Eemeren&Houtlosser paper The Contextuality of Fallacies 59 The Contextuality of Fallacies University of AmsterdamFRANS H. VAN EEMEREN PETER HOUTLOSSER Abstract: © Informal Logic Vol. 27, No. 1 (2007): pp. 59-67. Keywords: argumentative activity type, context-dependency, critical discussion, dialectical shift, dialogue type, fallacy, normative model, speech event Van Eemeren and Houtlosser observe that Walton’s (and Walton and Krabbe’s) notion of ‘dialogue type’ involves a mixture of an empirical notion on a par with a speech event or activity type and a normative notion such as the model of a critical discussion. Then they discuss Walton’s contextual analysis of fallacies as illegitimate dialectical shifts of dialogue types and offer an alternative in which both the empirical and the normative dimension are given their due 1. Walton’s treasures During his long career as an argumentation theorist Professor Douglas Walton has treated virtually every topic in the discipline. No other scholar has been so prolific as he is and no other scholar has been that broad. Being fully aware of these great merits, we are happy to observe a great many similarities between Walton’s work and the pragma-dialectical approach to argumentation we are engaged in. In particular after Woods and Walton’s scholarly ways separated in the 1980s, Walton’s approach has been characterized by a pragmatic streak that brings it close to ours, especially since Walton also seems to favor a dialectical perspective on argumentation. When a researcher manifests himself so extensively as Walton does, it is inevitable that not all his contributions are equally revolutionary. Neither can all contributions be equally refined. Some of Walton’s best ideas, however, are extremely interesting and highly stimulating, even if they are not always crystallized out to the full. In our view, they are like rough diamonds that need to be cut and polished. This goes, for instance, for his concept of “dialogical profiles” (1989), which has been taken up by Krabbe (1992) and inspired us to develop a kindred notion of Résumé: Van Eemeren et Houtlosser observent que la notion de Walton (et de Walton et Krabbe) de ‘type de dialogue’ implique un mélange d’une notion empirique à égalité avec soit un événement verbal ou soit un type d’activité et d’une notion normative telle que le modèle de discussion critique. Ils présentent ensuite l’analyse contextuelle des sophismes de Walton comme un virage dialectique illégitime dans les dialogues et offrent une alternative dans laquelle ils accordent mûre réflexion aux dimensions empiriques et normatives. 60 Frans H. van Eemeren & Peter Houtlosser “dialectical profiles” (van Eemeren, Houtlosser & Snoeck Henkemans 2005) that was our conceptual basis for making a systematic inventory of argumentative indicators in discourse (van Eemeren, Houtlosser & Snoeck Henkemans 2007). Another of Walton’s rough diamonds is his idea that fallacies are “contextual” in the sense that it is always conditional upon the background against which it appears whether an argumentative move is reasonable or not. Whether an argument is reasonable or fallacious in a particular case, according to Walton’s thesis, depends on the context of dialogue (1992: 140). In propounding this idea, Walton in fact revives, as happens so often in the study of argumentation, a classical Aristotelian idea, viz., that fallaciousness depends on the context of dialogue, not just on form (Walton 1992: 143). In Walton’s approach, which he developed further in collaboration with Krabbe (1995), the context of argumentative moves can be specified by describing the type of dialogue they are part of. Although we subscribe to Walton’s idea that fallaciousness in argumentation is eventually a matter of context, and even consider this idea a real treasure, we do not fully agree with his view of context as type of dialogue and norms of argumentation depending on type of dialogue. In this contribution we discuss Walton’s dialogue type approach to the fallacies in more detail and try to make some constructive amendments of our own. 2. Dialogue types and speech events Walton’s view that fallacies are context-dependent is based on his “pluralistic view” that argumentation needs to be judged as correct or incorrect “in relation to a multiplicity of different models of reasoned dialogue” (1992: 133). It is therefore imperative to have a proper understanding of what he means by a “type of dialogue.” This is not so easy to find out as it may seem, because on the one hand Walton assigns a normative status to dialogue types, but on the other hand he also appears to view dialogue types as empirical entities. Each type of dialogue, Walton and Krabbe say, “has its own distinctive rules and goals, its permitted types of move, and its conventions for managing the commitments incurred by the participants as a result of the moves they make” and each type of dialogue, they add, “exhibits a normative model, an enveloping structure that can aid us in evaluating the argumentative and other moves contained in it” (1995: 8-9). According to Walton and Krabbe, there are many different normative models of a dialogue, and each type of dialogue has its own distinctive goals and rules (1995: 65). It is important, they emphasize, not to confuse their account “with an account of how