7 32.2 Botting Reply F 
 © David Botting. Informal Logic, Vol. 32, No. 2 (2012), pp. 269-285. Reply Pragma-Dialectics Epistemologized: A Reply to Lumer DAVID BOTTING Instituto de Filosofia da Linguagem Faculdade de Ciências Sociais e Humanas Universidade Nova de Lisboa Av. de Berna, 26 - 4º piso 1069-061 Lisboa PORTUGAL davidbotting33@yahoo.co.uk I would like to thank Christoph Lumer for his illuminating comments (Lumer 2012) on my paper “The question of truth” (Botting 2010) and would like to reply on a few of the issues he raises. 1. On the Functional Claim The Functional Claim is that “The function of argumenta- tion is to resolve disputes on the basis of the better argu- ment.” In the middle of a protracted effort to show that I have misunderstood and distorted pragma-dialectics, Lumer (2012, 55) objects to the supplement “on the basis of the better argument” as an invention of my own and as something that the pragma-dialecticians themselves would not advocate. It is true that you will probably not find this phrase in any works of pragma-dialectics, but what you will find, over and over again, is that disputes should be resolved “on the merits” (e.g., van Eemeren 2010, 6-7). On the merits of what, if not the merits of the better argu- ment? What you will also find van Eemeren saying is that “the pragma-dialectical definition of argumentation refers the assessment of argumentation to a reasonable judge” where “reasonable” here is defined further as “using rea- son in an appropriate (‘well-considered’) way” (van Eemeren 2010, 29). What is epistemology if not the study of how to use reason in an appropriate way? So, there is an David Botting 
 270
 epistemic norm involved in the definition of pragma- dialectics, and in fact Huss (2005)—a more thoroughgoing consensualist than the pragma-dialecticians—objects to the introduction of the reasonable judge precisely because this concept introduces epistemic normativity into a place he thinks it doesn’t belong. The functional claim also serves as a definition of the speech act complex of argumentation. It is a part of the speech act conditions of argumentation that an arguer is aiming to resolve the dispute “on the merits.” For the pragma-dialecticians this condition is only regulative: the arguer does not argue well, by definition, if he is trying to resolve the dispute by means of an argument he does not consider the best, but he does still argue. This is also stated by the first of van Laar’s conditions on reasonable resolution: “Each has argued by giving what they consider their best arguments” (van Laar 2003, 2; Botting 2010, 418). For the pragma-linguist Bermejo-Luque, it is consti- tutive; for her such an arguer is not really arguing at all.1 Assuming, then, that arguers aim to argue well, they must give arguments that they believe to be epistemically sound, or to put it equivalently, they must give arguments they believe to epistemically objectively justify their con- clusions. When, in my previous paper, I made my claims and put forward arguments, I believed (and still believe) those arguments to objectively justify those claims. I could be wrong, and Lumer has put counter-arguments forward that he believes objectively justify his claim that I was wrong. But we are both aiming at objective justification. We, the arguers, are aiming at qualified consensus of the kind Lumer wants.2 The rules themselves do not aim at qualified consensus but at unqualified consensus in the 



























































 1
That I actually favour Bermejo-Luque’s constitutivist position is implied by what I say at (2010, 417) and especially in footnote 4. Her view that argumentation is constituted by an attempt to justify a target-claim is more transparently epistemological, perhaps, but pragma-dialectics has a similar commitment. While on the subject of transparency, I would like to point out that what I said regarding the rephrasing the Ten Commandments as prohibitions of argumentative moves was that it made them more transparently consistent with critical rationalism (Botting 2010, 415). I did not claim that this rephrasing made a substantive difference to their content. Lumer’s objection that it doesn’t (Lumer 2012, 63-64 ff. 11) may be correct, but this does not contradict what I actually said.
 2
This is, I think, what Lumer terms the external aim of the discussion (Lumer 2012, 58).
 



























