Encyclopedia of Language and Linguistics

Section Language and Politics – Ed. Ruth Wodak



Politics, Ideology and Discourse


Teun A. van Dijk

Universitat Pompeu Fabra, Barcelona


Second version, June 15, 2004




Few areas in the social sciences are as closely related as those of the study of politics, ideology and discourse. Politics is one of the social domains whose practices are virtually exclusively discursive; political cognition is by definition ideologically based; and political ideologies are largely reproduced by discourse. In this chapter we examine these relationships more closely.




The concept of ‘ideology’ is often used in the media and the social sciences, but it is notoriously vague. Its everyday usage is largely negative, and typically refers to the rigid, misguided or partisan ideas of others: we have the truth, and they have ideologies. This negative meaning goes back to Marx-Engels, for whom ideologies were a form of ‘false consciousness’; thus, the working class may have misguided ideas about the conditions of its existence as a result of their indoctrination by those who control the means of production. Throughout large part of the 20th century, and both in politics and in the social sciences, the notion of ideology continued to carry its negative connotation, and was often used in opposition to ‘objective’ knowledge (for histories of the notion of ideology, see, e.g., Billig, 1982; Eagleton, 1991; Larrain, 1979; for a useful collection of classical studies on ideology, see Zizek, 1994).

Originally, ‘ideology’ did not have this negative meaning. More than two hundred years ago, the French philosopher Destutt de Tracy introduced the term in order to denote a new discipline that would study ‘ideas’: idéologie. Also in contemporary political science, the notion is used in a more neutral, descriptive sense, e.g., to refer to political belief systems (Freeden, 1996).

One of the many dimensions highlighted in the classical approaches to ideology was their ‘dominant’ nature, in the sense that ideologies play a role in the legitimization of power abuse by dominant groups. One of the most efficient forms of ideological dominance is when also the dominated groups accept dominant ideologies as ‘natural’ or ‘commonsense’. Gramsci called such forms of ideological dominance ‘hegemony’ (Gramsci, 1971). Bourdieu does not use the notion of ideology very much (mainly because he thinks it is too vague and has often been abused to discredit others who do not agree with us; see Bourdieu & Eagleton, 1994), but rather speaks of symbolic power or symbolic violence. It should be stressed though that although related his uses of these terms are different from the (various) uses of the notion of ideology. His main interest lies in the social conditions of discursive and symbolic power, such as the authority and legitimacy of those who produce discourse.

To cut a long historical survey short, a specific concept of ideology will be used in this chapter, namely to describe specific, fundamental beliefs of groups of people. Our working definition of ideologies is therefore as follows: An ideology is the foundation of the social representations shared by a social group. Depending on one’s perspective, group membership or ethics, these group ideas may be valued ‘positively’, ‘negatively’ or not be valued at all. That is, we do not exclusively identify ideologies with dominant groups (see also the discussion in Abercrombie, Hill & Turner, 1980). Also dominated groups may have ideologies, namely ideologies of resistance and opposition. Ideologies more generally are associated with social groups, classes, casts or communities which thus represent their fundamental interests. The theory accounting for such ideological beliefs is complex and multidisciplinary, and may be summarized as follows (for details and many further references, see Van Dijk, 1998):


·        Ideologies have both social and cognitive properties which need to be accounted for in an integrated theory.

·        Cognitively, ideologies are a special kind of social belief systems, stored in Long Term Memory. (See Mental Representation).

·        Socially as well as cognitively, these ideological belief systems are socially shared by the members of specific social groups or ‘ideological communities’. (See Distributed cognition and communication).

·        Ideologies, like languages, are essentially social. There are no ‘personal’ or ‘individual’ ideologies – only personal or individual ‘uses’ of ideologies.

·        The identity of groups is not only based on their ‘structural’ properties, but also on their ideology.

·        Ideological belief systems – ideologies -- form the ‘axiomatic’ basis of the more specific beliefs or ‘social representations’ of a group, such as their group knowledge and group opinions (attitudes).

·        Unlike in most traditional approaches to ideology, ideologies are not necessarily ‘negative.’ They have similar structures and functions whether shared by dominant or dominated groups, ‘bad’ groups or ‘good’ groups. Thus, we may have negative as well as positive ideologies (‘utopias’), depending on the perspective, values or group membership of the one who evaluates them.

·        Not all socially shared beliefs of a group are ideological. Thus, ideologically different or opposed groups in the same society need to have beliefs in common in order to be able to communicate in the first place. This Common Ground consists of socioculturally shared knowledge, which by definition is pre-ideological within that society (although it may later or elsewhere be described as ideological knowledge).

·        Thus, the traditionally problematic relationship between knowledge and ideology is resolved as follows: General, sociocultural knowledge, shared by an epistemic community, forms the common ground for all social representations of all (ideological) groups in that community. However, each group may develop specific group knowledge (e.g., professional, religious or political knowledge) based on the ideology of the group. This knowledge is called ‘knowledge’ within the group because it is generally shared, certified and presupposed to be ‘true’. For other groups such knowledge may of course be called ‘mere belief’, superstition or religion. In other words, beliefs that are taken for granted, commonsense, non-disputed, etc. within a community, and shared by different ideological groups, is by definition non-ideological within that community.

·        Ideologies embody the general principles that control the overall coherence of the social representations shared by the members of a group. For instance, a racist ideology may control more specific attitudes about immigration or affirmative action.

·        Ideologically based social representations (such as feminist attitudes about abortion or glass ceilings on the job), are general and abstract. In order to relate to concrete social practices and discourses about specific events, they need to become contextualized and specified in mental models. These (ideological biased) mental models, stored in Episodic Memory, are the mental constructs that control discourse, interaction and other social practices. And conversely, it is through mental models that discourses are able to influence social representations and ideologies and reproduce these. (See Cognition, Emotion and Language).

·        Ideologies represent one of the dimensions of the social identity or self-image of groups.

·        Unlike less fundamental social representations and much more than variable personal models, ideologies are relatively stable. One does not become or cease to be a feminist, socialist or pacifist overnight. Many ideologies are acquired over many years and remain active for a lifetime of group members.

·        Ideologies are structured by a social schema consisting of a number of categories that cognitively represent the major social dimensions of groups, such as their distinguishing properties, membership criteria, typical actions, goals, norms and values, reference groups and basic resources or interests.

