Pragma-dialectics and Beyond

Transcription

1 Pragma-dialectics and Beyond DANIEL BONEVAC Department of Philosophy University of Texas at Austin U.S.A. ABSTRACT: Pragma-dialectics is dynamic, context-sensitive, and multi-agent; it promises theories of fallacy and argumentative structure. But pragma-dialectic theory and practice are not yet fully in harmony. Key definitions of the theory fall short of explicating the analyses that pragma-dialecticians actually do. Many discussions involve more than two participants with different and mutually incompatible standpoints. Success in such a discussion may be more than success against each opponent. Pragma-dialectics does well at analyzing arguments advanced by one party, directed at another party; it does much less well at analyzing arguments directed at several opponents at once or at convincing an audience. I suggest a strategy of construing fallacies as defeasible arguments relying on reasonable default principles but applying them in circumstances in which they are undercut or overridden. KEY WORDS: audience, defeasibility, fallacies, pragma-dialectics Pragma-dialectics, the theory of argumentation developed by Frans van Eemeren, Rob Grootendorst, M. Agnès van Rees, and others, has a number of advantages over other approaches to argumentation. Unlike most theories, it is: 1. dynamic. In the first half of the twentieth century, approaches to natural language generally, and argumentation more specifically, idealized the structure of arguments. Taking mathematical proofs as paradigms of successful argumentation, logical theories abstracted away from pragmatic elements of language entirely. Beginning in the 1940s, philosophers such as P. F. Strawson and J. L. Austin focused on pragmatic elements. Their work had little influence, however, on the increasingly formal semantic theories developed by Richard Montague and his followers. At best, as in the work of David Kaplan, pragmatic elements appeared as indices to which semantic values were relative. Since the work of Lauri Karttunen, Hans Kamp, and Irene Heim, we have seen that a dynamic theory of language can integrate pragmatic and semantic components into a unified theory. Pragma-dialectics fits nicely within such a dynamic framework. 2. context-sensitive. Any theory capable of accounting for pragmatic features of language must be sensitive to context. Dynamic theories build this in from the beginning; the meaning of a sentence as used in a context is understood in terms of the context that results from the assertion of Argumentation 17: , Kluwer Academic Publishers. Printed in the Netherlands.

2 452 DANIEL BONEVAC the sentence. The meaning of a sentence in general is understood as a mapping from contexts to contexts. Pragma-dialectics, while not advancing any general theory of meaning, fits well with such a theory by emphasizing the role of context in argument analysis. 3. multi-agent. Logic traditionally evaluates arguments as sound or unsound, valid or invalid. Sometimes, as in discussions of fallacies, the argumentative context enters the picture. But arguments are conceived as advanced by a faceless arguer toward a faceless audience. Rhetoric takes the roles of arguer and audience more seriously. Pragma-dialectics extends to dialogue a discussion or argument between two parties while analyzing a wide array of strategic moves. 4. a theory of fallacy and argumentative structure. Accounts of fallacies in logic and rhetoric hardly deserve to be called theories as they stand. They consist mostly of observations and generalizations which, however well justified or insightful they might be when taken individually, yield no overall theory that explains what either fallacies or nonfallacious arguments have in common. Pragma-dialectics, in contrast, takes the construction of a theory of fallacy as an explicit goal. Whether it can provide such a theory remains to be seen. But the recognition that a theory of fallacy is both lacking and desirable is itself a major step in the right direction. DOES PRAGMA-DIALECTICS GO FAR ENOUGH? My aim in this paper is to suggest some further fruitful directions in which pragma-dialectics might develop. To understand the points I am about to make, it is important to distinguish between the theory and practice of argument analysis. Pragma-dialectics consists both of highly sensitive analyses of particular arguments and a theoretical structure that underwrites those analyses. To put my criticism of pragma-dialectics in a nutshell: Pragma-dialectic theory and practice are not yet fully in harmony. Key definitions of the theory, especially as formulated by van Eemeren and Grootendorst, fall short of explicating the analyses that they, van Rees, and others actually do. The rules explicitly stated in the theory fall short of what is needed to underwrite analyses currently being performed especially van Rees s analyses of critical discussions not to mention others in different kinds of contexts. I hope to sketch some ways in which theory and practice might be brought together more closely. DIALOGUE DISCUSSION Pragma-dialectics advances significantly beyond typical logical or rhetorical approaches to the study of argument by recognizing that argumenta-

