INFORMAL LOGlC XILl, Winter 1990 Inconsistent Commitments and Commitment to Inconsistencies * ERIK C. W. KRABBE Groningen University It is no good to contradict oneself. For one thing, the perpetrator would flout the time- honored principium contradictionis (Law of Contradiction): "It is impossible for the same attribute at once to belong and not to belong to the same thing and in the same relation." (Aristotle, 1975, Metaph., lO05b I9 ). Aristotle assures us that "this is the most certain of all principles" and that to demand to have the law proved would show a "lack of education" (1006aS-7 ). On the other hand, if someone would be bold enough to challenge the principle, he can be refuted, if only he makes some state- ment. If not, we may sit back in comfort, for our opponent would be "really no bet- ter than a vegetable" (lO06a tS ). From the discussion that follows, it is clear that, ac- cording to Aristotle, a meaningful use of language presupposes the Law of Contradic- tion to hold. Ultimately, if your opponent denies everything he asserts, and asserts everything he denies, further discussion makes no sense. "For he says neither 'yes' nor 'no', but 'yes and no'; and again he denies both of these and says 'neither yes nor no' ... " (l 008a 3 t -34). This opponent does not really make a statement (another plant, so it seems). At this point one may ask whether serious opposition to the Law of Contradic- tion is at all feasible. Aristotle adds an argu- ment from practice: everyone judges one course of action better than another. ("Otherwise, why does a man walk to Megara and not stay at home, when he thinks he ought to make the journey?" (lO08b I4- IS». Hence "all men form un- qualified judgments, if not about all things, at least about what is better or worse." (lOO8b26-28 ). Consequently, it seems a practical impossibility to deny the Law of Contradiction in the strong sense of taking a contrary position (every attribute belongs and does not belong to the same things). How about a denial of the Law of Con- tradiction in the weaker sense of holding its contradictory (at least one attribute could at once belong and not belong to the same thing)? Let us skip a few centuries and have a look at another specimen of horror con- tradictionis. A serious opponent of the Law of Contradiction has to be deliberately in- consistent. Is this possible? The question is discussed in the first chapter of a well- known logic text: Is consistency of beliefs a virtue? To some extent this is a question in a vacuum. Nobody is deliberately inconsistent in his beliefs. It is simply impossible to believe, fully and without reservation, two things which you know are inconsistent with each other. Exercise I A. You know that human beings normally have two legs. Try to convince yourself that they normally have five. (Allow yourself at least a minute.) It seems we are obliged to believe only what we think is consistent, without having any real choice in the matter. In this way we are all logicians, simply because we are human. (Hodges. 1977: 15) We cannot get ourselves to hold inconsis- tent beliefs. But that is hardly surprising. For, probably, one is just not in a position to get oneself to believe anything, whether consistent or not. The more important point is whether Hodges is right in maintaining 34 E. C. W. Krabbe that nobody will deliberately stick to his in- consistent beliefs, once they are pointed out as such. This view would imply that we can- not seriously deny the Law of Contradic- tion at all. For, since the inconsistency of P and not-P is wholly explicit, there remains nothing to be pointed out, and no one can seriously believe both to be true. A further consequence would be that whosoever nevertheless asserts that both P and not-P hold, cannot honestly believe what he says, nor can he believe that his audience believes he does. He is flouting a Conversational Maxim (viz., Grice's First Maxim of Quali- ty, 1975:46). Certainly, that is not to say that contradictory assertions (P and not-P) are pointless, but rather that they can be us- ed only in indirect speech acts, if they are to make sense. "She loves me, no she loves me not" could express one's torment. "The cow exists and she doesn't exist" could be used to start a discussion on existence, or to show off sophistication. But let us leave these indirect speech acts aside. Is it really excluded that one seriously and literally asserts P and not-P? Incon- sistency and contradiction are in vogue nowadays, even among formal logicians. Rescher and Brandom refer to the Recent Period of Inconsistency Toleration (1980:56). They distinguish four modes: (A) (Weak Inconsistency) To accept the prospect that for some genuinely pos- sible world w: P true in wand not-P true in w, for some P. (B) (Strong Inconsistency) To accept [etc.]: (P and not-P) true in w, for some P. (C) (Hyperinconsistency) To accept [etc.]: (P and not-P) true in w, for all P. (D) (Logical Chaos) To accept [etc.]: P true in w, for all P (and accordingly P true in wand not-P true in w, for all P) (1980:24).1 Of these modes of inconsistency tolera- tion Rescher and Brandom champion only the first: Weak Inconsistency. Thus their tolerance appears to be quite restrained. They fulminate against the stronger modes in terms Aristotle would appreciate: ... surely nothing could (rationally) move us to accept an outright SELF-contradiction [viz,. (P and not -P)] . .. no cogent line of con- sideration could ever move us to accept such a contention ... In uttering an outright self- contradiction we literally 'cannot say' what we have in mind to assert. (1980:24,25) For a book aiming at the vindication of in- consistent (and incomplete) possible worlds this passage airs a remarkable degree of horror contradictionis. An unwarranted degree, so it seems, in view of the existence of more tolerant systems, like Minimal Logic. According to the semantics of Minimal Logic we can face Strong Incon- sistency, Hyperinconsistency and Logical Chaos without any fear of our logic getting chaotic. Further more, Rescher and Brandom won't tolerate any inconsistencies on the metalevel: "Keep your own discus- sion consistent" (1980:139). Though we may be brought to accept that 'both P is true in wand that not-P is true in w', we still must reject statements of the form 'P is true in w but P is also not true in w'. Granted that their toleration of incon- sistency is limited, the example of Rescher and Brandom at least shows that some types of inconsistency are seriously discussed by contemporary logical semanticists, and that not all inconsistencies are equally out of place. Even toleration of Weak Inconsisten- cy seems a long way from Aristotle (and from Hodges). Thus we confront a long tradition of horror contradictionis and re- cent steps towards inconsistency toleration in logical semantics, two trends that seem to conflict or to contradict each other. The problem of inconsistency can, I think, most profitably be approached from the perspective of dialogue and argumen- tation. Leaving aside psychology (Can I get myself to believe P and not-P?) and on- tology (Is there a possible world w such that both P and not-P are true in w?), we should foremost ask: What should happen if (ap- parent) contradictions or inconsistencies arise in dialogue? The question is nor- mative: What should an ideally constructed system of rules for argumentation (an op- timal system of dialectics) do about incon- sistencies? What about a party in dialogue that contradicts itself? Should this situation be ruled out beforehand? Should this party be penalized? Or should it be allowed to get away with it in some cases? Ultimately, such questions can be answered only in the context of a com- prehensive Theory of Argumentation. One cannot definitively tackle one problem (in- consistency) in isolation from other pro- blems (equivocation, evasion, irrelevance, filibustering, ad hominem, etc.). In an op- timal system of dialectics everything hangs together. 2 From the perspective of one pro- blem in isolation (here: inconsistency) one may, however, give a preview on how the rules of dialectic are to cope with certain situations, and that is as far as we shall get in this paper. Stating that P, in dialogue, amounts to committing oneself to P (propositional com- mitment). The content of this commitment depends on the type of dialogue (persuasion dialogue, negotiation, inquiry, ... ), on the role of the speaker (opponent, proponent, questioner, defendant, ... ) and on connec- tions with other utterances in the dialogue. In persuasion dialogue it may be that a state- ment of P commits one to offer arguments for P (if challenged), or it may be that one is merely committed to allow the other party to use P in an argument. Stating that not-P in the same dialogue leads to different com- mitments. Do these commitments conflict? Is commitment to both P and not-P to be compared to commitments to incompatible tasks (a quandary)? If so, who is to blame? The party thus committed, or the advocate of the dialectic system that allows this situa- tion to occur? Let us have a closer look at the 'anatomy of commitment'. 