Critical Review/Compte rendu critique 281 Critical Review Manifest Rationality A Pragmatic Theory of Argument Ralph H. Johnson Mahwah, N.J.: Lawrence Erlbaum Associates, 2000. Pp. xiv, 1-391 including References and Index. ISBN 0-8058-2173-2 (cloth: alk. paper), ISBN 0-8058-2174-0 (pbk: alk.paper). Reviewed by Trudy Govier In his energetic new book, Ralph Johnson takes on four fundamental tasks. First, having reviewed a number of central classical and contemporary ac- counts, Johnson argues that a philosophical theory of argument is necessary, and explains what the nature and purposes of such a theory would be and how it would be developed within informal logic. Second, he explains and argues for his own normative theory of argument, based on his belief that the pur- pose of argument is rational persuasion. Third, Johnson considers objections to informal logic as such, and to his own particular theory within it. Fourth, he sets forth a research agenda for informal logic, linking that agenda with rel- evant other disciplines and other areas of philosophy.Along the way, Johnson uses and comments on the work of many other theorists, including Stephen Toulmin, John McPeck, Mark Weinstein, Tony Blair, Leo Groarke, Frans van Eemeren and Rob Grootendorst, c.L. Hamblin, David Hitchcock, Douglas Walton, John Woods, and myself. Johnson describes his orientation as fol- lows: To say that the practice of argumentation is characterized by manifest rationality is to say that it is patently and openly rational. To whom? To the participants, whether they be arguer, critic, or those interested in the issue. They agree to do nothing that would compromise either the substance or the appearance of rationality. (MR, 63) Anyone interested in theories about the interpretation, criticism, and evalu- ation of arguments will benefit enormously from studying this book. Here I shall concentrate primarily on Johnson's development of his own theory. For the most part, I shall not address his criticisms of other authors and his suggested research agenda. But I make a partial exception for some of my own ideas-in an attempt to fulfill some of what Johnson would call my dialectical obligations. © Informal Logic Vol. 20, No.3 (2000): pp.281-291. 282 Govier review of Johnson, Manifest Rationality Johnson's Theory: Argnment as Manifest Rationality Johnson's view is that we offer arguments when we try to persuade an Other, on rational grounds, that a claim is true. The classic account of argument is a minimalist one: some claims, premises, are put forward to support a further claim, a conclusion. An argument is defined as a sequence of premise state- ments leading to a conclusion that those premises are supposed to support. What such accounts omit, according to Johnson, is the pragmatic dimension: why is it that the conclusion needs support? His answer is that the conclusion is in some respect controversial, and the arguer is trying to rationally persuade the Other of its truth. Implicit even in the minimalist account is a crucial teleological dimension, because an arguer puts forward premises in an effort to support a conclusion. If the conclusion were not in some respect contro- versial, it would not need support. In offering an argument for the conclusion, the arguer is implicitly addressing an Other who does not accept that conclu- sion as true; beginning with the premises, the arguer seeks to rationally per- suade the Other of -its truth. The case in which the arguer himself or herself also plays the role of Other is acknowledged; one may seek to persuade one- self of a claim that is in doubt, addressing arguments to oneself. But even in this limiting case, the role of Arguer and Other are both required for an argu- ment. The means of persuasion used in argument are rational: the Arguer does not seek merely to get the Other to change his or her mind in any way what- ever. Drugs, electro-therapy, brainwashing, threats, or hypnosis are not argu- mentative means. Rather, the Arguer appeals to the Other's beliefs and rea- soning. 1 The telos of argument is rational persuasion, and accordingly, the process of argument must both be and seem rational-hence, manifest ration- ality. The Arguer must make his or her argument responsive to the beliefs and queries of the Other. That means that the process of argumentation is in cru- cial respects dialectical. The arguments that are products ofthat process should reflect the dialectical contexts in which they are located. Because the conclusion is controversial, there will be alternative positions to it, and there will also be objections to it and to the argument or arguments in its favour. The dialectical dimension leads Johnson to advocate a two-tired account of argument. On the classical premise/conclusion (here PC) account, an argument is comprised of premises and a conclusion. Despite the occa- sional gesture toward radical scepticism about conclusion and premises, Johnson retains the classical PC account, referring to