Article 4

    --------------------------------------------------------

    This work attempts to carry out a discourse analysis of natural language data which involve conversations between American undergraduate students about issues of race. Many of these conversations about race reveal recurrent identifiable underlying structures and functions which lead to the conclusion that what is involved is a fairly well-defined speech activity. This speech activity shares many features in common with argumentative discourse. In particular, both types of discourse involve sequences of the same underlying ideational units (positions, supports and disputes) which are configured within larger units (arguments). On the other hand, despite this superficial resemblance, the data to be analyzed in this work are quite distinct from ordinary argumentative discourse on structural and functional grounds. The kind of discourse involved in the data is called “pseudo-argument” (PA). This is because the resemblance between the two discourse types is more than coincidence—it is argued that participants in PA borrow elements of ordinary argument in order to shield themselves against negative inferences which might arise among the participants. To put it another way, by making their talk appear argumentative, participants are able to appear rational, unbiased, and nonracist in what is an essentially one-sided discussion in which nonnormative beliefs and attitudes are continuously put forward and go unchallenged.
       Pseudo-argument arises in talk about delicate issues (such as race) in which all participants are allied in their nonnormative beliefs and at the same time are mindful of countervailing social norms and therefore work to forestall the negative impressions which can result from holding views that may not be entirely acceptable to the larger society. Pseudo-argument is a collaborative activity in which participants jointly construct arguments (in the sense of to “make” an argument), and also simulate arguments (in the sense of to “have” an argument) in which an absent antagonist is imported into the conversation and subsequently disputed. It is this collaboration within the organizational framework of PA which allows participants to diffuse, reinforce, practice and validate their beliefs and values, and for this reason it is claimed that the study of PA may contribute to our understanding of discursal processes having to do with the reproduction of the “modern racist” ideology in everyday face-to-face discourse.
       The conversations analyzed in this work show characteristics of what has variously been called “modern racism” (McConahay, 1986), “aversive racism” (Gaertner and Dovidio, 1986) and “symbolic racism” (Kinder and Sears, 1981) by social psychologists. Unlike “old-fashioned racism”, which may be characterized as the straightforward, open expression of hostility in both word and deed, “modern racism” involves a mix of conflicting values, beliefs and feelings:1
       In our view, aversive racism represents a particular type of ambivalence in which the conflict is between feelings and beliefs associated with a sincerely egalitarian value system and unacknowledged negative feelings and beliefs about blacks. (Gaertner and Dovidio, 1986: 62) According to Gaertner and Dovidio, because of their strong egalitarian values, modern racists in the United States support public policies which in principle promote racial equality. They may identify with a liberal political agenda and think of themselves as nonprejudiced. On the other hand, modern racists possess unacknowledged negative feelings and beliefs about Black people. In addition, modern racists are said to avoid expressing overtly anti-Black opinions, instead preferring to express their views in more subtle, sophisticated ways which may be defended by appeal to seemingly universally accepted egalitarian values and principles.
       According to McConahay (1986), the ideology of modern racism in the United States includes the following tenets: 1. Discrimination is a thing of the past because Blacks now have the freedom to compete in the marketplace and to enjoy those things they can afford. 2. Blacks are pushing too hard, too fast and into places where they are not wanted. 3. These tactics and demands are unfair. 4. Therefore, recent gains are undeserved and the prestige-granting institutions of society are giving Blacks more attention and concomitant status than they deserve. McConahay adds two more tenets adhered to by modern racists: 5. Racism is bad and the other beliefs do not constitute racism because these beliefs are empirical facts. 6. Racism, as defined by modern racists, is consistent only with the tenets and practices of old-fashioned racism: beliefs about Black intelligence, ambition, honesty, and other stereotyped characteristics, as well as support for segregation and support for acts of open discrimination. (pp. 92–3)
       Thus, aversive or modern racists in the United States are ambivalent or conflicted between strong egalitarian values and negative feelings and attitudes towards African Americans and other minorities. This ambivalence helps determine a set of beliefs which compose what may be called a modern racist ideology, which is characterized mainly by the conviction that discrimination and racism no longer exist, and therefore that any “advantages” given to African Americans are unfair and undeserved. American culture has historically been a racist one, and the continued presence of anti-Black sentiment in the United States is well documented (see reviews of Crosby et al., 1980; Dovidio and Gaertner, 1986; Katz et al., 1986). Although racial attitudes and beliefs in this country have progressed beyond the “old-fashioned” kind (Schuman et al., 1985), issues of race and discrimination are highly salient ones in contemporary American culture.
