A final version appeared in the Journal of Post Keynesian Economics, 25 (4), Summer 2003, 591-598. That journal, and M.E. Sharpe Inc., have reserved the rights to the version printed there.
Reconstructing
the Past: A Response to Four Readers
by
E. Roy Weintraub
In this article the author of How Economics Became a Mathematical Science (Duke University Press, 2002) responds to four papers (from Paul Davidson, Sheila Dow, Donald Katzner, and J. Barkley Rosser) written for a symposium on his book. In addition to specific responses to points raised by the readers, the author argues, for economists, the historical commonplace that there can be no privileged perspective from which to view the past; there is no right way to write history, only more or less interesting ways.
A recent discussion of the complex interrelation between author and reader noted that:
“In contemporary scholarship, the idea that there can be no absolute separation between a body of thought and its reception has become commonplace. Long before such notions became fashionable, the philosopher and critic Walter Benjamin formulated a related insight: `The work is the death of the intention.’ Once objectified, doctrines and ideas tend to defy the will of their author, taking on a life of their own. Often, commentary and interpretation outstrip proprietary assertions of authorial intention: rarely are authors the best judges of their own work.” (Wolin, 2001, pp.6-7)
My response to the four papers has its genesis in Wolin’s observation. Given the variety of views and perspectives the commentators have about my How Economics Became a Mathematical Science, it is not only difficult to knit my own reader-responses to their various comments into one overarching discussion, but it is not obvious why anyone should particularly attend to my views on reviews of, or responses to, my recent book. Nevertheless, out of my high personal and intellectual regard for the Editor, I will respond to his request by commenting on some of the separate concerns brought forward in the four papers. The first part of my “reply” will thus simply respond to material that the readers have separately “brought to the table.” Those remarks will then serve as an introduction to a short but more focused discussion of a point that all the papers address directly or indirectly, one of particular interest to Post Keynesian readers of this journal.
Let me
begin with Paul Davidson’s question “What does one mean by truth in economic
analysis? Is `truth’ reached merely by
the use of any axiomatic theory that is consistent relative to an extended set
of axioms? If a real world observation
(e.g. involuntary unemployment) is not `true’ within this extended axiom system
. . .” I submit that my own
argument, which suggested that new views
of mathematical truths were not unrelated to larger sets of the epistemological
issues, cuts somewhat deeper than Davidson’s own answer. In particular, my discussion to which Davidson
referred asked my reader to think very seriously about what a “real world
observation (e.g. involuntary unemployment)” might possibly mean. In its strongest form, it suggests a severe
antirealist perspective in the sense that involuntary unemployment as a fact is
not “out there”, but is a construct with reference to which theories exist and
have been so embedded-intertwined that we forget that the fact was constructed
to do some specific work in the first place.
Curiously, I first fully understood this many years ago when I taught
from Davidson and Smolensky’s (1964) Aggregate Supply and Demand Analysis.
In its Chapter 15 on social accounts, written by Charles Levin, national income
accounting was developed with a metaphor of a model of the demand for parking
at an American university. In that model the definition of student is not what
“reality” would suggest, but rather the fact of “student-ness” is developed as
a construct to do specific work in the parking model. So too for involuntary
unemployment. Alex Keyssar (1986)
for instance has argued, in his prize-winning study of the history of
unemployment programs in
This is a serious and contentious matter in the science studies literatures. For example, when Latour (1988) argued strenuously that, before Pasteur, there were no microbes, he was arguing that microbes have meaning and do work only within a particular system, within a specific framework, a framework of actors and their interrelationships in what he called networks. To talk about microbes before Pasteur is thus like talking about involuntary unemployment before Keynes’s General Theory. We can do it, but we need to be extremely careful about what it is that we think we are doing. Indeed, in my own attempt to provide constructivist readings (of a non-severe variety) in the book, I tried not to speak of the economy, or the real world, that way at all in order to bring maximum clarity to my historical reconstruction.