Pragma-Dialectics Epistemologized 
 271
 sense that they do not prescribe some inference rules and proscribe others; as rules of procedure they should not try to import substantive theses about what is epistemically or logically good.3 Yet I hoped to show that they supplied a means by which the arguers could discover for themselves when inference rules were epistemically non-normative provided only that the arguers can appeal to the evidence of the senses. This being so, I find it odd that Lumer should re- peatedly make the accusation that argumentation as pragma-dialectics conceives it is epistemically non- normative because it aims only at agreement and that pragma-dialecticians are simply not concerned with truth or justification. The problem, rather, is that the arguers themselves are not capable of distinguishing between what actually is objectively justified and what they believe to be objectively justified. The advice “Follow rules that are ob- jectively valid” is good advice only on the assumption that we know what is objectively valid, distinguishing it where necessary from what we only think is objectively valid. This advice is useless, because by our own lights—and assuming we are being reasonable by offering what we think are our best arguments—this is what we are doing anyway! This is a point well-made in Huss (2005) and discussed in my paper (Botting 2010, 427). Compare also the following well-known quote of Hamblin (1970, 244; italics in the original): The logician does not stand above and outside practical argumentation or, necessarily, pass judg- ment. He is not a judge or court of appeal, for there is no such judge or court: he is, at best, a trained advocate. It follows that it is not the logician's par- ticular job to declare the truth of any statement, or the validity of any argument. On the one hand this seems absurd: if it is not the job of the logician to determine the validity of arguments, what on earth is? But Hamblin's point is about the role that should be played by the logician, and by logic as the logi- cian knows it, in the actual argumentation taking place; and Hamblin says that the logician cannot stand outside the argumentation and dictate as an external criterion that 



























































 3
This, I think, corresponds to what Lumer terms the weak meaning of “unqualified consensus” (Lumer 2012, 75).
 David Botting 
 272
 arguers must use argument-forms that the logician himself may know (but perhaps the arguers themselves do not know) to be valid. What goes for the logician goes equally for the epistemologist, and comes down to whether there is an epistemically privileged viewpoint, a viewpoint from which epistemic goodness can be applied as an external criterion. Much of the debate between Kaufmann and Nagel is over precisely this point. Lumer comes down on the side of Nagel: “epistemological correctness is not relative to current rules—but to objective epistemological principles that guarantee the conclusion’s truth or epistemic accept- ability” (Lumer 2012, 71); “the reason magic is inferior to science is not because there is a shared rule whose appli- cation shows that it is, but because the former fails in terms of an objective criterion” (Lumer 2012, 74); and most tellingly (Lumer 2012, 72-73): Given how these operators are defined it is impos- sible that if the premises of a valid inference are true its conclusion is false.... The validity of these rules, however, i.e., that such rules from true prem- ises always lead to a true conclusion, of course, is independent of whether a particular interlocutor ... is willing to concede the validity of logical rules ... [and] independent of the historical discovery of the valid inferences and of formal logic ... In this sense the validity of these rules is objective. So valid logical inferences rules exist as Platonic entities as a consequence of the definitions of the logical op- erators and independent of their discovery. One can concede that something can be objectively true and objectively valid quite independently of any- body’s agreeing or noticing that it is objectively true and objectively valid. Kaufmann never says otherwise, and neither do I; Lumer has effectively changed the subject, taking “relative to rules” to be equivalent to “relative to consensus” or some such. Kaufmann’s point is rather that ultimately “objectivity” and “epistemic correctness” them- selves are defined relative to higher-order rules and ulti- mately relative to how we choose to use the word “know- ledge.” The rules we discover are not objective uncondi- tionally, but conditionally. Substitute the word “argumentation” for the word “knowledge” and we have the whole debate between the 



























