"The End of Certainties in the Social Sciences"
by Immanuel Wallerstein (iwaller@binghamton.edu)
© Immanuel Wallerstein 1997.
[You are free to download this paper or send it electronically to others. If you wish to translate it into another language, or to publish it in a printed medium or on another web site, you must obtain formal authorization from the author.]
[Talk at the Seminario, "Conceptos en Ciencias y Humanidades," Mexico City, Oct. 16, 1998, co-sponsored by CEEICH (UNAM), Fernand Braudel Center (Binghamton), Maison des Sciences de l'Homme (Paris), ZiF (Bielefeld)]
How to live with uncertainty is probably one of the oldest social problems humankind has faced. Ten thousand years ago, when human beings knew less about the physical and social world than we do today, the uncertainties of life were probably quite frightening to humans. They could not anticipate very well either short-run or long-run changes in their natural environment. They were unsure about whether they and their families would find even the necessary food and safety to survive immediately. They were unsure how soon, and in what form, they would be facing deadly enemies, animal or human. They may perhaps even have blamed themselves for bringing about this unwanted uncertainty. This is what we probably implied in the myth of the "expulsion from the Garden of Eden."
Such uncertainty was doubtless very socially destabilizing, and the doubts themselves could only have reinforced the dangers. To reduce the dangers, therefore, humans appealed to sources of certainty: magic and magicians, the gods and their priests, collective and communal authority and those who incarnated and exercised this authority. This worked, up to a point. It reduced doubts and fears, and thereby helped to stabilize social structures. But of course, all these sources of certainty were able to predict correctly, or even explain retroactively, only up to a point. There were endless surprises, and some of them were quite serious. Still this was a self-referential model of analysis. When predictions turned out to be wrong, when unexpected and seriously disturbing happenings did occur, what was most often blamed was not the possibility of certainty nor even the system of ascertaining certainty but usually the practitioners of the art of certainty - the magicians, the priests, even the emperors (who, in the Chinese phrase, were deemed to have lost the mandate of heaven).
The advent of the modern world-system, the capitalist world-economy, required for its efficient operation a somewhat higher level of accuracy in forecasting, without which the investment process so central to its functioning would never have acquired the extensiveness and level of risk-taking that has enabled it to expand and flourish. There was consequently considerable social support and social sanction for a new mode of certifying truth, the mode we have come to call science, or more accurately modern science.
The scientists had to create space for science within a world whose cultural values were still based on prior modes of validating certainties. this creation of room for science was actually a two-stage process. First, the philosophers attacked the meaningfulness of revealed truths (those that could be known only, or at least best of all, to priests and clerical structures). The philosophers insisted that all humans had the innate possibility to reason individually and thereby to perceive truth. To be sure, they acknowledged that this was not easy, and that some persons (namely, the philosophers) were able to do this much better than most other people. But the philosophers were primarily interested in denying the right of either religious authorities or political authorities to proclaim truth. One might say that this was the primary cultural message of what we now call modernity, and that this message was more or less successfully translated into everyday belief over the last five centuries, and accepted by most people throughout the world.
Science, as a cultural activity, represented a further specification of the philosophers' claim to universal human rationality as the source of truth statements. The scientists raised the question, how do we know whether any given person's claim to have discovered truth by reasoning is valid, especially given that there are multiple and competing claims to discoveries of truth? The answer of the scientists was that truth claims had to be validated by empirical evidence, which had to be collected in specific ways that could be replicated by fellow members of the scientific subcommunity. In effect, the scientists were insisting that it was not any philosophers but only a subgroup of them, the scientists, who had the moral right to receive recognition as the source of valid truth-claims. In fact, by the nineteenth century, the scientists had largely won out in this cultural war and did indeed receive social acceptance as the only persons to whose thisworldly truth-claims deference should be paid.
There was a curious feature to this scientific credo. On the one hand, they proclaimed theoretically a total rejection of authority as the basis of legitimating truth-claims, and a total egalitarianism about who might put forward legitimate truth-claims.
There was, it was argued, a free marketplace of ideas. Any and all persons might offer their truth-claims in this market and present whatever evidence they had to persuade others. And then, somehow, the truth-claims might be accepted as valid by the community or be rejected. There were no a priori limits, and the antiquity of a prior truth-claim, now contested, was no argument in its favor.
