Friedman’s essay is not only rhetorical itself; it also makes a rhetorical point

draft of

The Rhetorical Ethos of Positive Economics

Arjo Klamer

Cultural Economics, Erasmus University Rotterdam [[e-mail]]
and Deirdre McCloskey
Economics, History, English, and Communication, University of Illinois at Chicago; Tinbergen Gasthoogleraar of Philosophy, Economics, and Cultural Economics, Erasmus [[e-mail]]

Crude preliminary version: comments welcome

From Klamer and McCloskey a claim that Friedman’s essay “is rhetorical” should come as no surprise. For twenty years, after all, we have been arguing ad nauseum that almost all writing is rhetorical, necessarily and inevitably (McCloskey, 1983, 1986, Klamer, McCloskey and Solow, 1989; see the bibliographies in the 2^nd edition of The Rhetoric of Economics, 1998). Friedman’s essay is, too. That is, it attempts to persuade, as any scientific paper does; or indeed as most human speech acts do. Even the speech act of “stating,” such as stating that the time is 3:45 pm, or that the cat is on the mat, or that demand curves slope down, which sound so blessedly free of human purpose (at least compared with speech acts like “pleading” or “defending”), can and usually do if actually uttered outside a philosophy seminar have a pragmatic intent: to express cooperativeness in the face of a question, “What time is it?”; or to express a quantitative assurance that the cat is just about the only thing in the mat; or to express a qualitative assurance that the speaker in fact believes demand curves slopes down and has adequate evidence that they do.¹

More challenging is the claim that the rhetoric of Friedman’s essay has been unusually effective as far as the meta-discussions in economics are concerned. Viewing it as itself a speech act, its perlocutionary force has been immense. Of this there is no doubt, even considered as a mere statement. To this day, for example, fifty years on, the first-year students in the graduate program at Indiana University are provided with a copy and told that this is The Methodology of Economics. But a separate question is whether the essay has also had an impact on practice. That Friedman---who at the time he wrote it was viewed with considerable alarm in most portions of economics---succeeded in writing a holy text is hard to deny; one might deny, however, that the believers have followed the holy text. It would not be the first time.

Our central claim is that the essay makes a rhetorical point. That is, it is “rhetorical” in a meta-methodological sense: it talks about how economists talk, saying in the 1950s what we started saying in the 1980s. It is rhetorical in Aristotle’s sense, the study of the available means of [unforced] persuasion. This is not a point that others have taken away from reading it (cf Maki 2003), although de Marchi and Hirsch come close. Read around the essay, however, and one finds a powerful story of rhetoric, which tells a great deal about the profession, a little about Friedman himself, and has an especially startling moral for the enterprise called economic methodology.

The essay makes a rhetorical point

Though Friedman in our view made a rhetorical point, a good deal of confusion has ensued because he claimed he was making a point of logic. He was not. His point is the same as in “The Rhetoric of Economics” (McCloskey, 1983; and Friedman in fact commented favorably on McCloskey’s essay in personal correspondence): said Friedman in effect, economics is rhetorical, since it makes for example extensive use of metaphor, among numerous other rhetorical devices.

Friedman did not have the concept of metaphor available—it was not a current concept in the vocabulary of economists, certainly, and not even in that of sophisticated philosophers of science. The 20^th-century movement to precision in philosophical language had not yet reached the point of realizing that metaphor creates special problems for a mechanical view of semantics. Friedman’s claims are simply reasoning by means of a metaphor, or by analogy if you wish. A metaphor connects the phenomenon of one domain to another pertaining to another domain. “Time” belongs to the domain of “the passing of it, clocks, history, past, present, future, and so on”; “money” belongs to the entirely different domain of “cash, M1, banks, financial markets, prices, costs, wealth, opportunity costs” and so on. By connecting the two you can produce literally absurd assertions such as “time is money”---and that is a metaphor, bringing the two domains into a relationship (which one might call “functional” in the mathematical sense). The linguistic problem is that metaphors are absurd lies. Time is not literally money. Men are not literally wolves. A metaphor is never literally true. It can’t be. (At this point the reader of the Friedman text should be seeing the point.)

