INFORMAL LOGIC XIII.3, Fall 1991 Book Review Larry Wright's Practical Reasoning RICHARD FELDMAN University oj Rochester Wright, Larry. (1989). Practical Reasoning. San Diego, CA: Harcourt Brace Jovano- vich. Pp. xii, 1-435. ISBN 0-15-571041-9. Paper. Larry Wright's Practical Reasoning is a textbook for reasoning courses. Wright focuses on techniques for analyzing argu- ments in which the conclusion provides an explanation of the data reported in the premises. He draws on ideas and issues from philosophy of science and, to a lesser extent, epistemology. The book is decid- edly not one which applies the skills devel- oped in traditional symbolic logic courses to everyday arguments. In Chapter 1 Wright describes briefly the nature and purpose of arguments and introduces a standard format for schema- tizing arguments drawn from ordinary prose passages. He says that the two main functions of arguments are "to assure our- selves and to persuade somebody else" (23). Both functions receive attention throughout the text, with more emphasis placed on the former. In Chapter 1 Wright also introduces a principle of charity to use in schematizing arguments. In Chapter 2, "Induction and Deduc- tion", Wright divides arguments into two categories. In his terminology, an argu- ment is deductive when the conclusion is "contained within the supporting state- ments" (39). He proposes, as equivalent accounts of deduction, the ideas that deductive arguments are those in which "if you affirm the support and deny the con- clusion you have contradicted yourself' (40) and that they are ones in which "you can't get out ofthe conclusion without giv- ing up something in the support" (41). He is thus working with a notion of semantic validity. He says little about the formal structures of deductive arguments. According to Wright's classification, all non-deductive arguments are inductive. He says that every argument has an implicit question (Why did X happen? Who did it? etc.), and the conclusion of the argument is one answer to that question. When an argument is inductive, there are rival conclusions, or alternative answers to the implicit question, which the premises do not rule out. Wright explains and illus- trates the ways in which rival conclusions can be organized and generated. His dis- cussion of rival conclusions is clear and useful. I found the example Wright uses to introduce and illustrate the notion of a deductive argument unfortunately difficult. It is: S) The defendant had limited authority to write checks on his company's funds. S2 That authority did not extend to using company funds to cover his personal expenditures. S3 He did direct company funds into his private account for the purpose of vacationing in the Bahamas. C The defendant embezzled money from his employer. It is surely not obvious that this argument is deductive, in Wright's sense. If the defendant properly arranged for a loan from his company or directed company funds into his account in an amount equal to back pay owed to him, the premises are all true but he may not be guilty of embez- zlement. Thoughtful students who think of these rival conclusions will miss the point of the example. Students who don't think of these rival conclusions may still get the wrong idea about just what sort of connec- tion is required between the premises and conclusion of deductive arguments. Wright does go on to introduce more standard examples concerning Socrates' mortality to explain deduction, but I fear that using the example he does as the starting point may mislead students. Wright does not devote much attention to explaining the terms of argument evalu- ation. For example, he introduces no termi- nology to cover arguments with false premises or unjustified premises. He says that a "deductive argument can be a bad, weak argument, can fail to provide much support for its conclusion, but in only one way: by having support-claims that are implausible, clearly false, or just plain silly" (48-49). Yet later in the book (274-75) he discusses "question begging" arguments, which may be bad deductive arguments yet lack these flaws. Wright also introduces a notion of a sound argu- ment, but he says that an inductive argu- ment is sound when its conclusion is a better answer to its implicit question than is any rival conclusion (l 03). Thus, Wright is willing to call an argument sound even when it has a false premise. He thereby departs from conventional terminology. This may pose problems for students who have learned, or go on to learn, other uses of this term. Chapter 3, "Diagnostic Induction", presents the heart of Wright's method of argument analysis. He distinguishes a kind of inductive argument in which the implicit question asks for an explanation of some facts presented in the premises. Review o/Wright 191 These diagnostic inductive arguments are distinguished from arguments that yield predictions or recommendations. Wright calls the facts to be explained by the con- clusion "trace data." He devotes most of this chapter to discussions of the distinc- tion between trace and non-trace data, ranking rival conclusions, dealing with new information, and other details involved in evaluating diagnostic argu- ments. The chapter includes a detailed dis- cussion of an example concerning the sinking of the Titanic. Wright relies heav- ily on our ordinary ability to rank rival explanations and make relative plausibility judgments. He does not attempt to identify and develop the accounts of explanatory value proposed by philosophers of science. I'll mention one reservation I have about the way Wright advises students to formulate arguments. He encourages them to omit obvious or "boring" items of back- ground information from schematizations of arguments (97). I regard this as a dangerous suggestion, since too often what students regard as obvious turns out to be a crucial and controversial part of an argument. In Chapter 4 Wright applies the