       The institutional and interactional mechanisms underlying the current forms of racism in the United States are less obvious and more subtle, often masked by appeal to values defensible on nonracial grounds. Therefore, the question of how modern racism is reproduced at different levels is one which is worthy of consideration. One facet of this question is how modern racism is reproduced at the level of face-to-face verbal interaction, since this activity seems to be a prominent venue for the diffusion and reinforcement of beliefs and attitudes associated with the modern racist ideology. Little attention has been given so far by scholars to how modern racist beliefs are diffused and reinforced in day-to-day face-to-face verbal interaction by members of the dominant White majority. With the exception of a few studies (e.g. Van Dijk, 1987; Essed, 1991; Wetherell and Potter, 1992; Verkuyten et al., 1994), there appears to be little known about how the ambivalence involved in modern racism becomes manifest in everyday discourse. That is, few scholars have addressed the questions of how the conflict of egalitarian values and anti-Black feeling shapes discourse in talk about race, and conversely how that particular discourse may serve as a vehicle for the reproduction of modern racist beliefs and attitudes in face-to-face interaction.
       It becomes apparent in the investigation to follow that the data of this study give strong evidence for the modern racist orientation described by social psychologists. In particular, the content of many of the conversations repeatedly involves appeals to egalitarian values in the face of “unfair advantages” given to African Americans. In addition, the conversations show subtle evidence of the negative feelings attributed to modern racists by the social psychologists mentioned earlier. Further, participants in the conversations often explicitly make statements such as “Racism is in the past”, “They get all the advantages”, “It’s so unfair”, “They’re always whining and complaining . . .”, etc.—all of which serve to confirm McConahay’s summary of the tenets of modern racism. Finally, the data provide evidence that participants are extremely concerned about avoiding the appearance of being “racist”, and that much of the organization and content of their talk is geared towards forestalling this impression.
       [....]
       DATA COLLECTION
The students and their respective friends who participated were undergraduates, most of whom were White, presumably middle class and young.
       All of the conversations recorded were between ethnic sames. That is, all conversations were either all-White or all-Black (and in a few cases all-Asian). We requested this because we wanted to record the most uninhibited, spontaneous speech possible, and mixing races (ethnic groups) would probably result in a (no less interesting but) less honest, less forthright discussion of the issues. Our method of having students interview their friends was deliberately chosen for the same reason. That is, this method allows for a minimizing of observer’s influence on the speech behavior of participants. Without the presence of an authoritative outsider as interviewer, we hoped that the conversations would approach more closely the natural, less self-conscious level of ordinary talk among friends. We hoped to capture spontaneous behavior, natural discourse and honest beliefs. These hopes were not disappointed—the conversations were surprisingly open and spontaneous. It seems that despite the presence of a tape-recorder, the groups of friends engaged in the kind of sincere, animated, heartfelt discussions which only take place behind closed doors. There are very few references in the transcriptions to the tape-recorder, and while the influence of recording cannot be ignored, it seems safe to say that the speech of the participants reflects their ordinary way of carrying on in similar situations.
       [....]
       THE STRUCTURE OF PSEUDO-ARGUMENT
Schiffrin (1985: 45) defines argument as “. . . a discourse genre through which individuals support disputable and disputed positions”. She distinguishes between ‘rhetorical’ argument and ‘oppositional’ argument (which parallels O’Keefe’s (1977) distinction between ‘argument 1 ’ and ‘argument 2 ’). She defines ‘rhetorical’ argument (O’Keefe’s argument 1 ) as “discourse through which a speaker presents an intact monologue support-ing a disputable position”, and defines ‘oppositional’ argument (O’Keefe’s argument 2 ) as “an interaction in which an opposition between speakers cre-ates an extended polarization that is negotiated through a conversation”. Most contemporary definitions by argument theorists make reference to two or more participants who must be present for an argument to occur. For those definitions which explicitly distinguish the two senses of argument (for example, Schiffrin’s), this condition applies to arguments 1 as well as to arguments 2 . Second, all of the definitions refer in one way or another to the condition that there be some opposition or disagreement between the participants: “It may seem trivial to say that arguments require dissensus, but this claim is one of the few points of agreement among Argumentation theorists” (Willard, 1989: 53).
