sign in sign up
<<
Home , Armen Alchian, Chaim Perelman, Economics (Aristotle), Frank Knight, Paul Samuelson, Richard Whately

American Economic Association

The of Author(s): Donald N. McCloskey Source: Journal of Economic Literature, Vol. 21, No. 2 (Jun., 1983), pp. 481-517 Published by: American Economic Association Stable URL: http://www.jstor.org/stable/2724987 Accessed: 24/10/2009 16:06

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=aea.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is a not-for- that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

American Economic Association is collaborating with JSTOR to digitize, preserve and extend access to Journal of Economic Literature.

http://www.jstor.org Journal of Economic Literature Vol. XXI (June 1983), pp. 481-517

The Rhetoric of Economics

By DONALD N. MCCLOSKEY The Universityof Iowa

The length of the acknowledgments here testifies to an unexplored feature of the rhetoric of economics, the role of the audience: like oratory, scholarship depends for its virtues on the virtues of its audience. I have been fortunate in mine. I must apologize for my amateurish understanding of what is happening in philosophy, , literary criticism, rhetorical studies, and other places beyond my competence, and ask that practitioners in these fields assist in my further education. For their early attempts I thank Evan Fales, Paul Hernadi, John Lyne, Michael McGee,Allan Megill, John Nelson, and Jay Semel of the Colloquium on Applied Rhetoric at The ; Wayne Booth, Ira Katznelson, and others at the University of in the program in , Rhetoric and , before which the earliest version was delivered; Robert Boynton, Bernard Cohn,John Comaroff;Otis Dudley Duncan, James 0. Freedman, Clifford Geertz, William Kruskal, Donald Levine, Laura McCloskey,Richard Rorty, Renato Rosaldo; and the Humani- ties Society at the University of Iowa. That the on whom I have inflicted the have reacted with such intelligent and generous encouragement suggests, as the paper does, that we are better scholars than our methodology would allow. I thank my colleagues in economics at Iowa, especially the Sanctuary Seminar in Economic Argument; Seminars at the and the National Science Foundation; my colleagues at the Institute of Advanced Studies and the Faculty of Economics at the Australian National University; seminars at the universities of Adelaide, Auck- land, Melbourne, New South Wales, Tasmania and WesternAustra- lia; at Monash, and Iowa State universities; Victoria University of Wellington; and an assemblage of economists elsewhere: William Breit, ,Arthur Diamond, Stanley Engerman, J. M. Fin- ger, , Allan Gibbard, Robert Goodin, Gary Hawke, , Albert Hirschman, Eric Jones, Arjo Klamer, Harvey Leibenstein, David Levy, Peter Lindert, Neil de Marchi, Michael McPherson, , , Larry Westphal, Gordon Winston, and Gavin Wright. Thomas Mayer's encouragement at an early stage and his detailed comments as refereefor this Journal at a later stage were exceptionally heartening and useful. 481 482 Journal of Economic Literature, Vol. XXI (June 1983)

ECONOMISTS DO NOT FOLLOW the ism labels such issues "meaningless" or of enquiry their methodologies lay "nonscientific" or "just matters of opin- down. A good thing, too. If they did they ion." Yet even positivists actually behave would stand silent on human , the as though the matters are discussable. In law of demand, random walks down Wall fact, most discussion in most sciences, and Street, the of demand for gaso- especially in economics, arises from them. line, and most other matters about which Nothing is gained from clinging to the Sci- they commonly speak. In view of the volu- entific Method, or to any methodology ex- bility of economists the many official cept honesty, clarity, and tolerance. Noth- methodologies are apparently not the ing is gained because the methodology grounds for their scientific conviction. does not describe the sciences it was once Economists in fact argue on wider thought to describe, such as or grounds, and should. Their genuine, mathematics; and because physics and workaday rhetoric, the way they argue in- mathematics are not good models for eco- side their heads or their seminar rooms, nomics anyway; and because the method- diverges from the official rhetoric. Econo- ology is now seen by many philosophers mists should become more self-conscious themselves to be uncompelling; and be- about their rhetoric, because they will cause economic science would stop pro- then better know why they agree or dis- gressing if the methodology were in fact agree, and will find it less easy to dismiss used; and, most important, because eco- contrary on merely methodo- nomics, like any field, should get its stan- logical grounds. Philosophy as a set of nar- dards of argument from itself, not from rowing rules of evidence should be set the legislation of philosopher kings. The aside in scientific argument, as even many real arguments would then be joined. philosophers have been saying now for fifty years. I. Rhetoric Is Disciplined Conversation Economics will not change much in sub- stance, of course, when economists recog- These points, elaborated below, amount nize that the economic emperor has posi- to an appeal to examine the rhetoric of tively no clothes. He is the same fellow economics. By "rhetoric" is not meant a whether philosophically naked or clothed, verbal shell game, as in "empty rhetoric" in reasonably good health aside from his or "mere rhetoric" (although form is not sartorial delusion. But the temper of argu- trivial, either: disdain for the form of ment among economists would improve words is evidence of a mind closed to the if they recognized on what grounds they varieties of argument). In Modern Dogma were arguing. They claim to be arguing and the Rhetoric of Assent Wayne Booth on grounds of certain limited matters of gives many useful definitions. Rhetoric is statistical inference, on grounds of posi- "the art of probing what men believe they tive economics, operationalism, behavior- ought to believe, rather than proving ism, and other positivistic enthusiasms of what is true according to abstract meth- the 1930s and 1940s. They believe that ods"; it is "the art of discovering good rea- these are the only grounds for science. But sons, finding what really warrants assent, in their actual scientific work they argue because any ought to about the aptness of economic metaphors, be persuaded"; it is "careful weighing of the relevance of historical precedents, more-or-less good reasons to arrive at the persuasiveness of introspections, the more-or-less probable or plausible conclu- power of authority, the charm of symme- sions-none too secure but better than try, the claims of . Crude positiv- would be arrived at by chance or unthink- McCloskey: The Rhetoric of Economics 483 ing impulse"; it is the "art of discovering century after Descartes rose from the warrantable beliefs and improving those dead. The faith built on these miracles is beliefs in shared discourse"; its purpose known in literary studies as the New Rhet- must not be "to talk someone else into a oric, new in the 1930s and 1940s from preconceived view; rather, it must be to the hands of I. A. Richards in Britain and engage in mutual inquiry" (Booth, 1974, Kenneth Burke in America (Richards, pp. xiii, xiv, 59, 137). It is what economists, 1936; Burke, 1950). In philosophy John like other dealers in ideas, do anyway: as Dewey and Ludwig Wittgenstein had al- Booth says elsewhere, "We believe in mu- ready begun to criticize Descartes' pro- tual as a way of life; we live gram of erecting belief on a foundation from conference to conference" (Booth, of skepticism. More recently Karl Popper, 1967, p. 13). Rhetoric is exploring thought Thomas Kuhn, and Imre Lakatos among by conversation. others have undermined the positivist The word "rhetoric"is doubtless an ob- supposition that scientific progress does in stacle to understanding the point, so de- fact follow Descartes' doubting rules of based has it become in common parlance. method. The literary, epistemological, If "pragmatism"and ""had not and methodological strands have not yet already suffered as much, unable to keep wound into one cord, but they belong to- clear of irrelevant associations with the gether. On the eve of the Cartesian revo- bottom line or the bomb, the title might lution the French philosopher and educa- better have been "Pragmatism'sConcep- tional reformer, Peter Ramus (fi. 1550), tion of in Economics" or "Outline brought to completion a medieval ten- of an Anarchistic Theory of Knowledge dency to relegate rhetoric to mere elo- in Economics" (WilliamJames, 1907; Paul quence, leaving in charge of reason. Feyerabend, 1975). But the enemies of so- In the textbooks that Descartes himself phisticated pragmatism and gentle anar- read as a boy probable argument was chism, as of honest rhetoric, have used the made thus for the first time wholly subser- weapons at hand. The results discourage vient to indubitable argument. Hostile to onlookers from satisfying the curiosity classical rhetoric, such a reorganization of they might have had about alternatives the liberal arts was well suited for the to coercion in philosophy, politics, or Cartesian program extending over the method. A title such as "How Economists next three centuries to put knowledge on Explain" (Mark Blaug, 1980; but see be- foundations built by philosophy and math- low) or "Why Methodology Is a Bad" ematics. The program failed, and in the would perhaps have been meeker and meantime probable argument languished. more persuasive.1Still, "rhetoric"like the In Richard Rorty's words, following others is a fine and ancient word, whose Dewey, the search for the foundations of proper use ought to be more widely knowledge by Descartes, Locke, Hume, known among economists and calculators. Kant, Russell, and Carnap was "the tri- The rhetoric here is that of , umph of the quest for certainty over the Cicero, and Quintilian among the an- quest for wisdom" (Rorty, 1979, p. 61; cf. cients, reincarnated in the Renaissance, John Dewey, 1929, pp. 33, 227). To rein- crucified by the Cartesiandogma that only state rhetoric properly understood is to the indubitable is true; which in the third reinstate wider and wiser reasoning. The reaction to the narrowing of argu- IAfter recognizing the intent, Colin Forster of ment by the Cartesian program is by now Australian National University suggested the title "The Last Paper on Methodology." But ambition broad. Its leading figures range from pro- must have limits. fessional philosophers (Stephen Toulmin, 484 Journal of Economic Literature, Vol. XXI (June 1983)

Paul Feyerabend, Richard Rorty) to a size its pervasiveness in modern thinking miscellany of practitioners-turned-philos- well beyond scholarship it is best labeled ophers in chemistry (Michael Polanyi), law simply "modernism," that is, the notion (Chaim Perelman), and literary criticism (as Booth puts it) that we know only what (Wayne Booth). The reach of the idea now- we cannot doubt and cannot really know adays that argument is more than syllo- what we can merely assent to. gism is illustrated well by the lucid treat- Among the precepts of modernism are: ment of it in what would seem an unlikely place, by Glenn Webster, Ada Jacox, and (1) Prediction (and control) is the- goal Beverly Baldwin in "Nursing Theory and of science. the Ghost of the Received View" (1981, (2) Only the observable implications (or pp. 25-35). The reach, however, has not predictions) of a theory matter to its extended to economics. Austrian, institu- truth. tionalist, and Marxist economists, to be (3) Observability entails objective, re- sure, have for a century been attacking producible experiments. certain parts of as the basis for (4) If (and only if) an experimental impli- economic knowledge. But they have cation of a theory proves false is the seized on other parts with redoubled fer- theory proved false. vor, and have so expressed their remain- (5) Objectivity is to be treasured;subjec- ing doubts as to make them unintelligible tive "observation" (introspection) is to anyone but themselves. In their own not scientific knowledge. way they have been as narrowing as thor- (6) Kelvin's Dictum: "When you cannot oughgoing positivists-the rejection of express it in numbers, your knowl- , for instance, would be rea- edge is of a meagre and unsatisfac- sonable only if its more naive claims were tory kind."2 taken seriously. For the rest, economists (7) Introspection, metaphysical belief, have let philosophical scribblers of a few aesthetics, and the like may well fig- years back supply their official thinking ure in the discovery of an hypothesis about what a good argument is. but cannot figure in its justification. (8) It is the business of methodology to demarcate scientific reasoning from II. The Methodology of Official non-scientific, positive from norma- Economics Is Modernist tive. Economists have two attitudes towards (9) A scientific explanation of an event discourse, the official and unofficial, the brings the event under a covering explicit and the implicit. The officialrheto- law. ric, to which they subscribe in the abstract (10) Scientists, for instance economic sci- and in methodological ruminations, de- entists, have nothing to say as scien- clares them to be scientists in the modern tists about values, whether of moral- mode. The credo of Scientific Method, ity or art. known mockingly among its many critics as the Received View, is an amalgam of 2From Sir William Thomson (Lord Kelvin), Popu- , behaviorism, operation- larAddresses, edition of 1888-1889, quoted in Kuhn, 1977, p. 178n. An approximation to this version is alism, and the hypothetico-deductive inscribed on the front of the Social Science Research model of science. Its leading idea is that Building at the . , all sure knowledge is modeled on the early the famous University of Iowa , is said to have remarked on it one day: "Yes, and when you 20th century's understanding of certain can express it in numbers your knowledge is of a pieces of 19th century physics. To empha- meagre and unsatisfactorykind." McCloskey: The Rhetoric of Economics 485

(11) Hume's Fork: "When we run over - say is part of the intellectual equipment braries, persuaded of these princi- of most economists, and its arguments ples, what havoc must we make? If come readily to their lips. we take in our hand any volume- Premeditated writings on method, not of divinity or school , for excluding Chicago's own, are more careful instance-let us ask, Does it contain than the remark in the course of other any abstract reasoning concerning business that reveals modernism in its quantity or number? No. Does it con- rawer form. In precept one can be vague tain any experimental reasoning con- enough to earn the assent of everyone; cerning matter offact and existence? in practice one must make enemies. Kal- No. Commit it then to the flames, for man Cohen and Richard Cyert, to take it can contain nothing but sophistry one among many examples of first-chapter and illusion" (italics his [1748], 1955, methodology in economics texts, present p. 173). in their book an outline of modernism, which they assert is the method "used in Few in philosophy now believe as many all scientific analyses" (1975, p. 17). The as half of these propositions. A substantial, "method" they then outline, with a bibli- respectable, and growing minority be- ography heavily weighted towards logical lieves none of them. But a large majority positivism and its allies, reduces to an ap- in economics believes them all. peal to be honest and thoughtful. Only For instance, the leading methodolo- when such a phrase as "at least in princi- gists in economics do. It is odd but true ple testable by experiment and observa- that modernism in economic methodol- tion" (p. 23) is given content by practice ogy is associated with the Chicago School.3 do we know what is at stake. To be sure, The main texts of economic modernism, vague precepts are not without their uses. such as Milton Friedman's "The Method- When Friedman wrote, for instance, the ology of " (1953) or practice of economics was split into theory and 's "De without fact and fact without theory. His Gustibus Non Est Disputandum" (1977), modernist incantations, supported by bear a Chicago postmark; and the more choruses of philosophers, were at the time extreme interpretations of the texts flour- probably good for the souls of all con- ish among economists bearing a Chicago cerned. degree. What is odd about it is that a group Friedman's essay was even then more so annoying to other economists in most post-modernist than one might suppose of its activities should have their assent from slight acquaintance with its ideas. He in the matter of official method: Oddly, did, for example, mention with approval a watered down version of Friedman's es- the aesthetic criteria of simplicity and might use 3Nothing in this essay is meant to give comfort fruitfulness that an economist to the enemies of Chicago. Having long been a vic- to select among a multiplicity of theories tim of their anti-Chicago dogmatism, I am not im- with the same predictions, though in the pressed by the assertion that Chicago economics is next sentence he attempted to reduce peculiarly dogmatic. Chicago is merely a particularly clear and candid version of a dogmatic impulse com- them to objective matters of prediction mon to all economics, expressing itself in methodo- (p. 10). He accepted that questionnaires, logical imperatives. Economists appear to believe that economics is too important to be left to the forbidden to the modernist in economics, open-minded, and especially must never be left to are useful for suggesting hypotheses, anyone lacking faith in some approved formula for though in the next sentence he asserted achieving knowledge. Chicago is no worse than the rest. Immo, civis Chicagonussum, subspecies TP (cf that they are "almost entirely useless as Melvin Reder, 1982). a means of testing the validity of eco- 486 Journal of Economic Literature, Vol. XXI (June 1983)