You are to think of time in terms of the domain of money. What will you do? Do you connect the “clock” with “prices,” with a fleeting vision perhaps of the clocks at Alsmeer in the Netherlands counting down the auction prices of tulips, or do you pick another combination? What you do shows in your action. When we tell you that time is money, you get the message, and will hurry—our time is “precious” and we do not like to “waste” it. Note how easy it is to spin out the metaphor. A metaphor is “productive” in the same way that a good scientific theory is (for the good reason that theories are metaphors [Cartwright]). Whether you quite realize it or not, you as an economist connect “the passing of time” with the notion of “opportunity costs” in the domain of money. Outside of an audience of professional economists it is not the most obvious functional equivalence. But that is how the metaphor performed in the field, earning for example one of Milton’s students, Gary Becker, a Nobel prize.

In the vocabulary invented by Max Black, a philosophical student of metaphors, time is the primary subject and money the subsidiary subject. Black, with the literary critic I. A. Richards and many other students of metaphor in the 1940s and 1950s, dismissed the analytical strategy of transforming metaphors into semantically literal sentences. Such a transformation, they argued, does not do justice to the human importance of metaphors. Metaphors have a cognitive effect, as our conceptions of how the two terms interact with each other. Metaphors jolt our mind by steering our thoughts in unusual directions. This is especially the case with good poetic metaphors. It is their power; but it is also their weakness, at least in scientific settings, because a rich metaphor carries a variety of possible meanings. It “carries over and beyond,” which after all is the literal Greek meaning of the word. Equivocation (the chief analytic sin), thy name is metaphor. What prevents you, for example, from thinking of “time” as “wealth” instead of “money”? But such an alteration changes the meaning. Instead of rush, rushing you relax, happy in the thought: “Someone with plenty of time is rich, and leisured.”

Because of the instability of metaphors and because of their apparent falseness scientists are wont to attack them, though tangled in them necessarily. The empiricists of the 17^th century were especially vitriolic. Bacon, Descartes, Spinoza, Hobbes all write eloquently against eloquence. Consider Locke, who attacked rhetoric, and its “figures,” in a famous passage in the Essay Concerning Human Understanding (by the time he wrote, 1690, the trope of argument of attacking metaphor metaphorically was conventional): “If we would speak of Things as they are, we must allow, that the art of Rhetorick, besides Order and Clearness, all the artificial and figurative application of Words Eloquence hath invented, are for nothing else but to insinuate wrong Ideas, move the Passions, and therby mislead the Judgement; and so indeed are perfect cheat” (Essay, II,X,1 ). In view of the rhetorical flourish and exaggeration common in the writing of the time his railing against metaphors is understandable. Consider the following purple prose: “Eloquence, like the fair Sex, has too prevailing Beauties in it, to suffer it self ever to be spoken against. And 'tis vain to find fault with those Arts of Deceiving, wherein Men find pleasure to be Deceived.” The passage, of course, is Locke's own (Essay, III, X, 34, as cited by Bennington 1987, p. 106). Is Locke wittingly deploying one of the very master tropes he deprecates; or does his innocence demonstrate the impossibility of an altogether non-metaphorical language? Take your pick.

The attack is to no avail, for science is inconceivable without metaphor. Physicists imagine solar systems to think about atoms and its electrons; they think of strings and forces to think about gravity. Economists propose to consider education as a market. Education is obviously not a market, at least in the Alsmeer sense, but by stating it is you are invited to think as if it is and so are made to think of it in terms of products, demand, supply, and price, the elements of the domain of the market. The game is another popular metaphor among economists nowadays, and another item in Becker’s Nobel prize. Again, where most people perceive competition, economists think in terms of a game. You are invited to think of pay-off schemes, optimal strategies, and Nash equilibria. Such metaphors are heuristic: scientists use them to get started in their thinking about a problem (see Klamer and Leonard 19*). The heuristic metaphor frames their thinking and points them analytical directions to pursue.

Scientists differ from poets in making a concerted effort to control the instability of their heuristic metaphors. It is for good reason that we speak of a “discipline.” Associating game theory with play (Homo ludens, as a great Dutch historian put it) is ruled out of order. When someone associates markets with commerce and greed the economists shrug him off: the association is of no relevance, they think, for the analysis.

But the reasoning in any case starts with a metaphor. Peirce spoke therefore of abduction, that is reasoning by way of analogy, as the core of scientific reasoning---more so than induction and deduction. (Analogy is an expanded metaphor with the emphasis on comparisons between relations rather than attributes of the principal and subsidiary subjects. When particle physicists use the analogy with the solar system they consider the movements and not the attributes of the planets.)

