Rationality, Reasonableness, and Critical Rationalism: Problems with the Pragma-dialectical View

Abstract

A major virtue of the Pragma-Dialectical theory of argumentation is its commitment to reasonableness and rationality as central criteria of argumentative quality. However, the account of these key notions offered by the originators of this theory, Frans van Eemeren and Rob Grootendorst, seems to us problematic in several respects. In what follows we criticize that account and suggest an alternative, offered elsewhere, that seems to us to be both independently preferable and more in keeping with the epistemic approach to arguments and argumentation we favor.

A major virtue of the Pragma-Dialectical theory of argumentation¹ is its commitment to reasonableness and rationality as central criteria of argumentative quality. However, the account of these key notions offered by the originators of this theory, Frans van Eemeren and Rob Grootendorst, seems to us problematic in several respects. In what follows we criticize that account and suggest an alternative, developed elsewhere, that seems to us to be both independently preferable and more in keeping with the epistemic approach to arguments and argumentation we favor.²

1 The Reasonable Rabbi

In their most recent systematic discussion of these matters (2004), van Eemeren and Grootendorst define argumentation as “a verbal, social, and rational activity aimed at convincing a reasonable critic of the acceptability of a standpoint by putting forward a constellation of propositions justifying or refuting the proposition expressed in the standpoint.” (2004, p. 1) On this view, rationality is an essential aspect of argumentation, and by saying that argumentation is “a rational activity,” van Eemeren and Grootendorst mean that it is “a complex speech act aimed at convincing a reasonable critic,” one that is “generally based on intellectual considerations” (2004, p. 2, emphases in original):

When someone advances argumentation, that person makes an implicit appeal to reasonableness: He or she tacitly assumes that the listener or reader will act as a reasonable critic when evaluating the argumentation. Otherwise, there would be no point in advancing argumentation. (ibid.)

As van Eemeren and Grootendorst make clear, the pragma-dialectical view attempts to combine descriptive and normative approaches to the study of argumentation under the heading of ‘normative pragmatics.’ (2004, pp. 9–11) The normative dimension is captured by their accounts of acceptability, which concerns the appropriateness or acceptability (or otherwise) of argumentative moves or claims, and of reasonableness, which concerns the discussion rules in accordance with which judgments of acceptability are ideally made. They invoke the image or model of “an extremely wise man—say, a rabbi,” whose position is “that of a rational critic who judges reasonably.” (2004, p. 12) The rabbi asks himself: “When should I, as a rational critic who judges reasonably, regard an argumentation as acceptable?” (2004, p. 13) And if he adopts “the critical-rationalistic view of reasonableness” (2004, p. 17, emphasis in original) that van Eemeren and Grootendorst favor, he answers that “an argumentation may be regarded as acceptable” just in so far as it “is an effective means of resolving a difference of opinion in accordance with discussion rules acceptable to the parties involved.” (2004, p. 16) So, argumentations (argumentative moves, i.e., particular speech acts) are evaluated in terms of acceptability, which is itself a matter of instrumental efficacy: an argumentation is acceptable if it is “an effective means of resolving a difference of opinion in accordance with discussion rules” and conforms to procedures that the parties accept.³ Such rules are in turn deemed reasonable to the extent that they are adequate for resolving the relevant difference of opinion. Thus it is argumentations that are or are not acceptable, and discussion rules (and/or the procedure in which they play a role) that are or are not reasonable:

The extent to which a particular rule is considered reasonable depends on the adequacy of that rule, as part of a procedure for conducting a critical discussion, for solving the problem at hand. (2004, p. 16)

So, “[o]ur rabbi...asks himself which theoretical instruments are, or can be made, available to him to systematically arrive at a solution of his problem regarding the acceptability of argumentation.” (2004, p. 19) To pass judgments about the acceptability of argumentations, the rabbi, if he embraces the pragma-dialectical approach, uses “an ideal model of a critical discussion and a procedure for how speech acts should be presented in order to be constructive moves in such a discussion.” (2004, p. 20) Accordingly, the rabbi’s judgments concerning the acceptability of argumentations will be based on the reasonableness of the discussion rules that license the argumentations in question. The rules are deemed reasonable just in so far as they conduce to the resolution of the relevant difference of opinion.