Pragma-Dialectics Epistemologized 
 273
 pragma-dialecticians and epistemologists—and for that matter, the debate between any two competing instrumen- talist conceptions—in a nutshell. Choosing between these conceptions requires, Reichenbach says in the case of “knowledge,” not further philosophical analysis but a de- cision in light of its consequences to define it one way ra- ther than another. Perhaps Lumer would claim that how we choose to define “knowledge” is independent of what knowledge actually is, that the concept of knowledge is itself a Platonic entity awaiting discovery. This is a kind of essentialism that a justificationist such as Reichenbach would strongly reject. If the geometrical concept of know- ledge has become inadequate to our use of the word ‘knowledge,’ especially our best use as exemplified in sci- ence, then, Reichenbach argues, what we need to do is change our concept of knowledge. Such a problem does not call for further philosophical analysis but for a deci- sion, a co-ordinative definition. Relative to that definition, our entailed decisions follow analytically and define what is objective, but this falls short of the kind of objectivity that Lumer seems to want. Trying to provide further philo- sophical reasons will be futile and simply talk past its competitors. (I will come back to this point when discuss- ing Lumer’s claim that my account is intuitionistic.) Having defined ‘knowledge’ a certain way, rules can be discovered that are, relative to that definition, analyti- cally true and “objective.” The definitions of the logical operators are such rules, and Lumer is quite right to say that the validity of inferences is a consequence of how they are defined. The thing is that their definitions are a consequence of the definition of knowledge. In one sense, the definitions are neither objective nor non-objective, for they define what it is to be objective at the level of rules at which they occur. Their objectivity or correctness can only be decided relative to rules and definitions belonging to a higher level. Luckily for us, most of us mean much the same thing by ‘knowledge’ and have mostly the same rules, and this allows us to argue with each other and re- solve our disagreements without, for the most part, bring- ing exactly what we mean by ‘knowledge’ into question, but these are internal rules that we share as a matter of empirical fact, perhaps because of how we are constituted as epistemic agents. Rules are not external criteria or “ob- jective epistemological principles” in the way Lumer wants to use that phrase. David Botting 
 274
 Progress is measured by how many of the proposi- tions that we want to incorporate—in particular, observa- tion statements—actually are incorporated in an internally coherent corpus. To the degree that we can agree on these, we can agree that one system of rules performs better than another. Sometimes one system of rules performs so well that it gets given a special status and gets called “objec- tive.” Explaining progress in terms of objectivity would then be circular. Objectivity of the kind Lumer wants is an entirely fugitive quality. We can compare this with what Kaufmann (1948) says about truth; truth as it is conceived by the correspondence theory lacks “procedural signifi- cance” and becomes significant only on a coherence theory, or in other words, we call those propositions “true” that are coherent with other propositions we call “true” and we can do no better than this. Truth as the corres- pondence theory defines it is also a fugitive quality. Both Popper and Reichenbach revise the concept of knowledge so that its extension actually includes most of the things we would normally take ourselves to know, which the geometrical conception, after the attacks of Car- tesian doubt, did not. Their revised conceptions are actu- ally extremely similar. Reichenbach claims that know- ledge is a system of posits, Popper that it is a system of conjectures. It is important to realise that neither posits nor conjectures can objectively be given either a truth-value or a probability-value. Posits can be appraised conditionally on higher-level posits, but the highest-level posit— knowledge itself—is a blind posit that does not even have a probability-value. 2. On justification What is justification? One description is that it is a norm about what you should believe. What should you believe? Ideally what is true, and generally what is likely to be true. But according to Hume, matters of fact cannot be given a truth-value or a probability-value. This being so, when we infer some further proposition from matters of fact we do not say that what has been inferred is true or likely to be true either. Inference only preserves truth or probability- values and cannot preserve what is not there in the first place. Should we say that what has been inferred is justi- fied by what it has been inferred from? If by saying this 



