nomic hypotheses"(p. 31n). He empha- III. Modernism Is a Poor Method sized the role of the rhetorical community to which the scientist speaks in producing Modernism Is Obsolete in Philosophy conviction-whether made up of sociolo- There are a great many things wrong gists, say, or of economists-though in the with modernism as a methodology for sci- next sentence he returned to an "objec- ence or for economic science.4 Even when tive" theory of testing. Like Karl Popper, philosophically 'inclined, economists ap- Friedman appeared to be struggling to es- pear to read about as much in professional cape the grip of positivism and its intellec- philosophy as philosophers do in profes- tual traditions, though with only sporadic sional economics. It is unsurprising, then, success. Perhaps that the locus classicus that the news of the decline of modernism of economic modernism contains so much has not reached all ears. From a philoso- that is anti-modernist indicates that mod- pher's point of view the worst flaw in the ernism cannot survive intelligent discus- hostility to the "metaphysics" that mod- sion even by its best advocates. ernism sees everywhere is that the hostil- The unpremeditated remark in the heat ity is itself metaphysical. If metaphysics of economic argument, however, usually is to be cast into the flames, then the meth- has a crudely modernist content, often in odological declarations of the modernist Friedman's very words. An article by family from Descartes through Hume and Richard Roll and Stephen Ross on , to and and for instance, asserts that "the theory Comte Russell Hempel Popper will be the first to For this and other should be tested by its conclusions, not go. good reasons philosophers agree that strict by its assumptions" and that "similarly, logical positivism is dead, raising the ques- one should not reject the conclusions de- tion whether economists are wise to carry rived from firm profit maximization on the on with their necrophilia.5 basis of sample surveys in which managers In the economic case the metaphysical claim that they off profit for social position akin to logical positivism is not good" (1980, p. 1093 and footnote). The well argued, probably because its roots lie same can be found elsewhere, in nearly more in the philosophizing of physicists identical terms, all dating back to Fried- from Mach to Bridgeman than in the par- man's essay: William Sharpe (1970, p. 77), allel thinking of professional philosophers. for instance, writing on the same matter It is at least obscure what might be the as Roll and Ross, takes it as a rule of polite appeal of "operationallymeaningful state- scientific behavior that "the realism of the ments" (, 1947, p. 3 and assumptions matters little. If the implica- tions are reasonably consistent with ob- served 4 The overdiscussedquestion of whether there can phenomena, the theory can be said be a -free social science will not be much dis- to 'explain'reality" (1970, p. 77). Repeated cussed here, but it must be accounted one of the often, and exhibiting modernism as well chief failings of modernism that it places moral argu- in their ment outside the pale of rational discussion. In this devotion to objective evidence, connection it should be more widely known that quantifiable tests, positive analysis, and MorrisSchlick, the founder of the Vienna Circle of other articles of the faith, such phrases logical positivism and a vigorous lecturer on the have the ring of theme that moral knowledge is no knowledge at all, incantation. Modernism was murdered in 1936 by one of his students. is influential in economics, but not be- 6See John Passmore, 1967. Karl Popper quotes cause its premises have been examined Passmore with approval for the motto of a chapter of his own entitled "Who Killed Logical Positivism?" carefully and found good. It is a revealed, (Popper, 1976, pp. 87-90), in which he confesses to not a reasoned, religion. the murder. McCloskey: The Rhetoric of Economics 487

throughout)or "valid and meaningful pre- of failures of premises irrelevant to the dictions about phenomena not yet ob- main hypothesis in question. The hypothe- served" (Friedman, p. 7) as standards sis in question is insulated from crucial test against which all but mathematical asser- by the ancillary hypotheses necessary to tions are to be judged. Samuelson, Fried- bring it to a test. This is no mere possibility man, or their followers do not present rea- but the substance of most scientific dis- sons for adopting such metaphysical agreement: "Your experiment was not positions, except for confident assertions, properly controlled"; "You have not at the time correct, that they were the solved the identification problem"; "You received views of philosophers on the have used an equilibrium (competitive, method of science. The trust in philosophy single-equation) model when a disequi- was a tactical error, for the philosophy it- librium (monopolistic, 500-equation) self has since changed. Some philosophers model is relevant." And even if the one now doubt the entire enterprise of epis- hypothesis in question could be isolated, temology and its claim to provide founda- the probabilistic nature of hypotheses, tions for knowledge (Richard Rorty, most especially in economics, makes cru- 1982b). A great many doubt the prescrip- cial experiments non-crucial:chance is the tions of modernist methodology. ever present alternative, the Hn that spoils falsificationism. Falsification Is Not Cogent Prediction Is Impossible in Economics A prescription that economic methodol- ogists have in common, for instance, is an The common claim that prediction is emphasis on the crucial falsifying test, sup- the defining feature of a real science, and posedly the hallmark of scientific reason- that economics possesses the feature, is ing. But philosophers have recognized for equally open to doubt. It is a cliche among many decades that falsification runs afoul philosophers and historians of science, for of a criticism made by the physicist and instance, that one of the most successful philosopher Pierre Duhem in 1906, evi- of all scientific theories, the theory of evo- dent at once without philosophical read- lution, has no predictions in the normal ing to an economist who has tried to use sense, and is therefore unfalsifiableby pre- falsification for science. Suppose that the diction. It is at least suggestive of some- hypothesis H ("British businessmen per- thing odd in prediction as a criterion for formed very poorly relative to Americans useful economics that Darwin's theory and Germans in the late 19th century") was inspired by , a sys- implies a testing observation 0 ("Mea- tem as it happens erroneous in most of sures of total factor productivity in iron the predictions it made. With no apparent and steel show a large difference between awareness of the incongruity, Friedman British and foreign steelmaking"); it im- quoted Alchian's revival of the connection plies it, that is, not by itself, but only with (, 1950) in the midst of his the addition of ancillary hypotheses H1, most famous piece of predictionist meta- H2, and so forth that make the measure- physics ("the leaves are positioned as if ment possible ("Marginal productivity each leaf deliberately sought to maximize theory applies to Britain 1870-1913"; the amount of sunlight it receives"). "British steel had no hidden inputs offset- In any event, predicting the economic ting poor business leaderships"; and so future is, as put it, "be- forth). Then of course not-0 implies not- yond the power of any mortal man" (1949, H-or not-H1 or not-H2 or any number p. 867). What puts it beyond his power 488 Journal of Economic Literature, Vol. XXI (June 1983) is the very economics he uses to make the corrosive a standard of belief for a real prediction. When for a big human. As the chemist and philosopher bank predicts lower rates after Michael Polanyi put it, the methodology Christmas, and has not before the predic- of modernism sets up "quixotic standards tion placed his worth in margin loans of valid meaning which, if rigorouslyprac- on bonds, properly hedged and insured ticed, would reduce us all to voluntary against variance, he is behaving either ir- imbecility" (1962, p. 88). Modernism rationally or self-deceivingly. If he knows promises knowledge free from doubt, the expected value of the future, he for metaphysics, morals, and personal convic- some reason chooses not to take the unlim- tion; what it delivers merely renames as ited that such Faustian knowledge Scientific Method the scientist's and espe- can surely bring, and is willing for some cially the economic scientist's metaphys- reason instead to dissipate the opportunity ics, morals, and personal convictions. It by the act of telling others about it. If he cannot, and should not, deliver what it does not really know, then he faces no promises. Scientific knowledge is no dif- such unexploited opportunity. But then ferent from other personal knowledge he has perhaps no business talking as (Polanyi, 1962). Trying to make it differ- though he does. Predictionism cannot be ent, instead of simply better, is the death rescued by remarking that the big bank of science. economist makes only conditional predic- In other words, the literal application tions. Conditional predictions sell for their of modernist methodology cannot give a value in a soft : if the sea were to useful economics. The best proofs are his- disappear, a rock would accelerate in fall- torical. In his Against Method (1975) Paul ing from sea level to the sea floor at about Feyerabend uses an interpretation of Gal- 32.17 feet per second per second. But a ileo's career to attack the claims of pre- serious prediction has serious boundary scriptive methodology in physics; the conditions. If it does it must answer again same point can be made about economics. the American Question: If you're so smart Had the modernist criterion of persuasion why aren't you rich? At the margin (be- been adopted by Galileo's contemporar- cause that is where economics works) and ies, he argues, the Galilean case would on average (because some people are have failed. A grant proposal to use the lucky) the industry of making economic strange premise that terrestrial optics ap- predictions, which includes universities, plied also to the celestial sphere, to assert earns only normal returns. that the tides were the sloshing of water on a mobile earth, and to suppose that ModernismItself Is Impossible, and Is Not the fuzzy views of Jupiter's alleged moons Followed would prove, by a wild analogy, that the The most damaging, however, of these planets, too, went around the sun as did lesser criticisms of the modernist method- the moons around Jupiter would not have ology is that if taken at its word it is narrow survived the first round of peer review to absurdity. Consider again the steps to in a National Science Foundation of 1632, modernist knowledge, from predictionism at any rate if that one (unlike ours) were through Kelvin's Dictum to Hume's Fork. wedded to modernist ideology. The argu- If economists (or physicists) confined ment applies widely to the history of phys- themselves to economic (or physical) ics: observational anomalies in the experi- propositions that literally conformed to ments testing Einstein's theories were such steps they would have nothing to say. ignored for many years, to be revealed Cartesian or Humean skepticism is too as errors of measurement long after the McCloskey: The Rhetoric of Economics 489 theories had been embraced, embraced the castle. An important blow for mone- on grounds of "the reason of the matter," tarism was Friedman and Schwartz' big as Einstein was fond of saying (Feyera- book, A Monetary History of the United bend, 1975, pp. 56-57). States, 1867-1960. It established a correla- Historians of biology have uncovered tion, which Keynesians would not deny, one after another case of cooking the sta- between and money income. The tistical results to fit modernist precepts of significance of the correlation, however, what counts as evidence, from Pasteur and depended on the assumption that money Mendel down to the present. The mea- caused and that money was deter- surement of IQ has been a scandal of self- minable by the monetary authority (in deception and bold fraud in the name of 1929-1933, for example) despite the open- scientific method from its beginning (Ste- ness of the American to trade phen Jay Gould, 1981). Perhaps modern- in both and money itself. Nonethe- ism fits poorly the complexities of biology less, what was telling in the debate was and psychology: straining after evidence the sheer bulk of the book-the richness of a sort typically available only in the sim- and intelligence of its arguments, how- plest experiments in physics may not suit ever irrelevant most of the arguments their frontiers. It suits the frontiers of eco- were to the main point. nomics poorly enough. For better or A modernist method thoroughly ap- worse the in eco- plied, in other words, would probably stop nomics would not have happened under advances in economics. What empirical the modernist legislation recommended anomaly in the traditional tale inspired for the method of science. The Keynesian the new labor economics or the new eco- insights were not formulated as statistical nomic history? None: they were merely propositions until the early 1950s, well af- realizations that the logic of economics ter the bulk of younger economists had had not exhausted its applicability at con- become persuaded they were true. By the ventional borders. What observable impli- early 1960s liquidity traps and accelerator cations justify the of intellect models of investment, despite failures in since 1950 in general equilibrium theory? their statistical implementations, were For all the modernist talk common among taught to first-yearstudents of economics its theorists, none; but so what? Could ap- as matters of scientific routine. Modernist plications of economics to legal questions methodology would have stopped all this rely entirely on objective evidence? No; in 1936: where was the evidence of an but why would one wish to limit the play objective, statistical, controlled kind? of understanding? And so forth. There is Nor was the monetarist counterrevolu- nothing to be gained and a great deal to tion a success for modernist methodology, be lost by adopting modernism in eco- though so powerful had the methodology nomic methodology. become by the 1960s in the minds of econ- The very point is economic. In order omists and especially of monetarist econo- for an economic theory to be tested, Ron- mists that most of the explicit debate took ald Coase points out, some economists place in its terms. Yet in truth crude ex- must care enough about it to bother. They periments and big books won the day, by care only when it is believed by some in- their very crudeness and bigness. The vestigators-they and their allies or some Kennedy tax cut boosted the Keynesians significantgroup of opponents. Only when to their peak of prestige; the of many believe is there a demand for tests. the 1970s brought them down again, leav- Fortunately, "economists, or at any rate ing the monetarists as temporary kings of enough of them, do not wait to discover 490 Journal of Economic Literature, Vol. XXI (June 1983)

