CRITICAL STUDY The Place of Emotion in Argument DOUGLAS WALTON University Park, PA: Pennsylvania State University Press, 1992. Pp. xiv, 1-294. ISBN 0-271-00833-4. Cloth US $45.00. Paper US $14.95. Reviewed by John Deigh Walton, in this book, applies an unconventional method of analyzing fallacies to the study of appeals to emotion. Following good sense and in opposition to views one can find in some textbooks in introductory logic, he denies that such appeals are always fallacious. Accordingly, he examines both how an appeal to emotion can serve to advance an argument and what makes it a fallacy when it is fallacious. His method of analysis draws on ideas from the field of pragmatics in linguistics. He sees arguments as contributions to dialogues and takes the cogency of an argument to depend on the type of dialogue to which it contributes. Since different types of dialogue have different purposes and different rules of engagement, this means that he takes an argument's cogency to be relative to the purpose and rules of the dialogue to which it contributes. Thus, on Walton's method, an argument may be fallacious as a contribution to one type of dialogue and reasonable as a contribution to another. The method differs strikingly from the standard textbook treatment in which models of sound reasoning taken from formal logic are applied to arguments without regard to their practical contexts. And to capture this difference Walton describes his method as "pragma- dialectical" (p. 16) and the standard treatment as "deductive-semantic" (p. 19). The book consists of eight chapters. The middle four, which make up its core, cover four types of appeal to emotion that, in the standard treatment, are typically classified as fallacies of relevance. These, under their familiar Latin rubrics, are argumenta ad populum, ad misericordiam, ad baculum, and ad hominem. In the first two chapters Walton explains his method of analysis and discusses how it applies to two other appeals that are typically classified as fallacies of relevance, appeal to authority (argumentum ad verecundiam) and appeal to ignorance (argumentum ad ignorantiam). The last two chapters contain effluent analyses and commentary. Chapter seven concerns complications that arise when more than one of these appeals occur in a single argument. Chapter eight offers some general, concluding remarks about when in an argument it is reasonable and proper to appeal to pity, say, or popular sentiment and when such an appeal is sophistical and manipulative. In this review, I will concentrate on whether Walton's unconventional analysis of emotional appeals is necessary to uphold his general thesis that not every appeal to emotion is fallacious and to what extent his analysis sheds light on the character of the fallacies to which these appeals are liable. Though his particular analyses in the middle four Informal Logic Vol. 17, No.1 (Winter 1995): 113-121 © Informal Logic 1995 114 John Deigh chapters merit separate consideration, I will, for lack of space, have to restrict my discussion of them to comments that bear on these two questions. Walton's general thesis is unassailable. It should be obvious to anyone, upon reflection, that context often makes the difference between cogency and subterfuge when it comes to assessing an argument that is based on appeal to emotion. An argument calling for individual or collective action in the face of truly wretched or alarming circumstances could not be faulted for premisses designed to convey pity or fear, if it were otherwise sound. Indeed, to argue for the same conclusion using premisses that described these circumstances affectlessly would be as logically inappropriate as it would be self-defeating, for proper description of the circumstances, given the practical exigencies they involve, requires terms that elicit the appropriate emotional response. Failure to use such terms would actually be misleading. It would convey false ideas about the nature of the circumstances. By the same token, it would be misleading to use these terms in circumstances in which the misfortunes or dangers the premisses described were minor. To call an endurable loss "a tragedy" or a manageable risk of disruption "a potential disaster" is to use terms for circumstances in reaction to which powerful emotions would be warranted when the circumstances in fact do not warrant such emotions. Either would be an example of appeal to inflated emotion-sympathy-mongering, in the one case, alarmism in the other, and no argument based on such an appeal could be sound. To be sure, neither appeal would exemplify a fallacy of relevance, since the argument's fault would lie in the falsity of its premisses rather than their irrelevance 10 the conclusion. And if 'fallacy,' as a term of criticism, is reserved for errors of reasoning as distinct from false assumptions, then neither would exemplify the fallacies that the terms argumentum ad misericordiam and argumentum ad baculum are commonly used to denote. These form a distinct subclass of faulty appeals to emotion. They work by eliciting misplaced rather than inflated emotion. The difference can be illustrated by the different ways a defense lawyer might argue, to a judge or jury, against sending his client to