In a similar vein, Davidson begins a detailed query with “But why, we may ask, if refutation not possible in economics …” His own argument is based on the non-ergodicity of economic data. But there is another response to the paucity of refutations in economics, one which cuts much deeper, and which, of course, does not limit itself to economics. It is that such refutations find their place primarily in the Popperian framework of crucial experiments. But that idea itself is, if perhaps not total nonsense, at least part of a project fraught with difficulty. Harry Collins (1985) for instance, in his lucid discussions of the “experimenter’s regress”, shows us how, in order that a refutation or a critical experiment be framed, we need to know precisely the conditions under which we are to accept the experiment’s conclusions. But the terms under which the conclusions are to be accepted are themselves negotiated in the experimental process. How do we know then, in a fundamental epistemological sense, that we do not have noise, or a variation in one of the experimental conditions, etc.? We have to say “Oh, those don’t matter”. But to say those don’t matter, and to seek agreement on the point, we have to have prior agreement on what matters or not, while such agreement is precisely what the experiment was designed to educe. It is not, I submit, any harder to do an experiment or to falsify a conjecture or more generally to reject or to refute a theory or argument in economics than it is in any other science. The standards for refutation are, as always, community standards. Here again, the constructivist’s message does some work that may be unpalatable to many economists: there is no appeal to rules or standards uncongenial to the community’s own standards. After all, as Davidson himself is fond of remarking pace Keynes, "In economics you cannot convict your opponent of error, you can only convince him of it.” In fact, I think Keynes rather understated the matter, for economics is no different from any other science in that respect, no different from any other field of inquiry.
Sheila Dow’s essay engages many of my arguments in a manner with which I feel great sympathy. I do want however to take up a couple of her points. Early on she notes that I refer to Phipps’s argument about the empty set as “lunacy” but suggests that this “is clearly a matter of opinion.” I have no quarrel with this, since I tried to write consistently from the belief that all such matters of right and wrong in science were negotiated locally and contingently within the appropriate communities. My point was that Phipps’ colleagues believed rejection of the idea of the empty set was lunacy, so much so that he was literally laughed at while giving such a paper at a meeting of the Southern Mathematics Association, and that was my purpose in using that term.
Late in her
paper, in her discussion of consistency, Dow runs up against a very serious
issue, one that is worth dwelling on for a moment. Noting that “even within a closed formal
system, then, it is not clear that consistency can be established” she goes on
to say that “this conclusion is unsatisfying because it leaves the whole
question of how we regard conventions completely up in the air.” Her own approach is to return to Russell and
Whitehead. I think, however, that the
matter is even more serious than such a solution suggests. Recall that in
Don Katzner in his paper moves quickly to the epistemological issue of the changing nature of mathematical knowledge and the concomitant changes in economic knowledge. He though goes on to say “While Weintraub’s attention is directed primarily to the history of how economics became mathematical, the question might be raised as to why it became mathematical. What kind of forces or motivations might have propelled economists in the mathematical direction?” As Katzner recognizes, these kinds of “why” questions were not part of my project.
Katzner’s own several answers to the question of why mathematics took hold include not only individuals seeking to exploit their superior mathematical training, but issues of the desire for scientific respectability, as well as questions of the putative certification of knowledge claims framed mathematically versus the questionable nature of knowledge claims framed otherwise. As Katzner rightly notes, this latter point is quite loaded and fraught with historical difficulty. Moreover, the question of scientific respectability is a historicized claim. Why prior to the glorification of science in the post World War II period would it have occurred to individuals to think that science was going to be the majority perspective as opposed to just another approach among a set of interesting approaches?
For myself, the “why” issue is closely bound to the question, “who is asking why?” That is, what constitutes an acceptable answer to the question “Why do economists engage mathematics in their explanatory frameworks?” in one community will differ from the answer given in another. What my own account attempted to address was the variety of answers that those engaged in the mathematization process itself would have given to the question of “why”. Thus Don Patinkin’s “why” was not Kenneth Arrow’s, which was not Gerard Debreu’s, which was not Griffith Conrad Evans’s, which was not Sidney Weintraub’s.
Several small points in Barkley Rosser’s essay deserve a response[1], although the larger set of issues that his essay raises will provide a good entrée to my more general comment. First, Rosser seems to think that I believe “economics is really not a science and therefore cannot have real paradigms or true crises.” This is a curious misreading. The concern whether economics is or isn’t a science is, as I discuss at length in my chapter nine, a reflection of some old fights about demarcation criteria that much exercised philosophers of science in the post-Popperian years, but which little engage philosophers of science these days. From a science studies perspective however, the issue becomes not one of demarcating science from non-science, but studying the elements that particular fields provide with respect to evidence, standards, socialization, argumentation, etc. I took no position on the demarcation question because I think it is a waste of time discussing whether X is or is not REALLY a science or whether economics is more like physics or more like biology or more like philology or more like professional tennis. I find such questions intellectually stultifying.