that structure as the illative core of an argument. To rationally persuade the Other of a claim, an arguer will need to do more than put forward some statements as premises for it. He or she will need to respond to objections and alternative positions. Be- cause it does not address objections and alternative positions, the illative core by itself cannot constitute a complete argument. Accordingly, Johnson argues for a "second tier" of argument, which he calls the dialectical tier. Critical Review/Compte rendu critique 283 An argument is a type of discourse or text-the distillate of the practice of argumentation-in which the arguer seeks to persuade the Other(s) of the truth of a thesis by producing the reasons that support it. In addition to this illative core, an argument possesses a dialectical tier in which the arguer dis- charges his dialectical obligations.(MR, 168) An Arguer has dialectical obligations to respond to objections and address alternative positions. Johnson understands informal logic as oriented toward developing a theory of argument that will provide an account of the nature of argument, and derive from it a coherent and usable set of norms for their appraisal. He argues for a distinction within the theory of argument, between the tasks of analysis and those of appraisal, believing that some have tended erroneously to conflate these tasks. A coherent account of the norms of argument appraisal must be derived from the fundamental conception of argument as designed for rational persuasion. Johnson defends standards that any theory of argument would have to meet. (I) It must have an account of argument and the elements that constitute it. (2) It must contain an account of the structure of arguments and display it in a clear and precise manner. (3) It must contain a theory of ap- praisal that recognizes that there can be good arguments for a given position as well as good arguments against it. (4) The theory of appraisal must allow for strong arguments, weak arguments, and intermediate degrees of merit. (5) Its criteria of appraisal must be "user-friendly" in the sense that an ordinary reasoner could decide whether they are satisfied. (6) The theory of appraisal must allow for fruitful criticism. (7) The criteria of appraisal must be justified in terms of the theory as to what argument is and how argument analysis is to proceed. These standards strike me as having considerable independent interest and plausibility (see MR, 52-56 and 78-81). Johnson appeals to them to argue effectively that formal deductivism (FDL )--according to which a good argu- ment is a sound one with true premises from which its conclusion may be inferred according to rules offormal deductive validity--does not provide an adequate theory of argument. Johnson argues that a good argument must satisfy criteria of evaluation both for its illative core and for its dialectical tier, and that the appraisal criteria he defends for the theory as a whole will meet his own seven standards. A. Illative Core. After a detailed discussion considering his own previous work with J.A. Blair and arguments by Hamblin, Pinto, Allen, and myself, Johnson reaches the view that premises must be acceptable, relevant to the conclusion, true and sufficient to support the conclusion. (ARTS) B. Dialec- tical Tier. The arguer must properly take account of the "dialectical realities" by dealing with objections and alternative positions. In practice, this will mean satisfactorily dealing with Standard Objections to the premises, conclusion or argument itself, as well as showing that the conclusion is preferable to (stand- ard?) alternative positions. 284 Govier review of Johnson, Manifest Rationality The adequacy ofthe illative core requires satisfying ARTS, and this makes no reference to such dialectical dimensions as answering objections. For the illative core, Johnson is able to develop his points against a wealth of detailed material in informal logic, whereas when it comes to the dialectical tier, he finds himself in the role of a pioneer. Readers should not be surprised, then, to find that standards for the dialectical tier seem under-developed or even rather loose. Although Johnson addresses problems of how to select which objec- tions must be addressed, he makes little attempt to elaborate on what he finally appeals to, the notion of Standard Objections. He does not explore the issues of how alternative positions are to be specified and distinguished, or which ones an arguer might be dialectically obliged to consider. Nor does he fully acknowledge the regress problems that would seem to arise because argu- ments on the dialectical tier will themselves give rise to objections to which the response will be further arguments which themselves will be good only if objections to them are addressed. In addition, there will be positions alterna- tive to those taken in the arguments used to respond to objections and to address alternative positions. In the end, the apparently