specific formulation, linked to
the example I shall discuss, in the section below on political discourse analysis.

Stage 2: Identify obstacles to addressing the social wrong

Stage 2 approaches the social wrong in a rather indirect way by asking what it is about
the way in which social life is structured and organized that prevents it from being
addressed. This requires bringing in analyses of the social order, and one point of entry
into this analysis can be semiotic, which entails selecting and analysing relevant texts
and addressing the dialectical relations between semiosis and other social elements.
Steps 13 can be formulated as follows:

1. Analyse dialectical relations between semiosis and other social elements: between
orders of discourse and other elements of social practices, between texts and other
elements of events.
2. Select texts, and focuses and categories for their analysis, in the light of and
appropriate to the constitution of the object of research.
3. Carry out analyses of texts, both interdiscursive analysis, and linguistic/semiotic
analysis.

Taken together, these three steps indicate an important feature of this version of CDA:
textual analysis is only a part of semiotic analysis (discourse analysis), and the former
must be adequately framed within the latter. The aim is to develop a specifically semiotic
point of entry into objects of research which are constituted in a transdisciplinary way,
through dialogue between different theories and disciplines. The analysis of texts can
effectively contribute to this only in so far as it is located within a wider analysis of the
object of research in terms of dialectical relations between semiotic and other elements
which comprehend relations between the level of social practices and the level of events
(and between orders of discourse and texts).
I shall not elaborate much on the three steps at this stage, because I think they will be
clearer when I work through them using the example below.
There is one point about Step 3 however. I said above that although the particular
methods of textual analysis used in a specific case depend upon the object of research,
this version of CDA does have a general method of analysis. I alluded to this in the first
section: textual analysis includes both linguistic analysis (and if relevant analysis of other
semiotic forms, such as visual images) and interdiscursive analysis (analysis of which
genres, discourses and styles are drawn upon, and how they are articulated together).
Moreover, interdiscursive analysis has the crucial effect of constituting a mediating
interlevel which connects both linguistic analysis with relevant forms of social analysis,
and the analysis of the text as part of an event with the analysis of social practices in
more general terms, the analysis of event (action, strategy) with the analysis of structure.
Why so? Because interdiscursive analysis compares how genres, discourses and styles
are articulated together in a text as part of a specific event, and in more stable and
durable orders of discourses as part of networks of practices, which (qua social
practices) are objects of various forms of social analysis.

Stage 3: Consider whether the social order needs the social wrong

It is not awfully obvious what this means, and I shall try to clarify it by again anticipating
the example. I indicated above that the social wrong I shall focus on when I get to the
example is the suppression of political differences over the global economy and national
responses to it in favour of seeking to create a national consensus, which is substantively
realized in discourse. In what sense might the social order need this? Perhaps in the
sense again anticipating the discussion below that the internationally dominant
strategy for globalizing an economic order based upon neoliberal principles requires that
states be able to operate in support of this strategy without being encumbered by the old
adversarial politics. Stage 3 leads us to consider whether the social wrong in focus is
inherent to the social order, whether it can be addressed within it, or only by changing it.
It is a way of linking is to ought: if a social order can be shown to inherently give rise
to major social wrongs, then that is a reason for thinking that perhaps it should be
changed. It also connects with questions of ideology: discourse is ideological in so far as
it contributes to sustaining particular relations of power and domination.
Stage 4: Identify possible ways past the obstacles

Stage 4 moves the analysis from negative to positive critique: identifying, with a focus on
dialectical relations between semiosis and other elements, possibilities within the
existing social process for overcoming obstacles to addressing the social wrong in
question. This includes developing a semiotic point of entry into research on the ways
in which these obstacles are actually tested, challenged and resisted, be it within
organized political or social groups or movements, or more informally by people in the
course of their ordinary working, social and domestic lives. A specifically semiotic focus
would include ways in which dominant discourse is reacted to, contested, criticized and
opposed (in its argumentation, its construal of the world, its construal of social identities
and so forth).
To conclude this section, let me list the core analytical categories of this approach to
CDA which I have introduced so far:

semiosis (and other social elements)

discourse/genre/style, order of discourse (and social practices)
text (and social event)
interdiscursivity (and interdiscursive analysis)
recontextualization
operationalization (enactment, inculcation, materialization).

An example: political discourse analysis

The texts I shall discuss below are political texts: the Foreword to a government
document written by former British Prime Minister Tony Blair, and a critique of Blairs
New Labour government by two former members of the Labour Party. As I have said,
how a research topic is constituted as an object of research determines both the selection
of texts for analysis and the nature of the analysis. In this section, I shall suggest a more
specific formulation of the object of research for the research topic anticipated above
(adapting national strategy and policy for the global economy), which entails some
discussion of political theories of the contemporary political condition, and the main
issues and priorities it suggests for analysis of politics and political discourse. I shall
discuss the theoretical perspectives on the character of contemporary politics and the
State especially in advanced capitalist countries like Britain, but I should emphasize that
this discussion is necessarily partial given the spatial limitations of the Chapter. The
material in this section will also help with Step 1 of Stage 2 of the methodology when we
get to the texts analysing dialectical relations between semiosis and other elements,
especially at the level of social practices and orders of discourse.
 Let me begin with a highly condensed summary analysis of the contemporary political
condition, in the form of four major claims:

Globalization in its dominant neoliberal form has been associated with changes in
the State and national (as well as international) politics (Harvey, 2003; Pieterse,
2004).
There is a tendency of the State to become a competition state with the primary
objective of securing competitive advantage for the capital based within its borders
(Jessop, 2002).
There is an associated tendency within mainstream politics for the political division
and contestation (e.g. between political parties) characteristic of the previous period
to weaken, and for consensus to emerge on the main strategy and policy issues
(Rancire, 2006).
This tendency constitutes a fundamental political danger; not only is it a threat to
democracy, it also creates a vacuum which can be filled by nationalism and
xenophobia (Mouffe, 2005; Rancire, 1995).

The fourth point is based upon particular views of the general character of (democratic)
politics and of politics in modern democracies. I shall refer specifically to Rancires
version. He argues that democracies, both ancient and modern, are mixed forms, as
anticipated by Aristotle when he characterized a good regime as a mixture of
constitutions there should appear to be elements of both (oligarchy and democracy) yet
at the same time of neither the oligarch sees oligarchy and the democrat democracy
(see Aristotle, Politics IV: 1294b, cited in Everson [1996]). This follows from the fact
that the question of politics begins in every city with the existence of the mass of the
aporoi, those who have no means, and the small number of the euporoi, those who have
them (Rancire, 1995: 13). The task of politics is to calm and control the irreducible
conflict between rich and poor, which means curbing the excesses of democracy. What
we now call democracies are actually oligarchies in which government is exercised by
the minority over the majority. What makes them specifically democratic is that the power
of oligarchies rests upon the power of the people, most obviously because governments
are elected. In democracies, oligarchy and democracy are opposing principles in tension,
and any regime is an unstable compromise between them. The public sphere is the sphere
of encounters and conflicts between these principles: governments tend to reduce and
appropriate the public sphere, relegating non-State actors to the private sphere;
democracy is the struggle against this privatization, to enlarge the public sphere and
oppose the public/private division imposed by government.
In contemporary democracies, the conflictual equilibrium associated with popular
sovereignity is being undermined. The oligarchic system is being combined with a
consensual vision on the claim that contemporary reality, the global economy and the
prospect of endless growth which it promises, do not leave us with a choice.
Government is the business of managing the local effects of global necessity, which
requires consensus and an end to the archaic indulgence of political division.
Oligarchies are tempted by the vision of governing without the people, i.e. without the
division of the people, which means effectively without politics, rendering popular
sovereignty problematic. But the suppressed division inevitably returns, both in the form
of mobilization outside the political system (e.g. against the negative effects of neoliberal
globalization or the Iraq war) and in the dangerous form of extreme-right nationalism and
xenophobia.
A priority for political analysis is consequently contemporary processes of
depoliticization, which is by no means a new strategy (according to Rancire (1995), it
is the oldest task of politics) but is now emerging in a particularly profound and
threatening form. Depoliticization is the exclusion of issues and/or of people from
processes of political deliberation and decision placing them outside politics. But
politicization is equally a priority if we are to analyse the tension between the principles
of oligarchy and democracy, the democratic response to depoliticization, and how
responses might develop a momentum capable of contesting the push towards
depoliticization. Others have also identified depoliticization and politicization as
priorities (Hay, 2007; Muntigl, 2002b; Palonen, 1993; Sondermann, 1997), but from
different theoretical perspectives.
This prioritization provides a basis for questioning the centrality which has been
attributed to other problems and issues. Let me briefly mention two. First, the centrality
attributed to sub-politics or life politics by theorists of reflexive modernity, which is
linked to the recent prominence of identity politics. This accords with the perspective
above in giving prominence to grassroots political action, but clashes with it in
construing such politics as an alternative to adversarial politics centred around the
political system. The grassroots politics of politicization is both defined and limited by
the opposing logic of depoliticization, which means that state-and government-focused
adversarial politics is by no means outdated. Second, the centrality attributed by, for
instance, those influenced by Habermas to deliberative democracy also tends to be
associated with the assumption that adversarial politics can be superseded and to
construe political dialogue as a rational process of consensus-formation, rather than a
process which allows divisions, differences and conflicts to be contained within a shared
political community without the assumption that these are just problems waiting to be
solved. In different theoretical terms, we could say: these are contradictions, and
although they can be managed, they cannot be solved within the parameters of the existing
system (Jessop, 2002). This does not diminish or ignore cooperation in politics: conflict
in political dialogue requires cooperation (only those who are cooperating at a certain
level can stage a conflict), and adversarial politics necessarily includes cooperative
moments (e.g. the formation of alliances).
We can fruitfully develop a specifically semiotic point of entry into analysing the
processes of depoliticization and politicization. I shall illustrate this below in my
analysis of the texts. This does not exclude other issues and associated categories which
have tended to receive more attention in political discourse analysis, and indeed I shall
refer to some (legitimation, manipulation, ideology, cooperation and identity). But it does
imply a different mapping of the relations between categories which may lead to
reconceptualizing or changing some of them.
Politicization and depoliticization are high-level strategies or macro-strategies; so
are legitimation and delegitimation. Strategies combine goals and means, and these
macro-strategies are both means for achieving oligarchic or democratic goals (e.g.
governing with minimal interference from political divisions, or pushing political
differences into the public sphere), and goals in their own right associated with further
strategies as means. We can identify strategies for (de)politicization and (de)legitimation
for instance, authorization and rationalization have been suggested as legitimation
strategies (Van Leeuwen, 2007; Van Leeuwen and Wodak, 1999). All of these are
political strategies, not semiotic (or discourse) strategies, though they are generally
realized semiotically.
I suggested above that the object of research could be broadly formulated as the
complex relationship between discourse and reality in adapting national strategy and
policy for the global economy. We can now reformulate it more precisely: semiotic
realizations of strategies of depoliticization and politicization in national responses to the
global economy, focusing on the competitiveness policy in the UK.

An illustration: analysing political texts

I come now to the analysis of two sample texts. The one I shall begin with is the
Foreword written by the former British Prime Minister Tony Blair to the Department of
Trade and Industrys White Paper on Competitiveness (DTI, 1998 see Appendix 1). I
shall organize my comments according to the stages and steps listed in the Methodology
section, but I have just been effectively discussing aspects of Stage 1 so I shall keep my
comments on it brief.

Stage 1: Focus upon a social wrong, in its semiotic aspect

The social wrong I shall focus upon is the suppression or marginalization of political
differences over important issues of strategy and policy how to respond nationally to
radical international economic changes (and the prior question of what the changes
actually are) in favour of creating a consensus, which is, as I indicated above, a social
wrong in that it undermines democracy but also poses the danger that dissent, which
cannot be politically articulated, may emerge in nationalist or xenophobic forms. A
semiotic point of entry is possible and fruitful, focusing upon semiotic realizations of the
macro-strategy of depoliticization, in accordance with the construction of the object of
research which I have discussed above. The second text, an extract from a book (Brown
and Coates, 1996) written by former members of the Labour Party criticizing Blairs
New Labour government, exemplifies semiotic realizations of the macro-strategy of
politicization. (Note that both macro-strategies may however be at work in the same text.)
Blairs text is representative of the dominant tendency of the times towards
depoliticization; but this tendency coexists with politicizing responses such as that of the
second text, even if the latter often have a relatively marginal effect on government
strategy and policy. I have already discussed steps 1 and 2 above, on the construction of
an object of research for the research topic, in anticipation of the illustration, so we can
move on to Stage 2.

