прагматика и медиа дискурс / Teun A van Dijk - Ideology
.pdfContext |
223 |
positive self-presentation (face-keeping) or negative other-presentation, argumenta (against opponents, or in support of allies, etc.) and favourable or unfavourable rhetoric.
Note that even within this category there are levels or layers of roles. In everyday conversations or parliamentary debates opponents may be 'direct' and confront each other face to face. In a debate on the op-ed page, opponents may also be confronting each other personally, but not face-to- face and not at the same time.' But, as is the case for various communicative roles, there may be indirect, long-term addressees or relations. Opposing a speaker may stand for opposing her or bis boss or organization, and speaking as a member of an action group may be interpreted as advocating the stance of the action group itself. Tbus, in parliament, speakers may oppose what the previous speaker has said, they may, more broadly, oppose the bill being proposed by someone, and by so doing they may oppose the party to which that person belongs, and at the same time they of course advocate their own position, and/or that of their own party (which need not be identical), and as political representatives they may at the same time represent or oppose the 'special interests' outside of parliament. In other words, deeper and more sophisticated analyses of contexts in principie uncover complex sets or levels of various roles.
We have seen that besides membership of groups and organizations, ideologies typically involve polarization, struggle, conflict or competition, and these relationships precisely map onto the social roles being introduced into the context here. Ingroups and outgroups and their associated ideologies thus manifest and reproduce themselves precisely by the 'position' their members take in situations of debate and conflict, also in communication. Arguing in favour of a bill that restricts immigration, may by its very stance be part of the reproduction of nationalism or ethnocentrism. An anticommunist speech in parliament is thus taking a stance in an ideological conflict. In other words, social roles are contextually variable enactments of positions, including ideological positions.
Affiliation
Participants in professional roles often doñ t speak 'for themselves, but as representatives of an organization or institution, and as representatives who in principie can be replaced by any other institutional member. That is, their affiliation plays a prominent role in the context: confessing to a police officer or in court, doing an exam, making a declaration for a tax auditor, and so on, are the kind of speech events that are often appropriately accomplished in the presence of any representative (in the same professional role) of the organization. People speak in parliament or congress, but usually do so as representatives of their partes, as they do when listening to such speeches. More generally, these events and their participants are also integrated in a web of institutional affiliations. Some of these may be very
224 |
Discourse |
strict, and legally well described (also with respect to kinds of communicative events), while others may be looser and open to variation and negotiation. Thus, teachers will usually have more leeway in the accomplishment of their communicative and professional roles than speakers of parliament orjudges.
One of the many implications of the institutional or organizational affiliation of communicative events is precisely the fact that such participants take part as representativas of the institution, and hence often carry the institutional ideologies, if any, into the ongoing context. Indeed, the representatives of an organization are by habit, norm or law entitled or obliged to represent the Interests' of the organization, and hence their talk and text will multiply index or signal such ideological comtnitments. Thus, a teacher may thus implement the educational ideology of the school or university, the joumalist an ideology of the presa, and so on. Such ideologies may pertain to the content of text or talk (such as newsworthiness of events for news interviews or news reports), but also to the very nature of the interaction itself. Educational or medical ideologies may or may not allow a more or less independent and autonomous initiative to students or patients, depending on whether the ideology is more authoritarian or permissive.
Membership
More generally, participants may speak, write, listen or read (also) as membeis of groups or in addition to the organizational affiliation and the various roles described aboye. People may be male or female, white or black, old or young, and so on, and either they themselves or their co-participants will categorize them as such, and act (speak, write) accordingly. S ince such social groups and categories are the basis of ideologies, these ideologies will in principie also exhibit in the relevant communicative social practices in which group members engage. That group membership affects the structures of text and talk themselves has been shown in much socio-linguistic research, for example on intonation, lexical items, topics, rhetoric or interactional moves, as discussed in the previous chapter. In terms of the context, people of different social groups or categories are defined and treated as such, also in the communicative event
— they may be given preference in tutti taking, freedom in topic selection or style, but they may also directly be discriminated against along the same fines, only because they are a member of a specific group. Probably more than any other category of the context, thus, social group membership is what projects ideologies into communicative events. Later we will see how this is being accomplished in text and talk. 8
It should at this point be emphasized again that roles, affiliation and group membership are not always 'given' in social situations, and this is a fortiori the case in subjectively construed models of such social situations. That is, such social 'positions' may be negotiated, changed, oriented to, deviated from, ígnored, forgotten or otherwise become less (or more) relevant in a
Context |
225 |
specific situation. That is, a dynamic theory of discourse emphasizes such situational and personal fiexibility. The same will be true for the ideological conditions and consequences of the ways such categories are constructed in the current context by the participants. Men may temporarily disassocíate from their group and speak on behalf of women; speakers may defend the position of their opponents when they act as devil' s advocate; and dissidents are by definition speaking in defiance of dominant group ideologies.