Pragma-Dialectics Epistemologized 
 275
 we are committed to saying that we should believe what has been inferred, then we should not say that it has been justified, for this would amount to saying that it is true or probably true or approximately true, and we cannot say this. This seems to be Popper’s position and that of pragma-dialectics also, with the result that logic can be used only “critically.” Note, though, that foregoing justification does not mean that we cannot use inference or that we have to use inference rules differently. It means only that the attitude we take to what has been inferred in this way is not that of belief or even tentative belief except in the case of falsifi- cation, when one should not believe what has been falsi- fied. It is instead a practical, methodological attitude that can be called acceptance. Having accepted p and that q can be inferred from p, a falsificationist is still forced to accept q. The difference between falsification and justifi- cation is not that falsification can use modus tollens but not modus ponens whereas justification can use both. The main difference between them that concerns us here (of course there are others) is over what attitude we should take to what has been inferred. This is important because it means that as far as modelling them dialectically is con- cerned, they are indistinguishable, because both beliefs and acceptances have the same externalization. My claim that critical discussions model the process of conjecture and refutation was perhaps a bit misleading in this respect; it models justification equally. What makes it falsification- ist is only that the starting-points are analogous to Popper's basic statements and do not need to have any epistemic status; they are a matter of consensus only, mere agree- ment being sufficient to establish them. Reichenbach’s blind posits are actually in the same position. All the probabilities that he uses are actually conditional probabilities. Simplifying somewhat, what he ends up saying is that justification is not a norm of belief but a norm of our best wager, and conditional probabilities can tell us what is our best wager given some other propo- sition. Conceived this way, what has been inferred is justi- fied because it has a high conditional probability (in the case of a deductively valid inference, this is unity) relative to what it has been inferred from. What does Lumer mean by justification? He (2012, 67) says: “... van Eemeren and Grootendorst ... incoher- ently ... accept positive justification, albeit simply on the David Botting 
 276
 basis of shared premises.” Despite his denials, this seems to me to be a verbal matter. For Popper, “justification” on the basis of shared premises isn’t justification at all, be- cause sharing or agreeing to a premise is not the same as saying that it is true or likely to be true, as Lumer is so fond of pointing out. This doesn’t mean that Popper can- not say that we should also share or agree to what has been “justified” in this way, as already said. This kind of positive justification (a kind that Lumer seems to endorse) is one that it is in no way incoherent for critical rationalists to appeal to. Thus, the reason that I do not comment on the Münchhausen Trilemma—for which Lumer (2012, 67) takes me to task and accuses me of not answering to the point—is that I believe that it is irrelevant to the moral that the epistemological critics erroneously draw from it, namely that pragma-dialectics helps itself to a notion of positive justification that it is not entitled to and by doing so is incoherent. It is a purely logical point irrespective of its truth or falsity that the Münchhausen Trilemma is not incompatible with what answers to Lumer’s definition of justification. The appearance of incoherence is because Lumer has a more permissive usage of the term ‘justifica- tion’ (one that seems to encompass any use of modus po- nens whatever) than critical rationalists do. Labouring the point, the difference between justifi- cation and falsification is a question of what kind of atti- tude should be taken to propositions: belief or some kind of practical attitude such as acceptance. It is not a pro- cedural distinction and is lost in a dialectical model where beliefs and acceptances may be externalized equally as assertives. In both cases we make predictions and then see if they come true. Hence, if Lumer’s accusation against pragma-dialectics is that what it actually models is justifi- cation and not falsification, he finds a difference where there isn’t one. As said above, his own use of the term ‘justification’ collapses the distinction anyway. There is another point that should be made here. Supposing for the sake of argument that the Münchhausen Trilemma is false and that justificationism is possible, it does not follow that critical testing procedures are epis- temically non-normative. Granted, it may be epistemically inferior to justificationism, although generally speaking the best confirmed hypotheses will be ipso facto the best corroborated. Yet Lumer makes stronger claims than this: he claims that there is no connection at all between rea- 



























