"F. A. Hayek and the Rebirth of Classical Liberalism"
By John N. Gray
In the recent revival of public and scholarly interest in the values of limited government and the market order, no one has been more centrally significant than Friedrich A. Hayek. His works have figured as a constant point of reference in the discussions both of the libertarian and conservative theories of the market economy; they have also provided a focal point of attack for interventionist and collectivist critics of the market. Hayek’s return to such a pivotal position in intellectual life is remarkable when we recall that for several decades his work was subjected to neglect and obscurity. It was not until 1974 at the age of 75 that he was belatedly acknowledged by being awarded the Nobel Memorial Prize in Economic Science. During the three decades after 1945, when certain Keynesian ideas seemed to have been vindicated by the prevailing government policies of economic interventionism, Hayek may have seemed an intransigent and isolated figure, whose chief importance was that of an indefatigable critic of the spirit of the age. It was, however, during these very same years, in which he turned from economic theory to political thought, that Hayek made his greatest contributions thus far to the formulation of a public philosophy, including most notably his
Constitution of Liberty (1960), surely the most powerful and profound defense of individual freedom in our time. It is noteworthy that, in the revival of interest in Hayek’s work, his contributions to political philosophy have attracted as much interest as have his works in economic theory…. [From the text]
First Pub. Date
Literature of Liberty. vol. v, no. 4, pp. 19-101. Arlington, VA: Institute for Humane Studies
Gray, John N. (Jesus College, Oxford)
The text of this edition is copyright ©1982, The Institute for Humane Studies. Republished with permission of original copyright holders.
Hayek’s conception of knowledge, when taken in conjunction with the idea of a spontaneous social order, has important implications for the proper method for the practice of social science. To begin with, Hayek’s affirmation of “the primacy of the abstract” in all human knowledge means that social science is always a theory-laden activity and can never aspire to an exhaustive description of concrete social facts. More, the predictive aspirations of social science must be qualified: not even the most developed of the social sciences, economics, can ever do more than predict the occurrence of general classes of events. Indeed, in his strong emphasis on the primacy of the abstract, Hayek goes so far as to question the adequacy of the nomothetic or nomological model of science (i.e. exact prediction through ‘laws’), including social science. At least in respect of complex phenomena, all science can aim at is an “explanation of the principle,” or the recognition of a pattern—”the explanation not of the individual events but merely of the appearance of certain patterns or orders. Whether we call these mere explanations of the principle or mere pattern predictions or higher level theories does not matter.” Such recognitions of orders or pattern predictions are, Hayek observes, fully theoretical claims, testable and falsifiable: but they correspond badly with the usual cause-effect structure of nomothetic or law-governed explanation.
In his most important later statement on these questions, “The Theory of Complex Phenomena,” [bibliography, A-109], Hayek tells us that, because social life is made up of complex phenomena, “economic theory is confined to describing kinds of patterns which will appear if certain general conditions are satisfied, but can rarely if ever derive from this knowledge any predictions of specific phenomena.” If we ask why it is that social phenomena are complex phenomena, part of the reason at any rate lies in what Hayek earlier characterized as the subjectivity of the data of the social sciences: social objects are not like natural objects whose properties are highly invariant relatively to our beliefs and perceptions; rather, social objects are in large measure actually constituted by our beliefs and judgments. Social phenomena are non-physical, and Hayek has stated that “Non-physical phenomena are more complex because we call physical phenomena what can be described by relatively simple formulae.” And, because of the subjectivity of its data, social life always eludes such simple formulae.
A number of points may be made briefly about Hayek’s conception of method in social and economic theory. First, whereas he follows his great teachers in the Austrian tradition in emphasizing the subjective aspects of social phenomena, Hayek’s methodology of social and economic science does not belong to that Austrian tradition in which social theory is conceived as an enterprise yielding apodictic truths. Specifically—contrary to T. W. Hutchinson, who periodizes Hayek’s work into an Austrian praxeological and a post-Austrian Popperian period, and also contrary to Norman P. Barry who sees both trends running right through Hayek’s writings—Hayek never accepted the Misesian conception of a praxeological science of human action which would take as its point of departure a few axioms about the distinctive features of purposeful behavior over time. In the Introduction to Collectivist Economic Planning [E-5, 1935] and elsewhere in his early writings, Hayek had (as Hutchinson notes) insisted that economics yields ” ‘general laws,’ that is, ‘inherent necessities determined by the permanent nature of the constituent elements.’ “ As Hutchinson himself acknowledges in passing, however, such laws or necessities function in Hayek’s writings as postulates (rather than as axioms), and they continue to do so even in his later writings, in which (as I have already noted) a suspicion of the nomothetic paradigm of social science is expressed. It is clear from the context of the quotations cited by Hutchinson that, in speaking of the general laws or inherent necessities of social and economic life, Hayek meant to controvert the excessive voluntarism of historicism, which insinuates that social life contains no unalterable necessities of any sort, rather than to embrace the view that there can be an apriori science of society or human action. To this extent Barry is right in his observation that, “there is a basic continuity in Hayek’s writings on methodology.” Certainly there seems little substance in a periodization of Hayek’s methodological writings by reference to the supposedly Popperian paper of 1937 on “Economics and Knowledge” (A-34).
At the same time, there seems little warrant for Barry’s claim that throughout his work Hayek tries “to combine two rather different philosophies of social science; the Austrian praxeological school with its subjectivism and rejection of testability in favour of axiomatic reasoning, and the hypothetico-deductive approach of contemporary science with its emphasis on falsifiability and empirical content.” For there is no evidence, so far as I know, that Hayek ever endorsed the Misesian conception of an axiomatic or apriori science of human action grounded in apodictic certainties. Again, as we have seen, Hayek’s view that the social sciences are throughout deductive in form antedates Popper’s influence and is evidenced in the Introduction to Collectivist Economic Planning [E-5, 1935].