·        Both cognitively as well as socially, ideologies develop especially as socially shared resources for intragroup cohesion and cooperation, as well as for efficient means for intergroup relations.

·        Many – but not all -- ideologies are relevant in situations of competition, conflict, domination and resistance between groups, that is, as part of a social struggle. This also explains why many of the mental structures of ideologies and ideological practices are polarized on the basis of an ingroup-outgroup differentiation, typically between Us and Them¸ as also ideological discourses show.

·        Because individual people may be members of several groups, they may ‘participate’ in various ideologies. Thus, someone may be a nationalist, socialist, feminist journalist, and thus share in the ideologies of these different kinds of social and professional ideologies. Obviously, when activated (used) at the same time, in discourse or other social practices, this may sometimes lead to conflicts.

·        The social practices, and hence discourses, of group members may be (indirectly) controlled by group ideologies, but are usually mediated by more specific social representations at the group level and by concrete, personal mental models at the individual level.

·        Conversely, ideologies are personally acquired and socially reproduced by the social practices, and especially the discourses, of a group.

·        Groups may organize the discursive acquisition and reproduction of ideologies, for instance through special forms of education, indoctrination, job training, or catechesis, and by specialized group members (ideologues, priests, teachers, etc.) and in special institutions.

·        Not all group members have – nor need to have – the same level of ideological knowledge or expertise, nor need their ideological knowledge always be very explicit. Using an ideology is like being able to use a language without being able to formulate the grammar of that language. Many men are sexist and their sexist ideology may control much of their discourse and other social practices, but they need not always have explicit access to the contents of their ideologies.

·        However, since many social ideologies develop as part of group relations, conflict or domination and resistance, and hence involve ideological debate that is often published in the mass media, many group members know at least the main ideological tenets of their group – and of other groups. Indeed, when their interests are threatened they often know how and why to protect these.


These are some of the main properties of ideologies as formulated in a multidisciplinary, socio-cognitive theory. Thus, ideologies are the ‘axiomatic’ basis of the social representations of a group and – through specific social attitudes and then through personal mental models -- control the individual discourses and other social practices of group members. In this way they also are the necessary resource of ingroup cooperation, coordination and cohesion, as well as for the management of intergroup relations, competition, conflict or struggle. It is only within such a theory that we are able to account for ideological discourse and other social practices, namely as being derived from ideologically based social representations, and as instantiations of social relations between groups.

More than traditional approaches, this multidisciplinary approach not only emphasizes the social and political nature of ideologies, but also their sociocognitive nature. It should be emphasized though that this does not mean that especially or only this cognitive dimension is important. Unlike traditional social or socioeconomic approaches, the theory emphasizes that –trivially-- ideologies have to do with ideas of some kind, and hence also need a cognitive account besides a social theory of groups and group relations, power and interests. The point is that these different approaches need and can be integrated in one multidisciplinary theory. Hence, this approach does imply that a theory of ideology without an explicit cognitive component is incomplete: dealing about ideologies without talking about the nature and functions of socially shared ‘ideas’ is theoretically unsatisfactory.

We see that ideological social practices are by definition based on ideologies defined as shared mental representations of some kind, in a way that might be compared with the way language use is ‘based’ on a shared grammar or discourse and conversation rules. It is in this sense that ideologies as socially shared cognitive resources are fundamental for social practices, interaction and intra- and intergroup relations. Conversely, the general social functions of ideological practices must hence be represented as part of their ‘underlying’ ideologies. This is one of the many reasons why cognitive and social approaches to ideology need to be integrated.

The theory proposed here accounts for both the relatively stable as well as the flexible, dynamic, changing, contextualized and subjective aspects of ideology. The first dimension is explained in terms of relatively stable, socially shared mental representations of groups. The second dimension is accounted for by ideologically based, specific, subjective mental models of group members that control discourse and other social practices in each situation. Unlike other approaches, for instance in discursive psychology and other constructionist approaches (Billig, 1988, 1991; Potter, 1996), this theory does not attribute the flexible, subjective or contextually variable aspects of ideological practices to the nature of ideology itself, but to its ‘uses’ by individual members. Again, the comparison with relatively stable – and slowly changing – grammars of natural languages, and their variable, contextualized, personal uses, suggests itself. For the same reason, ideologies are not reduced to their ‘observable’ uses, discourses or other social practices, but defined as members’ socially shared underlying representations or resources that govern such practices. Nor do we reduce ideologies to discourses, because obviously they control also other social practices, such as forms of discrimination or violence. In sum, the theory presented here is not only multidisciplinary, but also non-reductionist.

Finally, ideologies are accounted for in sociocognitive rather than in emotional terms, because they are by definition socially shared, and in our definition of emotions, only individual persons and not groups can have, bodily based, emotions. When we sometimes speak of ideologies of ‘hate’, as is the case for racist or sexist ideologies, we are not speaking of emotions but of shared negative evaluations (opinions). Emotions are temporal, contextual and personal, physiologically based, and cognitively interpreted events. Thus one can have and share a more or less permanent negative opinion about immigrants, but one cannot, in the strict sense of the term be permanently ‘angry’ about immigrants, nor literally share an emotion with others. Thus, since ideologies are socially shared, they by definition cannot be ‘emotional’. However, their uses or applications by individual group members in concrete situations may of course trigger and be expressed as emotions. Also for this reasons it is essential to analytically distinguish between ideologies and their actual uses or manifestations in discourse, interaction and other social practices.


Ideology and politics


The general theory of ideology summarized above needs to be specified for the huge social ‘field’ of politics, that is, for politicians, political cognition, political processes, political practices and political discourse as characterizing political groups, such as political parties, members of parliaments or social movements. As soon as ideologies not only have general social functions but more specifically (also) political functions in the field of politics, we shall call them political ideologies. Thus, socialism is more obviously a ‘political’ ideology than the professional ideology of dentists, as long as we interpret ‘political’ here as describing processes in the field of politics – and not as part of the fields of health care, education or justice, among others. Thus, one way of classifying ideologies – as well as discourses -- is by the social field in which they function. That is, we have political, educational, legal, religious and health care ideologies, among others.