But on the other hand, in practice scientists did not mean it. They did not actually believe that all authority was illegitimate, and that any and all persons might put forward truth-claims in the marketplace of ideas. They actually believed that the small subcommunities of specialists in every defined subfield of science constituted a collective authority which, although not infallible, should be deemed correct unless there were overwhelming evidence to refute them. And they actually believed that, with rare exceptions, only persons with certain specialized training were entitled to be taken seriously when and if put forward truth-claims in the marketplace of ideas. The scientists, when challenged, pointed to the fact that although, in reality, they were limiting entry, they were in principle (and every once in a while in practice) ready to make exceptions. Still, to a distant observer from another planet, the deference paid to scientists in the twentieth century might not have seemed all that different from the deference paid to magicians, priests, and communal authorities in times of yore.
This applied as well to the question of whether or not the truth-claims they were making represented "certainties" or merely "probabilities." Whereas scientists promulgated obstinately the virtues of skepticism ("how do you know what you allege to know?"), and the limitations of current knowledge ("all truths are provisional"), they also insisted that intrinsically certainty was possible, and that therefore there could come a day when everything would be known about everything. This was the image of the deterministic world, so central to what we call modern science. God does not play dice with the universe, said Einstein, expressing the deepest commitment of modern science. Determinism was the centerpiece of Newtonian mechanics, itself in turn long considered to be the fundamental scientific program, the model for all other scientific efforts. Determinism was conjoined with linearity, equilibrium, and reversibility to add up to a set of minimal criteria for calling theoretical explanations "scientific."
We are all aware that, in the last 100 years, and particularly in the last 30 years, this Newtonian model of science has been under sustained and severe challenge within the belly of the beast, within physics and mathematics themselves. It is not my role here to outline this challenge. I shall simply point to the counterslogans of this challenge: in place of certainties, probabilities; in place of determinism, deterministic chaos; in place of linearity, the tendency to move far from equilibrium and towards bifurcation; in place of integer dimensions, fractals; in place of reversibility, the arrow of time. And I should add, in place of science as fundamentally different from humanistic thought, science as part of culture.
I wish to talk about the impact on social science of this challenge to Newtonian science that comes from within science. And I wish to speak of what kind of social science we can build that is grounded in the realization that we have reached the cultural end of certainties. Social science was institutionalized in the late nineteenth century in the shadow of the cultural dominance of Newtonian science. I shall not review here the internal methodological struggles of the social sciences, as they sought to carve a space for themselves amidst the two-culture split between science and the humanities. Suffice it to remember that, in this so-called Methodenstreit, the three principal disciplines that were created to deal with the modern world - economics, political science, and sociology - all opted to be nomothetic, by which they meant replicating to the extent possible the methods and epistemological worldview of Newtonian mechanics. The other social sciences thought of themselves as more humanistic and narrative but were nevertheless attempting in their own manner to be "scientific." The humanistic scholars embraced the emphasis on empirical data but cavilled at the idea of "generalizations."
In the period after 1945, all the social sciences, but especially the nomothetic trio, became increasingly quantitative and insisted very strongly on the presupposition of a deterministic social universe. The object of social science, they reiterated, was to discern universal covering laws akin to those they believed physics had been able to state. The main problem they encountered was that, in practice, they were unable to make even short-run predictions that turned out to be sufficiently correct to merit social applause. When pressed, social scientists attributed these failures to their collective immaturity as serious science. In short they blamed themselves and not their theorizing.