       It is shown here that although the conversations which comprise the data resemble argumentation, they clearly deviate from ordinary argumentation in many respects. While the conversations give evidence for the continuous presence of positions, supports and disputes, there is no opposition or disagreement between participants—the disputes which occur are aimed not at other (allied) participants, but rather at an imagined nonpresent antagonist whose positions and supports are imported into the conversa-tion. In addition, unlike ordinary arguments 1 where positions and their respective supports must be provided by the same speaker, in the data of this study, a single position may be collaboratively supported by two or more participants. It is argued in this section that PA deserves to be treated separately from ordinary argument (OA) in part at least because it has different structural properties. With the aim of distinguishing PA from OA, it is shown that although these types of discourse share the same underly-ing units (positions, supports and disputes), they differ in how these units are configured and distributed among participants.
       Following Schiffrin (1987) and Preston (1993), it is assumed here that positions, supports and disputes are the underlying (ideational) units which serve as the building blocks of argumentation in discourse. Moreover, treating positions, supports and disputes as the minimally necessary under-lying units of argument structure does not help to distinguish OA from PA since both contain these units (it is on this point that they overlap). It is shown that distinguishing these types of discourse on structural grounds requires reference not only to the ideational units, but also to how these units are configured in the sequential organization of the discourse and how they are distributed among the participants. In what follows, it is argued that the data give evidence for two distinct configurations or patterns of ideational units in PA, neither of which is found in OA. The first configu-ration, called collaborative argument 1 (CA1), shows features of argument 1 , whereas the second, called collaborative argument 2 (CA2), resembles but is clearly different from argument 2 . More specifically, the first pattern of ideational units involves the collaborative support of a single position, whereas the second involves the collaborative dispute of an absent antagonist’s imported position and support. It is shown that these two configurations are recurrent in and indicative of PA, and allow the analyst as well as participants to identify the talk as such.
       [....]
       PSEUDO-ARGUMENT [PA] vs ORDINARY ARGUMENT [OA]
Although positions, supports and disputes occur in ordinary arguments, neither CA1 nor CA2 occur in ordinary arguments. In PA, participants are busy collaborating either in building cases for their own positions (CA1), or else in identifying and destroying the antagonists’ arguments 1 (CA2). These collaborative activities are analogous to what happens in OA where an individual presents his or her own arguments 1 (position and supports) and attempts to destroy his or her opponent’s arguments (argument 2 ): “. . . when two individuals dispute each other’s positions in oppositional argu-ment, their talk is directed not only to making their own points but also to challenging their interlocutor’s points” (Schiffrin, 1985: 45).
       [....]
       THE FUNCTION OF PSEUDO-ARGUMENT [PA]
The kind of structural analysis offered so far in this paper does not address the central (functional) question of why such structures occur in the first place. That is, what is it about the context, topic, and the goals and motives of participants which combine to produce such structures? It is argued that much of the structure of PA (including the structure of CA1 and CA2) is due to identity work on the part of participants in an effort to forestall negative inferences by others, and to project an image of rationality, objectivity and fairness. That is, argument structures are employed by participants in PA primarily for the purpose of face-work and only secondarily for the instrumental purpose of persuasion (as in OA). It is later argued that beyond this facade of fairness and rationality is an activity whose main purpose is to allow participants to simulate, exchange, reinforce and practice beliefs and arguments in a nonhostile environment. Goffman (1967) observed that individuals do not just go about serving their own instrumental goals in interaction, but at the same time maintain a vigilant concern for the “face” of themselves and others. According to Goffman, there is a preference for maintaining a positive public image (“face”) in interaction, both for ourselves and others, while pursuing overt objectives. Brown and Levinson (1987) elaborated Goffman’s notion of “face” in their theory of politeness which embodies the notion that inter-actants use linguistic politeness as a form of social currency in the carrying out of sometimes face-threatening actions. The works of Goffman (1967) and Brown and Levinson (1987) illustrate that the prevalent, universal con-cern of individuals for protecting their own self-image and the image of others plays into every interaction and is a powerful force in the minutiae of everyday discourse.