Hayek’s real debts to Popper are, I think, different from those attributed to him by Hutchinson and Barry. It is not that Hayek under Popper’s influence abandoned an apodictic-deductive method that was endorsed (in different versions, Kantian and Aristotelian) by Mises and Menger, but rather that he came to adopt Popper’s proposal that falsifiability be treated as a demarcation criterion of science from non-science. Again, Hayek follows Popper in abandoning his earlier Austrian conviction that there is a radical dualism of method as between natural and social science: this conviction, he tells us, depended on an erroneous conception of method in the natural sciences: as a result of what Popper has taught him, Hayek says, “the differences between the two groups of disciplines has thereby been greatly narrowed.” Hayek’s debts to Popper are, then, in his seeing that it is the falsifiability of an hypothesis rather than its verifiability which makes it testable and empirical, and, secondly, in his acknowledging the unity of method in all the sciences, natural and social, where this method is seen clearly to be hypothetico-deductive.
Even in these Popperian influences, it is to be noted, there are differences of emphasis from Popper himself. Hayek anticipates Lakatos in perceiving that the theoretical sciences may contain a “hard core” of hypotheses, well-confirmed and valuable in promoting understanding of the phenomena under investigation, which are highly resistant to testing and refutation. And Hayek explicitly states that in some fields Popper’s ideas of maximum empirical content and falsifiability may be inappropriate:
It is undoubtedly a drawback to have to work with theories which can be refuted only by statements of a high degree of complexity, because anything below that degree of complexity is on that ground alone permitted by our theory. Yet it is still possible that in some fields the more generic theories are the more useful ones… Where only the most general patterns can be observed in a considerable number of instances, the endeavour to become more ‘scientific’ by further narrowing down our formulae may well be a waste of effort…
In general, then, it seems fair to hold that Hayek acknowledges that the proper method in social and economic studies, as elsewhere, is the hypothetico-deductive method of conjectures and refutations set out by Popper. On the other hand, he continues to recognize that in respect of complex phenomena such as are found in the social studies, testability may be a somewhat high-level and protracted process, and the ideal of high empirical content captured in a nomothetic framework—a demanding and sometimes unattainable ideal.
Hayek’s view that we can at best attain abstract models of social processes, whereas the concrete details of social life will always largely elude theoretical formulation, has large and radical implications in the field of public policy. In brief, it entails that the object of public policy should be confined to the design or reform of institutions within which unknown individuals make and execute their own, largely unpredictable plans of life. In a free society, in fact, whereas there may be a legal policy in respect of economic institutions, there cannot be such a thing as economic policy as it is presently understood, for adherence to the rule of law precludes anything resembling macroeconomic management. Here I do not wish to take up this point, which I will consider later, but rather to spell out the connection between Hayek’s methodological views and his belief that most, if not all economic policy as practiced in the postwar world has had a self-defeating effect.
We have seen that, for Hayek, the most we can hope for in understanding social life is that we will recognize recurring patterns. Hayek goes on to observe:
Predictions of a pattern are… both testable and valuable. Since the theory tells us under which general conditions a pattern of this sort will form itself, it will enable us to create such conditions and to observe whether a pattern of the kind predicted will appear. And since the theory tells us that this pattern assures a maximisation of output in a certain sense, it also enables us to create the general conditions which will assure such a maximisation, though we are ignorant of many of the particular circumstances which will determine the pattern that will appear.
Hayek’s view stands in sharp opposition to any idea of a policy science or a political technology aimed at producing specific desired effects. Such a policy science demands the impossible of its practitioners, a detailed knowledge of a changing and complex order in society. Even Popper’s conception of “piecemeal social engineering,” Hayek tells us, “suggests to me too much a technological problem of reconstruction on the basis of the total knowledge of the physical facts, while the essential point about the practical improvement is an experimental attempt to improve the functioning of some part without a full comprehension of the structure of the whole.” Indeed Hayek’s central point is that understanding the primacy of the abstract in human knowledge means that we must altogether renounce the modern ideal of consciously controlling social life: a better ideal is that of cultivating the general conditions in which beneficial results may be expected to emerge.
Hayek’s critique of the constructivistic or engineering approach to social life parallels in an intriguing way that of Michael Oakeshott and of the Wittgensteinian philosopher Rush Rhees. Consider Oakeshott’s statement: “The assimilation of politics to engineering is, indeed, what may be called the myth of rationalist politics.” Or Rhees’ observation (made in criticism of Popper): “There is nothing about human societies which makes it reasonable to speak of the application of engineering to them. Even the most important ‘problems of production’ are not problems in engineering.” The conception of social life which talk of social engineering expresses is at fault not only because it presupposes an agreement on goals or ends which nowhere exists but also because it promotes the illusion that political life may become subject to a sort of technical or theoretical control.
These general views illuminate much of the rationale of Hayek’s opposition not only to Keynesian policies of macroeconomic demand management but also to Friedmanite monetarism. Of course, in the great debates of the Thirties, Hayek had argued forcefully that Keynes in no way provided a general theory of economic discoordination. Again, Hayek always argued that the policies Keynes suggested, depending as they did for their success upon institutional and psychological irrationalities which their very operation would undermine, were bound over the longer run to be self-defeating. In particular, Hayek maintained that Keynesian policies of deficit financing depended for their success upon a widespread money illusion which the policies themselves could not help but erode. Hayek’s further objection to Keynesian policies is that, in part because they depend on a defective understanding of the business cycle (which is seen as expressing itself in aggregative variations in total economic activity rather than in a discoordination of relative price structures brought about by a governmental distortion of the structure of interest rates) Keynesian policy-makers, because of their holistic and aggregative bias, find it hard to avoid committing a sort of fallacy of conceptual realism: statistical artefacts or logical fictions are allowed to blot out the subtle and complex relationships which make up the real economy.