participants really behave in instances of real dialogue that take place” (1995: 67). Their normative models “are not meant to be descriptive” (1995: 117). Yet, Walton and Krabbe’s conception of a dialogue type has elements that are unmistakably empirical. In their observation that “discussions that occur in everyday The Contextuality of Fallacies 61 experience are complex speech events that [...] can be mixed composites of more than one normative model” (1995: 82, our italics), normative and empirical perspectives get conflated. The impression that dialogue types also have an empirical status is strengthened by the fact that Walton refers explicitly to Gumperz when explaining that during the course of a conversation between two or more parties there can be a change in the context of argumentation from one type of dialogue to another: “Empirical studies of [Blom and] Gumperz (1972) showed, for example, that students knew from a change of ‘register’ when a class-room conversation changed from [the speech event of] an academic discussion to [that of] a personal chat” (1992: 137). When anthropologists such as Gumperz and Blom talk about speech events—viewed by Gumperz (1972: 16-17) as “the basic unit for the analysis of verbal interaction in speech communities”—they emphatically refer to empirical entities: “communicative routines” which language users view as “distinct wholes, separate from other types of discourse” that often carry special names (Gumperz 1972: 17). Although their spiritual father assures us that dialogue types are to be regarded as normative models, it is not entirely clear what the basis of the normativity of the dialogue types is. Are we talking “etically” about an “external” critical normativity based on analytic considerations of the theorist concerning the (goals and procedures of the) various dialogue types or “emically” about an “internal” empirically-based normativity amounting to a reconstruction of what is regarded sound in practice? In either case we would like to learn more about the background of the normativity and its rationale. Lack of clarity about these issues also affects the appreciation of the criteria used to distinguish between the various types of dialogues. The question here not only is where these criteria come from but also, and more importantly, how they can be justified or accounted for. What, for instance, are Walton’s grounds for observing that “in negotiation dialogue, the aim for both parties is to ‘make a deal’” or that “the quarrel is a type of dialogue where the goal of each participant is to verbally ‘hit out’ at the other, and if possible, defeat and humiliate the other party” (1992: 135)? And what can be said about the criteria in the case in which two types of dialogue overlap, resulting in a “mixed dialogue”? This last query, by the way, is not answered by Walton and Krabbe’s observation that for deciding whether a verbal exchange in a particular case belongs to the one type of dialogue or to another “a good deal of evidence of the text and context of discourse may have to be judged” (1995: 81)—however true this may be, it does not concern a conceptual problem but a problem in the reconstruction process. 3. Dialectical shifts and fallacies Immediately connected with the notion of types of dialogue is the notion of “dialectical shift.” As Walton observes, “during the course of a conversation between two or more parties there can be a change in the context of argumentation or 62 Frans H. van Eemeren & Peter Houtlosser dialectical shift from one type of dialogue to another” (1992: 137). Such a shift can take place from an argumentative context of dialogue to a non-argumentative context, or vice versa, but also from “one context of argument to another” (1992: 144-145). In addition, there are shifts in “flavor,” where the type of dialogue remains basically the same, “but elements from another type of dialogue begin to intrude to some noticeable extent,” and internal shifts, where the type of dialogue also remains the same but there is merely “a shift in its subject matter or setting” (Walton & Krabbe 1995: 120). When dialectical shifts are instant changes Walton speaks of a “déplacement” of the one type of dialogue by the other and when they are gradual he speaks of a “glissement” from the one type of dialogue to the other. In some cases, the one dialogue is “‘sandwiched in’ [embedded] between the prior and subsequent parts of an enveloping sequence of dialogue of another type” and the transition is not so clearly marked (Walton 1992: 137-138).1 Walton links the dialectical shifts with the analysis of fallacies. In Walton and Krabbe’s view, “a fallacy tends to be associated with a shift or transition from one of these contexts to another” (1995: 7). Later on they weaken this view on two levels. First, not all dialectical shifts are fallacious, only illegitimate—“illicit”— shifts are. Second, illicit shifts are primarily related to informal fallacies: “Some dialectical shifts […] are illicit, and these illicit shifts are often associated with informal fallacies” (Walton & Krabbe 1995: 102).2 What does being “illicit” in the sense of fallacious involve? Without explaining whether they are necessary or sufficient criteria, and without explaining how these criteria must be applied, Walton mentions as criteria for legitimacy of