       That particular discursal strategies may be employed to satisfy compet-ing instrumental and identity (face) goals has been noted by Kline (1987), with respect to (ordinary) argumentative discourse:
       Since both parties simultaneously pursue both instrumental and identity aims, agreement on situational identities is constantly negotiated and reaf-firmed. Thus a crucial problem for communicators becomes one of using strategies that allow for the accomplishment of instrumental goals while simultaneously actualizing a desired identity for oneself and other. (p. 243) The work of Kline shows that in OA, which is inherently face-threaten-ing, a variety of rhetorical strategies is employed by arguers which allows them to balance their conversational goals (to alter the adversary’s beliefs) with identity goals (to avoid affront, to present a positive self-image, etc.). I would like to propose that it is this concern for identity and self-image that motivates the structures of PA. It is argued in what follows that the structures of PA stem primarily from a strategy by participants which allows them to achieve their instrumental goals while at the same time fore-stall negative inferences by others. This functional explanation will appeal to the nonnormative nature of the content of the conversations coupled with strong social pressures against expressing what might be considered “racist” views. PA structures are designed to allow participants to put for-ward their ideas without incurring negative inferences.
       Following Van Dijk (1987), it is assumed here that higher level discourse strategies are routinely employed in discussions about race which allow a speaker to balance competing instrumental and identity goals. Van Dijk (1987: 86) frames the issue as follows: . . . prejudiced talk about minorities, among other things, has the overall goal of negative other-presentation, while at the same time preserving a positive impression (or avoiding the loss of face). These goals are some-times in conflict, for instance, when social norms do not allow uninhibited negative talk about minority groups. Therefore, expedient strategies are in order to reconcile these real or apparent inconsistencies. These strategies are accomplished by sequences of moves that try to realize both goals as effectively as possible, for instance, with a maximum of negative other-presentation and a minimum of negative self-presentation. In other words, speakers persistently try to manage or control the social inferences the recipient is bound to make about them on the basis of what they say . . . I follow Van Dijk (1987: 76) in claiming that higher level argument struc-tures themselves may be used to ward off negative attributions: The expression of delicate or controversial social opinions in conversation is routinely expected to be backed up with arguments . . . Within the com-bined strategies of positive self-presentation and negative other-presen-tation, such arguments have the fundamental function of protecting the speaker against unwanted inferences about his or her ethnic attitudes. In this way then, according to Van Dijk, the provision of supports for positions in talk about race is fundamentally to protect the face of the speaker.
       THE FUNCTION OF COLLABORATIVE ARGUMENT 1
Why do participants in CA1 bother to contribute supports for a position that all participants already accept? Supports are given because arguments ward off negative inferences about the speaker. There is ample evidence in the data that the participants are anxious to avoid the attribution of “racist” given the content of their talk.
       [....]
       The tendency to explicitly dissociate oneself from potential racist attribution is a common element of modern racism (Gaertner and Dovidio, 1986; Billig, 1988). According to Gaertner and Dovidio, since, for modern racists, negative feelings and beliefs about minority groups are usually excluded from awareness, “When a situation or event threatens to make the negative portion of their attitude salient, aversive racists are motivated to repudiate or dissociate these feelings from their self-image . . .” (p. 62). This accounts for why the statement of repudiation made by K in lines 38–40 in the previous excerpt occurs at a moment in the interaction where the preceding talk appears to have become openly hostile and pejorative (see Sample 2).
       In sum, the function of CA1 is to balance both instrumental and identity goals—the provision of supports allows the expression of nonnormative positions while at the same time protecting those involved from their own negative attributions.
       THE FUNCTION OF COLLABORATIVE ARGUMENT 2
Why do participants bother to import the positions and supports of an absent antagonist? That is, what function or functions are served by bringing in the views of “the other side” within CA2? I would like to suggest that a similar motive of identity management is involved for participants—by presenting the views of an absent antagonist, participants project an image of fairness, balance and objectivity, and therefore shield themselves from negative inference. A CA1 has the undesirable effect of appearing one-sided and biased since there is no opposition to balance the unified stance of the participants. Importing the supposed views of an antagonist thus helps to dissolve the appearance of bias and unfairness, and so gives the impression that participants are simply seeking objectively arrived at conclusions, tested by a fair consideration of both sides of the issue at hand. By presenting the “other side”, participants in PA attempt to dissolve the impression that they are simply airing their negative attitudes. However, the CA2s in the data hardly involve unbiased presentations of an antagonist’s ar-guments. Rather, the presentation of the antagonist’s arguments amounts to no more than a minimal token more aimed to serve impression management than objectivity. Consider once again Sample 3, which involves a CA2. In this excerpt, the argument of the antagonist is indeed presented, in reported speech: “Well—you owe it to us because—you know—you had us as slaves—thousands of years ago, and blah blah blah:, and you OWE it to us. Blah Blah”. It would be difficult to make a case that B’s presentation of the antagonist’s position is unbiased here. On the contrary, the direct quote itself is loaded with negative evaluation. So although the speaker bothers to supply an argument on behalf of the conjured antagonist, it does not appear to be done with the aim of neutral, objective consideration. Rather, it is a token gesture which functions to ward off the attribution of one-sidedness. Note that in this case, and in all cases found in the data, the provision of the antagonist’s argument is minimal, that is, it rarely goes beyond giving one or two supports, although certainly other supports are available. An objective presentation would more likely present more than one or two supports for each position. In addition, the supports provided are not necessarily the most persuasive or effective ones which could have been selected; it is often the case that the imported supports are transparently weak and stereotypical.