whether a theory's predictions are accu- make it fit a philosopher's idea of the ulti- rate before making up their minds"; to mate good. wait in proper modernist style "would re- It is hard to take such claims seriously. sult in the paralysis of scientific activity" Einstein remarked that "Whoever under- (Coase, 1982, p. 14) because no one would takes to set himself up as a judge in the have an incentive to choose one out of field of Truth and Knowledge is ship- the infinite number of hypotheses for test. wrecked by the laughter of the gods" (Ein- Even quantitative studies, he argues, rely stein, 1953, p. 38). Modernism sets up a heavily on pre-quantitative arguments court of the Red Queen ("Normativeargu- founding belief, and he quotes with ap- ment," she says, "off with his head"), and proval T. S. Kuhn's remark that "the road the gods laugh merrily. Any methodology from scientific law to scientific measure- that is law-making and limiting will do so. ment can rarely be traveled in the reverse It will do so with the noblest intentions, direction" (Coase, p. 18, quoting Kuhn, but economists are fond of pointing out 1977, p. 219). The laws come from a rheto- in like cases that noble intentions can have ric of tradition or introspection, and in bad consequences. The methodologist physics as in economics "quantitative fancies himself the judge of the practi- studies . . . are explorations with the aid tioner. His proper business, though, is an of a theory" (Coase, p. 17), searches for anarchistic one, resisting the rigidity and numbers with which to make specific a pretension of rules. I. A. Richards applied theory already believed on other grounds the point to the theory of metaphor: "Its (see Edward Leamer, 1978, and the business is not to replace practice, or to discussion below). Modernism is imprac- tell us how to do what we cannot do al- tical. ready; but to protect our natural skill from the interference of unnecessarily crude Any Method Is Arrogant and Pretentious views about it" (1936, p. 116). The objections to modernist method so The crudeness of modernist methodol- far, however, are lesser ones. The greater ogy, or of any methodology reducible to objection is simply that modernism is a rigid precept, is bad; but that it is allowed method. It sets up laws of argument drawn to interfere with practice is worse. The from an ideal science or the underlying custom of methodological papers in eco- or the essence of knowl- nomics is to scold economists for not allow- edge. The claim is that the philosopher ing it to interfere more. MarkBlaug's use- of science can tell what makes for good, ful book summarizing the state of play of useful, fruitful, progressive science. He in 1980, The knows this so confidently that he can limit Methodology of Economics: Or How Econ- arguments that worthy scientists make omists Explain, is a recent case in point. spontaneously, casting out some as un- It would be better subtitled "How the scientific, or at best placing them firmly Young Karl Popper Explained," since it in the "context of discovery." The philoso- repeatedly attacks extant arguments in pher undertakes to second-guess the sci- economics for failing to comply with the entific community. In economics the rules Popper laid down in Logik der For- claim of methodological legislation is that schung in 1934. Blaug's exordium is typi- the legislator is not merely expert in all cal of the best of the methodologists in branches of economic knowledge within economics: "Economists have long been sound of his proclamations but expert in aware of the need to defend 'correct' prin- all possible future economics, limiting the ciples of reasoning in their subject; al- growth of economics now in order to though actual practice may bear little McCloskey: The Rhetoric of Economics 491 relationship to what is preached, the the word "evidence." Should it all be "ob- preaching is worth considering on its own jective," "experimental," "positive," "ob- ground" (Blaug, p. xii). Words like these servable"? Can it be? In The Open Society flow easily from a modernist's pen. But and Its Enemies (1945) Popper closes the why would preaching unrelated to actual borders of his society to psychoanalysts practice be worth considering at all? Why and Marxistson the grounds that they do do economists have to defend in the ab- not conform to the modernist notion of stract their principles of reasoning, and evidence prevalent there. He would also before what tribunal? A case for having have to close it to physicists from Galileo a methodology-whether logical positivist Galilei to particle charmers. An economist or Popperian or Austrian or Marxist- bracero, surely, would be deported on the would be expected to give answers to the next truck from such an open intellectual questions of why, but commonly does not. society. Recent and ordinary Other Sciences Do Not Follow Modernist good sense suggest that it cannot. Blaug's Methods peroration is frankly prescriptive, taking economic rhetoric directly from philoso- For all its claims to the scientific priest- phy: hood, then, economics is different from What methodology can do is to provide criteria the man-in-the-street's image of Science. for the acceptance and rejection of research Economists should be glad that their sub- programs, setting standards that will help us ject fits poorly with this image and well to discriminate between wheat and chaff. The with the New Rhetoric, as do studies long ultimate question we can and indeed must pose foreign to economics such as the study of about any research program is the one made familiarby Popper:what events, if they materi- literature or law or politics. Economics, alize, would lead us to reject that program? in other words, is not a Science in the way A program that cannot meet that question has we came to understand that word in high fallen short of the highest standardsthat scien- school. tific knowledge can attain [1980, p. 264]. But neither, really, are other sciences. It sounds grand, but Einstein's gods are Other sciences, even the other mathemat- rolling in the aisles. Why should a dubious ical sciences, even the Queen herself, are epistemological principle be any test of rhetorical. Mathematics appears to an in- practice, much less the ultimate test? And cognoscento to be the limiting example doesn't science take place most of the time of objectivity, explicitness, and demon- well short of the ultimate? strability.Surely here is bedrock for belief. Anyone would commend the vision of Yet standards of mathematical demon- science that Popper and his followers stration change. The last fifty years have have-of science as a self-correcting ex- been a disappointment to followers of Da- ploration verging on the other- vid Hilbert and his program to put mathe- wise so foreign to the analytic tradition matics on indubitable foundations. The in philosophy. For an economic scientist historian of mathematics, Morris Kline, to adopt an obdurate refusal to consider wrote recently that "it is now apparent objections and to resist offering hostages that the concept of a universally accepted, to evidence, though as common in mod- infallible body of reasoning-the majestic ernist as in nonmodernist circles, is not mathematics of 1800 and the pride of merely unscientific; it is cowardly. So man-is a grand illusion." Or again: much one can take from the idea of falsifi- There is no rigorousdefinition of rigor. A proof cation by evidence. The problem comes, is accepted if it obtains the endorsement of and the modernist preaching begins, with the leading specialists of the time and employs 492 Journal of Economic Literature, Vol. XXI (June 1983)

the principles that are fashionable at the mo- mathematician becomes totally invisible. This ment. But no standardis universallyacceptable is the problem of giving a philosophicalaccount today [1980, pp. 6, 315]. ... of preformal mathematics.. . , including an examination of how [it] relates to and is af- The recent flap over a computerized proof fected by formalization [1981, p. 344]. of the four-color proposition is one exam- They assert that "informal mathematics ple. The more fundamental example is is mathematics. Formalization is only an said to be Kurt G6del's proof fifty years abstract possibility which no one would ago that some true and statable proposi- want or be able actually to carry out" (p. tions in mathematics are unprovable. The 349). Real proofs "are established by 'con- point is controversial.John van Heijenoort sensus of the qualified"' and are "not writes that "the bearing of G6del's results checkable . . . by any mathematician not on epistemological remains un- privy to the gestalt, the mode of thought certain. . . . [T]hey should not be rashly in the particular field. . . . It may take called upon to establish the primacy of generations to detect an error" (p. 354). some act of intuition that would dispense They conclude: with formalization" (1967, p. 357). To be sure. But one need not dispense with for- The actual experience of all schools-and the malization and flee to an unexamined act actual daily experience of mathematicians- of intuition to think that formalization has shows that mathematical truth, like other kinds of truth, is fallible and corrigible.... It is rea- limits. sonable to propose a different task for mathe- Kline's opinions are somewhat loosely matical philosophy, not to seek indubitable expressed, and unpopular among mathe- truth, but to give an account of mathematical maticians. Apparently less so are those of knowledge as it really is-fallible, corrigible, Philip J. Davis and Reuben Hersh, whose tentative, and evolving, as is every other kind of human knowledge [p. 406]. recent book The Mathematical Experi- ence (1981) was described in the American Not much in this line has been done, Mathematical Monthly as "one of the though one astounding piece has shown masterpieces of our age." Davis and Hersh what can be: Imre Lakatos' Proofs and speak of the crisis of confidence in modern Refutations: The Logic of Mathematical mathematical philosophy, however, in Discovery gives an account for a theorem terms nearly identical with Kline's. In the in topology of the rhetoric of mathemat- work of The Ideal Mathematician "the ics. line between complete and incomplete It appears, then, that some deep prob- proof is always somewhat fuzzy, and often lems facing mathematics are problems of controversial"(p. 34; cf. p. 40). They quote rhetoric, problems in "the art of probing a living Ideal Mathematician, Solomon what men believe they ought to believe." Feferman, who writes "it is also clear that Similar points can be made about other the search for ultimate foundations via for- sciences, such as paleontology. The sud- mal systems has failed to arrive at any con- den proliferation of species at the begin- vincing conclusion" (p. 357). Without us- ning of the Cambrian period, one of the ing the word, Davis and Hersh argue that great puzzles in evolution, was explained what is required is a rhetoric of mathemat- by Steven Stanley in 1973 by supposing ics: the sudden arrival of forms of life that fed on other forms of life, single-celled her- The dominant style of Anglo-Americanphiloso- bivores, as it were, in a grassy sea. Their phy ... tends to perpetuate identification'of on dominant forms allowed the philosophy of mathematics with logic and grazing the the study of formal systems. From this stand- new forms to survive the point, a problem of principal concern to the from the previously dominant ones, which McCloskey: The Rhetoric of Economics 493 in turn resulted in new grazers. Stephen do not know why. Any economist believes Jay Gould remarks of the arguments of- more than his evidence of a suitably mod- fered in support of this brilliant and per- ernist and objective sort implies. A recent suasive theory that: poll of economists, for example, found that only three percent of those surveyed flatly . . .they do not correspond to the simplistic notions about scientific progress that are taught disagreed with the assertion that "tariffs in most high schools and advanced by most me- and import quotas reduce general eco- dia. Stanley does not invoke proof by new infor- nomic welfare." Only two percent dis- mation obtained from rigorousexperiment. His agreed with the assertion that "a ceiling second criterion is a methodological presump- on rents reduces the quantity and quality tion, the third a philosophical , the fourth an application of prior theory. . . . Sci- of housing available." Only eight percent ence, at its best, interposes human judgment disagreed with the assertion that "the tax- and ingenuity upon all its proceedings. It is, ing and spending of government has a sig- after all (although we sometimes forget it), nificant impact on the income of a partly practiced by human beings [Gould, 1977, p. idle economy" (J. R. Kearl, Clayne Pope, 125]. Gordon Whiting, and Larry Wimmer, One can even say the same of physics, 1979). You probably fall into the 97, 98, that favorite of outsiders seeking a pre- and 92 percent majorities. The evidence scription for real, objective, positive, for the assertions, however, is obscure. predictive science. The sequence Carnap- How do economists know these state- Popper-Lakatos-Kuhn-Feyerabendrepre- ments are true? Where did they acquire sents in the history and philosophy of such confidence? The usual answer is that physics a descent, accelerating recently, "theory tells us." But great social ques- from the frigid peaks of scientific absolut- tions are not answered by looking at a dia- ism to the sweet valleys of anarchic rheto- gram on a blackboard, because it is trivi- ric (see Popper, 1934, 1976; Lakatos, 1970; ally easy to draw a diagram that yields Kuhn, 1970; Feyerabend, 1975, 1978). If the opposite answer. The factual experi- economics should imitate other sciences, ence of the economy, certainly, has little imitate even the majesty of physics and to do with their confidence. No study has mathematics (there is, to be sure, consid- shown in ways that would satisfy a consis- erable doubt that it should), then it should tent modernist, for example, that high tar- officially open itself to a wider range of iffs in America during the 19th century, discourse. on balance, hurt Americans. Yet it is be- lieved that tariffs hurt then and now.6 No IV. The Unofficial Rhetoric Is Honorable study has shown that an inadvertent pol- But Unexamined icy of fiscal ease brought down War. Yet it is believed Econometric Rhetoric Is Too Narrow during the on all sides. Economists have not consid- But unofficiallyit does. The second atti- ered their rhetoric. tude towards discourse is that adopted in Everywhere in the literature of eco- actual scientific work in economics. It is nomics one is met with premises that are different from the official,modernist rhet- unargued, tricks of style masquerading as oric. What is alarming about the workaday reason ("it is evident that"), forms of evi- rhetoric is not its content but that it is dence that ignore the concerns of the au- unexamined, and that in consequence the official rhetoric pops up in mischievous 6Charles Peirce, the founder of pragmatism, re- lated in 1877 how he had been "entreated not to ways. Economists agree or disagree-their read a certain newspaper lest it might change my disagreements are exaggerated-but they opinion upon free-trade" (p. 101). 494 Journal of Economic Literature, Vol. XXI (June 1983) dience, and other symptoms of a lack of lines are fitted to someone else's collection self-consciousness in rhetoric. The lack is of facts. So far the official and workaday most evident in quarrels across research rhetoric correspond, and the one might paradigms. Some economists, (I am one) with be called a guide to the other. believe that peasants are rational. The Presently, however, they diverge. If the mass of modernist proofs, originating results of the fitting to the data are reason- largely from Chicago, that resistance to able, on grounds that are not themselves the "Green Revolution" or persistence in subject to examination, the article is sent scattering plots of land are rational leave off to a journal. If the results are unreason- many other economists cold. Some econo- able, the hypothesis is consigned to a do mists (I am one) believe that competition loop: the economic scientist returns to the is a robust characterization of the modern hypotheses or the specifications, altering American economy. The mass of mod- them until a publishable article emerges. ernist proofs, originating largely from Chi- The product may or may not have value, cago, that, for instance, advertising has but it does not acquire its value from its small effects on profits leaves the others adherence to the official rhetoric. It vio- cold. Why? Why do Chicago proofs leave lates the official rhetoric blatantly. Texas institutionalistsor NYU Austriansor But why shouldn't it? Even at the level Massachusetts Marxists or even Berkeley of tests of statistical significance the neoclassicists cold? The non-Chicago workaday rhetoric violates the philoso- economists, of course, believe they have pher's law. But so what? It is a cliche of modernist evidence of their own. But part cynicism in economics and related statisti- of the problem is that they also believe, cal fields to point out that a result signifi- without thinking about it much, that they cantly different from one that may have have evidence of a non-modernist sort: been gotten by chance does not have the stories of peasants and their lumpish char- significance it claims if the hypothesis has acter in the flesh; self-awareness of the been manipulated to fit the data. That force of advertising. A good part of the only significant results get published has disagreement is over evidence that is not long been a scandal among statistical pur- brought openly into the discussion, ists: they fear with some reason that at though it is used. the five percent level of significance some- Even in the most narrowly technical thing like five percent of the computer matters of scientific discussion economists runs will be successful. The scandal is not, have a shared set of convictions about however, the failure to achieve modernist what makes an argument strong, but a set standards of scientific purity. The scandal which they have not examined, which is the failure to articulate reasons why one they can communicate to graduate stu- might want to ignore them. dents only tacitly, and which contains It would be arrogant to suppose that many elements embarrassing to the offi- one knew better than thousands of intelli- cial rhetoric. A good example is the typical gent and honest economic scholars what procedure in econometrics. From eco- the proper form of argument was. The nomic theory, politics, and the workings Received View is arrogant in this way, lay- of the economist's psyche, all of which are ing down legislation for science on the ba- in the rhetorical sense unexamined, come sis of epistemological convictions held hypotheses about some bit of the econ- with vehemence inversely proportionalto omy. The hypotheses are then specified the amount of evidence that they work. as straight lines, linear models being those Better to look hard at what is in fact done. most easily manipulated. The straight In an important book that is an exception McCloskey: The Rhetoric of Economics 495