Stage 2: Identify obstacles to addressing the social wrong

I shall discuss Stage 2 by taking each of the three steps it includes in turn.
Step 1: Analyse dialectical relations between semiosis and other social elements (orders of discourse and elements of
social practices, texts and elements of events).
Step 1 also implicitly includes the dialectic between structures (at the intermediate level
of social practices) and events (and strategies). I have already (in the previous section)
given an indication of the social practices and orders of discourse at issue here, but let
me fill this out a little with respect to the re-structuring and re-scaling (Jessop, 2002)
tendencies associated with contemporary capitalism, and a brief note on New Labour in
Britain.
Re-structuring is changes in structural relations, notably between economic and non-
economic fields, which include extensive colonization of the latter (including politics
and the State) by the former; re-scaling is changing relations between global, regional,
national and local scales of social life, including changes in government and governance.
Analysing these tendencies would help contextualize the UK strategies and policies which
are in focus i.e. help determine what they are a part of. National governments are
increasingly incorporated within larger networks which include not only other
governments but also international agencies (e.g. the European Union, the World Bank,
the IMF), business networks and so forth. Governments, according to Castells (1996), are
increasingly coming to function as nodes within a transnational network based upon a
businessgovernment complex, whose central functions are focused upon creating the
conditions (financial, fiscal, legal, human capital, etc.) for successful competition in the
global economy. If the government strategies and policies in focus here are locked into
this powerful network, this in itself constitutes a substantial obstacle to addressing the
social wrong.
But these processes of re-structuring and re-scaling have an important semiotic
dimension: the networks of social practices which they entail are also orders of discourse
which themselves cut across structural and scalar boundaries. For example, the dominant
neoliberal discourse of globalization illustrated in the first text is dominant in education
as well as politics, and in the European Union, the World Bank and many other countries
apart from Britain. There are also genres and styles which are disseminated structurally
and in scale in a similar way (Fairclough, 2006). Moreover, the semiotic dimension is
fundamental to re-structuring and re-scaling, in the sense that these processes are
semiotically driven. They begin as discourses which constitute imaginaries (Jessop,
2004, 2008) imaginary projections for new relations of structure and scale in
economies, government, education and so forth; these may become hegemonic, or
dominant, and may be widely recontextualized; in so far as they do become hegemonic,
they are operationalized in new structures, practices, relations and institutions; and the
operationalization itself has a partly semiotic aspect in the emergence and dissemination
of genres and genre networks (see below), which enable the governance of these
complex new networks, as well as styles. The semiotic dimension, deeply embedded
within and constitutive of the new structural and scalar relations, is itself a part of the
obstacles to addressing the social wrong.
With respect to the dialectic between texts and other elements of social events, the
general point is that political texts are not some superficial embroidery upon political
events but a fundamental, constitutive part of them. In this case, for example, the strategies
and policies of the Blair government for building British competitiveness in adapting to
the global economy have a clearly textual character. They are formed, disseminated and
legitimized within complex chains and networks of events (committee meetings, reports,
parliamentary debates, press statements and press conferences, etc.) which are largely
chains and networks of texts i.e. different types of texts which are regularly and
systematically linked together. They are linked for instance in accordance with the genre
networks I referred to above systematically linked genres (e.g. discussion, report,
debate) which semiotically constitute procedures in this case, procedures of governance
(on chains of events, texts and genres, see Fairclough, 2003). These strategy and policy
processes thus have a largely textual character, and require textual analysis. The
illustrative examples are just two small samples from the complex networks of texts
involved.
The analysis would need to go into some detail about politics and social change in
Britain. I have no space for such detail here, but let me make a couple of points (see
further Fairclough, 2000a). First, New Labour abandoned the traditional social
democracy of the British Labour Party to embrace the neoliberalism of preceding
Conservative governments (those of Margaret Thatcher and John Major). The effect was
to produce a neoliberal consensus on major policy issues within mainstream politics and
a common political discourse the associated tendency to exclude opposition is
precisely the social wrong I am addressing. Second, the infamous preoccupation of
New Labour with media spin (close management and manipulation of the presentation
of policies and events in the media) indicates the growing importance of semiotic
processes (political communication) in government. Thus, the form of politics which
has developed with New Labour poses specifically semiotic obstacles to addressing the
social wrong at issue.
Step 2: Select texts and categories for analysis
With respect to Step 2, the constitution of the object of research indicates the selection of
texts in which the macro-strategies of depoliticization and politicization are semiotically
realized. My examples here are both written texts, but one would also want to include, for
instance, not only discussions, debates and interviews on TV and radio, and websites, but
also material from campaigns, protests and demonstrations centred upon the global
economy and government strategy and policy oriented towards it, and material
representing how people experience and react to the drive for competitiveness in a
variety of situated contexts (e.g. conversations and discussions within workplaces).
Appropriate focuses and categories for the analysis include semiotic strategies which
realize de/politicization, including argumentation and rhetorical strategies, as well as
semiotic aspects and realizations of legitimation, manipulation, ideology, cooperation and
identity. I shall be more specific about some of these in discussing the texts.
Step 3: Carry out analyses of texts
The first text is structured as an argument whose structure we can schematically
reconstruct as follows:

The modern world is changing.

Premises: There are opportunities to succeed and prosper in the modern world.
If we want to succeed and prosper, we must compete effectively.
Implicit premise: (We do want to succeed and prosper.)
Conclusion: Therefore, we must compete (more) effectively.

The argumentation realizes semiotically the macro-strategy of legitimation, and

specifically the strategy of rationalization: it is an example of the governments attempt to
legitimize its political strategy and the policies associated with it as necessary responses
to the situation.
The argument is formally valid, but whether it is sound or not (i.e. whether it is a
reasonable argument) depends upon the truth of its premises. We can challenge the
argument, argue that it is fallacious, by challenging the truth of its premises (Iet cu, 2006).
I want to specifically question the premises on the grounds that they (a) predicate the
possible success of a problematic identity category as subject (we), and (b) falsely
claim that the change attributed to the modern world is simply an inevitable fact of life
which we must accept. Both of these flaws in the premises can be associated with the
macro-strategy of depoliticization.
 With respect to the first flaw, the identity category we is problematic in that it is
based upon a false equation between we = Britain and we = all the citizens of
Britain: if Britain achieves success or prosperity, it does not follow that all of its
citizens do. This is the fallacy of division, when a general category has properties
which are mistakenly attributed to each of its parts. One sentence clearly implies that this
does follow: That is the route to commercial success and prosperity for all. This fallacy
is a banal feature of governmental discourse, but it is fundamental to the macro-strategy of
depoliticization, whose basic strategic goal is to dedifferentiate potentially antagonistic
identities the internal division of the political community into us and them. In this
sense, identity and the semiotic construal of identities are a major focus in analysis which
prioritizes depoliticization.
The issue in semiotic terms is personal deixis. There are two personal deictic
centres, or positionings of the author (Blair) with respect to identity: he positions himself
within two group identities we = the government, and we = the country. It is
commonplace in the literature on identity that identity entails difference we entails
they (Connolly, 1991). We might say that we = the government is implicitly construed
in opposition to they = previous governments which pursued strategies which are
rejected because they did not and cannot work: old-fashioned state intervention and
naive reliance on markets; whereas we = the country is construed in opposition to
competitors. But notice that the construal of personal deixis excludes a we/they
division both within the political community (Britain) and within the contemporary
political field (political system), where no contemporaneous political opposition is
construed. The implication is that there is consensus within both the political community
and the political field. This is depoliticization.
Texts semiotically construe identities and simultaneously seek to make these construals
persuasive. The fact that we can show fallacies in Blairs argument does not mean that it
will be widely perceived as fallacious, and we must consider what might make the
argument and construal of identities persuasive. This brings us to the second flaw, in the
construal of world change.
Dominant construals of the new global order have certain predictable linguistic
characteristics (on the linguistic categories I mention below, see Fairclough, 2003):
processes of change are construed without responsible social agents; they are construed
in a timeless, ahistorical present; statements about the new economy (which are often very
familiar truisms) are construed categorically and authoritatively as unmodalized truths,
and there is a movement from the is of the economic to the ought of the political
from what is categorically the case to what we ought to do in response; the new
economic reality is construed as indifferent to place; and series of evidences or
appearances in the new economy are construed paratactically as lists. I have shown
elsewhere (Fairclough, 2000b) that these features are sustained through
recontextualization, appearing in economic texts (e.g. texts of the World Bank), political
texts, educational texts and so forth, as well as on different scales.
 They are also evident in Blairs text, and they can be seen as aspects of the semiotic
realization of depoliticization. In the construal of economic change in the modern world,
there is an absence of responsible social agents. Agents of material processes are abstract
or inanimate. In the first paragraph, change is the agent in the first (passive) sentence,
and new technologies and new markets are agents in the second agents, notice, of
intransitive verbs (emerge, open up) which construe change as happenings or
processes without agents. The third sentence is existential new competitors and new
opportunities are merely claimed to exist, not located within processes of change. Notice
also that in the third paragraph, the inanimate this new world is the agent of
challenges, construing change itself as articulating what responses to it are necessary.
By contrast, when it comes to national responses to these implacable and impersonal
processes of world change, social agents are fully present business, the government, the
DTI and especially we.
Turning to time, tense and modality, world change is construed in the ahistorical
timeless present tense, as indeed are national responses, and, in terms of modality,
through authoritative categorical assertions of truisms (e.g. The modern world is swept
by change, and indeed all five statements in the first paragraph). The only historical
reference is to the old-fashioned strategies in paragraph 4. There is a movement from
is to ought. Ought is implicit in paragraphs 2 and 3: our success depends on how
well we exploit our most valuable assets implies that we should exploit them; this new
world challenges business to be innovative and government to create imply that
business and government should do these things. From paragraph 5 onwards, ought is
explicit and recurrent the modal verb must occurs six times. The domain of is is
world change; the domain of ought is national responses: a divide is textually
constructed between economics and politics (there is an industrial policy, but focused
on enabling the economic process rather than radically shaping it), fact and value, which
excludes the former from the latter. This differs from the social democratic tradition from
which New Labour has come; earlier Labour governments used political power to change
the economy, for example by nationalizing private industries, taking them into state
control. In contrast with economic processes, political processes do have responsible
social agents: the agent in processes modalized with must is in five cases we and in
one case the government. Summing up, world change is a process without a history
which we must respond to. Moreover, world change is implicitly construed as
indifferent to place there are no place expressions in the first or third paragraphs.
The syntax is paratactic,4 in relations between both sentences and phrases within
sentences. The first paragraph, for instance, consists of three paratactically related
sentences (the second and third contain paratactically related clauses), listing evidences
of world change. The same is true of the second paragraph. Notice that the sequencing of
these sentences is not significant and is changeable (with minor rewording) without any
substantive meaning change. Indeed, what is included in this list of evidences is
somewhat arbitrary; for instance, the second sentence of the first paragraph might have
been Huge amounts of money move across the globe in a fraction of a second, and even
our family cat, Socks, has his own homepage on the World Wide Web. The second
clause is fanciful only in that Blair does not have a cat called Socks. It was actually
included in a very similar list in a book by Bill Clinton. What is significant, rhetorically,
is the relentless accumulation of evidences of change what Clarke and Newman (1998)
call the cascade of change which persuasively (and manipulatively) establishes the
new economy as simple fact, what we must live with and respond to.
Summing up, change is authoritatively construed as lists of known appearances (and
truisms) in the present which are indifferent to place and whose social agency is effaced,
and which must be responded to in certain ways. These features together construe the new
economy as a simple fact to which there is no alternative. They locate the global
economy within the realm of necessity, and therefore outside the realm of contingency
and deliberation, i.e. outside the realm of politics, semiotically realizing the macro-
strategy of depoliticization (Hay, 2007). We can say that in so far as this sort of discourse
achieves significant public acceptance, which it has, it is part of the obstacles to
addressing the social wrong.
Let me briefly comment on interdiscursive analysis. One can see Blairs text as
recontextualizing analyses of the global economy more fully elaborated in texts
produced such as within the World Bank, and their particular discourse (construals of,
narratives of and arguments about the global economy). Blairs text is not primarily an
analytical text but an advocative text, arguing for necessary policies. But it is
interdiscursively complex in grounding this advocative argument in the recontextualized
analysis, combining analytical and advocative genres (as well as economic and political
discourses). This type of recontextualization and interdiscursive hybridity is common as a
semiotic realization of a favoured legitimation strategy: legitimizing by appeal to expert
knowledge. Notice that the expert discourse is not the same here as it might be in
specialist economic texts. For instance, in the first paragraph, the construal of change in
the global economy is stripped down to three short sentences which furthermore
incorporate characteristic features of political rhetoric (the dramatic metaphor swept by
change, the antithesis of new competitors but also great new opportunities), and which
constitute dramatic and potentially persuasive formulations of premises in the argument.
Recontextualization involves transformation to suit the new context, which affects forms
of interdiscursive hybridity.
In discussing Stage 2, I have identified a number of obstacles to addressing the social
wrong at issue, and shown that they are partly semiotic in nature. Let me summarize them:
the national and international networks that government strategies and policies are
embedded within; the consensual character of mainstream politics in Britain; and an
influential political discourse, exemplified in the Blair text, which in various ways
contributes to depoliticizing the global economy and national responses to it.
Stage 3: Consider whether the social order needs the social wrong

