4) These factors are discussed in T. S. Kuhn, The Copernican Revolution: Planetary Astronomy in the Development of Western Thought (Cambridge, Mass., 1957), pp. 123–32, 270–71. Other effects of external intellectual and economic conditions upon the substantive development of science are described in the following papers of mine: “Conservation of Energy as an Example of Simultaneous Discovery,” in Critical Problems in the History of Science, ed. Marshall Clagett (Madison, Wis., 1959), pp. 321–56; “Engineering Precedent for the Work of Sadi Carnot,” Archives internationales d’histoire des sciences, XIII (1960), 247–51; “Sadi Carnot and the Cagnard Engine,” Isis, LII (1961), 567–74. Therefore, the slight treatment given to the role of external factors is limited to the problems discussed in this essay.
I. Introduction: The Role of History
Introduction: A Role for History
If history is regarded as a repository filled with more than anecdote or chronology, it may bring about a decisive transformation in the image of science now available to us. That image has been formed chiefly from the study of completed scientific achievements, as recorded in the classics in earlier times, even by scientists themselves, and, more recently, in the textbooks from which each new generation of scientists learns its training. But the purpose of such works is necessarily persuasive as well as pedagogical. The concept of science drawn from them fits actual activity no better than an image of a nation’s culture drawn from a tourist guidebook or a language textbook. This essay is fundamentally intended to show that we have been misled by such books. Its aim is to sketch an entirely new conception of science that can emerge from the historical record of research activity itself.
But if one seeks out and examines historical data in order to answer questions posed by the nonhistorical clichés derived from science textbooks, no new conception will emerge from looking at history. As for such textbooks, they often give the impression that the content of science is uniquely exemplified by the observations, laws, and theories explained within them. Almost without exception, these books are written as though scientific methods were simply exemplified by the manual techniques used in gathering textbook data, and as though the logical operations applied in relating those data to the theoretical generalizations of the textbook were what is meant by scientific method. The result has been a concept of science with profound implications for the nature and development of science.
If science is the collection of facts, theories, and methods contained in present-day textbooks, then the scientist becomes a person who, successfully or not, strives with all his might to add one or two elements to that particular collection. Scientific development becomes a piecemeal process in which these items, singly or in groups, are added to the ever-growing heap of materials that make up scientific technique and knowledge. And the history of science becomes the discipline that records the chronology of these transmitted increments and of the obstacles that have hindered their accumulation. In that case, the historian of science has two principal tasks concerning scientific development. On the one hand, he must determine, one by one, when and by whom the scientific facts, laws, and theories of the age were discovered or invented. On the other hand, he must identify and explain the heaps of error, myth, and superstition that have delayed the more rapid accumulation of the contents of modern science textbooks. Much research has been directed toward these goals, and some of it still proceeds in this way.
Recently, however, a number of historians of science have come to feel that the concept of development-by-accumulation makes it increasingly difficult for them to fulfill the functions assigned to them. In other words, as chroniclers of an incremental process, they find that the deeper they dig, the more difficult it becomes to answer such questions as: When was oxygen discovered? Who first discovered the conservation of energy? Some historians of science increasingly suspect that these may be the wrong sorts of questions even to ask. Perhaps science does not develop through the accumulation of individual discoveries and inventions. At the same time, these very scholars are finding it increasingly troublesome to distinguish the scientific elements in past observations and beliefs from those things that their predecessors did not hesitate to label “error” and “superstition.” For example, the more closely historians of science study Aristotelian dynamics, phlogistic chemistry, or caloric thermodynamics, the more they feel that the views of nature that prevailed at those times, taken as a whole, were neither less scientific than those accepted today nor the product of human idiosyncrasy. If these outdated beliefs are to be called myths, then myths can be formed by the same kinds of methods and produced by the same kinds of reason that now lead to scientific knowledge. If, on the other hand, they are to be called science, then science includes bodies of belief that are quite incompatible with those we now possess. Given this choice, the historian must choose the latter. Outdated theories are not in principle unscientific simply because they have been discarded. But this choice makes it difficult to view scientific development as a process of cumulative accumulation. The very historical research that reveals the difficulty of isolating individual inventions and discoveries becomes the source of serious doubt about the cumulative process through which these individual contributions to science have been thought to have been compounded.
