Advances in Pragma-Dialectics

Frans H. van Eemeren, editor

Amsterdam: Sic Sat / Newport News, VA: Vale Press, 2002. pp. viii, 1-340. ISBN 90-74049-06-0. Paperbound: USD 34.95.


Reviewed by David Hitchcock

Reviewed by David HitchcockReviewed by David Hitchcock

Pragma-dialectics is the approach to the study of argumentation developed by Frans van Eemeren and the late Rob Grootendorst of the department of speech communication, argumentation theory and rhetoric at the University of Amsterdam. According to this approach, canonically articulated in English in three successive works (1984, 1992, 2004), argumentation is a complex speech act, with the usual four types of “felicity conditions” (Searle, 1969): propositional content and essential conditions (together called identity conditions), and preparatory and sincerity conditions (together called correctness conditions). Since speech act theory belongs to pragmatics (the study of the uses of language), the construal of argumentation as a speech act explains the “pragma” in the label “pragma-dialectics”.

The “dialectics” part of the label comes from the treatment of the speech act of argumentation as occurring in a conversation between two people. Ideally the conversation proceeds in four stages: a confrontation stage in which the two people verbalize a difference of opinion, an opening stage in which they agree on procedural and substantive starting points for resolving this difference of opinion, an argumentation stage in which argumentation is advanced and responded to, and a concluding stage in which the parties decide jointly whether and how their difference of opinion has been resolved. In the simplest case one party expresses doubt about an assertion by the other party. It is then up to this latter party, the “protagonist”, to justify the assertion to the satisfaction of the other party, the “antagonist”, using the starting points agreed to at the opening stage.

Van Eemeren and Grootendorst (1992) propose ten rules for the conduct of this type of conversation, called a “critical discussion”. A fallacy is a violation of one of the ten rules. Generically, such fallacies are moves which disrupt or “derail” the process of rationally resolving an expressed difference of opinion.

Pragma-dialecticians apply their model not only to explicit back-and-forth argumentative discussions but also to monological argumentation, which is analyzed as if it were a contribution to a critical discussion, with the author anticipating the responses of a rational critic.

The volume under review, a successor to (van Eemeren & Grootendorst, 1994), brings together 20 papers, eight by members of the Amsterdam department where pragma-dialectics originated, ten by speech communication theorists in the United States, and two by philosophers (Leo Groarke and Erik Krabbe). The editor arranges the papers in groups corresponding to five themes, for which I will use the italicized word as a title: comparison of pragma-dialectical norms with “naïve” and other norms, enlarging the scope of pragma-dialectics, indicators in discourse of argumentative commitments, argumentative discourse and personal interaction, argumentation in institutional contexts. This review considers the papers according to a different plan of organization, which is likely to be more perspicuous to readers of this journal: current research topics in the Amsterdam group, application of pragma-dialectics to specific types of discussions or arguments, non-ideal higher-order conditions for critical discussion, profiles of dialogue, justification and persuasiveness.

1. Current research topics in the Amsterdam group

Strategic manoeuvring: Recently van Eemeren and Houtlosser (1999, 2001) have introduced rhetorical considerations into the dialectical framework of pragma-dialectics. Participants in argumentative discussions can use various devices, collectively called “strategic manoeuvring”, to make it more likely that the discussion will conclude with a victory for their side. Such manoeuvres may or may not violate a rule of critical discussion. In the first chapter of the volume under review, included in the comparison group, van Eemeren and Houtlosser describe ways in which each party in a critical discussion can at each stage manoeuvre to alter the burden of proof so as to favour their side. For example, in the opening stage the protagonist can highlight those concessions of the antagonist which provide the strongest basis for defending the protagonist’s standpoint. Or in the argumentation stage the antagonist can repeat a critical question relevant to the protagonist’s argumentation scheme, even though the protagonist has already answered it. This latter manoeuvre is a fallacy, whereas the former one is not. The chapter also includes a comparison of the pragma-dialectical position on the burden of proof in argumentative discussion with positions on this issue articulated by other authors.

Since it remains subordinated to the critical rationalism which undergirds the rules for critical discussion, the pragma-dialectical accommodation of victory-oriented rhetorical manoeuvres will not satisfy those theorists of argumentation who construe argumentation as fundamentally a persuasive activity, to be judged by its effectiveness. For those who construe argumentation as fundamentally a matter of justification or proof, on the other hand, the addition of the concept of strategic manoeuvring makes pragma-dialectics empirically more realistic and more helpful, both analytically and practically. It provides scope for further analytical, empirical and practically oriented research.

Empirical studies of the conventional validity of pragma-dialectical norms and taxonomy: Pragma-dialectical researchers have begun to investigate empirically ordinary arguers’ standards for reasonableness. In the third chapter, included in the comparison group, van Eemeren and Bert Meuffels report on a test of ordinary (i.e. untrained) arguers’ judgments on ad hominem fallacies. In pragma-dialectics, such fallacies violate discussion rule 1, the so-called “confrontation rule”: “Parties must not prevent each other from advancing standpoints or casting doubt on standpoints.” (47, n. 4) That is, ad hominem fallacies (whether abusive, circumstantial or tu quoque) are treated as attempts to disqualify their target from participation in the discussion, rather than as attempts to show that the target’s conclusion is false. This analysis enables pragma-dialectics to draw a clear line between ad hominem fallacies and relevant personal attacks. Van Eemeren and Meuffels found that the 92 Dutch secondary school students in their sample on average rated fallacious personal attacks in short dialogues as much less reasonable than non-fallacious ones (3.75 ± .46 compared to 5.29 ± .64, on a 7-point scale with 1 as very unreasonable and 7 as very reasonable). On average, they considered the abusive ad hominem most unreasonable, the circumstantial variant next most unreasonable, and the tu quoque least unreasonable. And on average they considered ad hominem fallacies more unreasonable in a scientific discussion than in a political or domestic discussion. These results demonstrate the conventional validity (i.e. conformity to conventional judgments of reasonableness) of the standard textbook taxonomy of ad hominem fallacies, which pragma-dialectics appropriates.