reasonable require- ment of a dialectical tier leads to a wealth of interesting problems-hopefully, but not obviously, not intractable ones.2 Some Matters of Detail In this ambitious work, there are many matters of detail to inspire interest. I select just three for special comment. These are: the distinction between evalu- ation and criticism; the argument that there can be no such thing as a conclu- sive argument; and the inclusion of truth along with acceptability in the criteria for the illative core. The distinction between evaluation and criticism is developed in Chapter Eight. We evaluate, say, a movie, if we pronounce it good or bad-and when we do so, we presumably have some standards in mind. But to evaluate a movie is not yet to criticize it. To criticize it, we would have to articulate our standards, show evidence as to why the movie did or did not meet them, and put our comments into some kind of coherent perspective. To evaluate some- thing is to pronounce it good, bad, or indifferent--or somewhere along the spectrum. To criticize it is to develop an account of its strengths and weak- nesses, an account that shows some discrimination between more and less significant strengths or weaknesses and can give assistance as to how the product might be improved. By criticism on the other hand, I understand the articulated and reasoned evaluation of something communicated to the creator with the view that it will improve the product. Criticism goes beyond evaluation in that it must take into account the strengths as well as the weaknesses of the product and is in- tended for the one who produced the argument as a vehicle whereby the Critical Review/Compte rendu critique 285 argument may be improved. Thus it may be said that criticism is part of a dialectical process, whether evaluation is or not (MR, 219). Informal logicians have tended to speak of the evaluation of argument, which is an important task in many contexts. But criticism goes further, and is fundamentally important; Johnson argues plausibly that informal logic should also develop a theory of criticism. As to the possibility of conclusive arguments, Johnson argues that there is no such thing. For an argument to be conclusive, it would have to satisfy three internal conditions and one external one. Internally, its premises would have to be unimpeachable or uncriticizable; the connection between those premises and the conclusion would have to be unimpeachable, "the strongest possible;" and the argument would have to successfully and rationally resist every attempt at legitimate criticism. Externally, it would have to be regarded as a conclusive argument. Johnson argues that these demands are so strong that realistically no argument can satisfy all of them. .. (I)f there were such an argument, when people were taught about arguments, this argument would be given them (as) an exemplar or model. It would be celebrated in texts. Everyone would know it by name. But so far as I know there is no such argument. Because no argument has satisfied all these conditions (the internal and external properties for being a conclusive argument), I conclude that there are no conclusive arguments. (MR, 237) Johnson believes that proofs are conclusive, but he distinguishes between proof (one important case being mathematical proof) and argument. I ques- tion whether this qualification is necessary. Proofs start somewhere, and their starting point can always be questioned from some point of view. And proofs proceed on the basis of some reasoning; thus, similarly, the link be- tween what is cited to provide proof and the point that it is supposed to prove can always be questioned from some perspective or other. (Even in math- ematics, someone can raise issues. Is this formal system the right one to apply to this area? Do we have good grounds for accepting its axioms and rules?) Now to the matter of the acceptability and/or truth of the premises. In Logical Self-Defence, Johnson and Blair proposed acceptability, relevance, and sufficiency as criteria for a good argument, dealing with the illative core level only. I have used essentially these criteria in various editions of my text, A Practical Study of Argument, changing the terminology to acceptability, relevance, and good grounds, and proposing the acronym ARG. Many others in informal logic have also used them. The main reasons for using acceptabil- ity and not requiring that the premises of a good argument be true were two- fold. First, a truth standard seemed to impose a high standard of knowledge and evidence for premises and perhaps even a "God's eye view." Secondly, the (rational) acceptability ofthose premises to the Other was believed to be more pertinent to the task of rationally persuading the Other than their truth. 