Consider the following paradoxes. A welfare system designed by well-meaning politicians guided by the advice of the wisest sociologists and economists available, costing billions of dollars, whose net effect is radically to increase the numbers of the poor, especially women and children, and to deepen their misery, incapacity, and despair. A stock market which rises because the statistical instruments designed to detect similarities with previous rises are causing investors to make it rise in the same pattern and which helps to generate the financial conditions it predicts. A social polity expressly created to ensure the equality of all citizens, which produces an archipelago of concentration camps across a continent and whose theoretical dismissal of the concrete effectuality of theorizing unleashes real social forces of unparalleled savagery. A foreign policy which depends for its effectiveness on the government’s ignorance of its existence and implementation. An economy which attracts foreign investment by borrowing so much money that it is able to remain politically stable and thus economically healthy.

More and more of our collective life seems now to be populated with such logical monsters, such scyllas and charybdises of reflection and feedback. Yet good as well as evil can be compounded by the peculiar kind of interest which they offer; unfairly, unto him who has much, much shall be given and the kingdom of heaven is like a mustard seed: something that will grow quite unpredictably all over the place.

But these monstrosities are the despair of any “scientific” sociology or historiography. And now physics itself seems to have caught the plague; and even that purest sanctum of linear logic, mathematics. Those positive knowledges to which modernist history and sociology appealed for a model now seem almost as messy and chaotic as the seething life of human culture.

This new vision of the “positive sciences” has emerged from the brilliant new studies of chaotic, nondeterministic, recursive, fractal, dissipative, catastrophic, period-doubling, and feedback-governed systems, associated with the names of Mitchell Feigenbaum, Ilya Prigogine, Benoit Mandelbrot, and Rene Thom. Perhaps some of these terms require a brief (and necessarily incomplete) explanation. An algorithm is a mathematical method for doing something—say, generating a geometrical shape in a computer graphic. A recursive algorithm is one which possesses an internal loop, such that the solution arrived at by one passage through the loop is fed back again into the beginning of the loop, “adding,” as Benoit Mandelbrot puts it, “fresh detail to what has been drawn on previous runs.” Mandelbrot also gave the name fractals to a family of shapes, irregular and fragmented surfaces, curves, and “dusts,” generated by recursive algorithms based on a random or arbitrary numerical “seed,” which repeat their own form or type of form at different scales of magnification, so as to pack into their details at one scale a microcosm of the next larger scale. The space-filling curves of Peano are only one example. Mandelbrot sees these forms everywhere in Nature: in trees, cloudscapes, coastlines, the bronchi of the lungs, corals, star clusters, waves, craters, and so on. A dissipative system is one which maintains its form not despite its tendency to decay but by means of it. Dissipative systems can be self-organizing; I shall discuss some examples later, such as certain forms of turbulence. The term is Prigogine’s. A catastrophe is a discontinuity, as when the gradual increase of some factor suddenly crosses a threshold in which some entirely different state is precipitated; it can be observed when a cooling supersaturated solution suddenly crystallizes, or when an animal’s behavior suddenly changes during a gradual change in the stress it is undergoing, or when a gradual change in economic factors triggers a massive move in the stock market index. Rene Thom was first able to describe such discontinuities or catastrophes mathematically, in his catastrophe graphs. Period doubling—Feigenbaum’s term—is what happens when certain ordered systems break up into chaotic ones, like a smoke-ring dissolving in the air; out of such chaotic situations, however, new forms of order can arise spontaneously, given the right circumstances.

The lawfulness governing such systems is of a radically different kind from the rules that govern classical deterministic systems embodied in the empirical causal logic of the modern scholarly humanistic disciplines, hi other words, if even the sciences themselves no longer insist on a causal mechanism for events (and its attendant rules of objective and positivistic empirical evidence), then it is high time the social, historical, and humane studies reevaluated their scholarly methods. The indeterminacy of quantum physics was hard enough for the academy to swallow. The new indeterminacy is of quite a different kind.

