how should we approach the history of economic thought, fact, fiction or moral tale?

HOW SHOULD WE APPROACHTHE HISTORY OF

ECONOMIC THOUGHT,FACT, FICTION OR MORAL TALE?

BY

ROGER E. BACKHOUSE

I. INTRODUCTION

This paper reviews the way in which constructivist or anti-representa-tionalist arguments have been used as an argument in favor of changing theway we write the history of economic thought. It is argued that though sucharguments provide some important new perspectives on the subject, theiruse as a comprehensive methodological critique of "traditional" approachesto the subject rests on the theses that a non-foundationalist methodology isimpossible, and that we can assume that contemporary economics is in ahealthy state. If these theses are not accepted, the case against "traditional"histories collapses.

For students of my generation, the classic statement of alternative viewson how to approach the history of economic thought was the first chapterof Mark Blaug's Economic Theory in Retrospect (1985). The crucial choice,he argued, was between relativism and absolutism.

The relativist regards every single theory put forward in the past asa more or less faithful expression and reflection of contemporaryconditions, each theory being in principle equally justified in itsown context; the absolutist has eyes only for the strictly intellectualdevelopment of the subject regarded as a steady progression fromerror to truth. Relativists cannot rank the theories of different

University of Birmingham. Parts of this paper were contained in a paper presented to theESRC Political Economy Study Group in May 1990 and to the History of Economic ThoughtConference at the University of East Anglia in September 1990. I am grateful to participantsin both meetings for useful comments and advice. I also wish to thank Mark Blaug, RichardDavies, Chris Hookway, Philip Mirowski, Mary Morgan, Denis O'Brien, Margaret Schabas,Roy Weintraub, Donald Winch, the editor and three anonymous referees for helpfulcomments on various versions of the paper. The usual caveat applies, absolving them fromall responsibility for any remaining inadequacies.

Journal of the History of Economic Thought, 14, Spring 1992.e1992 by the History of Economics Society.

18

FACT, FICTION OR MORAL TALE? 19

periods in terms of better or worse; absolutists cannot help doingso (Blaug 1985, p. 2).

Faced between the choice of Werner Stark (1944) and Leo Rogin (1956) onone side, or Mark Blaug and J. A. Schumpeter (1954) on the other, therewas no contest! The notion that the history of econorpic thought wasconcerned with evaluating ideas received a further boost from the spate ofresearch influenced, directly or indirectly, by Imre Lakatos's methodologyof scientific research programmes, with its fascinating hypothesis (hismethodology of historical research programmes) that one might look at thehistory of science in order to decide amongst competing methodologies.

From this perspective, suggestions such as those recently made byDonald Walker concerning the way in which we should approach thesubject seem unproblematic. We should judge the works of past economistsaccording to the standards of modern economic theory; we should askwhether the writings of the economists we consider were original, influen-tial and logically consistent; we should criticize economists for claiming tohave shown things which they did not show; and we should praise writers forexplaining reality well (Walker 1988, pp. 101-2).

This viewpoint, however, has recently come in for strong criticism. Whatmakes it important to take recent criticisms seriously is the fact that theyrepresent not merely a reworking of old ideas, but involve the introductionof new ideas taken primarily from philosophy and literary criticism.1 Thesecriticisms may sound very much like the relativism that Blaug discussed inthe passage quoted above, but they are not the same. The main argumentemploys the idea that knowledge is socially constructed, knowledge-claimsbeing the property of specific discourse communities, making little senseoutside those communities. The clearest statement of such a position is thatof E. R. Weintraub, who is concerned, in his Stabilizing Dynamics (1991a),to develop the notion that knowledge is constructed, not found. Thisobservation, he explains applies at two levels: at one level it means that "thework of the economist is not as well-understood as the work of an explorerand finder, a creator of theories and assembler of facts"; at another level itmeans that "histories of economics are constructed, not chanced upon"(ibid., pp. 3-4). Such a perspective, claim its proponents, leads to anapproach to the history of economic thought that is significantly differentfrom those advocated by Blaug and Walker.

A different line of argument about how the history of economic thoughtshould be studied has been taken up by Margaret Schabas (1992), who has

1. The descriptions 'new' and 'old' refer to newness within the history of economic thought.Many of the ideas have a long history in other fields. For example, the view that we shouldtake "conversation" as our model goes back at least to Oakeshott and probably a long waybefore him. Even within the history of economic thought it can be argued that some of theseideas came in via Thomas Kuhn's Structure of Scientific Revolutions (1970).

20 JOURNAL OF THE HISTORY OF ECONOMIC THOUGHT

argued that historians of economics "should learn to embrace their histori-cal sensibilities," leaving the task of modeling to economists, and reducingtheir reliance on philosophy. Though her focus is on the appropriateinstitutional affiliation for historians of economic thought, completelyavoiding the methodological questions that Weintraub addresses, herappeal to develop a greater historical awareness implies a certain criticismof the way the subject is usually approached. It has much in common withWeintraub's claim that "the most subversive feature of More Heat than Light[Mirowski 1990] is its historiographic stance, one which denies the separatedisciplinary status of the history of economic thought" forcing "economiststo become historians first, economists second" (Weintraub 1991a, pp. 1-2).2

Weintraub's purpose is to argue for a particular way of doing history, hisinspiration coming from the sociology of science and, above all, fromliterary criticism. His argument about knowledge being the product ofspecific discourse (interpretive) communities has, however, much in com-mon with Richard Rorty's views on how the history of philosophy should bewritten. Rorty's position is worth considering here for two reasons. Thefirst is that although it is based on arguments similar to those used byWeintraub, Rorty takes the argument significantly further, providing amore comprehensive critique of certain types of history. The second is thatRorty's classification of four genres has been taken up by Philip Mirowski(1987) and Blaug (1990, 1991), but has not received adequate criticalattention.