Pragma-Dialectics Epistemologized 
 277
 sonable resolution on the merits and epistemic virtue, which amounts to saying that there is no connection be- tween trying to achieve something and actually achieving it. This, I tried to argue, was based on an over-demanding conception of epistemic virtue and of what is sufficient to make a process epistemically non-normative. This leads me to one of the main epistemological criticisms of pragma-dialectics made most pointedly by Biro and Siegel and which I was trying to address in my paper. Their argument goes something like this: (1) Arguers can agree to use argumentation-schemes that are not epistemically normative, i.e., that are falla- cious. (2) Therefore, the resolution that they reach will also not be epistemically normative (except by accident). (3) Therefore, the rules of critical discussion, since they place no epistemic constraint on what argumentation- schemes are agreed to, are likewise not epistemically normative. This argument contains a subtle equivocation. It is not the same thing for rules to be non-normative as it is for a result of applying the rules to be non-normative. For rules to be epistemically normative they must converge on the truth on repeated application, and to show that they are not epistemically normative it must be shown that they do not do so. Showing only that they do not do so for some arbitrary case is logically insufficient. This, again, is a purely logical point against Biro and Siegel’s argument that I have yet to see addressed. Kaufmann offers a positive argument that a proced- ure of this kind does converge on the epistemically better results provided that both parties share a rule in common that appeals in some way to the evidence of the senses. This was the point of the discussion of the merits of sci- ence over magic (Botting 2010, 429-431). Note that the common rule is not a direct appeal to experience, since Kaufmann denies that there is any such thing. Experience functions as the grounds of propositions (derived from the experience by a rule) whose inclusion in the corpus creates too many contradictions and leads eventually to modifica- tion of the rules. If magic is observed to not “come true” then science can suggest its alternative set of rules as be- David Botting 
 278
 ing more able to incorporate propositions that both sides agree should be in the corpus. This point does not mean that without the alternative of science, magic could go on indefinitely without any problems; it comes under an internal tension created by its own rules. The same, I argued, would happen to a group of people who started off agreeing that the gambler’s fal- lacy is a valid inference rule. When their prediction failed to come true, they would either have to reject the evidence of their senses (implausible), take the trials so far to be an unrepresentative sample (which is increasingly unlikely as the sample size increases, i.e., when the rules are applied iteratively in a progressive series of critical discussions) or tweak the rule so that the outcome is what would have been expected (ad hoc, and will soon lead to the same problem again). Of course, logically speaking, one can by such tweaks protect oneself forever from falsification, as Duhem pointed out—before pointing out that although it was logically possible, this did not mean that it was ra- tional to continue to do so. It might be objected that this argument itself com- mits the gambler’s fallacy of assuming that a random pro- cess will self-correct. This would be amusing if it were true. But argumentation is not a random process; unlike throwing dice, each critical discussion is not statisti- cally/causally independent of those preceding it. It is a cumulative process, as I think Lumer would agree—one where the participants may learn from the past.4 This explains “why historical development merely on the basis of shared premises and methodological rules produces a progress in these terms” (Lumer 2012, 71). Now Lumer will say that this is explained by the fact that appeal to experience just happens to be objectively norma- tive. I am not sure about this; it seems to me that learning from experience is a presupposition of epistemology and of knowing, rather than something that epistemology has 



























































 4

 Learning from past experience does not mean constructing a frequency series from which one then draws probabilistic conclusions. People learned from past experience long before the frequency interpretation of probability, and arguably before they had the concept of probability itself. The claim that a system of magical thinking could not self-correct because it could not actually formulate this kind of probabilistic argument is wide of the mark (Lumer 2012, 74). This does not prevent us from explaining such learning in terms of such a probabilistic argument, however.
 



