What the new science shows us is that the operation of fairly simple processes—the period-doubling mechanism of turbulence, for instance, or the random walk of particles precipitating a crystal—can very rapidly bring about states of a system that are utterly unpredictable from their initial conditions. In a computer simulation of planetary orbits, for instance, there is an unstable zone in which the velocity and proximity of a satellite to its primary is critical to whether it will settle into a stable orbit, whether it will escape altogether, or whether it will adopt an eccentric, continually changing looping orbit around its primary. Each time the initial velocity or position is defined to a further decimal place, the resultant orbit is radically different—not different so as to form a convergent series homing in on an asymptote, but utterly and unpredictably different. Thus the accuracy by which the world is defined makes a total difference to the nature of the world itself. A seacoast measured with a one-mile ruler might be hundreds of miles long; if measured by a foot ruler, thousands; if by a micron ruler, millions; and each level of magnitude has its own lawfulness and predicts its own pattern of wave action as the surf rolls in.

Given their unpredictability, such processes might be expected to bring about mere chaos, mere ugly inchoateness. But no. Often enough they resolve themselves into extremely beautiful, complex, and stable structures, to which I shall give the generic term “paisleys.” Such forms are coming to replace the classical shapes of ideal geometry—lines, triangles, circles, regular solids—as the governing imagery of the scientific visual imagination. Examples range from the convection cells of a good rolling boil in a teakettle or the planetary pattern of trade winds or Jupiter’s Great Red Spot—a storm that has raged for hundreds of years—to the forms of electrical discharge, crystals, river drainage systems, and organic structures. These systems forget their causes, and indeed if their causal determination were the only language in which they could be understood, they would be inherently unintelligible. The “modeling” or “generative” logic by which they are now understood is profoundly new as mathematical formalism but, as I hope to show, very ancient as an intuitive human-activity.

The test of whether we truly understand such a system is no longer our ability to predict it but our ability to construct another system that does the same sort of thing as the original. Perhaps we could say that we still test by prediction, but what we are predicting is not a certain future state of the system but the general type of behavior of the whole system itself. In other words, we are not predicting along a line of time, but across a sort of plane. And this notion, of other temporal geometries than the linear, has enormous implications not only for the study of history but for the arts and humanities in general.

The common feature in all these systems is feedback. The simplest forms of feedback are given in the initial conditions, for instance the setting of the thermostat of a home heating furnace. In this case the only unpredictable element is the precise value of some parameter—in our example, the temperature of the house at a specific time, in the course of its wanderings up and down around the “attractor” or average temperature we have set for it. More complex feedback systems can set their own parameters or even create the sort of parameters toward which they aspire or around which they oscillate. That is, their state at any given moment is the resultant of ordering processes that have arisen within the system itself.

Now the supreme example of such self-organizing systems is life. And we may go further and say that the evolution of life has been the evolution of more and more autonomous and complex and unpredictable—because inventive—feedback systems. The human species is, as far as we know, the most advanced state of this process, where it shows in its most paradigmatic and articulated form the general tendency it always exhibited. Nature strives toward freedom, in the sense of autonomy, as the clearest expression of its essence. And here of course we return to our subject, which is the present crisis in history and social studies.

For surely, a fortiori, the collective activities of human beings are of all phenomena in the world the most fully governed by the principles of complex feedback systems. Social game theory takes us part of the way. Consider a simple dyadic predicting contest—a little marital spat on a Sunday in some large American city. There is tension between Jack and Jill; they haven’t been able to talk much recently, because they have both been working hard. Jack, as usual, intends to get Sunday lunch. Jill, however, knowing what Jack has in mind, intends to shop for lunch at the deli instead, in order to upset him. But Jack knows his Jill. Guessing that she intends to go out to the deli in order to forestall Jack’s usual lunch. Jack plans to claim that he is feeling ill and doesn’t want lunch. Jill, though, expects the “I feel sick” ploy and finds occasion to joke pleasantly about Jack’s past propensity to use pretended illness to get out of things. Jack, recognizing that the game has got too complex at this level, changes levels by deliberately randomizing his own behavior. He starts a tedious conversation about fatal illnesses. Jill is flummoxed only for a moment, then recognizes the paradigm—or genre-switch. She plays the same game, but without any pretense at normal conversation, breaking in with some earnest remarks about chickadee nesting habits. She has thus thematized the issue of avoiding the subject and changed the ground rules once again. Jack now steps outside of the conversation and looks at it as a stranger might; no longer as “Jack versus Jill” but “Jack and Jill versus the outside world.” He sees how absurd they sound, catches Jill’s eye, and they both collapse in laughter. No doubt they will go out to lunch at their favorite cafe. Or maybe not.