Now there is plainly much in Hayek’s subtle account of the business cycle, and in his contributions to capital theory, which is difficult and disputable, and to comment on such questions is in any case beyond my expertise. Quite apart from its technical details, however, it is clear that Hayek’s critique of Keynesian policies is of a piece with his emphasis on the primacy of the abstract and with his insight into the indispensability of conventions for the orderly conduct of social life. Policies of macroeconomic demand management ask more in the way of concrete knowledge of the real relationships which govern the economy than any administrator could conceivably acquire, and their operation is in the longer run self-defeating. More generally, Hayek’s challenge to Keynesian theory is a demand that Keynesians specify in detail the mechanisms whereby an unhampered market could be expected to develop severe discoordination. Only if such mechanisms could be clearly described and (crucially) given a plausible historical application, would a serious challenge to Hayek’s own Austrian view—in which it is governmental intervention in the economy which is principally responsible for discoordination—enter the realm of critical debate.
In respect to Friedman’s proposals for monetary regulation by a fixed rule, Hayek has argued that in a modern democracy no governmental or quasi-governmental agency can preserve the independence of action essential if such a monetary rule is to be operated consistently. More fundamentally, such a policy of adopting a fixed rule in the supply of money is opposed by Hayek on methodological grounds. Such a policy calls for an exactitude in modeling and measuring economic life, and an unambiguity in the definition of money, which it is beyond our powers to attain. Hayek’s own objection to Friedman’s monetarist proposals is, then, most substantially that money is not the sort of social object that we can define precisely or control comprehensively; Hayek has even suggested that, in recognition of the elusiveness of the monetary phenomenon, we should treat “money” as an adjectival expression, applicable to indefinitely many distinct and disparate instruments. Hayek’s proposals in this area clearly open up technical questions in monetary theory which I am unqualified to adjudicate. It seems clear, though, that Hayek’s proposal favoring currency competition by the private issuance of money would be found objectionable by Friedmanites (who would argue that Hayek exaggerates the effect such competition would have in preventing currency debasement) and by advocates of the classical gold standard. It is clear, nonetheless, that in arguing for the establishment of a monetary catallaxy Hayek has illuminated questions both in monetary theory and in political economy which had hitherto gone largely neglected, but which it is critical that supporters of the market order now examine.
One objection to Hayek’s view may be worth addressing at this point. There is much in Hayek’s account of the business cycle, as in his more general account of spontaneous social order, to suggest that he believes economic discoordination results always from institutional factors, so that at any rate large-scale disequilibrium would be impossible in a catallaxy of wholly unhampered markets. Against this view, Hayek’s brilliant and largely neglected pupil, G. L. S. Shackle, has argued that the subjectivity of expectations must infect the market process with an ineradicable tendency to disequilibrium. It must be allowed that, if we accept Hayek’s view of equilibrium as a process in which men’s plans are coordinated by trial and error over time, there can be nothing apodictically certain about this process: conceivably, under some conditions of uncertainty in which hitherto reliable expectations are repeatedly confounded, large scale discoordination could occur in the market process.
Three counter-observations are in order, however. First, nothing in Shackle’s argument tells against the point, defensible both on theoretical grounds and as an historical interpretation, that in practice by far the most destabilizing factor in the market process is provided by governmental intervention. Secondly, and relatedly, it is unclear that the kind of disequilibrium of which Shackle speaks—disequilibrium generated by divergency in subjective expectations—could amount to anything resembling the classical business cycle, which is more plausibly accounted for in Austrian and Hayekian terms as a consequence of governmental intervention in the interest rate structure.
And thirdly, it is unclear that Shackle’s argument shows the presence in the market process of any tendency to disequilibrium. What we have in the market process is admittedly a ‘kaleidic’ world, in which expectations, tastes, and beliefs constantly and unpredictably mutate. Yet, providing market adaptation is unhampered, what we can expect from the market process is an uninterrupted series of monetary equilibrium tendencies, each of them asymptotic—never quite reaching equilibrium—and each of them soon overtaken by its successor. In this kaleidic world there may well be no apodictic certainty that we shall never face large-scale, endogenous discoordination, but we are nevertheless on safe ground in preferring that the self-regulating tendencies of the process be accorded unhampered freedom and that governmental intervention be recognized as the major disruptive factor in the market process. We are on safe ground, then, in discerning in the tendency to equilibrium in the market process the formation of spontaneous order in the economic realm.
Given that we recognize governmental intervention to be the greatest subverter of spontaneous order in the realm of economic exchange, what legal framework is to be adopted for the regulation of economic life? Here we come to one of the most fascinating and controversial of Hayek’s contributions to social philosophy, his account of individual liberty under the rule of law. Before we can address ourselves to some of the problems surrounding Hayek’s contribution to philosophical jurisprudence, however, a few words must be said about Hayek’s moral theory, since few aspects of Hayek’s work are so often misunderstood. Hayek has been characterized as a moral relativist, an exponent of evolutionary ethics and, less implausibly but nonetheless incorrectly, as a rule-utilitarian. Let us see if we can dissipate the confusion.
In the first place, moral life for Hayek is itself a manifestation of spontaneous order. Like language and law, morality emerged undesigned from the life of men with one another: it is so much bound up with human life, indeed, as to be partly constitutive of it. The maxims of morality, then, in no way presuppose an authority, human or divine, from which they emanate, and they antedate the institutions of the state. But, secondly, the detailed content of the moral conventions which spring up unplanned in society is not immutable or invariant. Moral conventions change, often slowly and almost imperceptibly, in accordance with the needs and circumstances of the men who subscribe to them. Moral conventions must (or Hayek’s account of them) be seen as part of the evolving social order itself.
Now at this point it is likely that a charge of ethical relativism or evolutionism will at once be levelled against Hayek, but there is little substance to such criticisms. He has gone out of his way to distinguish his standpoint from any sort of evolutionary ethics. As he put it in his Constitution of Liberty:
It is a fact which we must recognize that even what we regard as good or beautiful is changeable—if not in any recognizable manner that would entitle us to take a relativistic position, then in the sense that in many respects we do not know what will appear as good or beautiful to another generation… It is not only in his knowledge, but also in his aims and values, that man is the creature of his civilization; in the last resort, it is the relevance of these individual wishes to the perpetuation of the group or the species that will determine whether they persist or change. It is, of course, a mistake to believe that we can draw conclusions about what our values ought to be simply because we realize that they are a product of evolution. But we cannot reasonably doubt that these values are created and altered by the same evolutionary forces that have produced our intelligence.