The purpose of this paper is to provide a critical examination of theseclaims.

II. THE CASE FOR "CULTURAL" HISTORY3

The constructedness of economic knowledge. Economic knowledge is,like any other knowledge, socially constructed — it is produced by specificcommunities. Weintraub therefore takes as his starting point "the fact thateconomics is a social activity and it is carried out, is done, in communities"(Weintraub 1991a, p. 6). Quoting the literary critic Stanley Fish, he arguesthat such interpretive communities are made up of "those who shareinterpretive strategies, not for reading but for writing texts, for constituting

2. On reading such remarks one wonders whether it was prescience that caused BernardCorry (1975, p. 252) to write the following paragraph: "Rumor has it that somewhere outWest (Mid?) there is a township of Eeonsvilie which announces to the visitor 'Abandon HopeAll Ye That Enter Herein.' The story has it —although even I doubt the substance — that abit further on we hit Hopesville, which proclaims (although few listen) 'Abandon EconomicsAll Ye Who Enter Herein.'"

3. The term "cultural history" is being used in Passmore's sense, discussed below. Theprecise term used is not important; the terms "historical reconstruction" (Rorty 1984) or"thick history" (McCloskey 1988) could equally well have been used.


their properties" (Fish 1980, p. 14; quoted in Weintraub 1991a, p. 6). Thesignificance of this definition is that it marks out a position in whichmeaning is neither subjective (where the reader is free to determine themeaning of a text) nor objective (where meaning is inherent in a particulartext). The community creates meaning, for it is the community whichdetermines the rules according to which texts are written and interpreted.Continuing Weintraub's quotation from Fish,

[Interpretive] strategies exist prior to the act of reading and there-fore determine the shape of what is read rather than, as is usuallyassumed, the other way round... [But] an interpretive communityis not objective because as a bundle of interests, of particularpurposes and goals, its perspective is interested rather than neutral;but by the same reasoning the meanings and texts produced by aninterpretive community are not subjective because they do notproceed from an isolated individual but from a public and conven-tional point of view.... [M]embers of the same community willnecessarily agree because they will see (and by seeing, make)everything in relation to that community's assumed purposes andgoals; and conversely, members of different communities willdisagree because from their respective positions the other "simply"cannot see what is obviously and inescapably there (Fish 1980, pp.14-16; quoted in Weintraub 1991a, p. 6).

In place of the dramatic choice posed by Blaug, Weintraub, following Fish,seems to have found a third way.

This view of knowledge as being constructed by particular communities'interpretive strategies has, it has been claimed, implications for the way weview economics, for different economists see the economy in differentways. "Where Marx 'saw' a falling rate of profit, Clark 'saw' a marginalproduct of capital.... Where Solow 'sees' a market failure, Lucas 'sees'rational competitive activity" (Weintraub 1991a, p. 150). Claims abouteconomic phenomena (such as that the economy is in a stable equilibrium)can be understood only as imposing a discursive order on phenomena.When an economist makes such a claim he or she is simply inviting others"to participate in a (Wittgensteinian) language game to be played accordingto a particular set of rules" (ibid.). It thus makes no sense for us to befalsificationist in our economic methodology.4

Implications for the history of economic thought. If we adopt such aperspective on economic knowledge, it has enormous implications for theway we study the history of economic thought. As Weintraub puts it, "If we

4. This claim is explored in detail in Backhouse 1991. In that paper it is argued that thereare ways in which methodology, even falsificationist methodology, can be defended againstthese criticisms.


are not falsificationist in our conceptualization of economics, we are under nocompulsion to remain falsificationist about our histories of economics" (1991a,p. 4; emphasis in original). Like the economics which forms their subject-matter, histories of economics are constructed —they are not there to befound. "We must accept that history is not presented to us raw, as a neutralcase or data source, upon which we may perform tests of our methodologi-cal theories or how scientific knowledge is gained. History is not 'out there'waiting to answer our questions or corroborate our hypotheses. History isnot found. Instead, history is written, and is itself as much a creativeenterprise as is the 'theory' it is often 'meant' to describe" (Weintraub1991a, p. 4).

The notion that everything we see is constructed extends even to histori-cal texts themselves. Texts are not "fixed," for they cannot be read withoutbeing interpreted. This carries the implication that even relatively straight-forward issues, such as the attribution of priority, involve interpretive actson the part of the historian: "Historians construct by providingreinterpretations of apparently 'fixed' texts. So too 'priority' is constructed,not discovered and then awarded" (Weintraub 1991c, p. 385).

There is, therefore, no basis for "Whig" histories, as advocated bySamuelson (1987) and Walker (1988) in which "history must be a moralexemplar showing how scientists came to get it right, eventually, where theright stuff is called Truth" (Weintraub 1991a, p. 149). Not only is it notpossible to know what is the Truth, but even the notion of Truth isproblematic. Truth, Weintraub argues, can be seen only in terms of faithfulrepresentation of the true state of the economy, but all we can possibly haveis constructions. Referring to economists seeing different things in theeconomy, he claims "such 'seeing' is a construction, not a representation inthe sense of a discursive act that is closer to the truth than any otherconstruction. For me, I cannot confidently appreciate what such an idea of'representation' might mean" (Weintraub 1991a, p. 150).