We have already noted our worries concerning the instrumental view of acceptability built into the pragma-dialectical account. In what follows, we want to register our doubts concerning the view of reasonableness that van Eemeren and Grootendorst endorse.

2 The Pragma-dialectical View of Reasonableness

In Biro and Siegel (2006), we suggested that van Eemeren and Grootendorst’s account of argumentative normativity is defective in that the ‘dialectical’ account of reasonableness they offer fails to capture that normativity. We briefly summarize our case for that judgment next.

Van Eemeren and Grootendorst distinguish ‘rational’ and ‘reasonable’ as follows: “[W]e shall use the term rational for the use of the faculty of reasoning and the term reasonable for the sound use of the faculty of reasoning.” (2004, p. 124, emphases in original) They articulate their preferred, dialectical view of the ‘sound use of the faculty of reasoning,’ i.e., reasonableness, in the following way:

In our view, it is necessary to depart radically from the justificationism of the geometrical and anthropological approaches to reasonableness and to replace these conceptions of reasonableness with a different one. We do so by adopting the view of a critical rationalist who proceeds on the basis of the fundamental fallibility of all human thought. To critical rationalists, the idea of a systematic critical scrutiny of all fields of human thought and activity is the principle that serves as the starting point for the resolution of problems. In this approach, conducting a critical discussion is made the point of departure for the conception of reasonableness—which implies the adoption of a dialectical approach. As we have indicated, argumentation in a dialectical approach is regarded as part of a procedure for resolving a difference of opinion on the acceptability of one or more standpoints by means of a critical discussion... The reasonableness of the procedure is derived from the possibility it creates to resolve differences of opinion (its problem validity) in combination with its acceptability to the discussants (its conventional validity). In this connection, the rules of discussion and argumentation developed in a dialectical theory of argumentation must be scrutinized in terms of both their problem-solving effectiveness and their intersubjective acceptability. (2004, pp. 131–132)

We applaud the pragma-dialectical commitment to fallibilism. Nevertheless, there is a major problem with the view of reasonableness expressed here. According to it, a move in a critical discussion is acceptable if it comports with the rules governing critical discussions: those rules are reasonable if they are both ‘problem-valid,’ i.e., tend to produce a resolution of the difference of opinion in question, and ‘conventional-valid,’ i.e., are embedded in a procedure that is acceptable to the discussants. What of the resolution itself? If the parties resolve their difference of opinion by making acceptable argumentative moves, in accordance with reasonable (i.e., problem- and conventional-valid) rules, and, in doing so, come to agree, is the new belief on the part of one of them reasonable? Van Eemeren and Grootendorst are committed to an affirmative answer to this question, independently of any consideration of the probative strength of the reasons offered. This is manifestly not the way that ‘reasonableness,’ and normativity more generally, are understood in either philosophical or every-day discourse. Nor should they be, since it is clear that disputes resolved in accordance with the pragma-dialectical rules can result in new beliefs that are not reasonable in the straightforward sense that the reasons offered in their support establish their truth or enhance their justificatory status. For example, if you and I are white racists and are engaged in a critical discussion about the wisdom of voting for a black candidate—I plan to vote for him because, despite his skin color, he reminds me of my father, say—your reminding me of my general attitude concerning the abilities of blacks, in moves that comport perfectly well with the pragma-dialectical rules, might well resolve our difference of opinion in accordance with rules we both accept, but my new belief that I should not vote for this candidate is still not justified by my racist prejudices, despite our agreement on the matter and the appropriateness of the procedure by which I arrived at it.⁴ Thus ‘dialectical reasonableness’ as articulated by van Eemeren and Grootendorst fails to establish particular resolutions of differences of opinion as reasonable in any serious sense, since a ‘dialectically reasonable’ resolution may nevertheless be completely unreasonable insofar as there is no good reason for either discussant to accept or believe it.⁵

Van Eemeren and Grootendorst will not be troubled by this criticism, however, since they argue that any non-dialectical view of reasonableness—such as the one we just invoked, according to which reasonableness is a function of the epistemic or probative force of reasons—founders on the famous ‘Münchhausen trilemma.’ Let us consider their case for this claim next.

