TWO KINDS OF ORDER

John Marks

'An age of superstitions is a time when people imagine they know more than they do'

(Hayek, Law, Legislation and Liberty, Vol. III, p. 176)

I wish to draw attention to some ideas of Friedrich von Hayek concerning phenomena which are, 'the result of human action, but not the execution of any human design' (Adam Ferguson An Essay on the History of Civil Society, London, 1767; quoted in Law, Legislation and Liberty, (hereafter LLL) Vol. I, p. 20 and note 19, p. 150). Hayek distinguishes three sources of human values - those 'which are genetically determined and therefore innate'; those which are 'products of rational thought'; and those which are the result of cultural evolution. He emphasises the importance of the third of these - cultural evolution - and he considers that many people either ignore or underrate its significance. (LLL, III, p. 153). I believe that these ideas of Hayek's can help us to understand the market, science, and language; they also cast light on the structures of both liberal and socialist societies.

Hayek distinguishes two kinds of rationalism; what he has called constructive rationalism and evolutionary rationalism. And he associates these with two kinds of order: designed or made orders and spontaneous orders. Constructive rationalism derives from Descartes with his twin emphases on logical or mathematical deduction from explicit premises, and on machines as appropriate models for explaining natural phenomena, however complex. According to constructive rationalism, rational actions are those which are determined entirely by known and demonstrable truths, and rational social institutions are those which are deliberately designed to achieve specific, defined purposes.

Constructive rationalism gives rise to designed or made orders, like cars, or silicon chips, buildings or factories, armies or planned economies. All of these have been designed for one or several definite purposes. It is the very success of constructive rationalism in some of these examples - particularly in the less complex situations - that leads to the assumption that all social institutions and all other human productions are, and ought to be, the product of deliberate design.

But such design is neither actual nor feasible. It is not possible for any individual or small group to know all the relevant facts needed to design complex social institutions. To think that this is possible is to suffer from what Hayek calls the synoptic delusion. And many of the social institutions which are indispensable in a modern industrial society have not been consciously designed.

Hence we need to recognise the importance of evolutionary rationalism and of self-generating or spontaneous orders to which the ideas of purpose and design do not apply. Organisms, languages, market economies, societies are orders which were not designed: they evolved. Evolutionary rationalists insist on the distinction between designed and spontaneous orders, especially in understanding man and society.

Man is seen as a rule-following animal as well as a purposive one, and human culture as partly an order of rules which we inherit, and only partly as an order of rules which have been either designed or fully explicated. Many rules and institutions have evolved, and have been strengthened and refined by selection. Man has often been successful because he observed rules, not because he understood why de did so. It is not in any way irrational to follow rules we do not clearly understand. For example, even today we have only a small understanding of the structure of language - yet without language virtually nothing of our culture would exist. So evolutionary rationalists argue that the evolution of social rules and institutions is as important for understanding man and society as is biological evolution for understanding man as a species.

But to recognise this is not to deny the importance of constructive rationalism in limited areas. In almost all real situations, both kinds of rationalism are involved. If we recognise this, we can appreciate more accurately the potential benefits and limitations of conscious design. We shall also, be able, I hope, to distinguish situations where the constructive rationalist model will be most fruitful. A complex self-generating order of individuals, institutions and organisations, which is a modern society, makes continual use of constructive rationalism in limited areas., but in its totality such a society bears little resemblance to a machine. However, if we adopt a constructive rationalist approach and attempt to plan the whole of our society - just as if it were a machine - then we will be moving in a totalitarian direction.

The examples I wish to discuss - the market, science, language, and the structures of liberal societies - all show common features.

1. They all make use of constructive rationalism in limited areas. It is very difficult to think of pure-bred examples of either kind of rationalism.

2. The systems involved are so complex that it is inconceivable that any individual could know all the facts which are relevant to their functioning.

3. Consequently these spontaneous orders have evolved decentralised mechanisms for transmitting information which overcome the limitations of individual knowledge. The development of such mechanisms is a necessary condition for the formation of complex spontaneous orders.

4. A framework of rules is required if the information transmission mechanism of a spontaneous order is to function and the order to survive. These rules are partly explicit and partly tacit, and may in some cases be reinforced by a commonly accepted system of values.