286 Govier review of Johnson, Manifest Rationality True premises not accepted by the Other, and not acceptable by him or her, could not serve the purpose of rational persuasion; hence truth was not suffi- cient to make premises persuasive to another. And it did not seem clear that truth was necessary either. Johnson now argues that his earlier view needs to be revised. Although he stilI acknowledges that true premises not known by the Other to be true can- not serve the purpose of rational persuasion of the Other, he argues that in addition to being rationally acceptable the premises of a good argument must be true. (Hence ARTS, not ARG.) An argument with one or more false premises is not a good argument. Johnson contends that truth need not be a God's eye notion. He also claims that philosophers-including, formerly, Blair and him- self-have continued to appeal to truth (and falsity) in expounding central concepts such as relevance, consistency, and sufficiency, and in explaining various fallacies. Johnson contends that such appeals indicate that the truth criterion was being tacitly relied on and should not have been eliminated. This argument strikes me as somewhat questionable. First, there is no inconsist- ency in omitting truth as a condition for argument evaluation while continuing to use it for some other expository purposes such as defining relevance, con- sistency, or sufficiency. One can claim that truth is not a criterion for argu- ment merit without being committed to the view that the notion of truth should be altogether relinquished. Second, as to appeals that might indicate a truth criterion is tacitly presumed yet explicitly denied, there would be a problem here if one were to appeal to the truth or falsity of certain premises, or types of premises, in (say) offering an account of a fallacy. At that point, one would be presuming a conception of a good argument as containing true premises, inconsistently with an acceptability based account. If such appeals were made by those officially committed to acceptability instead of truth, they were in- deed lapses, and we must ask whether they amounted to mere slips, correct- able by shifting back to acceptability or unacceptability, or were, rather, una- voidable appeals to notions of truth or falsity that did not properly fit. Only if the latter is the case will truth be creeping inappropriately through the back door, as Johnson alleges. Johnson acknowledges that ARTS allows for possibilities of conflict be- tween the truth condition and the acceptability condition. First, a premise might be true (T) and yet not acceptable to the Other; second, it might be acceptable to the Other (A) and yet not true. Johnson discusses such conflicts between A and T interestingly, and resolves them in a plausible and resolutely non-relativistic way. In the first case, the Arguer should provide a sub-argu- ment in an effort to re'nder the true premise acceptable to the Other. In the second case, the Arguer should not employ the premise, since an argument with a false premise will not be a good argument, even if it does turn out to persuade the Other of the conclusion. This discussion strikes me as perfectly sensible. It is readily extended to take into account the perspective of a third Critical Review/Compte rendu critique 287 party, the Evaluator, if he or she is distinct from the Other. But it should be noted that perfectly parallel issues arise if we remain within the domain of rational acceptability and consider acceptability from the different points of the Arguer, the Other, and the Evaluator. If an Arguer believes that a premise, PI, is rationally acceptable, but sees that the Other does not, he cannot use that premise in an attempt to rationally persuade the Other unless he offers a sub-argument to render it rationally acceptable to the other. If an Arguer be- lieves that PI is not rationally acceptable, but nevertheless believes that the Other will deem it to be so, he should hold back from using PI, because if he does so he will be putting forward what he himself believes to be a bad argu- ment. If an Evaluator finds that PI is not rationally acceptable, even though both the Arguer and the Other deemed it to be so, then the Evaluator will judge the argument not to be good, because it has an unacceptable premise. And so on. Johnson's arguments on the need for a truth criterion as well as accept- ability do not seem to be definitive. Slightly Polemical Additions: Johnson on Govier Johnson makes extensive use of my earlier work, particularly my 1987 mono- graph, Problems in Argument Analysis and Evaluation (P AAE). I am grati- fied to receive this attention, though I cannot help but feel that P AAE is a product of my younger days. A dialectically pertinent fact is that it felt like a pioneering and anti-establishment effort when I wrote it in 1983-86. At that time there was virtually no sophisticated philosophical discussion about the non-formal normative theory of argument. Informal logic was widely regarded as an intellectual slum in which fallacies and sloppy theorists together led a symbiotic but careless and bleak existence. In some areas, such as the discussion of Deductivism in P AAE, the