I anticipated this example in discussing Stage 3 in the Methodology section, where I

suggested how the suppression of political differences in favour of consensus might be
interpreted as necessary for states to operate effectively within the hegemonic, neoliberal
strategy. We might add that achieving a broad consensus within the political system
depends upon semiotic conditions achieving semiotic hegemony, or broad acceptance of
the sort of discourse we have here. And as I noted above, this can be interpreted in terms
of ideology as the naturalization of meanings which sustain relations of power and
domination. So it seems plausible that the social order does need the social wrong in
this case addressing it might require wider changes in the social order and that, since
the wrong has a partly semiotic character, it also needs certain characteristics of
contemporary political discourse.

Stage 4: Identify possible ways past the obstacles

At this point, I shall introduce the second text (see Appendix 2), an extract from a book
(Brown and Coates, 1996) written by two longstanding members of the Labour Party
about New Labours view of what they call capitalist globalization. This will allow
some necessarily brief, partial and sketchy comments on the other main macro-strategy
politicization.
I mentioned one adversarial feature in the first text: a rejection of the old-fashioned
state intervention and the naive reliance on markets of previous governments, while
implying that there were no contemporaneous divisions on the nature of world change or
the national strategies needed to adjust to it. The second text, by contrast, enters into
adversarial dialogue with contemporaries, specifically Blairites. The macro-strategy of
politicization is semiotically realized in the texts dialogicality. Specifically, there are
claims which are denials of claims made elsewhere, by New Labour politicians among
others: What has changed is not that capital is more mobile and it is not true that
national governments and by extension the European Union are totally lacking in
powers to employ against the arbitrary actions of transnational capital. In this respect,
the strategy is to politicize by construing the nature of world change and government
responses as controversial matters, subject to political difference and division.
Text 2 also politicizes by counterposing to the New Labour narrative of collaboration
between government and business a narrative of conflict between government and
business, capital and labour. Notice that both texts construe the global(ized) economy as a
reality which countries need to adjust to, but in radically different ways. In the second but
not the first, the construal of the global(ized) economy does include responsible social
agents: the companies, whose actions are construed in general and negative terms
(moving internationally from bases , the arbitrary actions of transnational capital,
divide and conquer). The text also construes relations between the companies and
national governments, contrasting the clientelist relations which tend to exist and which
New Labour advocates (nation-states clients of transnational companies) with
adversarial relations which could and by implication should exist (employing their
powers against the arbitrary actions of transnational capital, making or withholding
tax concessions, bargaining). The same contrast between what is and what
could/should be is construed in relations between the EU and national governments
(reinforcing the status of nation states as clients of the companies, versus offering a
lead and challenge to the nation states).
In sum, whereas text 1 depoliticizes by construing a consensus on the global economy
as an inevitable fact of life and building national competitiveness as a necessary
response, text 2 politicizes by construing the globalized economy as a stake in struggles
between governments and transnationals, and capital and labour, and by opposing that
construal to the governments consensualist construal. But the mere existence of texts
which politicize in this way does not amount to ways past the obstacles. This text offers
an imaginary for a different, politicizing strategy in response to a differently conceived
global(ized) economy; it shows that different imaginaries are possible and indeed exist,
but we would also need to consider how feasible it would be to operationalize this or
some other imaginary in a strategy which could actually succeed and be implemented in
the face of the sort of obstacles I have begun to indicate. Its not impossible, but its
difficult to see how at present: there are abundant alternative imaginaries, but there is
currently no clear counter-hegemonic strategy. A fuller treatment than I have space for
would include an analysis of attempts to develop oppositional strategies and their
semiotic dimensions.

Discussion
The theoretical claim that relations between semiosis and other social elements are
dialectical in character, and the methodological focus on these relations rather than on
semiosis as such, mean that this approach to CDA is particularly attuned to
transdisciplinary research, to working with the grain of various bodies of social theory
and research, but at the same time bringing to them an enhancement of their capacity to
address often neglected semiotic dimensions of their research objects, as well as taking
from them perspectives and research logics which can contribute to the further
development of the dialecticalrelational approach itself.
As with any approach, there are things about which the dialecticalrelational approach
has little to say. We should distinguish however between issues and problems it has not
got around to because others seemed more pressing or more interesting or simply because
life is short, and issues and problems which fall outside its remit and are thus not issues
and problems for it (though they may be for other approaches). An example of the former
is a relative emphasis on the workings of power rather than the workings of reception,
reaction and resistance to power I stress relative because the latter have not been
entirely neglected (see, for instance, Fairclough, 2006). Critics might reasonably say that
I have done it again in this chapter, spending more time on depoliticization than
politicization. This has been a bias in my work, perhaps partly because of the sort of left-
wing politics I was involved with in the 1970s, but it is not in my opinion a limitation of
the approach as such. An example of the latter is a lack of attention to psychological and
cognitive matters. I would agree that cognitively oriented research on discourse can
complement the dialecticalrelational approach, but I would not accept that an absence of
attention to cognitive issues is a blindspot in the approach, still less that it in some
sense invalidates the approach.
Chilton, for example, has suggested that a proper understanding of the cognitive
capacities of humans may lead to the conclusion that CDA is trying to teach people what
they already know. Put bluntly, if people have a natural ability to treat verbal input
critically, in what sense can CDA either reveal in discourse what people can already
detect for themselves or educate them to detect it for themselves? (Chilton, 2005a). Yet
the closing sentences of Chilton (2004) note that if people are indeed political animals
then they are also in principle capable of doing their own political critique. The
important question is whether they are free to do so. I agree. Chilton (2005a) argues that
although there are various conditions under which people are not free, it is doubtful that
any of them can be elucidated by purely linguistic or discourse-analytical means. For they
would seem to have to do with economic forces or sociopolitical institutions. The main
problem with this argument is indicated by the contrast between purely linguistic or
discourse-analytical factors and economic forces or sociopolitical institutions. From a
dialecticalrelational perspective, economic forces and sociopolitical institutions are in
part semiotic, and analysis has to be in part semiotic analysis. The fact that people have
cognitive capacities which make them in principle capable of seeing through manipulative
intentions and even doing their own political critique (which CDA, far from discounting,
presupposes) does not mean that they are generally capable in practice of seeing through
the complex dialectical relations between semiotic and non-semiotic elements which
constitute the social, political and economic conditions of their lives.

FURTHER READING
Chouliaraki, L. and Fairclough, N. (1999) Discourse in Late Modernity. Edinburgh: Edinburgh University Press.
This book shows the relationships of an earlier version of this approach to various sources and influences in social
theory and research.

Fairclough, N. (2000a) New Labour, New Language? London: Routledge.

A popular introduction to analysis of political discourse, based upon a simplified version of this approach to CDA.
 Fairclough, N. (2003) Analysing Discourse: Textual Analysis for Social Research. London: Routledge.
This book focuses on using textual analysis in social research within the dialecticalrelational approach, with many
examples of possible applications.

Fairclough, N. (2005a) Critical discourse analysis, Marges Linguistiques 9: 7694.

A recent overview of the dialecticalrelational approach in an electronic journal available free of charge on the
internet.

Fairclough, N. (2006) Language and Globalization. London: Routledge.

This text exemplifies the application of the dialecticalrelational approach in transdisciplinary research on
globalization.

Appendix 1 Building the knowledge-driven economy

Foreword by the Prime Minister

The modern world is swept by change. New technologies emerge constantly; new markets
are opening up. There are new competitors but also great new opportunities.
Our success depends on how well we exploit our most valuable assets: our knowledge,
skills and creativity. These are the key to designing high-value goods and services and
advanced business practices. They are at the heart of a modern, knowledge-driven
economy.
This new world challenges business to be innovative and creative, to improve
performance continuously, to build new alliances and ventures. But it also challenges
government: to create and execute a new approach to industrial policy.
This is the purpose of this White Paper. Old-fashioned state intervention did not and
cannot work. But neither does naive reliance on markets.
The government must promote competition, stimulating enterprise, flexibility and
innovation by opening markets. But we must also invest in British capabilities when
companies alone cannot: in education, in science and in the creation of a culture of
enterprise. And we must promote creative partnerships which help companies: to
collaborate for competitive advantage; to promote a long-term vision in a world of short-
term pressures; to benchmark their performance against the best in the world; and to forge
alliances with other businesses and employees. All this is the DTIs role.
We will not meet our objectives overnight. The White Paper creates a policy
framework for the next ten years. We must compete effectively in todays tough markets if
we are to prosper in the markets of tomorrow.
In government, in business, in our universities and throughout society, we must do much
more to foster an entrepreneurial spirit: equipping ourselves for the long term, prepared
to seize opportunities, committed to constant innovation and enhanced performance. That
is the route to commercial success and prosperity for all. We must put the future on
Britains side.
The Rt Hon. Tony Blair MP, Prime Minister

Appendix 2
Capital has always been global, moving internationally from bases in the industrialized
countries. What has changed is not that capital is more mobile but that the national
bases are less important as markets and production centres. In other words, the big
transnational companies are not only bigger but more free-standing The European
Union, far from offering a lead and a challenge to the nation-states of Europe, reinforces
their status as clients of the transnational companies. Indeed, this clientism applies not
only to companies based in Europe While it is true that a national capitalism is no
longer possible in a globalized economy, it is not true that national governments and by
extension the European Union are totally lacking in powers to employ against the
arbitrary actions of transnational capital. There is much that governments can do in
bargaining in making or withholding tax concessions, for example But such
bargaining has to have an international dimension or the transnational companies can
simply continue to divide and conquer New Labour appears to have abandoned what
remained of Labours internationalist traditions Yet the ICTFU, the European TUC and
the Geneva trade groups all offer potential allies for strengthening the response of British
labour to international capital (Brown and Coates, 1996: 1724).