3 PRAGMA-DIALECTICS AND BEYOND 453 tive discourses are often multi-agent. Discussions often take place between more than two parties. Each party to the discussion, moreover, generally has a distinct position. In political contexts, in particular, there tend to be many discussants with many distinct positions; discourses are hard to understand without recognizing appeals to various groups intended to build coalitions, head off potential objections from different quarters, and so on. Former President Bill Clinton, for example, famously maneuvered by triangulation, seeking to distance himself from both Democrats and Republicans in Congress. His argumentative strategies cannot be understood without recognizing that he was implicitly or explicitly criticizing two different opponents and seeking to appeal to groups of voters positioned between them on the political spectrum. Even in the simplest cases, then with just one group of target voters an analysis of Clinton s rhetoric must involve at least four participants. Politicians in European countries typically face even more complex rhetorical tasks, criticizing competitors from various parties inside and outside government and appealing to groups of voters affiliated with those parties. A nice example of a nonpolitical discussion that is essentially multiagent is Arend Heyting s Disputation, in which he explains and defends basic principles of intuitionistic logic and mathematics. The disputation takes place among several participants espousing various positions in the philosophy of mathematics. Those appearing below are Int, an intuitionist; Class, a classical mathematician; Form, a formalist; Sign, a semioticist; and Prag, a pragmatist. Each has a distinct position. This passage reveals some of the flavor of the dispute: Int: Class: Sign: Class: Sign: Form:... intuitionism proceeds independently of the formalization, which can but follow after the mathematical construction. What puzzles me most is that you both [the intuitionist and the formalist] seem to start from nothing at all. You seem to be building castles in the air. How can you know if your reasoning is sound if you do not have at your disposal the infallible criterion given by logic? Yesterday I talked with Sign, who is still more of a relativist than either of you. He is so slippery that no argument gets hold of him, and he never comes to any somewhat solid conclusion. I fear this fate for anybody who discards the support of logic, that is, of common sense. Speak of the devil and his imp appears. Were you speaking ill of me? I alluded to yesterday s discussion. To-day I am attacking these other two damned relativists. I should like to join you in that job, but first let us hear the reply of your opponents. Please meet my friend Prag; he will be interested in the discussion. How do you do? Are you also a philosopher of science?

4 454 DANIEL BONEVAC Prag: Int: I hate metaphysics. Welcome, brother! Each discussant has a distinct position; arguments are often advanced against several opponents at once. Now I fully recognize that pragma-dialectics means to take such multiagent discourses into account. My question, however, is whether the theoretical definitions adopted in the theory are flexible enough to permit that. Argumentation is defined as involving two agents. Consider: According to pragma-dialectical theory, argumentation is always part of an explicit or implicit dialogue in which one party attempts to convince the other party of the acceptability of his standpoint.... The protagonist s argumentation is seen as a complex whole made up of statements put forward to deal with real or anticipated critical reactions from an antagonist. (Henkemans, 1994, p. 69) This is typical: What is pragma-dialectical about such an analysis? The dialectical aspect consists in the assumption that there are two parties who attempt to resolve a difference of opinion by means of a systematic exchange of moves in a discussion. (van Eemeren and Grootendorst, 1990, p. 154) The assumption that a two-party analysis suffices for argumentative discourse is built into the theory in a variety of ways. Van Eemeren and Grootendorst, for example, in detailing the correctness conditions for argumentation, characterize justification in terms of a duality of argumentative roles; to justify p, they assert, is to convince the listener of the acceptability of his standpoint with respect to p (van Eemeren and Grootendorst, 1992, p. 31). They outline preparatory conditions in terms of the speaker and the listener. In general, they proclaim, In argumentative discourse there are, in principle, always two parties involved (13n). The assumption of duality is not confined solely to argumentative discourse. In critical discussion, say van Eemeren and Grootendorst, One party takes the role of protagonist, which means that he is prepared to defend his standpoint by means of argumentation; the other party takes the role of antagonist, which means that he is prepared to challenge the protagonist systematically to defend his standpoint. (1992, p. 35) Similar points are made by van Eemeren, Grootendorst, Jackson, and Jacobs (1993, p. 26) and van Rees (2001, p. 464). Even in critical discussion, then, the analysis proceeds by isolating argumentative discourses which are in turn to be analyzed dualistically. The ideal for resolving disputes is essentially dualistic. This works in many cases. Even seemingly monological arguments can be seen as dialogues with a skeptic or evil demon. But many discussions, as I have noted, involve more than two participants with different and mutually incompatible standpoints. Why does this matter? Success in such a discussion may be more than success against each opponent. Someone