5. By their very nature, spontaneous orders evolve diverse mechanisms for correcting errors or imperfections. These self-correcting mechanisms, which in some ways resemble 'negative feedback loops' in mechanical or electrical systems or homeostatic mechanisms in biological systems, operate at many levels in a spontaneous order, in ways which are scarcely possible in a made order.

The classic example of an evolved or spontaneous order is a market economy. As Hayek puts it- 'We have never designed our economic system. We were not intelligent enough for that'. (LLL, III, p. 164). A market economy is one in which many individuals or organisations may plan their detailed activities by using a constructive rationalistic approach, but which in its totality forms a complex order whose details are not known to anybody. This complex order functions effectively because of one singular fact: the market prices obtaining under competition transmit information throughout the whole system. It is this information, transmitted through prices, which enables each part of the system to respond to the rest and to plan its own detailed activities.

If such an economy is to function effectively, central government must maintain the conditions required for the decentralised transmission of information. Governments must maintain order both internally and externally, and create a legal framework - including laws of property, contract and tort - which allows the market order to function. But the attempt to do more than that may endanger the price mechanism and its crucial role of transmitting information. How much more is permissible on the part of government has therefore been a matter for argument ever since the time of Adam Smith. Questions like, 'Which social services are public goods that cannot be provided by the market?' and, 'How large can the public sector be?' are not easy to answer. What we can say is that governments must not succumb to the synoptic delusion by adopting detailed economic planning, or by administering prices. For they can neither possess the knowledge required to do this, nor explicitly understand the detailed mechanisms which enable modern complex economies to function. Thus, ' . . . if anything is certain it is that no person who was not already familiar with the market could have designed the economic order which is capable of maintaining the present numbers of mankind'. (LLL, III, p. 164).

So far, what I have been saying is neither especially new nor controversial. But I suspect that my next example of evolutionary rationalism - modern science - may be much more contentious. For it is a widely accepted image that modern science is a particularly successful example of constructive rationalism. This image, which dates from the mechanical philosophy of the 17th century and from the writings of Descartes in particular, is influential, superficially plausible, but wrong. More than that; I believe it to be almost the opposite of the truth.

The image has gained its hold for two reasons. First, constructive rationalism has been very successfully applied in science in limited areas; this is especially true of the mathematical sciences. Secondly, scientific papers are written, for very good reasons, in ways which overemphasise the constructive rationalist aspects of scientific work (cf. Sir Peter Medawar's well-known essay - 'Is the Scientific Paper a Fraud?').

But when we consider the whole body of scientific knowledge, and the complex mode of operation of the international scientific community, we see that the constructive rationalist model is inadequate. The body of scientific knowledge is made up of a vast network of theories, experiments and observations, recorded in multifarious papers, monographs and textbooks. It is so vast and complex that nobody can be aware of more than a fraction of it. At the same time, it is not just a random assortment of separate and disconnected items of knowledge. The parts are interrelated in many complex ways, and the whole has a structure which has been designed or made or planned. The whole body of knowledge has evolved in many diverse, unplanned and unpredictable ways, and it has many of the characteristics of a spontaneous order.

These characteristics are even more apparent when we consider the structure, not of scientific knowledge, but of the scientific community and the way it operates. The scientific community is made up of individuals, scientific societies, learned journals and institutions of higher education and research, all of which are in continual interaction. The structure is now very complex: for example, there are estimated to be around 30,000 scientific journals currently being published. This structure has also grown and evolved in a spontaneous way from its rudimentary origins in the 17th century, in parallel with the growth in the body of scientific knowledge over the same period.

Its primary purpose is the transmission of reliable information throughout the whole complex system. It achieves this by using two main criteria; first, logical coherence, and the use of all available relevant evidence in judging claims to knowledge; and secondly, the practice of exposing such claims to open, public criticism.

The second main function of the scientific community is to maintain the conditions which enable this decentralised process of information transmission to operate. It does this by institutionalising the criteria of logic and evidence, and by fostering the values of openness, criticism, pluralism, individual academic freedom and tentativeness. In this way, the whole scientific community can generate and use a vastly greater stock of collectively validated knowledge than any individual or centralised group. At the same time, the system of continual public criticism and evaluation has the great advantage of enabling faulty work to be eliminated.