What the assault on Newtonian mechanics has opened up in the collective psychology of social scientists is the possibility that the poor results are due not to the failings of the social scientists as empirical researchers but to the methods and theoretical assumptions they had taken over from Newtonian mechanics. In short, social scientists are now able to consider seriously for the first time the commonsense proposition they had so rigorously rejected: that the social world is intrinsically an uncertain arena. I call this a commonsense proposition in that most persons, if asked, would see this assertion as so self-evident that they could not imagine that anyone (even social scientists) could think otherwise. What I would like to explore is what happens to social science if we take this commonsense proposition, now argued as a scientific proposition by Prigogine and many others, as the basis of our work in social science. Let us start with the ancient image that the universe is like a flowing river, in eternal, endless change. "You can't step into the same river twice." How can this image be reconciled with the idea that there are covering laws that govern the universe in every detail? It can only be reconciled if one assumes Laplace's demon, and presumes that a being outside the entire universe, knowing these laws, could have predicted every twist and turn in the river. What happens if we substitute for Laplace's demon the assumption that all processes tend to move away from their equilibria and that, when they move sufficiently far, they bifurcate - that is, that the system has arrived as a point such that there are two or more solutions to the equation describing the process? What happens in my view is that, while we continue to be required to search for regularities of processes within systems, the systems themselves are constantly moving far from equilibrium, and therefore at some point will be transformed, such that the regularities we have observed no longer hold, even as an approximation of reality. That is to say, we have to live with a contradiction. On the one hand, all our "truths" we discern hold only within certain TimeSpace parameters, and that therefore very little of much interest can be stated that is "universal." On the other hand, although all is ever-changing, the world is quite obviously not without some patterns of explanations of these changes, and that the changes themselves fall into two different kinds - those that are an intrinsic part of the regularities of the system, and those that involve the transition to or transformation into a different systemic context.
Prigogine tells us the following about bifurcations in natural phenomena:
Bifurcations are a source of symmetry breaking....Bifurcations are the manifestation of an intrinsic differentiation between parts of the system itself and the system and its environment. Once a dissipative structure is formed, the homogeneity of time...or space...,or both, is broken.
In general, we have a succession of bifurcations....The temporal description of any such systems involves both deterministic processes (between bifurcations) and probabilistic processes (in the choice of the branches). There is also a historical dimension involved. If we observe that the system is in state d2, that means it has gone through the state of b1 and c1 (Prigogine, 1997, 69-70).
Allow me to translate that view into the language we use in the social sciences. I shall call the systems with which we deal in the human social world "historical (social) systems," by which I shall mean some social whole which has spatial boundaries (even if they change over time) and which evolves historically over time. In order for us to call this social whole a system, we would have to argue that it was relatively self-sufficient and that, during its evolution, it maintained certain essential features intact. In order for us to speak of a system evolving, we would have to identify a moment that it came into existence as a system. In order for us to speak of a systemic bifurcation, we would have to identify a moment (in the past, because we could never identify such a moment in the future for an existing system) when this system came into systemic crisis. In short, we would need to analyze three different time-periods: the period of genesis; the period of the normal operation and evolution of the system; the period of bifurcation, or systemic crisis.
A close look at these defining characteristics of an historical system immediately makes apparent that each criterion involves measurements that are extremely difficult for us to make, at least at our present state of knowledge in the social sciences. How can one operationalize "relatively self-sufficient," given the fact that no system (perhaps not even the entire universe) is exempt from existing in a larger context that has at least occasional impact on its functioning (and perhaps much more than that)? How do we decide which features of a system can be considered essential and how do we measure that they are intact? How do we operationalize boundaries of a social system, especially given that we assume they are constantly changing? And how do we operationalize the moment of genesis of a system, or the moment of bifurcation/systemic crisis?
The answer has to be that none of these scientific tasks is the least bit easy, and all of them would give rise, in the case of any specific analysis of an historical social system, to enormous debate among social scientists. Still, the fact that there would be controversy over the validity and reliability of the data presented does not invalidate the theoretical model nor does it mean we can escape the responsibility of seeking to obtain such data. The fact that the data that could be obtained, or supplied by various sources, might give rise to quite different interpretations by persons differently placed in the historical social structure, does not mean that it is hopeless to arrive at relative and interim accords on the greater persuasiveness of some interpretations than others. It merely means that we must be self-conscious of the inevitable and unavoidable social biases of all the interpreters, and enter into our mental operations correctives that amend the results to reduce the effect of the multiple biases. In short, what we need is a methodological roadmap, itself filled with uncertainties, in order to ascertain plausible interpretations of uncertain social realities.
I think that one crucial element in this roadmap is a specification of how we use terms like change, crisis, and bifurcation. We can reduce the utility of such terms to nil if their definition becomes too vague. The great danger is that, since change is eternal, everything becomes defined as a crisis, every twist and turn in the road a critical bifurcation. Of course, in the sense that there is a irreversible historical process and that all past events are part of present reality and explain why the empirical realities of the present could not be other than what they are since reality passed through a given pathway, following an infinite series of social choices, large and small, everything is a crisis and bifurcations do occur at every microsecond. But, at that point, we have said nothing more than "you can't step into the same river twice." Using such definitions vitiates all analysis, and reduces our scholarly task to that of rerunning the infinite history of the universe, a task that is both impossible and totally without point.