5 PRAGMA-DIALECTICS AND BEYOND 455 seeking to defend a position against a variety of opponents at once, for example, must meet a number of constraints that cannot be understood as conjunctions of constraints applied to each dispute taken individually. For example: 1. Consistency: The arguments offered must cohere with one another. One cannot argue that p against one discussant and not p against another. 2. Issues: Disagreements among various disputants may raise issues that a protagonist would not otherwise have had to confront, tackling each oneby-one. 3. Individuation: How do we tell which counter-argument addresses which disputant? a. This complicates the coordinative/multiple argumentation distinction: a counter-argument might serve to make another argument more acceptable to one disputant without doing so to other disputants. b. It also complicates analysis of counterarguments: Against how many disputants, for example, does Thomas Aquinas argue in the Summa Theologiae? Against how many disputants is a specific remark in Heyting s Disputation directed? c. It complicates the application of rules of relevance: an argument may be relevant to issues raised by one disputant but not to those raised by others. d. It complicates analysis of the common ground: there may be little or nothing that all participants share, while each pair shows broad agreement. e. Participants may make metamoves: This is going nowhere! We re repeating ourselves. He s hopeless. I don t want to talk about that now. These may not be fallacious, but rather attempts to regulate discussion entry and exit. The set of disputants, that is, may itself shift during the discussion; the disputants are partly (and rightly) responsible for this. WHAT IS THE GOAL OF ARGUMENTATION? Pragma-dialectics maintains, in Henkemans s words, that one party attempts to convince the other party of the acceptability of his standpoint. Complex discussions are more complicated, but still fall into moves by a protagonist and counter-moves by an antagonist. But think again about political discourse: Clinton s triangulation, the Shell ad analyzed by van Eemeren, or any election campaign. The goal is not to convince one s opponent or opponents but to convince other parties. In a discussion between A, B, and C, say, the important question often is not What should the group {A, B, C} conclude? but rather What should their audience conclude? In short, pragma-dialectics does well at analyzing arguments advanced by one party, directed at another party; it does much

6 456 DANIEL BONEVAC less well at combining such considerations with traditional rhetorical concerns about audience. To put it another way, goals of resolving disputes and settling disputes are often intertwined in complex ways. In such circumstances, neither pragma-dialectics nor rhetoric can provide a comprehensive analysis. The solution, I maintain, is to think of an argumentative discourse and a critical discussion not as a dialogue between a protagonist and an antagonist, but as a discourse with a set of participants (the disputants) and a set of observers (the audience). The participants have standpoints which they seek to advance and defend; the audience members have standpoints, but no obligation to advance or defend them. The participants may try to resolve disputes with fellow participants but also to settle disputes by seeking the approval of observers. Different discussions will give participants different incentives about resolving or settling disputes. In a formal debate setting, for example, or in a court of law, there may be little hope of resolving anything with the opposing team; the goal is to get the observer, the judge or jury, on one s side. In a political campaign, the goal is not to convince the opponent but to win approval from the voters. In various kinds of group discussions, there may be attempts to seek common ground with other participants, resolve disputes with them, and sway observers to one s standpoint, in a complex mixture. If that is right, there are some important implications: 1. Participants try to position themselves not only with respect to the issues but also with respect to the other participants. They construct standpoints to be maximally defensible given the set of disputants and the intended audience which must not in general be identified with that set. 2. Participants must take into account not just assumptions they share with other participants, and issues dividing them from other participants, but assumptions they share with the audience and issues dividing them from the audience. To an audience, credibility is often a critical question. Participants often make seemingly fallacious moves in order to enhance their own credibility or undermine the credibility of their opponents. Whether this is fallacious depends, in large part, on whether someone has advanced an argument based on premises in the common ground of the discussion. Credibility is reasonably at stake if someone s testimony is at issue, but not otherwise. In the words of Samuel Johnson, Nay, Sir, argument is argument. You cannot help paying regard to their arguments if they are good. If it were testimony, you might disregard it.... Testimony is like an arrow shot from a long bow; the force of it depends on the strength of the hand that draws it. Argument is like an arrow shot from a cross bow, which has equal force though shot by a child. (Quoted in Fischer, 282) But distinguishing the evaluation of argument from the evaluation of testimony is not easy.