Hayek’s argument here, then, is manifestly not that we can invoke the trend of social evolution as a standard for the resolution of moral dilemmas, but rather that we are bound to recognize in our current moral conventions the outcome of an evolutionary process. Admittedly, inasmuch as nothing in the detailed content of our moral conventions is unchanging or unalterable, this means that we are compelled to abandon the idea that they have about them any character of universality or fixity, but this is a long way from any doctrine of moral relativism. As Hayek observes in his remarks on the ambiguity of relativism:
… our present values exist only as the elements of a particular cultural tradition and are significant only for some more or less long phase of evolution—whether this phase includes some of our pre-human ancestors or is confined to certain periods of human civilization. We have no more ground to ascribe to them eternal existence than to the human race itself. There is thus one possible sense in which we may legitimately regard human values as relative and speak of the probability of their further evolution.
But it is a far cry from this general insight to the claims of the ethical, cultural or historical relativists or of evolutionary ethics. To put it crudely, while we know that all these values are relative to something, we do not know to what they are relative. We may be able to indicate the general class of circumstances which have made them what they are, but we do not know the particular conditions to which the values we hold are due, or what our values would be if those circumstances had been different. Most of the illegitimate conclusions are the result of erroneous interpretation of the theory of evolution as the empirical establishment of a trend. Once we recognize that it gives us no more than a scheme of explanation which might be sufficient to explain particular phenomena if we knew all the facts which have operated in the course of history, it becomes evident that the claims of the various kinds of relativists (and of evolutionary ethics) are unfounded.
Hayek does not, then subscribe to any sort of ethical relativism or evolutionism, but it is not altogether clear from these statements if he thinks humanity’s changing moral conventions have in fact any invariant core or constant content. In order to consider this last question, and to attain a better general understanding of Hayek’s conception of morality, we need to look at his debts to David Hume, whose influence upon Hayek’s moral and political philosophy is ubiquitous and profound.
Hayek follows Hume in supposing that, in virtue of certain general facts about the human predicament, the moral conventions which spring up spontaneously among men all have certain features in common or (in other words) exhibit some shared principles. Among the general facts that Hume mentions in his Treatise, and which Hayek cites in “The Legal and Political Philosophy of David Hume” (in B-13), are men’s limited generosity and intellectual imperfection and the unalterable scarcity of the means of satisfying human needs. As Hayek puts it succinctly: “It is thus the nature of the(se) circumstances, what Hume calls ‘the necessity of human society,’ that gives rise to the ‘three fundamental laws of nature’: those of ‘the stability of possessions, of its transference by consent, and of the performance of promises.’ ” And Hayek glosses this passage with a fuller citation from Hume’s Treatise: “Though the rules of justice be artificial, they are not arbitrary. Nor is the expression improper to call them Laws of Nature; if by natural we understand what is common to any species, or even if we confine it to mean what is inseparable from the species.”
Hume’s three rules of justice or laws of nature, then, give a constant content to Hayek’s conception of an evolving morality. They frame what the distinguished Oxford jurist, H. L. A. Hart, was illuminatingly to call “the minimum content of natural law.” The justification of these fundamental rules of justice, and of the detailed and changing content of the less permanent elements of morality, is (in Hayek’s view as in Hume’s) that they form indispensable conditions for the promotion of human welfare. There is in Hayek as in Hume, accordingly, a fundamental utilitarian commitment in their theories of morality. It is a very indirect utilitarianism that they espouse, however, more akin to that of the late nineteenth-century Cambridge moralist Henry Sidgwick (1838-1900) than it is to Jeremy Bentham or John Stuart Mill. The utilitarian component of Hayek’s conception of morality is indirect in that it is never supposed by him that we ought or could invoke a utilitarian principle in order to settle practical questions: for, given the great partiality and fallibility of our understanding, we are in general better advised to follow the code of behavior accepted in our own society. That code can, in turn, Hayek believes, never properly be the subject of a rationalist reconstruction in Benthamite fashion, but only reformed piecemeal and slowly. In repudiating the claims that utilitarian principles can govern specific actions and that utility may yield new social rules, Hayek shows himself to be an indirect or system utilitarian, for whom the proper role of utility is not prescriptive or practical but rather as a standard of evaluation for the assessment of whole systems of rules.
Again however, Hayek’s utilitarian outlook is distinctive in that he explicitly repudiates any hedonistic conception of the content of utility itself. How, then, does he understand utilitarian welfare? Just how are we to assess different systems of rules in regard to their welfare-promoting effects? Here Hayek comes close to modern preference utilitarianism, but gives that view an original formulation, in arguing that the test of any system of rules is whether it maximizes an anonymous individual’s chance of achieving his unknown purposes. In Hayek’s conception, we are not bound to accept the historical body of social rules just as we find it: it may be reformed in order to improve the chances of the unknown man’s achieving his goals. It will be seen that this is a maximizing conception, but not one that represents utility as a sort of neutral stuff, a container of intrinsic value whose magnitude may vary. Indeed, in taking as the point of comparison an hypothesized unknown individual, Hayek’s conception (as he recognizes) parallels John Rawls’ model of rational choice behind a veil of ignorance as presented in Rawls’ Theory of Justice.
Mention of Rawls’ contractarian derivation of principles of justice at once raises the question of how Hayek’s indirect or system utilitarian argument is supposed to ground the rules of justice he defends, and, in particular, how Hayek’s defense of the priority of liberty squares with his utilitarian outlook.