Pragma-Dialectics Epistemologized 
 279
 to establish. I find it difficult to imagine a critical discus- sion or system of rules that does not have a rule by which thinkers accept into their corpus propositions derived from the evidence of their senses. One can perhaps ignore or put aside this evidence for a while (the first option described above) but this cannot go on indefinitely while remaining rational. Is such a scenario coherent even as a thought- experiment? For if one is to agree with someone about something, that thing must be made present to conscious- ness, either to perception or to reflection. Indeed, the intersubjective procedures that pragma-dialectics puts forward to decide, for instance, whether inference rules have been used correctly, seems to presuppose that this correctness is jointly observable in some sense. The argument above, it should be noted, does not rely on giving any special status or epistemic privilege to science. The point is only that science does not often come under this kind of internal tension (although it can do and in times of revolutionary science, it does). What Lumer (2012, 70) calls S5 (“Transfer to discussions in general”) and his criticism of it (2012, 75) actually misses the point of the argument. In summary, I claim that the rules of pragma- dialectics are epistemically normative because they pro- vide a self-correcting mechanism on those material start- ing-points and inference rules agreed in the opening stage of the critical discussion when used iteratively. Contrary to what Lumer says in (2012, 77 ff.16), it is because of this ability to self-correct, a consequence of following the pragma-dialectical rules, that the starting-points and infer- ence rules we agree to are, most of the time, epistemically good. It is not a mere happy accident. The rules have epi- stemic normativity inherently and not incidentally (Bot- ting 2010, 414). 3. On intuition I am curious where Lumer got the idea for my so-called intuitionistic defence of the Functional Claim. A search through my paper reveals no occurrence of the words ‘in- tuitionistic’ or ‘intuition,’ and only one occurrence of ‘in- tuitive’ (used not by me but in a quotation from Habermas in footnote 7). The argument was in response to a chal- lenge that I have always thought got things backwards, David Botting 
 280
 namely whether pragma-dialectics was justified. Pragma- dialectics is not justified, but this is not a defect; the whole point of critical rationalism is that you don’t justify your hypotheses (because you can’t) but put them forward as conjectures—or “mere allegations” if Lumer (2012, 57) finds this phrase more to his liking—and then try to falsify them. A good hypothesis, the critical rationalist continues, is one that provides its own means of falsification. Also, so as not to be thought to be designed in an ad hoc way to fit the evidence so far, it should predict new phenomena. We can only falsify a hypothesis via establishing proposi- tions that can be established independently of the hypoth- esis. This does not mean that we have to establish them by intuition. The Functional and Instrumental Claim combined is a good hypothesis. It is critically tested by comparing it against how well it captures our judgments about good and bad reasoning (which may be intuitive but are not necessa- rily so; may be culture-bound but are not necessarily so). Some forms of bad reasoning occur often enough to be labelled as fallacies. Our judgments may recognize them to be fallacies rather than merely bad, but may not; we need not presume familiarity with fallacies. But if the only way of determining whether or not some piece of reason- ing is fallacious or bad is by appealing to the hypothesis then this is obviously no kind of test at all (see the quota- tion from Habermas in Botting 2010, 420 note7, for an elucidation of what independence really amounts to). Cer- tainly, ‘fallacy’ is a theoretical term and, if the fallacy is a fallacy, this is an analytic consequence of the hypothesis, but we can still test it empirically provided that we are able to identify extensionally what the theory must ac- count for intensionally. Are we actually able to identify fallacies extension- ally? This is perhaps the most contentious point in my ar- gument and I admit it is one against which one can take pot-shots. For instance, the evidence is that we are very poor probabilistic reasoners, so we are not able to identify probabilistic/inductive fallacies very well and often take reasoning to be good that is bad by the lights of proba- bility theory (although some naturalized epistemologists might challenge whether this reasoning is really bad for beings of limited cognitive resources—I myself do not be- lieve that the inclusion fallacy, for example, is really a fal- lacy). Also, at different points in history theorists have 



