Friedman justifies the practice of abduction without knowing the concept (there is no more evidence that he had read pragmatic philosophers than that he had read positivist philosophers, and good deal that he had not---for example, his own testimony). To get us thinking about the expert billiard player he invites us to think as if “he [the player] knew the complicated mathematical formulas that would give the optimum directions of travel, could estimate accurately by eye the angles etc., describing the locations of the balls, could make lightning calculations from the formulas, and could then make the balls travel in the direction indicated by the formulas” (p.21). Of course the player does not know all that. But the analogy might get the analysis started. Economists, so Friedman continues, do something similar when they get thinking about consumers, investors, and workers. They begin by stating that such people are able to solve Lagrangians. Everyone knows that can’t. Most people have difficulty adding and subtracting. But it does not matter, not more than a grasp of physics of riding a bicycle is necessary for riding it (see Polanyi, p. ). Planets have volume and are not literally collapsed into a mathematical a point; nobody ever saw a force in and of itself, and so forth.

Friedman put great emphasis on the pragmatics of economic argumentation. Again, he does not use these terms, but he just as well could have (an instance of as-if thinking applied to methodology itself). He is a practitioner telling fellow practitioners how to reason. Set up your heuristic metaphor, do not worry about its realisticness (cf Maki), or the literal truth of what you are saying. Focus on how it works for you. A practitioner is not preoccupied with the truth, as an analytical philosopher wants him to be. Truth is not the principle that guides him. What matters is what works. At times you will want to consider the plausibility of an assumption, for the pragmatic reason that wildly implausible assumptions, or assumptions that come out of the blue and are not well connected with the ongoing research program, will not fly. Friedman goes to great length to make this clear. “Confidence in the maximization-of-returns hypothesis is justified by evidence of a very different character.” So forget about what business people say that they do, he tells his fellow practitioners. The assumption makes sense for the same reasons that the assumptions about the expert billiard player does: if business men were not trying to maximize returns they would not last. Such an argument makes the assumption plausible, by appealing to the other master metaphor of price theory: survival of the fittest, as Friedman points out explicitly. It does not tell whether the “analogy” works. A good test, he says, is the predictive accuracy of the hypothesis.

There you have it, a 3-by-5 card of prescriptions for good practice. It is not the teoria of scientific practice but the praxis, guided by common sense, phronesis.

What Friedman means by predictive accuracy is not immediately clear. As deMarchi and Hirsch stress, Friedman had extensive experience with statistical work. He was always empirically minded, and came equipped with respect for the facts. He knew all about the tinkering and experimenting that characterizes empirical work. He had little enthusiasm for rigid theoretical frameworks of the sort that Paul Samuelson was selling about that time, but he also saw as fruitless empirical work like surveys among businessmen to find out what they thought they are doing. The point is not to get photographic reproductions. “Marshall [the hero and model for Good Old Chicago School economics] took the world as it is; he sought to construct an `engine’ to analyze it, not a photographic reproduction of it” (p 35). Note that “the engine” is a metaphor for the heuristic metaphor. It is a practical point. Friedman is thinking of long articles that had filled numerous pages of journals with endless descriptive statistics. All the information had produced virtually no lasting economic knowledge---though Friedman, a student of Arthur Burns and of Wesley Mitchell, had respect for what blackboard economists (such as Tjallings Koopmans) attacked as “measurement without theory.” “The ultimate goal of a positive science is the development of a `theory’ or `hypothesis’ that yield valid and meaningful (i.e. not truistic) predictions about phenomena not yet observed” (p. 7). At the end he declares that “Economics as positive science is a body of tentatively accepted generalizations about economic phenomena that can be used to predict the consequences of changes in circumstances” (p. 39). If he had known about Popper he would have used Popper’s maxim of bold conjectures followed by the pursuit of refutations (economists have misunderstood Friedman by imagining that he did read German and had read Popper’s Ph.D. dissertation before its translation into English; on the other hand, such ideas were in the air of the Mt. Pelerin Society). The bold conjectures would come with heuristic metaphors that would be literally false, but that could get the thinking going. It is a matter of phronesis, the rhetorician would say, that practical wisdom. Starting with full-blown accounts of what business people actually do and think would not get us nowhere, Friedman the practitioner had come to realize. It is better to start with bold conjectures and see where they get us. The proof of the pudding, of course, is in the eating.

The Rhetorical Performance

Most interpreters give the essay a philosophical reading. They try to determine what it actually states, expose its flaws, inconsistencies, and shortcomings. So we know that Friedman does conflate idealizations and abstractions, is mixed up about types of assumptions, cares about certain assumptions anyway, and is vague about what he means by predictive accuracy. Perhaps that is why Milton long ago repudiated the essay. None of the philosophical readings, however, tell us about how it performs, why it had so much effect on the meta-discourse of economists yet ended up ineffectual as to the actual practice of economics.