meth- ods described in Chapter 3 to arguments about causation and correlation, testimony, samples, and enumerative induction. All of this works rather well, I think, since one can usefully construe these arguments as arguments whose conclusions explain the statistical or testimonial evidence in their premises, and it is easy to generate rival conclusions. For example, when dealing with arguments for general causal conclu~ sions, eg., smoking causes lung cancer, rival explanations of the data would be that lung cancer causes smoking, that both fac- tors are independent products of a com- mon cause, and that the correlation is purely accidental. Wright says rather little about statistical arguments, and does not go into much detail about the various ways in which arguments based on surveys and samples can go wrong. In Chapter 5 he 192 Richard Feldman extends the method to arguments whose conclusions are predictions and recom- mendations. Wright encourages students to divide such arguments into sub-argu- ments, one part of which is a diagnostic argument of the sort he's already dis- cussed. He pays relatively little attention to the evaluation of the non-diagnostic parts of these arguments. Chapter 6 is a mercifully brief chapter in which Wright discusses some of the crit- icisms of arguments often described under the heading "informal fallacies." Those who like to give this material prominence in their courses will find Wright's discus- sion too brief. Those who prefer to avoid this murky area altogether will wish that this chapter had been omitted. I will briefly describe two of the problems with Wright's treatment of this material. Wright says that I beg the question when "I simply presuppose the conclusion I am trying to support" (274). He illustrates this with the following argument: S] Unbelievers will face bad consequences on Judgment Day. C They should believe in God. He says that in the context in which this argument is offered, "the only person who would accept S I is somebody who has already accepted C. So the argument begs the question: It can convince only those who are already convinced" (275). This strikes me as a potentially confus- ing discussion. For one thing, the notion of an argument that begs the question is char- acterized in several different ways: as an argument given by a person who "presup- poses" the conclusion; as an argument whose premise would only be accepted by someone who already accepts the conclu- sion; and as an argument that can convince only those who are already convinced. These characterizations differ from one another. The notion of a presupposition is sufficiently obscure to make the first account difficult to apply. The second account seems to imply that any simple deductive argument is question begging. A person who did not already accept the con- clusion that Socrates is mortal probably would not accept the premises that all men are mortal and Socrates is a man. So this account suggests that this argument is question begging. Since someone could be convinced by virtually any argument, the third character- ization counts rather little as question beg- ging. There are also difficulties with the example Wright uses here. It may be that only people who do believe in God will accept S 1, but it's not at all clear that the only people who will accept S 1 already accept C. Some people who accept the premise may doubt that you should believe in God. They may think that belief is a free choice about which we have no obliga- tions. Some may think that belief is invol- untary and thus not the sort of thing to which "should" applies. So, some people, at least initially, may accept the premise but not accept the conclusion. Reflection on the argument may even lead them to revise their views, perhaps leading them to see the argument as a practical argument (a variation on Pascal's Wager). It merits more consideration than a casual invoca- tion of the "begging the question" label. I think that it is better to discourage students from using this label and to advise them to consider this argument, like any other argument, on its merits. They should be taught to ask themselves whether they have reason to accept the premise and whether the premise supports the conclusion. In Chapter 6 Wright also discusses "loaded descriptions," which lead to another kind of flaw in arguments. He says that "a description is loaded if it contains an evaluation (good or bad) that not every- one would share" (275). Taken literally, this suggests that nearly every evaluation is loaded, since few evaluations are univer- sally shared. Giving students this ground for criticizing arguments is particularly dangerous since it is so open to abuse. Chapters 7 and 8 depart significantly from the style and tone of the previous chapters. They are not primarily about the analysis of arguments. Instead, Chapter 7 is a discussion of our linguistic skills, describing in a fairly abstract way our remarkable ability to use and understand language. The main way in which Wright relates this to arguments is to point out that we can take a proposed interpretation of a sentence to be the conclusion of a diagnos- tic argument whose premises describe what someone said and the background and context in which it was said. The chap- ter includes discussion of ambiguity, vagueness, and other important features of language. The emphasis here is on how well we are able to get by with a language containing these features, and on the practical need for vagueness. There is not much discussion of how ambiguity and vagueness affect argument analysis. Chapter 8 is primarily an attack on the idea that there is one proper use of each term and a warning about the dangers of revising the language to achieve additional precision. The chapter also includes a very good discussion of the uselessness of attempts to resolve genuine controversies by legislating definitions of ordinary terms. This point is