It is beyond the scope of this brief chapter to define and theorize in detail about what characterizes the field of politics (see, e.g., Goodin & Klingemann, 1996). However, apart from being defined by its prototypical participants (politicians), this field may briefly – and somewhat traditionally – be defined by


·      its overall systems (democracy, dictatorship, etc.),

·      special social macro actions, such as government, legislation, elections, or decision making,

·      and their micro practices, interactions or discourses such as parliamentary debates, canvassing or demonstrations,

·      its special social relations, such as those of institutional power,

·      its special norms and values (e.g., freedom, equality, etc.)

·      its political cognitions, such as political ideologies.


If there is one social field that is ideological, it is that of politics. This is not surprising because it is eminently here that different and opposed groups, power, struggle and interests are at stake. In order to be able to compete, political groups need to be ideologically conscious and organized. Few ideological groups besides political parties have ‘programs’ that formulate their ideologies explicitly, and that compete for new members or supporters on that basis. Few ideologies are as explicitly defended and contested as political ideologies, as we know from the history of socialism, communism, liberalism, and so on. In other words, the political process is essentially an ideological process, and political cognition often simply identified with ideology (see Freeden, 1996; Ball & Dagger, 1999; Eatwell, 1999; Leach, 2002; Seliger, 1976)

The social organization of the field of politics, and hence of politicians and political groups, is largely based on ideological differences, alliances and similarities. The overall organization of social beliefs as a struggle between the Left and the Right is the result of the underlying polarization of political ideologies that has permeated society as a whole. Elections, parliaments, political campaigns, propaganda, demonstrations, and many other phenomena of the political field are thus profoundly ideological. Debates in parliament pitch opposed political ideologies as basis for political policies, measures, decisions or actions. One’s political identity, stances and allegiances are not so much defined in structural group membership, such as membership of a political party, but rather in terms of one’s ideology. Most socialists or neoliberals do not have a membership card. The same is true for other social ideologies that have profound political implications, such as feminism, pacifism, ecologism or racism. 

Although primarily defined in sociocognitive terms, political ideologies permeate the whole political field, such as overall systems such as democracies (based on democratic ideologies), overall acts and processes (such as government, coalition building, or elections), everyday political practices (such as parliamentary debates or demonstrations), group relations (such as domination and resistance, government or opposition), fundamental norms and values (such as equality and independence that are constitutive categories of ideologies), as well as more specific political attitudes (for instance about legislation about abortion or divorce) that are controlled by ideologies.


Political discourse and ideology


If the political field is thoroughly ideological, then so are its political practices, and hence its discourses (among the many books on political discourse, see, e.g., Chilton, 1995, 2004; Chilton & Schäffner, 2002; Wilson, 1990; Wodak & Menz, 1990; see also the other contributions to this section). Indeed, political ideologies not only are involved in the production or understanding of political discourses and other political practices, but are also (re)produced by them. In a sense discourses make ideologies ‘observable’ in the sense that it is only in discourse that they may be explicitly ‘expressed’ and ‘formulated’. Other political practices only implicitly show or experience ideologies, for instance in practices of discrimination on the basis of sexist, racist or political ideologies. It is in discourse that we need to explicitly explain that such discrimination occurs ‘because she is a woman’, ‘because he is black’ or ‘because they are socialists’.

Thus, it is largely through discourse that political ideologies are acquired, expressed, learned, propagated and contested. The rest of this chapter will be about these relationships between political discourse and political ideologies. Interestingly, despite the vast literature on ideology (thousands of books in English alone), there are virtually no monographs that explore the details of the relations between discourse and ideology, although many books in critical linguistics and critical discourse analysis deal with at least some aspects of this relationship (see, e.g., Fairclough, 1989, 1995; Fowler, Hodge, Kress & Trew, 1979; Fowler, 1991; Hodge & Kress, 1993; Pêcheux, 1982; Van Dijk, 1998; Wodak, 1989; Wodak et al., 1987; Wodak & Menz, 1990; Wodak & Meyer, 2001).


Political situations and contexts


The relations between discourse and political ideologies are usually studied in terms of the structures of political discourse, such as the use of ‘biased’ lexical items, syntactic structures such as actives and passives, pronouns such as us and them, metaphors or topoi, arguments, implications and many others properties of discourse (see the references given at the end of the last section).

It should be emphasized though that discourse should be conceptualized also in terms of its context structures (Duranti & Goodwin, 1992). It is not sufficient to observe, for instance, that political discourse often features the well-known ‘political’ pronoun we. It is crucial to relate such use to such categories as who is speaking, when, where and with/to whom, that is, to specific aspects of the political situation.

Since such political situations do not simply ‘cause’ political actors to speak in such a way, we again need a cognitive interface between such a situation and talk or text, that is, a mental model of the political situation (van Dijk, 1999, 2001, 2003). Such mental models define how participants experience, interpret and represent the for-them-relevant aspects of the political situation. These specific mental models are called contexts. In other words, contexts are subjective participant definitions of communicative situations. They control all aspects of discourse production and comprehension.

Political discourse, thus, is not only defined in terms of political discourse structures but also in terms of political contexts. Thus, acting an MP, prime minister, party leader, or demonstrator will typically be perceived by speakers or recipients as a political relevant context category in political discourse, whereas being a dentist or a doorkeeper much less so. Similarly, political contexts may be defined by special settings, featuring locations such as parliamentary buildings or events such as debates or meetings, as often controlled by precise timing, as is the case in parliamentary debates. And political discourses and their structures will only be able to have the political functions they have when they are enacting political acts or processes, such as governing, legislation or ‘doing’ opposition, and with very specific political aims in mind, such as defending or defeating a bill or getting elected. And finally, political actors obviously do not participate mindlessly in political situations, but have political knowledge, share political norms and values, as well as political ideologies. Indeed, it is through this form of contextualization that we are able to link the ideologies of the participants to their discourses (Gumperz, 1982). Text or talk ‘show’ ideologies discursively, but it is people, politicians or protesters, who ‘have’ ideologies – not only in this social practice or discourse, but typically also in others.

These then are some of the kind of categories that make up our political context models – that is, political categories that we use to define political situations of text and talk. In the same way as discourses may be ideological when based on ideologies, also the structures and practices of political contexts may have such an ideological basis. Obviously, being an MP presupposes a parliamentary system and hence a democratic ideology, whereas being a dictator presupposes another ideology.