It is clear that any attempt to plan or direct the course of future scientific development is likely to fail - not only because future discoveries, by their very nature, cannot be predicted, but also because individuals or small groups cannot possess all the details even of our present knowledge; nor can they explicitly understand and control the detailed mechanisms by which scientific knowledge is acquired.

Thus scientific knowledge and the scientific community together form a complex spontaneous order which has many similar features to a market economy. It has evolved rather than been designed, it operates through a decentralised system of information transmission whose maintenance is the prime function of its governing body, and it contains many self-correcting mechanisms which enable errors or mistakes to be eliminated. Moreover, as with the market, it is inconceivable that anyone could have designed the scientific order which over the last three centuries has created the most significant and universal body of interlocking theoretical and empirical human knowledge yet achieved.

At first sight it may seem surprising that two such different areas of human activity as a market economy and modern science should show so many common features. But I believe that this is no accident. I also believe that there are other suggestive parallels and analogies which are worth exploring - analogies between science and language, and between evolutionary and constructive rationalism and the structures and values of liberal and socialist societies respectively. I realise that these are deep waters and therefore I put forward my ideas tentatively in what I hope is the true Popperian spirit of conjecture and refutation. In other words, I offer my conjectures, and invite my readers to refute them.

Human languages are spontaneous orders which have many important similarities with science - particularly if we accept the view of science as a redescription of the world. Natural languages are not designed; they are the products mainly of evolutionary rather than constructive rationalism. The artificial languages for which this is not true - mathematical and computer languages - are special cases, which lack many of the essential characteristics of natural languages.

Languages deal with and describe the natural world, a world which is so complex that any individual attempt to describe it, and make sense of it, can only capture part of it. In order to survive, each individual must make some sense of his environment, most fundamentally by acquiring a language. But the language of each individual (his idiolect) only functions effectively if it forms part of a wider structure such as the language of a group, a region or a nation. So our languages are complex decentralised mechanisms for transmitting information. And we use them confidently without much explicit understanding of their structure or of how they develop. A framework of rules governs the forms which natural languages take. But these rules are concerned with the structure rather than the content of what we say. This is true both for the socially transmitted grammatical rules which are specific to each language, and for the deep structures or linguistic universals which Chomsky claims are common to all human languages (cf. Geoffrey Sampson, Liberty and Language, Oxford, 1979). Hence languages exhibit the same common features as the other spontaneous orders I have mentioned. Nor is the analogy destroyed by the fact that in this case part of the governing framework of rules - Chomsky's deep structure - appears to be innate rather than man-made.

There is a further respect in which languages resemble markets and science. Like these last two institutions, they incorporate self-correcting or error-eliminating mechanisms. Linguistic forms which do not lead to effective communication tend not to persist; more successful forms persist, and gradually evolve into more permanent features of the language. Natural languages have the interesting property that partially incoherent or ungrammatical statements can transmit some information, although they may also lead to misunderstandings. This is not a property of such products of constructive rationalism as computer languages, where one error is often enough to destroy completely the meaning of a statement. Again, unlike computer languages, the very complexity of natural languages may lead to differences in interpretation and a lack of communication - as when two people use the same word in different ways.

Yet another similarity exists between language and other institutions of spontaneous order. This is the essentially personal character of language: the fact that much of our linguistic experience and knowledge is personal, tacit and never fully conveyed to anyone else. Similarly, much of scientific knowledge and experience had a strong personal element of this kind (a point particularly emphasised by Michael Polanyi in his books The Tacit Dimension and Personal Knowledge).

I have already stressed the importance of evolutionary rationalism in understanding modern complex societies. I indicated that such societies, while they may use constructive rationalism in limited areas, in their totality have evolved rather than been designed. Moreover, they are so complex that they need to develop and maintain decentralised mechanisms for information transmission; and these mechanisms provide such societies with adaptive or self-correcting properties.

I want now to try to illustrate these points by contrasting two very different kinds of society - liberal (capitalist) societies and socialist (Marxist) societies. I suggest that the fundamental differences between liberal and socialist societies arise because liberal societies depend primarily on evolutionary rationalism and spontaneous order, whereas the structures of socialist societies take constructive rationalism and designated order as their model.