Several observations are apposite here. First, Hayek undoubtedly follows Hume in believing that, because they constitute an indispensable condition for the promotion of general welfare, the rules of justice are bound to take priority over any specific claim to welfare. Again, it is to be noted that Hume’s second rule of justice, the transference of property by consent, itself frames a protected domain and so promotes individual liberty. Finally, Hayek argues forcefully that, if individuals are to be free to use their own knowledge and resources to best advantage, they must do so in a context of known and predictable rules governed by law. It is in a framework of liberty under the rule of law, Hayek contends, that justice and general welfare are both served. Indeed, under the rule of law, justice and the general welfare are convergent and not conflicting goals or values.
These claims regarding the relations between justice, liberty, and the rule of law encompass the most controversial and the most often attacked portion of Hayek’s social philosophy. Common to all criticisms of it is the objection that Hayek expects too much of the rule of law itself, which is only one of the virtues a legal order may display, and a rather abstract notion at that. Among classical liberals and libertarians, this objection has acquired a more specific character. It has been argued that upholding the rule of law cannot by itself protect liberty or secure justice, for these values will be promoted only if the individual rights are respected. Hayek’s theory is at the very least radically incomplete, according to these critics, inasmuch as his conception of the rule of law will have the classical liberal implications he expects of it, only if it incorporates a conception of individual rights, which he seems explicitly to disavow. All these liberals and libertarians fasten upon Hayek’s use of a Kantian test of universalizability to argue that such a test is almost without substance, in that highly oppressive and discriminatory laws will survive it, so long as their framers are ingenious enough to avoid mentioning particular groups or named individuals in them. The upshot of this criticism is that, in virtue of the absence in his theory of any strong conception of moral rights, Hayek is constrained to demand more of the largely formal test of universalizability than it can possibly deliver, and so to conflate the ideal of the rule of law with other political goods and virtues.
This fundamental criticism of Hayek, stated powerfully by Hamowy and Raz and endorsed in earlier writings of my own, now seems to me to express an impoverished and mistaken view of the nature and role of Kantian universalizability in Hayek’s philosophical jurisprudence. It embodies the error that, in Hayek or indeed in Kant, universalizability is a wholly formal test.
In his “Principles of a Liberal Social Order,” (A-115, in B-13) Hayek tells us: “The test of the justice of a rule is usually (since Kant) described as that of its ‘universalizability,’ i.e. of the possibility of willing that rules should be applied to all instances that correspond to the conditions stated in it (the ‘categorical imperative’).” As an historical gloss, Hayek observes that:
It is sometimes suggested that Kant developed his theory of the Rechtstaat by applying to public affairs his conception of the categorical imperative. It was probably the other way round, and Kant developed his theory of the categorical imperative by applying to morals the concept of the rule of law which he found ready made (in the writings of Hume).
Hayek’s own argument, that applying Kantian universalizability to the maxims that make up the legal order yields liberal principles of justice which confer maximum equal freedom upon all, has been found wanting by nearly all his critics and interpreters. Thus Raz quotes Hayek as follows:
“The conception of freedom under the law that is the chief concern of this book rests on the contention that when we obey laws, in the sense of general abstract rules laid down irrespective of their application to us, we are not subject to another man’s will and are therefore free. It is because the judge who applies them has no choice in drawing the conclusions that follow from the existing body of rules and the particular facts of the case, that it can be said that laws and not men rule… As a true law should not name any particulars, so it should especially not single out any specific persons or group of persons.”
Raz comments on this passage: “Then, aware of the absurdity to which this passage leads, he modifies his line, still trying to present the rule of law as the supreme guarantee of freedom…”
Similarly, discussing Hayek’s criteria that laws should not mention proper names and that the distinctions which the law makes be supported both within and without the group which is the subject of legislation, Hamowy comments:
That no proper name be mentioned in a law does not protect against particular persons or groups being either harassed by laws which discriminate against them or granted privileges denied the rest of the population. A prohibition of this sort on the form laws may take is a specious guarantee of legal equality, since it is always possible to contrive a set of descriptive terms which will apply exclusively to a person or group without recourse to proper names…
How are these standard objections to be rebutted?
We must first of all note that, even in Kant and in Kantian writers other than Hayek, such as R. M. Hare and John Rawls, the test of universalizability does far more than rule out reference to particular persons or special groups. The test of universalizability does indeed, in the first instance, impose a demand of consistency as between similar cases, and in that sense imposes a merely formal requirement of non-discrimination. This is the first stage or element of universalization, the irrelevance of numerical differences. But the next stage of universalization is that of asking whether one can assent to the maxim being assessed coming to govern the conduct of others towards oneself: this is the demand of impartiality between agents, the demand that one put oneself in the other man’s place. And this element or implication of universalizability leads on to a third, that we be impartial as between the preferences of others, regardless of our own tastes or ideals of life—a requirement of moral neutrality. I do not need to ask here exactly how these elements of universalizability are related to one another, to ask (most obviously) if the second is entailed by the first in any logically inexorable way, or similarly the third by the second. It is enough to note that there is a powerful Kantian tradition according to which strong implications do link the three phases of universalization, and that this is a tradition to which Hayek himself has always subscribed.
Applying the full test of universalizability to the maxims that go towards making a legal order, we find that, not only are references to particulars ruled out, but the maxims must be impartial in respect of the interests of all concerned, and they must be neutral in respect of their tastes or ideals of life. If it be once allowed that the test of universalizability may be fleshed out in this fashion, it will be seen as a more full-blooded standard of criticism than is ordinarily allowed, and Hayek’s heavy reliance on it will seem less misplaced. For, when construed in this fashion, the universalizability test will rule out (for example) most if not all policies of economic intervention as prejudicial to the interests of some and will fell all policies of legal moralism. Two large classes of liberal policy, supposedly allowable under an Hayekian rule of law, thus turn out to be prohibited by it.
