Добавил:
Upload Опубликованный материал нарушает ваши авторские права? Сообщите нам.
Вуз: Предмет: Файл:

прагматика и медиа дискурс / Teun A van Dijk - Communicating Racism

.pdf
Скачиваний:
148
Добавлен:
08.06.2015
Размер:
5.13 Mб
Скачать

302 Communicating Racism

Processing, thus, takes place "on line," that is, incoming information is decoded and interpreted immediately, unit by unit (e. g., words, sentences, propositions) at several levels, together with the application of general knowledge and attitudes or personal situation models. The results of this on-line interpretation are hypothetical representations that may be revised by following interpretation steps. This is also true for the construction of a speaker model: Previous evaluations of the speaker, as well as initial impressions based on the first fragments of the conversation, may be continually revised later in the conversation, even when "first impressions" usually lead to powerful (macro)evaluations that may be used to monitor further processing. Here too, the revision of initial hypotheses may be difficult: People are usually inclined to seek information that confirms rather than disconfirms such early hypotheses.

The processing steps just summarized presuppose vast amounts of personal and social knowledge, beliefs, experiences, opinions, and attitudes (Eagly, 1981). That is, the persuasive "effects" of prejudiced speakers are not inscribed on some attitudinal tabula rasa. In our earlier account of the structures of attitude and attitude change, we also stressed the fact that attitudes are very complex, and seldom changed, as a whole, due to one communicative event. Cognitive "changes" may be limited to isolated models about personal experiences, or may lead to very indirect and often delayed changes of model-based ethnic opinions. It simply does not occur very often in an ethnically complex society that social members receive completely new information about ethnic minority groups. It is one of the aims of this section, therefore, to show how existing beliefs and opinions interact with information conveyed in talk.

3.1. CONTEXT UNDERSTANDING

We have shown earlier that speech participants represent the communicative situation in a specific, episodic memory model. This model features the current participants (including self), the overall goals and categorization of the interaction type, as well as relevant social constraints on the interaction, and various setting features. For discourse production, thus, it appeared relevant that the speaker in our interviews models the recipient as an institutional representative (interviewer), as a stranger, as a young (male or female) student, and so on. The converse happens with the interviewer, who represents the speaker primarily as selected "interviewee," but also categorizes the speaker socially according to gender, social class, occupation, status, or other characteristics.

Similar processes are at work in informal talk about ethnic affairs. The interview situation fairly closely simulates talk among strangers,

Interpersonal Communication

303

as may be the case for talk in public places with unknown people in the neighborhood (on the street, in shops), in offices, or in public transportation. Our evidence about various sources people mention (see Chapter 3) suggests that there is frequent conversational interaction about ethnic affairs in such settings and situations. Strategic representation of others in talk may be crucial for the interpretation of their utterances. For instance, if the recipient believes that elite members are not usually racist, he or she will not expect prejudiced talk, and may even tend to ignore those expressions that might be heard as such from other speakers. And conversely, when the recipient already has information about the speaker and his or her beliefs and opinions, as is the case for press or TV interviews with people who admittedly vote for a racist party, understanding their talk will, of course, be monitored by such a provisional person interpretation.

In both cases, however, a fundamental category in the speaker model will be ethnicity. Speakers who are categorized as belonging to the White in-group are in principie represented as in-group members, which also allows inferences about probable opinions and attitudes about out-group members. Direct or media represented talk of ethnic group members, thus, may be less credible for recipients when their opinions are interpreted as being in their own interests. Discounting strategies or other strategies of denying the implications of information given by out-group members may then start to operate during understanding, as we can see in the following example:

(12) *I-F-1 /2 (Man/woman, 50/50, hi-con, P4/P5)

(Not they but the Dutch are being discriminated against)

M:I don't know whether you heard that radio report last Friday on VPRO [broadcast network]?

I:No.

M:There was a Surinamese woman from Utrecht ... well she really took on on the radio she was being discriminated and she lived in a dump and she had to see a city councilman and well VPRO had arranged that she could talk to one of the councilmen. And the only thing that carne from her rnouth was 1 am being discriminated and the Dutch all have good housing, well it is a big lie, it is not

true.