prison. Consider first an argument based on the client's poor health. The lawyer in making this argument might, by exaggerating the seriousness of his client's condition, seek to elicit more sympathy from the judge or jury than the condition warrants. Though the appeal in this case would be to inflated· emotion, it would nonetheless rest on a relevant consideration since incarceration works greater hardships on the sick than on the healthy. It would not, then, exemplify the fallacy of argumentum ad misericordiam. By contrast, an argument against sending the client to prison that was based on the hard life of his ancestors would exemplify this fallacy, for it would proceed from considerations that were irrelevant to its conclusion, though of course their irrelevance would be masked if they succeeded in arousing misplaced sympathy. The commonplace view of emotional appeals that Walton rejects, the view that they are necessarily fallacious, is thus doubly mistaken. Not only are there cogent emotional appealS, but there are also faulty emotional appeals that are not fallacious. Review of Walton 115 Presumably, this commonplace view comes from a misconstruction of the age-old opposition between reason and passion that permeates our culture. Whatever truth there may be to this opposition, it does not vitiate the kind of understanding of ourselves and the world around us that our susceptibility to emotions gives us. Our emotional susceptibilities, like our sensory ones, supply materials from which we can draw inferences, which is to say, they supply materials on which reason can operate. In this way, reason and emotion work, not as opposing faculties, but as cooperative ones whose joint product is belief and practical judgment. Of course, an emotional response can bring a distorted understanding of one's circumstances, and it can encourage one to reason badly from the understanding of one's circumstances it brings. And this liability to distortion and bad reasoning no doubt explains the association of appeals to emotion with sophistry. The liability thus justifies the logician's special concern with appeals to emotion as types of argument that can easily go awry or be used sophistically. It does not, however, justify wholesale dismissal of them as always and necessarily fallacious. Statistics too are liable to be misunderstood and to produce erroneous inferences, but this hardly justifies wholesale dismissal of statistical arguments as always and necessarily fallacious. Walton attributes the commonplace view he rejects to the modern logician's excessive deference to science (pp. 31-33). Science and its conception of knowledge, he maintains, so thoroughly dominate modern logic's view of rigorous argument and sound judgment that modern logicians commonly discount other forms of argument and judgment, particularly those that contain or result from an emotional appeal, as grossly inferior if not worthless (pp. 67-68). This view of rigorous argument and sound judgment, Walton argues, is mistaken, for there are many contexts and occasions outside those of scientific inquiry which are equally important and exemplary sites of good reasoning and in which emotional appeals are at home (ibid.). There is, I believe, much to be said for his position. Science, in our culture, is the institutionalization of reason and its ~ractice is often idealized as immune to emotion when its pursuit is pu~e. This Idealization corresponds to the idea of emotion as invariably an opponent and therefore corrupter of reason, an idea that, as I argued above, is seriously flawed. :Val:on's interest, :hen, in fitting models of argument to the reasoning expressed 10 discourses outSide that of science, in the discourse of social criticism and public policy, for example, is certainly to be applauded. At the same time, his program is not beyond criticism. Walton's idea is to expound a model of argument that is different from the model he attributes to science. Inspired by Aristotle's distinction between demonstration and dialectic he puts the arguments of science into a category that he thinks of as descendin~ fr?m the former and the arguments of ethics, politics, criticism, and law, what I will call the arguments of unscientific discourse, into a category that he thinks of as descending from the latter (pp. 70-71). The model he expounds is thus meant to be an analogue of Aristotle's dialectic. It resembles the latter in two ways. Its Context is that of a dialogue, and its premisses need have no stronger warrant than 116 John Deigh general acceptance by the participants in the dialogue it presupposes. This model, Walton holds, unlike the model that fits the arguments of science, accommodates arguments based on emotional appeals (ibid.). That is, such argu~ents, when understood according to this model, do not automatically qualify as unsound. The question of their soundness then becomes a matter of investigating the types of dialogue in which they occur. The first problem with this program is its implication that emotional appeals are always inappropriate to scientific arguments. This implication is either false or trivial. It is false if by 'a scientific argument' one means an argument that a scientist puts forth as part of his or her work. For