The result of all these questions and difficulties is a historiographic revolution in the study of science, one that still remains in its early stages. Generally, and often without being aware that they are doing so, historians of science have begun to pose new kinds of questions and to trace different, often noncumulative lines of development in science. Rather than asking what lasting contribution older science made to us in the present, historians try to reveal the historical integrity of that science in its own time. For example, they ask not about the relation between the viewpoint of modern science and that of Galileo, but about the relation between his views and those of his group—that is, his teachers, his contemporaries, and his immediate successors working in the sciences. Moreover, historians insist that the views of that group and of other similar groups must be studied on the basis of a viewpoint—generally quite different from that of modern science—that gives those views the greatest coherence and the closest fit to nature. When one examines the studies obtained in this way, perhaps best exemplified in the writings of Alexandre Koyré, science no longer appears to be quite the activity discussed by compilers who followed the older traditions of historical writing. Implicitly, these historical inquiries suggest the possibility of a new image of science. This essay has been written in order to outline that image by making explicit the implications suggested by the new historiography of science.
In the course of this attempt, what aspects of science will stand out most prominently? At least in the order of exposition, the first is the insufficiency of methodological directives to determine, by themselves, conclusions unique to many kinds of scientific questions. If a person is instructed to investigate electrical or chemical phenomena, and if he does not know these fields but does know what it means to be scientific, he may legitimately arrive at any one of several incompatible conclusions. Which particular conclusion he reaches among these valid possibilities will perhaps be determined by prior experience in another field, by the accidents of his investigation, or by his own personal characteristics. For example, what beliefs about the stars will he bring to bear upon the study of chemistry or electricity? Among the various plausible experiments relevant to the new field, which will he decide to undertake first? And which aspects of the complex phenomena obtained there will seem to him especially relevant to revealing the nature of chemical change or electrical affinity? At least for the individual, and sometimes for the scientific community, the answers to these questions often constitute crucial determinants of scientific development. As we shall see in Section II, for example, the early developmental stages of most sciences are characterized by continual competition among differing views of nature, each of which is derived in part from the dictates of scientific observation and method and most of which are roughly compatible with those dictates. What distinguishes these various schools is not some failure of method—they were all “scientific”—but something that must be called incommensurable ways of seeing the world and of practicing science within it. Observation and experience can, and must, severely restrict the range of admissible scientific belief; otherwise there would be no science. But observation and experience alone cannot determine a particular body of such belief. An apparently arbitrary element, composed of personal and historical accidents, always enters as a component of the beliefs embraced by any given scientific community at any given time.
But this element of arbitrariness does not mean that a group of scientists can carry on its scientific activity without any beliefs it can accept as its own. Nor does this element lessen the necessity of the particular existing body of knowledge upon which that group in a given age actually depends. Effective research rarely begins before a scientific community thinks it has secured firm answers to such questions as: What are the fundamental entities that compose the universe?
How do these entities interact with one another and with human perception? What questions may legitimately be raised about such entities, and what techniques should be applied in seeking their solutions?
At least in the mature sciences, the answers to these questions—or desirable substitutes for answers—are firmly embedded in the educational transmission that prepares and qualifies students for professional activity. Because that education is so rigorous and so assured, these answers come to exert a profound force upon the scientific mind. The fact that they can do so explains both the distinctive efficiency of normal scientific research and the direction in which it proceeds at any given time. When we examine normal science in Sections III, IV, and V, we shall ultimately be tempted to describe such research as a strenuous and devoted attempt to force nature into the conceptual boxes supplied by professional education. At the same time, whatever element of arbitrariness may lie in their historical origins and subsequent development, we shall come to doubt whether research could proceed at all without such boxes.
In any case, the element of arbitrariness exists, and it too has a significant effect on scientific development; this will be treated in detail in Sections VI, VII, and VIII. Normal science, the activity to which most scientists necessarily devote almost all their time, rests on the assumption that the scientific community knows what the world is like. Much of the success of scientific activity derives from that community’s willingness to defend this assumption, if necessary at considerable cost. Normal science, for example, often suppresses fundamental novelty, because such novelty is bound to subvert its basic commitments. Nevertheless, so long as those commitments sustain an element of arbitrariness, the very nature of normal science guarantees that the new will not be suppressed for very long. Sometimes a normal problem—that is, a problem to be solved by existing rules and procedures—resists solution despite repeated attacks by the most able scholars, those best equipped to solve it. In other cases, an apparatus designed and constructed for the purposes of normal research reveals an anomaly that does not fit the predicted results. In these and other ways, normal science repeatedly goes astray. And when it does so—when, in other words, a specialty can no longer evade anomalies that destroy the existing tradition of scientific activity—there begins the extraordinary inquiry that finally leads the specialty to new commitments, to a new basis for the practice of science. The extraordinary episodes in which such shifts in professional commitment occur are the events called scientific revolutions in this essay. Scientific revolutions are the tradition-shattering complement to the tradition-bound activity of normal science.