Hayek himself is explicit that the test of universalizability means more than the sheerly formal absence of reference to particulars. As he puts it:
The test of the justice of a rule is usually (since Kant) described as that of its ‘universalizability,’ i.e. of the possibility of willing that the rules should be applied to all instances that correspond to the conditions stated in it (the ‘categorical imperative’). What this amounts to is that in applying it to any concrete circumstances it will not conflict with any other accepted rules. The test is thus in the last resort one of the compatibility or non-contradictoriness of the whole system of rules, not merely in a logical sense but in the sense that the system of actions which the rules permit will not lead to conflict.
The maxims tested by the principle of universalizability, then, must be integrated into a system of nonconflictable or (in Leibniz’ terminology) compossible rules, before any of them can be said to have survived the test.
Again, the compatibility between the several rules is not one that holds in any possible world, but rather that which obtains in the world in which we live. It is here that Hayek draws heavily on Hume’s account of the fundamental laws of justice, which he thinks to be, not merely compatible with, but in a large measure the inspiration for Kant’s political philosophy. As I have already observed, the practical content of the basic rules of justice is given in Hume by anthropological claims, by claims of general fact about the human circumstance. It is by interpreting the demands of universalizability in the framework of the permanent necessities of human social life that we derive Hume’s three laws of natural justice.
Note again that, in Hume, as in Hayek, the laws of justice are commended as being the indispensable condition for the promotion of general welfare, i.e. their ultimate justification is utilitarian. But in order to achieve this result, neither Hayek nor Hume need offer any argument in favor of our adopting a Principle of Utility. Rather, very much in the spirit of R. M. Hare’s Kantian reconstruction of utilitarian ethics, Hayek’s claim is that an impartial concern for the general welfare is itself one of the demands of universalizability. A utilitarian concern for general welfare is yielded by the Kantian method itself and is not superadded to it afterwards. Hayek’s thesis, like Hume’s, is that a clear view of the circumstances of human life shows justice to be the primary condition needed to promote general welfare. But, like Hare and Kant, he thinks concern for both justice and the general welfare to be dictated by universalizability itself.
Hayek’s argument, then, is that the maxims of liberal justice are yielded by applying the Kantian universalizability test to the principles of the legal order. As he puts it:
It will be noticed that only purpose-independent (‘formal’) rules pass this (Kantian) test because, as rules which have originally been developed in small purpose-connected groups (‘organizations’) are progressively extended to larger and larger groups and finally universalized to apply to the relations between any members of an Open Society who have no concrete purposes in common and merely submit to the same abstract rules, they will in the process have to shed all reference to particular purposes.
Again, in listing the essential points of his conception of justice Hayek asserts:
… a) that justice can be meaningfully attributed only to human actions and not to any state of affairs as such without reference to the question whether it has been, or could have been, deliberately brought about by somebody; b) that the rules of justice have essentially the nature of prohibitions, or, in other words, that injustice is really the primary concept and the aim of rules of just conduct is to prevent unjust action; c) that the injustice to be prevented is the infringement of the protected domain of one’s fellow men, a domain which is to be ascertained by means of these rules of justice; and d) that these rules of just conduct which are in themselves negative can be developed by consistently applying to whatever such rules a society has inherited the equally negative test of universal applicability—a test which, in the last resort, is nothing less than the self-consistency of the actions which these rules allow if applied to the circumstances of the real world.
There seem to be several elements, then, in Hayek’s contention that applying the Kantian test to the legal framework yields a liberal order. First, though he does not explicitly distinguish the three stages or phases of universalization I mentioned earlier, he is clear that the universalizability test is not only formal, and that it comprehends the requirement that the scheme of activities it permits in the real world would be conflict-free. Second, at any rate in a society whose members have few if any common purposes, law must have a largely formal character, stipulating terms under which men may pursue their self-chosen activities rather than enjoining any specific activities on them; in the term Hayek adopts from Oakeshott, the form of legal rule appropriate to such an abstract or open society is “nomocratic” rather than “teleocratic,” purpose-neutral rather than purpose-dependent. Third, in a society whose members lack common purposes or common concrete knowledge, only abstract rules conferring a protected domain on each can qualify as rules facilitating a conflict-free pattern of activities. This means that the conditions of our abstract or open society will themselves compel adoption of a rule conferring just claims to liberty and private property—which Hayek rightly sees as indissolubly linked—once these conditions are treated as the appropriate background for the Kantian test.
One crucially important implication of this last point, noted in all of Hayek’s political writings over the last twenty years but spelled out most systematically in the second volume of his recent trilogy, Law, Legislation and Liberty, is that the rules of justice which survive the Kantian test can prescribe justice only in the procedures and never in end-states. As Hayek puts it, explicating Hume: “There can be no rules for rewarding merit, or no rules of distributive justice, because there are no circumstances which may not affect merit, while rules always single out some circumstances as the only relevant ones.”
This pattern of argument is an important and striking one, worth examining in detail on its merits, and not capable of being dismissed as prima facie unworkable. One important point may be worth canvassing, however. Hayek argues that once the legal framework has been reformed in Kantian fashion, it must of necessity be one that maximizes liberty. Hamowy goes so far as to assert that Hayek defines liberty as conformity with the rule of law. Now, whereas not every aspect of Hayek’s treatment of freedom and coercion is clear or defensible, it seems a misinterpretation to say that he ever defines freedom as consisting solely in conformity with the rule of law. Rather, he takes such conformity to be a necessary condition of a free order. His thesis is that applying the Kantian test to the legal order will of itself yield a maxim according equal freedom to all men. So it is not that the rule of law contains freedom as part of its definition, but rather that a freedom-maximizing rule is unavoidably yielded by it. In other terms, we may say that, whereas moral rights do not come into Hayek’s theory as primordial moral facts, the right to a protected domain is yielded by his conception as a theorem of it.