3 Every commitment has the following three aspects: (1) a subject, (2) a commitment bond, (3) an object (or content). The subject of a commitment is a person or a collective (union, institution, party, etc): the one who is bound by the commitment. The bond itself can be Inconsistency and Commitment 35 characterized as ethical, legal, conventional, or by refinements of such terms. The im- portant thing in characterizing a type of commitment bond, is the sanction: what will happen if I fail to live up to my commit- ment? The object of a commitment, final- ly, is characterized by the imperative tell- ing the subject what to do in order to live up to his commitment. Here is an example: A student (Sarah) hands in a paper with her professor (Pamela) to be graded. In ordinary circumstances this act of Sarah's leads to a commitment, with Pamela as the subject and 'Grade Sarah's paper!' as the associated imperative. The bond can be characterized as based upon professional standards. If Pamela should fail to grade Sarah's paper, that would amount to her performing bad- ly in her job. (That would be the sanction.) The close connection between com- mitments and imperatives allows us to apply Hamblin's theory of imperatives, especially his 'Action-State-Semantics' (1987). Accord- ingly, let us identify the object of a com- mitment with the set of (partial) strategies one of which the subject should follow (or at least try to follow) in order to wholeheartedly satisfy the imperative associated with the commitment. Let us par- tition the objects of commitment according to the nature of the commitment bond (the sanction). Two objects of commitment of a subject go into the same sanction-set if and only if the commitments are enforced by the same sanction. The set of all of a subject's sanction sets may be called its agenda (at a certain moment of time). The present theory of commitment in general applies to propositional commitment as well. Propositional commitments, however, are special in two ways. First, as to their content: the associated imperatives of propositional commitments tell the sub- ject to perform, or refrain from, certain moves in dialogue (Defend P! Do not challenge P! Criticize P!). Second, these dialogical moves, or acts, all pertain to a proposition P (which mayor may not have been uttered in the dialogue). 36 E. C. W. Krabbe In persuasion dialogues we must distinguish (at least) two sanction sets. Delinquency as to commitment of the first set consists of breaches of the rules of the specific type of persuasion dialogue in which the commitment was incurred. Following a recent proposal I shall call such breaches: fallacies. 4 Delinquency as to com- mitments of the second set consists of weak play in dialogue: unneccessary weak arguments and weak criticism. I shall refer to such acts as weaknesses or, if a strong term is needed, blunders. Thus we may compare committing a fallacy in dialogue with a 'move' in contravention of the rules of some game, like chess, and a blunder in dialogue with a permitted but poor move in the same game. It is well known, both from moral philosophy and practical life, that com- mitments that seem to be consistent at first, may nevertheless by a conjunction of cir- cumstances get one into serious trouble. For instance, suppose Pamela is bound to grade Sarah's paper, but also to help her aunt in case of an emergency. These commitments need by no means conflict. But suppose Pamela's aunt falls seriously ill all of a sud- den, unfortunately on the very last night available for grading Sarah's paper. (Sup- pose the grading of the paper would at least take one night.) Under these circumstances Pamela's commitments clash. There is a prima facie conflict of duties. But perhaps Pamela is able to assign priority to one com- mitment over another. In that case she will be able to resolve the conflict and to act ac- cording to some partial strategy her prefer- red commitment provides. Otherwise, she is in a quandary. The concept of a quandary is discussed extensively by Hamblin (1987, Ch.5: The Consistency of a Set of Imperatives). Especially interesting is Hamblin's descrip- tion of four "kinds of quandary-freedom" a set of imperatives may possess (1987, 177 -181). Translated to quandaries that arise from commitments in dialogues, we get the following list of degrees of quandary-freedom a set of dialogue rules may guarantee (1 have added 2a): (1) Absolute Quandary-Freedom: the rules of dialogue do not allow quandaries to arise under any circumstances. Of course one cannot stipulate quandary- freedom in the rules. The rules must be carefully formulated (and, if