Obviously these categories are culturally variable: members of parliament, prime ministers or party secretaries are not exactly universal political participant categories. Other cultures may have their own specific political event types, political actions, participants, locations, time management, and of course their own political knowledge, attitudes, ideologies, norms and values.

A detailed explanation of the cognitive processes involved in the way context models control political discourse is beyond the scope of this chapter. Suffice it to say that the information in the various categories of the (pragmatic) context model – for instance who are participating in the communicative situation – first of all controls the speech acts and other acts of the current situation. Thus, the current utterance may be defined as a political promise or as a threat, depending on the power or relationships of the participants, their political position (government or opposition, my party or your party), as well as the intentions to help or harm the recipient. Secondly, pragmatic context models control the selection of information in the (semantic) mental model that (inter)subjectively defines what participants talk about, such as the war in Iraq. Thus, an MP or minister addressing his or her peers in parliament will express and presuppose very different knowledge than does a politician giving a speech or an interview. Thirdly, context models control all levels of style of political discourse, such as lexical choice, pronouns, syntactic structure and other grammatical choices that depend on how situations are defined. Thus, lexical and syntactic style in a parliamentary debate will be much more formal than an informal political meeting of party members or a propaganda leaflet. And finally, context models control the overall ‘format’ or ‘schema’ of political discourse, such as the formal turn-taking organization, openings and closings of a debate in parliament, the conversational structure of a political interview, the overall organization of a party program or the layout of a political advertisement in a magazine or on a billboard. For instance, only the Speaker, as specific participant category in the British House of Commons, may open and close parliamentary sessions and debates, distribute turns and decide when interruptions or questions will be allowed, among many other things. Thus the rules and structures of parliamentary interaction and their participants are closely related to the structures of the discursive structures of the debate being engaged in by the MPs.

Relevant for our discussion in this case is that it is especially the political ideology of the participants that not only controls much of what they say themselves, but also how they will understand other speakers. Thus, a call to limit immigration by an extremist right wing party member will typically be heard and commented upon as ‘racist’, whereas similar proposals by left wing MPs of our own party will obviously seldom be interpreted as such.


Political discourse and political ideology


If political ideologies are relevant properties of political situations, namely as being shared by participants, then how are they expressed and reproduced by the structures of text and talk? 

A first question we need to deal with is whether all properties of political discourse are influenced by ‘underlying’ ideologies? The response to that question is: obviously not, because only those properties of discourse can be influenced by ideologies that can be contextually variable in the first place. Thus, choice of more or less polite pronouns is contextually variable, whereas much of syntactic structure, such as the position of articles in front of nouns in English is not. People of different ideologies do not have different grammars, although they use such grammars sometimes a bit differently. Sociocultural knowledge, including language, defines communities and not ideological groups. In that respect, the left or the right, socialists or neoliberals, racists or antiracists, will not speak or write very differently. This suggests that ideological differences should rather be sought in what people say, rather than in how they say it. Political ideas may be persuasively defended by the Right or the Left, so ideologically differences will hardly be defined only in terms of rhetoric. Thus, although there are probably political uses of discourse forms such as the use of pronouns as ingroup and outgroup markers, or rhetorical means of persuasion, it is likely that most ideological variation will be found at the levels of meaning.

In order to avoid a rather arbitrary discovery procedure of the potentially huge amount of ideologically variable structures of text and talk, it is more useful to proceed in a more systematic and theory-driven way. Thus, we have seen that ideologies often have a polarized structure, reflecting competing or conflicting group membership and categorization in ingroups and outgroups. These underlying structures also appear in more specific political attitudes – for instance racist attitudes about immigration – and ultimately in the biased personal mental models of group members. These mental models control the ‘contents’ of discourse, and if they are polarized, it is likely that discourse will thus also show various types of polarization. Thus, much research has shown that ideological discourse often features the following overall strategies of what might be called the ‘ideological square’:


·                 Emphasize Our good things

·                 Emphasize Their bad things

·                 De-emphasize Our bad things

·                 De-emphasize Their good things.


These overall strategies may be applied to all levels of action, meaning and form of text and talk. Thus, political speeches, interviews, programs or propaganda typically focus on the preferred topics of ‘our’ group or party, on what we have done well, and associates political opponents with negative topics, such as war, violence, drugs, lack of freedom, and so on. Thus, many politicians and media associate immigrants or minorities with problems or delinquency. For decades communism was associated with aggression, lack of freedom and rigid ideology. Similarly, if communism is good or better than ‘us’ in the area of social services, health care or education, anticommunist discourse will typically ignore or downplay such ‘good things’ of its opponent.

What is true for meanings or topics also hold for form or structure: we may enhance meanings in many ways by intonation or stress, visual or graphical means, word order, headlining, topicalization, repetition, and so on. The opposite will occur when we want to downplay our bad things. Very bad things of our arch enemies – like a terrorist attack-- will thus appear on the front page, in a big article with big negative headlines, or in an emergency debate in parliament, and so on.

In other words, there are systematic means to examine discourse at various levels when looking for ways ideologies are (not) expressed or enacted in such discourse: Whenever a meaning is associated with good things it will tend to be associated with the ingroup of the speaker, and all structural properties of the discourse may be brought to bear to emphasize such meanings. And the opposite will be the case for Others, Opponents or Enemies.

Besides the general, combined strategy of positive self-presentation and negative other-presentation, ideological discourse structures may appear as expressions of other underlying ideology structures, and not only as an expression of the polarized relationship between (opposed) ideological groups. Thus, if identity, characteristic actions, aims, norms, values, group relations and resources are fundamental ideological categories, we may expect that references to the ‘contents’ of such categories will be prominent in the discourses of ideological group members. Thus, if people talk as group members, in terms of ‘we’, and positively evaluate their own actions, norms and values, and defend the resources or other interests of their group, then such talk will also usually be ideological. Again, such will typically be true for the meaning or context of discourse, but the ways such meanings are expressed and especially persuasively conveyed may of course also involve many formal aspects of grammar, discourse and conversation.

These are the general strategies of ideological discourse production and also a handy discovery or recognition procedure for ideological analysis of political discourse. The more detailed and subtle ideological discourse structures will be examined in a concrete example.