Taking history seriously. The conclusions Weintraub draws for the historyof economic thought are straightforward. He argues that the constructednessof all economic knowledge implies that it makes no sense to view the historyof economic thought through methodological spectacles. The only way towrite history is to treat appraisal as a complex process. "We have to askmore complex questions of a theory and its interpretations: How was itdeveloped? How was it presented? What do its terms mean? Who was itsaudience? We seek to understand the way the interpretive community hasread the economy text, and what makes the community more likely torespond to one interpretation rather than another" (Weintraub 1991a, p.

7).In his history of the analysis of stability of general equilibrium, therefore,

he is concerned not with documenting progress, in the sense of the


achievement of better, more general or more rigorous proofs, but withquestions such as how economists' understanding of concepts such asequilibrium and stability evolved over time. Appraisal, understood infalsificationist terms, is not on the agenda. This approach is similar to thatof Dorothy Ross, who describes herself as writing: "that kind of intellectualhistory which seeks to reconstruct the discourse within which social scien-tists worked. I understand discourse as conversation, developed over time,centering around certain problems, setting the terms of discussion for thosewho enter into it, and at the same time responding to the differentintentions of participants" (Ross 1991, pp. xviii-xix).

In such a historical reconstruction or cultural history, it is clearly vital totake account of developments outside economics, whether these are doc-trines of American exceptional ism (Ross 1991) or energetics (Mirowski1990).5 It explains Weintraub's opinion that Mirowski's work is "nothingless than a redefinition of what might constitute a convincing account of somefeature, or other, of economic analysis" (1991b, p. 3, emphasis in original).To borrow an expression from Donald McCloskey, Weintraub is saying thatafter Mirowski, the old stories no longer have satisfactory endings(McCloskey 1990).

Historians of economic thought, argues Weintraub, have been "lookingfor love in all the wrong places" (Weintraub 1991a, p. 11). Accepting theconstructivist position, he claims, means that we can afford to lose interestin methodology, and that we should "give up the wish for approval fromphilosophers of science." Similarly, Schabas (1992) criticizes historians ofeconomics for persisting in a "longlasting love affair with Kuhn, Lakatosand Popper" long after mainstream historians of science have turned theirattentions elsewhere. She attributes this partly to "physics envy" but also toa preoccupation with modeling and theory. Her conclusion is that weshould embrace our historical sensibilities, leaving the task of modeling toeconomists, and that we should "break away" from economics. She believesthat the benefits of self-sufficiency (perhaps in the company of historians ofscience) would be worth the cost.6

5. Whether More Heal than Light is best viewed as historical or rational reconstruction isa moot point, and in any event beyond the scope of this paper.

6. There are also institutional factors underlying the suggestion that historians ofeconomics should break away from economics departments. Here, however, we are con-cerned only with the intellectual case. It is worth noting at this point that Weintraub arguesthat historians of economics should not jump out of the arms of philosophers into those ofhistorians of science. As this sentence implies, Schabas is noncommittal on the question,though she clearly sees advantages in cooperation with historians of science.


III. RORTY'S CASE AGAINST DOXOGRAPHY

Four genres in the history of philosophy. Mirowski (1987) and Blaug (1990,1991) have urged us to think of the history of economic thought in terms ofRichard Rorty's four genres (Rorty 1984): (i) historical reconstructions,(ii) rational reconstructions, (iii) Geistesgeschichte and (iv) doxography.This classification is similar to one proposed nearly two decades earlier byJohn Passmore (1965), who distinguished between (i) polemical, (ii) cul-tural and (iii) elucidatory histories, subdividing the last of these into threesub-types: (iiia) doxographical, (iiib) retrospective and (iiic) problematichistories. Though our main focus is on Rorty, reference to Passmore's moreprecise classification helps clear up some of the ambiguities in Rorty'sdefinitions, particularly in his catch-all category of doxography.

Historical reconstructions or cultural histories are accounts of the work ofdead philosophers which set them in the context of their own times. Thisis what McCloskey (1988) describes as "thick" history. Though they mightdissent from some of the details of Rorty's account, the type of historyadvocated by Weintraub and Dorothy Ross fits in here. Rational reconstruc-tions or retrospective histories are accounts of past philosophers' ideas whichview them in the context of contemporary philosophy. Geistesgeschichte7

are larger, more sweeping stories than either historical or rational recon-structions. They are concerned with problematics rather than with solu-tions to problems, being histories which seek "to justify the historian and hisfriends in having the sort of philosophical problems which they have"(Rorty 1984, p. 57). They seek to justify particular views of what philosophyis, and are concerned with questions such as: Why should anyone havemade such-and-such a question central to his thought? or Why did anyonetake such-and-such a problem seriously? (ibid.).8

Doxography is defined by Rorty as "the attempt to impose a problematicon a canon drawn up without reference to that problematic, or, conversely,to impose a canon on a problematic constructed without reference to thatcanon" (ibid). It is, he argues, "exemplified by books which start fromThales or Descartes and wind up with some figure roughly contemporarywith the author, ticking off what various figures traditionally called 'philoso-phers' had to say about problems traditionally called 'philosophical'"(Rorty 1984, p. 62).