So, right away, we have to distinguish somehow between minor changes and major changes, between cyclical downturns and systemic crises, between choices and bifurcations. This is of course clearly implied in Prigogine's distinction between the deterministic processes that occur between bifurcations and the probabilistic processes in the choice of the branches. But, when this concept is applied to social science, there seems to be a lot of slippage, and a lot of forgetting of the basic distinction. Let me therefore return to my sense that there are three moments of time in the analysis of any historical social system - that of genesis, that of ongoing operation, and that of systemic crisis. None of these three moments lasts a mere microsecond, although clearly when we speak of the ongoing operation of a system, we are normally referring to a much longer time span than the time span of either a genesis or a systemic crisis.
Most social science is in fact written about what is happening in historical systems during their ongoing operations. Of course, in some cases, this is in full consciousness of the fact that the onset and sunset of an historical system are special, different moments. But in many other cases (more than one would hope), the fact that historical systems have a finite temporal duration is lost from view and scholars use data to compare situations in various instances that are located in quite different historical systems. Such comparisons can readily and easily lead to dubious, even quite erroneous and/or irrelevant, conclusions. It is here that the classic idiographic critiques of the generalizers can point to many instances of egregious scholarly misjudgment. However, if the analyses are kept within a single historical system, we can also with relative ease come up with sets of generalizations that seem plausible, and seem to be replicable. And it is here that the defenders of a nomothetic mode find their greatest justification, allowing them to feel that their underlying epistemology has been validated.
I would like to offer a methodological guideline which would enable us to observe simultaneously the ways in which so-called essential features of the system remain intact and the ways in which the system is evolving in a direction far from equilibrium such that it will have to bifurcate at some point. I call this methodological guideline the search for cyclical rhythms and secular trends. The concept of cyclical rhythms assumes some kind of equilibrium, albeit usually a moving equilibrium. It also assumes that there is always "noise" in any process, such that there are always fluctuations, and that these fluctuations, when graphed, take the form of multiple cycles of varying lengths. Since there is always so-called noise, such cycles are inherent in all systems, whether physical or social, and they can be measured. Of course, there is no presumption that the pattern involves defined and unchanging time intervals. Quite the contrary: it is assumed that, in all complex systems, the most that can be described, because the most that really exists, are approximate standards, which show a high probability of recurrence. However, what needs to be demonstrated is that there is something in the process that makes these fluctuations inevitable and recurrent, and that this something can be adequately delineated.
Of course, the number of rhythms that exist in a given historical system may be quite large, and one may not only wish to show that some are more important and/or longer than others, but also to explain what are the particular consequences of the more important rhythms. However, it may also be that, to explain other particularities, one would be better off noticing some of the so-called less important rhythms. In any case, the description of the rhythms is the description of the operational features of the system. They are what allow us to call a system a system. Historical social systems are not special in this regard, only more complex, and it is therefore more difficult to measure, even approximately, these rhythms.
Since, in fact, the rhythms are always imperfectly symmetrical in reality because of the ever-changing detailed happenings of the system, the equilibrium is always moving, and it can usually be seen to be moving in a certain direction. This gives us what is often described in deterministic literature as the linearity of the system, and what we are calling secular trends. What is often left out of the analysis is that most trends, at least most trends in historical social systems, cannot be extended infinitely because they reach certain types of in-built limits. Let me illustrate with an obvious one. I suppose the population on the earth could expand indefinitely through biological reproduction. But at some point, one would literally run out of space for the population to stand up. And at some no doubt earlier point, one might run out of food supply. Something would thereupon happen which would reduce the size of the world population. So it is not in fact true that this linear vector could expand indefinitely. It would be very easy to make a long list of such indefinitely linear vectors that are in practice impossible to achieve.