necessary, assigned priorities) in order to provide a legal (nonfallacious) act for each disputant in every possible situation (even if fallacies are essential for getting into that situation). (2) Legislative Quandary-Freedom: a. Personal: As long as a party acts legally (does not commit any fallacy) this party will never be put in a quandary (even if the other par- ty commits fallacies). b. Collective: If all parties act legally (avoid fallacies), no quandaries will arise at all. (3) Strategic Quandary-Freedom: For each party there is at least one legal strategy that will keep this party out of quandaries. (4) Minimal Quandary-Freedom: It is possible for all parties to act legally and in such a way that no quandaries will arise at all. It seems desirable that the rules of a system of dialectics would guarantee a high degree of quandary-freedom, certainly 2b, but preferably 2a or 1. For fallacies run counter to the ends of dialogue (ends the rules should help us to achieve), and quandaries leave no option but to commit a fallacy. If a commitment to inconsistencies would lead to inconsistent commitments (i.e., a quandary), there is every reason to try to banish them. Altogether, we now have reached three subquestions to the question about how to handle inconsistencies in dialogue: (l) Should we consider inconsistent statements by one and the same party to constitute a fallacy? (What rule of dialogue would be violated?) (2) Or should we rather consider them to constitute a weakness or blunder? (3) Do such statements lead to a quandary? (And would that be a reason to declare them fallacious?) As to (1) it may be noted that, tradi- tionally, inconsistency is seldom counted among the fallacies. Exceptions are Gautama in the Nyaya sutra,5 Mackie (1967:176,177) Johnson and Blair (1983:59-63), and Rescher (1987). But Rescher rejects the characterization of in- consistency as a fallacy. Mackie and Johnson and Blair do not discuss dialectical rules, and consequently cannot draw a dialectical distinction between 'blunder' and 'fallacy'. If their treatments were 'put in dialectical terms', inconsistency might, with these authors, come out as a blunder rather than a fallacy. 6 Thus tradition seems to put 'inconsistency' apart from the subject of fallacy. This is not surprising, for one of the aims of dialectic is to unearth hidden inconsistencies. And how could this be done if inconsistencies couldn't be expressed within the dialogue? So we cannot expect the rules to proscribe all inconsistencies. Suppose that, according to some system of dialectical rules, one's position is refuted as soon as it is developed to a point where both P and not-P are stated. 7 We can add a rule to ban such explicit inconsistencies: if P was stated first (and not retracted) one would not be allowed to state not -P and vice versa. If we now add a rule to the effect that one's position is refuted as soon as each legal move would bring about a situation of explicit inconsistency, refutation will oc- cur one step earlier in dialogue, but other- wise the new system will be equivalent to the old one. Thus we may continue for some time, banning some less explicit inconsisten- cies, such as P together with Q and not (P or Q), etc. But somewhere we must stop. We cannot stipulate that as soon as one's position becomes inconsistent it is refuted. For in order to apply such a rule it must be obvious, or easily checked, whether a position is inconsistent. And that is simply not the case. 8 Consider for instance, the well-known Barber Pseudoparadox: (B) In a certain village there lives a barber who shaves exactly those inhabitants of the village that do not shave themselves. Inconsistency and Commitment 37 Does the barber shave himself? He/she does if and only if he/she doesn't. (B) is self- contradictory (It is not a paradox, since the village in question does not exist.) The in- consistency in such riddles is not "ob- vious", otherwise they would be no fun. Perhaps one might say that in cases like (B) it is "easily checked" whether there is an inconsistency. So a viable dialectics could ban (B) and its cognates. If so, change the example to any desired degree of com- plexity. Frege's system in Grundgesetze der Arithmetik (1893,1903) was shown to be inconsistent by Russell (Russell's paradox). Certainly this was not "obvious", nor was Frege's system "easily checked" for con- sistency. For rich languages (predicate logic and beyond) consistency is an undecidable concept (i.e., no mechanical test can exist for it), and therefore totally insuitable for figuring in the formulation of dialectical rules. If there is to be a fallacy of inconsistency it must be limited to some restricted class of inconsistencies (e.g., fully explicit ones). One may, however, also opt for not hav- ing any fallacy of inconsistency at all. In that case positions containing both P and not-P are not ruled out in dialogue. Such positions would be merely refutable: they would lead to a loss of part of the argument, or to a forced retraction. 