Jack and Jill have become a “we” by internally modeling each other’s motivations and each other’s image of the other. In the process they have touched on a broad and sensitive range of values and value judgments. Their story is not atypical; the narrative and dramatic arts are full of this sort of thing, and Erving Goffman and Thomas Scheff, among others, have provided close analysis of it. Now imagine it extended to the billions of dyadic relationships in the human world, and the trillions of larger group relationships. Nor are such strategies unique to intimate relationships. One could easily demonstrate the same sort of strategy at work between, say, two opposing generals in the wars of the Austrian Succession. All human interactions significantly involve such halls of mirrors.

Our first observation, that it was an error to apply to human affairs the deterministic logic of classical science, would if left on its own imply that we might as well give up on such studies altogether. But there is now a body of theory and concept that can put them on a new footing. In the remarks that follow, I shall pay special attention to historical and social studies, but I do so to let part stand for whole; much of what I have to say applies broadly to the humanities, the academic study of the arts, and the human sciences as well.

For want of a nail, the kingdom was lost. Imagine the predicament of a historian, reminded, by some trivial historical episode with momentous effects, of the insecurity of the discipline of history. The tiniest event can snowball into the most gigantic consequences, just as the minutest subatomic difference in a flow can result in an utterly altered pattern of turbulence. Philosophers of history have never been able to demonstrate that this snowball effect cannot take place. Perhaps every event that occurs is just as crucial, and just as insignificant, because undifferentiated in importance from everything else. Perhaps this is the specter, the existential cackle of empty laughter, that haunts certain historians, that drives them to construct their elaborate deterministic edifices of economic and social history, class struggle, invisible oppressors, conspiracy theories. Like Casaubon in Middlemarch, they set out to uncover the Key to All the Mythologies.

Another kind of historian with a different temperament, confronted with the appalling indifference of historical significance, will seek to enumerate all the primary sources, to recite all the “facts,” to deal with all exceptions to all rules, all special eases, all the statistics, and to do it without bias, without giving any one fact more significance than any other. It is as if one should seek an understanding of a turbulent flow by listing and mapping all the positions of all the particles in the flow at all times. Perhaps if the map is on a fine enough scale, the answer will emerge. The mapmakers of Borges’ mad dictator who made a map of the country so perfect that when opened it covered the country itself and brought on its economic ruin, or the mole in Kafka’s story who kept building new tunnels to keep watch on the entrances of his burrow, are literary examples of this mindset. Such heroic historians, fixated on the old intellectual modes, accumulate a tragically meaningless scholarship. Given this approach, why should not a life of scholarship which devoted itself to a descriptive catalog of every spot on the library wall—and there are enough spots on any library wall, if we choose small enough parameters—why should not such a labor be just as valid as Darwin’s collection of evidence for evolution, or Albert Lord’s for oral composition? Why is one fact more significant than another?

Not unlike the collecting moles in their theory, but different from them in practice, are those postmodernists and deconstructionists who accept the complexity and interdependence of the world but refuse to recognize the real stable order that it also generates. Unable to escape their own Oedipal and patriarchal model of knowledge, which insists that the identity of something derives from what originated it, they regard order as an illusion because that order originated out of chaos. Self-made men and women, they are as horrified by the idea of the self-made—by the made—as by the primal scene itself. For them to make, to create, is a fascist imposition of a totalizing structure upon the freeplay of the world. They thus abolish the idea of the writer, the maker, the text, the made thing, even the reader, even the world. They seize on quantum theory as a sort of warrant for a deconstructed and valueless universe—quite erroneously, of course, as indeterminate particles happily clump together to make very determinate pieces of matter. Those determinate pieces of matter indeed evolved sometimes into self-ordering and even free systems; but this does not help the deconstructionists’ case.