Notes
1 I am grateful to Isabela Iet cu, Michael Meyer and Ruth Wodak for commenting on a draft version of the
chapter.
2 Critical realism is a realist philosophy of science and social science which has been developed especially in the
work of Roy Bhaskar (Bhaskar, 1986). Cultural political economy is a version of political economy which
claims that economic processes and systems are culturally and semiotically conditioned and embedded, as well as
politically.
3 In the first edition of this book and in other publications, I referred to social problems rather than wrongs. I
have changed this because I think that construing all wrongs as problems which need solutions which can in
principle be provided even if they have not been so far in practice is part of the self-justifying (and one might
say ideological) discourse of contemporary social systems in countries like Britain. The objection to it is that some
wrongs are produced by systems and are not resolvable within them.
4 Paratacticsyntactic relations are relations between sentences, clauses or phrases which are grammatically
equal, and are coordinated; they contrast with hypotactic relations, where there is one main sentence, clause or
phrase, and others are subordinated.
References

Agar, M. (2002) The Professional Stranger. 3rd unrevised edn. San Diego, CA: Academic Press.
Alexander, J. C., Marx, G. T. and Williams, C. L. (eds) (2004) Self, Social Structure, and Beliefs:
Explorations in sociology. Berkeley: University of California Press.
Altheide, D. L. and Johnson, J. M. (1994) Criteria for Assessing Interpretive Validity in Qualitative Research, in
N. K. Denzin (ed.), Handbook of Qualitative Research. Thousand Oaks: Sage, pp. 485499.
Anthonissen, C. (2001) On the Effectivity of Media Censorship: An analysis of linguistic, paralinguistic and other
communicative devices used to defy media restrictions. Unpublished PhD thesis, University of Vienna.
Anthonissen, C. and Blommaert, J. (eds) (2007) Discourse and Human Rights Violations. Amsterdam:
Benjamins.
Argyle, M., Furnham, A. and Graham, J. A. (1981) Social Situations. Cambridge: Cambridge University Press.
Auer, P. (ed.) (1992) The Contextualization of Language. Amsterdam: Benjamins.
Augoustinos, M. and Walker, I. (1995) Social Cognition: An integrated introduction. London: Sage.
Baker, P. (2006) Using Corpora in Discourse Analysis. London: Continuum.
Baker, P. and McEnery, T. (2005) A corpus-based approach to discourses of refugees and asylum seekers in
UN and newspaper texts, Journal of Language and Politics 4(2): 197226.
Baker, P., McEnery, T. and Gabrielatos, C. (2007) Using collocation analysis to reveal the construction of
minority groups: the case of refugees, asylum seekers and immigrants in the UK press, paper given at Corpus
Linguistics 2007, University of Birmingham, 2830 July. Available at: eprints.lancs.ac.uk/602/.
Baker, P., Gabrielatos, C., Khozravinik, M., Krzy anowski, M., McEnery, T., and Wodak, R. (2008) A useful
methodological synergy? Combining critical discourse analysis and corpus linguistics to examine discourses of
refugees and asylum seekers in the UK press, Discourse and Society 19(3): 273305.
Balke, F. (1998) Was zu denken zwingt. Gilles Deleuze, Felxi Guattari und das Auen der Philosophie, in J. Jurt
(ed.), Zeitgenssische Franzsische Denker: Eine Bilanz. Freiburg im Breisgau: Rombach Litterae, pp. 187
210.
Bauer, M. W. and Aarts, B. (2000) Corpus Construction: A Principle for Qualitative Data Collection, in M. W.
Bauer and G. Gaskell (eds), Qualitative Researching with Text, Image and Sound. London: Sage, pp. 1937.
Bell, A. (1994) Climate of opinion: public and media discourse on the global environment, Discourse and
Society 5(1): 3364.
Bellah, R. N. (1973) Emile Durkheim: On morality and society selected writings. Chicago: University of
Chicago Press.
Berger, P. and Luckmann, T. (1966) The Social Construction of Reality. Harmondsworth: Penguin.
Bernstein, B. (1981) Codes, modalities and the process of cultural reproduction: a model, Language and
Society 19: 327363.
Bernstein, B. (1986) On Pedagogic Discourse, in J. Richardson (ed.), Handbook for Theory and Research in
the Sociology of Education. Connecticut: Greenwood Press.
Bernstein, B. (1990) The Structuring of Pedagogic Discourse. London: Routledge.
Bhaskar, R. (1986) Scientific Realism and Human Emancipation. London: Verso.
Billig, M. (2003) Critical Discourse Analysis and the Development of New Science, in G. Weiss and R. Wodak
(eds), Critical Discourse Analysis: Theory and Interdisciplinarity. London: MacMillan, pp. 3546.
Billig, M. (2008) Nominalizing and de-nominalizing: a reply, Discourse and Society, in press.
Blau, P. M. (1964) Exchange and Power in Social Life. New York: Wiley.
Blommaert, J. (2005) Discourse. Cambridge: Cambridge University Press.
Blommaert, J. and Bulcaen, C. (2000) Critical Discourse Analysis, Annual Review of Anthropology, 29: 447
466.
Boden, D. and Zimmerman, D. H. (eds) (1991) Talk and Social Structure: Studies in ethnomethodology and
conversation analysis. Berkeley: University of California Press.
Bourdieu, P. (1977) Outline of a Theory of Practice. Cambridge: Cambridge University Press.
Bourdieu, P. (1984) Homo Academics. Paris: Les ditions de Minuit.
Bourdieu, P. (1989) La Noblesse dtat: Grandes coles et esprit de corps. Paris: ditiones de Minuit.
Bourdieu, P. (1991) Language and Symbolic Power. Cambridge: Polity Press.
Bourdieu, P. and Wacquant, L. (1992) An Invitation to Reflexive Sociology. Cambridge: Polity Press.
Boykoff, M. T. and Boykoff, J. M. (2004) Balance as bias: global warming and the US prestige press, Global
Environmental Changes Human and Policy Dimensions 14: 125136.
Brown, B. M. and Coates, K. (1996) The Blair Revelation: Deliverance for whom? Nottingham: Spokesman.
Brown, G. and Yule, G. (1983) Discourse Analysis. Cambridge: Cambridge University Press.
Brnner, G. and Graefen, G. (eds) (1994) Texte und Diskurse. Opladen: Westdeutscher Verlag.
Bublitz, H. (1999) Foucaults Archologie des kulturellen Unbewuten: Zum Wissensarchiv und
Wissensbegehren moderner Gesellschaften. Frankfurt am Main: Campus.
Burr, V. (2003) Social Constructionism, 2nd edn. Hove: Routledge.
Caborn, J. (1999) Die Presse und die Hauptstadtdebatte: Konstrukte der deutschen Einheit, in U. Kreft, H.
Uske and S. Jger (eds), Kassensturz: Politische Hypotheken der Berliner Republik. Duisburg: DISS, pp. 61
84.
Caborn, J. (2006) Schleichende Wende: Diskurse von Nation und Erinnerung bei der Konstituierung der
Berliner Republik. Mnster: Unrast.
Castells, M. (1996) The Rise of the Network Society. Oxford: Blackwell.
Charteris-Black, J. (2006) Politics and Rhetoric. Basingstoke: Palgrave.
Chilton, P. (1994a) La plaie quil convient de fermer: Les mtaphores du discours raciste, Journal of
Pragmatics 21(6): 583619.
Chilton, P. (ed.) (1994b) Schmas cognitifs du discours raciste francais. Vol. 4. Rotterdam: Institute for Social
Policy Research.
Chilton, P. (ed.) (1996a) The Meaning of Security. East Lansing: Michigan State University Press.
Chilton, P. (1996b) Security Metaphors: Cold War discourse from containment to common European home.
Bern: Peter Lang.
Chilton, P. (2004) Analysing Political Discourse: Theory and practice. London: Routledge.
Chilton, P. (2005a) Missing Links in Mainstream CDA: Modules, Blends and the Critical Instinct, in R. Wodak
and P. Chilton (eds), A New Research Agenda in (Critical) Discourse Analysis: Theory and
Interdisciplinarity. (Reprinted in 2nd edn, 2007.) Amsterdam: Benjamins, pp. 1952.
Chilton, P. (2005b) Manipulations, Memes and Metaphors: The Case of Mein Kampf, in L. D. Saussure and P.
Schulz (eds), Manipulation. Amsterdam: Benjamins.
Chilton, P. (2007) Is it Possible to Compare Political Rhetoric across Cultures? International Round Table on
Discourse. Hong Kong: City University.
Chilton, P. (2008) Critical Discourse Analysis, in Cambridge Encyclopedia of the Language Sciences.
Cambridge: Cambridge University Press, in press.
Chilton, P. and Lakoff, G. (eds) (1995) Metaphor in Foreign Policy Discourse. Dartmouth: Aldershot.
Chilton, P. and Wodak, R. (2007) Preface, in R. Wodak and P. Chilton (eds), A New Research Agenda in
(Critical) Discourse Analysis: Theory and Interdisciplinarity. Amsterdam: Benjamins, pp. xixviii. (Revised
2nd edn.) Chilton, P., Tian, H. and Wodak, R. (eds) (forthcoming) New Discourses in Contemporary China.
Hong Kong: Hong Kong University Press.
Chouliaraki, L. (1995)Regulation and Heteroglossia in One Institutional Context: The Case of a Progressivist
English Classroom. Unpublished PhD thesis, University of Lancaster.
Chouliaraki, L. and Fairclough, N. (1999) Discourse in Late Modernity. Edinburgh: Edinburgh University Press.
Church, K. and Hanks, P. (1990) Word association norms, mutual information, and lexicography,
Computational Linguistics 16(1): 2229.
Clarke, J. and Newman, J. (1998) A Modern British People? New Labour and the reconstruction of social
welfare. Copenhagen: Department of Intercultural Communication and Management, Copenhagen Business
School.
Clear, J. (1993) From Firth Principles: Computational Tools for the Study of Collocation, in M. Baker, G. Francis
and E. Tognini-Bonelli (eds), Text and Technology: In Honour of John Sinclair. Amsterdam: Benjamins, pp.
271292.
Connerton, P. (1976) How Societies Remember. Cambridge: Cambridge University Press. (Reprinted 1996.)
Connolly, W. (1991) Identity/Difference. Ithaca: Cornell University Press.
Cotterill, J. (2001) Domestic discord, rocky relationships: semantic prosodies in representations of marital violence
in the O. J. Simpson trial, Discourse and Society 12(3): 291312.
de Beaugrande, R. (1997) The Story of Discourse Analysis, in T. A. van Dijk (ed.), Discourse as Structure
and Process (Discourse Studies: A Multidisciplinary Introduction. Vol. 1). London: Sage, pp. 3562.
de Cillia, R. and Wodak, R. (eds) (2008) Gedenken im Gedankenjahr. Innsbruck: Studienverlag.
Delanty, G., Wodak, R. and Jones, P. (eds) (2008) Migration, Identity and Belonging. Liverpool: Liverpool
University Press.
Deleuze, G. (1988) Foucault. Minneapolis: University of Minnesota Press.
Denzin, N. K. (1970) The Research Act in Sociology. London: Butterworth.
Department of Trade and Industry (DTI) (1998) White Paper on Competitiveness. London: The Stationery
Office.
Dring, H. and Hirschauer, S. (1997) Die Biographie der Dinge: Eine Ethnographie musealer Representation, in
S. Hirschauer and K. Amann (eds), Die Befremdung der eigenen Kultur. Frankfurt am Main: Suhrkamp, pp.
267297.
Drews, A., Gerhard, U. and Link, J. (1985) Moderne Kollektivsymbolik: Eine diskurstheoretisch orientierte
Einfhrung mit Auswahlbiographie, Internationales Archiv fr Sozialgeschichte der deutschen Literatur
(IASL). Tbingen: Sonderheft Forschungsreferate, pp. 256375.
Dreyfus, H. L. and Rabinow, P. (1982) Michel Foucault: Beyond Structuralism and Hermeneutics. Sussex:
The Harvester Press.
Duranti, A. and Goodwin, C. (eds) (1992) Rethinking Context: Language as an Interactive Phenomenon.
Cambridge: Cambridge University Press.
Durkheim, E. (1933) The Division of Labor in Society. New York: The Free Press.
Durkheim, E. (1976) The Elementary Forms of Religious Life. London: Allen and Unwin.
Durkheim, E. and Mauss, M. (1963) Primitive Classification. London: Cohen and West.
Duveen, G. and Lloyd, B. (eds) (1990) Social Representations and the Development of Knowledge.