7. The nearest we can get to an English translation is "histories of the spirit." The word"Geist" evokes G. Hegel's work, which is paradigmatic of this genre (Rorty 1984, p. 56).

8. Although we equated Passmore's retrospective history with Rorty's rational reconstruc-tion, it has something in common with Geistesgeschichte, for retrospective histories fiteverything into a continuous pattern "as if a particular philosophy had been from thebeginning of time the 'one far-off divine event, to which creation moves'" (Passmore 1965, p.23). Though it can be used with care, the distinction between rational reconstructions andGeistesgeschichte is not clear cut.


As defined by Rorty, however, doxography would seem to cover three ofPassmore's categories. Polemical histories use a discussion of historicalfigures to make certain philosophical points. They arise out of a particulartradition in the teaching of philosophy, and are neither histories norattempts to advance our knowledge of the history of philosophy (cf.Passmore 1965, p. 7). We can agree with Passmore that these are nothistories, but teaching tools; they need not detain us. Next comedoxographicalhistories, which "are concerned to 'tell us what actually happened' withinphilosophy" (Passmore 1965, p. 18). In them, "the major conclusions ofphilosophers are concisely set out, generally to the accompaniment of acertain amount of biographical detail. The chronological framework isprovided by the idea of a 'succession,' or a school, alleged to indicate a lineof influence" (ibid., p. 19). They contain basic information but give us nounderstanding (ibid., p. 22). Finally, there are problematic histories,Passmore's favoured genre. Philosophers, he argues, do not set out toconstruct systems; they set out to solve problems. Thus to understand aphilosopher's work we must retrace the steps by which he or she solved, orsought to solve, certain problems. The history of philosophy can thus beseen as a series of attempts to solve particular problems.

The critique of doxography in the history of philosophy. Rorty considershistory not as an end in itself, but rather as serving particular purposes.Historical reconstructions make us aware of the contingency of humanideas. Rational reconstructions enable us to see progress in history. Thelatter may be anachronistic, but that is acceptable provided that we areaware of what we are doing. Geistesgeschichte tell us what is important inphilosophy and who are the important philosophers. But why should this benecessary? Rorty suggests four answers. We need to believe that we arebetter off by becoming aware of new problems; the authors of historical andrational reconstructions need to know that they are discussing the majorfigures; philosophy has an important honorific use, so we need to know onwhom to confer the title of "philosopher"; we need heroes. The genre ofGeistesgeschichte fulfills these needs.

This leaves doxography which, Rorty claims, should be allowed to witheraway: it "inspires boredom and despair"; its existence is "calamity itself, andnot the mere risk of it"; and the works involved unintentionally mummifyand "decorticate" the thinkers they discuss (Rorty 1984, p. 62). There aretwo main features of doxography to which he takes such exception. The firstis that it is unimaginative. It never alters the canon in the way Geistesgeschichtedo (ibid., p. 67).9 It eschews the question of how we are better or worse offthan our predecessors (ibid., p. 72). The chapter headings of doxographies,

9. His comment that experimental alterations of the canon make it possible to avoiddoxography suggests that he regards this as a necessary characteristic of doxography.


he argues, are known in advance. Certain chapters are missing for lack ofmaterial, and important thinkers are skipped over because they had nothingto say on the "central problems" (ibid., p. 65). He suggests that doxographyis badly-done, perhaps simplistic, history, as when he claims that "the realtrouble with doxography is that it is a half-hearted attempt to tell a new storyof intellectual progress...half-hearted because it lacks the courage to re-adjust the canon to suit the new discoveries" (ibid., p. 63). These areobjections that apply to textbooks; that is, to Passmore's doxographicalhistories.10 The second feature is that doxography is based on "the idea that'philosophy' is the name of a natural kind — the name of a discipline which,in all ages and places, has managed to dig down to the same deep,fundamental questions," implying that anyone identified as a "great phi-losopher" has to be described as studying these questions (ibid., p. 63). Thisis a more fundamental criticism of doxography. Starting from the positionthat knowledge is the property of specific discourse communities, Rorty isdrawing the conclusion that there are no continuing, clearly defined, central

philosophical problems. Different communities construct different prob-lems.

The argument so far. Rorty's evaluation of alternative types of history isbased on his pragmatist perspective. His approval of historical reconstruc-tions, rational reconstructions and Geistesgeschichte is based on his assess-ment of their usefulness to the philosophical community. Similarly, herejects doxography on the grounds that it contributes nothing. Oneresponse, therefore, is to take issue with his conclusions on pragmaticgrounds. One might argue, for example, that he underestimates theimportance of "unimaginative" histories for teaching, or for the processwhereby knowledge is consolidated. Alternatively, it is arguable that thecategory of doxography is too loose to be usable, the boundaries betweenrational reconstructions, Geistesgeschichte and doxographies being veryunclear. The most important aspect of Rorty's critique, however, is hisargument that there are no continuing central problems of philosophy. Itis the key argument against problematic histories, and rests on the notionthat knowledge pertains only to specific communities. This is the argumentthat Weintraub uses to criticize Whig history. We turn to it in the nextsection.

IV. NON-CONSTRUCTIVIST HISTORIOGRAPHY

The possibility of a non-foundationalist epistemology." In Rorty's writing

10. A good example of the histories to which such objections apply is the history of thetheory of the consumption function found in most macro textbooks.