9 But even that is not necessary. The example of Minimal Dialectics (the dialectics of Minimal Logic) shows that one may hold P and not-P without refutation (and without being forced to assert and deny every proposition, pace Aristotle). 10 So even explicit inconsisten- cies need not be declared fallacious. If not a fallacy, would it be correct to describe inconsistency as a weakness or blunder? Certainly, sometimes it would be. Consider the following example: Sue: That greasy spot you left on it just ruins my book! Bill: How come! I didn't borrow your book at all! What's more, the spot was already there. You shouldn't shout at me: I returned it spotless. 38 E. C. W. Krabbe Bill's argument to the effect that he has not been gUilty ofleaving a greasy spot on Sue's book contains obviously inconsistent statements. We may wonder whether Bill is serious, or whether he merely wants to poke fun at Sue (an indirect speech act). Assuming that Bill is serious, it is certain- ly a weak argument, and by presenting it Bill gets into a vulnerable position. On the other hand the argument isn't entirely worth- less either, since any part of it could perhaps be substantiated, and would then suffice to establish Bill's claim of innocence. A stronger argument would leave the rest out, for sure. Even though it may often be correct to frown upon an inconsistency as constituting a weakness (or even a blunder) in arguing, there are other situations where it would not be correct to describe it that way. For one thing, inconsistencies can be hidden. We saw already how they can be hidden by complexity. This was a reason not to count them among fallacies. However, a complex inconsistency could still be said to constitute at least a weakness, assuming that a better (consistent) way of arguing is available. It is another matter if one's position is incon- sistent to start with. This should be called a weakness in one's position rather than in one's arguing. Thus being inconsistent was a weakness of Frege's system (position) in the Grundgesetze. Only if we view the whole system as an argument for logicism, may we say that Frege presented an un- necessary weak argument (assuming that better arguments are available). Since it is one of the ends of dialectic to unearth in- consistencies and other weaknesses in posi- tions, it is no weakness in dialogue if an initial position is inconsistent. Perhaps this is even clearer so if part of an initial posi- tion is hidden, not by complexity, but by virtue of its not being articulated in the form of uttered statements. There may be pro- positional commitments on the dark-side of an agenda of which the subject itself is not fully aware (the so-called dark-side of the commitment-store, cf. Walton, 1984: 247fO. Thus one may run into inconsisten- cies, without any deficiency in the dialec- tic process. The following example is Walton's: II George is a committed socialist but frustrated by the continual postal strikes, and he argues that the Post Office should be taken over by private enterprise. His friend Bob replies: "How can you say that, George? After all you are a socialist, and socialists do not think that state functions should be run by private enterprise." As the discussion gets on, it may become clear to both Bob and George that certain socialist tenets to which George is deeply committed do indeed conflict with George's uttered position. These dark-side com- mitments of George may come to the light in the course of, and as a consequence of, the discussion. Finally George may have to withdraw his uttered proposal, or modify his commitment to socialism. Yet all the time George could be arguing according to optimal strategies, without any fallacies, weaknesses, or blunders. We should now consider explicit incon- sistencies once more. Above it was main- tained that even a statement of an explicit inconsistency does not constitute a fallacy. Would it always constitute a blunder? We can't plead complexity, or dark-side com- mitment in this case. Yet the utterance of an explicit inconsistency could be entirely alright. First, it may be that no better moves are available (a party may be driven by the dialectic process to proceed from hidden in- consistencies to explicit ones). Second, such a position could be tenable in at least some systems of dialectic (e.g., Minimal Dialec- tics), as we remarked above. Third, an ex- plicit contradiction (P and not-P) could be used as a provocative thesis, i.e., a thesis not reflecting the position of its proponent, but the proponent's point of view that its adversary should, given its position, accept this thesis. 