Cambridge: Cambridge University Press.
Ehlich, K. (1983) Text und sprachliches Handeln: Die Entstehung von Texten aus dem Bedrfnis nach
berlieferung, in A. Assmann, J. Assmann and C. Hardmeier (eds), Schrift und Gedchtnis: Beitrge zur
Archologie der literarischen Kommunikation. Munich: Fink, pp. 2443.
Emerson, R. M. (1962) Power-dependence relations, American Sociological Review 27: 3141.
Emerson, R. M. (1975) Social exchange theory, Annual Review of Sociology 2: 335362.
Emerson, R. M., Fretz, R. I. and Shaw, L. L. (1995) Writing Ethnographic Fieldnotes. Chicago: University of
Chicago Press.
Ensink, T. and Sauer, C. (eds) (2003) The Art of Commemoration. Amsterdam: Benjamins.
Everson, S. (ed.) (1996) The Politics and the Constitution of Athens. Cambridge: Cambridge University Press.
Fairclough, N. (1989/1991) Language and Power. London: Longman. (Revised 2nd edn, 2001.)
Fairclough, N. (1992a) Discourse and Social Change. Cambridge: Polity Press.
Fairclough, N. (1992b) Critical Language Awareness. London: Longman.
Fairclough, N. (1993) Critical discourse analysis and the marketization of public discourse, Discourse and
Society 4(2): 133168.
Fairclough, N. (1995a) Critical Discourse Analysis. London: Longman.
Fairclough, N. (1995b) Media Discourse. London: Edward Arnold.
Fairclough, N. (1996) A reply to Henry Widdowsons Discourse analysis: a critical view, Language and
Literature 5(1): 4956. (Reprinted in M. Toolan [ed.] [2002] Critical Discourse Analysis. Vol. 3. London and
New York: Routledge, pp. 148155.)
Fairclough, N. (2000a) New Labour, New Language? London: Routledge.
Fairclough, N. (2000b) Represenciones del cambio en discurso neoliberal, Cuadernos de Relaciones
Laborales 16: 1336.
Fairclough, N. (2003) Analysing Discourse: Text Analysis for Social Research. London: Routledge.
Fairclough, N. (2005a) Critical discourse analysis, Marges Linguistiques 9: 7694.
Fairclough, N. (2005b) Critical Discourse Analysis in Transdisciplinary Research, in R. Wodak and P. Chilton
(eds), A New Agenda in (Critical) Discourse Analysis. Amsterdam: Benjamins, pp. 5370.
Fairclough, N. (2006) Language and Globalization. London: Routledge.
Fairclough, N. (forthcoming) Language, Reality and Power, in J. Culpeper, F. Katamba, P. Kerswill and R.
Wodak (eds), English Language and Linguistics. London: Palgrave.
Fairclough, N. and Wodak, R. (1997) Critical Discourse Analysis, in T. A. van Dijk (ed.), Discourse as Social
Interaction (Discourse Studies: A Multidisciplinary Introduction. Vol. 2). London: Sage, pp. 258284.
Fairclough, N., Jessop, B. and Sayer A. (2004) Critical Realism and Semiosis, in J. Joseph and J. Roberts (eds),
Realism, Discourse and Deconstruction. London: Routledge, pp. 2342.
Falkner, G., Treib, O., Hartlapp, M. and Leiber, S. (2005) Complying with Europe. Cambridge: Cambridge
University Press.
Fay, B. (1987) Critical Social Science. London: Polity Press.
Fiske, S.T. and Taylor, S. E. (1991) Social Cognition. 2nd edn. New York: McGraw-Hill.
Foucault, M. (1975) Surveiller et punir: Naissance de la prison. Paris: Gallimard.
Foucault, M. (1977) Language, Counter-Memory, Practice. (D. F. Bouchard and S. Simon, eds.). Ithaca:
Cornell University Press.
Foucault, M. (1979) Discipline and Punish: The Birth of the Prison. Harmondsworth: Penguin Books.
Foucault, M. (1980a) Truth and Power, in C. Gordon (ed.), Power/Knowledge: Selected Interviews and Other
Writings 19721977 by Michel Foucault. New York: Pantheon Books, pp. 107133.
Foucault, M. (1980b) The Confession of the Flesh, in C. Gordon (ed.), Power/Knowledge: Selected Interviews
and Other Writings 19721977 by Michel Foucault. New York: Pantheon Books, pp. 194228.
Foucault, M. (1983) Der Wille zum Wissen: Sexualitt und Wahrheit. Frankfurt am Main: Suhrkamp.
Foucault, M. (1991) Remarks on Marx: Conversations with Duccio Trombadori. Transl. by R. J. Goldstein
and J. Cascaito. New York: Semiotext(e).
Foucault, M. (1996) What is Critique?, in J. Schmidt (ed.), What is Enlightenment? Eighteenth-century
Answers and Twentieth-century Questions. Berkeley: University of California Press, pp. 382398.
Foucault, M. (2002) The Archaeology of Knowledge. London: Routledge. (1st English edn, 1972.)
Fowler, R., Hodge, R., Kress, G. and Trew, T. (eds) (1979) Language and Control. London: Routledge.
French, J. R. P. and Raven, B. (1959) The Bases of Social Power, in D. Cartwright (ed.), Studies in Social
Power. Ann Arbor: University of Michigan, pp. 150167.
Froschauer, U. (2002) Artefaktanalyse, in S. Khl and P. Strodtholz (eds), Methoden der
Organisationsforschung. Reinbek: Rororo, pp. 361395.
Gee, J. (2004) Discourse Analysis: Theory and method. London: Routledge.
Giddens, A. (1984) The Constitution of Society: Outline of the theory of structuration. Cambridge: Polity
Press.
Girnth, H. (1996) Texte im politischen Diskurs: Ein Vorschlag zur diskursorientierten Beschreibung von
Textsorten, Muttersprache 106(1): 6680.
Glaser, B. and Strauss, A. L. (1967) The Discovery of Grounded Theory: Strategies for qualitative research.
Chicago: Weidenfeld and Nicolson.
Gleason, H.A.J. (1973) Contrastive Analysis in Discourse Structure, in A. Makkai and D. G. Lockwood (eds),
Readings in Stratificational Linguistics. Tuscaloosa, AL: University of Alabama Press.
Goffman, E. (1970) Strategic Interaction. Oxford: Basil Blackwell.
Gore, A. (2007) Angriff auf die Vernunft. Munich: Riemann.
Graesser, A. C., Gernsbacher, M. A. and Goldman, S. R. (eds) (2003) Handbook of Discourse Processes.
Mahwah, NJ: L. Erlbaum.
Graham, P. (2002) Space and Cyberspace: On the enclosure of consciousness, in J. Armitage and J. Roberts
(eds), Living with Cyberspace: Technology and Society in the 21st Century. London: The Athlone Press.
Graham, P. and Paulsen, N. (2002) Third-sector discourses and the future of (un)employment: skilled labor, new
technologies, and the meaning of work, Text 22(3): 443467.
Grant, D., Hardy, C., Oswick, C. and Putnam, L. (eds) (2004) The Sage Handbook of Organizational
Discourse. London: Sage.
Grimes, J. (1975) The Thread of Discourse. The Hague: Mouton.
Gruber, H. (1991) Antisemitismus und Mediendiskurs. Die Affre Waldheim in der Tagespresse.
Wiesbaden: Deutscher Universittsverlag.
Gumperz, J. J. (1982) Discourse Strategies. Cambridge: Cambridge University Press.
Habermas, J. (1967) Erkenntnis und Interesse. Frankfurt am Main: Suhrkamp.
Habermas, J. (1996) Die Einbeziehung des Anderen. Studien zur politischen Theorie. Frankfurt am Main:
Suhrkamp.
Halliday, M. A. K. (1978) Language as Social Semiotic. London: Arnold.
Halliday, M. A. K. (1985) An Introduction to Functional Grammar. London: Arnold. (2nd edn, 1994.)
Hammersley, M. and Atkinson, P. (2007) Ethnography: Principles in Practice. 3rd edn. London: Routledge.
Hardt-Mautner, G. (1995) Only Connect: Critical Discourse Analysis and Corpus Linguistics. UCREL Technical
Paper 6, University of Lancaster, Lancaster. Available at www.comp.lancs.ac.uk/ucrel/tech_papers.html.
Harvey, D. (1996) Justice, Nature and the Geography of Difference. Oxford: Blackwell. Harvey, D. (2003)
The New Imperialism. Oxford: Oxford University Press.
Hay, C. (2007) Why We Hate Politics. Cambridge: Polity Press.
Heer, H., Manoschek, W., Pollak, A. and Wodak, R. (eds) (2008) The Discursive Construction of History:
Remembering the Wehrmachts War of Annihilation. Basingstoke: Palgrave.
Horkheimer, M. and Adorno, T. W. (1969/1991 [1944]) Dialektik der Aufklrung: Philosophische Fragmente.
Frankfurt am Main: Fischer.
Hunston, S. (2002) Corpora in Applied Linguistics. Cambridge: Cambridge University Press.
Hunston, S. (2004) Counting the Uncountable: Problems of identifying evaluation in a text and in a corpus, in A.
Partington, J. Morley and L. Haarman (eds), Corpora and Discourse. Bern: Peter Lang, pp. 157188.
Iedema, R. (1997) Interactional Dynamics and Social Change: Planning as morpho-genesis. Unpublished PhD
thesis, University of Sydney.
Iedema, R. and Wodak, R. (1999) Introduction: organizational discourse and practices, Discourse and Society
10(1): 519.
Iet cu, I. (2006) Discourse Analysis and Argumentation Theory. Bucharest: Editura Universit ii din Bucures ti.
IPCC (Intergovernmental Panel on Climate Change) (2007a) Climate Change 2007: Synthesis report,
Intergovernmental Panel on Climate Change Fourth Assessment Report. Topic 1 (Downloaded from
www.ipcc.ch/pdf/assessment-report/ar4/syr/ar4_syr_topic1.pdf on 9 February 2008).
IPCC (Intergovernmental Panel on Climate Change) (2007b) Climate Change 2007: Synthesis report,
Intergovernmental Panel on Climate Change Fourth Assessment Report. Topic 2 (Downloaded from
www.ipcc.ch/pdf/assessment-report/ar4/syr/ar4_syr_topic2.pdf on 9 February 2008).
Jger, M. (1996) Fatale Effekte: Die Kritik am Patriarchat im Einwanderungsdiskurs. Duisburg: DISS.
Jger, S. (2001a) Dispositiv, in M. S. Kleiner (ed.), Michel Foucault: Eine Einfhrung in sein Denken.
Frankfurt am Main: Campus, pp. 7289.
Jger, S. (2001b) Discourse and Knowledge: Theoretical and methodological aspects of a critical discourse and
dispositive analysis, in R. Wodak and M. Meyer (eds), Methods of Critical Discourse Analysis. London: Sage,
pp. 3262.
Jger, S. (2004) Kritische Diskursanalyse. 4th unrevised edn. Mnster: Unrast.
Jger, S. (ed.) (2008) Wie kritisch ist die Kritische Diskursanalyse? Mnster: Unrast.
Jger, M. and Jger, S. (2007) Deutungskmpfe: Theorie und Praxis Kritischer Diskursanalyse. Wiesbaden:
VS Verlag.
Jessop, B. (2002) The Future of the Capitalist State. Cambridge: Polity Press.
Jessop, B. (2004) Critical semiotic analysis and cultural political economy, Critical Discourse Studies 1(2):
159174.
Jessop, B. (2008) The cultural political economy of competitiveness and its implications for higher education, in
B. Jessop, N. Fairclough and R. Wodak (eds), Education and the Knowledge-based Economy in Europe.
Amsterdam: Sense Publishers, pp. 1341.
Jessop, R., Fairclough, N. and Wodak, R. (eds) (2008) Higher Education and the Knowledge Based Economy
in Europe. Rotterdam: Sense Publishers.
Johnson-Laird, P. N. (1983) Mental Models. Cambridge: Cambridge University Press.
Katz, E. and Lazarsfeld, P. F. (1955) Personal Influence. Glencoe: Free Press.
Kienpointner, M. (1992) Alltagslogik: Struktur und Funktion von Argumentationsmustern. Stuttgart-Bad
Cannstatt: Frommann-Holzboog.
Kienpointner, M. (1996) Vernnftig Argumentieren: Regeln und Techniken der Diskussion. Hamburg:
Rowohlt.
Kienpointner, M. and Kindt, W. (1997) On the problem of bias in political argumentation: an investigation into
discussions about political asylum in Germany and Austria, Journal of Pragmatics 27: 555585.
Kindt, W. (1992) Argumentation und Konfliktaustragung in uerungen ber den Golfkrieg, Zeitschrift fr
Sprachwissenschaft 11: 189215.
Kintsch, W. (1998) Comprehension: A paradigm for cognition. New York: Cambridge University Press.
Klaus, V. (2007) Blauer Planet in grnen Fesseln!Was ist bedroht: Klima oder Freiheit? Vienna: Carl
Gerolds Sohn Verlagsbuchhandlung.
Klemperer, V. (1975/1947) LTI Lingua Tertii Imperii: Die Sprache des Dritten Reiches. Leipzig: Reclam.
Klemperer, V. (2000/2006 [1947]) The Language of the Third Reich: LTI Lingua Tertii Imperii: A
Philologists Notebook. London: Continuum.
Klemperer, V. (2001) I Will Bear Witness: A diary of the Nazi years. New York: Modern Library.
Knight, K. (2006) Transformations of the concept of ideology in the twentieth century, American Political