The essay (Friedman actually speaks of it as a “paper”) starts off in his exordium stating the case of positive science. He uses the authority of John Neville Keynes, who which in the early 1950s was quite an elderly citation. He frames science as a problem of choice: “how to decide whether a suggested hypothesis or theory should be tentatively accepted as part of the `body of systematized knowledge concerning what is’” (p. 3). His distinction between positive and normative economics ends on an upbeat note:

I venture the judgment . . . that currently in the Western world, and especially in the United States, differences about economic policy among disinterested citizens derive predominantly from different predictions about economic consequences of taking action—differences that can in principle can be eliminated by the progress of positive economics—rather than from fundamental differences in basic values, differences from about which men can ultimately fight.

We see here the traces of a narrative in which the reader is invited to sympathize with the positive scientist, the protagonist, who is able to overcome differences and point the way to agreement. It is especially the American positive scientists whom Friedman evokes. The Europeans are in his view apparently too much affected by the Marxist bug, or something of the sort: anyway, an insufficient attachment to Getting the Job Done. The antagonists of the story do not appear in the main body of the text, but are identified in footnotes. (Apparently Friedman does not want them too much status by addressing them directly.) He singles out (in a footnote) “some economists, particularly a group connected with the Cowles Commission in Economics at the University of Chicago” as the bad guys. (The Cowles Commission was of course at Chicago then, but did not stay long after Friedman’s arrival.) Another target are economists who take to the theory of monopolistic and imperfect competition (p. 15), principally Edward Chamberlain himself, who had already identified the Chicago view as an obstacle to his theory (one of us was Chamberlain’s student in 1962 and wrote for the course a pro-Harvard attack on Chamerlain’s enemies). The narrative here tells about the pitfalls and distractions for the hero, promising that he will win out over the enemies if he is pure and good. As is routinely the case in scientific persuasion, the implied narrative is one that Odysseus would have understood.

The narrative is intended to motivate, to spur to a specific action, as all good narrative does. Its theme is the goal of positive science: “The ultimate goal of a positive science is . . . “, etc., as before. The narrative tells which practices are good and which are bad. (Methodology is of course an ethical practice.) It tells us that our theories are always tentative, claiming as is usual in such productions (there are two possibilities and this is one of them) the ethos of the modest speaker: “Any theory is necessarily provisional and subject to change with the advance of knowledge” (p 41).

Friedman ends with a romantic characterization of the creative scientist: “the construction of hypotheses is a creative act of inspiration, intuition, invention; its essence is the vision of something new in familiar material. The process must be discussed in psychological, not logical, categories; studied in autobiographies and biographies, not treatises on scientific method; and promoted by maxim and example, not syllogism or theorem” (p.43). The ending is particularly revealing. In his identification with the practicing economist Friedman expects little insight from the logical accounts and analyses that his argument would inspire. To see how science works he suggests we consider how scientists work. Heuristic metaphors from psychology would do better than the ones that analytical philosophers provide. It is not a blackboard matter.

Get the gist of Friedman’s narrative and you get his point. Philosophical readings extract the logic from the argumentation, and subject it to scrutiny. Fine. But what is gained? A rhetorical reading remains at the surface and looks at how the paper performs---by means of a romantic narrative of a hero, for example. The narrative tells that Friedman had the practitioner in mind, and not the economic methodologist (apart from Neville Keynes, as we have noted, and later John Stuart Mill, he cites none—he does not seem to have even known of Terence Hutchison’s work bringing Continental logical positivism c. 1920 into the conversation of economists) or the philosopher of science. He cites no work of philosophers. None. He cites only the work of fellow practitioners. George Stigler, Arthur Burns, and Dorothy Brady, all practicing economists, were his scientific peers.

Accordingly Friedman did not intend the essay to be a serious methodological or philosophical one. He is trying to settle a dispute with opponents in research, not to come to agreement with one philosophical position or another. Friedman underscored the point by his action. Never would he mingle in the methodological discussions that his essay triggered. Later he would profess utter lack of interest in such discussions. He did not care about the ins and outs of assumptions, about whether his was a Popperian argument or not. He had made his meta-point and moved on, practicing economics: shortly after he produced his magnum opus, A Theory of the Consumption Function.