dialectical shifts that the new dialogue should not block the goals of the old dialogue and that the shift should be agreed upon by the speech partners (1992: 139).3 A shift is considered “unilateral” when one of the parties quietly abandons the type of dialogue that both parties—explicitly or implicitly—agreed to be engaged in and slips into a different type of dialogue. In such a case, the move the shifting party makes is incorrect when viewed in the perspective of the original dialogue, but supposedly correct in the new type of dialogue. In this new type of dialogue, it “may have a surface appearance of correctness to the uncritical respondent or observer” (Walton 1992: 146). We wonder what the added value is of saying that a unilateral shift makes a dialogue move fallacious (Walton 1992: 146). Why does it not suffice to say that the move itself is inappropriate in the dialogue the parties are engaged in? One does not get the impression that the fact that there is a dialectical shift plays a decisive role in the analysis.4 Walton’s admission that particular moves are fallacious per se in one type of dialogue but not in another type adds to this impression. According to Walton, in a “negotiation type of dialogue,” for example, “threats and appeals to force or sanctions are quite typical and characteristic.” However, if the context is supposed to be that of a “persuasion dialogue,” the same kind of argumentation The Contextuality of Fallacies 63 becomes “highly fallacious” (1992: 141). This leaves us with the mere observation that a particular dialogue move can be sound—or, as Walton puts it empirically [!], “quite typical and characteristic”— in one type of dialogue but fallacious in another.5 Apart from the fact that these queries shed some doubt on the necessity and usefulness of employing the concept of a “dialectical shift” in the analysis of fallacies, there is—again—the identification problem. According to Walton, to judge whether in a particular case of argumentation a certain shift was licit or illicit, “you first have to ask what the original context of dialogue was supposed to be. Then you have to identify the new context, and ask whether the shift was licit or illicit by looking backwards, and judging by the goals and standards of the original context” (1992: 139). If there is not a flavor of circularity to these questions, it is in any case not immediately clear how to get to an answer. Whatever problems the specifics of Walton’s (and Walton and Krabbe’s) analysis of—informal—fallacies may present, their account has one great advantage: It illuminates the context-dependency of all judgments concerning the fallaciousness of particular dialogue moves. Still, as we already suggested in our introduction, in our view this account needs some further refinement. We leave it, of course, to Walton to provide this refinement in his own terms. In the last section of our contribution we describe the refinement that we think appropriate from our pragma- dialectical perspective. 4. A different view of the contextuality of fallacies In our view, the theorizing about fallacies has to start from a general and coherent perspective on reasonable argumentative discourse that provides a common rationale to the study of the fallacies. In the pragma-dialectical theory of argumentation, such a perspective is offered by means of the ideal model of a “critical discussion” that specifies the stages and the types of speech acts that are instrumental in resolving a difference of opinion on the merits (van Eemeren & Grootendorst 1984, 2004). The rules for conducting a critical discussion express the norms that are pertinent to the process of critically testing the acceptability of the standpoint at issue. Any move constituting an infringement of any of the rules is a possible threat to concluding the difference in a reasonable way and must therefore (and in this particular sense) be regarded as fallacious. In the various argumentative activity types that can be discerned in argumentative practice, it is generally not the arguers’ sole aim to conduct the discussion in a way that is considered reasonable, but also to achieve the outcome that suits them best. In our view, the arguers’ “rhetorical” attempts to make things go their way are, in principle, incorporated in their “dialectical” efforts to resolve the difference of opinion in accordance with proper standards for a critical discussion (van Eemeren & Houtlosser 2002). In their efforts to reconcile the simultaneous pursuit of a dialectical and a rhetorical aim, the arguers make use of strategic maneuvering directed at diminishing the potential tension between these two aims. If a party 64 Frans H. van Eemeren & Peter Houtlosser allows its commitment to a critical exchange of argumentative moves to be overruled by the aim of persuading the opponent, the strategic maneuvering “derails.” Because derailments of strategic maneuvering always involve violating a rule for critical discussion, they are on a par with the wrong moves in argumentative discourse designated as fallacies. The difference between legitimate and fallacious manifestations of strategic maneuvering is that in the latter case one or more rules for critical discussion have been violated because the soundness conditions have not been met that apply to that type of strategic maneuvering