to the general neglect of rhetorical consid- "flimsy" ones, for instance that money erations in economics Edward Leamer causes inflation, shows the strength of asks what purpose the workaday proce- prior beliefs. (The beliefs can be reversed dures in econometrics may be serving without changing the example.) To leave (Leamer, 1978, esp. p. 17). Instead of com- the discussion at prior beliefs, however, paring them with a doctrine in the philos- perhaps formalizing them as prior proba- ophy of science he compares them with bility distributions, is to perpetuate the reasons that ought to persuade a reasona- fact-value split of modernism, leaving ble person, with what really warrants as- most of what matters in science to squeals sent, with, in short, economic rhetoric. As of pleasure or pain. What is required is Christopher Sims points out in a review, an examination of the workaday rhetoric "there is a myth that there are only two that leads to the prior beliefs. It is not of knowledge about the enough, as Thomas F. Cooley and Stephen world-'the' model, given to us by 'eco- F. LeRoy do in their recent, penetrating nomic theory,' without , and paper on "Identification and Estimation the parameters, about which we know of Money Demand" (1981) to merely nothing except what the data, via objec- stand appalled at the infection of econo- tively specified econometric methods, metric conclusions by prior beliefs. If tells us. . . . The sooner Leamer's cogent econometric argument does not persuade writings can lead us to abandon this myth, it is because the field of argument is too to recognize that nearly all applied work narrow, not because the impulse towards is shot through with applications of uncer- thoughtfulness and explicitness which it tain, subjective knowledge, and to make embodies is wrong. The arguments need the role of such knowledge more explicit to be broadened, not merely dismissed. and more effective, the better" (Sims, The Controversy Over Purchasing Power 1979, p. 567). Yes. The very title of Lea- Parity Is an Example of Unexamined mer's book is an outline of rhetoric in Rhetoric econometrics: Specification Searches: Ad Hoc Inference with Nonexperimental A good example of how the officialrhet- Data. oric-in the absence of an examination of Examples of the search abound. It is the workaday rhetoric-can lead a litera- common in a seminar in economics for ture in economics astray, especially in the speaker to present a statistical result, econometric matters, is the debate about apparently irrefutable by the rules of posi- . It is worth exam- tive economics, and to be met by a chorus ining in detail as a case study in unex- of "I can't believe it" or "It doesn't make amined rhetoric and the need to broaden sense." Milton Friedman's own Money it (Donald McCloskey and J. Richard Workshopat Chicago in the late 1960s and Zecher, 1982). The question is: is the inter- the early 1970s was a case in point. Put national economy more like the economy in statistical language, the rhetorical con- of the Midwest, in which Iowa City and text that creates such skepticism can be Madison and Champaign all face given called a priori beliefs and can be analyzed prices for goods; or is it more like the solar in Bayesian terms. It seldom is, but such system, in which each planet's economy a step would not be enough even if it were is properly thought of in isolation? If the taken. That the rhetorical community in Iowa City view is correct, then the prices economics might reject "solid"results, for of all goods will move together every- instance that oil prices appear in a regres- where, allowing for exchange rates. If the sion explaining inflation, and accept Martian view is correct they will move 496 Journal of Economic Literature, Vol. XXI (June 1983) differently. If the Iowa City view is cor- States (of steel or of goods-in-general, in rect, then all closed models of , levels or in differences) against the corre- whether Keynesian or monetarist or ra- sponding abroad, allowing for the tionally expecting, are wrong; if the Mar- exchange rate. If the slope coefficient is tian view is correct, then economists can 1.00 the hypothesis of purchasing power (as they do) go on testing macroeconomic parity is said to be confirmed; if not, not. faiths against American experience since Kravisand Lipsey perform such a test. Be- the War. ing good economists they are evidently The question of whether prices are made a little uncomfortable by the rheto- closely connected internationally, then, is ric involved. They admit that "Each ana- important. The official rhetoric does not lyst will have to decide in the light of his leave much doubt as to what is required purposes whether the purchasing power to answer it: collect facts on prices in, say, parity relationships fall close enough to the United States and Canada and . . . 1.00 to satisfy the theories" (p. 214). Pre- well . . . test the hypothesis (derived in cisely. In the next sentence, however, they orthodox fashion from a higher order hy- lose sight of the need for an explicit stan- pothesis, using objective data, looking dard if their argument is to be cogent: only at observable facts, controlling the "As a matter of general judgment we ex- experiment as much as possible, and so press our opinion that the results do not forth, according to the received view). A support the notion of a tightly integrated large number of economists have done international price structure." They do this. Half of them conclude that purchas- not say what a "general judgment" is or ing power parity works;the other half con- how one might recognize it. The purpose clude that it fails. A misleading but none- of an explicit economic rhetoric would be theless superb paper by Irving Kravis and to provide guidance. The guidance Kravis Robert Lipsey on the subject concludes and Lipsey provide for evaluating their that it fails, in terms that are worth repeat- general judgment is a footnote (p. 214) ing: reporting the general judgments of Hou- thakker, Haberler, and Johnson that devi- We think it unlikely that the high degree of ations from parity of anything under 10 national and international commodity arbi- trage that many versions of the monetarist [sic] to 20 percent are acceptable to the hy- theory of the contemplate pothesis. It happens, incidentally, that the is typical of the real world. This is not to deny bulk of the evidence offered by Kravisand that the price structuresof the advanced indus- Lipsey passes rather than fails such a test, trial countries are linked together, but it is to belying their conclusions. But accepting suggest that the links are loose rather than rigid [1978, p. 243, italics added]. or rejecting one unargued standard by comparing it with another unargued stan- Every italicized word involves a com- dard does not much advance the art of parison against some standard of what argument in economics. constitutes unlikelihood or highness or Kravis and Lipsey, to be fair, are unusu- typicality or being linked or looseness or ally sensitive to the case for having some rigidity. Yet here and elsewhere in the standard, more sensitive than are most tortured literature of purchasing power economists working the field. They return parity no standard is proposed. repeatedly to the question of a standard, The narrowest test of purchasing power though without resolving it. On page 204 parity, and the one that springs most read- they reject in one irrelevant sentence the ily to a mind trained in the official rheto- only standard proposed in the literature ric, is to regress the price in the United so far, the Genberg-Zecher criterion, de- McCloskey: The Rhetoric of Economics 497

scribed below. On pages 204-05 and on a standard error of .0000001 was signifi- 235 and again on 242 they draw a distinc- cantly different from unity in a significant tion between the statistical and the eco- meaning of significance. Such common nomic significance of their results. So fre- sense should be applied to findings of quently do they make the point that it slopes of .90 or 1.20. It is not.7 must be counted one of the major ones The irrelevance of the merely statistical in the paper. On page 205 they remark, standard of fit does not undermine only for example, that even small differences that half of the empirical literature that between domestic and export prices can finds purchasing power parity to be make a big difference to the incentive to wrong. Towards the end of a fine article export: "this is a case in which statistical favorable to purchasing power parity, significance [that is, a correlation of the writes: two prices near 1.0, which one mis- might There are several ways in which we might try takenly suppose to imply that they were to evaluate purchasing power parity as a the- insignificantly different] does not neces- ory. We can ask how much it explains [that sarily connote economic significance."Yet is, R-square];we can ask how large the devia- they do not turn the sword on themselves. tions from purchasingpower parity are in some No wonder: without a absolute sense; and we can ask whether the rhetoric of eco- deviations from purchasing power parity are nomic significance, and in the face of a in some sense systematic [1978, p. 405]. modernist rhetoric of statistical signifi- cance with the prestige of alleged science The defensive usage "in some absolute behind it, they are unaware they are sense" and "in some sense" betrays his wielding it. unease, which is in the event justified. The abuse of the word "significant"in There is no "absolute sense" in which a connection with statistical arguments in description is good or bad. The sense must economics is universal. Statistical signifi- be comparative to a standard, and the cance seems to give a standard by which standard must be argued. to judge whether a hypothesis is true or Similarly, Jacob Frenkel, an enthusiast false that is independent of any tiresome for purchasing power parity as such things consideration of how true a hypothesis go among economists but momentarily must be to be true enough. The point in bewitched by the ceremony of regression, the present case is that the "failure" of says that "if the market is efficient and if purchasing power parity in a regression the forward exchange rate is an unbiased of the usual type is not measured against forecast of the future spot exchange rate, a standard. How close does the slope have the constant [in a regression of the spot to be to the ideal of 1.00 to say that pur- rate today on the future rate for today chasing power parity succeeds? The litera- quoted yesterday] . . . should not differ ture is silent. The standard used is the ir- significantly from unity." In a footnote relevant one of statistical significance. A on the next page, speaking of the standard sample size of a million yielding a tight estimate that the slope was .9999, 7An example is J. D. Richardson'spaper, "Some "signifi- on Commodity Arbitrage and cantly" different from 1.00000, could be the " (1978). He regressesCanadian produced as evidence that purchasing on American prices multiplied by the exchange rate power parity had "failed," at for a number of industries and concludes: "It is nota- least if the ble that the 'law of one price' fails uniformly. The logic of the usual method were to be fol- hypothesis of perfect commodity arbitrage is re- lowed consistently. Common sense, pre- jected with 95 percent confidence for every com- sumably, would rescue the scholar from modity group" (p. 347, italics added). The question is, why in an imperfect world would it matter that asserting that an estimate of .9999 with perfect arbitrage is rejected? 498 Journal of Economic Literature, Vol. XXI (June 1983)

errors of the estimates for such an equa- we are not told whether it matters to the tiori in the 1920s, he argues that "while truth of purchasing power parity where these results indicate that markets were such limits of confidence are placed. efficient and that on average forward rates Silence on the matter is not confined were unbiased forecasts of future spot to the literature of purchasing power par- rates, the 2-8 percent errors were signifi- ity. Most texts on econometrics do not cant"(1978, pp. 175-76, italics added). He mention that the goodness or badness of evidently has forgotten his usage of "sig- a hypothesis is not ascertainableon merely nificant" in another signification. What he statistical grounds. Statisticians them- appears to mean is that he judges a 2-8 selves are more self-conscious, although percent error to be large in some unspeci- the transition from principle to practice fied economic sense, perhaps as offering is sometimes awkward. A practical diffi- significant profits for lucky guessers of the culty in the way of using the Neyman and correct spot rate. In any event, it is un- Pearson theory in pure form, A. F. Mood clear what his results imply about their and F. A. Graybill say, is that subject, purchasing power parity, because the loss function is not known at all or else it significance in statistics, however useful is not known accurately enough to warrant its it is as an input into economic significance, use. If the loss function is not known, it seems is not the same thing as economic signifi- that a decision function that in some sense min- cance. imizes the error probabilitieswill be a reason- The point is not that levels of signifi- able procedure [1963, p. 278]. cance are arbitrary. Of course they are. The phrase "in some sense" appears to The point is that it is not known whether be a marker of unexplored rhetoric in the the range picked out by the level of signifi- works of intellectually honest scholars. In cance affirms or denies the hypothesis. any event, the procedure they suggest Nor is the point that econometric tests are might be reasonable for a general statisti- to be disdained. Quite the contrary. The cian, who makes no claim to know what point is that the econometric tests have is a good or bad approximation to truth not followed their own rhetoric of hypoth- in fields outside statistics. It is not reason- esis testing. Nowhere in the literature of able for a specialist in tests of purchasing power parity does or . If the loss function there appear a loss function. We do not is not known it should be discovered. And know how much it will cost in policy that will entail a study of the question's wrecked or analysis misapplied or reputa- rhetoric. tion ruined if purchasing power parity is One standard of economic significance said to be true when by the measure of in questions of parity, for example, might the slope coefficient it is only, say, 85 per- be the degree to which the customary re- cent true. That is, the argument due to gressions between countries resembled Neyman and Pearson that undergirds similar regressions within a single country. modern econometrics has been set aside We agree, for purposes of argument, that here as elsewhere in favor of a merely the United States is to be treated as a sin- statistical standard, and an irrelevant one gle point in space, as one economy across related to sampling error at that. We are which distances are said not to matter for told how improbable it is that a slope coef- the purposes of thinking about inflation ficient of .90 came from a cen- or the balance of payments. Having done tered on 1.00 in view of the one kind of so we have a standard:is Canadaeconomi- error we claim we know about (unbiased cally speaking just as closely integrated sampling error with finite variance), but with the United States as is Californiawith McCloskey: The Rhetoric of Economics 499