If the responses of the historical determinists, the obsessive collectors, and the rebels without a cause are inadequate to the problem of history, what approach might really work? The beginnings of an answer to this question are what this essay proposes; to get there we must follow a somewhat winding path of dialectical reasoning.

Any analysis of historical events that we make, or any theory of social behavior we formulate, is itself one of the determining factors in the situation it describes. Thus there is no “meta” position, no detached Olympian viewpoint from which objective assessments can be made, and therefore no escape from the apparent chaos of mutual feedback. We are all revolutionaries and reactionaries, whatever our claims as historians or social scientists. Economists are just another group of competitors over what constitutes value.

Not that this struggle for ontological control is a blind one. We would be totally ineffective at it if we were not able to assess the motives and assume the world view of others. And even this would not be enough. Our imaginative model of the other must contain its own image of oneself—the gift, said Robert Burns, is to see ourselves as others see us; and that image itself must contain its own assessment of the other. And our outer negotiations take place not just between our own persons but also among the entire dramatis personae of the inner drama by which we estimate the future. The confusion is not one of blindness but of too much sight; not of randomness but of an excess of determinants; not of chaos but of an order too complex to be explained before the next complicating event comes along—of which the next, complicating, event is the best explanation.

Indeed, this capacity to impose our interpretations on things is not only our predicament but also what enabled us to second-guess, predict, and control the simpler systems of nature, such as the biological, chemical, and physical ones. We bought our power over the rest of nature with the essential uncontrollability of human events. We can control nature to the extent that we stay one step of reflexivity ahead of it. Nor is even nature innocent, but itself the resultant and living history of a cosmic evolution which pitted many forms of reflection against each other; the marvelous cooperation of nature is a prudent and subtle form of mutual feedback. Even so, when we find we can reduce another organism to a successfully testable set of laws and predictions, it is a sign that we are dealing with a lower order of reflections than our own.

Thus to attempt to do so with human beings, to educe and apply the laws governing them and to predict their actions, is, in human terms, a viciously aggressive act, an attempt to get control at the expense of others’ freedom. It implicitly reduces human beings to the level of things, of lower animals. But this indeed is what much social and economic history, much sociology and progressive political theory, has attempted to do. The promise such studies held out was not lost on those with the sweet thirst for power. Transformed into political programs, those systems appeared in our century as the great totalizing regimes—Marxism, Fascism, National Socialism, International Socialism. We should not be surprised at the vigorous counterreaction of human cultures against such systems.

In the light of this analysis it now becomes clear why with the best will in the world all principled revolutions have ended up diminishing human variety and freedom in their societies. For a revolution to be truly freeing it must be unprincipled, in the sense that its intentions do not rest on a predictive theory of human social behavior. Such a revolution was the American, whose ideas, enshrined in the Constitution, really amount to a declaration of regulated intellectual anarchy. The principle of separation of powers, which is, more than equality and more even than democracy, the central message of the Constitution and the thematic undertone of every article, is an intuitive recognition of the reflexive, self-organizing, unpredictable, feedback nature of history, which by reinterpreting its initial conditions is able to forget them.

Separation of powers makes politics into a drama, not a sermon. Perhaps the true hidden presence behind the Constitution is William Shakespeare. All the world’s a stage. We are all actors, in both senses of the word. Our inherent value derives from that condition, not from Kant’s notion that we are ends in ourselves. We can still keep our dignity even if we are, for immediate purposes, means, as long as we are actors in the drama. Even if their function is to serve, the crusty boatman or witty nurse or pushy saleslady are interpreting the world from their own center, are characters, dramatis personae, to be ignored by others at their peril, and are thus free. We might, parenthetically, therefore view with alarm the tendency in modern and postmodern theater to get rid of characters altogether. But of course even this formulation which I have made is itself a part of the situation it describes; it is a speech in the play, to be evaluated by your own reflexive processes of assessment. Let us see whether the line of thought it prompts is more or less liberating than its competitors.