The rhetorical performance of the essay shows in its use of three examples. They function as anecdotes, and tend to stick in people’s head. The billiard player and the leaves have become exemplary in the conversation of economists. Many participants know about them and what they are supposed to illustrate: “Ah yes, the realism of assumptions don’t matter! What a blessing!” The length of the essay invites the reader to skip through the more cumbersome passages. What is left with are some key phrases. They help account for its effect in the meta-discourse of economics

The Impact on the Meta-discourse of Economics

Has there ever been a methodological essay in economics with an impact of this one? Of course not. The essay received extraordinary attention, all the more extraordinary in a field impatient of philosophical thinking. “As-if reasoning”(again there is no evidence that Friedman was a student of the phrase in philsopophy), “the realism of assumptions does not matter,” and “prediction is all that matters” have become commonest of topoi for those who do economics. Practicing economics will cite as a mantra, or an apotropaic incantation, when challenged methodologically. Have someone question the plausibility of one of your assumptions, say “the realism of assumptions does not matter” and the criticism backfires. Just like that. The phrases are conversation stoppers, magical.

The essay received its share of criticisms for its supposed apologetic stance towards Chicago economics, but that only accounts part of the extraordinary attention: after it, it was easy at the time to simply dismiss Chicago-School optimism about markets by mere sneering. The criticism did not affect its practical success, especially remarkable when we realize that a similar argument of someone like Machlup was much more sophisticated---philosophically, that is. Fritz Machlup most certainly did read German. Virtual no practicing economist nowadays knows of Machlup’s methodological writings---yet every one of them believe they have read Friedman’s essay and know its conversation stoppers. In Conversations with Economists (Klamer, 1983) three interviewees spontaneously evoke the “as-if” phrase. Anyone who frequents economic seminars will hear it stated time and again, critically or approvingly, but always without attention to its grounding. Friedman’s phrases are part of the buzz in economics.

One reason for his impact might be his ethos, his standing to speak. Although he was already known to be an outlier in what had become on its frontiers a Keynesian field, he had made a mark with his empirical work. His work on the theory of uncertainty with Savage, furthermore, had been widely noted; and with I.M.D. Little he had put paid to the notion that direct taxes could be proven superior to indirect on a blackboard. Friedman was a coming man in the world of economists. He had the attention of his colleagues. When he had something to say about their practice, they listened, just as they would if someone like Paul Samuelson were to speak up, or (later) Robert Solow. The rhetorician would say that Friedman had the ethos to get attention for his arguments. Still, there is something odd here.

Most remarkable is the way in which the essay has preoccupied those who specialize in economic methodology. At the time the field of economic methodology was all but non-existent. A few economists spent a little of their time thinking and writing about the methodological principles of their science, but it was rare to find someone specialized in it. Terence Hutchison was an exception, yet he, too, was mainly occupied with the history of economic thought. Friedman’s essay changed all this, giving rise to the field of economic methodology. It has spurned a whole industry of bickering about what the essay actually said and what it does not say, about its shortcomings, flaws, and stupidities, about its Popperian connections; and so on. To some extent the essay has defined the field of economic methodology. All the basic issues are there, like the issue of rules (cf. Hands), of theory choice, of the logical properties of an economic theory, of the role and the nature of assumptions, of the empirical foundations, of the proportion between theory and empirical work, and so on.

Its impact on the field of economic methodology is especially remarkable as the author had, as we have noted, no such aspiration. He had no desire to influence the methodological literature, and did not bother therefore to read it. The methodological field internalized the argumentation and made it its own. The efforts had little to do with what Friedman had in mind: affecting the practice of economics. Whereas methodology is preoccupied with probing the system, Friedman develops thoughts and norms that can be followed in the life world.
The pick-up game of basketball is common practice in the U.S. Having gathered at a court two teams are formed by having two people alternately pick the players. Another method is for everyone to shoot a foul shot: the first five to make their shots form team one; the second five become team two. The people left over form a third team that plays against the winner of the first game. (The rule is somewhat strange in that the better players tend to end up on one team and may continue winning till the end.) Before the start of the game the players agree on some basic rules like “looser takes out” (the team that got scored against gets the ball). Most rules, however, are tacit. For example when someone fouls you you are supposed to make the call yourself. There is no referee. Usually the calls are honored, but when someone appears to abuse the rule people will make comments, argue, get mad, or even get into a fight (note that it is men who play by such tacit rules, not women).