       Thus it seems that although the views of the antagonist are supplied, they serve some purpose other than the pursuit of objectively arrived at conclu-sions. The data suggest that participants import an antagonist’s argument in order to set up a straw man which can be collaboratively disputed. That is, just as in OA in which one tries to destroy the antagonist’s arguments in order to win, CA2s set up and destroy the imported antagonist’s arguments. Furthermore, the antagonist’s arguments are presented in such a way that they are designed to be destroyed in order to allow the participants to “win”. To put it another way, in CA2, the imported arguments are selected for their weakness so that the participants may win the PA. In addition, it is likely that the participants select imported arguments to which they already have in advance some ready-made attack—there are no cases in the data in which an imported antagonist’s argument is given and no retort/dis-pute is immediately provided.
       In sum, there appear to be two main functions of CA2s. The first has to do with identity work—by presenting the views of the antagonist, partici-pants ward off the potential impression that they are biased and simply airing negative attitudes. By giving the point of view of the other side, the discussion appears fair, balanced and objective. The second has to do with an instrumental goal—setting up the antagonist’s position in order to dispute it. Despite appearances, CA2s are designed to defeat the antagon-ist from the start; weak, stereotypical imported supports are selected with ready-made disputes. Winning an argument (whether real or pseudo-) includes not only presenting one’s own arguments, but also destroying the arguments of the opponent.
       The simulation of an argument 2 in PA allows participants the benefit of having considered the views of the antagonist and therefore shielding themselves from the appearance of bias. It appears that the goal of each participant in CA2 is not to change the opponent’s view (the opponent is absent), but rather to reinforce the “rightness” of their own side’s views. This is accomplished by jointly disputing each and every argument pre-sented on behalf of the antagonist—there are no instances in the data where the absent antagonist “wins” in PA. Also, unlike in ordinary argu-ment 2 in which very often no resolution is reached (Willard, 1989), in PA resolution is reached in every case, with the victory of the protagonists and the vanquishment of the antagonist.
       [....]
       In sum, PA earns its name because it simulates OA in various ways in order to borrow certain appearances which derive from ordinary argumen-tative discourse. In other words, PA looks like OA (and is meant to), since the use of argument structures (including argument 1 and argument 2 ) lends the appearance of fairness, objectivity and rationality, and wards off nega-tive inferences about the participants. In addition, PA is “pseudo” in the sense that despite these appearances, the discourse is anything but fair and objective. It has been argued that participants engage in PA not primarily for the purposes of persuasion, but rather in order to reinforce their mutu-ally held nonnormative beliefs and perhaps more importantly to reassure each other of their (perhaps threatened) ideological alliance and solidarity. The use of argument structures in pseudo-argumentative discourse then is “pseudo” in that it accomplishes something other than what it purports to accomplish on the surface. Rather than actually providing fairness and objectivity, it mainly functions to control negative inference in an essen-tially one-sided, attitude-laden discussion. PA is, in sum, a type of discourse which allows participants to secure their alliance and strengthen mutually held beliefs in the face of strong social pressures which threaten those beliefs.
       [....]
       I contend that pseudo-argumentative discourse plays an important role in the reproduction of the modern racist ideology in everyday talk, since it provides for its users a suitable staging ground for the practice, diffusion, and reinforcement of beliefs and arguments about minority groups, and further allows for the balancing of both instrumental and identity goals. That is, it is a type of discourse which allows participants to exercise nonnormative beliefs with relative impunity. Modern racist beliefs and arguments are rarely expressed and carried out in real-life oppositional situations. Rather, more often they are rehearsed between people who are close to and similar to each other in a nonhostile environment, where pos-itions and supports can be shared, honed, evaluated and reinforced.
       [....]