12 A provocative thesis P and not- P is equivalent to a claim that the position of the other party is inconsistent. It is essen- tial that such claims be admitted in dialogues. Sometimes a particular dialec- tic constant (absurdum. falsum) is used for this purpose. Thus Russell, confronting Frege as the author of the Grundgesetze, may be called a proponent of the absurd. A provocative thesis need not be absurd or self-contradictory. Even if it is not, it may contradict the position of its proponent. Take Haddock's conclusion as an example: The two Johnsons doubt whether Captain Haddock did accurately calculate the ship's position. They show the captain some im- proved calculations of their own. Haddock: "You are right... I was wrong,., Gentlemen, hats off!" [takes off cap, and stands in prayer for quite some time], Johnson: "Hey, captain, what's that for?" Haddock: "That's to say, gentlemen, that, according to your calculations, we are now in St. Peter's Cathedral, Rome!" (paraphrasing Herge, 1947) Haddock's provocative thesis confronting the two Johnsons in their capacity of calculators, determines the ship's position as St. Peter's Cathedral, Rome. This is what the captain should defend, if challenged, starting from the other party's concessions, in this case the improved calculations made by the Johnsons. Obviously, Haddock himself does not hold the ship to be in that position. Let us turn to quandaries. We want, on the one hand, the rules of a dialectical system to guarantee at least General Legislative Quandary-Freedom (with respect to commitments in the first sanction set). On the other hand, we want to admit explicit inconsistencies as legal positions in dialogue. The consequence is that the dialecticians should take care that situations in which a party is committed to both P and not-P do not lead to quandaries. If a party is to defend both P and not-P one would expect the rules to stipulate which defense takes preference. This can be done quite simply, e.g., by stipulating that one should defend first whatever statement is challenged first. In general: if more than one defense obliga- tion is allowed to be in force at the same time, these obligations should be ordered. 13 Inconsistency and Commitment 39 As long as a dialogue proceeds without fallacies, it seems impossible for quandaries with respect to the second sanction set to occur. As long as there are permitted moves, some of these must be optimal. If fallacies occur one has to appeal to higher order rules to regulate the situation. To avoid quandaries one could have a rule to the effect that a fallacy should be pointed out as soon as it is committed, and that the responsible party should in such a case withdraw its last move and substitute a legal one. Thus one would always have a legal move at his disposal (legal at the level of these higher order rules). By such means quandaries may be excluded from a dialogue: commitments to inconsistencies will not lead to inconsistent commitments. However, as we shall see, there is still a danger of quandary if a party is involved in several dialogues at the same time. Sometimes inconsistent statements are so mutually isolated that they cannot in- terfere one with the other. For instance in systems of formal dialectic (Barth and Krabbe, 1982) a concession P in one chain or aguments does not interfere with another concession not-P, if the latter should take place in a different chain of arguments. (Such chains of arguments may occur, for instance, in a discussion with not-P or (P and Q) as a concession.) In such a case there is no fallacy, no blunder, no quandary. Perhaps one should say there is no real in- consistency. (And perhaps 'being in the same chain of arguments' could figure as a dialectical interpretation of Aristotle's 'in the same relation' .14) Another case, where the isolation is less evident, is provided by the Preface Paradox (Following Makinson, 1964): An author has completed a learned work containing propositions S], through Sm He adds a preface containing the statement (P): "In view of human fallibility and the com- plexity of the present work, there is no doubt that I must have erred