Massachusetts? Is the Atlantic Economy ric, mere game playing in aid of ego grati- as closely integrated as the American fication; the serious business of Science Economy? The standardis called the Gen- will come on that happy day when the berg-Zecher criterion, after its inventors information theory of or the vul- (Hans Genberg, 1976; McCloskey and gar Marxist theory of the state is brought Zecher, 1976). It is not the only conceiv- to a critical test under the auspices of able one. The degree of market integra- naive falsificationism. tion in some golden age (1880-1913 per- Economists can do better if they will haps; or 1950-1970) might be a standard; look soberly at the varieties of their argu- the profits from arbitrage above normal ments. The varieties examined here can profits might be another standard;the de- only be crude preliminaries to a fuller gree to which an X percentage deviation study, a study that might dissect samples from purchasing power parity does or of economic argument, noting in the man- does not disturb some assertion about the ner of a literary or philosophical causes of inflation might be still another. exactly how the arguments sought to con- The point is to have standards of argu- vince the reader. It is not obvious a priori ment, to go beyond the inconclusive rhet- what the categories might be; in view of oric provided by the pseudo-scientific the methodological range of modern eco- ceremony of hypothesis-regression-test- nomics they doubtless would vary much publish in most of modern economics. from author to author. A good place to start might be the categories of classical V. The Rhetoric of Economics Is a rhetoric, Aristotle's divisions into inven- Literary Matter tion, arrangement, delivery, and style, for instance, with his paired sub-headings of Even a Modernist Uses, and Must Use, artificial (i.e., argumentative) and inartifi- Literary Devices cial (i.e., factual) proofs, and ex- The mere recognition that the official ample, and the like. A good place to con- rhetoric might be dubious, then, frees the tinue would be the procedures of modern reason to examine how economists really literary critics, bright people who make argue. Obscured by the official rhetoric their living thinking about the rhetoric of the workaday rhetoric has not received texts. the attention it deserves, and the knowl- The purpose would not be to make the edge of it is therefore contained only in author look foolish or to uncover fallacies seminar traditions, advice to assistant pro- for punishment by ridicule-fallacymon- fessors, referee reports (a promising pri- gering is evidence of a legislative attitude mary source for its study), and jokes. It towards method, and it is no surprise that is significant that George Stigler, a leader , confident of his ability of modernism in economics, chose to ex- to legislate for others in matters of method press his observations about the rhetoric as in education, prisons, and government, of economics in a brilliantly funny "The had compiled from his notes The Book of Conference Handbook" ("Introductory Fallacies (1824). David Hackett Fischer's Remark number E: 'I can be very sympa- book, Historians'Fallacies (1970), has this thetic with the author; until 2 years ago flaw: that it takes as fallacious what may I was thinking along similarlines' ") rather be merely probable and supporting argu- than as a serious study in one of the several ment. fields he has mastered, the history of eco- The purpose of literary scrutiny of eco- nomic thought (1977). The attitude of the nomic argument would be to see beyond Handbook is that rhetoric is mere rheto- the received view on its content. Two 500 Journal of Economic Literature, Vol. XXI (June 1983) pages (pp. 122-23) chosen literally at around 1983-those who have acquired random from that premier text of the re- a skill at partitioned matrices and eigen- ceived view, and a local maximum in eco- values should have charge of a great econ- nomic scholarship, Samuelson's Founda- omy. The argument is not absurd or a "fal- tions, will suffice for illustration: lacy" or "mere rhetoric." Virtuosity is (1) To begin with he gives a general some evidence of virtue.8 mathematical form from which detailed (2) There are six instances of appeal to results in can be ob- authority (C. E. V. Leser, Keynes, Hicks, tained by reading across a line. The impli- Aristotle, Knight, and Samuelson; appeal cation of the lack of elaboration of the to authorities is something of a Samuel- mathematics is that the details are trivial sonian specialty). Appeal to authority is (leading one to wonder why they are men- often reckoned as the worst kind of tioned at all). An "interesting"special case "mere" rhetoric. Yet it is a common and is left "as an exercise to the interested often legitimate argument, as here. No sci- reader," drawing on the rhetorical tradi- ence would advance without it, because tions of applied mathematics to direct the no scientist can redo every previous argu- reader's mind in the right directions. The ment. We stand on the shoulders of giants, mathematics is presented in an offhand and it is a perfectly legitimate and persua- way, with an assumption that we all can sive argument to point this out from time read off partitioned matrices at a glance, to time. inconsistent with the level of mathematics (3) There are several appeals to relax- in other passages. The air of easy mathe- ation of assumptions. The demand for matical mastery was important to the in- money is "really interesting . . . when un- fluence of the book, by contrast with the certainty . . . is admitted." Again, the im- embarrassed modesty with which British plicit assumption in Hicks that money writers at the time (Hicks most notably) bears no interest is relaxed, unhitching the pushed mathematics off into appendices. from the zero return on Samuelson's skill at mathematics in the money. Relaxation of assumptions is the eyes of his readers, an impression nur- literature generating function of modern tured at every turn, is itself an important economics. In the absence of quantitative and persuasive argument. He presents evidence on the importance of the as- himself as an authority, with good reason. sumption relaxed it is no modernist evi- That the mathematics is so often pointless, dence at all. Samuelson is careful to stick as here, is beside the point. Being able to the subjunctive mood of theory (money to do such a difficult thing (so it would "would pass out of use"), but no doubt have seemed to the typical economist wants his strictures on a theory of the in- reading in 1947) is warrant of expertise. terest rate based merely on liquidity pref- The argument is similar in force to that erence (that is, on risk) to be taken seri- of a classical education conspicuously dis- ously as comments on the actual world. played. To read Latin like one's mother They are, surely, but not on the operation- tongue and Greek like one's aunt's tongue alist grounds he articulates when preach- is extremely difficult, requiring applica- ing methodology. tion well beyond the ordinary;therefore- (4) There are several appeals to hypo- or so it seemed to Englishmen around 1900-men who have acquired such a skill 8 The limiting case is spelling. Most college teach- ers will agree that those who do not know how to should have charge of a great empire. spell "consensus" lose some of their authority to Likewise-or so it seemed to economists speak on it. McCloskey: The Rhetoric of Economics 501

thetical toy economies, constrained to one they are easily misclassified. A common or two sectors, from which practical re- argument in economics, for example, is sults are derived. This has since Ricardo one from verbal suggestiveness. The prop- been among the commonest forms of eco- osition that "the economy is basically com- nomic argument, the Ricardian vice. It is petitive" may well be simply an invitation no vice if done reasonably. "It would be to look at it this way, on the assurance quite possible to have an economy in that to do so will be illuminating. In the which money did not exist, and in which same way a psychologist might say "we there was still a substantial rate of inter- are all neurotic"-it does not mean that est." Yes, of course. 95 percent of a randomly selected sample (5) There is, finally, one explicit appeal of us will exhibit compulsive handwash- to analogy, which is said to be "not ... ing; rather, it is merely a recommendation superficial."Analogy, as will be shown in that we focus attention on the neurotic detail in a moment, pervades economic ingredient "in us all" (Passmore, 1966, p. thinking, even when it is not openly 438). To misunderstand the expression as analogical: transaction "friction," yield a properly modernist hypothesis would be "spread," securities "circulating," money to invite much useless testing. The case "withering away" are inexplicit examples is similar to MV=PT understood as an here from one paragraph of live or only identity. The equation is the same term- half-dead metaphors. Yet analogy and for-term as the equation of state of an ideal metaphor, like most of the other pieces gas, and has the same status as an irrefuta- of Samuelson's rhetoric, have no standing ble but useful notion in chemistry as it in the official canon. has in economics. The identity can be ar- gued against, but not on grounds of "fail- Most of the Devices Are Only Dimly a The arguments against it will Recognized ing test." deny its capacity to illuminate, not its The range of persuasive discourse in modernist truth. economics is wide, ignored in precept Another common argument in econom- while potent in practice. At the broadest ics with no status in the official rhetoric level it is worth noting that the practice is philosophical consistency: "If you as- of economic debate often takes the form sume the firm knows its own cost curve of legal reasoning, for, as Booth put it, you might as well assume it knows its pro- "the processes developed in the law are duction function, too: it is no more dubi- codifications of reasonable processes that ous that it knows one than the other." The we follow in every part of our lives, argument, usually inexplicit though sig- even the scientific" (1974, p. 157). Econ- nalled by such a phrase as "it is natural omists would do well to study jurispru- to assume," is in fact characteristic of dence, then, with some other aim than philosophical discourse (Passmore, 1970). subordinating it to economic theory. For It is analogous to symmetry as a criterion instance, economists, like jurists, argue by of plausibility, which appears in many example, by what Edward Levi calls "the forms and forums. A labor economist tells controlling similarity between the pres- a seminar about compensating differen- ent and prior case" (1967, p. 7). tials for the risk of unemployment, refer- The details of the pleading of cases at ring only to the functions of the economic law have little to do with the workers. An auditor remarks that the official scientific method. Without self- value of unemployment on the demand consciousness about workaday rhetoric side (that is, to the firm) is not included. 502 Journal of Economic Literature, Vol. XXI (June 1983)

The remark is felt to be powerful, and a any part of the range. They can deliver long discussion ensues of how the demand summary grunts of belief or disbelief but side might alter the conclusions. The argu- find it difficult to articulate their reasons ment from the-other-side-is-empty is per- in a disciplined way. suasive in economics, but economists are unaware of how persuasive it is. VI. Economics Is Heavily Metaphorical Likewise (and here we reach the border of self-consciousnessin rhetoric), "ad hoc- Models Are Metaphors cery" is universally condemned by semi- nar audiences. An economist will cheer- The most important example of eco- fully accept a poor R2 and terrible and nomic rhetoric, however, falls well outside understated standard errors if only she the border of self-consciousness. It is the "has a theory" for the inclusion of such- language economists use, and in particular and-such a variable in the regressions. its metaphors. To say that markets can be "Having a theory" is not so open and shut represented by as it might seem, depending for instance "curves" is no less a metaphor than to say on what reasoning is prestigious at the mo- that the west wind is "the breath of au- ment. Anyone who threw accumulated tumn's being." A more obvious example past into an equation explaining is "," the very name being productivity change before 1962 would a metaphor. It is obviously useful to have have been accused of ad hoccery. But in one's head the notion that the arms race after Arrow's essay on "The Economics is a two-person, negative-sum cooperative of Learning By Doing" (which as it hap- "game." Its persuasiveness is instantly ob- pened had little connection with max- vious, as are some of its limitations. Each imizing behavior or other higher order step in economic reasoning, even the rea- hypotheses in economics), there was sud- soning of the officialrhetoric, is metaphor. denly a warrant for doing it. The world is said to be "like" a complex An example of the rhetoric of econom- model, and its measurements are said to ics which falls well within the border of be like the easily measured proxy variable self-consciousness is simulation. Econo- to hand. The complex model is said to be mists will commonly make an argument like a simpler model for actual thinking, for the importance of this or that variable which is in turn like an even simpler by showing its potency in a model with model for calculation. For purposes of per- back-of-the-envelope estimates of the pa- suading doubters the model is said to be rameters. Common though it is, few books like a toy model that can be manipulated or articles are devoted to its explication quickly inside the doubter's head while (but, see Richard Zeckhauser and Edith listening to the seminar. John Gardner Stokey, 1978). It would be as though stu- wrote: dents learned econometrics entirely by studying examples of it-no bad way to There is a game-in the 1950s it used to be learn, but not self-conscious in grounding played by the members of the Iowa Writers' Workshop-called 'Smoke.' The player who is the arguments. What is legitimate simula- 'it' [thinks of] some famous person .. . and tion? Between A. C. Harberger's modest then each of the other players in turn asks one little triangles of distortion and Jeffrey question. .. such as 'What kind of weather Williamson's immense multiequation are you?'. . . Marlon Brando, if weather, would models of or econ- be sultry and uncertain.... To understand the American Japanese that Marlon Brando is a certain kind of weather omies since 1870 is a broad range. Econo- is to discover something (though something mists have no vocabulary for criticizing neither useful nor demonstrable) and in McCloskey: The Rhetoric of Economics 503

the same instant to communicate something mind-stretching fancy; "depression" was [Gardner, 1978, pp. 118-19]. depressing; "equilibrium" compared an to an apple in a bowl, a set- On the contrary, in economics the com- economy idea; "competition" once induced parable discovery is useful and by re- tling thoughts of horseraces; money's "veloc- course to rhetorical standards demonstr- ity" thoughts of swirling bits of paper. able. Much of the vocabulary of economics con- Metaphors in Economics Are Not sists of dead metaphors taken from non- Ornamental economic spheres. Comparing noneconomic with eco- Metaphor, though, is commonly viewed nomic matters is another sort of novelty, as mere ornament. From Aristotle until apparent in the imperialism of the new the 1930s even literary critics viewed it economics of law, history, politics, crime, this way, as an amusing comparison able and the rest, and most apparent in the to affect the emotions but inessential for work of that Kipling of the economic em- thought. "Men are beasts": if we cared pire, Gary Becker. Among the least bi- to be flat-footed about it, the notion was, zarre of his many metaphors, for instance, we could say in what literal way we is that children are durable goods. The thought them beastly, removing the orna- philosopher Max Black pointed out that ment to reveal the core of plain meaning "a memorable metaphor has the power underneath. The notion was in 1958 com- to bring two separate domains into cogni- mon in philosophy, too: tive and emotional relation by using lan- guage directly appropriate to the one as With the decline of metaphysics, philosophers a lens for seeing the other" (1962, p. 236). have grown less and less concerned about God- liness and more and more obsessed with clean- So here: the subject (a child) is viewed liness, aspiringto ever higher levels of linguistic through the lens of the modifier (a durable hygiene. In consequence, there has been a ten- good). A beginning at literal translation dency for metaphors to fall into disfavour, the would say, "A child is costly to acquire common opinion being that they are a frequent initially, lasts for a long time, gives flows source of infection [H. J. N. Horsburgh, 1958, p. 231]. of pleasure during that time, is expensive to maintain and repair, has an imperfect Such suspicion toward metaphor is widely second-hand market.... Likewise, a du- recognized by now to be unnecessary, rable good, such as a refrigerator.. even harmful. That the very idea of "re- That the list of similarities could be ex- moving" an "ornament" to "reveal" a tended further and further, gradually re- "plain" meaning is itself a metaphor sug- vealing the differences as well-"children, gests why. Perhaps thinking is metaphori- like durable goods, are not objects of affec- cal. Perhaps to remove metaphor is to re- tion and concern"; "children, like durable move thought. The operation on the goods, do not have their own opinions"- metaphoric growth would in this case be is one reason that, as Black says, "meta- worse than the disease. phorical thought is a distinctive mode of The question is whether economic achieving insight, not to be construed as thought is metaphorical in some nonorna- an ornamental substitute for plain mental sense. The more obvious meta- thought" (p. 237). The literal translation phors in economics are those used to con- of an important metaphor is never fin- vey novel thoughts, one sort of novelty ished. In this respect and in others an im- being to compare economic with noneco- portant metaphor in economics has the nomic matters. "Elasticity" was once a quality admired in a successful scientific 504 Journal of Economic Literature, Vol. XXI (June 1983)

theory, a capacity to astonish us with im- both fields was improved, labor economics plications yet unseen.9 by recognizing that skills, for all their in- But it is not merely the pregnant quality tangibility, arise from from of economic metaphors that makes them ; capital theory by recogniz- important for economic thinking. The lit- ing that skills, for all their lack of capitali- erary critic I. A. Richards was among the zation, compete with other first to make the point, in 1936, that meta- for a claim to abstention. Notice by con- phor is "two thoughts of different things trast that because economists are experts active together, . . . whose meaning is a only in durable goods and have few (or resultant of their interaction" (Richards, at any rate conventional) thoughts about 1936, p. 93, my italics; Black, 1962, p. 46; children, the metaphor that children are Owen Barfield, 1947, p. 54). A metaphor durable goods has so to speak only one is not merely a verbal trick, Richards con- direction of flow. The gains from the trade tinues, but "a borrowing between and were earned mostly by the theory of chil- intercourse of thoughts, a transaction be- dren (fertility, nuptiality, inheritance), tween contexts" (p. 94, his italics). Econo- gaining from the theory of durable goods, mists will have no trouble seeing the point not the other way around. of his economic metaphor, one of mutu- ally advantageous exchange. The opposite Economic Metaphors Constitute a notion, that ideas and their words are in- of Economics variant lumps unaltered by combination, What is successful in economic meta- like bricks (Richards,p. 97), is analogous phor is what is successful in poetry, and to believing that an economy is a mere is analyzable in similar terms. Concerning aggregation of Robinson Crusoes. But the the best metaphors in the best poetry,- point of economics since Smith has been comparing thee to a summer's day or com- that an island-fullof Crusoes trading is dif- paring A to B, argued Owen Barfield, ferent from and often better off than the mere aggregation. We feel that B, which is actually said, ought to be necessary, even inevitable in some way. Another of Becker's favorite metaphors, It ought to be in some sense the best, if not "," illustrates how two sets the only way, of expressing A satisfactorily.The of ideas, in this case both drawn from in- mind should dwell on it as well as on A and side economics, can thus mutually illumi- thus the two should be somehow inevitably nate each other by exchanging connota- fused together into one simple meaning [Bar- field, 1947, p. 54]. tions. In the phrase "human capital" the field in economics treating human skills If the modifier B (a summer's day, a refrig- was at a stroke unified with the field treat- erator, a piece of capital) were trite-in ing investment in machines. Thought in these cases it is not, although in the poem Shakespeare was more self-critical of his 9 A good metaphor depends on the ability of its simile than economists usually are of audience to suppress incongruities, or to wish to. theirs-it would become as it were de- Booth gives the example of: tached from A, a mechanical and unillu- 'All the world's a stage' . . . [The reader must make a minating correspondence. If essential, it choice only if the incongruities-failures of fit-come too soon]. Usually they arrive late and without much fuses with A, to become a master meta- strength.... [W]e have no difficultyruling from our phor of the science, the idea of "human attention, in the life-stage metaphor, the selling of tick- capital," the idea of "equilibrium," the ets, fire insurance laws, the necessity for footlights [1961, p. 22f]. idea of "entry and exit," the idea of "com- petition." The metaphor, quoth the poet, The appreciation of "human capital" requires the same suspension of disbelief. is the "consummation of identity." McCloskey: The Rhetoric of Economics 505