part of a scientist's work inclu.des criticism .ofthe work of others, and while most scientists might regularly refram from emotIOnal appeals when criticizing the work of their peers, they are far less restrained in their criticisms of what they see as quackery. Ad hominem attacks, which belittle the training, credentials, or expertise of their targets, are common to such criticisms, and they may be as sound or fallacious as ad hominem attacks in other contexts. Similarly for the appeals to fear that are comn:only made (e.g., in medicine) when the influence of a crackpot theory is perceived as not merely unfortunate but dangerous. One might of course try to define 'scientific argument' in a way that excludes these and other examples of emotional appeal in the arguments scientists put forth as part of their work, but recourse to a definition would appear only to shift the problem to the other horn of the dilemma. For it is hard to see how one could give such a definition without trivializing the issue. What leads to the belief that a genuinely scientific argument must be free of emotional appeal is the idea of a scientific argument as an argument that is based entirely on hard facts and unshakable truths. The idea recalls Aristotle's notion of demonstration and guides Walton as he develops his program (ibid.; also pp. 42-43). The metaphors at its core, however, have to be cashed, and the same dilemma recurs when one tries to cash them. No one, looking at the liveliest controversies . i~ contemporary science, controversies in evolutionary biology, say, or cogllltive psychology, would describe the arguments the disputants advance as based entirely on hard facts and unshakable truths. Either the idea of scientific argument as based entirely on hard facts and unshakable truths is false or it reduces to a merely verbal point. ' The guiding influence of Aristotle's distinction between demonstration and dialectic produces a second and still more serious problem in Walton's program. The problem arises from the looseness of Walton's thinking, which one can reconstruct as follows. If scientific arguments are arguments based entirely on hard facts and unshakable truths and the arguments of unscientific discourse do not fit this idea, then they must be based on something squishier and shakier. Aristotle's notion of dialectic, where the warrant for introducing a premiss is acceptan~e by all participants in the dialogue and where the acceptance of a premiss IS always presumptive (which is to say, that a premiss, even after its Review a/Walton 117 initial acceptance, is open to being challenged), offers a fitting alternative. Hence, Walton proposes, we should adopt this alternative model according to which the arguments of unscientific discourse are to be understood and assessed as contributions to dialogues. The proposal, however, comes rather quickly. Nothing in what Walton says ever justifies construing these arguments generally as contributions to dialogues. What is more, it does not appear to be needed to correct the distorted opinion of such arguments that misleading idealizations of science have caused. As we saw in the account I gave above of how the context of an argument based on emotional appeal can affect its cogency, one can, consistently with the "deductive-semantic" method, correct such distortion. There is reason, then, to think that Walton's dialectical model is useless for understanding many of 'the arguments of unscientific discourse and uniIluminating as a general conception intended to apply to all. Everyday we read and hear arguments by people with whom we are not in the least engaged in dialogue: writers of newspaper editorials and op-ed pieces, commercial advertisers selling their merchandise on radio or TV, politicians and advocates of political causes whose literature appears in our mailboxes, etc. Sometimes, to be sure, these arguments are contributions to a discussion whose other contributions we can also read or hear. But not always, and in any event much of the time we can understand and assess these arguments as if they were directed solely to an audience of listeners or readers. Such one-way argumentation is common in modern life. Accordingly, as critical listeners or readers, we analyze the arguments we receive into their premisses and inferences and assess the believability of the former and the cogency of the latter, all without recourse to the complications and irrelevancies that regarding them as contributions to dialogues would bring. Surely, fictionalizing their context or our relations to their authors would throw no further light on them. Since we are simply trying to make up our own minds about whether to be convinced by these arguments, nothing is to be learned from pretending that we are engaged in a dialogue with the arguer. Nothing, for instance, is to be learned from determining how much of an answer the arguer owes to some question we could raise or whether the burden of proof lies with him or us on some matter we could dispute. Whatever importance dialectic had to the study of argument in Aristotle's time, th~ subsequent inventions of Guttenberg and Marconi have made it largely ullimportant to the study of argument in ours. Let us turn next to Walton's account