The social others
So far, the relevant participant roles discussed aboye pertain to people involved in various capacities in the communicative context itself. However, text and talk are often also about other people, usually people who are not present in the' ongoing context at all. Strictly speaking this is a property of the nteaning of discourse, and hence part of a semantic and not of a (pragmatic) context analysis. That is, discourse referents are .not part of the context model, but part of the event model (partly) expressed by the discourse. Thus, men routinely speak about women, whites about blacks, and doctors about patients, and these social others are thus the referents of their talk. It is also in this way that the ideologies relating communicative participants to the social others, as members of outgroups, are projected into the meanings of a discourse. Yet, one might also argue that these social others are some kind of 'absent participants' in the context.
Racist talk addressed to other whites may obliquely be addressed, in a broader social context, to the social others, and thus not only be semantically relevant, but also pragmatically, that is, as an inherent element of the act of discriminatory talk, as a form of reproduction of racist ideologies. That is, the social others, as part of the targeted outgroup, may be talked about but at the same time indirectly, socially and ideologically addressed. That is, acts of discrimination also may be categorized and interpreted as such when the discriminated party is not present — yet, they are somehow 'party' to such communicative interaction.
Social representations
Most context categories discussed aboye have a proper social nature, and are typically made explicit in sociological tercos. Obviously, however, participants not only have positions, rights, duties and relations in social situations, institutions and overall social structures, but also share social representations, such as knowledge, attitudes and ideologies. Some of these mental dimensions of participants have been discussed in terms of intentions and goals, which are more individual and contextual.
However, especially as members of various social groups and institutions, communicative participants also share social representations that have an impact on ongoing interaction, text and talk. Thus, crucial for all communicative events is the respective knowledge of the participants, both personal as well as social and cultural. Thus, speakers have knowledge about each
226 |
Discourse |
other (that is, they have a model about Self as well as about others), and such knowledge may instantiate more general knowledge and beliefs about the group to which the others belong.
Similarly, ideologies of participants in many ways affect the ongoing definition of the communicative situation, the various actions, participant roles, as well as the discourse itself. The same is true for the socially shared attitudes monitored by these ideologies. Indeed, these attitudes may even be specific and tailored to the communicative event at hand. Thus, trivially, in an informal discussion about abortion, or a parliamentary debate about nuclear energy, speakers bring to bear their specific actitudes about these issues, and such attitudes multiply influence the event and context models that monitor ongoing talk: who is defined as proponent or opponent, whether a speaker is seen as a representative of a social group (man or woman in the abortion debate), who .will be treated more or less politely, and so on.
In sum, all social aspects of the complex communicative event are variously monitored by the social representations of the participants as members of groups, categories or institutions. Knowledge will be mutually presupposed accordingly, for example when doctors or lawyers speak with members of the same professional group, or when women presuppose both knowledge, attitudes and ideologies of other women of the same feminist movement. Indeed, most of the communicative context and the discourse need not be made explicit because of such presupposed sharing of social representations within the same group, society or culture.
Together with mental models of individuals, social representations are part of the cognitive interface between social structure, group membership and discourse. If people speak or write as members of groups, their group membership will largely be brought to bear in the current context in tercos of the social representations shared with the group, that is, as instantiations of group knowledge, attitudes and ideologies.
This does not mean, incidentally, that social representations, including ideologies, cause or determine text and context. It has been explained in some detall in Part I, that there is still a vast 'mental distance' between social representations, and hence the influence of social groups, on the one hand, and discourse structures (including context) on the other hand. Most crucially, although variably so in different situations, speakers are also individuals with their own biography, goals, preferences, plans and emotions
— that is, with their own personal models. Intentionally or unwittingly, such models may instantiate shared elements of social representations, but even then the context and the individual and hence their text or talk will be unique. If not, and as suggested before, all members of a group would say or write the same thing in the same situation. This is also one of the reasons why I include relevant aspects of personal models (e.g. intentions and purposes) in the current context.