11. A more comprehensive treatment of this question is provided in Backhouse 1991.


the reader is presented with a stark choice: either accept that knowledge issocially constructed, and abandon any attempt at epistemology; or adopt afoundationalist approach, according to which people have a secure, com-pletely reliable, privileged source of knowledge that can be used to passauthoritative judgement on the truth or falsity of knowledge claims. As hehas put it, "no description of how things are from a God's-eye point of view,no skyhook provided by some contemporary or yet-to-be-developed sci-ence, is going to free us from the contingency of having been acculturatedas we were" (Rorty 1991, p. 13). Though expressed in different language,many of the critics of economic methodology offer us just this choice.McCloskey invites us to choose between rhetoric and modernism, definingthe latter as "the notion that we know only what we cannot doubt"(McCloskey 1986, p. 5). Similarly, Mirowski attacks the quest for certaintyas the "Cartesian Vice" (Mirowski 1988, p. 119), whilst Weintraub refers toepistemology being concerned with "refuting the sceptic" (Weintraub1991a, p. 150). Faced with such a choice, who could resist Rorty's position?

This is, however, a false choice. Methodology does not have to be basedon indubitable foundations. The Popperian approach, for example, isemphatically not Cartesian, for it involves accepting that we do not have anycertain knowledge, and that all knowledge is provisional. For Popper, asmuch as for Rorty, all knowledge ultimately rests on a conventionalfoundation. "There are no uninterpreted 'data'; there is nothing simply'given' to us uninterpreted; nothing to be taken as a basis. All our knowledgeis interpretation in the light of our expectations, our theories, and istherefore hypothetical in some way or other" (Popper 1983, p. 102). An-other example is the literature on the sociology of science, where, thoughthe emphasis is frequently on how scientists construct theories, it is recog-nized that there are certain things that scientists cannot do (see, forexample, Bazerman 1989, or Collins 1985). There are empirical constraintsunder which scientists are working. Such constraints may not give us a"God's-eye view," but they may be a sufficient basis on which to constructa methodology. Hacking (1983), for example, uses experimentation as thebasis for a form of scientific realism. These examples are enough to makethe point that there are alternatives to the anti-epistemological position ofRorty and Fish.

The need for evaluative histories of economic thought.n Rorty argues thatit is inappropriate to ask whether a discipline is in good shape:

12. The term "evaluative" is used, not to denote a genre additional to those distinguishedby Rorty, but as a shorthand for the approaches to the history of economic thought whichevaluate the economics of the past according to criteria the writer believes to be appropriate.It may cover rational reconstructions, Geistesgeschichte, and even some types of what Rortyterms doxography.


Those who wish to reduce objectivity to solidarity —call them"pragmatists" — do not require either a metaphysics or an episte-mology. They view truth as...what is good for us to believe....Insofar as a person is seeking solidarity, she does not ask about therelation between the practices of the chosen community andsomething outside that community.... [Pragmatists do not see] thegap between truth and justification as something to be bridged byisolating a natural and transcultural sort of rationality which can beused to criticize certain cultures and praise others (Rorty 1991, pp.21, 22-23).

One appraises economic theories only in relation to the standards and aimsof the community of economists. One does not introduce "outside" criteriawhich might bring those standards into question. It is compatible withMcCloskey's view that the "economy of intellect [is] running just fine"(McCloskey 1986, p. 28).

If we can do nothing other than assume that contemporary economics isin good shape, we can safely abandon not only methodology, but also theidea of appraising the history of economic thought. Many economists,however, have argued that there are serious problems with contemporaryeconomics — that we cannot assume the discipline to be in good shape.13 Ifwe accept that there are problems with contemporary economics we canhardly avoid the task of inquiring into the nature of the malady andexploring possible cures. This, almost by definition, involves methodology.It is then a short and defensible step from appraising contemporaryeconomics to appraising past economics — indeed, where does the history ofthought begin?14 Though it is important not to treat past ideas anachro-nistically, it could be argued that we have little choice but to use the historyof economic thought as a testing-ground for methodology. If we avoidmethodology, we avoid evaluation.

This view of the history of economic thought requires that the subjectretain its links with both philosophy and economics. The argument forbreaking away clearly has its attractions. Historical sensibilities are impor-tant, and there would be great freedom to be derived from studying thesubject simply for its own sake; and Weintraub and Schabas may even beright when they say that historians of economic thought have paid too muchattention to philosophers. However, to jump from this to a position thatrules out, virtually a priori, any application of ideas from another disciplinewould seem to be going too far. Physics may not be the only discipline worthtaking into account —Schabas is almost certainly right, for example, toreprimand historians of economics for not paying more attention to psy-

13. See, for example, Colander 1990 and Leonlief 1970. See also Backhouse 1993.14. See, for example, Walker 1988.


chology —but it has undoubtedly had great successes. Any claim thateconomists have nothing to learn from physics has to be argued on historicalgrounds, on the basis of detailed arguments about the nature of the twodisciplines.

The argument that the history of economic thought should becomeautonomous, studied for its own sake along lines such as those suggested byWeintraub and Schabas, therefore, pays insufficient attention to the criticalrole the discipline should play within economics.15 Economists may notwish to listen to methodologists and historians of economic thought.Sometimes this will be for good reasons, but on other occasions, lessonsfrom the history of economic thought may be important. If we turn ourbacks on evaluative histories, the usefulness of our histories will be severelylimited.