Science Review 100(4): 619626.
Koller, V. (2008) Identity, Image, Impression: Corporate self-presentation and public reactions, in R. Wodak and
V. Koller (eds), Handbook of Communication in the Public Sphere (Handbook of Linguistics, Vol. 4).
Berlin: de Gruyter, pp. 155180.
Koller, V. and Mautner, G. (2004) Computer Applications in Critical Discourse Analysis, in A. Hewings, C.
Coffin and K. OHalloran (eds), Applying English Grammar. London: Arnold, pp. 216228.
Koller, V. and Davidson, P. (2008) Social exclusion as conceptual and grammatical metaphor: a cross-genre
study of British policymaking, Discourse and Society 19(3): 307331.
Kovcs, A. and Wodak, R. (eds) (2003) Nato, Neutrality and National Identity: The case of Austria and
Hungary. Vienna: Bhlau.
Kress, G. and Hodge, R. (1979) Language as Ideology. London: Routledge.
Kress, G. and Van Leeuwen, T. (1996) Reading Images. London: Routledge.
Krings, H., Baumgartner, H. and Wild, C. (1973) Handbuch philosophischer Grundbegriffe. Munich: Ksel.
Krishnamurthy, R. (1996) Ethnic, Racial and Tribal: The language of racism? in C. R. Caldas-Coulthard and M.
Coulthard (eds), Texts and Practices: Readings in Critical Discourse Analysis. London: Routledge, pp. 129
149.
Krzy anowski, M. and Oberhuber, F. (2007) (Un)Doing Europe. Bern: Peter Lang.
Krzy anowski, M. and Wodak, R. (2008) The Politics of Exclusion: Debating migration in Austria. New
Brunswick, NJ: Transaction Press.
Laclau, E. (1980) Populist rupture and discourse, Screen Education 34: 8793.
Lakoff, G. (1987) Women, Fire and Dangerous Things. Chicago: University of Chicago Press.
Lakoff, G. and Johnson, M. (1980) Metaphors We Live By. Chicago: University of Chicago Press.
Lakoff, G. and Johnson, M. (1999) Philosophy in the Flesh. New York: Basic Books.
Lalouschek, J., Menz, F. and Wodak, R. (1990) Alltag in der Ambulanz: Gesprche zwischen rzten,
Schwestern und Patienten. Tbingen: Narr.
Lazar, M. (ed.) (2005) Feminist Critical Discourse Analysis. Basingstoke: Palgrave.
Le, E. (2006) The Spiral of Anti-Other Rhetoric. Amsterdam: Benjamins.
Lemke, J. (1995) Textual Politics: Discourse and social dynamics. London: Taylor and Francis.
Lemke, J. (2001) Discursive technologies and the social organization of meaning, Special Issue of Folia
Linguistica 35(12): 7996.
Lemke, J. (2002) Multimedia Genres for Science Education and Scientific Literacy, in M. Schleppegrell and C.
Colombi (eds), Developing Advanced Literacy in First and Second Languages. Mahwah, NJ: Lawrence
Erlbaum, pp. 2144.
Leontev, A. N. (1978) Activity, Consciousness, and Personality. Englewood Cliffs: Prentice Hall.
Leontev, A. N. (1982) Ttigkeit, Bewutsein, Persnlichkeit. Cologne: Pahl-Rugenstein.
Levinson, S. C. (1983) Pragmatics. Cambridge: Cambridge University Press.
Lvi-Strauss, C. (1964) Totemism. Harmondsworth: Penguin.
Link, J. (1982) Kollektivsymbolik und Mediendiskurse, kulturRRevolution 1: 621.
Link, J. (1983) Was ist und was bringt Diskurstaktik, kulturRRevolution 2: 6066.
Link, J. (1992) Die Analyse der symbolischen Komponente realer Ereignisse: Ein Beitrag der Diskurstheorie zur
Analyse neorassistischer uerungen, in S. Jger and F. Januschek (eds), Der Diskurs des Rassismus.
Oldenburg: Osnabrcker Beitrge zur Sprachtheorie 46, pp. 3752.
Link, J. (1996) Wie ideologisch war der Ideologiebegriff von Marx? Zur verkannten Materialitt der Diskurse
und Subjektivitten im Marxschen Materialismus, in R. Scholz and K.-M. Bogdal (eds), Literaturtheorie und
Geschichte: Zur Diskussion materialistischer Literaturwissenschaft. Opladen: Westdeutscher Verlag, pp.
132148.
Link, J. and Link-Heer, U. (1990) Diskurs/Interdiskurs und Literaturanalyse, Zeitschrift fr Linguistik und
Literaturwissenschaft (LiLi) 77: 8899.
Louw, B. (1993) Irony in the Text or Insincerity in the Writer? The diagnostic potential of semantic prosodies, in
M. Baker, G. Francis and E. Tognini-Bonelli (eds), Text and Technology: In honour of John Sinclair.
Amsterdam: Benjamins, pp. 157176.
Lueger, M. (2004) Grundlagen qualitativer Feldforschung. 2nd edn. Vienna: WUV.
Lutz, B. and Wodak, R. (1987) Information fr Informierte. Vienna: Verlag der sterr. Akademie der
Wissenschaften.
Maas, U. (1984) Als der Geist der Gemeinschaft eine Sprache fand: Sprache im Nationalsozialismus.
Opladen: Westdeutscher Verlag.
Maas, U. (1989a) Sprachpolitik und politische Sprachwissenschaft. Frankfurt am Main: Suhrkamp.
Maas, U. (1989b) Sprache im Nationalsozialismus: Analyse einer Rede eines Studentenfunktionrs, in K. Ehlich
(ed.), Sprache im Faschismus. Frankfurt am Main: Suhrkamp, pp. 162197.
Maier, F. (forthcoming) Doing Hair Doing Age: Perspectives of emancipated ageing, in M. Beisheim, B.
Gusenbauer, L. Kreil and F. Maier (eds), Perspectives of Womens Age at the Work Place. Bern: Peter Lang.
Malinowski, B. (1923) The Problem of Meaning in Primitive Languages, in C. K. Ogden and I.A. Richards
(eds), The Meaning of Meaning. London: Routledge and Kegan Paul.
Malinowski, B. (1935) Coral Gardens and their Magic. Vol. 2. London: Allen and Unwin.
Martin, J. R. (1984) Lexical Cohesion, Field and Genre: Parceling experience and discourse goals, in J. E.
Copeland (ed.), Linguistics and Semiotics: Text semantics and discourse semantics. Proceedings of the
Second Rice Symposium. Houston: Rice University Press.
Martin, J. R. (1992) English Text: System and structure. Amsterdam: Benjamins.
Martin, J. and Wodak, R. (eds) (2003) Re/reading the Past. Amsterdam: Benjamins.
Marx, K. (1992) Capital: A critique of political economy. London: Penguin. (1st German edn,1867.)
Matouschek, B., Wodak, R. and Januschek, F. (1995) Notwendige Manahmen gegen Fremde? Genese und
Formen von rassistischen Diskursen der Differenz. Vienna: Passagen Verlag.
Matsumoto, Y. (2003) Lexical Knowledge Acquisition, in R. Mitkov (ed.), The Oxford Handbook of
Computational Linguistics. Oxford: Oxford University Press, pp. 395413.
Mautner, G. (2005) Time to get wired: using web-based corpora in critical discourse analysis, Discourse and
Society 16(6): 809828.
Mautner, G. (2007) Mining large corpora for social information: the case of elderly, Language in Society 36(1):
5172.
Mautner, G. (2008) Analysing newspapers, magazines and other print media, in R. Wodak and M. Krzy
anowski, (eds), Qualitative Discourse Analysis in the Social Sciences. Basingstoke: Palgrave, pp. 3053.
McEnery, T. and Wilson, A. (2001) Corpus Linguistics. 2nd edn. Edinburgh: Edinburgh University Press.
McEnery, T., Xiao, R. and Tono, Y. (2006) Corpus-based Language Studies: An advanced resource book.
London and New York: Routledge.
Merton, R. K. (1967) On Theoretical Sociology. New York: The Free Press.
Mitten, R. (1992) The Politics of Antisemitic Prejudice: The Waldheim phenomenon in Austria. Boulder, CO:
Westview Press.
Morse, J. M., Barrett, M., Mayan, M., Olson, K. and Spiers, J. (2002) Verification strategies for establishing
reliability and validity in qualitative research, International Journal of Qualitative Methods 1(2): 119.
Moscovici, S. (1982) On Social Representations, in J. P. Forgas (ed.), Social Cognition: Perspectives on
everyday understanding. London: Academic Press, pp. 181209.
Moscovici, S. (2000) Social Representations. Cambridge: Polity Press.
Mouffe, C. (2005) On the Political. London: Routledge.
Mouzelis, N. P. (1995) Sociological Theory: What went wrong? Diagnoses and remedies. London: Routledge.
Mulderrig, J. (2006) The Governance of Education: A Corpus-Based Critical Discourse Analysis of UK
Education Policy Texts 1972 to 2005. Unpublished PhD thesis, Department of Linguistics, Lancaster University.
Mller, M., Fuentes, U. and Kohl, H. (2007) Der UN-Weltklimareport: Bericht ber eine aufhaltsame
Katastrophe. Cologne: Kiepenheuer und Witsch.
Mullins, W. E. (1972) On the concept of ideology in political science, American Political Science Review 66:
498510.
Muntigl, P. (2002a) Policy, politics, and social control: a systemic functional linguistic analysis of EU employment
policy, Text 22(3): 393441.
Muntigl, P. (2002b) Politicization and Depoliticization: Employment policy in the European Union, in P. Chilton
and Ch. Schffner (eds), Politics as Text and Talk. Amsterdam: Benjamins, pp. 4579.
Muntigl, P., Weiss, G. and Wodak, R. (2000) European Union Discourses of Un/employment: An
interdisciplinary approach to employment policymaking and organizational change. Amsterdam:
Benjamins.
Musolff, A. (2004) Metaphors and Political Discourse. Basingstoke: Palgrave.
Nelson, M. (2005) Semantic associations in Business English: a corpus-based analysis, English for Specific
Purposes 25: 217234.
Ng, S. H. and Bradac, J. J. (1993) Power in Language. Newbury Park: Sage.
OHalloran, K. and Coffin, C. (2004) Checking Overinterpretation and Underinterpretation: Help from corpora in
critical linguistics, in A. Hewings, C. Coffin and K. OHalloran (eds), Applying English Grammar. London:
Arnold, pp. 275297.
Oevermann, U., Allert, T., Konau, E. and Krambeck, J. (1979) Die Methodologie einer objektiven Hermeneutik
und ihre allgemeine forschungslogische Bedeutung in den Sozialwissenschaften, in H.-G. Soeffner (ed.),
Interpretative Verfahren in den Sozialund Textwissenschaften. Stuttgart: Metzler, pp. 352434.
Oreskes, N. (2004) Beyond the ivory tower: the scientific consensus on climate change, Science 306: 1686
1690.
Orpin, D. (2005) Corpus linguistics and critical discourse analysis: examining the ideology of sleaze,
International Journal of Corpus Linguistics 10(1): 3761.
Palonen, K. (1993) Introduction: From policy and polity to politicking and politicization, in K. Paolonen and T.
Parvikko (eds), Reading the Political: Exploring the margins of politics. Helsinki: FPSA, pp. 616.
Panagl, O. and Wodak, R. (eds) (2004) Text und Kontext. Munich: Knigshausen & Neumann.
Parsons, T. (1977) The Structure of Social Action. Chicago: The Free Press.
Parsons, T. and Shils, E. (eds) (1951) Towards a General Theory of Action. Cambridge, MA: University Press.
Partington, A. (2004) Utterly content in each others company: semantic prosody and semantic preference,
International Journal of Corpus Linguistics 9(1): 131156.
Pelinka, A. and Wodak, R. (eds) (2002) Dreck am Stecken: Politik der Ausgrenzung. Vienna: Czernin.
Pieterse, J. (2004) Globalization or Empire? London: Routledge.
Popitz, H. (1992) Phnomene der Macht. Tbingen: Mohr.
Rahmstorf, S. and Schellnhuber, H. J. (2007) Der Klimawandel, 4th edition. Munich: Beck.
Rancire, J. (1995) On the Shores of Politics. London: Verso.
Rancire, J. (2006) Hatred of Democracy. London: Verso.
Rapley, J. (2004) Globalization and Inequality: Neoliberalisms downward spiral. Boulder, CO: Lynne
Rienner.
Reeves, F. (1989) British Racial Discourse: A study of British political discourse about race and related
matters. Cambridge: Cambridge University Press.
Reisigl, M. (2003) Wie man eine Nation herbeiredet. Eine diskursanalytische Untersuchung zur sprachlichen
Konstruktion der stereichischen Identitt in politischen Gedenkreden. Unpublished PhD thesis, University of
Vienna.
Reisigl, M. (2007) Nationale Rhetorik in Fest-und Gedenkreden: Eine diskursanalytische Studie zum
sterreichischen Millennium in den Jahren 1946 und 1996. Tbingen: Stauffenburg.
Reisigl, M. and Wodak, R. (2001) Discourse and Discrimination: Rhetorics of racism and antisemitism.
London: Routledge.