Extent ever spinning "parables" or telling "sto- Explicit- ries." The word "story" has in fact come ness Short Long to have a technical meaning in mathemat- ical economics, though usually spoken in Explicit simile tiresome caution seminars rather than written in papers. (The firm behaves (repeat the simile) as if it were one It means an extended example of the eco- mind maximiz- nomic reasoning underlying the mathe- ing its dis- matics, often a simplified version of the counted value.) situation in the real world that the mathe- Middling metaphor allegory matics is meant to characterize. It is an (human capital) (economics of edu- cation using hu- allegory, shading into extended symbol- man capital) ism. The literary theories of narrative Implicit symbol a symbol system: a could make economists self-conscious (income) mathematics; a about what use the story serves. Here the (demand curve) theory (Keynes- story is the modifier, the mathematics the ian theory of in- come determi- subject. A tale of market days, traders with nation) (supply bins of shmoos, and customers with costs and demand of travel between bins illuminates a fixed analysis) point theorem. Figure 1. AnalogicalThinking Has Two Dimensions Even Mathematical Reasoning Is (And Economic Cases in Point) Metaphorical Few would deny, then, that economists The critical question is whether the op- frequently use figurative language. Much posite trick, modifying human behavior of the pitiful humor available in a science with mathematics, is also metaphorical. If devoted to calculations of profit and loss it were not, one might acknowledge the comes from talking about "islands"in the metaphorical element in verbal econom- labor market or "putty-clay"in the capital ics about the "entrepreneur,"for instance, market or "lemons" in the commodity or more plainly of the "," market. The more austere the subject the yet argue that the linguistic hygiene of more fanciful the language. We have mathematics leaves behind such fancies. "turnpikes"and "golden rules" in growth This indeed was the belief of the advanced theory, for instance, and long disquisitions thinkers of the 1920s and 1930s who in- on what to do with the "auctioneer" in spired the now-received view in economic general equilibrium theory. A literary method. Most economists subscribe to the man with advanced training in mathemat- belief without doubt or comment or ics and statistics stumbling into Econome- thought. When engaging in verbal eco- trica would be astonished at the meta- nomics we are more or less loose, it is said, phors surrounding him, lost in a land of taking literary license with our "story"; allegory. but when we do mathematics we put away Allegory is merely long-winded meta- childish things. phor, and all such figures are analogies. But mathematical theorizing in eco- Analogies can be arrayed in terms of ex- nomics is metaphorical, and literary. Con- plicitness, with simile ("as if") the most sider, for example, a relatively simple explicit and symbol ("the demand curve") case, the theory of production functions. the least explicit; and they can be arrayed Its vocabulary is intrinsically metaphori- by extent. cal. "Aggregate capital" involves an anal- Economists, especially theorists, are for- ogy of "capital" (itself analogical) with 506 Journal of Economic Literature, Vol. XXI (June 1983) something-sand, bricks, shmoos-that Even when the metaphors of one's eco- can be "added" in a meaningful way; so nomics appear to stay well and truly dead does "aggregate labor," with the addi- there is no escape from literary questions. tional peculiarity that the thing added is The literary man C. S. Lewis pointed out no thing, but hours of conscientious atten- in 1939 that any talk beyond the level tiveness; the very idea of a "production of the-cow-standing-here-is-in-fact-purple, function" involves the astonishing analogy any talk of "causes, relations, of mental of the subject, the fabrication of things, states or acts . . . [is] incurably metaphori- about which it is appropriate to think in cal" (1962, p. 47). For such talk he enunci- terms of ingenuity, discipline, and plan- ated what may be called Screwtape's ning, with the modifier, a mathematical Theorem on Metaphor, the first corollary function, about which it is appropriate to of which is that the escape from verbal think in terms of height, shape, and single into mathematical metaphor is not an es- valuedness. cape: The metaphorical content of these ideas when a man claims to think independently of was alive to its inventors in the 19th cen- the buried metaphor in one of his words, his tury. It is largely dead to 20th-century claim may .. . [be] allowed only in so far as economists, but deadness does not elimi- he could really supply the place of that buried nate the metaphorical element. The meta- metaphor.... [TMhisnew apprehension will phor got out of its coffin in an alarming usually turn out to be itself metaphorical [p. 46]. fashion in the Debate of the Two Cam- bridges in the 1960s. The debate is testa- If economists forget and then stoutly deny ment, which could be 'multiplied, to the that the production function is a meta- importance of metaphorical questions to phor, yet continue talking about it, the economics. The very violence of the com- result is mere verbiage. The word "pro- bat suggests that it was about something duction function" will be used in ways beyond mathematics or fact. The com- satisfying grammatical rules, but will not batants hurled mathematical reasoning signify anything. The charge of meaning- and institutional facts at each other, but lessness applied so freely by modernists the important questions were those one to forms of argument they do not under- would ask of a metaphor-is it illuminat- stand or like sticks in this way to them- ing, is it satisfying, is it apt? How do you selves. Lewis' second corollary is that "the know? How does it compare with other meaning in any given composition is in economic poetry? After some tactical re- inverse ratio to the author's belief in his treats by Cambridge, Massachusetts on own literalness" (p. 27). An economist points of ultimate metaphysics irrelevant speaking "literally" about the demand to these important questions, mutual ex- curve, the national income, or the stability haustion set in, without decision. The rea- of the economy is engaging in mere syn- son there was no decision was that the tax. Lewis cuts close to the bone here, important questions were literary, not though sparing himself from the car- mathematical or statistical.The continued nage: of the idea of an vitality aggregate produc- The percentage of mere syntax masquerading tion function in the face of mathematical as meaning may vary from something like 100 proofs of its impossibility and the equal percent in political writers,journalists, psychol- vitality of the idea of aggregate economics ogists, and economists, to something like forty as practiced in parts of Cambridge, En- percent in the writers of children'sstories.. in The mathematician, who seldom forgets that gland the face of statistical proofs of his symbols are symbolic, may often rise for its impracticality would otherwise be a short stretches to ninety percent of meaning great mystery. and ten of verbiage [p. 49]. McCloskey: The Rhetoric of Economics 507