The case for problematic histories of economics. Of the genres discussedin section III, the one against which the constructivist criticisms are mostclearly aimed is that of problematic history.16 These are histories concernedwith what Schumpeter called "the filiation of economic ideas" (1944, p. 6).Such accounts may appear to be doxography, for they may discuss what eachof a series of writers has to say on a particular problem. It is assumed thatthere is a particular problem (such the theory of value), and that each of thewriters considered has something to say on this particular problem. Tojustify such accounts we need to defend the notion that there can becommon problems in the history of economics.17 Once we do this, thedistinctions between Rorty's various genres become blurred.18

Although some very sweeping claims have been made concerning theimplications of the constructivist perspective, Rorty himself is more mod-est. First, he argues that it may indeed be possible to find common problemsover limited periods of time (say a century or two). It may make sense, forexample, to write "a story of the steps which led from Descartes to Kant"(Rorty 1984, p. 65).19 Second, Rorty claims only that his arguments applyto philosophy, suggesting that they may not apply to "science." Specifically,he argues that Geistesgeschichte are not needed in science, for four reasons:

15. This is similar to the conclusion reached by Walker (1992).16. We leave aside the question of how far this genre overlaps with rational reconstructions

and geistesgeschichte.17. Insofar as there are common problems, it will be valid to use contemporary concepts

in historical reconstructions. See Blaug 1990, pp. 28-9.18. To say this is not to say that there will always be common problems. It is to say that

the question of whether or not there are common problems can be settled only in particularcases and on the basis of textual and other evidence. Similarly, it is not being claimed thatunimaginative, doxographic histories do not exist; of course they do, but this should not bemuddled up with historiographic arguments.

19. Interestingly, this example covers about 175 years, comparable with the period fromthe Wealth of Nations to the 1950s.


science does not need the legitimation provided by Geistesgeschichte; canonformation is not a problem in science; there exist uncontroversial parts ofany science, which can be taken as the goal towards which the history ofscience leads; and there is no need to ask whether, say, contemporarybiochemistry is in good shape (Rorty 1984, pp. 57-59). This contrast isimportant not because such a sharp contrast between philosophy andscience is defensible,20 but because it makes the important points thatdisciplines may differ, and that the way in which the history of a subject istold should depend on the nature of the discipline concerned. This bringsus back to methodology, for we cannot discuss the nature of a disciplinewithout discussing methodological issues.

Of particular importance is the role of empirical evidence. Whereempirical evidence imposes constraints on the economic theories that canbe constructed, the existence of perennial "real world" economic problemsmay explain the persistence of certain problems in economic theory. Forexample, one might argue that it is reasonable to write a history of thedifferent ways in which unemployment has been explained. That is a clearexample of problematic history. In such a history one might stress the roleof empirical evidence. The fact that the concept of unemployment isconstructed and has meant different things in different periods, does notinvalidate such an approach. Indeed, the very purpose of such a history maybe to show how the meaning of the concept has changed over time.21

The importance of methodological questions concerning the role ofempirical evidence will clearly vary according to the branch of economicsbeing considered. It would appear to be no accident that the first consis-tently constructivist history of economic thought is concerned with generalequilibrium theory. Weintraub found he was able completely to dispensewith methodology in Stabilizing Dynamics. In part this reflects the questionsto which he wanted to find answers. It is, however, natural to wonderwhether it would have been as easy to avoid appraisal had he beeninvestigating a branch of applied economics, or a branch of economicsclosely tied up with economic policy-making. General equilibrium theoryhas been criticized for having become little more than a branch of puremathematics distinguished by its origins, so it is hardly surprising if ques-tions of empirical testing, falsifiability, empirical progress and so on proveto be irrelevant. It can equally be argued that the literature on the stabilityof general equilibrium is irrelevant to most economic investigation.

20. There are many people who would take exception to his characterizations of bothphilosophy and science.

21. See, for example, Backhouse 1988, chapter 5.


V. CONCLUSIONS

This paper has been critical of the constructivist approach to the historyof economic thought. In making these criticisms it has been taken forgranted that the constructivist perspective on the history of economicthought has much to contribute. It forces us to ask new questions and tolook at history in a new light. In Stabilizing Dynamics, for example,Weintraub has provided an example of the insight that can be obtained byadopting such a perspective. What this paper has taken issue with is theclaim that the philosophical arguments underlying that perspective invali-date other ways of writing the history of economic thought. This impliesthat we need many types of history, including not only cultural histories, orhistorical reconstructions, such as Weintraub and Ross wish to write, butalso the evaluative histories that Walker advocates.22 Such a pluralistposition is compatible with the view expressed by Rorty: "There is, in ourview, nothing general to be said in answer to the question 'How should thehistory of philosophy be written?' except 'As self-consciously as one can —in as full awareness as possible of the variety of contemporary concerns towhich a past figure may be relevant'" (Rorty, Schneewind and Skinner1984, p. 11).

Such a pluralistic approach, recognizing the many purposes served byhistories of economic thought, means we require not only "thick" histories,rich in detail, but also "thinner" histories, which seek to see how far certainaspects of the historical record can be explained in terms of a few boldgeneralizations. McCloskey is certainly right in saying that philosophy ofscience is thin compared with the history of science, and it is easy to lampoonthe philosophy of science on these grounds.23 But that is not the point. Inthe sense that a good theory abstracts from irrelevant detail to focus onimportant relationships, Friedman was right to say that unrealisticness is asign of a good theory. The test of a theory is not descriptive accuracy, butwhether or not it works according to whatever criteria we are employing. Ifour theories become too thick, therefore, they will lack power, an issuewhich has emerged in debates over the merits of analytical versus narrativehistory (see, for example, Stone 1979; Dray 1985).