Renkema, J. (2004) Introduction to Discourse Studies. Amsterdam: Benjamins.
Richardson, J. (2004) (Mis)representing Islam. Amsterdam: Benjamins.
Richardson, J. and Wodak, R. (2008) Visual argumentation in racist discourse, Controversies (in press).
Rose, N. (1999) Powers of Freedom: Reframing political thought. Cambridge: Cambridge University Press.
Rydgren, J. (2005) Movements of Exclusion. New York: Nova.
Sacks, H. (1992) Lectures on Conversation. 2 volumes. Oxford: Blackwell.
Sarasin, P. (2003) Geschichtswissenschaft und Diskursanalyse. Frankfurt am Main: Suhrkamp.
Schank, R. and Abelson, R. (1977) Scripts, Plans, Goals and Understanding. Hillsdale, NJ: Lawrence
Erlbaum.
Scherer, K. R. and Giles, H. (1979) Social Markers in Speech. Cambridge: Cambridge University Press.
Schiffrin, D. (1994) Approaches to Discourse. Oxford: Blackwell.
Schiffrin, D., Tannen, D. and Hamilton, H. E. (eds) (2001) The Handbook of Discourse Analysis. Malden, MA:
Blackwell Publishers.
Silverman, D. (1993) Interpreting Qualitative Data. London: Sage.
Sondermann, K. (1997) Reading Politically: National anthems as textual Icons, in T. Carver and M. Hyvrinen
(eds), Interpreting the Political: New methodologies. London: Routledge, pp. 128142.
Spradley, J. P. (1979) The Ethnographic Interview. New York: Holt, Rinehart and Winston.
Spradley, J. P. (1980) Participant Observation. Fort Worth: Harcourt Brace Jovanovich College Publishers.
Strauss, A. (1987) Qualitative Analysis for Social Scientists. Cambridge: Cambridge University Press.
Strauss, A. and Corbin, J. (1990) Basics of Qualitative Research. Newbury Park: Sage.
Stubbs, M. (1996) Text and Corpus Analysis: Computer-assisted studies of language and culture. Oxford:
Blackwell.
Stubbs, M. (2001) Words and Phrases: Corpus studies of lexical semantics. Oxford: Blackwell.
Talbot, M. R. (2003) Language and Power in the Modern World. Edinburgh: Edinburgh University Press.
Ten Have, P. (1999) Doing Conversation Analysis: A practical guide. London: Sage.
Teubert, W. (1999) Zum Verlust von Pluralitt im politisch-gesellschaftlichen Diskurs: Das Beispiel
Besitzstnde, in U. Kreft, H. Uske and S. Jger (eds), Kassensturz: Politische Hypotheken der Berliner
Republik. Duisburg: DISS, pp. 2948.
Thompson, J. B. (1988) Critical Hermeneutics (4th edn.) Cambridge: Cambridge University Press.
Thompson, J. B. (1990) Ideology and Modern Culture. Cambridge: Cambridge University Press.
Titscher, S., Meyer, M., Wodak, R. and Vetter, E. (2000) Methods of Text and Discourse Analysis. London:
Sage.
Tognini-Bonelli, E. (2001) Corpus Linguistics at Work. Amsterdam: Benjamins.
Toolan, M. (ed.) (2002) Critical Discourse Analysis: Critical concepts in linguistics. 4 vols. London:
Routledge.
Triandafyllidou, A., Wodak, R. and Krzy anowski, M. (eds) (2009) European Media in Crisis and the
European Public Sphere. Basingstoke: Palgrave.
Tricento, T. (ed.) (2005) Introduction to the Study of Language Policies. Oxford: Blackwell.
Tulving, E. (1983) Elements of Episodic Memory. Oxford: Oxford University Press.
US Congress (2007) Hearing on Perspectives on Climate Change (downloaded from
frwebgate.access.gpo.gov/cgi-bin/getdoc.
cgi?dbname=110_house_hearings&docid=f:37579.pdf on 10 April 2008).
Van Dijk, T. A. (1984) Prejudice in Discourse: An analysis of ethnic prejudice in cognition and
conversation. Amsterdam: Benjamins.
Van Dijk, T. A. (1987) Communicating Racism: Ethnic prejudice in thought and talk. Newbury Park, CA:
Sage.
Van Dijk, T. A. (1991) Racism and the Press. London: Routledge.
Van Dijk, T. A. (1993a) Elite Discourse and Racism. Newbury Park, CA: Sage.
Van Dijk, T. A. (1993b) Principles of critical discourse analysis, Discourse and Society 4(2): 249283.
Van Dijk, T. A. (ed.) (1997) Discourse as Structure and Process (Discourse Studies: A Multidisciplinary
Introduction. 2 vols.) London: Sage.
Van Dijk, T. A. (1998) Ideology: A multidisciplinary approach. London: Sage.
Van Dijk, T. A. (2005) Racism and discourse in Spain and Latin America. Amsterdam: Benjamins.
Van Dijk, T. A. (ed.) (2007a) Discourse Studies. 5 vols. London: Sage.
Van Dijk, T. A. (ed.) (2007b) Racismo y Discurso en Amrica Latina. Barcelona: Gedisa. (English version to
appear with Lexington Books.)
Van Dijk, T. A. (2008a) Discourse and Context. A sociocognitive approach. Cambridge: Cambridge
University Press.
Van Dijk, T. A. (2008b) Discourse and Power. Basingstoke: Palgrave.
Van Dijk, T. A. (2009) Society and Discourse: How context controls text and talk. Cambridge: Cambridge
University Press.
Van Dijk, T. A. and Kintsch, W. (1983) Strategies of Discourse Comprehension. New York: Academic Press.
Van Eeemeren, F. H. and Grootendorst, R. (1992) Argumentation, Communication and Fallacies: A
pragmadialectical perspective. Hillsdale, NJ: Lawrence Erlbaum.
Van Leeuwen, T. (2005) Introducing Social Semiotics. London: Routledge.
Van Leeuwen, T. (2006) Critical Discourse Analysis, in K. Brown (ed.), Encyclopedia of Language and
Linguistics. 2nd edn. Vol. 3. Oxford: Elsevier, pp. 290294.
Van Leeuwen, T. (2007) Legitimation in discourse and communication, Discourse and Communication 1(1):
91112.
Van Leeuwen, T. (2008) Discourse and Practice: New tools for critical discourse analysis. New York:
Oxford University Press.
Van Leeuwen, T. and Wodak, R. (1999) Legitimizing immigration control: a discourse-historical analysis,
Discourse Studies 1(1): 83118.
Van Oostendorp, H. and Goldman, S. R. (eds) (1999) The Construction of Mental Representations During
Reading. Mahwah, NJ: Lawrence Erlbaum.
Vass, E. (1992) Diskursanalyse als interdisziplinres Forschungsgebiet. Unpublished MA thesis, University of
Vienna.
Verdoolaege, A. (2008) Reconciliation Discourse. Amsterdam: Benjamins.
Viehver, W. (2003) Die Wissenschaft und die Wiederverzauberung des sublunaren Raumes. Der Klimadiskurs
im Licht der narrativen Diskursanalyse, in R. Keller, A. Hirseland, W. Schneider and W. Viehver (eds),
Handbuch Sozialwissensch-aftliche Diskursanalyse. Vol. 2: Forschungspraxis. Opladen: Leske and Budrich,
pp. 233270.
Vygotsky, L. S. (1986) Thought and Language. Revised edn. Cambridge, MA: The MIT Press. (Originally
published in 1934.)
Wacquant, L. (2004) Body and Soul: Notebooks of an apprentice boxer. Oxford: Oxford University Press.
Wagner, W. (1994) Alltagsdiskurs: Die Theorie sozialer Reprsentationen. Gttingen: Hogrefe.
Waldenfels, B. (1991) Ordnung in Diskursen, in F. Ewald and B. Waldenfels (eds), Spiele der Wahrheit:
Michel Foucaults Denken. Frankfurt am Main: Suhrkamp, pp. 277297.
Webb, E. J., Campbell, D. T., Schwartz, R. D. and Seclrest, L. (1966) Unobtrusive Measures: Nonreactive
research in the social sciences. Chicago: Rand-McNally.
Weber, M. (1980) Wirtschaft und Gesellschaft. 5th edn. Tbingen: Mohr.
Weiss, G. and Wodak, R. (eds) (2003) Critical Discourse Analysis: Theory and interdisciplinarity.
Basingstoke: Palgrave.
Wengeler, M. (2003) Topos und Diskurs. Tbingen: Niemeyer.
Wetherell, M., Taylor, S. and Simeon, Y. (eds) (2001) Discourse as Data. London: Sage
Whorf, B. L. (1956) Language, Thought and Reality. Cambridge, MA: MIT Press.
Widdowson, H. G. (1995) Discourse analysis: a critical view, Language and Literature 4(3): 157172.
(Republished in M. Toolan [ed.] [2002] Critical Discourse Analysis. Vol. 3. London and New York: Routledge,
pp. 131147.)
Widdowson, H. G. (2004) Text, Context, Pretext. Oxford: Blackwell.
Wittgenstein, L. (1989 [1952]) Philosophische Untersuchungen, in L. Wittgenstein (1989) Werkausgabe Band
1: Tractatus logico-philosophicus Tagebcher 19141916 Philosophische Untersuchungen. Frankfurt
am Main: Suhrkamp, pp. 224580.
Wodak, R. (1986) Language Behavior in Therapy Groups. Los Angeles: California University Press.
Wodak, R. (ed.) (1989) Language, Power and Ideology. Amsterdam: Benjamins.
Wodak, R. (1996) Disorders of Discourse. London: Longman.
Wodak, R. (2004) Critical Discourse Analysis, in C. Seale, G. Gobo, J. Gubrium and D. Silverman (eds),
Qualitative Research Practice. London: Sage, pp. 197213.
Wodak, R. (2006a) Dilemmas of discourse (analysis), Language in Society 35: 595611.
Wodak, R. (2006b) Critical linguistics and critical discourse analysis, in J.-O. stman and J. Verschueren (eds),
Handbook of Pragmatics. Amsterdam: Benjamins, pp. 124.
Wodak, R. (2007) Pragmatics and critical discourse analysis, Pragmatics and Cognition 15(1): 203225.
Wodak, R. (2008a) Introduction: Terms and concepts, in R. Wodak and M. Krzy anowski (eds), Qualitative
Discourse Analysis in the Social Sciences. Basingstoke: Palgrave, pp. 142.
Wodak, R. (2008b) The Contribution of Critical Linguistics to the Analysis of Discriminatory Prejudices and
Stereotypes in the Language of Politics, in R. Wodak and V. Koller (eds), Handbook of Communication in the
Public Sphere (Handbook of Linguistics, Vol. 4). Berlin: de Gruyter, pp. 291316.
Wodak, R. (2008c) Us and Them: Inclusion/Exclusion discrimination via discourse, in G. Delanty, R.
Wodak and P. Jones (eds), Migration, Identity and Belonging. Liverpool: Liverpool University Press, pp. 54
78.
Wodak, R. (2009) Politics as Usual: The discursive construction and representation of politics in action.
Basingstoke: Palgrave.
Wodak, R. and Chilton, P. (eds) (2005) A New Agenda in (Critical) Discourse Analysis. Amsterdam:
Benjamins. (2nd edn 2007.)
Wodak, R. and de Cillia, R. (2006) Politics and Language: Overview, in K. Brown (ed.), Encyclopedia of
Language and Linguistics. 2nd edn. Vol. 9. Oxford: Elsevier, pp. 706719.
Wodak, R. and de Cillia, R. (2007) Commemorating the past: the discursive construction of official narratives
about the rebirth of the second Austrian Republic, Discourse and Communication 1(3): 337363.
Wodak, R. and Koller, V. (eds) (2008) Handbook of Communication in the Public Sphere. Berlin: de Gruyter.
Wodak, R. and Krzy anowski, M. (eds) (2008) Qualitative Discourse Analysis in the Social Sciences.
Basingstoke: Palgrave.
Wodak, R. and Meyer, M. (eds) (2001) Methods of Critical Discourse Analysis. London: Sage.
Wodak, R. and Pelinka, A. (2002) From Waldheim to Haider: An introduction, in R. Wodak and A. Pelinka
(eds), The Haider Phenomenon in Austria. New Brunswick: Transaction Press, pp. viixxvii.
Wodak, R. and Schulz, M. (1986) The Language of Love and Guilt. Amsterdam: Benjamins.
Wodak, R. and van Dijk, T. A. (eds) (2000) Racism at the Top. Klagenfurt: Drava.
Wodak, R. and van Leeuwen, T. (2002) Discourses of un/employment in Europe: the Austrian case, Text 22(3):
345367.
Wodak, R., de Cillia, R., Reisigl, M. and Liebhart, K. (1999) The Discursive Construction of National Identity.
Edinburgh: Edinburgh University Press. (2nd edn 2009.)
Wodak, R., Menz, F., Lutz, B. and Gruber, H. (1985) Die Sprache der Mchtigen und Ohnmchtigen: Der
Fall Hainburg. Vienna: Arbeitsgemeinschaft fr staatsbrgerliche Erziehung und politische Bildung.
Wodak, R., Nowak, P., Pelikan, J., Gruber, H., de Cillia, R. and Mitten, R. (1990) Wir sind alle unschuldige
Tter!: Diskurshistorische Studien zum Nachkriegsantisemitismus. Frankfurt am Main: Suhrkamp.
Wright, W. (1975) Sixguns and Society: A Structural Study of the Western. Berkeley: University of California
Press.
Wuthnow, R. (1989) Communities of Discourse: Ideology and social structure in the Reformation, the
Enlightenment, and European socialism. Cambridge, MA: Harvard University Press.
Young, L. and Fitzgerald, B. (2006) The Power of Language: How discourse influences society. London:
Equinox.
Index