Let me move on, though, to a larger issue, one that Rosser frames well, and that seems to produce some discomfiture in other readers besides Rosser. A systematic difficulty that Rosser has with my book can be introduced with his arguments that: “[Weintraub] ultimately leaves us hanging on certain crucial points with an ambivalence that is never resolved. . . he refuses, most of the time, to discuss the superiority or inferiority of particular schools or approaches to economics. . . . Unfortunately, a central contradiction for his own analysis is that he is ultimately ambivalent regarding the most formalistic approach to mathematical economics. He both defends it even while demonstrating that the ultra-abstract and axiomatic form of mathematics known as Bourbakism has for some time now been out of favor among mathematicians themselves. His failure to tell us what the successor forms of mathematics that might be or are superior leaves us in an unsettled and somewhat as unsatisfactory position at the end of the book. . .”
This comment concerns me because Rosser should not be able to find any evidence, in anything I’ve written in the book, which amounts to my “defense” of any economics, nor evidence that in fact do I oppose any economics. What Rosser refers to as “ambivalence” is in fact nothing of the sort. My dictionary notes that ambivalence is associated with simultaneous conflicting feelings. Presumably, Rosser believes that I both like and dislike neoclassical economics, I both like and dislike Bourbakist mathematics, etc. and that this ambivalence pervades the work. But it was not my task in this book, however, to stake out and defend a position on neoclassical economics, or on mathematization. Like any particular human institution, modern economics stands there and we scrutinize it and narrate it as historians by historicizing it.
I cannot emphasize too strongly
that what I described in the last chapter as my historiographic approach was
not idle or casual. Part of the
difficulty in historical writing by economists is that most such writing is
guided by current interests. A desire to
tell a success story, or a desire to provide historical evidence for the claim
that mainstream economics is snare and a delusion, is quite pervasive. Suppose I were writing a history of the
Those who write history of science in other fields are not so burdened with strange expectations about the purpose of the history. A history of Mesopotamian astronomy would surely not be regarded as ambivalent in its account of the successes and failures of those astronomical practices. Writing a history of the mathematization of mechanics would not leave the author open to charges of ambivalence about whether or not it was a “good thing.” Why should matters be different in economics? Davidson appears wistful when he writes that “Weintraub does not give us a clue however as to when this soul searching will occur in economics and what will replace the Bourbaki scourge.” That Davidson, a leading proponent of the view that uncertainty guides practice, and that community practices are our collective attempt to shape the future, should worry that I do not tell my readers how the economics profession will behave in the future seems to me to be a bit off. Katzner does not fall into that particular pit, fully recognizing as he does that my project asks “how” not “why”. And Dow, likewise understanding the historiographic approach, asks how historical material can be utilized by non-historians, an entirely appropriate response for this symposium.
Dow’s
comment thus opens some appropriate discursive space, acknowledging as it does
that there can be no privileged perspective from which to view the past. There is no right way to write history. There are many different approaches. Most of them will engage the past
differently. One of the difficulties
faced by economists writing about the mathematization of economics, or how
economics became a science in the 20th century, is that they are
concerned to justify or to denigrate the state of affairs. Yet even some of these histories are in my
view extremely useful, even profound:
how can we turn away from the wonderful accounts by Mirowski or Ingrao
and
Collins,
H. (1985). Changing Order.
Dauben,
J. (1992). “Are There Revolutions in Mathematics?” in The Space of Mathematics.
J. Echeverria, A. Ibarra and T. Mormann.
Davidson, P and Smolensky, E. (1964) Aggregate
Supply and Demand Analysis.
Keyssar,
A. (1986). Out of work : the first century of
unemployment in Massachusetts.
Latour,
B. (1988). The Pasteurization of
Weintraub,
E. R. (1991). Stabilizing Dynamics.
Wolin, R. (2001). Heidegger’s Children.
[1] Two minor points: I must reply to the claim that Rosser makes that the
book’s chapters book were originally stand-alone articles that now have been
knit together into a volume. This is not
true. The project was a book from the
beginning (in Venice in 1992), but in order to get feedback, and to get my way
paid to various conferences and lovely European capitals to give papers,
portions of the large project got carved up in different ways at different
times. Second, in a footnote, Rosser
identifies what he calls an “outright error” in my reference to René Thom as a
Bourbakist. To claim that one is a
Bourbakist is to assert that their intellectual energies and mathematical
sensibilities were shaped, formed, and operated in conjunction with the
Bourbaki approach to mathematics.
Certainly at the time Thom won a Field medal for his work in algebraic
geometry, he was a Bourbakist. Rosser is
though correct that he later, following his work on catastrophe theory, moved
away from, and in some cases repudiated, the Bourbaki movement.