The man's denial of the stated experiences of the Black woman suggests how he might have represented the woman as well as her opinions. Because it is obviously not true that all Dutch people have good housing (conflict with his general knowledge), he may easily reject that opinion at the same time as the intended opinion that Black people are being discriminated against in housing. It is the latter opinion that conflicts with his own, stated, opinion that Dutch people and not minority groups are being discriminated against. By giving an example of a communicative

304 Communicating Racism

situation in which a Black woman is represented as telling lies, he need not even make explicit the speaker model assumption that Black people will either lie in general, or show prejudice against Dutch people. In general, then, we assume that in a racist society, ethnic minority group members are represented as being less credible, in particular when speaking about discrimination and racism, because in that case their topics and goals of speech may be represented as self-serving. This presumption of "bias" is also reflected in the representation of such talk in the media: Accusations of discrimination or racism by minority group members are usually accompanied by quotation marks, or other distance markers. Also in our own data, foreigners represented are seldom modeled as truthful speakers. It is interesting to note, however, that when they are represented as denying racism or as critics of their own group, their credibility is very high, and their utterances often taken as good evidence for the opinions of the speaker, as we see in the following example:

(13) I-C-6 (Woman, 60, lo-con, P6)

Alright, you also have good ones because my husband works with uhh ... well Surinamese and all that, who do a very fine job, but they say themselves: Yes uhh it is a pity but most of those who come here aren't worth much already in their own country ... .

Note that the credibility of the Black speakers is further emphasized by the woman when she interpolates that the Surinamese her husband work with "do a very fine job." This is surprising when we also note that in general the woman speaks very negatively about ethnic minority groups. This suggests that speaker representation may be flexible and context bound. Obviously, such positive (ad hoc) models have a strategic nature in persuasion. The same speaker would probably not be represented as positively in other situations by the same recipient. If speakers show belief similarity then there will be a tendency to represent them more positively (sometimes even across ethnic boundaries), which is in accordance with well-known experimental fmdings (Rokeach, 1968). Note, though, that perceived belief similarity need not generalize to the group of the speaker as a whole, nor to overall positive attitudes. In our case, thus, Surinamese are found (more) credible only when they díscredit their own group. The converse will be true for White speakers who denounce racism and discrimination of ethnic groups. In our earlier examples (section 2.2), we saw, for instance, that a highly prejudiced woman resents the positive treatment of "minorities" by the government. This suggests that authorities and the elite may be represented by members of nonelíte groups as being biased in favor of ethnic groups and, hence, less credible. Here is a good example:

Interpersonal Communication

305

(14)I-G-7 (Man, 45, market vendor, P6)

A few months ago 1 saw on TV, there was a minister who tells a Turk, a Turkish girl it was 1 believe in Sonja Barend's show [well-known talkshow on Dutch TV] and she says yes but the Dutch they put us off, then that minister says, 1 don't remember his narre, but he says but the DUTCH have to adapt to those foreigners. 1 ask you, where are we heading like this?

In this example, ethnic group and class factors interact in the ways the man represents the talk of a Turkish girl and a cabinet minister. Note, though, that in general, White (elite) speakers who defend the rights of minority group members or protest against racism are represented as less credible, or even as "traitors." Whereas they might be more credible due to their elite status (position, expertise), their different ethnic attitude might be seen as especially damaging to the in-group. This explains why White antiracists tend to be portrayed very negatively in (other) elite discourse, such as in the media.

From this discussion and the examples we may provisionally conclude that speaker modeis in understanding are topic and context bound. When "foreigners" are the topic of talk, it matters which group member is speaking, and what the possible interests of the speaker are. For prejudiced recipients (whether elite or nonelite), "pro-foreigner" in-group members especially tend to be represented very negatively, and their utterances, therefore, become much less credible. A final, personal example may illustrate this point. In a letter from the editor of a well-known Dutch elite weekly (Intermediair), read by managers and social scientists, to the author, an article about news analysis and the portrayal of ethnic group members in the press was being rejected on the basis of the following revealing arguments (note that the statements rejected were being supported by much empirical evidence and references to similar findings in many other publications):

(15)Especially what has been stated for news reports about minorities remains unproven andan unacceptable caricature of reality. The thesis "that the tenet of most reports is that ethnic groups provide problems for us" is in my opinion not only not proven, but simply incorrect.