By way of examples I shall use some fragments from a debate in the British House of Commons on asylum seekers, held on March 5, 1997. Mrs. Gorman, representative of Billericay for the Conservative Party, then still in power, had taken the initiative for this debate, which she opens with a critique of the alleged costs of asylum seekers, costs she claims are being paid by poor old English “ratepayers”. Among those who oppose her is Jeremy Corbyn, of the Labour Party.

In order to enhance the usefulness of our analysis, we shall assign an analytical category to each example, and order the categories alphabetically. After the category name I shall add the domain of discourse analysis to which the category belongs (e.g., meaning, argumentation, etc.). The main point of the analysis is to show how various ideologies, especially those of racism and antiracism, are expressed in various kinds of structures. There are in principle hundreds of such categories, so we make a small selection (for details, see a more detailed ideological analysis in Van Dijk, 2000; no further references are given to the many hundreds of studies that deal with the respective analytical categories mentioned above; see Van Dijk, 1997, for general introduction to many of these notions).




ACTOR DESCRIPTION (MEANING). The way actors are described in discourses also depends on our ideologies. Typically we tend to describe ingroup members in a neutral or positive way, and outgroup members in a negative way. Similarly, we will mitigate negative descriptions of members of our own group, and emphasize the attributed negative characteristics of Others. Here is how Mrs. Gorman describes a Romanian asylum seeker:


(1) In one case, a man from Romania, who came over here on a coach tour for a football match (…) decided that he did not want to go back, declared himself an asylum seeker and is still here four years later. He has never done a stroke of work in his life (Gorman).


AUTHORITY (ARGUMENTATION). Many speakers in an argument, also in parliament, have recourse to the fallacy of mentioning authorities to support their case, usually organizations or people who are above the fray of party politics, or who are generally recognized experts or moral leaders. International organizations (such as the United Nations, or Amnesty), scholars, the media, the church or the courts often have that role. People of different ideologies typically cite different authorities. Thus, Mr. Corbyn ironically asks Mrs. Gorman whether she has not read the reports of Amnesty or Helsinki Watch.


BURDEN (TOPOS). Argumentation against immigration is often based on various standard arguments, or topoi, which represent premises that are taken for granted, as self-evident and as sufficient reasons to accept the conclusion. One of the topoi of anti-immigration discourse is that asylum seekers are a financial ‘burden’ for ‘us’:


(2) It is wrong that ratepayers in the London area should bear an undue proportion of the burden of expenditure that those people are causing (Gorman).


CATEGORIZATION (MEANING). As we also know from social psychology, people tend to categorize people, and so do speakers in parliament, especially when Others (immigrants, refugees, etc.) are involved.  Most typical in this debate is the (sub)categorization of asylum seekers into "genuine" political refugees, and "bogus" asylum seekers, a categorization formulated in the following ways:


(3) There are, of course, asylum seekers and asylum seekers (Gorman).


(4) ... those people, many of whom could reasonably be called economic migrants and some of whom are just benefit seekers on holiday, to remain in Britain (Gorman)



COMPARISON (MEANING, ARGUMENTATION). Different from rhetorical similes, comparisons as intended here typically occur in talk about refugees or minorities, namely when speakers compare ingroups and outgroups. In racist talk, outgroups are compared negatively, and ingroups positively. In antiracist talk we may negatively compare our country or government with loathsome undemocratic regimes. In the following example, Mr. Corbyn uses an argumentative comparison with the Second World War to emphasize the plight of asylum seekers:


(5) Many soldiers who were tortured during the second world war found it difficult to talk about their experiences for years. That is no different from the position of people who have been tortured in Iran, Iraq, West Africa or anywhere else. (Corbyn).


CONSENSUS (POLITICAL STRATEGY). To claim or insist on cross-party or national consensus is a well-known political strategy in situations where the country is threatened, for instance by outside attack. Immigration is often seen as such a threat. Thus, Mrs. Gorman insists that the current immigration law is the fruit of consensus, and hence should not be tampered with:


(6) The Government, with cross-party backing, decided to do something about the matter (Gorman, C).


COUNTERFACTUALS (MEANING, ARGUMENTATION). (See also Counterfactuals) “What would happen, if…”, the typical expression of a counterfactual, is often used in this debate by the Labour opposition in order to suggest that the conservatives try to imagine what it would be like to be in the situation of asylum seekers, an persuasive argumentative move that is also is related to the move of asking for empathy:


(7) I suggest that he start to think more seriously about human rights issues. Suppose he had to flee this country because an oppressive regime had taken over. Where would he go? Presumably he would not want help from anyone else, because he does not believe that help should be given to anyone else (Corbyn).


(8) If that happened in another country under a regime of which we disapproved, the British Government would say that it was a terrible indictment on the human rights record of that regime that prisoners were forced to undertake a hunger strike to draw attention to their situation (Corbyn).



DISCLAIMERS (MEANING). A well-known combination of the ideologically based strategy of positive self-presentation and negative other-presentation, are the many types of disclaimers. Note that disclaimers in these debates are not usually an expression of attitudinal ambiguity, in which both positive and negative aspects of immigration are mentioned, or in which humanitarian values are endorsed on the one hand, but the "burden" of refugees is beyond our means. Rather, disclaimers briefly save face by mentioning Our positive characteristics, but then focus rather exclusively, on Their negative attributes. Hence our qualification of the positive part of the disclaimer as 'Apparent', as in Apparent Denials, Concessions, Empathy, etc.:


(9) [Apparent Empathy] I understand that many people want to come to Britain to work, but there is a procedure whereby people can legitimately become part of our community (Gorman).


(10) [Apparent Denial] I did not say that every eastern European's application for asylum in this country was bogus. However... (Gorman).



EUPHEMISM (RHETORIC; MEANING). (See Taboo, euphemism and political correctness). The well-known rhetorical figure of euphemism, a semantic move of mitigation, plays an important role in talk about immigrants. Within the broader framework of the strategy of positive self-presentation, and especially its correlate, the avoidance of negative impression formation, negative opinions about immigrants are often mitigated, especially in foreign talk. The same is true for the negative acts of the own group. Thus, racism or discrimination will typically be mitigated as "resentment" or "unequal treatment", respectively. Similarly Ms. Gorman in this debate uses the word "discourage" ("to discourage the growing number of people from abroad...") in order to refer to the harsh immigration policies of the government, and thus mitigates the actions of the conservative government she supports. Similarly, the Labour (Corbyn) opposition finds the condemnation of oppressive regimes by the Government "very muted" instead of using more critical terms. Obviously, such mitigation of the use of euphemisms may be explained both in ideological terms (ingroup protection), as well as in contextual terms, e.g., as part of politeness conditions or other interactional rules that are typical for parliamentary debates.