3 Critical Rationalism, ‘Justificationism’ and The Münchhausen Trilemma

Van Eemeren and Grootendorst suggest that any non-dialectical conception of reasonableness will inevitably founder on the ‘justificationism’ that Karl Popper, Hans Albert and their fellow critical rationalists famously argued leads inevitably to this allegedly irresolvable trilemma (henceforth MT):

A crucial objection that applies to both the geometrical and the anthropological norm of reasonableness is that they are both based on “justificationism”: Both approaches assume that reasonableness is concerned exclusively with legitimizing standpoints definitively. Justificationism of any kind, however, can never escape the so-called Münchhausen Trilemma, because in the last resort the justification has to choose from the following three alternatives: (1) ending up in an infinite regress of new justifications (regressus in infinitum); (2) going round in a circle of mutually supporting arguments; (3) breaking off the justificatory process at an arbitrary point. None of these three alternatives is really satisfactory. (2004, p. 131)

We wish to make two points concerning this argument.

(i) Justificationism. ‘Justificationism’ is understood in this passage as a matter of “legitimizing standpoints definitively.” What does ‘definitively’ mean here? If it means ‘proving,’ ‘justifying with certainty,’ ‘establishing once and for all, with no possibility of reconsideration,’ and the like, we agree that it should be rejected. But if it means, rather, justifying by adequate reasons and evidence, then we do not agree.⁶

It is significant that Popper himself, though he frequently uses ‘justificationism’ in the first way, understands it in this latter way when he famously rejects induction and confirmation. He claims that theories can be refuted, though not justified or supported, by evidence; in his hands the rejection of ‘justificationism’ is tantamount to the rejection of the very possibility of supporting evidence and justification.

It is easy to find passages in which Popper, in rejecting justification, seems to be rejecting it in the first, certainty-involving sense:

The Greeks’ discovery of the critical method gave rise at first to the mistaken hope that it would lead to the solution of all the great old problems; that it would establish certainty; that it would help to prove our theories, to justify them. But this hope was a residue of the dogmatic way of thinking; in fact nothing can be justified or proved (outside of mathematics and logic). (Popper 1963, p. 51, emphases in original)

Here Popper clearly regards justification as requiring proof, or certainty. We agree with Popper that scientific (and other) theories cannot be proved in the logician’s sense of the term or established with certainty and thus cannot be justified in this sense. But this is not a controversial point. Nor does it adequately capture Popper’s philosophical program, since his philosophical opponents, e.g., the logical positivists, did not claim that scientific theories could be proved or established with certainty. Rather, they argued that theories could be justified, confirmed, and inductively supported by reasons and evidence, and they endeavored (among other things) to render such confirmatory relationships between theories and the evidence for or against them, or, more precisely, the propositions that express these, probabilistically precise. We do not wish to defend Reichenbach’s, Carnap’s, or anyone else’s version of inductive logic here. Rather, we are concerned to point out that the philosophical power of Popper’s falsificationist program depends upon understanding it as involving the second sense of ‘justificationism’ articulated above, i.e., as rejecting not just certainty, but the very possibility of positive evidential, confirmatory, justificatory support. Passages supporting this understanding of Popper’s rejection of ‘justificationism’ can also be readily found:

[Our conjectures] may survive these tests; but they can never be positively justified: they can neither be established as certainly true nor even as ‘probable’ (in the sense of the probability calculus).... None of [our theories] can be positively justified. (Popper 1963, p. vii)

...we do not establish anything by this procedure: we do not wish to ‘justify’ the ‘acceptance’ of anything, we only test our theories critically, in order to see whether or not we can bring a case against them. (Popper 1963, p. 388, emphasis in original)

We cannot justify our theories, but we can rationally criticize them...A scientific result cannot be justified. It can only be criticized, and tested. (Popper 1972, p. 265)

...there are no such things as good positive reasons; nor do we need such things. (Popper 1974, p. 1043, emphasis in original)⁷

Notice, first, that these passages straightforwardly and uncontroversially speak not just of the rejection of certainty, but also of the rejection of the very possibility of reasons and evidence that yield support/warrant/justification. Notice, next, that it is only under the latter interpretation that Popper’s famous rejection of confirmation and induction makes sense, for confirmation and induction are not normally thought to yield either certainty or proof. Finally, notice that this is exactly how Popper and critical rationalism are usually interpreted, by both their defenders and their critics:

What matters to a critical rationalist is whether the conjectures under debate are right, not whether there are reasons to suppose that they are... Arguments, according to critical rationalism, are always negative; they are always critical arguments, used only and needed only to unseat conjectures that have been earlier surmised. (Miller 1985, p. 10)

Naïvely one might think that one could at least have good reasons on occasion for thinking that one hypothesis or observation report is more likely to be true than false. Not so, says Popper... [This] amounts to the rejection of all inductive argumentation. That is, Popper denies the legitimacy of any argument in which the premises purport to support the conclusion without entailing it. (Newton-Smith 1981, p. 44)

[T]here are no such things as good reasons; that is, sufficient or even partly sufficient favourable (or positive) reasons for accepting a hypothesis rather than rejecting it, or for rejecting it rather than accepting it, or for implementing a policy, or for not doing so. (Miller 1994, p. 52)

[G]ood reasons do not exist; it is impossible to furnish a good reason in favour of any thesis or action whatever. (Miller 1994, p. 55)

Such citations could be multiplied; it is uncontroversial among Popper scholars, and especially defenders of critical rationalism, that in rejecting ‘justificationism’ Popper was rejecting not just the requirement of certainty, but the possibility of positive support tout court. With respect to this understanding of ‘justificationism,’ it is important to note that (a) Popper’s rejection of it did not survive critical scrutiny—even Popper himself famously admitted that his view required a “whiff of inductivism”⁸—and (b) van Eemeren and Grootendorst cannot comfortably join in rejecting the possibility of supporting evidence, insofar as they offer and rely upon reasons and evidence in support of their own claims. We briefly develop these points in turn.

(a) Can critical rationalists do without positive justification? Many scholars think that they cannot. Putnam argues that both the practice of science and the application of scientific theories and laws require induction and positive justification, and that without these, “science would be a wholly unimportant activity. It would be practically unimportant, because scientists would never tell us that any law or theory is safe to rely upon for practical purposes; and it would be unimportant for the purpose of understanding, since on Popper’s view, scientists never tell us that any law or theory is true or even probable” (Putnam 1974, pp. 222–223; see also pp. 224, 237). Lakatos’ thorough and incisive discussion (Lakatos 1974, pp. 256–263) equally insists on the need for a substantive inductive principle, because “the ‘logic of the growth of knowledge’ must include—in addition to Popper’s logico-metaphysical theory of verisimilitude—some speculative genuinely epistemological theory connecting scientific standards with verisimilitude” (Lakatos 1974, p. 261, emphases in original). Levison puts the difficulty this way:

Popper’s difficulty is that he cannot consistently hold that successfully surviving a wide range of experiments makes it likely that a theory will continue to survive such tests. Thus, to be consistent, he must deny that the claim that a test can be successfully repeated can be justified by argument. But, if so, he cannot claim consistently that he has solved the logical problem of induction, even as he defines it. Hume’s problem is not so much solved by Popper as it is transformed from the problem of justifying generalizations based on past observations to the problem of determining the comparative acceptability of explanatory theories and other scientific statements on the basis of experimental testing. The question that we are left with is why the fact that an empirical theory has survived a wide range of experimental tests, when other comparable theories have not survived those tests, gives us good reason for supposing that a predictive consequence of the former or corroborated theory is worthy of the confidence of reasonable men, while those of the latter are not worthy. (Levison 1974, p. 330)

Essentially the same point is made by Newton-Smith (1981, pp. 44–76), O’Hear (1980, pp. 36–67, see esp. p. 46), and others too numerous to mention. The problem, as all these authors suggest, is straightforward: if corroboration does not provide such ‘good reason,’ it is hard to see in what sense corroborated theories are preferable to non-corroborated or less-well-corroborated ones—why is corroboration an epistemic good, and a corroborated theory epistemically preferable to a non-corroborated one, or otherwise “worthy of the confidence of reasonable men”? But if corroboration does provide such good reason, it can do so only via induction and positive support. Or, to put the point slightly differently: Popper needs a connection between corroboration and verisimilitude for his theory to succeed, but the only sort of connection available is an inductive one. As Salmon pithily puts it: “Modus tollens without corroboration is empty; modus tollens with corroboration is induction.” (Salmon 1966, p. 26, emphases in original) Popper and his fellow critical rationalists simply cannot do without positive justification. In this sense, critical rationalism’s rejection of ‘justificationism’ fails. Critical rationalism denies the possibility of good reasons and justificatory support, yet itself requires it.⁹