We immediately run up against a large problem. Does this critique of historical and human studies mean that they must revert to the status of chronicle and appreciative observation? Like amateur naturalists, must their practitioners only be collectors, without testable hypotheses or laws? Should we just admire the exquisite coiled turbulence of human events, wonder, and move on? The French historian Fernand Braudel is almost such a historical naturalist; there are moments as one contemplates his great colorful slowly roiling paisley of mediterranean history, seemingly without direction or progress, that one could wish for little more out of history. Should not the historian be a sort of Giacomo Casanova, a picaro among the courts and sewers of eternal Europe or China, remarking the choice beauties to be seen on one’s travels?

A directionless view of history can be seductive. But even if the essential logic of the modern humane disciplines is utterly erroneous, it has nevertheless provided an impetus and direction for research and has led to the vigorous discovery of huge masses of information, at least some of which is interesting to everyone. The bias of that information, the preponderance of certain types of source and the direction of the researchers’ gaze, may be corrupting; but in itself we feel it to be valuable.

But let us explore the possible value of the naturalist’s or chronicler’s agnosticism. Although it might not wish to own up to it, deconstructionism actually presents a rather good case for this perspective—to the extent that a case as such, with all its theoretical baggage, can be made for so uncaselike an approach. Deconstruction is purposely not long on logic, and as such it is quite consistent. The bête noire of all deconstructionists is totalization. What does totalization mean? Once we have disposed of those cases of totalization which every sane person would deplore—Nazism, for example—we get into interesting territory. What makes deconstruction unique is its inability to distinguish between those forms of order we would all agree are evil and such things as the narrative structure of a text, marriage and family, the idea of the writer and the reader, even the very idea of the self or person. There is no plausible place for deconstructionists to stop on their slide into total inarticulacy. Poor Jacques Derrida, nailed recently to the wall by the inspectors of ideological purity on the subject of South Africa, was forced to squirm to reconcile the indifference of his skepticism towards all forms of order, his fundamental belief in the radical apartheid of all points of view, and his decent liberal distaste for the regime.

Certainly history and sociology would be easy meat for a deconstructionist’s acid test; but so would any human or indeed natural product, process, or action. Deconstructionism has now begun to turn its acids on itself; as it does so, it will encounter the paradox of what container to keep the perfect corrosive in. And if it is not the perfect corrosive, deconstruction must end up, like its old enemy Descartes, asserting with more totalizing violence than any other system that it is the one idea which is not subject to the deconstructive process: in Descartes’ system, the cogito ergo sum, “I think, therefore I am,” isolated by his skepticism about all else; in Derrida’s, that force or energy he perceives as prior to and underlying all difference.

But there is a rather benign factor to deconstructionism, to be found, for instance, in Jean-Francois Lyotard’s classical essay on postmodernism. Here he offers a way of thinking about human society that makes no generalizations and which recognizes all human activities and thoughts as flows in a great interacting soup of information. On the face of it, a very attractive vision; and it satisfies some of the criticisms we leveled earlier at history and the social “sciences.”

But in doing so it abdicates that very activity—holistic understanding and the enrichment of the world by interpretation—that characterizes the human Umwelt itself. The admonition not to totalize is the most totalitarian command of all, since it essentially dehumanizes history. The feedback process of human culture is a feedback of what deconstructionists would call totalizations. The open-endedness of history is created by the competition and accommodation of various candidates for the last word, the dernier cri, the formula of closure (including this one); it is an ecology of absolutisms. Nor is this ecology a random play of flows, without direction or growth; technology, records, and enduring works of art constitute ratchets which prevent any return to earlier, less complex states of the system, just as genetic inheritance did in earlier ages. Thus history is an evolutionary system, with the three factors required for evolution to take place: variation (provided by the unpredictable paisley of reflexive events), selection (provided by the competition and accommodation of “totalizations”), and inheritance, a conservative ratchet to prevent what is of advantage from being lost.