In the sixth chapter, also included in the comparison group, Bart Garssen reports the results of two empirical studies of ordinary arguers’ recognition of the three argumentation schemes which pragma-dialectics acknowledges: symptomatic argumentation from a symptom to its concomitant or vice versa, comparison argumentation from one thing to another which resembles it, and causal argumentation from a cause to its effect or vice versa. Respondents in both studies were much better at recognizing the comparison scheme than the other two. Garssen explains the discrepancy by the highly abstract nature of the argumentation schemes. An alternative explanation is that the symptomatic and causal schemes are overly general, and lump together quite different sorts of arguments.

Indicators of argumentative commitments: The Amsterdam pragma-dialecticians are also investigating discoursal indicators of argumentative commitments. In chapter 11, part of the indicators group, Peter Houtlosser uses the pragma-dialectical approach to good advantage in identifying verbal indications that an author is advancing a point of view, i.e. a standpoint which the author is committed to defend in a critical discussion. Advancing a point of view is construed as a kind of speech act, whose essential condition is the assumption of an obligation to defend a formulated opinion if asked to do so. The preparatory conditions for such an act are a belief that the addressee does not already accept the opinion and a belief that the author can justify it to the addressee with the help of arguments. Its sincerity conditions are a belief that the opinion is correct and an intention to justify it to the addressee if asked to do so. Houtlosser makes the interesting and perhaps surprising point that, in cases where the author does not explicitly claim to be advancing a point of view, the most useful clue that the author is in fact advancing a point of view is not the identity condition for this speech act but its first preparatory condition: the presumed doubt of the addressee.

Three types of textual clues indicate such a presumption of doubt.

First, the speaker can indicate anticipation of the addressee’s doubt by expressions like “I believe that” or “I think that” or “I am sure that” which indicate the author’s propositional attitude. Such expressions, Houtlosser claims, indicate anticipation of doubt in all cases except where speakers are describing their own state of mind in response to a specific request to describe it, or where they make a concession or express agreement in response to an interlocutor’s previous endorsement of the embedded proposition.

Second, an interlocutor may express doubt in response to a speaker’s assertive. Any assertive speech act becomes a point of view once an interlocutor has expressed doubt about its acceptability. Since the basic type of assertive has four correctness conditions (Searle & Vanderveken 1985), such doubt can be expressed in four different ways: explicit expression of doubt, request for evidence, denial of a presupposition, denial of an implication. In connection with the last-mentioned method of expressing doubt, Houtlosser somewhat controversially claims that the author of an assertive expresses a commitment to its strict implications, so that denying any of these would be a way of expressing doubt about the assertive’s acceptability. But any logical falsity strictly implies anything, and any proposition strictly implies any logical truth. Hence, on Houtlosser’s account, one could express a doubt about someone’s denial of the law of non-contradiction (which strictly implies that December 25 is not Christmas Day) by saying “December 25 is Christmas Day” and one could express a doubt about someone’s claim that Christmas Day is December 25 (which strictly implies the logical truth that it is not both raining and not raining here right now) by saying, “It is both raining and not raining here right now.” The implausibility of such analyses indicates the need for a more restrictive conception of the implications to which an assertive speech act commits its author.

Third, a speaker may follow up an assertive by supporting argumentation, which presupposes doubt that the addressee already accepts the assertive. Here it becomes necessary to distinguish an argument for the assertive from a causal explanation. For this purpose, two conditions distinguishing the speech acts of arguing and explaining are particularly helpful. First, only factual assertives can be explained, whereas any assertive can be argued for. (In fact, one can strengthen this point; only factual assertives which are not predictions can be explained.) Second, explaining presupposes an accepted explanandum, whereas arguing presupposes an unaccepted assertive. (This second difference is of course no help if one is still trying to decide if the follow-up indicates doubt that the addressee accepts the preceding assertive.)

If the word “because” introduces the speaker’s follow-up to a factual assertive, three possibilities need to be considered.

First, if this follow-up describes a consequence of the previously asserted situation rather than a cause of it (“She must be very brave, because she never complains”), then it is a justification of the assertive, which is therefore an advancement of a standpoint.

