Friedrich A. Hayek:The Hayekian Prism

De Catallaxia
Révision datée du 11 janvier 2009 à 11:51 par Copeau (discussion | contributions)
(diff) ← Version précédente | Voir la version actuelle (diff) | Version suivante → (diff)
Aller à la navigation Aller à la recherche
Friedrich A. Hayek
Auteur libéral classique
« La liberté, laissée à chacun d'utiliser les informations dont il dispose ou son environnement pour poursuivre ses propres desseins, est le seul système qui permette d'assurer la mobilisation la plus optimale possible de l'ensemble des connaissances dispersées dans le corps social. »
« Laisser la loi aux mains de gouvernants élus, c'est confier le pot de crème à la garde du chat. »
Articles internes

Liste de tous les articles

Friedrich A. Hayek:The Hayekian Prism
The Hayekian Prism

Analyse de David Gordon

Hayek's Challenge: An Intellectual Biography of F.A. Hayek, By Bruce Caldwell. University of Chicago Press, 2004. Xi + 489 pgs.

Bruce Caldwell has adopted a sensible strategy to cope with the formidable task he has set himself. Friedrich Hayek was not only one of the most eminent economists of the twentieth century, whose contributions ranged from capital and business cycle theory to the socialist calculation debate, his work in economics also led him to major contributions to political philosophy, the theory of knowledge, psychology, and history. How can an intellectual biographer of Hayek cope with this vast range of work? Must his breadth of knowledge rival Hayek's own, if he aims adequately to assess his subject's contributions?

Caldwell solves this problem through selection. Rather than attempt to follow Hayek on all his peregrinations, he chooses a few issues for analysis. The reader in search of an account of Hayek's capital theory, e.g., must go elsewhere; but the benefits in careful attention to detail that Caldwell's method allows outweigh its losses.[1]

Caldwell begins with an account of Carl Menger's battle with the German Historical School. As he rightly discerns, a grasp of this controversy is essential to understanding Hayek's work. But one of his remarks surprises me. He states: "Given later developments, it is interesting that both Carl Menger and Ludwig von Mises criticized the historical approach for its inability to produce predictions" (pp.96–97). Would it not be more accurate to say that Mises believed that the Historical School wrongly sought to substitute historical research for economic theory? Austrians contend that history, taken by itself, cannot generate theory: an Austrian would hardly then criticize the Historical School for failure to achieve an impossible goal.

Our author rightly devotes attention to Max Weber, whose views on methodology were of crucial importance for Mises and his successors, Hayek foremost among them; but I think his presentation of Weber at one point misses the mark. He notes that Weber maintains that the choice of facts a social scientist elects to study depends on the values he brings to the investigation. As Weber puts the matter,

Only a small portion of existing concrete reality is colored by our value-conditioned interest and it alone is important to us. . . . We cannot, however, discover what is meaningful to us by means of a 'presuppositionless' investigation of empirical data. Rather perception of its meaningfulness to us is the presupposition of its becoming an object of investigation." (p. 89, quoting Weber)

From this passage Caldwell draws a radical conclusion: "Even 'pure observation' is always observation from a point of view; there is no such thing as 'facts in themselves'. In modern terminology, what we take to be the facts are themselves 'theory laden' in that they reflect our own prior interests" (p. 82).

This goes altogether too far. Weber says that we select facts that interest us to study. But it does not follow from this that, once one has selected some object of study, the properties of the facts disclosed to us exist only because of the theory that we hold. If someone investigates the claim "all ravens are black," he presumably does so because of his interest in this hypothesis. Nevertheless, the data on black ravens seem entirely neutral: a raven is either black or it isn't, regardless of what theory an investigator may hold. The claim that facts are "theory laden," much promoted by Thomas Kuhn, Norwood Russell Hanson, et hoc genus omne, dissolves scientific investigation into a mere subjective clash of perspectives[2]. The point is so important that it bears repetition. If you select the facts you study according to your interests, it does not follow that your interests construct the facts, to any degree whatever.

When Caldwell turns to the work of Hayek himself, he continues the pattern so far suggested: he selects an important topic, has much of value to say about it, but fails to get matters quite right.

In earlier studies, Caldwell has been much preoccupied with the a priorism of Mises; so it comes as no surprise, then, that he devotes attention to Hayek's famous article of 1937, "Economics and Knowledge." Hayek later identified a covert theme in this article, his break with Mises over the a priori. Caldwell, a strong opponent of Mises's method, argues that Hayek had even before his article distanced himself from Mises[3].