7 PRAGMA-DIALECTICS AND BEYOND We must therefore rethink relevance. Things irrelevant to the goal of resolving an issue with other disputants may be relevant to resolving that issue with the audience, or vice versa. One person s red herring may be another person s crucial distinction. 4. The goal of argumentation is often not getting the other parties to agree but displaying your disagreements with them to your advantage that is, in a way that would rationally induce an audience member to agree. These are of course not exclusive; one may hope to get agreement from disputants and audience alike on certain matters, but hope at most to sway audience members to your standpoint on certain others. DEFEASIBILITY AND FALLACIES Even with the help of the theoretical apparatus of pragma-dialectics, it is easy to get the sense that fallacies are in the eye of the beholder: A fallacy is an argumentative move you don t like. Analyses of political arguments, such as the analysis of the Shell ad, often betray the analyst s view on the political issues in question. My point is not that analysts are biased, though that may sometimes be true. It is, rather, that in the absence of an adequate and reasonably precise theory argument analysis is unlikely to produce anything but further argument on the issues in question. Pragma-dialectics has yielded interesting insights into a variety of fallacies. What it has not yielded and, to be fair, what no other theory has yielded is a comprehensive and precise account of what fallacious reasoning is. Van Eemeren refers to fallacies as specific obstacles that can be impediments to resolving disputes. While helpful in some ways, this is surely too broad; one party s hardness of hearing or stubbornness may impede dispute resolution, but these hardly count as fallacies. Van Eemeren and Grootendorst (1992, pp ) clearly intend this in a more restrictive way, however; a fallacy is a violation of the pragma-dialectical rules for dispute resolution. How comprehensive and precise this definition is thus clearly depends on the comprehensiveness and precision of the pragma-dialectical rules. If those rules lack unity, furthermore, it may be hard to see what fallacies have in common what fallacious reasoning per se is. Evaluating the pragma-dialectical rules in particular, their comprehensiveness lies beyond the scope of this paper. But I would like to advance a proposal intended to unify at least much of the discussion of fallacies. I think we can understand argument structure more readily, and devise a better account of fallacies, if we realize that reasoning is often not deductive but defeasible. An opponent may object to an argument, not only by objecting to a premise or objecting to the argument s validity (the sufficiency of the