What this immediately shows is that a vector cannot be analyzed as though it were in an autonomous trajectory since its actual itinerary is the direct result of its interaction with other factors. Its development is dependent on certain specific conditions such that all systemic expansion has quantitative boundaries. I think, in fact, it is not very useful to measure vectors as absolutes but always as relations with other vectors. In short, we should rethink what it is we should be measuring as secular trends. I believe that what we should be measuring is the percentages to be found in processes one has determined to be crucial to the operation of a particular historical social system. For example, in the modern world-system, I would be interested not in the numbers of full-time wage workers but in the percentage of the system's population that is engaged more than half-time in wage labor. I will not argue here why this is important. I merely wish to point out that, once one has converted data into percentages, there always exists the asymptote of 100 percent. Nothing can be true of more than 100% of the population being measured. It follows that all secular trends tend to reach a point where they cannot continue in a linear fashion. This, it seems to me, is precisely where historical systems reach crisis points, and which therefore lead to bifurcations.
What then is the methodological relationship between cyclical rhythms and secular trends? It seems to me rather obvious. Why after all are cyclical rhythms in historical social systems cyclical? Evidently, because when the operation deviates too far above or too far below the equilibrium, it becomes in the interests of some social players to act in ways that push the system back towards the equilibrium. In everyday language, adjustments are being made. Of course, the nature of these adjustments is a function of the power structure of any system, and the priorities built into the operating mechanisms of the system. Of course, they do not occur smoothly because of the large number of actors and the large number of conflicting interests. But generally speaking, we can predict what is likely to happen, and therefore what has happened. This is basically what Braudel meant when he argued that "events are dust," and that, rather than recounting the sequences of events, we must seek to discover what has been happening in the longue durée.
I think it would therefore be useful, at least in discussing historical social systems, to distinguish between minor and major uncertainties. Minor uncertainties are ubiquitous. No one ever knows what is really going to happen at the next instant. The virtually infinite number of potential actors plus the ever-changing physical environment renders all exact forecasting an intrinsic impossibility. But much of such uncertainty can be minimized. We can estimate probabilities with some degree of error, and I suppose one can argue that the "normal science" (in Kuhn's phrase) of the Newtonian paradigm as applied to the social sciences is precisely the attempt in fact to make these estimates nearer and nearer to what eventually occurs. From the point of view of the larger social order, there is no doubt that a decrease in the degree of error of an estimate is a positive thing. Public policy is constantly being made on the basis of such estimates. Is it more important to invest social resources in expanding health facilities or improving earthquake detection devices? The answer depends in part on our estimates of degrees of potential danger, and to whom.
I no more wish to deny the utility of such social scientific work than Prigogine is denying the utility for many purposes of classical Newtonian equations. But..., and there is a but! We must bear in mind three things about this "normal" Newtonian science. First of all, its legitimation is to be found in policy outcomes. In the natural sciences, we would say its legitimacy is to be found in technology or in engineering. Can you build better bridges as a result of this scientific work? Can you make more intelligent policy decisions as a result? Up to now, the physicists and chemists have had a better score using this measuring rod than say the economists and sociologists. This is precisely what has impelled social scientists over the past century to try to "catch up" with the natural scientists. And, given both the external social pressure on them and the demands of their own superegos, it is quite understandable that social scientists have been attracted to the so-called nomothetic path. But given their really low score on social engineering, one might think they would look again at whether this route pays off.
The second thing to say about following this path is that we have been thereby blinded by the epistemological blinkers it has imposed upon us. It is the growing awareness of the negative effects of these epistemological blinkers that has fueled the knowledge movement within the physical sciences that bears the embracing label of "complexity studies." Once again, social scientists have been laggards, and are only now beginning to look again seriously at the epistemological assumptions that underlie their chosen methods. They are beginning to return to "philosophy," a domain they had loudly ejected from their purview as "unscientific." This trend is not at all a negative thing, and I shall return to this question.
The third problem about doing "normal" Newtonian science, even in a probabilistic mode, is that it obliterates all knowledge about, and therefore concern with, the larger uncertainties in social reality. The larger uncertainties do not occur every day, or even every year or every decade. They may, in the case of historical social systems, occur only once every 500 years. But it is these fundamental bifurcations that comprise the pattern of historical evolution of the human species, and tell us what we really want to know: where we have been, where we are, and where we are likely to be going; or rather, which of our possible futures we might reasonably seek to obtain because we prefer it.