(b) Can van Eemeren and Grootendorst do without positive justification? Here we can be brief. Van Eemeren and Grootendorst have made a major contribution to argumentation theory in developing their pragma-dialectical view, and they have done so by offering arguments intended to secure their central claims, such as those concerning the aims of argumentation, the legitimacy of their dialectical rules, and many others. To reject ‘justificationism’ in the sense of positive justification would be devastating to their project in at least two ways: it would undercut the possibility of any argument(ation) succeeding in the sense of providing interlocutors with good reasons to accept or reject any given standpoint at issue; and it would undercut their own many impressive efforts to defend the pragma-dialectical view they champion. Without the possibility of positive justification, there can be no possibility of their arguments establishing the epistemic worthiness of their own view.

If all this is right, van Eemeren and Grootendorst cannot do without ‘justificationism’ in the sense of positive support. They need it—but can they have it? Here we need to confront the critical rationalist’s case against it: the dreaded Münchhausen Trilemma.

(ii) The Münchhausen Trilemma and the Structure of Justification. MT reflects the fundamental problem of the structure of epistemological justification that has exercised epistemologists for centuries; we do not pretend to resolve this hoary difficulty here.¹⁰ Nevertheless, as articulated by van Eemeren and Grootendorst, we think MT can be readily overcome, in the following way: standpoints or claims can be ‘legitimated’ or justified by reasons or evidence. For example,

p: Cheney approved of and encouraged the torture of suspected ‘terrorists’ by U.S. agencies and personnel

is justified by (or receives strong evidential support from)

q: Cheney lobbied Senators to defeat the McCain amendment prohibiting torture by all US agencies and personnel

The first option in MT, infinite regress, is avoided by noting that sometimes evidence is sufficient for justification. In this example, q, if itself well justified, affords strong justification for p: that is, if we have good reason to believe that q, we have good reason to believe that p as well. Of course, questions concerning the belief-worthiness or justificatory status of q can always be raised, and such questioning pushes the justificatory chain back one step. If q is challenged, it can, in turn, be justified by some further evidence, e.g.,

r: Cheney called a news conference during which he admitted his lobbying efforts and resigned

In this case, r strongly supports q, which, in turn, strongly supports p. Of course, r could itself be challenged; in that case the discussion, and the chain of justificatory reasons and evidence, might be extended further back to some further consideration s. Absent some such further challenge, however, p is justified, and we are well within our epistemic rights to so take it, on the basis of evidence supplied by q (and, if needed, r). No regress is necessary in order that a given standpoint or claim be justified, and thus the first possibility in MT is avoided. The second, circularity, is also avoided, as the example makes clear. So, too, is the third, arbitrariness: absent a good reason to query q, it is not arbitrary to stop the justificatory chain there; and if there is such reason, but there is no good reason to query r, then it is not arbitrary to stop the justificatory chain at the latter. Cheney’s admission at the news conference, were such an event to take place, would be a non-arbitrary stopping point: it would give us very good reason (though not certainty or anything beyond further critical scrutiny) to believe that he had indeed lobbied the Senators, which would in turn give us good reason to believe that he approved of and encouraged torture by U.S. agencies and personnel. The key point here is that while arbitrariness in selecting stopping points is possible, it is not unavoidable; the example is meant to illustrate the possibility of a non-arbitrary, non-circular stopping point. There are often, as in this example, good reasons for stopping the chain of justification at a given point, in which case arbitrariness is avoided. If these reasons are thought to be unpersuasive, or if other, new reasons for extending the chain are advanced, it is always possible not to stop at that point but to push on and extend the justificatory chain further back. Doing so does not betoken or necessarily involve an infinite regress but rather an unwillingness to regard arbitrary stopping points as justificatory. Once the possibility of positive justification is acknowledged, the possibility of avoiding arbitrariness is, as well.¹¹

This resolution of MT depends on thinking of justification in the second, evidential, sense discussed above. We do not claim that standpoints can be justified or ‘legitimated’ definitively, if that is understood to require certainty; we agree with critical rationalism that fallibilism should be embraced. But we think that critical rationalists (and everyone else) have good reasons for embracing it—otherwise it would not be rational to do so. Taking this point seriously requires van Eemeren and Grootendorst and their fellow critical rationalists to acknowledge that there is good reason to embrace fallibilism—if not, their embrace of it is by their own admission arbitrary, and so unjustified—and so, to accept the possibility that standpoints can be ‘legitimated,’ i.e., justified, by reasons and evidence.