The clearest examples of scientific revolutions are those famous episodes in the history of scientific development that have often previously been described as revolutions. Thus, in Sections IX and X, where the nature of scientific revolutions is first examined directly, the major turning points in scientific development associated with the names of Copernicus, Newton, Lavoisier, and Einstein will be treated repeatedly. At least in the history of the physical sciences, more clearly than any other episodes, these cases reveal what scientific revolutions are. Each of them compelled the scientific community to reject a once-honored scientific theory in favor of another incompatible with it. Each revolution brought about a corresponding change in the problems that became the objects of scientific inquiry, and also in the standards by which the specialty determined what it would regard as an admissible problem or as a legitimate problem-solution. And each episode transformed the scientific imagination in a way that must ultimately be described as a transformation of the world within which scientific research was conducted. These changes, together with the controversies that almost invariably accompany them, become the defining characteristics of scientific revolutions.
These characteristics are especially evident in studies such as those of the Newtonian revolution or the chemical revolution. The basic thesis of this essay, however, is that such characteristics can also be grasped in the study of many other episodes whose nature appears far less certain. Maxwell’s equations were, for the much smaller group of specialists affected by them, no less revolutionary than Einstein’s equations, and for that reason they met with resistance. The creation of other new theories, too, regularly and quite naturally provokes the same kind of response from the specialists in the particular fields their domain affects. For such people, a new theory means a change in the rules that had governed the existing practice of normal science. It therefore inevitably affects much of the scientific work that had already been successfully completed. This is why a new theory, however specialized its range of application, is rarely, if ever, a merely cumulative supplement to what is already known. The assimilation of a new theory requires the reconstruction of existing theory and the reevaluation of existing fact, an intrinsically revolutionary process that is seldom completed by one person or overnight. It is therefore no surprise that historians of science find it difficult to date precisely this broad process, which their terminology compels them to treat as a separate, discrete event.
Nor is the creation of a new theory the only scientific event that casts a revolutionary shock upon the specialists in the domain where it occurs. The commitments regulating normal science not only specify what kinds of entities the universe contains; they also implicitly suggest what it does not contain.
Though this is a theme that must be developed further, it follows that discoveries such as oxygen or X-rays are not simply the addition of one more item to the scientist’s world. Ultimately a discovery does have such a result, but only after the professional community has reorganized the theoretical network through which it deals with the world, in the process of reevaluating traditional experimental procedures and remaking its long-familiar conceptions of entities. Scientific fact and theory are not categorically separable, except perhaps within a single tradition of normal scientific activity. That is why an unexpected discovery does not, in its meaning, end as a mere fact; and that is why the scientist’s world is not only quantitatively enriched but qualitatively transformed by the discovery of fundamental novelty in the realm of fact or theory.
It is from here that this extended conception of the nature of scientific revolutions is explained. Of course, such an extension distorts ordinary usage in some respects. Nevertheless, I shall continue to speak of discoveries as revolutions, because what makes this broadened conception seem so important to me is precisely the possibility of relating the structure of discovery to, for example, that of the Copernican revolution. The preceding discussion suggests how the complementary notions of normal science and scientific revolution will be developed in the nine sections that follow. The remainder of this essay attempts to settle three remaining central questions. Section XI, by discussing the textbook tradition, deals with why scientific revolutions have previously been so difficult to see. Section XII describes the revolutionary competition that takes place between the defenders of an old normal-scientific tradition and the supporters of a new one. Thus Section XII considers the route by which, in the theory of scientific inquiry, the processes of confirmation or falsification familiar from the conventional image of science are, in one way or another, replaced. Competition among factions within the scientific community is, in practice, the only historical process that results in the rejection of a previously accepted theory or the adoption of another. Finally, Section XII will ask how development through revolutions can be compatible with the distinctive character of scientific progress.