Why do we avoid studying the fundamental bifurcations? In part, we are afraid to do so because their outcome is truly uncertain. In part, our attention is deliberately diverted from doing so in order that we not apply our collective efforts to affecting the outcome of the bifurcation in some ways, thereby permitting a minority (normally the privileged minority) to make their inputs into the process unimpeded. If, however we decide consciously to look at systemic bifurcations, we require a quite clear picture of the considerable difference between everyday choices and systemic bifurcations. To put this into the language people use in daily life as well as in social science, we have to be aware that, historically, most so-called "revolutions" - whether political, economic, or other - have in fact been minor adjustments, and that the real upheaval of moving from one historical system to another may look very chaotic indeed and be very hard to classify.
At the present time, the modern world-system finds itself in one of these fundamental bifurcations. It is in systemic crisis, and consequently there is also a crisis in our structures of knowledge. Hence we are confronted not with one but with two major social uncertainties. One is what will be the nature of the new historical system that we are constructing. The second is what will be the epistemology of the new structures of knowledge that we are constructing. Both involve struggles, whose outcome is unpredictable, but each marks the end of the world as we know it. I am using know in its double sense: know, as to be acquainted with (cognoscere, conocer, connaître, kennen) and know, as to understand (scre, saber, savoir, wissen). The modern world-system, the capitalist world-economy, is in crisis. We no longer know it. It presents to us unfamiliar landscapes and uncertain horizons. The modern structures of knowledge, the division of knowledge into two competing epistemological spheres of the sciences and the humanities, is in crisis. We can no longer use them as adequate ways in which to gain knowledge of the world. We are confused by our inability to know in both these senses, and many fall back on dogmatisms. We are living in the eye of the hurricane.
I shall not discuss here the crisis in the capitalist world-economy. I have done it many times elsewhere.(1) Suffice it to say that I believe there exists today, as a result of long secular trends that have been moving away from the equilibrium, a massive profit squeeze which will block the continuation of an endless accumulation of capital, the motor force of capitalist development. This squeeze derives from at least three separate vectors: the secular rise of real wages across the world-economy as a whole; the growing destruction of the environment resulting from the institutionalized externalization of costs; and the fiscal crisis of states, which has been caused by the democratization of the world-system that has led to raising significantly the minimum levels of demands on the states for education, health care, and lifetime minimum income guarantees. In addition, there has been a collapse in the legitimacy of the state structures because of growing disillusionment with the possibility of reducing the polarization of the world-system, and this legitimation had long been a key mechanism in maintaining the equilibrium of the world-system. I cannot argue this case now and must simply assert that there appears to be no solution within the framework of the existing system that would bring the system back to an even temporarily stable equilibrium. Therefore systemic parameters are oscillating widely and a branching is occurring. We can assume that the elaboration of this bifurcation will occur over about half a century before a definitive choice is made, and a new system (or systems) come(s) into existence. We may further expect that this period will be a period of great social turmoil, in part because of the fluctuations of the system, in part because of the decline of legitimacy of the state structures, and in part because it will be a period of great conflict about the nature of the successor system.
I wish however to concentrate here on the implications of this systemic bifurcation for the structures of knowledge. Structures of knowledge are of course an integral part of the cultural underpinnings on any historical social system. While there always exists a sort of internal logic of such structures, and therefore a somewhat autonomous intellectual trajectory, they are part of a larger structure, having to fit into the logic of this larger structure, and constrained by the intellectual boundaries the larger system has put in place. Structures of knowledge are precisely that, structures, and as such they are brought into existence socially and can only survive socially if there continues to be long-term compatibility with the social environment.
Over the long period during which the modern world-system struggled to put in place a geoculture suitable for its optimal functioning, the epistemological unity of knowledge presumed in prior systems came under ever-increasing attack until, approximately in the last half of the eighteenth century, the so-called divorce between science and philosophy (or the humanities) was consummated. We can easily explain the social underpinnings and the historical process of this major reorganization of our conceptions of knowing. What is more relevant at the moment is the nature of the presumed division into "two cultures." It was argued by each of the two camps that constituted themselves as combatants in the arena that they had ways of knowing the world radically different from and significantly better than those of the other camp.