The question may be raised as to whether it is right to bracket thehistorians of economics we have discussed with Rorty and with the philo-sophical position that this implies. Surely, it might be argued, those writersare arguing simply that we should take history seriously, opposing simplistichistories such as those written by economists who uncritically accept a few

22. Blaug (1990) uses Rorty's categories to make this point, stressing the importance ofbeing clear about what type of history we are writing.

23. In just the same way we can ridicule neoclassical economic theory on the grounds thatits models are hopelessly simple compared with real economies.


ideas from the philosophy of science in the light of which they fit someepisode of economics into a "reconstruction." In this light, what we havepresented as arguments for cultural history are simply arguments to takethe task of,writing history seriously: to.pay attention to evidence, toalternative sources and so on.

Of the historians discussed in section II, the only one who completelyavoids entanglement in epistemological issues is Schabas. Her appeal tohistorians of economics can be seen simply as an appeal to do betterhistory.24 The suggestion that historians of economics should move out ofeconomics departments, however, implies that they should distance them-selves from the concerns of contemporary economists. Their role vis-a-visthe rest of economics would clearly be different. Next comes Ross, whoseemphasis is still very much on doing history, her sources of inspiration beingother historians of the social sciences rather than philosophers, but whoadopts a view, like Rorty's, of discourse as a conversation. At the otherextreme is Weintraub.25 Stabilizing Dynamics is not only a history of abranch of economics. It is also a manifesto for a particular way of doinghistory, grounded explicitly on a particular view of truth.26 Furthermore, hemakes it clear that in criticizing Walker and Paul Samuelson he is criticizing"the history of economic thought as usually presented" (1991a, p. 149,emphasis added). He may not buy Rorty's typology, but the epistemologicalarguments he uses are the same.

Constructivist historiography h,as been presented as a radical new depar-ture, subversive of traditional approaches to the history of economicthought. It is important, however, not to exaggerate the extent to which itis a new departure. Many of the issues raised, such as the theory-ladennessof facts, have long been recognized.27 In addition, constructivist historyneed not look very different from conventional histories. Weintraub forexample, stresses the constructed nature of priority, and argues that "we

24. Indeed, A World Ruled by Number (Schabas 1991) shows that it is possible to combinehistorical sensibility with the willingness to appraise for which Weintraub criticizes Walker.

25. It is tempting to place Mirowski alongside Weintraub, for he has explicitly endorsedRorty's critique of doxography, and he is hostile to traditional historiography. However, hispublications contain few explicit historiographical statements.

26. Weintraub docs stale that he has no interest in recapitulating the arguments for aconstructivist point of view, and that he hasonly a residual interest in philosophy. On the otherhand, he argues that there is nothing productive to be said aHouHhe connections betweentheory and reality, about the truth of theoretical propositions (1991a, p. 110); that he cannot"confidently appreciate" what it would mean to say that one discursive act is "closer to theTruth than any other construction" (ibid., p. 150); that statements such as "the economy is notstable" are simply invitations "to participate in a (Wittgensteinian) language game to beplayed according to a particular set of rules" (ibid.); that models, theories, stories and data"do not 'describe' except as they create that which is to be described" (ibid., p. 151).

27. See the essays reprinted in Blaug 1991.


may, of course, reconstruct the past as we wish" (Weintraub 1991c, p. 395).This does not preclude him from conducting an investigation into the pastrecord from which he concludes, unambiguously, that it is wrong to awardpriority to Allais for using a Liapunov function. What Allais used, Weintraubconcludes, is not "that modern proof at all" (ibid., p. 389); "The issue is, forassessing 'priority,' whether Allais's argument actually used a Liapunovfunction. Unambiguously the answer is no" (ibid., p. 393). There is nosuggestion here that we are free to interpret the historical evidence as wewish, or that different discourse communities might construct these factsdifferently: there are certain statements that are not allowable.28

In this paper it has been argued that the epistemological arguments onwhich constructivist historiography is based are not compelling, and that ifhistorians of economic thought lose interest in methodology and theappraisal of economic ideas, the importance of their studies will be greatlyreduced. Some accounts of the principles on which the history of economicthought were constructed may have been expressed in excessively positivistterms, and it may be that some need to be rewritten. However, to argue thatsuch accounts must be abandoned in favour of a perspective that precludeslooking for progress in economics, or investigating the health of contempo-rary economics, is to throw out the baby with the bathwater. Facts may beproblematic, but to treat everything simply as fiction is to abandon whatshould be one of the main tasks facing the historian of economic thought:to draw conclusions about the merits of different approaches to the studyof economic phenomena.

REFERENCES

Backhouse, Roger E. 1985. A History of Modern Economic Analysis,Blackwell, Oxford.

. 1988. Economists and the Economy, Blackwell, Oxford.

. 1991. "The constructivist critique of economic methodology,"University of Birmingham, Discussion Papers in Economics, no. 91/12.

. 1993. "Rhetoric and methodology," in R. F. Hebert, ed. Perspec-tives on the History of Economic Thought, 9, Edward Elgar, Aldershot.