Second, if the follow-up describes a cause, it may be either a causal argument or a causal explanation. Houtlosser claims that, if it is a causal argument, it justifies an action or mental state (e.g. emotion or belief) of the speaker, and gives as an example the sentence “I could not take pictures, because my camera didn’t work”. The interpretation of this example as a causal argument seems to depend on interpreting it as offered in some imagined context as an excuse or justification for the speaker’s failure to take pictures, rather than as evidence for the claim that the speaker was unable to take pictures. It is an indirect justification of the reasonableness of the speaker’s (in)action, by way of providing directly a causal explanation of the already accepted fact of this (in)action. Similarly with the general thesis that causal arguments of the form “p because q”, where q is represented as a cause of p, justify an action or mental state of the speaker. Consider the causal argument, “The ice cube has melted, because it was put into a sink full of hot water five minutes ago.” Suppose that neither speaker nor addressee has observed whether the ice cube has melted, but the speaker knows directly how long ago it was put into the sink and the addressee accepts it on the basis of the speaker’s say-so. On its face this argument does not justify an act or mental state of the speaker, but rather the proposition that the ice cube has melted. To maintain his thesis, Houtlosser must construe such arguments as justifying the speaker’s belief in the conclusion or act of asserting it. But it is only indirectly that the speaker’s belief or act is justified; directly, a proposition is justified, whose content does not describe any action or mental state of the speaker. A more accurate criterion for such cases seems to be the following: if “p because q” expresses that q causes p, then it is a causal argument (rather than an explanation) if and only if it is advanced as a justification of the claim that p, i.e. as evidence that p is true. Textual and contextual clues are of course required to decide whether this criterion is met in a given case.

Third, if the follow-up introduces a reason (i.e. a motive) for someone’s action or mental state, Houtlosser claims that the reason is justificatory if the person spoken about is the speaker but otherwise explanatory. Both these claims seem false.

On the one hand, the reverse seems true about reasons for one’s own action or mental state. If a speaker says, “I really enjoyed it, because it was brilliant” (Houtlosser’s example), in normal circumstances the addressee would accept the speaker’s claim to have enjoyed it on the speaker’s say-so—who would know better whether the speaker enjoyed it or not?—and would take its brilliance as explaining the speaker’s enjoyment. Similarly for parallel first-person examples. It is true that such statements often function simultaneously as justifications of the speaker’s reported action or mental state. For example, the statement “I chose this selection of Aristotle’s writings because it has consistent translations of his technical terms, with a glossary which you can consult” can function to justify the choice as a reasonable one. But it does so by providing an explanation of the act of choosing. What is justified is not the proposition that the speaker made this choice but the proposition that the choice was reasonable. Such a combination of causal explanation and justification is quite general in statements of the form “I did X because Y”. But it is not universal. One can consistently say, “I chose this selection because of its consistent translations of Aristotle’s technical terms, but it was a bad choice.” Such exceptions seem to be decisive counter-examples to Houtlosser’s claim.

On the other hand, giving the reasons for someone else’s action or mental state seems only sometimes explanatory. If someone says, “Hank did not come to the party, because he does not like large crowds”, the reason cited for Hank’s (non-)action can in certain contexts be advanced to justify the claim that he did not come, rather than to explain his already noticed non-attendance. Textual and contextual clues (e.g. prefacing the above statement with the qualifier “I’m pretty sure”) are needed to determine whether such “because” statements are justificatory or explanatory.

It is a difficult matter to determine whether a particular use of a particle like “because” is inferential or causal, and a fortiori a difficult matter to construct an adequate theoretical basis for making this determination. A speech-act approach takes us only part of the way, as Houtlosser makes clear. The difficult cases are those where the main clause in a “because” statement (or its analogue in other constructions) is a factual assertive which is not a prediction. Here one must fall back on appeal to contextual and textual clues as to whether the “because” clause (or its analogue) makes more sense as evidence for the other clause (or its analogue) or as an explanation of what it reports. The hardest cases to identify correctly are those where a known causal factor is being used to justify rather than explain the claim that some past or present factual state of affairs obtains (“I’m pretty sure that Hank did not come to the party, because he does not like large crowds”).

In the twelfth chapter, also included in the indicators group, Francisca Snoeck Henkemans lists indicators of “symptomatic” argumentation, i.e. argumentation that some individual X has a condition Y because of the presence in X of some sign or characteristics Z. Such lists are likely to be helpful to argument analysts but, as Snoeck Henkemans points out, they need to be used with judgment, since many of the indicators listed occur outside argumentative contexts or in other kinds of argumentation. A good example is the phrase “it is clear that”, which indicates symptomatic argumentation only if it introduces the conclusion of an argument, and even then not infallibly so. I do not repeat or discuss Snoeck Henkemans’ list, in part because few readers of this journal would distinguish symptomatic argumentation as a special kind.

Judicial argumentation: Pragma-dialectics has found particularly fruitful application to legal and especially judicial argumentation. In chapter 16, one of the institutional group of papers, Eveline Feteris develops a model for the analysis and evaluation of arguments by judges from the consequences of adopting a particular interpretation of a law or regulation. Such “pragmatic argumentation”, she claims, supports the (un)desirability of the interpretation by appeal to the (un)desirability of some supposed consequence of it. The (un)desirability of this consequence can in turn be supported by its (non-) accordance with some goal or value, in particular one implied by either legislative intention or the purport of the rule or general legal principles. Such pragmatic argumentation can be linked with other arguments, e.g. references to the text of the law being interpreted. It is acceptable to use such pragmatic argumentation, she claims, when such complementary arguments show that the decision is coherent and consistent with other legal rules, principles and goals. For the argumentation scheme to be applied correctly, the normative premiss of (un)desirability must be adequately defended and the empirical premiss of the supposed consequence must be correct. Feteris acknowledges that successful pragmatic argumentation requires consideration of alternative means of securing the desirable consequence (or avoiding the undesirable consequence) and other effects of adopting the proposed interpretation. But she is vague on how such considerations are to be accommodated in the evaluation of pragmatic argumentation in a legal context.