Theoretically, speaker models of recipients are organized in a way similar to that of persons or actors in general, such as in terms of a person schema featuring a number of relevant categories (see Hastie et al., 1980, for details). In addition to the group schema categories (Origin/ Appearance, Socioeconomic status, Sociocultural properties, and Personal characteristics) we have postulated in Chapter 4, individual person schemata may feature information about whether or not a person is

306 Communicating Racism

a friend or an enemy, or other categories that determine interpersonal communication and interaction. For speakers in opinion interviews or generally in talk about controversial issues, personal evaluative categories such as honesty, truthfulness, or accuracy may be involved as components under a higher-level category of "credibility." The general group membership categories provide stereotypical evidence for evaluating individual speakers, and we have seen that different ethnic group membership (known or derived from origin/appearance/language), conflicting interests, lower status, and cultural differences may add to the negative personal evaluations about the speaker.

The result of such a very complex speaker model within the overall context model the recipient is constructing of the conversational event is an equally complex interaction with the representation of what is being said by the speaker. For instance, high status may in principle make the speaker more credible, but different beliefs or group allegiances may again reduce credibility. The hearer, therefore, must decide from further text and context analysis which of these attributed characteristics is most relevant in the present context. Thus, when the topic is ethnic affairs, and when the hearer is prejudiced, the high-status speaker may get much attention, but what is said may be discredited on other grounds.

This suggests that other important components must also be represented in the context model. First, the goals of the speaker are, of course, relevant. If the speaker is perceived to be aiming at persuading the recipient, the activation of his or her own topic-relevant attitudes is more relevant than when the speaker is apparently criticizing others. Then, the context type may also be an important factor. When speaking in an official function, in public, or in a formal situation, the speaker may receive more attention, and his or her credibility may be enhanced or reduced when expressing opinions that are inconsistent with those of the recipient. This also explains why people tend to use the media more often as support for their opinions than the personal communications of other arbitrary sources.

The point of this section has been to show that before recipients even start interpreting talk from speakers, they already tend to represent their relevant beliefs about the speaker strategically, and to construct inferences that may predict what the speaker is likely to say. Relevant situation models, attitudes, or other cognitive information in that case are being "prepared" for actual understanding. These expectations may even lead to provisional "counterarguing" (and, hence, to self-confirmation) when the speaker is believed to have different opinions about an expected topic of talk (Cialdini & Petty, 1981). The context model, or at least its main information (macropropositions), finally becomes part of the memory Control System, which monitors the actual participation in, and, hence, understanding of the conversation.

Interpersonal Communication

307

3.2. UNDERSTANDING PREJUDICED DISCOURSE

Although (full) understanding of persuasive discourse appears to be neither a sufficient condition nor a good predictor of attitude change, we may safely assume that it is mostly a necessary condition (Eagly, 1974). (We ignore cases where people are already "persuaded" by a discourse before it is actually uttered and understood because, in that case, persuasion is based only on people's assumptions about future communicative events, not on those events themselves; see Cialdini & Petty, 1981.) No serious insight into the processes of communicative attitude change is possible, in our opinion, without an explicit account of the processes of discourse understanding on which they are based. The memory contents and structures of information derived from discourse are a function of the ways this discourse is understood in the first place. And we need to know about these contents and structures when we want to account for the formation of, or the interaction with, opinions and attitude structures. This also holds for cases in which people are being persuaded by discourse they understand only in part, such as news reports or other media messages and programs. This relevance of the cognitive processes of discourse understanding, of course, implies the fundamental interaction between representations of discourse structures with other information (models, scripts, attitudes, and so on) stored in memory. This is also the major reason why further persuasion processes, such as opinion formation or attitude change, cannot be explained or predicted from discourse contents or structures alone. Contrary to the classical studies that defined "persuasibility" in terms of personality characteristics of the recipient (e.g., authorianism, self-esteem), however, we emphasize the important role of the contents and structures of current models and attitude schemata of the recipient, as well as the (individually variable) applications of strategies for their change (see Eagly, 1981, for discussion).