Bazerman, Charles. 1988. Shaping Written Knowledge, University ofWisconsin Press, Madison.

Blaug, Mark. 1985. Economic Theory in Retrospect, Cambridge UniversityPress, Cambridge.

28. It is tempting to argue that constructivist historiography can have no consequences forthe writing of history, in the same sense thai Weinlraub has argued that methodology can haveno consequences for economic arguments (Weintraub 1989). It can, of course, like method-ology in Weinlraub 1989, have consequences in the sense that anything can have conse-quences if it makes people change their behavior.


. 1990. "On the Historiography of Economics "Journal of the Historyof Economic Thought, 12 Spring, 27-37.

. 1991. "Introduction" to The Historiography of Economics, Vol. 1 ofPioneers in Economics, Edward Elgar, Brookfield, VT., and Aldershot.

Colander, David. 1990. Why Aren'tEconomists as Important as Garbagemen?,M. E. Sharpe, London and Armonk, N.Y.

Collins, Harry. 1985. Changing Order: Replication and Induction inScientific Practice, Sage, London.

Corry, Bernard. 1975. "Should Economists Abandon HOPE?" History ofPolitical Economy, 7, no. 2,252-60.

Dray, W. H. 1985. "Narrative versus Analysis in History," Philosophy ofSocial Science, 15, 125-45.

Hacking, Ian. 1983. Representing and Intervening, Cambridge UniversityPress, Cambridge.

Klamer, Arjo, and Colander, David. 1990. The Making of an Economist,Westview Press, Boulder, San Francisco, and London.

Leontief. Wassily 1970. "Theoretical Assumptions and Non-observedFacts," American Economic Review, 61, March, 1-7.

McCloskey, Donald N. 1986. The Rhetoric of Economics, HarvesterWheatsheaf, Brighton.

. 1988. "Thick and Thin Methodologies in the History of EconomicThought" in Neil de Marchi, ed., ThePopperian Legacy in Economics,Cambridge University Press, Cambridge.

. 1990. If You're So Smart: The Narrative of Economic Expertise,Chicago University Press, Chicago.

Mirowski, Philip. 1987. Review of Backhouse 1985, Journal of EconomicLiterature, 25, no. 4,1858-59.

. 1988. "Shall I Compare Thee to a Minkowski-Ricardo-Leontief-Metzler Matrix of the Mosak-HicksType?" in Arjo Klamer, Donald N.McCloskey and Robert M. Solow, eds., The Consequences of EconomicRhetoric, Cambridge University Press, Cambridge.

. 1990. More Heat than Light, Cambridge University Press, Cam-bridge.

Passmore, John 1965. "The Idea of a History of Philosophy," History andTheory, 1964-65, Beiheft 5, 1-32.

Popper, Karl R. 1983. Realism and the Aim of Science, Hutchinson,London.

Rogin, Leo. 1956. The Meaning and Validity of Economic Theory, Harper,New York.

Rorty, Richard. 1980. Philosophy and the Mirror of Nature, Basil Blackwell,Oxford.

. 1984. "The Historiography of Philosophy: Four Genres," in R.Rorty, J. B. Schneewind and Q. Skinner, eds., Philosophy in History,


Cambridge University Press, Cambridge and New York.. 1991. Objectivity, Relativism, and Truth, Philosophical Papers, 1,

Cambridge University Press, Cambridge.Rorty, Richard, J. B. Schneewind, and Quentin Skinner, 1984. "Introduc-

tion" to Philosophy in History, Cambridge University Press, Cambridge.Ross, Dorothy. 1991. The Origins of American Social Science, Cambridge

University Press, Cambridge.Schabas, Margaret. 1991. A World Ruled by Number: William Stanley Jevons

and the Rise of Mathematical Economics, Princeton University Press,Princeton.

. 1992. "Breaking Away: History of Economics as History ofScience," History of Political Economy, 24, no. 1,187-203.

Schumpeter, J. A. 1954. History of Economic Analysis, Oxford UniversityPress, New York.

Stark, Werner. 1944. History of Economics in Relation to its SocialDevelopment, Kegan Paul, Trench and Trubner, London.

Stone, Lawrence. 1979. "The Revival of Narrative: Reflections on a NewOld History," Past and Present, 85, 3-24.

Walker, Donald. 1988. "Ten Major Problems in the Study of the HistoryofEconomic Thought," History of Economics Society Bulletin, 10, no. 2,99-115.

. 1992. "Comment on Margaret Schabas's Breaking Away: Historyof Economics as History of Science," History of Political Economy, 24,no. 1,243-45.

Weintraub, E. Roy. 1989. "Methodology Doesn't Matter, but the Historyof Thought Might," Scandinavian Journal of Economics, reprinted inSeppo Honkapohja, ed., The State of Macroeconomics, Basil Blackwell,Oxford, 263-79.

. 1991a. Stabilizing Dynamics: Constructing Economic Knowledge,Cambridge University Press, Cambridge.

. 1991b. "After Mirowski, What?," paper presented at the confer-ence on "Rethinking the history of economics after Philip Mirowski'sMore Heat than Light," Duke University, March, History of PoliticalEconomy, forthcoming.

. 1991c. "Allais, Stability and Liapunov Theory," History of PoliticalEconomy, 23, no. 3,383-96.

how should we approach the history of economic thought, fact, fiction or moral tale?

Documents