Feteris’ model illustrates both the strengths and the weaknesses of the pragma-dialectical approach. The pragmatic argumentation scheme is ready-to-hand in pragma-dialectics as a recognized sub-type of causal argumentation. So is the twofold evaluative framework of asking whether the argumentation scheme is an acceptable way to defend the “standpoint” (conclusion) and whether answers to the relevant “critical questions” indicate that the scheme is correctly applied. All that is required is some slight tweaking for the legal context of judicial interpretation of the law. But the pragma-dialectical framework provides few resources for addressing the hard questions about consequentialist justifications of judicial interpretation: How far can judges go in adjusting their “interpretations” of laws and regulations so as to produce (what they regard as) good consequences? Supposing that a consequentialist approach is legitimate for a particular interpretive question, what is required in addition to noting an actual or likely consequence of a contemplated interpretation? On what basis can judges legitimately evaluate such a consequence? To what extent do they need to take into account as well other consequences of the contemplated interpretation, as well as consequences of the competing interpretations under consideration? These difficult questions about the limits of judicial discretion can be the subject of a critical discussion conducted according to the procedural norms of pragma-dialectics, but pragma-dialectics itself does not answer them. In particular, Feteris’ analysis only presents options for supporting the claimed (un)desirability of a supposed consequence.

In chapter 17, also included in the institutional group of papers, José Plug addresses within the context of judicial decision-making a difficult question in argument analysis: When an arguer offers more than one reason in direct support of a conclusion, is the arguer to be taken as offering these premisses in joint support or as offering several independent arguments for the same conclusion? Each alternative has evaluative risks. An interpretation of joint support means that showing just one premiss to be false will undermine the argument. An interpretation of independent support means that at least one premiss must be sufficient all by itself to justify the conclusion. A particularly difficult interpretive decision arises when the arguer cites several separately relevant considerations, such as seven reasons for summarily dismissing an employee. Plug considers in turn Walton’s (1996) degree of support test, the principle of charity, and the pragma-dialectical principle of maximal argumentative analysis. According to the degree of support test, reasons are offered in joint support of a conclusion if and only if their combined weight is significantly greater than the sum of their force when taken separately. Plug objects rightly that it is unclear how to answer this question and that Walton’s test ignores the context. The principle of charity, in its moderate form favoured by Walton (1996) and Govier (1987), enjoins that, where the text leaves it unclear whether the premisses are offered in joint or independent support, the argument should be interpreted so as to be neither weaker nor stronger than intended. But, as Plug points out and as the above description of evaluative risks makes clear, each interpretive choice can work to the disadvantage of the arguer. The pragma-dialectical principle of maximal argumentative analysis requires one to interpret the premisses as independent arguments unless there are good reasons for treating them as offered in joint support. This analytical approach maximizes the targets of criticism and protects the argumentation from being undermined by the discovery of just one incorrect premiss, but it risks rejection of a case when no single reason is enough to justify the conclusion but the reasons taken together are enough (as perhaps in the case of the seven reasons for summarily dismissing an employee). Plug proposes, in an extension of the pragma-dialectical principle, that in some cases a judge should consider a phased interpretation, first considering whether each of the reasons is acceptable and sufficient by itself to justify the conclusion, then considering in the case of a negative answer to this question whether the reasons are together sufficient to justify the conclusion. The example which she uses to support this proposal, however, in fact involves clear reasons, both in the provisions of the law and in the wording of the lawyer’s case, for an interpretation of the reasons as offered in joint support; thus this particular example is no exception to the pragma-dialectical principle of maximal argumentative analysis. Nevertheless her suggestion of a phased interpretation has merit.

Plug’s discussion of this question, while intelligent, fails to consider the full range of positions in the literature, for example the work of Freeman (1991) and Vorobej (1994, 1995), and the argument of Pollock (1995) that, when independent arguments are given for the same conclusion, the degree of support given to the conclusion is the degree conferred on it by the strongest of those arguments, the other ones being therefore quite superfluous. Further, there is an instructive analogy between the difficulty of classifying multiple-premiss arguments as offering joint or independent support of their conclusion and the difficulty of classifying arguments as deductive or inductive. Just as the latter difficulty has led many argumentation theorists to reject the distinction altogether and to regard it as a task of evaluation to decide whether the premisses give deductive or inductive support to the conclusion, without first classifying the argument as deductive or inductive, so the former difficulty might lead argumentation theorists to reject altogether the distinction between joint arguments and independent arguments and to regard it as a task of evaluation to decide which of the offered premisses is correct and whether the correct ones provide enough support to the conclusion, either jointly or severally or in some hybrid way, without first classifying the discourse as a single argument with joined premisses or as several independent arguments. Many of the proposed tests for joint or independent support in fact require the “analyst” to evaluate the argument(s) in just these ways, so nothing is gained by labeling the text as a single argument with joint support or as independent arguments. The evaluative work has been done, and that is all that is needed. It should be noted, however, that positive evaluations of the argumentation are possible in more cases than those where a single correct premiss provides sufficient support and where all the premisses together are correct and provide sufficient support. The argumentation would prove its point also if only some of the premisses can be shown to be correct, and these can be shown to provide sufficient support in combination.