Hayek in his article restricted the sphere of a priori knowledge to the actions of single individuals. Interpersonal actions cannot be studied without resort to the empirical, because a person cannot know a priori the choices of others. But where did Mises ever say that one can? Mises's praxeology is confined entirely to the form of human action and what follows from this: it never attempts to deduce the particular choices of individuals. Mises, I suggest, was entirely on target in thinking that Hayek had not broken with him. Hayek's belief to the contrary rested on misunderstanding.

Hayek noted with surprise that Mises "took my critique silently and even approved the article as if he had not been aware that it was a criticism of his own views. I cannot explain this" (p. 221, quoting Hayek). Why did Hayek think that Mises denied that "the empirical element enters in people learning about what the other people do?" (p. 221, quoting Hayek).

The issue I have raised is of much more than biographical interest, since the nature of a priori reasoning is crucial to economics. To my amazement, Caldwell in one passage suggests that Hayek identifies a priori reasoning with introspection, when Hayek of course stands innocent of this gross fallacy. Hayek states: "We can derive from the knowledge of our own mind in an 'a priori' or 'deductive' or analytic' fashion an (at least in principle) exhaustive classification of all the possible forms of intelligible reasoning." Concerning this, Caldwell says that when Hayek "refers to knowledge that is derived '"a priori,"' he seems to mean by the phrase something like "knowledge that is gained by introspection" (Both quotations from p. 222).

Not at all! A priori reasoning concerns the logical relations among concepts. By thinking about the concept of choice, e.g., one sees that an actor will always choose his highest valued alternative. By contrast, introspection merely reports a particular person's thought. Relying on introspection, I report that I prefer vanilla ice cream to pistachio, but there is nothing a priori, deductive, or analytic about my assertion. Hayek surely did not mean anything like this. No wonder Caldwell views the a priori with misgiving: How can you gain knowledge about the world if you confine yourself to your private thoughts? Caldwell's view, I hope to have shown, is a confusion[4].

If Hayek was an epistemologist, a historian, and a political philosopher, he nevertheless always remained an economist; and Caldwell rightly calls attention to one of Hayek's final contributions to economic thought. In his 1981 talk, "The Flow of Goods and Services," Hayek renounced the notion of equilibrium. Equilibrium, Hayek maintained, was not a useful way to understand the market process; instead, Hayek suggested the metaphor, not altogether clear, of a stream. Caldwell sees the crucial point:

he rejected here even his own definition of equilibrium . . . as useful for understanding how the price system continually guides the formation of the capital stock. At a minimum, the paper suggests that, by the end of his life, Hayek's commitment to equilibrium theory as a metaphor for capturing the essence of the market process had reached its lowest point. (p. 227)

How is one to explain so fundamental a change? Equilibrium had long been at the center of Hayek's thought, a fact that Joseph Salerno has taught us better than anyone else. Now, suddenly, it is abandoned. Caldwell is puzzled. (Incidentally, I cannot think that Caldwell has adequately confronted Salerno's fundamental work on Hayek and equilibrium. See his brief and inadequate remarks at p. 143, n.14.)

I venture to suggest that had Caldwell devoted more attention to Mises, he might have found the answer to his puzzle. For Mises, equilibrium, in his terms the Evenly Rotating Economy, is a purely heuristic concept that can never be realized in practice. By no means is the market to be judged by how closely it matches this artificial construct. Is it too speculative to think that Hayek continued throughout his life to study the writings of someone who had influenced him so much in his youth? If so, perhaps Mises's doubts about equilibrium at some point hit home with Hayek. Certainly, Mises kept up with Hayek's work; and in Human Action, Mises comments at several places on Hayek's views on economic theory. (See, e.g., Mises's rejection of the Ricardo effect, famously championed by Hayek, in Human Action, Scholar's edition, pp. 767 ff.) Of course, this does not prove that Hayek read Mises; and my remarks are no more than conjecture[5].

On one point, I think, we can go further. Caldwell claims that Hayek's theory of the business cycle is "a paradigmatic example of equilibrium theory" (p. 228). If Hayek abandoned equilibrium, how could he continue to maintain, as he did so late as 1978, that the Austrian account of the cycle was in essence valid? Here Mises resolves the supposed difficulty. As he makes clear in Human Action, the Austrian account does not at all depend on the initial assumption of equilibrium, i.e., full employment of all factors of production. (See the section, "The Role Played by Unemployed Factors of Production in the First Stages of a Boom", pp.576–78.)