The scientists asserted that we could only know by empirical investigation (ideally, by experimentation) and that, from such empirical investigations we could develop theorems that were testable in rigorous ways. As long as these theorems continued to pass successive tests (which in principle were unending), they could be said to state, at least provisionally, universal truths. If one were able to present an adequately replicated and validated hypothesis, one could make the claim that this truth was certain. It is not always clear what we mean by something being certain, but minimally it surely meant that we could count on getting the same mathematical results each time we used the equation, the variation being only in the data that were inserted, the so-called "initial conditions." Insofar as the state of knowledge about any given object of enquiry was insufficient to assert such universal truths, this was said to be the fault of the scientists themselves, who had not yet been able to arrive at this point of knowledge. But the epistemological expectation was that eventually the community of scientists would bring forward members who could demonstrate the universal truths pertinent to the object of enquiry. Certainty of analysis was a certain prospect.
The epistemological divide of science and philosophy has been directly challenged by two knowledge movements that have gained strength in the last 25 years. One is the sciences of complexity. It is a movement of many strands, and Prigogine is one major figure in its intellectual development. I have already indicated the main lines of its differences from Newtonian mechanics in its basic assumptions. I would just like at this point to underline its relationship to social science. I have already indicated the social psychological impact of complexity studies on social science. It has undermined the cavalier argument of nomothetically-oriented social scientists that they represented the incarnation of the scientific method. By doing this, the sciences of complexity have opened up space within social science for a different approach to science, one centering around the end of certainties. And this is very healthy and will be fruitful in itself.
There is however a second thing to say about the relation of the sciences of complexity and social science. One of the central slogans of the sciences of complexity is the "arrow of time," a phrase invented by Arthur Eddington and taken up and disseminated widely by Prigogine. For Prigogine, this is the response to a central theme of Newtonian mechanics, reversibility. In social science, no one, not even the most hardy nomothetic advocates, has dared to argue reversibility. What they have done instead is ignore history, and indeed deplored "historicism."
By raising high the banner of the "arrow of time," by asserting in effect that even the tiniest units of physical matter have a historical trajectory, one which cannot be ignored, Prigogine has both reinforced those social scientists who have always insisted that there can be no social analysis that is not historical but he has moved physical science onto the central epistemological terrain of social science. He has renewed the call for a unified science, but not in the spirit of the analytic philosophers who wished everyone to adopt the premises of Newtonian mechanics and become social physicists. Rather, he has in effect suggested that the natural scientists become part of a larger family in which the sociocultural premises and links of all knowledge activity be its unifying theme, one in which we overcome the two cultures because science and philosophy are conjoined activities deriving from a common epistemological base.
Meanwhile, at the other end of the spectrum, among the humanities there has emerged a vibrant and extremely diverse knowledge movement, now usually called cultural studies. Once again, I cannot develop here either the historical origins of this movement or the reasons for its sudden explosion or the limitations of its analyses. I wish in this case too simply to indicate its relationship with social science. The importance of cultural studies is not that it has launched a critique of what they often refer to as Enlightenment views, but by which they centrally mean the cultural dominance of the premises of Newtonian mechanics. The sciences of complexity has made this critique better and more effectively. The real social contribution of cultural studies has been its critique of the humanities, as it had been institutionalized as a counterdogma to science.
The humanities were not historically concerned with science; that after all was the point of the so-called divorce. They therefore looked askance at social science, which they saw as too committed to science, and encouraged its practitioners, particularly the historians, to define themselves as humanists and to locate themselves in faculties of philosophy. What they were concerned with were the criteria of evaluation of quality in the arts, broadly defined, as well as engaging in an empathetic, hermeneutic perception of social reality. This led them on the path to the establishment of canons, lists of aesthetic achievements that could be held aloft and taught to successive generations. In a curious way, they had reached the same point as the most committed Newtonian scientists. They were interested in perfection: elegance in art rather than elegance in theoretical formulation. But the point for both was that the worth of this excellence was not measured either by social utility or by criteria outside the internal rules of the knowledge activity.
Cultural studies arose in rebellion to this kind of aesthetic ivory tower. The practitioners of this new knowledge movement insisted that all cultural activity occurred within a social context, that it was produced and appreciated differently according to the social location of the producer or the appreciator. And of course, the social location itself was a constantly evolving historical reality, such that how one appreciates a text today may be different from how the same person appreciates it tomorrow. Here, too, I wish to emphasize the relationship of these propositions to social science.
Social science has always been built around the assumption that we perceive social reality through socially-constructed lenses. Even the most nomothetically-oriented social scientists admit this, at least implicitly. They simply seek to overcome what they see as a limitation, whereas other social scientists take this so-called limitation as a permanent reality, one which actually promotes a richer understanding of the world. In any case, cultural studies, by laying emphasis on this central theme, has placed itself fully on the terrain of social science and has thereby aided in overcoming a false dichotomy between humanism and science.