EVIDENTIALITY (MEANING, ARGUMENTATION). (See also Evidentiality in grammar). Claims or points of view in argument are more plausible when speakers present some evidence or proof for their knowledge or opinions. This may happen by references to AUTHORITY figures or institutions (see above), or by various forms of Evidentiality: How or where did they get the information. Thus people may have read something in the paper, heard it from reliable spokespersons, or have seen something with their own eyes. Especially in debates on immigration, in which negative beliefs about immigrants may be heard as biased, evidentials are an important move to convey objectivity, reliability and hence credibility. In stories that are intended to provoke empathy, of course such evidence must be supplied by the victims themselves. When sources are actually being quoted, evidentiality is linked to INTERTEXTUALITY. Here are a two examples:


(11) This morning, I was reading a letter from a constituent of mine (..) (Gorman).


(12) The people who I met told me, chapter and verse, of how they had been treated by the regime in Iran (Corbyn).


EXAMPLE/ILLUSTRATION (ARGUMENTATION). A powerful move in argumentation is to give concrete examples, often in the form of a vignette or short story, illustrating or making more plausible a general point defended by the speaker. Concrete stories are usually better memorized than abstract arguments, and have more emotional impact, so they are argumentatively more persuasive. Of course, the right and the left each will have its own stories to tell:


(13) The Daily Mail today reports the case of a woman from Russia who has managed to stay in Britain for five years. According to the magistrates court yesterday, she has cost the British taxpayer £40,000. She was arrested, of course, for stealing (Gorman).


(14) The people who I met told me, chapter and verse, of how they had been treated by the regime in Iran--of how they had been summarily imprisoned, with no access to the courts; of how their families had been beaten up and abused while in prison; and of how the regime murdered one man's fiancée in front of him because he would not talk about the secret activities that he was supposed to be involved in (Corbyn).


GENERALIZATION (MEANING, ARGUMENTATION). Instead of providing concrete stories, speakers may also make generalizations, in racist discourse typically used to formulate prejudices about generalized negative characteristics of immigrants. Similarly, in a populist strategy, conservative speakers may generalize the negative feelings against asylum seekers:


 (15) Such things go on and they get up the noses of all constituents (Gorman).



HYPERBOLE (RHETORIC).  Hyperboles are semantic rhetorical devices for the enhancement of meaning. Within the overall strategy of positive self-presentation and negative other-presentation, we may thus expect in parliamentary debates about immigrants that the alleged bad actions or properties of the Others are expressed in hyperbolic terms (our bad actions in mitigated terms), and vice versa. Sometimes such forms of hyperbole are implied by the use of special METAPHORS, as we observe in Mrs. Gorman's use of "opening the floodgates" in order to refer to the arrival of many asylum seekers. And conversely, on the left, Labour speakers will of course emphasize the bad nature of authoritarian regimes, and like Mr. Corbyn, will call them "deeply oppressive", and the conditions of refugees coming from those countries "appalling".


IMPLICATION (MEANING). For many 'pragmatic' (contextual) reasons, speakers do not (need) to say everything they know or believe. Indeed, large part of discourse remains implicit, and such implicit information may be inferred by recipients from shared knowledge or attitudes and thus constructed as part of their mental models of the event or action represented in the discourse. In debates about immigration, implicitness may especially be used as a means to convey meanings whose explicit expression could be interpreted as biased or racist. Thus, when Ms. Gorman says that many refugees come from countries in Eastern Europe who have recently been "liberated", she is implying that people from such countries cannot be genuine asylum seekers because democratic countries do not oppress their citizens (a point later attacked by the Labour opposition). And the same is true when she describes these refugees as "able-bodied males", which implies that these need no help from us.


IRONY (RHETORIC). (See Irony). Accusations may come across as more effective when they are not made point blank (which may violate face constraints), but in apparently lighter forms of irony. There is much irony in the mutual critique and attacks of Conservatives and Labour, of course, and these characterize the proper interactional dimension of the debate. However, when speaking about immigrants, irony may also serve to derogate asylum seekers, as is the case for the phrase "suddenly discover" in the following example, implying that such a "sudden discovery" can only be bogus, since the asylum seekers allegedly knew all along that they came to the country to stay:


(16) Too many asylum seekers enter the country initially as family visitors, tourists, students and business people, and then suddenly discover that they want to remain as asylum seekers (Shaw).



LEXICALIZATION (STYLE). At the local level of analysis, debates on asylum seekers need to express underlying concepts and beliefs in specific lexical items. Similar meanings may thus be variably expressed in different words, depending on the position, role, goals, point of view or opinion of the speaker, that is, as a function of context features. In conservative discourse opposing liberal immigration policies, this will typically result in more or less blatantly negative expressions denoting refugees and their actions, thus implementing at the level of lexicalization the overall ideological strategy of negative other-presentation. Thus, also in this debate, we may typically find such as expressions as "economic immigrants", "bogus asylum seekers", or "benefit scroungers", as we also know them from the tabloid press in the UK. On the other hand, lexicalization in support of refugees may focus on the negative presentation of totalitarian regimes and their acts, such as "oppression", "crush", "torture", "abuse" or "injustice".


METAPHOR (MEANING, RHETORIC). (See the other chapters on metaphor). Few semantic-rhetorical figures are as persuasive as metaphors, also in debates on immigration. Abstract, complex, unfamiliar, new or emotional meanings may thus be made more familiar and more concrete. Virtually a standard metaphor (if not a topos) is the use of flood-metaphors to refer to refugees and their arrival, symbolizing the unstoppable threat of immigration, in which we would all "drown". Thus, Ms. Gorman warns for changes in the present law by saying that such changes would "open the floodgates again". Another notorious semantic realm of metaphors is to describe people in terms of (aggressive, repulsive, etc.) animals, for instance asylum seekers as “parasites”, as does Mrs. Gorman.