8 458 DANIEL BONEVAC premises) the two grounds recognized by the pragma-dialecticians but also by 1. Suggesting undercutting considerations: considerations that do not challenge the truth of the premises or the acceptability of the argument, but nevertheless constitute reasons for refusing to accept the conclusion. Suppose, to take a very simple example, that someone argues that Tweety flies on the grounds that Tweety is a bird and birds fly. If an opponent observes that Tweety injured her wing a while ago, we may withhold affirmation of the conclusion that Tweety flies. But we do not thereby deny that birds fly or that Tweety is a bird. 2. Suggesting overriding considerations: considerations that do not challenge the truth of the premises or the acceptability of the argument but nevertheless constitute reasons for denying the conclusion. If someone argues that Tweety flies, as above, and an opponent observes that Tweety is a penguin, we have reason to conclude that Tweety does not fly. Are responses to such arguments coordinative or multiple? This may vary with the case. But we might think of coordinative argumentative structures as, in part, intended to stave off such considerations. Once we realize that real-world arguments are often defeasible, allowing for the possibility of undercutting or overriding considerations, we are in a position to see something that fallacies might have in common. I offer the following bold (that is, probably false, but, I hope, containing some grain of truth) conjecture: Fallacies fall into two groups: fallacies of ambiguity and fallacies of misapplication. The latter involve applying a perfectly legitimate default principle in the presence of undercutting or overriding considerations. Here are some examples of fallacies and default principles: 1. appeal to pity ( Don t hurt people ) 2. appeal to force ( Do what s in your own self-interest ) 3. ad hominem abusive ( Don t trust knaves or fools ) 4. ad hominem circumstantial ( Don t trust people with vested interests ) 5. composition ( Judge wholes by their parts ) 6. division ( Judge parts by their wholes ) 7. appeal to the people ( Trust the judgments of your peers ) 8. appeal to authority ( Trust people you respect ) There is nothing wrong with these principles, in general. That is why allegedly fallacious arguments often persuade. The question is whether these considerations are undercut or overridden in particular cases. Consider, for example, the fallacy of composition. The fallacy, as generally understood, involves attributing a property to a whole on the basis of its application to the parts. A pin dropping makes no sound. We cannot

9 PRAGMA-DIALECTICS AND BEYOND 459 infer, however, that a thousand pins dropping makes no sound. So, the form of the argument seems to be: The (proper) parts of X have property P. So, X has P. Plainly, this is not valid for all X and P. All the proper parts of X, for instance, are proper parts of X; it does not follow that X is a proper part of X. Nevertheless, we routinely accept arguments of this form. We evaluate academic departments by evaluating the faculty members in them. We evaluate universities by evaluating their departments. We evaluate sports teams by evaluating their players. All other things being equal, if team A has excellent players, A is an excellent team. There is nothing wrong with the form of argument as such. It is, however, defeasible; the conclusion that A is excellent may be undercut or overridden by additional information. So, if we learn that the players on A do not get along well together, that A has a terrible coach, or that A has lost its last eight games, we may withdraw the conclusion. If this is right, it has an important consequence, which also explains why argument analyses so often appear to be not so much objective scientific analyses but, to a large extent, further contributions to the discussion of the issues at hand. The evaluation of an argument as fallacious is not really the business of a logician or even pragma-dialectician; it is part of the subject matter at hand. Whether dissension, poor coaching, or a losing streak should lead us to withdraw our assent to A s excellence depends on many things, including the sport in question. (Individual excellence outweighs most other things in baseball, but counts for less in basketball, and even less in football.) What counts as undercutting or overriding the above rules in particular argumentative contexts is context- and content-driven. There can be no general theory of that. REFERENCES Fischer, D. H.: 1970, Historians Fallacies, Harper and Row, New York. Henkemans, A. F. S.: 1994, Complex Argumentation in a Critical Discussion, in van F. H. Eemeren and R. Grootendorst (ed.), Studies in Pragma-dialectics, Vale Press, Newport News. Heyting, A.: 1951, Disputation, in Intuitionism, North-Holland, Amsterdam. Van Eemeren, F. H. and R. Grootendorst: 1990, Analyzing Argumentative Discourse, in R. Trapp and J. Schuetz (eds.), Perspectives on Argumentation: Essays in Honor of Wayne Brockriede, Waveland Press, Prospect Heights, IL. Van Eemeren, F. H. and R. Grootendorst: 1992, Argumentation, Communication, and Fallacies, Lawrence Erlbaum, Hillsdale, NJ. Van Eemeren, F. H., R. Grootendorst, S. Jackson and S. Jacobs: 1993, Reconstructing Argumentative Discourse, University of Alabama Press, Tuscaloosa, Ala., and London. Van Rees, M. A.: 2001, The Diagnostic Power of the Stages of Critical Discussion in the Analysis and Evaluation of Problem-Solving Discussions, Argumentation 15,

10