Social representations may not only apply to the semantic dimension of discourse (e.g. abortion as a topic of talk), but also to the discursive interaction itself: who may/must speak/write about what/whom, to whom, in
Context |
227 |
which way? Joumalists know how to interview news sources or news actors, how to write news reports and follow rules and strategies they have leamed as group members, and the same is true in afl other professional roles discussed aboye. Thus, both in conversation and in parliament, people instantiate the very ideological forms of membership that we routinely assign to speakers: he is a conservative, she is a liberal, and so on.
At this point we have come full circle. Ideologies may indirectly control the properties of all categories of context models for discourse. But it now appears that one of these categories itself pertains to the social beliefs, and hence the ideologies, of the participants. In other words, ideological control is, so to speak, not external nor deterministic, but internal, that is, through the beliefs of the participants themselves. Thus, I may participate in a conversation as an anti-racist, and this stance influences the way I construct the current context as well as what I say and how I say it. At the same time, both the recipient and I myself represent (part of) my anti-racist beliefs as part of our respective context models, of ourselves as well as about each other (indeed, I may know that my interlocutor knows that I am an antiracist, and may shape my talk accordingly).
There may even be a discrepancy between my role and my role as represented in my model of myself in the present context. People may speak as anti-racists without much self-control or self-monitoring and thus more or less directly express and enact their group membership. However, they may also do so by monitoring their current identity and by carefully managing their'imagé as an anti-racist, for instance for recipiente that are hostile to anti-racists. Also this subtle interplay between 'real' social identities of discourse participants, on the one hand, and those that are locally and intersubjectively represented in their current context models and displayed in their discourse, on the other hand, shows how complex the relations between ideology and discourse may be.
Concluding remark
The context analysis presented aboye shows that the discursive reproduction of ideologies also applies to the contextual aspects of communicative events. Contexts, or rather context models, explain personal, situational and social variations in the ways underlying ideologies may or may not affect text and talk. They thus serve as another layer of constraints, another interface, between ideology and discourse, and explain that ideologies are not 'deterministic' in the sense of necessarily affecting discourse structures — this will always, literally, depend on the context. Therefore, no discursive theory of ideological expression and reproduction can be adequate without a detailed analysis of context. We shall later spell out in somewhat more detail how exactly mental models of such contexts intervene between social representations, including ideologies, and structures of discourse.
23 Reproduction
What is reproduction?
It has often been argued, aboye, that ideologies are typically reproduced by social practices, and especially by discourse. What exactly does this mean? As with most general notions, the concept of reproduction is not very precise. In general, it implies that ideologies are 'continued', 'made to remain, last, persist.. .', and so on. Like its second part, however, it implies an active, human dimension: It is what people do, make happen, while also making something new, creating something. The repetitive'ré part implies that the act of production is being repeated. For social practices and discourse this usually implies that such acts of production take place every day, are routine, and are part of the definition of everyday life.
More specifically, however, when we refer to the reproduction of ideologies, we are dealing with an equally vague sociological notion, also used to denote the reproduction of groups, social structures, or even whole cultures. Again, reproduction here implies continuity of a system or structure as well as human agency. More theoretically, the notion is used to bridge the wellknown gap between the macro-level and the micro-level of social structure. Systems or abstract structures, such as ideologies, natural languages, and societal arrangements are thus said to be both manifested in, as well as made to persist as such through, social practices of social actors at the micro-level. A language Like English is reproduced, daily and by millions of people, by its everyday use. And so are capitalist, sexist or racist ideologies.'
The active concept of 'production' is relevant here because such systems are not only being 'applied','implemented' or passively 'used', but at the same time constituted and reconstituted, as well as gradually changed, by such contextual uses by many social actors. Indeed, also the gradual development of ideologies of a group is based on such social practices. That is, ideologies are (re)produced as well as (re)constructed by social practices.
There is another macro—micro dimension involved here. This time not just that of an abstract system on the one hand and actual practices on the other hand, but the distinction between the group and its members, and especially its new members. Just as groups are reproduced (also) by getting or recruiting new members, also ideologies are reproduced by getting new 'users,' as is also the case for natural languages. Whether by socialization or other processes of sharing social representations (initiation, teaching, train-
Reproduction |
229 |
ing, preaching, propaganda), ideologies are continually reproduced because new social members 'acquire' or learn to use' them.