So here we are today, on the verge of a major epistemological restructuring, a reunification of the methods of enquiry across the fields of knowledge, and one in which the terrain of social science will now be central, if not all-encompassing. Social science is after all the study of the most complex systems that exist, and therefore the most difficult to translate into systematic analysis. It is also the inevitable, if often unacknowledged, underpinning of what we have historically called humanistic studies. It is in fact everybody's necessary activity, from physicists to literary scholars. Far from being a call for social science imperialism, this is a call for universal entry into social science.
We desperately need a collective intellectual discussion, and whether we call this discussion science, philosophy, or social science is a matter of great indifference to me. We live with the knowledge that uncertainty, at least long-term uncertainty, seems to be the only unbudgeable reality. This means that self-reflective knowledge activities not only have to build this central reality into the practices they develop to expand our understanding of the world, but be ready to move from level to level of analysis in the search for more plausible explanations allowing for more informed choices. In the end, knowledge has to be about choices, and therefore about innovation, imagination, and possibilities. Choices involve responsibility, and scholars/scientists are precisely persons who, by opting for the activity, commit themselves to assuming the responsibilities of their assertions, their claims, their guesses, their suggestions of priorities. Isabelle Stengers, in her distinction between "requirements on" and "obligations of" the scientist, spells out the significance of assuming responsibilities:
The theme of "rationality" changes in effect its signification according to whether it is placed under the heading of a requirement, in which case it is usually a vector of arrogance and infamy, or is placed under the heading of an obligation, in which case it is a synonym of risk and submission to a test, a test not for the public or for the incompetent but for the one who chooses to place his work within a practice that claims to be rational (Cosmopolitique I: La guerre des sciences, Paris: La Découverte, 1996, 90).
If reality is uncertain, there is no way to avoid choices. If we cannot avoid choices, there is no way to separate the value commitments, preferences, and presuppositions of the analyst from entering the process of analysis. Even if we eliminate all such considerations at the conscious level, that is, we insist on affecting a stance of moral neutrality before the object of our knowledge activity, these factors come back in at the unconscious level and at the level of what is permissible social discourse. And even if we bring these latter to the surface, we discover that there is an endless regression of contextuality, personal and collective biographies, which can never be eliminated since they constitute the psyche of the analyst. In short, there is no search for truth which does not involve arguments about the good and the beautiful.
Is then science, scientia, an impossible dream? Far from it. I would argue the exact opposite. It is only when we accept the impossibility of the separation of knowing from wanting that we can begin to know better. What it requires is two things. First, we must be willing to expose premises, of ourselves and of others, in an analytic and not an accusatory way. We can then debate the question, would we have different results of our research if the premises were altered? There would be no taboo questions about premises.
And secondly, we need to have scientific communities that are composed of persons coming from every collective trajectory, in order to discover what would be proposed when persons with truly different biographies examine the same data, explore the same problem? In social science, this means the extensive and veritable internationalization of the social science community. We are still a long way away from there being one.
And finally, to return to my earlier point, we must learn to distinguish between minor and major bifurcations, between adjustments and systemic transformations, between explaining what is ongoing and what is exploding. This informs the issue of choices. For the kinds of choices one makes in adjusting an ongoing social system and in branching into two or more possible future social systems are not at all the same, and they cannot be made intelligently, either set of them, unless we are clear about the problems we are facing and examining.
I believe we live in a very exciting era in the world of knowledge, precisely because we are living in a systemic crisis which is forcing us to reopen the basic epistemological questions and look to structural reorganizations of the world of knowledge. It is uncertain whether we shall rise adequately to the intellectual challenge, but it is there for us to address. We engage our responsibility as scientists/scholars in the way in which we address the multiple issues before us at this turning-point in our structures of knowledge.
1. See in particular Utopistics: or, Historical Choices for the Twenty-First Century (New York: New Press, 1998); and After Liberalism (New York: New Press, 1995). [In Spanish: Utopística, forthcoming, Siglo XXI; and Después del Liberalismo, México: Siglo XXI, 1996]
(Go to top of list of papers)
(Go to Fernand Braudel Center Home Page)