If economists are not comparing a social be argued cautiously. Self-consciousness fact to a one-to-one mapping, thus bring- about metaphor in economics would be ing two separate domains into cognitive an improvement on many counts. Most and emotional relation, they are not think- obviously, unexamined metaphor is a sub- ing: stitute for thinking-which is a recom- I've never slapped a curved demand; mendation to examine the metaphors, not I never hope to slap one. to attempt the impossible by banishing But this thing I can tell you now: them.10 Richard Whately, D.D., Arch- I'd rather slap than map one. bishop of , publicist for for other pieces of classical political Literary Thinking Reunifies the Two as work Cultures economy, and author of the standard in the 19th century on The Elements of Metaphor, then, is essential to economic Rhetoric, drew attention to the metaphor thinking, even to economic thinking of of a state being like an individual, and the most formal kind. One may still doubt, therefore benefiting like an individual though, whether the fact matters. For it from free trade. But he devoted some at- is possible for rhetoricians as well as unre- tention, not all of it ironic, to the question constructed modernists to commit the of the aptness of the figure: Philosophizing Sin, to bring high-brow To this is it replied, that there is a great differ- considerations of the ultimate into discus- ence between a Nation and an Individual.And sions about how to fix a flat tire. Pushkin's so there is, in many circumstances. . . [he enu- poetry may be ultimately untranslatable, merates them, mentioning for instance the un- in view of the difference in language, to limited duration of a Nation] and, moreover, the transactions of each man, as far as he is be sure, but also the difference in situation left free, are regulated by the very person who between a Pushkin in Russia in the early is to be a gainer or loser by each,-the individ- 19th century and a bilingual translator in ual himself; who, though his vigilance is sharp- New York in the late 20th century. Be- ened by interest, and his judgment by exercise cause he was a different man speaking to in his own department, may chance to be a a different even bril- man of confined education, possessedof no gen- audience Nabokov's eral principles, and not pretending to be versed liant translation,in the words of the econ- in philosophical theories; whereas the affairs omist and litterateur, Alexander Ger- of a State are regulated by a Congress, Cham- schenkron, "can and indeed should be ber of Deputies, etc., consisting perhapsof men studied but . . . cannot be read" (in of extensive reading and speculative minds Steiner, 1975, p. 315). Our intrinsic loneli- [1894, p. 63]. ness will make some nuance dark. Yet The case for intervention cannot be put crude translation,even by machine, is use- better. And the metaphor is here an occa- ful for the workaday purposes of inform- sion for and instrument of thought, not ing the Central Intelligence Agency (for a substitute. instance, "out of sight, out of mind" = Metaphors,further, evoke attitudes that "blind madman"). Likewise, the intrinsic are better kept in the open and under metaphors of language may make it ulti- the control of reasoning. This is plain in mately impossible to communicate plain the ideological metaphors popular with meaning without flourishes-the flour- parties: the invisible hand is so very dis- ishes are the meaning. But the economist crete, so soothing, that we might be in- may be able to get along without full clined to accept its touch without protest; awareness of his meaning for the worka- day purposes of advising the Central Intel- 10 An example of a naive attack on economic meta- phors, and of a failure to realize that economic the- ligence Agency. ory is itself armed with metaphor, is the first page So it might be argued. But it should of McCloskey (1981). 508 Journal of Economic Literature, Vol. XXI (June 1983) the contradictions of are so in contexts where the economist is simul- very portentous, so scientifically precise, taneously fitting 50 other equations. The that we might be inclined to accept their response that the metaphor will be tested existence without inquiry. But it is true eventually by the facts is a stirring prom- even of metaphors of the middling sort. ise, but seldom fulfilled (see again Leamer, The metaphors of economics convey the throughout). A better response would be authority of Science, and often convey, that we like the metaphor of, say, the self- too, its claims to ethical neutrality. It is ishly economic person as calculating ma- no use complaining that we didn't mean chine on grounds of its prominence in ear- to introduce moral premises. We do. lier economic poetry plainly successful or "Marginal productivity" is a fine, round on grounds of its greater congruence with phrase, a precise mathematical metaphor introspection than alternative metaphors that encapsulates a most powerful piece (of people as religious dervishes, say, or of social description. Yet it brings with it as sober citizens). In The New Rhetoric: an air of having solved the moral problem A Treatise on Argumentation (1967), of distribution facing a society in which Chaim Perelman and L. Olbrechts-Tyteca people cooperate to produce things to- note that "acceptance of an analogy . . . gether instead of producing things alone. is often equivalent to a judgment as to It is irritating that it carries this message, the importance of the characteristics that because it may be far from the purpose the analogy brings to the fore" (p. 390). of the economist who uses it to show ap- What is remarkable about this unremark- proval for the distribution arising from able assertion is that it occurs in a discus- competition. It is better, though, to admit sion of purely literary matters, yet fits so that metaphors in economics can contain easily the matters of economic science. such a political message than to use the This is in the end the significance of jargon innocent of its potential. metaphors and of the other rhetorical ma- A metaphor, finally, selects certain re- chinery of argument in economics: econo- spects in which the subject is to be com- mists and other scientists are less separate pared with the modifier; in particular, it from the concerns of civilization than leaves out the other respects. Max Black, many think. Their modes of argument and speaking of the metaphor "men are the sources of their conviction-for in- wolves," notes that "any human traits that stance, their uses of metaphor-are not can without undue strain be talked about very different from Cicero's speeches or in 'wolf-language'will be rendered promi- Hardy's novels. This is a good thing. As nent, and any that cannot will be pushed Black wrote, discussing "archetypes" as into the background"(1962, p. 41). Econo- extended metaphors in science: "When mists will recognize this as the source of the understanding of scientific models and the annoying complaints from non-mathe- archetypes comes to be regarded as a rep- matical economists that mathematics utable part of scientific culture, the gap "leaves out" some feature of the truth or between the sciences and the humanities from non-economists that economics will have been partly filled" (p. 243). "leaves out" some feature of the truth. Such complaints are often trite and ill- VII. Be Not Afraid formed. The usual responses to them, The Alternative to Modernism Is Not are less so. The however, hardly response Irrationalism that the metaphor leaves out things in or- der to simplify the story temporarily is It will be apparent by now that the ob- disingenuous, occurring as it often does jectivity of economics is overstated and, McCloskey: The Rhetoric of Economics 509 what is more important, overrated. Preg- for wider rationality. The tactic is an old nant economic knowledge depends little one. RichardRorty notes that "the charges on, as Michael Polanyi put it, "a scientific of '' and 'irrationalism'once lev- rationalism that would permit us to be- eled against Dewey [were] merely the lieve only explicit statements based on mindless defensive reflexes of the philo- tangible data and derived from these by sophical tradition which he attacked" a formal inference, open to repeated test- (1979, p. 13; Rorty, 1982a, Ch. 9). The posi- ing" (1966, p. 62). A rhetoric of economics tion taken by the opponents of Dewey, makes plain what most economists know Polanyi, Kuhn, and the rest is "if the anyway about the richness and complexity choice is between Science and irrational- of economic argument but will not state ity, I'm for Science." But that's not the openly and will not examine explicitly. choice. The invitation to rhetoric, however, is The Barbarians Are Not at the Gates not an invitation to irrationality in argu- ment. Quite the contrary. It is an invita- Yet still the doubt remains. If we aban- tion to leave the irrationality of an artifi- don the notion that econometrics is by it- cially narrowed range of arguments and self a method of science in economics, if to move to the rationality of arguing like we admit that our arguments require human beings. It brings out into the open comparative standards, if we agree that the arguing that economists do anyway- personal knowledge of various sorts plays in the dark, for they must do it somewhere a part in economic knowledge, if we look and the various official leave at economic argument with a literary eye, them benighted. will we not be abandoning science to its The charge of irrationalismcomes easily enemies? Will not scientific questions to the lips of methodological authoritari- come to be decided by politics or whim? ans. The notion is that reasoning outside Is the routine of Scientific Method not a the constricted of modern- wall against irrational and authoritarian ism is no reasoning at all. Mark Blaug, for threats to inquiry? Are not the barbarians instance, charges that Paul Feyerabend's at the gates? book Against Method "amounts to replac- The fear is a surprisinglyold and persis- ing the philosophy of science by the phi- tent one. In classical times it was part of losophy of flower power" (1980, p. 44). the debate between philosophy and rheto- Feyerabend commonly attracts such dis- ric, evident in the unsympathetic way in missive remarks by his flamboyance. But which the sophists are portrayed in 's Stephen Toulmin and Michael Polanyi are dialogues. Cicero viewed himself as bring- nothing if not sweetly reasonable; Blaug ing the two together, disciplining rheto- lumps them with Feyerabend and attacks ric's tendency to become empty advocacy the Feyerabend-flavored whole. On a and trope on the one hand and disciplin- higher level of philosophical sophistica- ing philosophy's tendency to become use- tion Imre Lakatos' Methodology of Scien- less and inhuman speculation on the tific Research Programmes(1978, from ar- other. The classical problem was that rhet- ticles published from 1963 to 1976) oric was a powerful device easily misused repeatedly tars Polanyi, Kuhn, and Feyer- for evil ends, the atomic power of the clas- abend with "irrationalism" (e.g., Vol. 1, sical world, and like it the subject of wor- pp. 9nl, 76n6, 91nl, 130 and 130n3), em- rying about its proliferation. The solution phasizing their sometimes aggressively ex- was to insist that the orator be good as pressed case against rigid rationalism and well as clever: Cato defined him as "vir ignoring their moderately expressed case bonus dicendi peritus," the good man 510 Journal of Economic Literature, Vol. XXI (June 1983) skilled at speaking, a Ciceronian ideal as it must be admitted, are not so bad an well. Quintilian, a century and a half after argument for authoritarianism,at least un- Cicero, said that "he who would be an til one looks more closely at the results orator must not only appear to be a good of authoritarianism.Surely, though, the al- man, but cannot be an orator unless he ternative to blindered rules of modernism is a good man" (Institutio XII, 1, 3). The is not an irrational mob but a body of en- classical problem looks quaint to moderns, lightened scholars,,perhaps more enlight- who know well that regressions, radios, ened when freed to make arguments that computers, experiments, or any of the actually bear on the questions at issue. now canonized methods of persuasion can ThereIs No Good Reason to Wish to Make be and have been used as methods of de- As Against Plausible ceit. There is nothing about anaphora, "Scientific" Statements chiasmus, metonymy, or other pieces of classical rhetoric that make them more The other main objection to an openly subject to evil misuse than the modern rhetorical economics is not so pessimistic. methods. One can only note with regret It is the sunny view that scientific knowl- that the Greeks and Romans were more edge of a modernist sort may be hard to sensitive to the possibility, and less hypno- achieve, even impossible, but all will be tized by the claims of method to moral well on earth and in heaven if we strive neutrality. in our poor way to reach it. We should The 20th century's attachment to limit- have a standard of Truth beyond persua- ing rules of inquiry solves a German prob- sive rhetoric to which to aspire. In Figure lem. In the German Empire and Reich 2 all possible propositions about the world it was of course necessary to propound a are divided into objective and subjective, split of fact from values in the social sci- positive and normative, scientific and hu- ences if anything was to be accomplished manistic, hard and soft. The modernist free of political interference. And German supposes that the world comes divided speculative philosophy, one hears it said, nicely along such lines. warranted a logical positivist cure. The German habits, however, have spilled over into a quite different world. It is said scientific humanistic that if we are to avoid dread anarchy we fact opinion objective subjective cannot trust each scientist to be his own positive normative methodologist. We must legislate a uni- vigorous sloppy form though narrowing method to keep precise vague scholars from resorting to figurative and things words literal murder in aid of their ideas. We cognition intuition hard soft ourselves could be trusted with method- ological freedom, of course, but the others Figure 2. The Task of Science Is to Move the cannot. The argument is a strange and au- Line thoritarian one, uncomfortably similar to the argument of, say, the Polish authori- According to the modernist methodolo- ties against Solidarity or of the Chilean gist the scientist's job is not to decide authorities against free politics. It is odd whether propositions are useful for under- to hear intellectuals making it. Perhaps standing and changing the world but to their low opinion of the free play of ideas classify them into one or the other half, comes from experiences in the faculty sen- scientific or nonscientific, and to bring as ate: the results of academic , many as possible into the scientific por- McCloskey: The Rhetoric of Economics 511 tion. But why? Whole teams of philosoph- conceivable but practically impossible test ical surveyors have sweated long over the takes over the prestige of the real test, placing of the demarcation line between but free of its labor. Such a step needs scientific and other propositions, worrying to be challenged. It is identical to the one for instance about whether astrology can involved in equating as morally similar the be demarcated from astronomy;it was the actual compensation of those hurt during chief activity of the positivist movement a Pareto optimal move with a hypothetical for a century. It is not clear why anyone compensation not actually paid, as in the troubled to do so. People are persuaded Hicks-Kaldortest; and it is identically du- of things in many ways, as has been shown bious. A properly identified econometric for economic persuasion. It is not clear measurement of the out-of-sample prop- why they should labor at drawing lines erties of macroeconomic policy is "opera- on mental maps between one way and an- tional," that is, conceivable, but for all the other. scientific prestige the conceivability lends The modernists have long dealt with the to talk about it, there are grave doubts embarrassmentthat metaphor, case study, whether it is practically possible. While upbringing, authority, introspection, sim- economists are waiting for the ultimate plicity, symmetry, fashion, , and they might better seek wisdom in the hu- politics serve to convince scientists as they manism of historical evidence on regime do other folk by labeling these the "con- changes or of introspection about how in- text of discovery." The way scientists dis- vestors might react to announcements of cover hypotheses has been held to be dis- new monetary policies. And of course they tinct from the "context of justification," do. namely, proofs of a modernist sort. Tho- The point is that one cannot tell mas Kuhn's autobiographical reflections whether an assertion is persuasive by on the matter can stand for puzzlement knowing at which portion of the scientific/ in recent years about this ploy: humanistic circle it came from. One can tell whether it is persuasive only by think- Having been weaned intellectually on these ing about it. Not all regression analyses distinctions and others like them, I could moral scarcely be more aware of their import and are more persuasive than all argu- force. For many years I took them to be about ments; not all controlled experiments are the nature of knowledge, and . .. yet my at- more persuasive than all introspections. tempts to apply them, even grosso modo, to Economic intellectuals should not dis- the actual situations in which knowledge is criminate against propositions on the basis gained, accepted, and assimilated have made them seem extraordinarilyproblematic [Kuhn, of race, creed, or epistemological origin. 1970, p. 9]. There are some subjective, soft, vague propositions that are more persuasive The methodologist's claim is that "ulti- than some objective, hard, precise propo- mately" all knowledge in science can be sitions. brought into the hard and objective side Take, for instance, the law of demand. of Figure 2. Consequently, in certifying The economist is persuaded that he will propositions as really scientific there is buy less oil when its price doubles better great emphasis placed on "conceivable than he or anyone else is persuaded of falsification"and "somefuture test." The the age of the universe. He may reason- apparent standard is the Cartesian one ably be persuaded of it better than he is that we can find plausible only the things that the earth goes around the sun, be- we cannot possibly doubt. But even this cause not being an astronomer with direct curious standard is not in fact applied: a knowledge of the experiments involved 512 Journal of Economic Literature, Vol. XXI (June 1983) he has the astronomical facts only from cient on M in a regression of prices in 30 the testimony of people he trusts, a reli- countries over 30 years is insignificantly able though not of course infallible source different from 1.0") should take over the of knowledge.1"The economic fact he has whole of persuasiveness, leaving moral mostly from looking into himself and see- persuasiveness incomparablyinferior to it. ing it sitting there. The ceremony of the Arguments such as that "murder violates official rhetoric to the contrary, it is not the reasonable moral premise that we because the law of demand has predicted should not force other people to be means well or has passed some statistical test that to our ends" or that "from behind a prena- it is believed-although such further tests tal veil of ignorance of which side of the are not to be scorned. The "scientific" murderer's revolver we would be after character of the tests is irrelevant. It may birth we would enact laws against mur- be claimed in reply that people can agree der" are persuasive in comparable units. on precisely what a regression coefficient Not always, but sometimes, they are, in- means but cannot agree precisely on the deed, more persuasive, better, more prob- character of their introspection. Even if able (Toulmin, 1958, p. 34). We believe true (it is not) this is a poor argument for and act on what persuades us-not what ignoring introspection if the introspection persuades a majority of a badly chosen is persuasive and the regression coeffi- jury, but what persnades well educated cient, infected with identification prob- participants in our civilization and justly lems and errors in variables, is not. Preci- influential people in our field. To attempt sion means low variance of estimation; but to go beyond persuasive reasoning is to if the estimate is greatly biased it will tell let epistemology limit reasonable persua- precisely nothing. sion. An extreme case unnecessary for the ar- gument here will make the point clear. VIII. The Good of Rhetoric You are persuaded that it is wrong to mur- der better than you are persuaded that Better Writing inflation is always and everywhere a mon- Well, what of it? What is to be gained etary phenomenon. This is not to say that by taking the rhetoric of economics seri- similar techniques of persuasion will be ously? The question can be answered by applicable to both propositions. It says noting the burdens imposed by an unex- merely that each within its field, and each amined rhetoric. therefore subject to the methods of honest First of all, economics is badly written, persuasion appropriate to the field, the written by a formula for scientific prose. one achieves a greater certainty than the The situation is not so bad as it is in, say, other. To deny the comparison is to deny psychology, where papers that do not con- that reason and the partial certitude it can form to the formula (introduction, survey bring applies to nonscientific subjects, a of literature, experiment, discussion, and common but unreasonable position. There so forth) are in some journals not accepted. is no reason why specifically scientific per- But economists are stumbling towards suasiveness ("at the .05 level the coeffi- conventions of prose that are bad for clar- ity and honesty. The study of rhetoric, it "The astronomical "fact" that the earth goes around the sun, of course, is not even a properly must be said, does not guarantee the stu- modernist fact, though it is commonly treated as one dent a good English style. But at least it in such discussions. Which goes around which is a makes him blush at the disdain for the matter of the point of view one chooses. It is the aesthetics of the simpler theory, not the "facts," that reader that some economics exhibits (Wal- leads to heliocentrism. ter Salant, 1969). McCloskey: The Rhetoric of Economics 513

The economist's English contains a mes- is the tacit knowledge of the sort Polanyi- sage, usually that "I am a Scientist: give describes.12We know the economics, but way." Occasionally the message is more cannot say it, in the same way a musician genial: ' irony says "Do not knows the note he plays without con- make a fetish out of these methods I am sciously recalling the technique for exe- expounding: they are mere human arti- cuting it. A singer is a prime example, for fices." Milton Friedman's style, so careful there is no set of mechanical instructions and clear, has to an exceptional degree one can give to a singer on how to hit a the character of the Inquirer. We will not high C. Al Harberger often speaks of so- raise up a race of Dennis Robertsons, Rob- and-so being able to make an economic ert Solows, George Stiglers, or Robert Lu- argument "sing." Like the directions to cases by becoming more sensitive to the Carnegie Hall, the answer to the question real messages in scientific procedure and "how do you get to the Council of Eco- prose, but maybe we will stunt the growth nomic Advisors?" is "practice, practice." of the other kind. Better Foreign Relations Better Teaching A third burden placed on economics by A second burden is that economics is its modernist methodology is that eco- badly taught, not because its teachers are nomics is misunderstood and, when re- boring or stupid, but because they often garded at all, disliked by both humanists do not recognize the tacitness of economic and scientists. The humanists dislike it for knowledge, and therefore teach by axiom its baggage of antihumanist methodology. and proof instead of by problem-solv- The scientists dislike it because it does not ing and practice. To quote Polanyi yet in reality attain the rigor that its method- again: ology claims to achieve. The bad foreign relations have many costs. For instance, . . .the transmissionof knowledge from one as was noted above, economics has re- generation to the other must be predominately tacit.... The pupil must assume that a teach- cently become imperialistic. There is now ing which appears meaningless to start with an economics of history, of sociology, of has in fact a meaning that can be discovered law, of anthropology, of politics, of politi- by hitting on the same kind of indwelling [a cal philosophy, of . The flabby meth- favorite Polanyi expression] as the teacher is odology of modernist economics simply practicing [1966, p. 61]. makes this colonization more difficult, It is frustratingfor students to be told that raising irrelevant methodological doubts economics is not primarily a matter of in the minds of the colonized folk. memorizing formulas,but a matter of feel- Better Science ing the applicability of arguments, of see- ing analogies between one application and A fourth burden is that economists a superficially different one, of knowing pointlessly limit themselves to "objective" when to reason verbally and when mathe- facts, admitting the capabilities of one's matically, and of what implicit character- own or others' minds as merely sources ization of the world is most useful for cor- of hypotheses to be tested, not as them- rect economics. Life is hard. As a blind selves arguments for assenting to hypothe- man uses his stick as an extension of his ses. The modernist notion is that common body, so whenever we use a theory "we sense is nonsense, that knowledge must incorporate it in our body-or extend our somehow be objective, not verstehen or body to include it-so that we come to 120n this score, and some others, I can heartily dwell in it." Problem-solving in economics recommend McCloskey, 1982. 514 Journal of Economic Literature, Vol. XXI (June 1983)