well illustrated by an argument about abortion. Wright says that critics of the argument can concede to its proponents a definition of "person" that makes it true that fetuses are people. He shows that their objections to the argument just pop up in another place. Practical Reasoning is, on the whole, a good book. It effectively applies a general approach to argument analysis to a range of arguments. I believe that it is more diffi- cult and more abstract than the average text in the field, and it requires a more sophisti- cated reader than most of the texts with which I am familiar. The first six chapters conclude with study guides that some students will find useful. There are, how- ever, several noteworthy features of the book that may detract from its usefulness Review of Wright 193 for many instructors. I'll conclude by describing them. First, I suspect that most instructors would prefer a text with more exercises. Of the eight chapters, six have only one set of exercises, located at the end of the chapter. One (Chapter 6) has no exercises at all, and the last chapter has exercises at the end of each of its main sections. Although some of the exercise sets contain numerous good examples, I think that exercises and study questions at the ends of the main sections of all the chapters would be useful to instructors and would help students work their way through the book. Second, Wright says that the "aim of our examination of arguments will be to evaluate them-to say whether they are any good and why" (3). Early in the book he makes much of a principle of charity to be used in interpreting and evaluating arguments. He says, quite rightly, that the point of applying a principle of charity in formulating arguments is that it is in our interest to find the best argument we can in a passage. We learn the most by doing so, and we don't waste our time with inferior arguments. Given this view, Wright devotes surprisingly little space to helping students figure out just what the argument of a passage is. For example, in Chapter 4 when he discusses arguments for causal and statistical claims, it would be particu- larly helpful for him to work through some examples based on actual research reports of the sort students are likely to encounter. It is often not easy to retrieve the relevant information from these reports, and stu- dents who are to use Wright's method for analyzing the arguments based on these reports need advice about how to extract the crucial information from them. Third, there are important topics about which we argue frequently that get little attention in this book. Of course, no text can cover all kinds of arguments, but the omissions here are significant. For exam- ple, there is little about moral arguments. There is a discussion of arguments for 194 Richard Feldman recommendations-what we should do-in Chapter 5. Wright acknowledges that "[ c ]asual observation reveals that an abso- lute majority of arguments aired in public are for recommendations" (244). He says that these arguments often appeal to the consequences of actions and that they have premises evaluating these consequences as good or bad. However, he says almost nothing about how we make and defend these judgments about the merits of conse- quences. The main points Wright makes in his discussion of recommendations and predictions are that arguments for them often include diagnostic arguments for fac- tual claims, eg., that an action has a partic- ular consequence, and that the usual methods of argument analysis apply to these diagnostic arguments. Wright's long discussion of scientific method in Chapter 4 reveals another note- worthy omission. He devotes twenty-two pages to showing how we can view scien- tific inquiry as a diagnostic activity. How- ever, he writes about only the natural sciences, dismissing the social and behav- ioral sciences as "too controversial to be treated si~ply" (186, footnote 7). Since so much of what students are apt to read in • Teachin~ m Philosop y the popular press about scientific research concerns the social and behavioral sci- ences, this is an unfortunate omission. Fourth, significant parts of the book stray from the account of the method of argument analysis. In addition to the dis- cussions of scientific method in Chapter 4 and the discussions of language in Chapters 7 and 8 mentioned above, there is also a fairly long discourse on dispositions in Chapter 3. These discussivns take up almost one quarter of the book. I will not here raise any objections to the points Wright makes in these sections. I suspect that instructors who want to focus exclu- sively on argument analysis will regard these sections as extensive digressions. Others may find them to be welcome forays into philosophy. Instructors of reasoning courses who are looking for a challenging reasoning text and who would be satisfied with a text hav- ing the features just described should give Practical Reasoning careful consideration. RICHARD FELDMAN DEPARTMENT OF PHILOSOPHY UNIVERSITY OF ROCHESTER ROCHESTER, NY 14607 o Teaching Philosophy, edited by Arnold Wilson of the University of Cin- cinnati, provides an open forum for the exchange and evaluation of ideas, information and materials concerned with the teaching of philosophy. It is published quarterly and the subscription rates for 1992 are $55 (individu- als $22.50). back issues $15 (individuals $7). What the reviewers say: M A journal vital to our discipline .. .irreplaceable" David Kaplan. UCLA MUseful information and interesting ideas. No teacher of philosophy will read it without profit." Nicholas Rescher, University 0/ PiIlsburgh "Excellent articles and useful insights on both theory and practice. Unique and invaluable service to the profession.· Ernest Sosa, Brown University Send orders to: PHILOSOPHY DOCUMENTATION CENTER, Bowling Green State University, Bowling Green, OH 43403-0189 USA (419) 372- 2419 (800) 444-2419 Fax: (419) 372-6987