In chapter 18, another paper in the institutional group, the American speech communication theorist R. P. Alford argues that pragma-dialectics can function as an extra-legal source of expertise on the evaluation of argumentation by judges in common law jurisdictions. He proposes pragma-dialectics as a replacement for the post-modern relativistic approach typified by Stanley Fish (1990), according to which there is no theory independent of the practice of interpretive communities. Such relativism, Alford contends, leaves appellate judges no basis outside legal practice for judging the quality of argumentation at trial. Alford points out that common law procedure has dialectical features: it puts the burden of proof on the party which advances a standpoint, it provides for objections to such classically recognized dialectical fallacies as eliciting premisses by compound questions, and it bars lawyers from drawing faulty inferences during their closing arguments. Pragma-dialectics fits naturally into this dialectical framework, and is in fact superior to the quasi-Aristotelian dialectical framework within which the common law operates. Alford gives four examples of how pragma-dialectics could help judges in common law jurisdictions to evaluate argumentation. First, the pragmatically oriented reconstruction described in van Eemeren, Grootendorst, Jackson & Jacobs (1993) enables judges to discover the argumentative role of implicit content in the presentation of a case. Second, the four-stage model of a critical discussion can account more adequately than other theories for the damage done by personal attacks on opposing counsel: they are not mere irrelevant argumentation, but function at the confrontation stage to undermine the receptivity of judge or jury to the opposing attorney’s arguments. Third, a pragma-dialectical analytical overview of the argumentation in a complex case can reveal the effect on the quality of the argumentation if an argumentative error had not occurred, and thus help judges to decide if the error was harmless. Fourth, because the ten rules of critical discussion are comprehensive, they provide a better basis for outlawing unacceptable practices than the piecemeal phrase-by-phrase approach currently adopted by common law judges. Alford recommends case studies as a way to show legal theorists and jurists the potential benefits of pragma-dialectics to judicial review of questionable argumentative practices.

Pragma-dialectics is already applied to Dutch legal and judicial practice. Alford makes a compelling case for its applicability in common-law jurisdictions. He perhaps exaggerates the grip of post-modern relativistic theories on outsiders’ theorizing about legal argumentation.

2. Application of pragma-dialectics to specific types of discussions or arguments

Group deliberation: In the fifth chapter, included in the comparison group, M. A. van Rees proposes pragma-dialectics as a tool for analyzing and evaluating those parts of problem-solving discussions where interlocutors try to resolve differences of opinion, e.g. about what the problem is, what possible solutions should be considered, how they should be judged, and what solution should be adopted. Although pragma-dialectics applies only to some aspects of some parts of a problem-solving discussion, van Rees demonstrates its power by applying it sensitively to a discussion among managers of a hospital about future negotiations on collaboration with a neighboring hospital. The pragma-dialectical approach differs from existing approaches to problem-solving discussions in being normative as well as descriptive, in being a procedure for the resolution of differences of opinion rather than a problem-solving procedure like Dewey’s reflective thinking procedure, in using critical rather than epistemic rationality as a basis, in being rooted in linguistics rather than in the social sciences, in attending to socio-emotionally oriented aspects in combination with task-oriented aspects rather than attending to each aspect separately, in focusing on the process rather than the product, and in being applied to spontaneous natural interaction rather than to contrived discussions in a laboratory setting.

In the eighth chapter, included in the scope group, Mark Aakhus identifies the theory of argument reconstruction implicit in each of three types of “groupware”: commercially available software for helping groups to reach a decision by discussion. Issue-networking programs construe argumentation as a process of raising and connecting issues, with reference to which pros and cons are articulated. Such programs conform to the confrontation rule of pragma-dialectics, which prohibits restrictions on the advancing or critique of standpoints, but they are subject to drift and use problematic labeling of turns. Funneling programs, such as group decision support software, construe decision-oriented argumentation as a sequential process, moving from initial definition of the problem and open-ended contribution of ideas for its solution through organization of the ideas suggested to a choice among the possible solutions thus identified in the light of mutually agreed criteria. Such programs have the merit of separating claims from their authors, but in fact allow for very little argumentation. Reputation-management programs construe group decision-making as expert inquiry, where non-experts grade the answers of experts to their questions. Such programs are unique among groupware in tying positions inextricably to their authors, a feature which risks confusion of the quality of the ideas with the reputation of their author. Groupware in general, according to Aakhus, illustrates how software designed to meet challenges to group deliberation uses models of argumentation which are somewhere between a complete theory like that of pragma-dialectics and the naive beliefs of untrained arguers.

In fact, group deliberation beginning from an open-ended decision-making situation is different from the critical discussion envisaged in pragma-dialectics. The starting-point is not the expression of a difference of opinion but the posing of a question about what to do: “Where shall we go for dinner?” “What changes if any shall we make to our curriculum?” “How shall we ensure that residents in this jurisdiction do not get poisoned by the food they buy?” Critical discussions may be embedded in such contexts, when one participant expresses an opinion and another disagrees with it or requests a justification. But the overall framework is rather different. Analogues for group deliberation of the pragma-dialectical theoretical model for critical discussions may be helpful to groupware designers confronting practical obstacles to good joint deliberation.