introspection. But, to repeat, we have nomics. But self-testimony is not useless, much information immediately at our dis- even for the purpose of resolving the mar- posal about our own behavior as economic ginal cost- debate of the molecules, if we would only examine the 1930s. One could have asked "Has your grounds of our beliefs. The idea that obser- profit margin' always been the same?" vational proofs of the law of demand, such "What do you think when you find sales as the Rotterdam School's multi-equation lagging?" (Lower profit margin? Wait it approach, are more compelling than in- out?) Foolish inquiries into motives and trospection is especially odd. Even the foolish use of human informants will pro- econometrics itself would be better, as duce nonsense. But this is also true of fool- Christopher Sims has recently argued: ish use of the evidence more commonly admitted into the economist's study. If we think carefully about what we are doing, we will emerge, I think, both more confident Better Dispositions that much of applied econometrics is useful, despite its differences from physical science, A fifth and final burden is that scientific and more ready to adapt our language and debates in economics are long-lasting and methods to reflect what we are actually doing. ill-tempered. Journals in geology are not The result will be econometrics which is more filled with articles impugning the charac- scientific [by which he means "good"] if less superficiallysimilar to statistical methods used ter of other geologists. They are not filled in experimental sciences [Sims, 1982, p. 25]. with bitter controversies that drone on from one century to the next. No wonder. The curious status of survey research Economists do not have an official rhetoric in modern economics is a case in point. that persuasively describes what econo- Unlike other social scientists, economists mists find persuasive. The mathematical are extremely hostile towards question- and statistical tools that gave promise in naires and other self-descriptions. Second- the bright dawn of the 1930s and 1940s hand knowledge of a famous debate of ending economic dispute have not suc- among economists in the late 1930s is part ceeded, because too much has been asked of an economist's formal education. The of them. Believing mistakenly that opera- debate concerned the case of asking busi- tionalism is enough to end all dispute, the nessmen if they equalized economist assumes his opponent is dishon- to marginal revenue. It is revealing that est when he does not concede the point, the failure of such a study-never mind that he is motivated by some ideological whether that was indeed the study-is passion or by self-interest, or that he is supposed to convince economists to aban- simply stupid. It fits the naive fact-value don all self-testimony. One can literally split of modernism to attribute all dis- get an audience of economists to laugh agreements to political differences, since out loud by proposing ironically to send facts are alleged to be, unlike values, im- out a questionnaire on some disputed eco- possible to dispute. The extent of disagree- nomic point. Economists are so impressed ment among economists, as was men- by the confusions that might possibly re- tioned, is in fact exaggerated. The amount sult from questionnaires that they aban- of their agreement, however, makes all don them entirely, in favor of the confu- the more puzzling the venom they bring sions resulting from external observation. to relatively minor disputes. The assaults They are unthinkingly committed to the on Milton Friedman or on John Kenneth notion that only the externally observable Galbraith, for example, have a bitterness behavior of economic actors is admissable that is quite unreasonable. If one cannot evidence in arguments concerning eco- reason about values, and if most of what McCloskey: The Rhetoric of Economics 515

matters is placed in the value half of the Ed.: MARTIN STEINMANN, JR. NY: Scribner's, 1967. fact-value split, then it . Modern dogma and the rhetoric of assent. follows that one Chicago, IL: U. of Chicago Press, 1974. will embrace unreason when talking BURKE, KENNETH. A rhetoric of motives. Berkeley: about things that matter. The claims of U. of California Press, 1950. an overblown methodology of Science COASE, RONALD. "How Should Economists Choose?" The G. Warren Nutter LectUres in Politi- merely end conversation.13 cal Economy. Washington, DC: American Enter- prise Institute, 1982. A rhetorical cure for such disabilities COHEN, KALMAN AND CYERT, RICHARD. . 2nd ed. Englewood Cliffs, NJ: Prentice- would reject philosophy as a guide to sci- Hall, 1975. ence, or would reject at least a philosophy COOLEY, T. F. AND LEROY, S. F. "Identification and that pretended to legislate the Estimation of Money Demand," Amer. Econ. Rev., knowable. Dec. 1981, 71(5), pp. 825-44. The cure would not throw away the illu- DAVIS, PHILIP J. AND HERSH, REUBEN. The mathe- minating regression, the crucial experi- matical experience. Boston, MA: Houghton Mifflin, ment, the unexpected implication unex- 1981. DEWEY, JOHN. The questfor certainty. NY:Putnam, pectedly falsified. These too persuade [1929] 1960. reasonable scholars. Non-argument is the EINSTEIN, ALBERT. "Aphorisms for Leo Baeck," re- necessary alternative to narrow argument printed in Ideas and opinions. NY: Dell, [1953] 1973. only if one accepts the dichotomies of FEYERABEND, PAUL. Against method: Outline of an modernism. The cure would merely rec- anarchistic theory of knowledge. : Verso, ognize the good health of economics, dis- [1975] 1978. . Science in a free society. London: New Left guised now under the neurotic inhibitions Books, 1978. of an artificial methodology of Science. FISCHER, DAVID HACKETT. Historians' fallacies. NY: Harper & Row, 1970. FRENKEL, JACOB. "Purchasing Power Parity: Doc- REFERENCES trinal Perspectives and Evidence from the 1920s," ALCHIAN, ARMEN. "Uncertainty, Evolution, and J. Int. Econ., May 1978, 8(2), pp. 169-91. Economic Theory," J. Polit. Econ., June 1950, FRIEDMAN, MILTON. "The Methodology of Positive 58(3), pp. 211-21. Economics,"in in positive economics. Chi- BARFIELD, OWEN. "Poetic Diction and Legal Fic- cago, IL: U. of Chicago Press, 1953. tion," in Essays presented to Charles Williams. FRIEDMAN, MILTON AND SCHWARTZ, ANNA J. A London: Oxford U. Press, 1947; reprinted in The monetary history of the United States. Princeton, importance of language. Ed.: MAx BLACK,Engle- NJ: Princeton U. Press, 1963. wood Cliffs, NJ: Prentice-Hall, 1962, pp. 51-71. GARDNER, JOHN. On moralfiction. NY: Basic Books, BECKER,GARY S. AND STIGLER,GEORGE J. "De Gus- 1978. tibus Non Est Disputandum," Amer. Econ. Rev., GENBERG, A. HANS. "Aspects of the Monetary Ap- Mar. 1977, 67(2), pp., 76-90. proach to Balance-of-Payments Theory: An Empir- in monetary approach BENTHAM,JEREMY. The book of fallacies from un- ical Study of Sweden," The to of payments. Eds.: JACOB A. finished papers. London: Hunt, 1824. the balance BLACK,MAX. Models and metaphors:Studies in lan- FRENKEL AND HARRY G. JOHNSON. London: Allen guage and philosophy. Ithaca, NY: Cornell U. & Unwin, 1976. Press, 1962. GOULD, STEPHEN JAY. Ever since Darwin. NY: Nor- BLAUG, MARK. The methodology of economics: Or ton, 1977. how economists explain. Cambridge, U.K.: Cam- . The mismeasureof man. NY: Norton, 1981. bridge U. Press, 1980. VAN HEIJENOORT, JOHN. "Godel's Proof," in The en- of philosophy. NY: Macmillan & Free BOOTH,WAYNE C. The rhetoric of fiction. Chicago, cyclopedia IL: U. of Chicago Press, 1961. Press, 1967. HORSBURGH, H. J. N. "Philosophers Against Meta- __. "The Revival of Rhetoric," in New rhetorics. phor," Philosophical Quart., July 1958, 8(32), pp. 231-45. 13Listen to Harry Johnson: "The methodology of HUME, DAVID. An inquiry concerning human un- positive economics was an ideal methodology for derstanding. Ed.: CHARLES W. HENDEL. India- justifying work that produced apparently surprising napolis: Bobbs & Merrill, [1748] 1955. results without feeling obliged to explain why they JAMES, WILLIAM. "Pragmatism's Conception of occurred" (Johnson, 1971, p. 13). I do not need to Truth" reprinted in Essays in pragmatismby Wil- think about your evidence for widespread monopo- liam James. Ed.: CASTELL ALBUREY. NY: Hafner, listic competition because my methodology tells me [1907] 1948. the evidence is irrelevant. JOHNSON, HARRY G. "The Keynesian Revolution and 516 Journal of Economic Literature, Vol. XXI (June 1983)

the Monetarist Counterrevolution,"Amer. Econ. . Philosophical reasoning. 2nd ed. London: Rev., May 1971, 61(2), pp. 1-14. Duckworth, 1970. KEARL, J. R.; POPE, CLAYNE; WHITING, GORDON PEIRCE,CHARLES. "The Fixation of Belief," re- AND WIMMER, LARRY. "A Confusion of Econo- printed in Values in a universe of chance: Selected mists?" Amer. Econ. Rev., May 1979, 69(2), pp. writings of Charles S. Peirce. Ed.: P. P. WIENER. 28-37. Garden City, NJ: Doubleday, [1877] 1958. KLINE, MORRIS.Mathematics: The loss of certainty. PERELMAN, CHAIM AND OLBRECHTS-TYTECA,L. NY: Oxford, 1980. The new rhetoric: A treatise on argumentation. KRAVIS,IRVING B. AND LIPSEY, ROBERTE. "Price Eng. trans. Notre Dame: Notre Dame U. Press, Behavior in the Light of Balance of Payments The- [1958] 1967. ories," J Int. Econ., May 1978, 8(2), pp. 193-246. POLANYI, MICHAEL. Personal knowledge: Towards KRUGMAN,PAUL R., "PurchasingPower Parity and a post-critical philosophy Chicago, IL: U. of Chi- Exchange Rates:Another Look at the Evidence," cago Press, 1962. I. Int. Econ., Aug. 1978, 8(3), pp. 397-407. . The tacit dimension. Garden City, NY: Dou- KUHN, THOMAS. The structure of scientific revo- bleday, 1966. lutions. 2nd ed. Chicago: U. of Chicago Press, POPPER, KARL. The logic of scientific discovery. 1970. Eng. trans. NY: Harper, [1934] 1959. . The essential tradition: Selected studies in . The open society and its enemies. London: scientific traditionand change. Chicago: U. of Chi- , 1945. cago Press, 1977. . Unended quest: An intellectual autobiogra- LAKATOS,IMRE. Proofs and refutations: The logic phy. London: Collins, 1976. of mathematical discovery. Cambridge: Cam- QUINTILIAN,MARCUS F. Institutio oratoria. Cam- bridge U. Press, 1976. bridge, MA: Harvard U. Press, [c. 100 AD] 1920. _ The methodology of scientific research pro- REDER,MELVIN. "Chicago Economics:Permanence grammes. From articles 1963-1976. Cambridge, and Change," J. Econ. Lit., Mar. 1982, 20(1), pp. NY and London: Cambridge U. Press, 1978. 1-38. AND MUSGRAVE,ALAN. Criticism and the RICHARDS,I. A. The philosophy of rhetoric. NY:Ox- growth of knowledge. Cambridge: Cambridge U. ford U. Press, 1936. Press, 1970. RICHARDSON,J. D. "Some Empirical Evidence on LEAMER, EDWARD. Specification searches: Ad hoc Commodity Arbitrageand the Law of One Price," inferences with nonexperimentaldata. NY:Wiley, J. Inter. Econ., May 1978, 8(2), pp. 341-51. 1978. ROLL,RICHARD AND Ross, STEPHEN."An Empirical LEVI, EDWARD.An introduction to legal reasoning. Investigation of the Arbitrage Pricing Theory," J. Chicago, IL: U. of Chicago Press, [1948] 1967. Finance, Dec. 1980, 35, pp. 1073-1103. LEWIS,C. S. "Buspels and Flansferes," in Rehabilita- RORTY,RICHARD. Philosophy and the mirrorof na- tions and other essays. London: Oxford U. Press, ture. Princeton NJ: Princeton U. Press, 1979. 1939; reprinted in The importance of language. . Consequencesof pragmatism (Essays: 1972- Ed.: MAx BLACK.Englewood Cliffs, NJ: Prentice- 1980). Minneapolis:U. of Minnesota Press, 1982a. Hall, 1962. . "The Fate of Philosophy," TheNew Republic, MCCLOSKEY,DONALD N. "The Loss to Britain from Oct. 18, 1982b, 187(16), pp. 28-34. Foreign Industrialization,"reprinted in his Enter- SALANT, WALTER. "Writing and Reading in Eco- prise and trade in VictorianBritain. London: Al- nomics," J. Polit. Econ., July-Aug. 1969, 77(4, Pt. len & Unwin, [1970] 1981. I), pp. 545-58. . The applied theory of price. NY: Macmillan, SAMUELSON,P. A. The foundations of economic 1982. analysis. Cambridge, MA: Harvard U. Press, 1947. AND ZECHER,J. RICHARD. "How the Gold SEN, AMARTYA."Behaviour and the Concept of Pref- Standard Worked, 1880-1913," in The monetary erence," Inaugural Lecture. London School of approach to the balance of payments. Eds.: J. Economics and Political Science, 1973. FRENKELAND H. G. JOHNSON.London: Allen & SHARPE,WILLIAM. Portfolio theoryand capital mar- Unwin, 1976. kets. NY: McGraw Hill, 1970. AND ZECHER,J. R. "The Success of Purchas- SIMS, CHRISTOPHER. "Review of Specification ing Power Parity,"in A retrospectiveon the classi- searches:Ad hoc interference with nonexperimen- cal gold standard. Eds.: MICHAEL BORDO AND tal data. By Edward E. Leamer." J. Econ. Lit., ANNA J. SCHWARTZ. NBER conference, 1982. June 1979, 17(2), pp. 566-68. Forthcoming. . "Scientific Standards in Econometric Model- MISES, LUDWIG VON. Human action. New Haven, ing," Unpub. paper for the 25th anniversary of CT: Yale U. Press, 1949. the Rotterdam Econometrics Institute, Apr. 1982. MOOD, A. F. AND GRAYBILL,F. A. Introduction to STEINER,GEORGE. After Babel:Aspects of language. the theory of statistics. 2nd ed. NY: McGraw Hill, London: Oxford U. Press, 1975. 1963. STIGLER,GEORGE J. "The Conference Handbook," PASSMORE,JOHN. A hundred years of philosophy, J Polit. Econ., Apr. 1977, 85(2), pp. 441-43. 2nd ed. London: Penguin, 1966. TOULMIN, STEPHEN. The uses of argument. Cam- . "Logical Positivism," in The encyclopedia of bridge: Cambridge U. Press, 1958. philosophy. NY: Macmillan, 1967. . "The Construal of Reality: Criticism in Mod- McCloskey: The Rhetoric of Economics 517

ern and Postmodern Science," Critical Inquiry, Boston, MA: Blackwell Scientific, 1981, pp. 16- Autumn 1982, 9(1), pp. 93-110. 35. WEBSTER, GLENN; JACOX, ADA AND BALDWIN, BEV- WHATELY, RICHARD. Elements of rhetoric. 7th ed. ERLY. "Nursing Theory and the Ghost of the London: [1846] 1894. Received View," in Current issues in nursing. ZECKHAUSER, RICHARD AND STOKEY, EDITH. A Eds.: JOANNE MCCLOSKEY AND HELEN GRACE. primerfor policy analysis. NY: Norton, 1978.