somewhere. Conse- quently, not all assertions that follow can be correct, however much I tried to avoid er- ror." Clearly P implies that S" through Sm 40 E. C. W. Krabbe cannot be all true (given that S!, through Sm are the assertions that follow I5 ). Let S be the conjunction of SI, through Sm. Then P implies that not-S. The book cum preface is, therefore, ineonsistent. What is the paradox? On the one hand the author displays a reasonable modesty. After all, there may be very good reasons to hold P. On the other hand the author's position is obviously inconsistent, and for that reason often deemed unreasonable. The Preface Paradox is usually discussed from the point of view of reasonable belief or acceptance. 16 From the point of view of dialogue all we can say is that the author's position is weak: He can be refuted. Yet, from this point of view, it is not unreasonable to be in a refutable position. I? However, it may also be possible to isolate P from S by claiming that here we have two different fields or areas of discussion (epistomology and whatever the book itself is about). P and S would then no longer in- terfere with each other, at least not without more ado. Notes * The author would like to express his gratitude to the Netherlands Institute for Advanced Study in the Humanities and Social Sciences (NJ.A.S.) for its support, and to Mrs. J.P. van Breda-Burguef(o for technical assistance. I This quotation has been adapted so as to avoid the use of logical symbolism. 2 For an example of the complicated kind of reasoning involved if several purposes or norms are to be implemented in one system see Barth and Krabbe (1982). Ch. III. See D.N. Walton and E.C.W.Krabbe, Com- mitment in Dialogue, to be published. In the next few paragraphs, I summarize the contents of the first chapter of that book. 4 Barth and Martens, (1977), Van Eemeren and Grootendorst (1987). 5 See Hamblin (1970: 180-181) and also "Clin- cher" [3]. p. 186. There seems to be no other mention in Hamblin's historical survey of 'con- tradictory' reason or inconsistency as a fallacy. A final example of presumed isolation is the following: A social worker in a regional hospital has to deal with patients from all walks of life. Two patients, both seriously ill, are totally upset by the prospect of death. It is urgent that the social worker should have a word of comfort for each of them. However, in view of their difference in outlook, this can- not be the same word. Discussing death with the first patient she stresses that one may look forward to eternal bliss hereafter. In- volved in a dialogue with the second patient she points out that "as long as we are present death is not, and once death is, we are not. "18 We may assume that both dialogues are free of fallacies and blunders. Since the two statements occur in distinct dialogues, our social worker feels sure they will not lead to a quandary. Unfortunately, the two pa- tients meet shortly after in the hospital's conversation room. They get into a serious discussion on death and afterlife. The social worker happens to walk in and both par- ties claim her support. What should she do? The 'contradictory' reason fallacy in the Nyaya sutra is presented in a text on controversies (a dialectical setting we may say), so it deserves more attention than I can give it here. From Hamblin's report I'm unable to decide whether 'contradictory' reason could not as easily be classified as a weakness or blunder. 6 Johnson and Blair are much concerned with substantiating charges of fallacy. But this seems characteristic more of assaults on weaknesses or blunders than of charges of fallacy (in our sense). In the case of a fallacy (in our sense) it would ordinarily suffice to point out the par- ticular dialectical rule one's opponent has violated. Mackie explicitly admits that it would be a "fallacy to suppose that because your op- ponent has tried to have it both ways, he can- not have either way-that every part of an in- consistent position must be false." (1967: 177). So part of an inconsistent argument may sur- vive in further debate. This again is characteristic of weak arguments. 7 In Barth and Krabbe (1982) refutation is pushed even further back: to a situation of being in con- trary dialogue attitudes, followed by a win- ning remark of the other party. 8 A similar point was made by J. Woods in a lec- ture (1987). Cf. the Woods-Mackenzie discus- sion in Argumentation, Woods (l988a, 1988b), Mackenzie (1988). 9 This is the case in most systems of formal dialectics, whenever the Opponent holds such a position. Cf. Note 7. 