I have so far been rather critical of Caldwell; but he deserves full credit for his ability to explain clearly difficult concepts. To me, the high point of the book is Caldwell's expert journey through the labyrinthine complexities of The Sensory Order. He succinctly explains, e.g., the key assumptions that underlie one of Hayek's major arguments in that work:

One of the most important philosophical consequences of Hayek's psychological studies is the claim that it is impossible for the mind to explain itself. Hayek's argument hinged on the twin ideas that explanation is itself a sort of classification and that the explanation of a given classification system requires a classification system of greater complexity than the one being explained. (p. 344)

Caldwell also deserves praise for realizing that Hayek's views deserve further development: Hayek is no mere past figure whose insights have been absorbed into a supposed neoclassical consensus. But here our praise must decidedly be mixed. Caldwell gives a radically incomplete and distorted account of those who are attempting to continue the Austrian economics that Hayek championed throughout his life.

Although he mentions "signs of considerable vigor" in Austrian macroeconomics, citing in this connection the work of Steve Horwitz, Roger Garrison, Larry White, and George Selgin, (pp. 326–27) there is a rather noticeable omission in what he has to say. The name of Murray Rothbard nowhere appears. Even if one does not share my own view that Rothbard is the most important successor of Mises and Hayek, is it not strange that Caldwell devotes not a single word to this eminent Austrian? Hayek did not share Caldwell's evident opinion that Rothbard's work is insignificant. I well remember, in response to my query, his praise for Rothbard's America's Great Depression.

It is of course too much to expect Caldwell to acknowledge the contributions of Rothbard's successors, such as Hans Hoppe and Joseph Salerno. (The latter is mentioned only in a critical footnote.) Instead, our author lavishes attention on such exotica as the artificial society model Sugarscape, devised by Joshua Epstein and Richard Axtell. In this model, agents who follow simple rules generate various structures. Caldwell finds in this model analogies to Hayek's spontaneous orders and cultural evolution.

But our author ignores a crucial point. The patterns produced in this model are computer simulations. They are not deduced praxeologically from facts about individuals' actions. This model counts as an Austrian approach only in a most etiolated sense. (Incidentally, Robert Nozick was in his last years much interested in the Sugarscape model.) If Caldwell must go in for this sort of thing, he might also have mentioned the computational economy devised by Eric Baum and Igor Durdanovic, called The Hayek Machine. (For a discussion, see Eric Baum, What Is Thought? MIT Press, 2004, pp. 250–66.)

Caldwell's book is based on wide reading, but I did note a few errors. Bismarck was not "Prussian chancellor" (p. 57). Frederic Benham's joke about "given data" is I think misunderstood. The point is not that the economists who used the expression did not specify to whom the data are given, but rather that, since "data" means "things given", the expression "given data" is redundant (p. 212). Hayek's claim that the ultimate state of science is a self-contained, tautological system is not inconsistent with "evolutionary themes," (p. 275) since temporal predicates can be incorporated into such a system. (Caldwell does not pursue the obvious parallel with Leibniz.) It is unlikely that the remarks on evolutionary adaptation in The Sensory Order come from Ludwig von Bertalanffy. He was a strong critic of Darwinism, and I remember Hayek's dismissal in conversation of his anti-evolutionary views (p. 278). Why is Felix Ehrenhaft's amusing complaint about Hayek's incomprehensibility cited to support the view that social scientists found Hayek hard to understand (p. 261)? Ehrenhaft was an experimental physicist.

David Gordon covers new books in economics, politics, philosophy, and law for The Mises Review, the quarterly review of literature in the social sciences, published since 1995 by the Mises Institute.


  1. ^ Caldwell’s decision to ignore Hayek’s work on capital theory leads him to ignore at least one instance of material relevant to the topics he covers. In The Pure Theory of Capital, Hayek has a most valuable discussion of how misleading terms can impede economic analysis.
  2. ^ A supporter of the theory-ladenness of facts can escape subjectivism if he confines theory to the a priori. Mises took this way out in economics.
  3. ^ Caldwell states, "I do not now and never have in the past accepted Ludwig von Mises’s a priorism. Indeed, I spent a considerable amount of time during my post-doc year studying Austrian economics at NYU trying to convince a number of Austrians to abandon their commitment to apriorism." (pp.419–20). He does say, though, that if a priorism were compatible with fallibilism, this would remove one of his main objections.
  4. ^ I suggest that Quine’s influence on Hayek’s view of the a priori, a topic Caldwell does not explore, warrants attention. Hayek told me that he regarded Quine as "one of our most stimulating epistemologists."
  5. ^ Hayek told me in 1969 that he regarded Mises’s Theory and History as worth careful study.

Accédez d'un seul coup d’œil aux articles consacrés à Friedrich A. Hayek sur Catallaxia.