pio- neering context shows. I readily admit to hastiness and over-statement in some of my early arguments against Deductivism. But I would still argue that even Reconstructive Deductivism will not provide a satisfactory and empiri- cally plausible theory of argument. I welcome Johnson's contributions to the dispute between Groarke and myself about the tenability of Reconstructive Deductivism, a version of which Groarke defended. I would not be happy to appeal to such a notion as degrees of truth in order to render Deductivism compatible with the intuition that there should be a spectrum of goodness for arguments, ranging from strong to weak. J I remain puzzled by the Problem of Missing Premises. But that is not to say that I reject it as a false problem, or a pseudo-problem. I believed, and still believe, that deductivists long abused appeals to missing premises, tending to add them ad hoc to make any real argument that seemed to be a counter-example fit their theory. It was this sort of adhockery that I found objectionable and objected to in P AAE. Contrary to what Johnson says, I do not think that the Problem of Missing Premises is necessarily tied to Deductivism, and I never wished to suggest that it was so 288 Govier review of Johnson, Manifest Rationality tied. I believe that some real arguments actually are deductive and-as a dis- tinct, but related, phenomenon-some real arguments have missing premises.4 While pleased that Johnson uses and develops some of my early work, I do object to the way he describes some key themes. In P AAE, I termed the view that all arguments are either inductive or deductive 'Positivism.' Johnson changes the name for this view to Inductivism. This shift strikes me as mis- leading and unsatisfactory, because it suggests that the view represents all arguments as inductive. According to Positivism (which is still appealed to in many philosophical circles) some arguments are deductive and all the others are, in some broad sense, inductive. StilI more misleading is Johnson's labeling of my own Pluralist view as Conductivism. In P AAE I defended the claim that there are inductive, deductive, conductive, and analogical arguments. accord- ing to this Pluralist theory, there are at least these four types of argument, and no one type is properly reducible to any other. There may be further types: for instance, inference-to-the-best explanation (sometimes caIled abductive argu- ments) or "narrative argument" could be argued to be distincttypes. IfC.AJ. Coady is right in arguing that appeals to testimony are too basic to be induc- tively grounded, one might even wish to argue that these constitute a distinct type. Johnson's term 'Conductivism' suggests that according to my Pluralist theory all arguments are conductive--excluding inductive, deductive, and any other arguments. StilI more seriously-and rather mysteriously-Johnson's whole chronicle at this point omits to consider case-by-case reasoning, which was something I emphasized and found ~ascinating .5 What I called a priori analogies have an excellent claim to be. deemed a distinct type from deductive, inductive, and conductive arguments. The case-by-case theme was devel- oped in the fifties by John Wisdom in his unpublished "Virginia Lectures." Others, including for instance Stephen Barker and Jerry Bickenbach, have found a priori analogies or case-by-case arguments theoretically important, especially for what they suggest about the relation between particular judg- ments and generalizations. I still regard these issues of typology as highly central topics within the theory of argument, and as topics with important implications elsewhere, due to their implications for our understanding of justification and justifiability. Johnson says that typology is a theme for argument analysis, one which I and some others connected too closely to argument appraisal. He is right in judg- ing that I failed to emphasize the distinction between analysis and appraisal. But he is wrong ifhe means to imply that argument typology is not pertinent to argument appraisal. To determine whether and how premises are relevant and sufficient to support a conclusion (key stages for his evaluation of the illative core and thereby also for the evaluation and criticism of an argument), we need to know what sort of argument it is: deductive, inductive, conductive, analogical, or other. My early emphasis on what I called "the identification of arguments" arose from teaching experience, and especially from my surprise when I discovered Critical Review/Compte rendu critique 289 how difficult it was for students to pick out conclusions and premises from texts. In P AAE, I used such expressions as "naturally occurring" or "every- day" arguments. As Johnson points out, these expressions require clarifica- tion. How, if at all, do 'everyday' arguments in conversations or newspapers relate to arguments in the specialized disciplines? The importance of this issue emerges in discussing the work