Pragma-Dialectics Epistemologized 
 281
 categorized and explained fallacies differently, but this, it should be noted, is not necessarily an extensional differ- ence. Equally, as time has gone on, more fallacies have been added to the standard lists, although this does not mean to say that before they were added to the list people thought that they were good reasoning, and if their not thinking it bad is explained by the fact that the kind of rea- soning or argument had not been invented yet (as is the case, for example, of the base-rate fallacy) then obviously this is no objection either. What I think is needed is an ex- ample of some reasoning that everybody thought was bad and is now considered good. Even this would not be com- pelling, because, to make a similar point to one made earlier, our judgments may be normative even if they are wrong or unreliable about some particular cases. Lumer (2012, 58-59) claims that there are extensional differences. But if these differences are significant enough to support his contention, then this seems to amount to admitting that we can’t differentiate good reasoning from bad in the first place—in which case I am at a loss to know exactly what it is that the hypothesis is meant to explain. As well as ac- counting for the judgments we already have, the Func- tional and Instrumental Claim predicts new fallacies. These are empirical tests, as I make clear (Botting 2010, 419-20). I wonder what exactly Lumer (2012, 56) means when he says that this is only an idea and that I “in no way [proceed] to its realisation, showing that the pragma- dialectical rules and the Functional Claim would really be the outcome.” If he means that I have not actually con- ducted this empirical testing then he is, of course, correct.5 Theorizing does not occur in a vacuum but must have some pre-theoretical judgments from which an at- tempt to extract formal conditions can be made, that can then be tested against more judgments, and so on. Un- doubtedly, sometimes we end up concluding that some pre-theoretical judgments are wrong and do not stand up to considered reflection in light of the theory, but gener- ally we consider them at minimum a good starting point. I 



























































 5
Empirical testing of a kind has been done by the Amsterdam School and its results been published as “Fallacies and Judgments of Reasonableness” [van Eemeren, Garssen and Meuffels (2009)]. Curiously, however, what it seems they want to establish is only the conventional validity of some of the rules. They could argue towards problem-validity by assuming that most judgments are normatively correct, but for some reason they do not take this additional step.
 David Botting 
 282
 get the feeling that Lumer seems to want to start the other way around (see esp. Lumer 2012, 61 note 6, where he says that his theory does not aim at providing the best ex- planation of our “given” intuitions)! None of this, it should be clear, implies criticism of the instrumentalist approach that both the epistemologists and the pragma-dialecticians take, nor should it be taken as an alternative to that approach, as Lumer (2012, 57-60) seems to take it. (Although neither does it endorse that ap- proach; the issue is simply orthogonal to the points I wish to put forward or attack.) Describing one’s theoretical mo- tivation to take the instrumental goal of argumentation as this or that (as qualified or unqualified consensus) will not justify the choice of goal and will be persuasive only to those who already define ‘argumentation’ in a particular way and will beg the question against those who define it differently. This is why in the fictional dialogue “van Eemeren” immediately sets aside this kind of purely theo- retical discussion (Botting 2010, 421) to focus instead on what they have in common, considering the Ten Com- mandments and the rules “Lumer” suggests as rival hy- potheses where adjudication between them depends on the results of using those rules. Without some way of compar- ing their results (which I suggested was decided by how well they capture fallacies in real discourse) there is no way to choose between them, since relative to their own definitions they have equal claim to being right. They end up talking past each other and neither one will ever accept the other’s justification. My point is that this is entirely consistent with the critical rationalist approach and that engaging in the kind of theoretical discussion Lumer wants to engage van Eemeren in would not be consistent; Lumer’s insistence that they justify themselves in this way is illegitimate, and they would be being inconsistent if they did justify themselves in this way. Prior to being es- tablished empirically, arguments in favour of the problem- validity of the rules, and to convert others to their adop- tion, must be pragmatic in nature, and this is what van Eemeren gives.6 



























































 6

It seems to be in this vein that van Eemeren and Houtlosser write (2002, 133): On a philosophical level, would-be arguers have a ‘pragmatic’ rationale for accepting these procedures as guiding principles. The acceptability of the procedures is not derived from any external source of authority or some metaphysical necessity, but depends on their suitability for resolving a 



