NATIONAL SELF-GLORIFICATION (MEANING). Especially in parliamentary speeches on immigration, positive self-presentation may routinely be implemented by various forms of national self-glorification: Positive references to or praise for the own country, its principles, history and traditions. Racist ideologies may thus be combined with nationalist ideologies, as we have seen above. This kind of nationalist rhetoric is not the same in all countries. It is unabashed in the USA, quite common in France (especially on the right), and not uncommon in Germany. In the Netherlands and the UK, such self-glorification is less explicit. See, however, the following standard example -- probably even a topos:


(17) Britain has always honoured the Geneva convention, and has given sanctuary to people with a well-founded fear of persecution in the country from which they are fleeing and whose first safe country landing is in the United Kingdom (Wardle).


NEGATIVE OTHER-PRESENTATION (SEMANTIC MACRO-STRATEGY). As the previous examples have shown, the categorization of people in ingroups and outgroups, and even the division between 'good' and 'bad' outgroups, is not value-free, but imbued with ideologically based applications of norms and values. Thus, throughout this debate, Mrs. Gorman describes asylum seekers in terms of “benefit seekers” or “bogus” immigrants. Negative other-presentation is usually complimentary to positive self-presentation.


NORM EXPRESSION. Anti-racist discourse is of course strongly normative, and decries racism, discrimination, prejudice and anti-immigration policies in sometimes explicit norm-statements about what 'we' (in parliament, in the UK, in Europe, etc.) should or should not do:


(18) We should have a different attitude towards asylum seekers (Corbyn).



NUMBER GAME (RHETORIC, ARGUMENTATION). Much argument is oriented to enhancing credibility by moves that emphasize objectivity. Numbers and statistics are the primary means in our culture to persuasively display objectivity, and they routinely characterize news reports in the press. Arrivals of immigrants are usually accompanied by numbers, also in parliament. The same is true for the ‘costs’ of immigrants:


 (19) It would open the floodgates again, and presumably the £200 million a year cost that was estimated when the legislation was introduced (Gorman, C).



POLARIZATION, US-THEM CATEGORIZATION (MEANING). Few semantic strategies in debates about Others are as prevalent as the expression of polarized cognitions, and the categorical division of people in ingroup (US) and outgroup (THEM). This suggests that especially also talk and text about immigrants or refugees is strongly monitored by underlying social representations (attitudes, ideologies) of groups, rather than by models of unique events and individual people (unless these are used as illustrations to argue a general point). Polarization may also apply to 'good' and 'bad' sub-categories of outgroups, as is the case for friends and allies on the one hand, and enemies on the other. Note that polarization may be rhetorically enhanced when expressed as a clear contrast, that is, by attributing properties of US and THEM that are semantically each other's opposites. Examples in our debate abound, but we shall only give two typical examples:


(20) It is true that, in many cases, they have made careful provision for themselves in their old age, have a small additional pension as well as their old-age pension and pay all their rent and their bills and ask for nothing from the state. They are proud and happy to do so. Such people should not be exploited by people who are exploiting the system (Gorman, C).


POSITIVE SELF-PRESENTATION (SEMANTIC MACROSTRATEGY). Whether or not in combination with the derogation of outgroups, group-talk is often characterized by another overall strategy, namely that of ingroup favoritism or "positive self-presentation". This may take a more individual form of face-keeping or impression management, as we know them from familiar disclaimers ("I am not a racist, but..."), or a more collective form in which the speaker emphasizes the positive characteristics of the own group, such as the own party, or the own country. In the context of debates on immigration, such positive self-presentation will often manifest itself as an emphasis of own tolerance, hospitality, lack of bias, EMPATHY, support of human rights, or compliance with the law or international agreements. Positive self-presentation is essentially ideological, because they are based on the positive self-schema that defines the ideology of a group. Here is an example:


(21) I entirely support the policy of the Government to help genuine asylum seekers, but...(Gorman, C).


POPULISM (POLITICAL STRATEGY). One of the dominant overall strategies of conservative talk on immigration is that of populism. There are several variants and component moves of that strategy. The basic strategy is to claim (for instance against the Labour opposition) that "the people" (or "everybody") does not support further immigration, which is also a well-known argumentation fallacy. More specifically in this debate, the populism-strategy is combined with the topos of financial burden: Ordinary people (taxpayers) have to pay for refugees. Of the many instances of this strategy, we only cite the following:


(22) It is wrong that ratepayers in the London area should bear an undue proportion of the burden of expenditure that those people are causing (Gorman)



PRESUPPOSITION (MEANING). (See Presuppositions – Pragmatics). Discourses are like the proverbial icebergs: most of their meanings are not explicitly expressed by presupposed to be known, and inferrable from general sociocultural knowledge. Strategically, presuppositions are often used to assume the truth of some proposition when such truth is not established at all:


(23) I wonder whether the hon. Gentleman will tell the House what mandate he has from the British people to share their citizenship with foreigners? (Gill).



VAGUENESS (MEANING). Virtually in all contexts speakers may use 'vague' expressions, that is, expressions that do not have well-defined referents, or which refer to fuzzy sets. Vague quantifiers ('few', 'a lot'), adverbs ('very') nouns ('thing') and adjectives ('low', 'high'), among other expressions may be typical in such discourse. Given the normative constraints on biased speech, and the relevance of quantification in immigration debates, we may in particular expect various forms of Vagueness, as is the case for "Goodness knows how much", and "widespread" in the following examples:


(24) Goodness knows how much it costs for the legal aid that those people invoke to keep challenging the decision that they are not bona fide asylum seekers (Gorman, C).



VICTIMIZATION (MEANING). Together with DRAMATIZATION and POLARIZATION, discourse on immigration and ethnic relations is largely organized by the binary US-THEM pair of ingroups and outgroups. Thus, in order to emphasize the ‘bad’ nature of immigrants, people may tell horrible stories about poor nationals, such as the kids….


(25) Many of those people live in old-style housing association Peabody flats. They are on modest incomes. Many of them are elderly, managing on their state pension and perhaps also a little pension from their work. They pay their full rent and for all their own expenses. Now they are going to be asked to pay £35 to able-bodied males who have come over here on a prolonged holiday and now claim that the British taxpayer should support them.