The processes operating during understanding and memory representation of prejudiced talk are in part similar to those outlined for the production of such talk. Wejust suggested that these processes are monitored by three different kinds of information: (a) discourse structures and their memory representation (the "text model"), (b) the context model, representing the "facts" talked about, and (c) previous knowledge, beliefs, and opinions activated and applied during understanding. When understanding prejudiced talk, a special topic is being constructed and special knowledge and attitude structures are being activated, namely, those featuring "foreigners" or "ethnic groups" and related information as concepts. This may mean that nonprejudiced discourse may be interpreted in a "biased," prejudiced, way (or conversely), or

308 Communicating Racism

that prejudiced discourse is similarly or differently understood by a ,prejudiced recipient. We limit our analysis to an account of the understanding and further processing of prejudiced discourse, by both prejudiced and nonprejudiced recipients. In the first case, recipient prejudices may finally be confirmed or changed, and in the second case, they may be formed (or not) due to the prejudiced discourse.

Phases of Discourse Comprehension

Before we turn to the analysis of more specific cases, let us briefly summarize and further explain the major phases of cognitive discourse understanding. Although we mention these phases in a specific order, it should be recalled that understanding is a strategic process in which information may be processed from various sources and levels at the same time and in varying order, depending on the goals and information required by the strategic operations. In other words: Part of a process mentioned first may be attended to only in later stages of understanding, and conversely. To illustrate the various phases of understanding, we take a concrete example of our data, and theoretically "simulate" how a hearer could interpret such a passage.

(16)(MdU2, W, 50, Hi-Contact, P4)

W:What I find unreasonable sometimes, is uhh ... I am also a woman on my own. I moved from that other apartment to this one, so I needed money. So I went to welfare and asked if they could help me. But no. But the Surinamese who come to

this country they get a check for eight-nine thousand guilders, so that they can decorate their apartment. So then I threw in: "Why do THEY get it?" OK? "These are circumstances." I said "Aren't we circumstances?" Look,

those are things ... I was begging for my children's allowance and then a Surinamese puts, he puts a knife beside you and he gets it right away. And that, that are things .. .

I: That really happened like that? W: Yes, really, yes, yes, yes, yes

I: That a knife was put down ... on the table?

W: Because I jumped away. I thought, because he was so aggressive that man. They called him twice. Oh that is the man who was behaving like that last time as well. Well, he puts that

little knife on the table, and that little check was already waiting for him. Oh yes, and that indeed is ....1 am not making ... 1 don't tell fantasy stories, but that is true.

I: And that was at the welfare agency?

W:And then I said "Is my color not alright, or what?" Isn't it? Well, he got it. And that happens everyday. If you have

a small thing going, and think by yourself you want to redecorate your apartment ... then you are afraid they might

Interpersonal Communication

309

get to know that. And those guys they are driving around in big cars, they hang ... [makes a gesture indicating big jewels on their fingers].

This is a very typical story, told by a woman who does not seem very prejudiced on all accounts, but who resents the perceived favorable treatment of Surinamese immigrants. Because such stories give concrete "evidence" for such resentment, they are potentially very persuasive. Let us, therefore, examine how such a story is understood in the first place. For the moment we ignore the context model built by the recipient, as well as the—influential—text and situation models already constructed by the hearer on the basis of the previous parts of the conversation. The current topic is "living with foreigners in the neighborhood" (they get apartments easily, have many children, make a lot of noise, "but I have nothing against them").

Decoding and Analysis

Once perceived, the sounds of talk by the mechanisms of sensory stores, operations of decoding, and analysis are applied in Short Term Memory (STM), and yield morphematic (word) sequences and part of their syntactic structures (categories, functions, ordering). Although these processes are basically bottom-up, they may be strategically facilitated by topic-controlled local semantic analysis: Word meanings that are predictable from currently established semantic representations can be decoded and analyzed more easily. The knowledge applied here is basically general knowledge of the language and of the regional dialect. Yet,. pronunciation and syntax, of course, also yield signals that allow personal and social inferences that may form or change the model of the speaker (e.g., origin, class, and gender, in this case: a poor woman from Amsterdam, probably with little education). We have seen that this speaker model in turn may influence attention, understanding, and processes of opinion (trans)formation.