in that argumentative activity type.6 By “argumentative activity types” we mean verbal routines in the Gumperz’ sense that are more or less institutionalized in empirical reality and characterized by specific goals and conventions (van Eemeren & Houtlosser 2005). What, then, do the differences between our approach and Walton’s (and Walton and Krabbe’s) amount to? To begin with, Walton and Krabbe’s dialogue types are neither equivalent with our normative notion of a critical discussion nor with our empirical notion of an argumentative activity type. We make, in fact, a radical distinction between, on the one hand, the model of a critical discussion, which represents an analytic ideal, and, on the other hand, empirical argumentative activity types, which are to be found in the reality of argumentative praxis. In Walton and Krabbe’s dialogue types this distinction is blurred. Although we agree with Walton and Krabbe that fallacy judgments are in the end always contextual judgments that depend on the specific circumstances of situated argumentative acting, we do not agree that the norms underlying these judgments are context-dependent. In our view, the norms expressed in the rules for critical discussion are general—who knows even universal—norms for sound argumentation that are not limited to one particular type of argumentative activity —or “dialogue type.” It is true that argumentative discourse always takes part in a certain context of argumentative activity, but this does not mean that the soundness norms are automatically relative. The context-dependency of judgments of argumentative discourse lies in the way in which the conduct of argumentative discourse is conventionally disciplined in a certain activity type by specific criteria for determining whether or not a certain type of maneuvering agrees with the relevant norm, which criteria may vary to some extent per argumentative activity type— in a law case, for instance, different criteria apply to making a legitimate appeal to authority, e.g. by referring to a certain law code, than in a political debate. It may be one of Walton and Krabbe’s basic assumptions that “the critical discussion [...] is the most fundamental context of dialogue needed as a normative structure in which fallacies and other errors of reasoning can be analyzed and evaluated” (1995: 7), and this observation may seem to involve a recognition of the special position a critical discussion has in the study of argumentative discourse, but their concept of a critical discussion is different from the pragma-dialectical concept and lacks the special function the concept has in the pragma-dialectical The Contextuality of Fallacies 65 theorizing.7 The concept of a critical discussion is, of course, also crucial to us, and in a more fundamental way, but we do not consider it as belonging to the same category of “dialogue types” such as a negotiation and a quarrel, as Walton and Krabbe do, nor do we put it on the same (empirical) level as the argumentative activity types. The difference between Walton and Krabbe’s approach and ours is that between “a good argument is one that contributes to the specific goal of a type of dialogue” (Walton and Krabbe) versus “a good argument is one that complies with the general rules of critical discussion” (van Eemeren and Houtlosser). Using the rules for critical discussion as a context-independent standard, we take the peculiarities of the various argumentative activity types into account when we start evaluating whether these rules have been obeyed or violated. In our view, a clear understanding of the design of the various activity types, and the prevailing criteria for deciding whether a rule for critical discussion has been violated, is therefore an important step towards determining whether a certain type of strategic maneuvering has got derailed. Instead of mixing the normative and the empirical prematurely, in our approach we first separate these two dimensions, dealing with each dimension in its own right. Subsequently, we bring the two dimensions together in the reconstruction of argumentative discourse, making use of insight into the peculiarities of the various argumentative activity types in order to be able to apply the criteria that are needed to ensure a sound evaluation of the discourse in a critical sense. Thus, we think, we are in a better position than Walton to reconcile the real and the ideal. Notes 1 It may be no coincidence that, when Blom and Gumperz investigate the relation between a standard and a local dialect, they also observe “cases of shift from one form of speech to another” (1972: 30). 2 A few examples: “Various traditional fallacies can be viewed from the perspective of dialectical shifts. […] The argumentum ad populum […] is associated with a shift from persuasion to epidictic dialogue […]. The argumentum ad baculum […] is associated with shifts from persuasion dialogue to negotiation dialogue, or in some cases, even to quarrelling. […] Use of expert opinion […] can be perfectly legitimate and constructive, but it characteristically becomes fallacious when a participant in a persuasion dialogue illicitly shifts to an information-seeking dialogue […] (argumentum ad verecundiam). In general, a key element in judging argumentation as fallacious or non-fallacious, in a given case, is the context of dialogue surrounding the argument, especially if a dialectical shift has occurred during the cause of the argument” (Walton 1992: 121). 