Commercial and domestic disputes: In the second chapter, included in the comparison group, Scott Jacobs and Mark Aakhus consider a puzzle about the behaviour of third-party mediators in disputes. Disputants typically behave like participants in a critical discussion: they advance arguments to support their position, and they raise objections to the arguments of the other side. But mediators (at least in the United States) typically ignore these arguments, and indeed treat argumentation by a disputant as a sign of resistance to seeking a solution. Why do mediators so studiously fail to hold disputants to the ideals of critical discussion?  Because their model for dispute resolution is not critical discussion but bargaining, where the parties are to reach through a process of offer and counter-offer a settlement which accommodates reasonably well the interests of both parties. As Jacobs and Aakhus point out, such “positional bargaining” does not necessarily lead to the most reasonable settlement. The “principled negotiation” proposed by Fisher and Ury (1981) has much more in common with a critical discussion, in that the parties are to agree at the beginning, in something like the opening stage of pragma-dialectics, on common principles and on the interests each of them has. Such a starting-point, its proponents claim, is more likely to lead to an optimal settlement (e.g. one which is Pareto-optimal with respect to the parties’ interests) than positional bargaining. Thus the best procedure for settling disputes may not be as remote from the procedure of an ideal critical discussion as is the practice of many mediators.

In chapter 13, included in the interaction group, Harry Weger Jr. brings the resources of pragma-dialectics to bear on studies of verbal disputes between husbands and wives. Such studies focus on the interpersonal dimensions of the interaction, but Weger shows that behaviours like personal criticism, defensiveness and withdrawal not only damage the relationship but also violate the pragma-dialectical rules for critical discussion. He speculates that rational resolution of disputes with intimates according to these rules may be personally satisfying as well, and advocates research on the effect of violations of critical discussion rules not only on dispute resolution but also on decision quality and interpersonal outcomes. He also advocates attention by argumentation scholars (including pragma-dialecticians) to the relational as well as the content dimension of argumentative messages, on the ground that people who advance or respond to arguments are not only trying to arrive at a justified position but are also managing their identity and their relationship with the interlocutor. Analysts of written texts, speeches and conversations need to attend to these various dimensions of communication.

Somewhat more controversially, Weger claims that some conversational moves are only fallacies when the hearer responds in a certain way. In other words, the fallacy lies in the interaction, not in the contribution of one speaker. Specifically, Weger notes that personal criticism only harms the direction of the conversation if its target recognizes it as personally injurious and responds defensively. But the ad hominem seems to be a fallacy because of what its perpetrator is trying to do, even if the attempt is unsuccessful. According to pragma-dialectical analysis, any personal criticism whose function is to undermine its target’s ability to contribute to a critical discussion is a fallacy, regardless of how this person reacts to the criticism; in particular, a fallacious ad hominem attack remains a fallacy even if its target takes no notice of it. On the other hand, personal criticism is not a fallacy if it is relevant to a standpoint at issue in a critical discussion; a relevant personal criticism does not become a fallacy when its target takes offence or responds defensively. These evaluations seem quite reasonable.

Visual arguments: Leo Groarke argues in the ninth chapter, which is included in the scope group, that pragma-dialectics can accommodate visual argument as well as verbal argument, because it permits interpreters to reconstruct communicative acts on the assumption that they are comprehensible, sincere, relevant, efficacious and appropriately related to other speech acts. These principles provide the latitude for analyzing visual arguments as comprehensible and as making sense, both internally and externally. Groarke’s proposed extension of pragma-dialectics will no doubt be congenial to its proponents, but of course other approaches to argumentation may also accommodate visual arguments. Whatever the approach, one must be careful to distinguish different uses of images in argumentation. In some cases Groarke cites, the image is the production of physical evidence in support of a claim (e.g. juggling some balls to demonstrate that one can juggle). In others, the image works in combination with words to express a point of view, as in many political cartoons, without any supporting premisses or further conclusion. Advertising images which create associations with a product can only with strain be interpreted as arguments. Images can function to illustrate verbally expressed premisses, as when a diagram depicts the experimental setup described in a scientific paper. It is doubtful that there can be arguments consisting solely of visual images. Rather, verbal arguments sometimes have non-verbal visual components.

Question-answer discussions: In chapter 14, included in the interaction group, Leah Polcar evaluates from a pragma-dialectical perspective various sorts of non-straightforward answers to questions, such as those given by an office-seeking politician who does not wish to give a direct answer to an awkward question. Opt-out responses (such as “no comment” or “I am not at liberty to discuss this matter” or “I have already responded to that question”) are non-fallacious if the respondent is simply refusing to engage in a critical discussion, but violate the second rule of critical discussion (“a party that advances a standpoint is obliged to defend it if asked to do so”) if they are refusals to provide supporting argument for a previously declared position. Their obviousness, however, makes them easy to repair if they are fallacious. Indirect answers (such as a comment that the United States should work through international agencies in response to a question whether the interlocutor agrees with some recent unilateral action by the US President) implicate a direct answer through their obvious violation of Gricean cooperative maxims, and so are not fallacious. Treating such implicated answers as analogous to unexpressed but implicit premisses, Polcar takes subsequent denial of commitment to the implicated answer as a fallacy, a violation of the fifth rule of critical discussion (“a party may not disown a premiss that has been left implicit by that party or falsely present something as a premiss that has been left unexpressed by the other party”). Evasive answers (such as a statement that an athlete has a family bereavement in response to a question why she pulled out of a race, when the real reason for pulling out was that she failed a drug test) appear to be or implicate full answers when they do not really do so, and in an argumentative context can therefore violate the fourth rule of critical discussion which requires that premisses be relevant (“a party may defend a standpoint only by advancing argumentation related to that standpoint”). Besides, an evasive answer which appears to implicate a full answer, but is unclear about what is implicated, could violate the tenth rule, which requires clarity (“parties must not use any formulations that are insufficiently clear or confusingly ambiguous, and they must interpret the formulations of the other party as carefully and accurately as possible”). As Polcar acknowledges, not all non-straightforward answers occur in argumentative contexts, and, where they do occur in such contexts, further analysis is required of the extent to which an evasive answer in particular frustrates the purpose of a critical discussion.