abductive methodology, 30
ability modality, 153
abstraction, 157
activation of actions and reactions, 156
activity sequences, 147
activity theory, 423, 45
annotation process, 132
Anthropological Museum, Vienna, 5960
argumentation analysis, 17, 26, 96, 102, 110, 114
Aristotle, 172
artefact analysis, 5960
Austria, 49, 95
Aztec culture, 5960

Baker, P., 128, 138, 142

Barton, Joe, 923, 101
Bell, Allan, 99
Berger, P., 154
Bernstein, B., 147, 163
Bhaskar, R., 167
bias on the part of researchers, 123
Billig, Michael, 45
Blair, Tony, 144, 171, 1745, 17881
Bourdieu, Pierre, 7, 1467
bricolage, 45, 60
British National Corpus (BNC), 139, 141
Brown, B. M., 181
Bush, George W., 144

Caborn, J., 59
Castells, M., 175
catachreses, 48
Center for the Moral Defense of Capitalism, 6772, 76, 83
Chernobyl accident, 489
Chilton, Paul, 7, 1416, 25, 183
Chomsky, Noam, 5
Chouliaraki, L., 12, 166
Clarke, J., 180
climate change, 90120
Clinton, Bill, 180
Coates, K., 181
cognition, definition of, 646
cognitive theory, 1315, 76
coherence, referential and extensional, 76
collective symbols, 38, 44, 478, 53
commemorative events, 1819
competencies, 1456
computer games, 16
concordance programs, 122, 12530, 135, 13841
context, concept of, 201, 25
context models, 656, 69, 734, 779
corpora, types of, 1313
corpus linguistics, 12, 15, 267, 30, 12242
design issues in, 12931
limitations of, 142
corpus mark-up, 132, 141
Corts, Hernn, 5960
critical analysis (CrA), 623, 669, 73
critical discourse analysis (CDA), 132
aims of, 36
central questions of, 34
characteristics of, 21
and corpus linguistics, 1226, 12933, 13840
definition of, 4, 10
history of, 34
principles of, 3, 5, 19
research agenda of, 1119
and semiosis, 1635, 1823
school for, 5
theoretical influences on, 237, 32, 359
use of term, 12
critical discourse studies (CDS), 2, 6283
social aspects of, 7980
critical instinct, 14
critical linguistics, 1, 5, 7
critical social research, 163
critical theory, 67, 26
critique, concepts and definitions of, 878, 164

data-gathering for research, 278, 32

de-agentialization, 1567
de Beaugrande, R., 122
decontextualization, 150
deductive methodology, 19, 30
democracy deficit, 17
democratic politics, 1723
descriptivization, 156
dialectical-relational approach to CDA, 27, 301, 16283
methodology of, 16771
discourse
concepts and definitions of, 26, 25, 35, 45, 667, 1234, 144, 147, 1624
as distinct from text, 8990, 1445
macropolitics of, 114
orders of, 27, 1646
and power, 379, 89
and reality, 367, 41, 45
and social practice, 148
and social structures, 83
societal and global, 501
see also political discourse
discourse analysis, 14, 17, 24, 28, 31, 345
comparative, 14
completeness of, 51
foundational assumptions of, 129
multimedia type, 5960
quantitative, 51
sociocognitive, 645
toolbox for, 526
discourse fragments, 47
analysis of, 546
discourse planes, 4853
discourse prosody, 129
discourse semantics, 758; see also semantic macrostructures
discourse strands, 4654
entanglements of, 47, 54
history, present and future of, 51
discourse studies, 62
discourse theory, importance of, 169
Discourse, Politics, Identity (DPI) research centre, Vienna, 95
discourse-historical approach (DHA), 6, 18, 212, 2631, 87120
aims of, 88
strengths of, 11920
discourse-immanent critique, 88
discursive action, 81
discursive context, 49
discursive events, 489
discursive knots, 47
discursive limits, 47
discursive and nondiscursive practices, 43, 5860
discursive positions, 4950
discursive strategy, 94
dispositive analysis (DA), 25, 31, 345, 3942, 45, 5660
distillation, 15760
Durkheim, mile, 256, 1467

enactment of discourse, 165

enlightenment philosophy, 8
environmentalism, 49, 923, 11415, 118
epistemology, 24
ERASMUS programme, 4
European Union, 1812
event models, 769
Everson, S., 172
existentialization, 157

fallacies, 110, 11417, 178

false consciousness, 367
fascism, 17, 38
fields of action (Girnth), 901
fine analysis, 28
Foucault, Michel, 910, 17, 217, 3445, 51, 56, 60, 144
Fowler, R., 148
Frankfurt School, 6, 87

Gates, Bill, 85
generalization, 157
genres, 13, 15, 27, 901, 131, 1645, 176
Germany, 39, 4750; see also National Socialism
Gillespie family, 1334
Gleason, H. A. J., 147
globalization and the global economy, 12, 16882
Gore, Al, 101
Graham, Phil, 1213
grand theories, 236, 95
Grimes, J., 147
Grootendorst, R., 110
grounded theory, 27

Habermas, Jrgen, 6, 10, 173

Haider, Jrg, 489
Halliday, Michael, 5, 21, 148
Hastert, J. Dennis, 923, 101
Heer, H., 19
hegemony, concept of, 8
hermeneutics, 223, 28
Hitler, Adolf, 144
Hodge, R., 148
Horkheimer, Max, 6
hortatory genres, 12
House of Representatives Perspectives on Climate hearing, 10111
human rights, 634

ideology and ideological analysis, 8, 65, 789, 88

image fractures, 48; see also catachreses
immigration, 44, 478, 52, 81, 83
inculcation of discourse, 165
inductive theories, 19
interdisciplinary research, 21, 89; see also transdisciplinary research
interdiscursive analysis, 46, 902, 95, 166, 170, 180
intertextuality, 902, 95, 166
inversion, 158

Jessop, R., 11
journals on critical discourse analysis, 4, 11

Kintsch, Walter, 13
Klaus, Vclav, 92, 98118
Klemperer, Victor, 38, 60, 144
knowledge, concepts of, 58
knowledge-based economy (KBE) concept, 5, 1112
Koller, V., 15
Kress, Gunther, 3, 15, 148
Krings, H., 7
Krzy anowski, M., 17

Laclau, Ernesto, 25, 42

language
of fascism, 17; see also National Socialism
and power, 10
and society, 21
see also linguistics
leadership, discourses of, 1445, 152, 155, 15960
Lemke, Jay, 6, 1516
Leontev, A. N., 25, 423, 45
Levi-Strauss, C., 1467
linguistics, 24, 41, 166; see also language
Link, Jrgen, 25, 35, 45
Lomborg, Bjrn, 101
Lominger-International, 145, 153

Maas, Utz, 17
McEnery, T., 129
Major, John, 176
Malinowski, B., 147
Martin, J. R., 147
Marxist theory, 367, 87
materialization of discourse, 59, 165
media influences, 1213, 1718, 48, 99
mediation, concept of, 21
mental models, 689, 73, 768, 82
micro-sociological theories, 24
Microsoft, 6777, 836
middle-range theories, 24
moral evaluation, 151
Moscovici, Serge, 256
Mouzelis, N. P., 26
Mulderrig, Jane, 12
myths, 158

National Socialism, 17, 38, 95, 144

naturalization, 157
neoliberalism, 50, 11718, 170, 176, 181
New Labour, 12, 129, 166, 171, 17582
new public management, 165
Newman, J., 180

Oberhuber, F., 17
objectivization, 160
operationalization (of discourse), 165
overdetermination, 1578

Parsons, Talcott, 1467

participant observation, 59
patterned bonding, 126
personal deixis, 178
Plato, 145
political discourse, 1719, 44
politicization and depoliticization, 17383
populism, 18, 48
power, 6, 910, 20, 60, 63, 82, 889, 183
concept and definition of, 35, 88
and discourse, 379, 89
prospective critique, 88

qualitative research, 31

racism, 52, 82, 95

Rancire, J., 1723
reconstruction
of discourse, 1512
of knowledge, 57
recontextualization, 90, 95, 119, 145, 14853, 1635, 17880
Richardson, John, 13
Rose, Nikolas, 12

sampling, 27, 130, 1412

Scollon, Ron, 3
semantic macrostructures, 689
semantic preference, 128, 1358
semiosis and semiotics, 1516, 27, 16383
semiotic actions, 156
sexism, 82
Silverman, D., 31
social action and social actor analysis, 21, 267, 1546
social class divisions, 1378
social cognition, 78
social inclusion and exclusion, 13, 18
social representation theory, 256, 789
social structures, 1314, 24, 27, 823
social wrongs, 16771, 1746, 1801
sociocognitive approach, 21, 25, 2831
socio-diagnostic critique, 88
socio-psychological theories, 24
special discourses, 46
stem cell research, 534
structural analysis, 28, 536
Stubbs, M., 128, 1356
styles, semiotic, 1645
the Sun, 1338
symbolization, 158; see also collective symbols
synchronic analysis, 57
synoptic analysis, 56
systemic functional linguistics, 27

tacit knowledge, 45, 59

text
definition of, 6
discursive differences in, 10
textual analysis, 166
Thatcher, Margaret, 166, 176
Thompson, J. B., 88
Three Mile Island accident, 489
The Times, 125, 1368
Tognini-Bonelli, E., 128
topoi, 467, 101, 110, 11417
transdisciplinary research, 163, 169, 182; see also interdisciplinary research
transmedia, 16
Trew, T., 148
triangulation procedures, 31, 89, 93, 119

unemployment, 1238
United Nations, 97

van Eeemeren, F., 110

Viehver, W., 119
Voices questionnaire, 1456, 14954, 1578
Vygotsky, Lev, 42

Waldenfels, Bernhard, 44
Waldheim, Kurt, 18, 95
Weber, Max, 88
Whorf, B. L., 148
Wordbanks Online, 125, 127, 131, 1346, 13941
Wordsmith Tools, 12930
World Bank, 17880
World Wide Web, 132, 140
Wright, W., 158