As we shall see in more detall later, this may happen directly through explicit ideological discourse, or indirectly by making inferences from discourse and other social practices about what opinions other group members share. White people learn racism' by accepting general racist statements such as 131ack women are welfare queens' as expressed in conversations with friends or colleagues, or they infer such a belief from repeated stories in the media in which black women are portrayed as being on welfare, or because they overgeneralize from one or a few black women they know who are on welfare. This last case, as a personal experience, however, is usually told in stories to other group members, and the relevant inference may then be jointly produced in talk, as a conclusion suggested or accepted by co-participants. That is, sharing is usually not simply a onesided, passive event, but a complex, co-operative procedure, involving people who (already) 'know', as well as people who 'still don' t know'. In other words, reproduction also implies socialization, leaming, inculcation or adoption by young or new members, of the socially shared representations of a group.
And finally, besides its macro—micro (system—actions, group—members) dimensions, we also have the local and contextual versus the global and decontextualized dimension of reproduction. Members having learned how to make an inference from one case or example, or to express an ideological opinion in one context, are typically able to do so for similar cases and in similar contexts. That is, reproduction is not only top-down and bottom-up, but also allows for transition from token to type and from type to token, from today to tomorrow, and from here to elsewhere. Reproduction thus also implies generalization. Combined with the vertical relations between system and actions, this also explains the bottom-up nature of reproduction — social representations are not merely acquired directly, in an abstract (and usually discursive) manner, but also as generalizations from daily experiences. In specific social situations of ethnic inequality, such generalizations may be morally unacceptable overgeneralizations (prejudice), but they may also be forms of (correct, justified) social learning, for example when minorities learn to detect and interpret racist events as such, and thus acquire an antiracist ideology. 2
Summarizing these various aspects of the social reproduction of ideologies, we thus have the following dimensions.
1 System—Action: top-down application, use and implementation of general, abstract ideological beliefs in concrete social practices.
2 Action—System: bottom-up sustaining, 'continuing and changing the socially shared system by its daily uses in social practices. Along this dimension, ideologies are effectively being constructed, constituted and changed by social practices, including discourse.
230 |
Discourse |
3Group—Mernbers: ideological communication, inculcation, teaching, socialization and initiation of new members by (knowledgeable) group members.
4Members—Group: acceptance and compliance or non-acceptance, resistance or dissidence of one or some groups members, against the ideology of the group or its elites.
5Local—Global: generalization, extension, decontextualization of specific experiences and opinions to similar or abstract contexts, experiences, cases or circumstances; social leaming, overgeneralization, stereotyping, prejudice formation and ideology construction.
In 4 we see that the group—member relation may also be conversed, that is, when individual members reject, refuse or do not accept a group's ideology. This may not seem to be a dimension of reproduction, but it is necessary to account for personal variations and change of ideologies, which are also part of their reproduction. Obviously, as soon as most members reject ideologies or some ideological beliefs, then change may eventually lead to the abolition of ideologies. 3
Discourse and reproduction
Many of the types and modalities of reproduction discussed aboye appear to be discursive. Ideologies may be expressed in many genres and contexts of discourse and their respective structures as discussed in the previous chapters. Such ideological discourses have several functions, such as a display of group knowledge, membership and allegiance; comparison and normalization of values and evaluation criteria; evaluating social practices; socialization; or persuasion and manipulation. Some of these functions will be dealt with more specifically in the next chapters. Here I focus on some of the more general aspects of the discursive reproduction of ideologies.
Context
In the previous chapter we saw how ideologies may interven in the social construction or interpretations of the contextual categories which in turn constrain (or are influenced or constituted by) text and talk. Thus, participants may act as speakers, as proponents, as journalists, as representatives of an institution like a newspaper, and as members of various groups (age, gender, ethnicity, nationality, etc.). In all these roles, participants may enact (and sometimes disregard) the social representations, including the ideologies, related to their social identity. That is, social situations in general, and contexts of discourse in particular, are literally the site where ideologies are being enacted in society. As long as speech participants identify with or willingly or unwillingly (have to) represent the groups and institutions of which they are members, they thus by definition contribute to the use and the reproduction of the ideologies associated with these social formations.