In the last chapter, included in the institutional group, Alan Aldrich shows, through quotations from the US Congressional hearings at which Oliver North testified in 1987, that participants in such avowedly information-seeking question-answer dialogues appeal to Gricean maxims of conversational appropriateness in quality, quantity, relevance, and manner (Grice 1975) both to defend their own behaviour and to attack that of others. While these quotations illustrate the common acceptance of the Gricean maxims, they have little relevance to pragma-dialectics, which is a theory about critical discussion rather than about information-seeking dialogues.

3. Non-ideal higher-order conditions for critical discussion

In the seventh chapter, which introduces the scope group of papers, Sally Jackson explains how pragma-dialectics can be used in the design of argumentative space. The sort of critical discussion envisaged in the pragma-dialectical model requires higher-order conditions, such as absence of authority relations among the discussants and willingness to risk loss of face if one fails, which are rarely met in practice. It is possible to use knowledge of the required higher-order conditions to design circumstances of interaction which will facilitate critical discussion and improve its quality. In particular, people learn better if they produce arguments in response to disagreement, but the typical classroom situation has impediments to such argumentation: the authority of the teacher, student concerns about losing face, lack of self-confidence among some students, and time constraints. A design approach examines discourse practices, compares them to an ideal model, specifies designable features, and examines the change in practice which follows implementation of the design. Jackson describes three computer-mediated designs for stimulating argumentation in the context of learning, and their outcomes in practice. Designs which expanded opportunities for disagreement and redistributed authority or mitigated inequality were most useful for learning. Unfortunately, in Virtual Peer, after giving an answer to a physics problem, students exposed to an argument for a different answer were just as likely to switch from an initially correct answer to an incorrect one as vice versa—a result indicating a need to improve the design. Developing and reviewing argumentation designs not only contributes to improved practice but also provides a test of the underlying theory, which may need revision in the light of empirical results. The use of pragma-dialectics as a theoretical basis for (re)designing argumentative space is clearly a promising extension of the research program. As Jackson points out, redesigning argumentative space does not require computers; it can for example be implemented in any classroom.

In chapter 15, included in the interaction group. Dale Hample uses the practices of the medieval and Renaissance Inquisition to propose an addition to the pragma-dialectical approach to reconstructing arguments. According to van Eemeren, Grootendorst, Jackson & Jacobs (1993: 95), an argumentative discussion and its surrounding practical activity contain a “disagreement space”, a large complex of reconstructible commitments of one’s conversational partner, any of which may be reconstructed and challenged. Trials in the Inquisition, however, presupposed the guilt of the accused, and were oriented towards securing a confession. Hence in practice the disagreement space was severely constricted; for example, the accused and their attorneys were unable to challenge the claim of guilt. Between the virtual disagreement space of everything that is theoretically arguable and the actual disagreement space of the commitments which are called into question, Hample proposes to insert the “possible disagreement space” allowed by institutional and power constraints. Hample’s proposal is a salutary reminder that the higher-order conditions for an ideal critical discussion are often lacking in practice, the Inquisition being but an extreme example of a ubiquitous phenomenon.

In chapter 19, included in the institutional group, Dale Brashers, Stephen M. Haas and Judith L. Neidig describe how arguers overcome disadvantages due to the absence of required higher-order conditions for a critical discussion. They use as their database descriptions by adults with HIV or AIDS of what they would say to convince a reluctant physician to prescribe a desired new experimental treatment that is available only if a physician recommends it. Aside from the power imbalance in this situation (where only the physician has the power to prescribe the drug) and the pattern of physician dominance and patient submissiveness in physician-patient interactions in the United States (with even the most participatory decision-making models limiting patients’ roles to providing information about their personal circumstances and choosing from alternative treatments proposed by the physician), it is a situation of self-advocacy, where the fact that the arguer is arguing out of self-interest potentially conflicts with the requirement of objectivity in a critical discussion. Brashers and his co-authors hypothesized that arguers in such situations would attend directly to such deficiencies in higher-order conditions as the interlocutor’s pursuit of goals other than rational resolution of the dispute (e.g. managing identity and maintaining the relationship), the patient’s real or perceived lack of relevant expertise, time constraints, and the patient’s perceived self-interest. They claim that the strategies reported by their respondents address such deficiencies in the higher-order conditions. In fact, however, only three of the 11 reported strategies (establishing expertise, appealing to an agreed collaborative nature of the relationship, eliciting counter-arguments) appear to do so. Some strategies (conditional threat to change physician, establishing an obligation based on the fact that the patient is paying for the physician’s services, assertion of a right to bodily self-determination) appear to belong rather to a negotiation dialogue than to a critical discussion, with the aim being to get the physician to prescribe the drug independently of being rationally convinced of the reasonableness of doing so. One strategy, accepting whatever the physician recommends, opts out of either critical discussion or negotiation. Other strategies (establishing the facts, describing the severity of the consequences, promising to exercise caution) seem to belong rather to the substance of a critical discussion of the issue than to remedying the inadequacies of the higher-order conditions. In fact, Brashers and his co-authors report no strategies for dealing with the physician’s other goals and with the time constraints. More work is needed on the important issue of how to encourage critical discussion when the higher-order conditions for such discourse are less than ideal.