of Stephen Toulmin, John McPeck, and Mark Weinstein. I would now say that the purpose of a theory of argument is to account for the nature and norms of arguments wherever they might be found, provided that those arguments are either real or realistic. A real argument is one that some Arguer actually employed to address an Other, whether in eve- ryday life or in a more specialized domain (history, biology, physics, geology, anthropology ... ) or discussion. A realistic argument, though contrived as an illustration for some purpose of theory or pedagogy, is relevantly similar to a real argument; one can construct a plausible context in which it might plausi- bly be used by an Arguer to address an Other. I believe that a good theory of argument should offer coherent and justifiable guidelines for the interpreta- tion, evaluation, and criticism of real and realistic arguments. It need not ac- count for invented "arguments" that are unrealistic and have been contrived merely to fit a theory or provide counter-examples. Like others who became interested in informal logic during the seventies and early eighties, I had been shocked to discover that in writing textbooks and elsewhere, philosophers continued to cite either vague Positivism, Deductivist norms of classical soundness, or inconsistent combinations of the two. It seemed to me to be readily demonstrable that such norms rather rarely applied to real or realistic arguments. P AAE was the theoretical part of my response to this shock. What I wanted to argue-and still would argue-was that there was a need for a theory of argument that would apply to real or realistic arguments, and we did not have such a theory. Despite disagreeing with some details of my arguments and re-describing some of the develop- ments, Johnson is echoing this call. In Manifest Rationality, he proposes a theory in response. Johnson suggests at several points that there is no pre-theoretical way that arguments simply "appear," questioning my (naiVe?) assertion that Deductivism makes a distorting lens through which some people have tried to look at all arguments. He claims that we have no non-theoretic conception of what non- distortion would be, and thus no position from which we could accuse a priori Deductivism of distorting a pre-theoretic reality. These claims point to deep themes in meta-theory-and to some extent, I accept them. I acknowledge Johnson's point that theory affects perception, interpretation, description- and thereby our whole sense of what is "out there." However, in studying arguments, as in studying other phenomena, I would urge that there are limits to what theories can tolerably do with data. In the theory of argument, real and realistic arguments in all domains are our data. To construct a good theory, 290 Govier review of Johnson, Manifest Rationality we have somehow to react responsibly to that data-as I am sure Johnson would agree. I believed, and continue to believe, that, among other things, that means not taking reconstruction too far. Simplistic comments may no doubt be found in P AAE, in the light of the extensive literature about argument from the mid-eighties onward. But I remain convinced that there a pre-theoretic sense in which a detailed inductive analogy seems and is significantly different in structure and force from an argument of the constructive dilemma form- and that real arguments may be found, of both these types. Were there no apparent structural differences between real arguments, logicians and theo- rists of argument would have no data to have theories about. For anyone who is interested in real and realistic arguments, and current theorizing about them, Manifest Rationality will be an important book to study. Notes lOne might suggest that the Other's emotions are also appealed to; however this matter is not discussed in Johnson's work. I note it as a topical related theme. 2 Serious complications arise here, as I argued in "Becoming Dialectical: Two Tiers of Argument Appraisal?" and "Progress and Regress on the Dialectical Tier," both in The Philosophy of Argument (Newport News, V A: Vale Press, 1999), J Urged at one point by Leo Groarke in his defence of deductivism and apparently supported by Johnson, though in a different context. See MR, 198. 4 This stance is clearly indicated in the discussion in Chapter Two of A Practical Study of Argument, (Belmont, Ca: Wadsworth, 4'h edition, 1997). l Argued in "Two Unreceived Views of Reasoning and Argument," in Problems in Argument Analysis and Evaluation (Dordrecht: Foris 1987; now Berlin: De Gruyter); in "Analogies and Missing Premises" (Jriformal Logic 11 [1989J, 141- 152), later printed as "Euclid's Disease and Desperate Violinists: Do Analogies Have Missing Premises?" in The Philosophy of Argument; and also developed in all five editions of my text A Practical Study of Argument (Belmont, Ca: Wadsworth, 1985, 1988, 1992, 1997,2001). Trudy Govier 3207 Canmore Road, N W. Calgary, Alberta Canada T2M 4J8 govier@home,com