Think of economists playing a pick-up game. People like the methodologists are watching the game. And here is the difference. Some of the methodologists yell at the players, telling them how to play the game. We, after all, know the rules because we are philosophers and can tell a good move from a bad one. We at times give the impression of wanting to jump and participate as referees. The players, however, do not pay any attention. They continue to violate and vitiate all kinds of rules and do not seem to be bothered. Only when one of them stops and makes suggestions about one rule or another do they pay attention, as when Milton Friedman tells them that accurate predictions, not realistic assumptions, are key. Then the other players get into the discussion and an argument ensues. Subsequently the play changes a bit. Players will invoke the argument when disputes occur. (“Hey, listen, the realism of assumptions doesn’t matter, don’t you know? Let’s go!” “Wait a minute, Samuelson pointed out it does matter. So let’s see whether what we are doing here is right.” “C’mon man, let’s just play.”) In the meantime we methodologists are busy critiquing what the players said, and yell again that they are mistaken, that Friedman did not mean it that way, that the rules really are different. But the players do not hear and do not care.

I wonder what this says about the games of economists and methodologists. Apparently, the methodologists are involved in a game that has little to do with the game economists are playing. That does not mean it is irrelevant. The endless talk about and around a game like soccer may have no influence on how the game it is played, but it sure is important. The game of soccer would mean little without all that talk. Methodological talk may not be as important for the game of science, but science is inconceivable without it.

With the discursive turn, students of economics have become more modest and extensively study the games economists are playing while withholding judgment. They can only hope that the players will learn from such studies, are inspired by them, or may realize when they are repeating the discussions and even mistakes of previous players.

From the practitioner’s point of view, the inconsistencies, unclarities, and the like of the essay do not matter. They do not care. None of the recommendations that the critics have made, none of the most persuasive deconstructions made one iota difference. Put all the paper work that followed the essay through the grinder and it would not matter. One difference is the preoccupation with the Truth in all this literature. Friedman, the pragmatist, does not care about Truth. Neither do his fellow practitioners. They want to know what works.

The Irony of Unintended Consequences

In one sense Friedman lost the battle, rather badly. For the strategy that he advocates in the essay as the only right one, a Marshallian strategy in opposition to a Walrasian one (as he would further elaborate in a latter strategy), turned out to lose. His pragmatically minded and empirically oriented strategy would succumb to Samuelsonian economics. Samuelson lost the methodological debate, but his scientific approach won out.

Lipsey and Steiner were followers of Friedman. But their textbook lost from the one by Samuelson and clones thereof. Compare Coase and Stigler, two versions of Marshallian economics in Friedman’s style, both losers in method (except among a relatively small number of Chicago/UCLA/Virginia students of the old type).
We here break off in the drafts, intending to make a few more points:

• 
  Comparable articles: Hicks IS-LM, Muth, Becker "Time is Money," Coase 1960,
  
• 
  Friedman from 1943 to 1945 was a statistician for the Statistical Research Group of the Division of War Research at Columbia University (there is still a non-parametric test named after him). Listen to his experience with statistical vs. substantive significance:
  

One project for which we provided statistical assistance was the development of high-temperature alloys for use as the lining of jet engines and as blades of turbo superchargers---alloys mostly made of chrome, nickel, and other metals. . . . Raising the temperature a bit increases substantially the efficiency of the turbine, turbo supercharger, or jet engine . . . . I computed a multiple regression from a substantial body of data relating the strength of an alloy at various temperatures to its composition. My hope was that I could use the equations that I fitted to the data to determine the composition that would give the best result. On paper, my results were splendid. The equations fitted very well [note: statistically; with high R²] and they suggested that a hitherto untried alloy would be far stronger than any existing alloy. . . . The best of the alloys at that time were breaking at about ten or twenty hours; my equations predicted that the new alloys would last some two hundred hours. Really astounding results! . . . So I phoned the metallurgist we were working with at MIT and asked him to cook up a couple of alloys according to my specifications and test them. I had enough confidence in my equations to call them F1 and F2 but not enough to tell the metallurgist what breaking time the equations predicted. That caution proved wise, because the first one of those alloys broke in about two hours and the second one in about three.

(Friedman 1985).

Friedman learned that statistical significance is not the same as metallurgical significance.

• 
  Samuelson’s Multiplier/Accelerator article was much more influential as a model of how to do economics than was, say, Milton's AEA presidential address (which was, by the way, extracted from him: he had no lust for theorizing).
  
• 
  Lunch-time arguments. Not the culture at Harvard. Wife, Rose, Aaron Director.
  

1 Stephen C. Levinson, Pragmatics (Camrbridge University Press 1983), p. 101. “Pragmatics” is the linguist’s name for “rhetoric.”