Reproduction |
231 |
The examples mentioned in the previous chapter suggest, however, that such ideological alignment is not straightforward. First, language users may have their own personal models, and these may be more or less at variance with the social representations they share as group members, given the constraints of the present context. Indeed, their interests as group members may be less salient or less relevant than their current personal interests, and their intentions and goals may be formed accordingly. Second, language users are members of several social groups, and thus share in several social representations at the same time. Again, some of these may be more relevant or more powerful than others. The result is that the event and context models that monitor the communicative event may have contents and structures that in many ways are inconsistent with those expected of loyal group members.
If such is the case for models, this will also be the case for the discourse properties that are a function of these models, such as the meaning derived from event models (including specific opinions), as well as the surface structure, style, speech acts or interactional strategies that are controlled by context models.
The consequences of these complex and subtle acts of interactional and communicative management in specific social situations are that ideologies are not simply reproduced in talk and text by the members of the groups that share such ideologies. There is more or less substantial variation, there is explicit and intentional deviance, there are dilemmas, and there are personal and interpersonal conflicts that need to be negotiated and resolved. 4 Hence, not all news reports in a newspaper will show the ideology or political allegiances of that newspaper. Not all journalists always give priority to joumalistic ideologies in their reports, and not all racists will treat minorities always and everywhere with derogatory remarks.
The empirical picture emerging from this variation may be that ideologies do not seem to 'exise in the first place — the local and personal constraints of context may distort or prohibit their unfettered expression. The question is then in what respect we are able to speak of the 'reproduction' of ideologies, when social situations so often prevent their direct implementation. Theoretically, then, we are able to account for ideological reproduction only when we assume that across language users and contexts, there are 'enough' instances of ideological expression.
How much is 'enough'? Obviously, this may vary. Por instance, it may be assumed that journalists most of the time will have to follow the ideological principies of their profession. If not, they will not be hired or they will be fired. Exceptions will be allowed, especially for highly qualified or popular journalists, but there will be a margin of variation within which each journalist will have to remain when working for one of the mass media. In some cases, for example in public office, even one deviation from the ideological 'party liné may be enough for a politician to be marginalized, discredited or voted out of office.
Interestingly, quantity as such may not be the right measure. One public racist statement may be enough to conclude that someone is expressing a
232 |
Discourse |
racist ideology, even when in most other situations these expressions were better controlled. The rationale behind such a conclusion is that people who do not have a racist ideology will simply never make such a blatantly racist remark in the first place. In actual practice, there will be a broad range between regular and unique expressions of ideology, on which basis other participants and observers will be able to draw conclusions about the underlying ideologies of group members. Some of these expressions may be very indirect or subtle, and participants and observers may not even notice them if the ideology that inspires them is taken for granted. Thus, the quality press, including the liberal quality press, may not daily make blatant remarks about ethnic minorities or immigrants. Yet, more subtly and indirectly, for example by the choice of its topics (e.g. about crime, violence or cultural deviation), it may well slowly create a negative image of the cultural others, and thus contribute to the reproduction of an ethnocentric ideology.
Given the processes of memory, attention and recall, readers may selectively focus on and memorize even the occasional story in which minorities are represented negatively, and forget about the larger number of negative stories in which members of their own majority group are represented negatively. This is a familiar finding in differential attribution for ingroups and outgroups. 5
In sum, the conditions of reproduction are as complex as the structures of context and discourse, and the strategies of information processing and social representation, combined. Under what conditions specific text and talk is being attended to, read or listened to, understood, and represented in models, and under what conditions these models are accepted as true and generalized to more abstract social knowledge and beliefs, are all questions that need to be answered in a theory of reproduction.
All this also applies to the projection of ideologies in context models and hence in the enactment or interpretation of the context itself. Negative beliefs about minorities when uttered by prominent members of minority groups themselves or by a white cabinet minister of a respectable party, may be much more credible than those of a member of a racist party. That is, credibility is one element of the process of acceptability, and itself a function of the group membership of the speaker, that is, a category of the context.
Generally, thus, acceptability of beliefs, which is the core criterion in the reproduction of ideologies, depends also on the interpretation and the evaluation of context structures, and especially on the various roles and positions of the participants. Even the context categories of communicative domain, action type, and circumstances may be especially conducive to ideological reproduction, as is the case for classroorns and education, parliament and politics, newsrooms and the media. This is so first because of the credibility or the prestige of the social actors involved, as well as the mass-mediated consequences of text and talk. One 'unhappy' but widely publicized remark of a prominent politician about immigrants may contribute more to the reproduction of ethnic prejudices and ideologies than