Ideally, liberal societies are fundamentally decentralised. All the main spheres of human life - economic, political, educational, cultural, religious - are partially separated. And within each sphere, pluralism is, to some extent, both encouraged and realised - diverse centres of economic power, many political parties, diversity of provision in education and other cultural spheres, many coexisting religions. The resulting complex order is underpinned by the values of tolerance, pluralism and individual freedom, and, in particular, by freedom of expression, communication and access to information. It is clearly impossible to plan the overall development of such a society even if short-term developments in some areas can be planned.

By contrast, Marxist societies are centralised and monolithic. Their governments attempt to control all aspects of life. They have centrally planned economies and only one political party possesses any substantial reality. This party controls all educational and cultural activities. Freedom of expression and of access to information are expressly prevented, and the values which are encouraged are those of collective commitment to the centralised rule of the party. Anything which the party does is to be approved and anything which threatens its central role is condemned and suppressed. Marxist societies have clearly tried to organise themselves according to the principles of constructive rationalism. But, as I have argued, such principles are antithetical to the effective functioning of many essential institutions of modern complex societies. So there develops a great gulf between promise and performance in Marxist states. And this gulf persists because of the lack of self-correcting mechanisms in a system based on constructive rationalism.

This analysis may help to explain why the Gulag Archipelago has lasted so long and cost so many millions of lives, why the charlatan Lysenko ruled Soviet biology for more than forty years, and why the Great Soviet Encyclopaedia still does not list five of the eight men who have been premiers of the USSR since the revolution.

All the examples of spontaneous orders have one major common characteristic - they are all concerned with how we gradually learn about our environment and how we adapt to it. Successful learning and adaptation are facilitated if the environment is relatively stable. So this partly explains why the kind of knowledge we have been able to acquire about the relatively stable natural world - via language and science - is of a more durable and cumulative kind than that which we have acquired about the more plastic social and political worlds. And the knowledge or information transmitted by the market is the most transient of all, reflecting as it does the current economic situation. But this does not mean that all economic knowledge is transient any more than the transient nature of the spoken word implies that we can learn nothing enduring about language. If my analysis is valid, it should be just as possible to build up a cumulative body of knowledge in economics as it is in linguistics.

Therefore I would contend that we can aspire to appreciate how the market and other complex spontaneous orders function. The work of Hayek has made is possible for us to generalise some of the insights of Adam Smith, in ways which it may be vital for us to understand if our society is to survive. It has also displayed the dangers involved in trying to organise societies according to the constructive rationalist model, as socialists advocate and practise. I shall conclude with some of Hayek's own words:

If our civilization survives, which it will do only if it renounces those errors, I believe men will look back on our age as an age of superstition chiefly connected with the names of Karl Marx and Sigmund Freud. I believe people will discover that the most widely held ideas which dominated the twentieth century, those of a planned economy with a just distribution, a freeing ourselves from repressions and conventional morals, or permissive education as a way to freedom, and the replacement of the market by a rational arrangement of a body with coercive powers, were all based on superstitions in the strict sense of the word. An age of superstitions is a time when people imagine that they know more than they do. In this sense the twentieth century was certainly an outstanding age of superstition, and the cause of this is an over-estimation of what science has achieved - not in the field of the comparatively simple phenomena, where it has of course been extraordinarily successful, but in the field of complex phenomena, where the application of the techniques which proved so helpful with essentially simple phenomena has proved to be very misleading.

Ironically, these superstitions are largely an effect of our inheritance from the Age of Reason, that great enemy of all that it regarded as superstitions. If the enlightenment has discovered that the role assigned to human reason in intelligent construction had been too small in the past, we are discovering that the task which our age is assigning to the rational construction of new institutions is far too big. What the age of rationalism - and modern positivism - has taught us to regard as senseless and meaningless formations due to accident or human caprice, turn out in many instances to be the foundations on which our capacity for rational thought rests. Man is not and never will be the master of his fate: his very reason always progresses by leading him into the unknown and unforeseen where he learns new things.

To which it is fitting to add the words of Isaac Newton - the man whose scientific work inspired many of the men of the Enlightenment but who as a lifelong opponent of that prime advocate of constructive rationalism, Descartes:

I do not know what I may appear to the world; but to myself I seem to have been only like a boy playing on the sea-shore, and diverting myself in now and then finding a smoother pebble or a prettier shell than ordinary, while the great ocean of truth lay all undiscovered before me.

John Marks, London 1985