Our resolution of the difficulty raised by MT is not offered as a resolution of the old problem of the ‘structure of justification,’ for we are not offering any such ‘structure.’ We are arguing only that MT can be avoided, since a belief can be justified without involving an infinite regress, a vicious circle, or an arbitrary stopping point.¹²

In this way, we suggest, MT can be overcome, so that van Eemeren and Grootendorst are incorrect when they suggest that justificationists require “a premise... that is immune to criticism.” (2004, p. 131)¹³ Justificationists (in the second, evidential, sense) can and should be fallibilists, too; fallibilism is not the sole property of either critical rationalists or ‘dialectical-ists.’ For this reason, van Eemeren and Grootendorst err when they suggest that a ‘dialectical’ view of reasonableness is the only one that avoids MT:

In our view, it is necessary to depart radically from the justificationism of the geometrical and anthropological approaches to reasonableness and to replace these conceptions of reasonableness with a different one. We do so by adopting the view of a critical rationalist who proceeds on the basis of the fundamental fallibility of all human thought... In this approach, conducting a critical discussion is made the point of departure for the conception of reasonableness—which implies the adoption of a dialectical approach. (2004, pp. 131–132)

As we have seen, justificationism, if understood evidentially, resolves the difficulty. And, as we have argued elsewhere¹⁴, there are independent reasons for embracing an epistemic approach, both to reasonableness, in particular, and to argumentation, in general.

We have interpreted van Eemeren and Grootendorst’s embrace of critical rationalism as involving the rejection of ‘justificationism’ in the second sense discussed above—i.e., as involving the rejection not just of certainty, but of the very possibility of positive support—and argued that understanding it in this way is problematic for the reasons given thus far. Are we guilty of misinterpreting their position? Is the pragma-dialectical view more properly understood¹⁵ as involving the rejection of ‘justificationism’ in the first sense, i.e., as rejecting certainty but accepting the possibility of positive support? We grant that so understanding their position allows van Eemeren and Grootendorst to avoid the criticisms adduced thus far. Nevertheless, we do not think that this interpretation can succeed. First, it makes the invocation of MT a mystery, since MT is supposed by the critical rationalists who invoke it, especially by Popper, to demonstrate the impossibility of positive support, not merely the impossibility of achieving certainty. If van Eemeren and Grootendorst mean to reject ‘justificationism’ only in the first sense, their invocation of MT is pointless. Second, so interpreting their view makes it unclear why van Eemeren and Grootendorst are entitled to the mantle of ‘critical rationalism’—which they proudly wear—given that the label applies to Popper if it applies to anyone, and the hallmark of Popper’s understanding of ‘justificationism’—the rejection of which is central to his critical rationalism—is the second, rather than the first, sense. Finally, so interpreting van Eemeren and Grootendorst renders their account epistemic, rather than pragma-dialectical, since ‘reasonableness’ would on this interpretation rest ultimately on an account of the epistemic goodness of reasons, rather than on the problem- and conventional-validity which are the mainstays of the pragma-dialectical account of reasonableness. For these reasons, we think that interpreting van Eemeren and Grootendorst as rejecting ‘justificationism’ only in the first sense cannot succeed. And as we have argued throughout, it is equally problematic to interpret them as rejecting ‘justificationism’ in the second sense.

4 Conclusion: Toward Epistemic Accounts of Rationality, Reasonableness, and Argumentation

We have argued that van Eemeren and Grootendorst’s accounts of rationality, reasonableness and argumentation are inadequate, and their embrace of ‘critical rationalism’ problematic. The Popperian critique of ‘justificationism’ they endorse as a guide to the normative dimension of argumentation fails: it is right to reject a conception of justification that requires certainty or proof, but wrong to reject the possibility of justification or positive support altogether. Doing the latter makes it impossible to capture that normative dimension. A more adequate account of these matters, we maintain, is provided by the epistemic view we have defended elsewhere.