3 Walton’s (1992: 138) example of Karen’s legitimate shift from a critical discussion to an “action-directed dialogue,” when she and Doug are cycling near Sassenheim, seems to suggest that a unilateral shift can also be legitimate, so that being bilateral cannot be a necesssary condition. 4 Apart from the fact that it does not seem helpful to include the concept of a unilateral dialectical shift in the analysis of informal fallacies, including it also brings a danger with it: someone who is accused of having made a move that is fallacious in the current dialogue always has the option to act as if this move was not meant as part of this dialogue but as a proposal to 66 Frans H. van Eemeren & Peter Houtlosser start another type of dialogue. After all, “once a fallacy has been committed, it may not be fallacious from the norms of the new type of dialogue one has entered” (Walton & Krabbe 1995: 115). 5 An additional aspect of the lack of clarity we perceive in Walton’s shift approach is that it is not adequately explained whether a move is fallacious when there is a shift in flavor, when elements from other dialogue types intrude the original dialogue. This query applies all the more when the shift is merely internal, i.e., when only the subject matter or setting of the original dialogue changes. It may even happen that a move that is fallacious in a particular dialogue type still has a quality that makes it “less” fallacious than may seem to be the case at first sight: “In the eristic discussion, an outrageous fallacy may be the best and most successful technique to persuade your audience” (Walton & Krabbe 1995: 79). 6 A basic commonality between our approach to the fallacies and Walton’s is that both aim for an understanding of why fallacies can go so easily unnoticed in practice. Where Walton’s explanation seems to be that they involve dialectical shifts that are not perceived by the uncritical respondent or observer (1992: 146), so that they are obviously not bilateral, our explanation is that they are often deviations of rules for critical discussion that are hard to detect because there is a prevailing assumption of reasonableness and these fallacious moves involve derailments of strategic maneuvering in which a use of argumentative means that is known to be reasonable is “stretched” too far. What exactly “too far” means in a particular case must be decided in the context of the activity type concerned. 7 The confusion is well-expressed in the following quote: “One important type of dialogue is the critical discussion, well described by van Eemeren and Grootendorst (1984), which is a type of persuasion dialogue” (Walton & Krabbe 1995: 133). References Blom, J-P., and J.J. Gumperz (1972). Social meaning in linguistics structure: Code-switching in Norway. In: J.J. Gumperz and D. Hymes (Eds.), Directions in Sociolinguistics: The Ethnography of Communication (pp. 407-434). New York etc.: Holt, Rinehart and Winston. Eemeren, F.H. van, and R. Grootendorst (1984). Speech Acts in Argumentative Discussions: A Theoretical Model for the Analysis of Discussions Directed Towards Solving Conflicts of Opinion. Dordrecht/Berlin: Foris/Walter de Gruyter. Eemeren, F.H. van, and R. Grootendorst (2004). A Systematic Theory of Argumentation: The Pragma-Dialectical Approach. Cambridge: Cambridge University Press. Eemeren, F.H. van, and P. Houtlosser (2002). Strategic maneuvering: Maintaining a delicate balance. In: F.H. van Eemeren and P. Houtlosser (Eds.), Dialectic and Rhetoric: The Warp and Woof of Argumentation Analysis (pp. 131-159). Dordrecht: Kluwer Academic. Eemeren, F.H. van, and P. Houtlosser (2005). Theoretical construction and argumentative reality: An analytic model of critical discussion and conventionalized types of argumentative activity. In: D. Hitchcock (Ed.), The Uses of Argument: Proceedings of a Conference at McMaster University, 18-21 May 2005, Hamilton (pp. 75-84). St. Catharines, ON: OSSA. Eemeren, F.H. van, P. Houtlosser and A.F. Snoeck Henkemans (2005). Argumentatieve indicatoren in het Nederlands: Een pragma-dialectische studie. Amsterdam: Rozenberg Publishers. The Contextuality of Fallacies 67 Eemeren, F.H. van, P. Houtlosser and A.F. Snoeck Henkemans (2007). Argumentative Indicators in Discourse: A Pragma-Dialectical Study. Dordrecht: Springer. Gumperz, J.J. (1972). Introduction. In: J.J. Gumperz and D. Hymes (Eds.), Directions in Sociolinguistics: The Ethnography of Communication (pp. 1-25). New York etc.: Holt, Rinehart and Winston. Krabbe, E.C.W. (1992). So what? Profiles for relevance criticism in persuasion dialogues. Argumentation 6, 271-283. Walton, D.N. (1992). Types of dialogue, dialectical shifts and fallacies. In: F.H. van Eemeren, R. Grootendorst, J.A. Blair, and Ch.A. Willard (Eds.), Argumentation Illuminated (pp. 133-147). Amsterdam: Sic Sat. Walton, D.N., and E.C.W. Krabbe (1995). Commitment in Dialogue: Basic Concepts of Interpersonal Reasoning. Albany, N.Y.: State University of New York Press. Frans H. van Eemeren Peter Houtlosser Department of Speech Communication, Argumentation Theory and Rhetoric University of Amsterdam Spuistraat 134 1012 VB Amsterdam The Netherlands F.H.vanEemeren@uva.nl P.Houtlosser@uva.nl