If Hayek is right that his method shows the unacceptability of contemporary patterned conceptions of justice, for example, and if as I think, he has shown that only procedural justice can be squared with the liberal maxim demanding equal freedom of action, then we can begin to see the measure of his achievement. Certainly, his Kantian derivation of equal freedom deserves close and sympathetic scrutiny, and it cannot be assumed without argument that Hayek’s system cannot protect individual rights or claims to justice simply because such rights do not enter the system at a fundamental level. For the most original and striking claim of Hayek’s legal and political philosophy, which in this respect may be regarded as a synthesis of the theories of justice of Hume and Kant, is that applying the rational test of universalizability to the conditions of our world must of necessity yield a system of rules in which a protected domain of individual liberty is secured.
In regard to his theory of justice, the criticisms we have surveyed appear to be premature, or at least inconclusive. We have yet to consider a much more fundamental criticism of Hayek’s system, directed against it by thinkers in very different traditions, which attends to the highly ambiguous role in Hayek’s theory of the idea of spontaneous order.
One of the clearest and deepest statements of some of the difficulties in Hayek’s use of spontaneous order arguments may be found in James M. Buchanan’s writings. In an important paper, Buchanan observes that, in Hayek’s later writings we find:
the extension of the principle of spontaneous order, in its normative function, to the emergence of institutional structure itself. As applied to the market economy, that which emerges is defined by its very emergence to be that which is efficient. And this result implies, in its turn, a policy of nonintervention, properly so. There is no need, indeed there is no possibility, of evaluating the efficiency of observed outcomes independently of the process; there exists no external criterion that allows efficiency to be defined in objectively measurable dimensions. If this logic is extended to the structure of institutions (including law) that have emerged in some historical evolutionary process, the implication seems clear that that set which we observe necessarily embodies institutional or structural ‘efficiency.’ From this it follows, as before, that a policy of nonintervention in the process of emergence is dictated. There is no room left for the political economist, or for anyone else, who seeks to reform social structures, to change laws and rules, with an aim of security instead of efficiency in the large… Any ‘constructively rational’ interferences with the ‘rational’ processes of history are, therefore, to be avoided.
Buchanan’s criticism, then, is that Hayek’s apparent extension of spontaneous order or evolutionary arguments from the market processes to institutional structures is bound to disable the tasks of criticism and reform. We are left with no leverage in Hayek’s account which might be used against the outcomes of the historical process. Instead, it seems, we are bound to entrust ourselves to all the vagaries of mankind’s random walk in historical space.
In an earlier critique, Buchanan noted perceptively the phenomenon of “spontaneous disorder”—the emergence of patterns of activity that thwart the purposes and damage the interests of all who participate in them. Such “spontaneous disorder” is, after all, the core of the idea of the Prisoner’s Dilemma, which has been explored imaginatively in Buchanan’s writing in its political and constitutional applications. The neglect in Hayek’s political work in English of any treatment of the problem this Dilemma poses for his system invites the attempt to accommodate these fundamental objections.
It is clear, however, that as it stands Hayek’s conception of spontaneous order needs revision or at least refinement. Buchanan’s identification of certain states of affairs as manifesting spontaneous disorder suggests the question whether the idea of spontaneous order in Hayek is a value-free explanatory notion or else a moral notion of some sort. If the former—as Hayek’s examples of spontaneous order in nature suggest—then spontaneous order really functions as a cipher for invisible hand explanations of the sort brilliantly discussed by Robert Nozick in his Anarchy, State, and Utopia.
We might then be compelled to regard the growth of interventionism and of the welfare state, and even certain aspects of the functioning of totalitarian regimes, as exemplifying spontaneous order inasmuch as we might be able to explain these social phenomena as the unintended outcomes of human action. If, on the other hand, spontaneous orders are taken as embodying positive moral values—if, that is to say, the idea of a maleficient or destructive spontaneous order is repudiated as incoherent—then it seems clear that Hayek requires a far bolder moral theory than any he has advanced thus far. In particular, such a moral theory would need to bridge the gap between evaluative and descriptive language which is a feature of modern moral philosophy, and in this and other respects it would need to come much closer to natural law ethics than Hayek has ever himself done.
Buchanan’s critique is decisive, then, in compelling Hayek to clarify the idea of spontaneous order as being either a moral notion, which might plausibly be embedded only in some variant of natural law ethics, or else as a value-free explanatory concept whose political uses must then be made more explicit than Hayek has heretofore done.
Buchanan’s critique is important, again, in disclosing that Hayek’s attitude to rationalism is ambivalent and unstable. If we adopt the latter view of spontaneous order as a value-free explanatory idea, its uses in political argument depend upon two kinds of considerations. First, they must invoke a political ethics, which arguably is given by Hayek’s synthesis of Hume with Kant. More problematically, however, the use of an explanatory idea of spontaneous order in political argument presupposes that we have a genuine theoretical or synoptic knowledge of social life of just the sort that Hayek occasionally suggests is impossible. This is to say that, if we are to make use of the idea of spontaneous social order in framing or reforming social institutions so as to make best use of society’s spontaneous forces, we need to invoke a theoretical model of social structure and social process which gives some assurance as to the outcome of our reforms. To this extent, contrary to some of Hayek’s recommendations but in line with a part of his recent practice, we cannot avoid adopting a critical rationalist stance toward our inherited institutions and the historical process. This is true, whether we accept Hayek’s own effort at a political ethics, or Buchanan’s neo-Hobbesian contractarian constitutionalism.
These cited points are reinforced if we consider Michael Oakeshott’s attitude to Hayek’s work. Oakeshott is a more intrepid traditionalist than Hayek in that Oakeshott claims that we cannot in the end do anything but accept the traditions which we inherit in our society. Certainly, we cannot appraise our traditions by reference to any transcendental standard of reason or justice, since such standards (in Oakeshott’s view) necessarily turn out to be abridgements of our traditions themselves. Like Hayek, then, Oakeshott maintains that all moral or political criticism must be immanent criticism, but, unlike Hayek, he denies that there is any inherent or evolutionary tendency for the development of traditional practices to converge on liberal institutions. For this reason Oakeshott would insist that his conception of civil association or nomocracy—upon which, as we have already seen, Hayek draws in his conception of the juridical framework of the liberal order—is a description of a strand of practice in the modern European state and has no necessary application beyond the cultural milieu in which it came to birth. Oakeshott would accordingly repudiate the implicit universalism of Hayek’s argument for the liberal order.