of the fallacies to which emotional appeals are liable. Because Walton ties his conception of fallacy to his dialectical model of argument, worries about the usefulness of the latter spill over onto the former. Indeed, his analysis of fallacious emotional appeals is the least satisfactory part of his book. Walton, in laying out his method, defines fallacies as "misused techniques of argumentation that go against the goals of the dialogue the participants in a given argument are supposed to be engaged in" (p. 16); or again as "a technique of argumentation that may in principle be reasonable but that has been misused in a given case in such a way that it goes strongly against 118 John Deigh or hinders the goals of the dialogue" (p. 18). He then distinguishes a number of different types of dialogue, each with a different goal (pp. 19-23). His idea is that a technique of argumentation, which could go against the goal of some dialogue and thus qualify as a fallacy in that dialogue, need not be a fallacy every time it is used in a dialogue of that type, for sometimes it might be used in such a dialogue without going against the dialogue's goal (p. 18). Further, it need not be a fallacy when used in a dialogue of a different type, for it can go against the goal of one type of dialogue without going against the goals of all (p. 23). This idea would make sense if fallaciousness were a function of the goals of dialogues in which fallacious arguments could occur. But it plainly is not. Affirming the consequent is .a fallacy in virtue of the type of inference it is. Its fallaciousness does not vary with the type of dialogue in which it occurs And similarly for petitio principii and equivocation. Contrary to Walton's definition, then, the occurrence of a fallacy in the course of a dialogue does not necessarily make it harder for the participants in that dialogue to achieve its goals. It could even make it easier. Consider negotiation, a type of dialogue that is central to Walton's discussion of argumentum ad baculum. A question begging argument offered in the course of a negotiation might, by convincing an otherwise stubborn negotiator to give up some de~and, help to overcome an impasse and thus advance the dialogue's goal of reachlOg agreement. To be sure, the goal in this case would have been reached through error or deception, but that merely shows that error and deception are not always inconsistent with a dialogue's goal. The participants, after all, might be su~ficiently happy with the result as to be unconcerned about its having been arrIved at through error or deception. "Lucky for us you didn't see the fallacy in my argument," the negotiator who begged the question might say to his adversar~. "For we might still be arguing over sick-leave." And the latter, being happy with the result, could easily agree. Would this, then, show that the question-begging argument wasn't really fallacious, that the negotiator who advanc~d it had misdescribed his own argument? This would be a rather dotty conclusIOn to draw. Yet on Walton's definition, it would seem to follow. The defect in Walton's definition reflects the same penchant for excessive reconceptualization that we saw in his alternative model for understanding the arguments of unscientific discourse. Like that model, his definition is a response to ~roblems with treatments of appeals to emotion one can find in introductory logiC textbooks. Specifically, it is a response to dubious examples of fallacious appeals to emotion that these texts sometimes give. Walton here makes an acute an~ valuable observation. Greater attention to the context of these appeals, he pOlOts out, makes them less obviously qualified to be "textbook" examples of whatever fallacy they are presented as exemplifying (pp. 80-82). Thus, a textbook might give as an example of the fallacy of argumentum ad populum a f~agment of a politician's speech given on the hustings, a speech filled with the klOd of appeal to popular sentiment that virtually every politician in a democracy must make. To criticize the speech as being filled with fallacies, Walton Review o/Walton 119 observes, is to miss its point. When a politician is rallying his supporters with words that identify him with popular causes and ideals, he is doing little more than cheerleading, and one could hardly be a cheerleader without appealing to emotion. Similarly, Walton points out, textbooks commonly give as examples of argumentum ad baculum threats that are a normal part of the give-and-take between adversaries locked in serious negotiations. Diplomacy, for instance, invariably mixes offers and threats with genuine arguments. Convincing the other side of the truth of one's view or the rightness of one's position is only part of a typical strategy for settling a dispute on terms acceptable to one's own side. Threatening to make the other side's situation worse if they don't come around is another. And to dismiss such threats as fallacious appeals to fear shows a singular blindness to the nature of serious negotiations. Walton is therefore right, and importantly right, to criticize textbook treatments of examples like these that display such blindness. He goes wrong, however, in the lesson he draws from this criticism. The