10 For Minimal Dialectics, see Barth and Krabbe (1982). It See Note 3. 12 Cf. Krabbe (1982). 13 Technically, other solutions are possible. Cf. Classical Dialectics in Barth and Krabbe (1982). 14 Cf. the quotation from Aristotle at the begin- ning of this paper. Inconsistency and Commitment 41 15 It is argued by Keith Lehrer (1975) that P, S I, ... , Sm could be all true, because the book might have constained other statements than it actually does. (Lehrer doesn't frame the paradox in terms of a book, but that does not matter.) Lehrer is right if we assume that the phrase' 'all assertions that follow" in Prefers descriptively to the assertions that follow (dif- ferent ones in different possible worlds). Here, I have assumed this phrase to be no more than a shorthand for the list of propositions and hence to refer rigorously to them (in every possible world to the actual ones). 16 Makinson (1964), Lehrer (1975), Rescher and Brandom (1980). 17 Cf. Barth and Krabbe (1982: 71): "It is not ir- rational to lose a discussion ... But it is ... irra- tional not to admit that one has lost. .. " 18 Epicure. Letter to Menoeceus, ap. Diog. X 125. References Aristotle (1975) The Metaphysics. Cambridge, Mass. (Harvard D.P.) and London (William Heinemann), The Loeb Classical Library. Barth, E.M. and E.C.W. Krabbe (1982) From Axiom to Dialogue. A Philosophical Study of Logics and Argumentation. Berlin and New York (de Gruyter). Eemeren, F.H. and R. Grootendorst (1987) Fallacies in Pragma-Dialectical Perspective, Argumentation I, 283-301. Eemeren, F.H. van, R. Grootendorst and T. Kruiger (1986) Argumentieleer 2. Drogredenen. Groningen (Wolters-Noordhoft). Frege, G. (1893, 1903) Grundgesetze der Ari1hmetik, begriffsschriftlich abgeleitet. Jena. Volume I 1893, Volume II 1903. Grice, H.P. (1975) Logic and conversation. In: P. Cole and J.L. Morgan (eds.), Syntax and Semantics III: Speech Acts. New York, etc. (Academic Press), pp. 41-58. Hamblin, C.L. (1970) Fallacies, London (Methuen). (1987) Imperatives. Oxford and New York (Blackwell) . Herge (1947) De avonturen van Kuifie: De schat van Scharlaken Rackham. Tournai and Dronten (Casterman). [Les A ventures de Tintin: Le Tresor de Rackham le Rouge]. Hodges, W. (1977) Logic, Harmondsworth, etc. (Penguin: Pelican books). Johnson, R.H. and J.A., Blair (1983) Logical Self- Defense, 2nd ed., Toronto, etc. (McGraw-Hili Ryerson). [1977J. Krabbe, E.C.W. (1982) Formal Dialectics as Immanent Criticism of Philosophical Systems. In: E.M. Barth and 1.L. Martens (eds.), Argumentation: Approaches to Theory Formation: Containing the Contributions to the Groningen Conference on the Theory of Argumentation, October 1978, Amsterdam (John Benjamins), pp 233-243. Lehrer, K. (1975) Reason and Consistency. In K. Lehrer (ed), Analysis and Metaphysics. Essays in Honor of R.M. Chisholm, Dordrecht (Reidel) pp 57-74. Mackie, J.L. (1967) Fallacies. In: P. Edwards (ed.), The Encyclopedia of Philosophy m. New York and London (MacMillan), pp 169-179. Mackenzie, J. (1988) Woods on Ideals of Rationality in Dialogue, Argumentation 2,409-417. Makinson, D.C. (1964) The Paradox of the Preface, Analysis 25, 205-207. Rescher, N. and R. Brandom (1980) The Logic of Inconsistency. A Study in Non-Standard Possible-World Semantics and Ontology. Ox- ford (Blackwell). 42 E. C. W. Krabbe Walton, D.N. (1984) Logical Dialogue-Games and Fallacies. Lanham (Univ. Press of America). (1988b) Rationality Ideals and Mentality, Argumentation 2, 419-424. Woods, J. (1987) What Rationality Is Not. Lecture at the University of Amsterdam, February 22nd, 1987. (1988a) Ideals of Rationality in Dialogic, Argumentation 2, 395-408. KING'S COLLEGE ERIK C. W. KRABBE FILOSOFISCH INSTlTUUT RIJKSUNIVERSITElT. GRONINGEN WESTERSINGEL 19 9718 CA GRONINGEN THE NETHERLANDS THINKING SKILLS NEWSLETTER Thinking Skills Newsletter is devoted to providing practical information on methods of teaching critical thinking, decision making, problem solving and creative thinking, Our audience is composedofteachers and adminstrators from all levels of education and in all subject areas. * practical tips for teaching thinking skills * reviews of books, journal articles and conferences * information and advertisement of workshops and conferences We are partially supported by Pennsylvania Department of Education and Pennsylvania Association for Supervision and Curriculum to provide free subscriptions. However, we cannot afford to send subscriptions to nonreaders. To receive either a single review copy or an annual sub- scription, write: Max M. Thomas, Editor Thinking Skills Newsletter King's College Wilkes-Barre, PA 18711 o