The categories and examples shown above are not limited to racist or antiracist social ideologies, or to socialist or conservative political ideologies. Virtually all categories also apply to macho and feminist or pacifist or militarist ideologies and their discourses. That is they are rather general resources groups and their members acquire and use in order to account for and defend their ideas and social practices. Indeed, we need not learn totally new ways of ideological talk and text as soon as we become member of or identify with another social or political group.



There is a close relationship between discourse, ideology and politics, in the sense that politics is usually discursive as well as ideological, and ideologies are largely reproduced by text and talk. Traditionally, ideologies are vaguely and negatively defined in terms of “false consciousness”. In a more contemporary, multidisciplinary approach, ideologies are described in terms of the axiomatic foundation of the social representations shared by groups. Such general ideologies form the basis of more specific group attitudes, which in turn may influence group members’ individual opinions, constructions or interpretations of specific events, as well as the social practices and discourses in which group members engage. In politics, ideologies specifically play a role to define political systems, organizations, movements, political practices and political cognition, all enacted or reproduced by political discourse. Underlying political ideologies are typically expressed in political discourse by emphasizing Our good things and Their bad things, and by de-emphasizing Our bad things and their good things. Such a general strategy may be implemented at all levels of discourse. Thus, in examples from a debate on asylum seekers in British parliament we see that there are many ways ideologies may be expressed, for instance in the actor descriptions, fallacies, disclaimers, metaphors, comparisons, euphemisms, hyperboles, and so on.







Abercrombie, N., Hill, S., & Turner, B. S. (1980). The dominant ideology thesis. London Boston: G. Allen & Unwin.


Ball, T., & Dagger, R. (1999). Political ideologies and the democratic ideal. New York: Longman.


Billig, M. (1982). Ideology and Social Psychology. Oxford: Basil Blackwell.


Billig, M. (1988). Ideological dilemmas: A social psychology of everyday thinking. London Newbury Park: Sage Publications.


Billig, M. (1991). Ideology and opinions: Studies in rhetorical psychology. London Newbury Park, Calif.: Sage Publications.


Bourdieu, P., & Eagleton, T., (1994). Doxa and Common Life: An Interview. In S. Zizek, Mapping Ideology. (pp. 265-277). London: Verso.


Chilton, P. A. (1995). Security metaphors. Cold war discourse from containment to common house. New York: Lang.


Chilton, P. A. (2004) Analysing Political Discourse: Theory and Practice, London, Routledge.


Chilton, P. A. , & Schäffner, C. (Eds.). (2002) Politics as Text and Talk: Analytic Approaches to Political Discourse, John Benjamins, Amsterdam.


Duranti, A., & Goodwin, C., (Eds.). (1992). Rethinking Context: Language as an Interactive Phenomenon Cambridge: Cambridge University Press.


Eagleton, T. (1991). Ideology. An introduction. London: Verso Eds.


Eatwell, R. (Ed.). (1999). Contemporary political ideologies. New York: Pinter.


Fairclough, N. (1989). Language and power. London New York: Longman.


Fairclough, N. (1995). Critical discourse analysis. The critical study of language. London: Longman.


Fowler, R. (1991). Language in the news: Discourse and ideology in the British press. London New York: Routledge.


Fowler, R., Hodge, B., Kress, G., & Trew, T. (1979). Language and control. London: Routledge & Kegan Paul.


Freeden, M. (1996). Ideologies and political theory. A conceptual approach. Oxford: Clarendon Press.


Goodin, R. E., & Klingemann, H. D. (Eds.). (1996). A New handbook of political science. New York, NY: Oxford University Press.


Gramsci, A. (1971). Prison notebooks. New York: International Publishers.


Gumperz, J. J. (1982). Language and social identity. Cambridge (England New York: Cambridge University Press.



Hodge, B., & Kress, G. R. (1993). Language as ideology. London New York: Routledge.


Larraín, J. (1979). The concept of ideology. London: Hutchinson.


Leach, R. (2002). Political ideology in Britain. New York: Palgrave.


Pêcheux, M. (1982). Language, semantics, and ideology. New York: St. Martin's Press.


Potter, J. (1996). Representing reality: Discourse, rhetoric and social construction. London Thousand Oaks, CA: Sage.


Seliger, M. (1976). Ideology and politics. London: Allen & Unwin.


Van Dijk, T. A. (1998). Ideology: A multidisciplinary approach. London: Sage.


 Van Dijk, T. A. (1999). Context models in discourse processing. In: van Oostendorp, Herre, &  Goldman, Susan R. (Eds.), The construction of mental representations during reading. (pp. 123-148). Mahwah, NJ, USA: Lawrence Erlbaum Associates.


Van Dijk, T. A. (2001). Discourse, ideology and Context. Folia Linguistica XXXV/1-2, 11-40.


Van Dijk, T. A. (2003). Text and context of parliamentary debates. In Paul Bayley (Ed.),  Cross-Cultural Perspectives on Parliamentary Discourse. (pp. 339-372). Amsterdam: Benjamins.


Wilson, J. (1990). Politically speaking. Oxford: Blackwell.


Wodak, R. (Ed.). (1989). Language, power, and ideology studies in political discourse. Amsterdam Philadelphia: J. Benjamins Co.


Wodak, R., & Menz, F. (Eds.). (1990). Sprache in der Politik - Politik in der Sprache. Analysen zum öffentlichen Sprachgebrauch. (Language in politics - politics in language. Analyses of public language use). Klagenfurt: Drava.


Wodak, R., & Meyer, M. (Eds.). (2001). Methods of critical discourse analysis. London: Sage.


Zizek, S. (1994). Mapping ideology. London: Verso.





Discourse, political discourse, ideology, categories of description, parliamentary debates.



Context, communicative

Context, sociocultural

Critical Discourse Analysis

Parliamentary Discourses




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 takes a more "critical" perspective and deals 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 edited The Handbook of Discourse Analysis (4 vols, 1985) as well as the introduction Discourse Studies (2 vols., 1997). He founded 4 international journals, Poetics, Text, Discourse & Society, and Discourse Studies, of which he still edits the latter two. His last monograph is Ideology (1998), and his last edited book (with Ruth Wodak), Racism at the Top (2000). He is currently working on a new book on the theory of context. Teun van Dijk, who holds two honorary doctorates, has lectured widely in many countries, especially also in Latin America. For a list of publications, recent articles, resources for discourse studies and other information, see his homepage: www.discourse-in-society.org.