The only way open is to seek principles of understanding and descriptive categories that are proper to our own level of reflexive complexity. To do this is essentially an artistic, constructive, performative, and religious activity, and it cannot fully depend on the capacity for calculation by which we claim to understand the rest of the natural world. (Even this claim must yield at a certain point. Ultimately scientists appeal to the beauty of a theory to justify it before the infinite plenum of its equally consistent rivals.) History is an art, even a technology, or a liturgy as much as it is a science; and it is so not only in the activity of historiography as research.

In other words, I am proposing a change in our fundamental paradigm of historical and human study. And here another set of major scientific advances comes into play. Most workers in the historical and sociological fields still accept the cultural determinism that was one of the first naive responses of the West to the cultural diversity of the newly discovered non-Western world. Thus for them the units of historical study, human beings, are tabulae rasae, to be inscribed by cultural conditioning or economic pressures.

More recently, however, in fields as diverse as cultural anthropology, linguistics, twin-studies, paleoanthropology, human evolution, psychophysics, performance studies, neuroanatomy, neurochemistry, folklore and mythology, and ethology, it is becoming clear that we human beings bring to history and society an enormously rich set of innate capacities, tendencies, and exclusive potentials. We uncannily choose, again and again, the same kinds of poetic meters, kinship classifications, calendars, myths, funerals, stories, decorative patterns, musical scales, performance traditions, rituals, food-preparation concepts, grammars, and symbolisms. We are not natureless. Indeed, our natures include, genetically, much of the cultural experience of our species in that period of one to five million years of nature-culture overlap during which our biological evolution had not ceased, while our cultural evolution had already begun: the period in which unwittingly we domesticated and bred ourselves into our humanity. The shape and chemistry of our brains is in part a cultural artifact. We are deeply written and inscribed already, we have our own characters, so to speak, when we come from the womb.

So having taken away one kind of rationality from historical and human studies, we may be able to replace it with another. But in so doing are we not committing the very sin—reducing a self-organizing and unpredictable order to a set of deterministic laws—of which we accuse the determinist historians? Are we not replacing cultural or economic determinism with biological determinism? Not at all. First, to understand the principles governing the individual elements of a complex system is, as we have seen, not sufficient to be able to educe laws to predict the behavior of the ensemble. The beautiful paisleys of atmospheric turbulence are not explained by the most precise understanding of the individual properties—atomic weight, chemical structure, specific heat, and so on—of its elements. Second, the peculiar understanding of the human being that we are coming to is that of a creature programmed rather rigidly and in certain specific ways to do something which is totally open-ended—that is, to learn and to create. Our hardwiring—whose proper development we neglect in our education at great peril—is designed to make us infinitely inventive. Our nature is a grammar that we must learn to use correctly and which, if we do, makes us linguistically into protean gods, able to say anything in the world or out of it.

Thus the paradigm change which this line of argument suggests is from one in which a social universe of natureless, culturally determined units is governed by a set of causal laws which, given precise input, will generate accurate predictions, to one in which a cultural universe of complex-natured but knowable individuals, by the interaction and feedback of their intentions, generates an ever-changing social pattern or paisley, which can be modeled but not predicted. The meaning of understanding would change from being able to give a discursive or mathematical account of something to being able to set up a working model that can do the same sorts of things as the original.

Fundamental political concepts like freedom, war, civil order, equality, literacy, power, justice, sovereignty, and so on would no longer be defined in terms of a set of objective abstract conditions but as living activities in a one-way unrepeatable process of historical change. It would be such a revaluation as occurred in literary criticism in the 19th century, when tragedy came to be defined as a process, an organic and recognizable activity, rather than as conforming to such rules as the Three Unities. Conceivably the automobile has done as much to create political freedom as any set of laws; yet historical and political scholarship is taken by surprise by such relationships. Imagine how governable the Soviet Union would be if everyone had a car—or a personal computer with a modem; glasnost may be more than what the party bargained for. Is not justice very much a matter of talent and personality? Blake said: One law for the ox and the lion is oppression. Might there not have been more equality of certain valid kinds between a gentleman and his valet than between an employer and employee in a classless society? Is not power the most questionable and fugitive of all concepts, seemingly so solid at one moment but blown away by unpopularity the next? What is war in an age of terrorism, export dumping, military computer games, and nuclear standoff?

Such questions are not intended to indu