To some extent, of course, Hayek concedes that there cannot be universal scope for liberal principles when he allows that the Great or Open Society is itself an evolutionary emergence from rude beginnings. Where he differs from Oakeshott is in affirming that the Great or Open Society in which liberal principles are uniquely appropriate represents the future of all mankind. In this respect, Hayek continues to subscribe to an Enlightenment doctrine of universal human progress which Oakeshott has abandoned. I do not mean that Hayek has ever endorsed the belief that historical change is governed by a law of progressive development, but rather that he seems to take for granted (what surely is most disputable) that the unhampered natural selection of rival practices and traditions will result in a general convergence on liberal society.
A contrast of Hayek’s thought with that of Oakeshott revives one of the commonest criticisms of Hayek’s work, namely, that it straddles incompatible conservative and libertarian stand-points. The upshot of my discussion thus far may support this standard criticism in that it suggests that Hayek’s system is poised uneasily between the constructivist (but not uncritical) rationalism of a Buchanan and the out-and-out traditionalism of an Oakeshott.
At the same time, however, elements of Hayek’s conception of social evolution via the competitive selection of rival traditions may provide a point of convergence, if not of fusion, for some libertarian and conservative concerns. One central argument in contemporary neo-conservatism, after all, is in the claim that the stability of the free society depends upon its containing strong supportive traditions. Modern neo-conservatives such as Irving Kristol and Daniel Bell take up the doubts expressed by writers of the Scottish Enlightenment such as Smith and Ferguson about the effect on society’s moral traditions of the workings of the commercial marketplace itself. A major difficulty in the neo-conservative analysis is the lack of any very convincing prognosis: if free markets have corrosive effects in respect of the moral traditions which support them, so that capitalism institutions contain cultural contradictions which make them over the long run self-destroying, what is to be done?
This is an especially hard question if we recognize (as some of the neo-conservatives themselves sometimes fail to do) that merely capturing positions of power in the apparatus of the contemporary democratic state affords no longrun security for the market order.
There is in Hayek’s work an argument for voluntaristic traditionalism which goes some way toward answering this question. Hayek sees that the principal cause of the erosion of definitive moral traditions in advanced societies is not so much the market itself, but rather interventionist policies sponsored by governments. Often with the support of business, governments have contributed to the erosion of moral traditions by their educational, housing, and welfare policies. Hayek’s argument for a voluntaristic traditionalism distinguishes him from neo-conservatives, firstly in that he would argue that it is government interventionism which causes much of the contemporary moral malaise and because he would not seek to use government power to prop up faltering traditions. Rather, he seeks to establish something like a market in traditions, in the hope that the traditions which would emerge from an unhampered social life would be most congenial to the stability of the market order itself. In his argument for a competitive and voluntaristic traditionalism, Hayek plainly treats particular traditional communities as filter devices for social practices of the sort Robert Nozick discusses in his fascinating and profound account of the framework of utopia.
It cannot be said unequivocably that Hayek’s libertarian traditionalism answers the most profoundly disturbing doubts of the neo-conservatives. In particular, Hayek’s advocacy of procedural justice, with the role of chance in distributing incomes being recognized clearly, confronts the difficulty that the moral defense of capitalism has chiefly been conducted by reference to the notion of desert. By comparison with this traditional defense, Hayek’s apologia for the market order may be, as Kristol observes, “nihilistic.”
Against this criticism Hayek may justifiably maintain that there is a sheer conflict between traditional sentiments of desert and merit and any clear-sighted defense of the market order—a conflict which the neo-conservative endorsement of the market order does nothing to resolve.
Kristol’s criticism of Hayek has other, and perhaps profounder aspects, however. Hayek recognizes that contemporary moral sentiment is by no means uniformly, or even generally, favorable to the market order, and, both in his writings on Mandeville and elsewhere, Hayek has implicitly acknowledged that the spontaneous growth of moral norms may not, in fact, yield results congenial to a stable market order. At the same time, Hayek continues to advocate a strong form of moral conventionalism, resisting the claims of those who see modern morality as in need of radical reform. There is thus a tension, perhaps irresolvable in terms of Hayek’s system, between his Mandevillian moral iconoclasm and his moral conservatism.
In his argument for a voluntaristic traditionalism, Hayek (as we have seen) answers some of the concerns of contemporary conservatives. His argument for a market in traditions may be vulnerable to criticism, inasmuch as the growth of anti-market ethics over the past centuries seems to belie his expectation that natural selection of moral traditions will filter out those unfriendly to the market process. In recognition of this, Hayek would in consistency be compelled to adopt, in respect of moral convention, a more “rationalist” stance than he usually recommends. He would need to undertake a systematic criticism of modern morality in regard to its viability as part of an ongoing market order. In so doing, he would be resuming the task undertaken by those moderate rationalists, Bernard Mandeville and David Hume, whom Hayek rightly sees as the fountainheads of classical liberalism. Even if his own system of ideas should prove unstable, it recalls to us the insights of the great classical liberals, and intimates the most powerful research program in classical liberal political philosophy. And, in recalling that intellectual tradition from what had sometimes seemed an irrecoverable oblivion, Hayek’s work is a hopeful augury for an uncertain future.
I have learned much from three studies by Jeremy Shearmur: (1) “Abstract Institutions in an Open Society,” in H. Berghel and others, eds. Wittgenstein, the Vienna Circle and Critical Materialism, Vienna: Holder-Pichler-Tempsky, 1979, pp. 349-354; (2) “The Austrian Connection: F. A. von Hayek and the Thought of Carl Menger,” in B. Smith and W. Grassl, eds., Austrian Philosophy and Austrian Politics, Munich: Philosophia Verlag, forthcoming; and (3) Adam Smith’sSecond Thoughts (pamphlet), London: Adam Smith Club, 1982.