Pragma-Dialectics Epistemologized 
 283
 4. Conclusion I will now sum up what I think is most important in my discussion. First of all, the via negativa: (a) Epistemically principled adoption of pragma-dialectics is compatible with its being “unjustified.” (b) A non-normative result is compatible with a normative set of rules. (c) “Positive justification” in Lumer’s sense is compatible with the truth of the Münchhausen Trilemma. These are purely logical defects in the arguments of the epistemologists and I feel quite entitled to point them out even though I am not, as Lumer says, a member of the Amsterdam school. I have yet to see these points ad- dressed. My via positiva is this: (d) What makes adoption of pragma-dialectics principled is that it submits itself to critical testing against em- pirical data that can be established independently, viz., judgments of reasonableness. (e) What makes the rules normative is that pragma- dialectics self-corrects agreement to inference rules that are epistemically bad, provided that the senses are agreed to be an epistemically good source of evi- dence. To be fair, Lumer gives what seems a recapitulation of my argument where the problem of standing outside of the argumentation and imposing epistemic constraints on the discussion rules themselves is stressed. He responds to this problem by pointing out (1) that the discourse rules he gives are open to revision and designed by experts or peo- ple in epistemic authority, and (2) that pragma-dialectics does not have such provisions because it is focussed in- stead on achieving agreement (Lumer 2012, 79-80). (Note that his discourse rules contain substantive epistemic con- straints about the goodness or badness of certain argumen- 




















































































































 difference of opinion on the merits for which they were designed (van Eemeren & Grootendorst, 1988). Viewed philosophically, this rationale for judging intersubjective or conventional validity may be called pragmatic because pragmatists characteristically decide the value of any proposal on its contribution to solving they are out to solve.
 David Botting 
 284
 tation-schemes, and are thus oriented towards qualified consensus and against even the weak kind of unqualified consensus I defend.) Now, if the epistemically constrained discourse rules really are epistemically normative, then it is a tautology that the consensus it produces is more likely to be epistemically normative than one produced by purely procedural rules that are not so constrained. Although a tautology, it is a tautology relative to a higher-level defini- tion, and the objective normativity of the rules is estab- lished only via agreement that its results are, indeed, true, at least most of the time (see what I said previously con- cerning the procedural significance of objectivity and truth). Notwithstanding this whiff of circularity I am (as Lumer notes) ready to concede or at least leave open the epistemic superiority of a system of rules of this type, but I wanted more modestly to claim that a system of rules not so constrained could still be genuinely, inherently norma- tive in the sense of being truth-conducive, and that argu- ments given to the contrary were logically insufficient. I also don’t object to some argumentation schemes being buttressed, for example, by the best practice of experts. My feeling, however, is that these things are among the material starting-points—the accumulated wisdom of past critical discussions—and not built into the discussion rules. Acknowledgment: The author would like to acknowledge financial support received from the FCT under the grant ‘‘Argumentation, Communication and Context’’ PTDC/FIL-FIL/110117/2009. References Bermejo-Luque, L. (2011). Giving reasons. Dordrecht, Heidelberg, London, New York: Springer. Botting, D. (2010). A pragma-dialectical default on the question of truth. Informal Logic 30: 413-434. Eemeren, F. H. van and Houtlosser, Peter. (2002). Stra- tegic maneuvering. In Dialectic and rhetoric: the warp and woof of argumentation analysis. Frans H. van Eemeren, Frans H. and Peter Houtlosser (Eds.). Dordrecht, Heidelberg, London, New York: Springer. 



























Pragma-Dialectics Epistemologized 
 285
 Eemeren, F. H. van, Garssen, Bart and Meuffels, Bert. (2009). Fallacies and judgments of reasonableness. Dordrecht, Heidelberg, London, New York: Springer. Eemeren, F. H. van. (2010). Strategic maneuvering in ar- gumentative discourse. Amsterdam/Philadelphia: John Benjamins Publishing Company. Hamblin, C. L. (1970). Fallacies. London: Methuen. Huss, B. (2005). Useful argumentation: a critique of the epistemological approach. Informal Logic 25(3): 261- 275. Kaufmann, F. (1948). Three meanings of “truth”. The Journal of Philosophy 45 (13): 337-350 Laar, J. A. van (2003). The dialectics of ambiguity. [Available on- line.]http://dissertations.ub.rug.nl/faculties/fil/2003/j.a .van.laar. Last accessed 18th May 2010. Lumer, C. (2012). The epistemic inferiority of pragma- dialectics. Informal Logic 32(1): 51-82.