In Chapter 2, we have found that talk about ethnic affairs also has a number of surface structure features that might be rather specific, such as pauses, false starts, repairs, hesitations, and so on, for instance, when a specific ethnic group or its properties are being described. We may assume that such signals are partially observable and interpretable by hearers, for instance, as indications of uncertainty and hesitation, or as manifestations of social control in speech. These interpretations may be added to the speaker model, and in general, we may assume that they will reduce the overall credibility of the speaker. Foreigners and opinions about them typically are a delicate topic, for which much social

310 Communicating Racism

control and self-monitoring may be needed, and such control processes may show in the speech phenomenajust mentioned, which allows hearers to make provisional inferences from them about the speaker. We may also hypothesize that these inferences may be more negative for hearers who are less prejudiced than the speaker because, in that case, the speech production "errors" can be interpreted as "uncertainty" or "hesitation" about ethnic affairs in general and about the current events told about in particular, rather than as concern for accuracy and adequate formulation. In other words, before actual semantic interpretation, a hearer can already make many personal and social inferences about the possible social status and ethnic opinions of the speaker on the basis of superficial speech characteristics (Scherer & Giles, 1979).

Semantic Interpretation

Central in understanding, however, is the semantic interpretation of the decoded and analyzed forms of speech: Words, phrases, clauses, sentences, and sequences of sentences are assigned semantic representations, that is, concepts and propositional structures. Words and phrases are interpreted strategically by a complex information input, such as syntactic and intonational information, knowledge of the lexicon of the language, meanings of previous words and sentences, the overall topic, as weil as social and personal information from the speaker model. Clausal and sentential interpretation in addition requires propositional schemata and an interpretation of coherence relations, such as functional or conditional relationships between them.

At this point the situation model starts to operate: A sentence can be understood only if the hearer can build a semantic representation that corresponds to a possible state of affairs (events, acts, and so forth) in a relevant situation model. That is, ultimately there is no (full) understanding when people cannot "imagine" what a text or talk is about. This does not mean that partial understanding, for instance, of meanings (as represented in the text model), or of reference (a fragmentary model of the situation), is impossible. Complete understanding may even be only an ideal goal of interpretation, at least in the sense of constructing a model exactly as intended by the speaker. Because ah l understanding is subjective, thus, communication cannot possibly be "complete" in this sense. Also, there is theoretically no boundary to the completeness of the model (much general knowledge and many previous personal experiences may be inserted into the model) and, hence, no limit to the "depth" or "breadth" of processing. However, contextual relevance and time limitations, of course, usually constrain such processing to manageable proportions.

Interpersonal Communication

311

In our example, the hearer already knows that the woman speaks about foreigners as part of the main topic, and also that she is particularly concerned about their favorable treatment. Unlike other stories, however, this story is not preceded by a summary statement expressing its planned topic (although a general topic of unfavorable treatment is expressed earlier in the interview). So, the hearer cannot yet use such a topic to interpret, top-down, the initial clauses of this passage. Yet, instead of the summary statement about the topic, there is a summary statement of a major category of this kind of story, namely, of the Evaluation. Stories about minorities often start with an Evaluation. This suggests that their point is not primarily a macrostructurally represented event or action, but rather a (negative) opinion. In our example, the first sentence denotes an opinion ("I find") with a specific evaluation ("unreasonable"). The hearer may now expect further information about what the woman finds unreasonable, and because similar concepts have been activated before, the hearer may also provisionally infer that the unreasonableness may be connected to what foreigners do, or what Dutch people do for foreigners.

At this point also, stylistic choice, such as lexicalization, becomes important. Note, for instance, that the use of a word such as unreasonable and the added temporal quantifier (sometimes) mitigate the opinion, which may be interpreted by the hearer (a) literally, asan expression of an opinion about a minor problem, (b) as a weakly negative opinion,

(c) as rhetorical understatement of a strong opinion about a big problem, and/or (d) as an indication of strategically monitored self-presentation in the interaction: Opinions about foreigners are mitigated (as the women did before, more explicitly). Due to several factors, thus, this beginning could result in creating a slightly (more) positive impression of the woman. We see that the cognitive processing of impression management is a continuous, strategic process, also in the hearer, who may constantly "adapt" his or her model of the speaker after each discourse unit (e. g., a sentence). This model will then feed back to the interpretation of the rest of the conversational contributions of the speaker. The strategic, that is, hypothetical, result of the semantic interpretation of the first clause, then, is a high-level evaluative proposition dominating the interpretation of the rest of the story. The woman does not simply tell the story to have something interesting to say, or to portray herself as a hero; rather, the story is presented as an example and as further evidence for what she finds unreasonable about the ethnic situation in the neighborhood and the city. The resulting textual representation of this conversational fragment in episodic memory, that is, the text (story) model, thus, has an evaluative proposition as part of its highest-level macroproposition. This theoretically explains findings in social psychology that