4. Profiles of dialogue

The remaining two papers in this collection are only tangentially related to pragma-dialectics. In the tenth chapter, included in the indicators group, Erik Krabbe argues by example for the usefulness of profiles of dialogue in analysing and evaluating argumentative discussions. Profiles of dialogue indicate abstractly how a fragment of dialogue does or might proceed, and thus facilitate evaluation of actual exchanges. They differ from the systems of formal dialectic elaborated by such philosophers as Charles Hamblin (1970) and Paul Lorenzen and Kuno Lorenz (1978) in being merely partial. An example is the following unembedded sequence:

Protagonist: p.

Antagonist: Why p?

Protagonist: Why not p?

Here the antagonist has initiated a critical discussion, but the protagonist frustrates its purpose of rationally resolving the expressed conflict of opinion by demanding that the antagonist provide a justification for the opposite position, to which the antagonist has not expressed a commitment. The profile of dialogue, by abstracting from the content, shows what is fallacious about the protagonist’s move. Krabbe shows how a branching normative profile can exhibit the possible legitimate moves at each stage of an ongoing critical discussion; the profile can include fallacy allegations, which initiate a critical discussion at the next highest level. Krabbe’s examples show how profiles of dialogue can exhibit the dialectical structure of imagined conversational exchanges of considerable complexity, and can provide tools for evaluating them. It remains to be seen how useful such profiles are for analysing and evaluating actual argumentation, including monological argumentation, which not only pragma-dialecticians but also formal dialecticians like Krabbe regard as implicitly dialogical.

5. Justification and persuasion

In the fourth chapter, included in the comparison group, Daniel J. O’Keefe reports on the results of a meta-analysis of studies of the persuasive effects of leaving one’s conclusion unstated or incompletely specified, techniques which violate discussion rule 10: “Parties must not use any formulations that are insufficiently clear or confusingly ambiguous, and they must interpret the formulations of the other party as carefully and accurately as possible.” (47, n. 4) Somewhat surprisingly to this reviewer, in the 35 studies that met O’Keefe’s inclusion criteria, with a total of 14,215 participants, arguments with explicitly stated conclusions were significantly more persuasive than arguments with unstated conclusions, and arguments with more specified conclusions were significantly more persuasive than arguments with less specified conclusions. Thus normative and instrumental considerations happily coincide: both for fairness and for effectiveness, arguers should make their conclusions explicit and specific.

The book ends with a useful index of terms and an index of names.


            I thank Peter Houtlosser for his helpful comments on an earlier draft of this review. I am of course responsible for any flaws in this version, and indeed Houtlosser disagrees with some of my positions expressed above.


Eemeren, F. H. van, & Grootendorst, R. (1984). Speech acts in argumentative discussions: a theoretical model for the analysis of discussions directed towards solving conflicts of opinion. Dordrecht, The Netherlands and Cinnaminson, NJ: Foris.

Eemeren, F. H. van, & Grootendorst, R. (1992). Argumentation, communication and fallacies: a pragma-dialectical perspective. Hillsdale, NJ: Lawrence Erlbaum Associates.

Eemeren, F. H. van, & Grootendorst, R., eds. (1994). Studies in pragma-dialectics. Amsterdam: Sic Sat.

Eemeren, F. H. van, & Grootendorst, R. (2004). A systematic theory of argumentation: The pragma-dialectical approach. Cambridge: Cambridge University Press.

Eemeren, F. H. van, Grootendorst, R., Jackson, S., & Jacobs, S. (1993). Reconstructing argumentative discourse. Tuscaloosa, AL: University of Alabama Press.

Eemeren, F. H. van, & Houtlosser, P. (1999). Strategic manoeuvring in argumentative discourse. Discourse Studies 1, 479-497.

Eemeren, F. H. van, & Houtlosser, P. (2001). Managing disagreement: rhetorical analysis within a dialectical framework. Argumentation and Advocacy 37, 150-157.

Fish, S. (1990). Doing what comes naturally: Change, rhetoric and the practice of theory in literary and legal studies. Oxford: Oxford University Press.

Fisher, R., & Ury, W. (1981). Getting to YES: negotiating agreement without giving in. Boston: Houghton Mifflin.

Freeman, J. B. 1991. Dialectics and the macrostructure of arguments: A theory of argument structure. Berlin: Foris.

Govier, T. (1987). Problems in argument analysis and evaluation. Dordrecht: Foris.

Grice, H. P. (1975). Logic and conversation. In Cole, P., & Morgan, J. (eds.), Syntax and semantics, vol. 3: Speech acts (pp. 41-58). New York: Academic Press.

Hamblin, C. L. (1970). Fallacies. London: Methuen.

Lorezen, P., & Lorenz, K. (1978). Dialogische Logik. Darmstadt: Wissenschaftliche Buchgesellschaft.

Pollock, J. L. (1995). Cognitive carpentry: A blueprint for how to build a person. Cambridge, MA: MIT Press.

Searle, J. R. (1969) Speech acts: an essay in the philosophy of language. London: Cambridge University Press.

Searle, J. R., & Vanderveken, D. (1985). Foundations of illocutionary logic. Cambridge: Cambridge University Press.

Vorobej, M. (1994) The TRUE test of linkage. Informal Logic, 16, 147-157.

Vorobej, M. (1995). Hybrid arguments. Informal Logic, 17, 289-296.

Walton, D. (1996). Argument structure: A pragmatic theory. Toronto/ Buffalo: University of Toronto Press.