textbooks' mishandling of these examples is not due to a failure on the part of the books' authors to comprehend how the occurrence of a fallacy in an argument depends on the type of dialogue to which the argument contributes. It is due rather to their failure to recognize the irrelevance to a discussion of fallacious appeals to emotion of emotional appeals that serve other purposes than advancing an argument. Fallacies are defects in arguments. Specifically, they are ~efects in the reasoning an argument represents. Hence, an emotional appeal that IS not part of an argument, that is not part of an attempt to win assent to or acceptance of a proposition ostensibly through the exercise of reason as represented by a series of other propositions, cannot be fallacious. The imperceptiveness of criticizing the politician and the diplomat, in the above examples, for making fallacious appeals to emotion consists in a failure to appreciate that neither means to be giving an argument when he appeals to popular sentiment or fear. Cheerleading and unvarnished coercion are not exercises in argumentation. Curiously, Walton denies this (p. 147). Drawing on an essay by John Woods, he explicitly asserts that even simple and direct threats are arguments. Following Woods, he describes them as "prudential arguments" and uses an example of being robbed to support his claim. What he says, though, shows only that he has confused the reasoning the victim of a robbery goes through when ordered, at the point of a gun, say, to turn over his wallet with the content of the robber's speech. Such confusion is clear in the analogue in which I conclude that ~ ought to retreat back into my ho~se when a growling, jittery Doberman turns up oos.e on my front porch. While I may reason prudentially, the Doberman obVIOusly isn't making an argument by growling and baring his teeth. Why then should we think an argument is being made when the menace issues a threat and brandishes a gun instead? Walton, again following Woods, conjures up an ~rgument for the robber to intend to convey in issuing his threat (thus unwittingly InViting us to imagine variations on a famous scene in Woody Allen's Take the 120 John Deigh Money and Run). But attributing such an argument to a robber is fanciful. Most robbers, it is safe to assume, operate on the simple thought, "People will do what I say when I point a gun at them." The chief consequence of this confusion is that Walton, like the authors of the textbooks he criticizes, offers as examples of "fallacious appeals to emotion" threats, entreaties, and other pieces of speech that do not warrant the description. Thus his own mistaken examples of fallacious appeals to fear include a criminal defendant's threatening a potential witness with bodily injury if he testifies against him (pp. 159-160) and another defendant's implicitly threatening the jurors at his trial by asking the judge about the arrangements for their security (pp. 179-180). The fallaciousness of these appeals, Walton declares, consists in their being violations of the rules governing the conduct of criminal trials, as if showing such contempt for the rules of justice were illogical. They are fallacious on his analysis because, being obstructions of justice, they go against the goals of the dialogue that, on his view, a criminal trial represents. But whether or not they go against the goals of some dialogue represented by a criminal trial, they are not, in either case, part of an argument that one or the other defendant is making. Hence, they cannot exemplify a fallacy. To mistake issuing an immoral threat for making an illogical argument, as Walton does, indicates just how far from the subject Walton's use of the pragma- dialectical method takes him. In the opening chapter, Walton proposes, for the purpose of studying fallacious appeals to emotion, replacing as the standards of sound reasoning whose violation is implicit in the definition of fallacy the principles of deductive logic with the norms that define various speech acts and the conditions of their felicitous performance. This replacement initiates Walton's unconventional approach, which he then follows through the examination of the several types of emotional appeal that are the main topics of his discussion. By the end of the book, however, one comes to see that the replacement has not yielded a distinctive approach to the study of fallacies so much as turned that study into the study of something else. Along the way, Walton offers some very perceptive criticisms and points on the subject of emotional appeals. His critique of the standard textbook treatment of the subject is a valuable counter to the excesses of the conventional wisdom that this treatment perpetuates, and his observations on the necessity of considering context when assessing the cogency of arguments based on emotional appeals throw important light on the subtleties of such assessments. Review a/Walton 121 In these and other ways his book corrects and enlarges our understanding of informal fallacies. Walton's success at improving our understanding, however, comes in spite of his use of the pragma-dialectical method and not because of it. JOHNDEIGH DEPARTMENT OF P HILOSOP HY NORTHWESTERN UNIVERSITY EVANSTON, IL 60208-1315 o