On this question, however, this essay will offer no more than the broad outline of an answer, for the answer depends on characteristics of the scientific community that require far more investigation and study.
Clearly, some readers will already be wondering whether historical study can in fact bring about the kind of conceptual transformation aimed at here. The entire stock of available dichotomies suggests that it cannot properly do so. History, as we commonly say, is a purely descriptive discipline. The theses presented above, however, are sometimes interpretive and sometimes normative. Again, many of my generalizations concern the sociology or social psychology of scientists. Yet at least some of my conclusions traditionally belong to logic or epistemology. In the preceding paragraph, I may even have seemed to violate the powerful modern distinction between the “context of discovery” and the “context of justification.” What, other than serious confusion, can be suggested by such a mixing of diverse fields and interests?
It would be difficult, by intellectually working through these distinctions and others like them, to illuminate their meaning and force more effectively. For many years I have regarded them as being concerned with the nature of knowledge, and I still think that, if properly revised, they can tell us something important. Nevertheless, my attempt to apply them, even in a rough way, to the actual situations in which knowledge is acquired, accepted, and assimilated has made those distinctions appear extremely problematic. Rather than possessing basic logical or methodological features and thereby taking precedence over the analysis of scientific knowledge, they seem instead to be essential elements of the traditional conception that provides substantive answers to the very questions developed so far. Such circularity by no means invalidates them. But it makes them part of a theory, and in doing so subjects them to the same kind of investigation regularly applied to theories in other fields. If they are to have, as their content, anything more than purely abstract concepts, then that content must be discovered by observing them in application to the data to be explained. How could the history of science(
history of science) fail to become the source of the phenomena to which epistemologies may legitimately be required to apply?
II. The Route to Normal Science
The Route to Normal Science
In this essay, “normal science” means research firmly based upon one or more past scientific achievements—achievements that some particular scientific community acknowledges, for a time, as supplying the foundation for its further practice. Today, of course, such achievements are set forth in detail, though not in their original form, by elementary and advanced science textbooks.
These textbooks explain in detail the gist of the accepted theory, illustrate it by citing many or all of its successful applications, and compare those applications with exemplary observations and experiments. Before such books became widespread in the early nineteenth century (and, in the newly matured sciences, even more recently), many of the celebrated classics of science served a function much like that of textbooks. Aristotle’s “Physica,” Ptolemy’s “Almagest,” Newton’s “Principia” and “Optics,” Franklin’s “Electricity,” Lavoisier’s “Chemistry,” Lyell’s “Geology,” and many other works, for a certain period of time, implicitly defined the legitimate problems and methods of a field of research for the next generation of practitioners. They could do so because these writings shared two essential characteristics. Their achievements were sufficiently unprecedented to draw an enduring group of adherents away from competing modes of scientific activity. At the same time, they were sufficiently flexible to leave all sorts of problems to be solved by the reconstituted group of practitioners.
Achievements that possess these two characteristics I shall henceforth call “paradigms,” a term closely related to “normal science.” In choosing this term, I wish to suggest that certain accepted examples of actual scientific practice—examples that include laws, theories, instrumentation, and the like—provide models from which particular coherent traditions of scientific research arise. These are the traditions historians of science describe under such headings as “Ptolemaic astronomy” (or “Copernican astronomy”), “Aristotelian dynamics” (or “Newtonian dynamics”), and “corpuscular optics” [or “wave optics”]. The study of paradigms, including many far more specialized designations than those named above for descriptive purposes, is what prepares the student of science to become a member of the particular scientific community in which he will later carry out scientific work. Since the student thereby joins people who learned the foundations of their field from those same established models, the work that follows will rarely give rise to overt disagreement over scientific activity. Those whose research is based on shared paradigms observe the same rules and standards for scientific practice. Such commitment, and the evident consensus it fosters, are indispensable elements in the emergence and continuation of normal science, that is, of a particular research tradition.
Since in this essay the concept of the paradigm will often be used in place of various familiar notions, the reasons for introducing it require further explanation. Why should a firm scientific achievement, as the locus of professional commitment, take precedence over the various concepts, laws, theories, and points of view abstracted from it? In what sense does a shared paradigm become the basic unit for the student of scientific development—that is, a unit that cannot be fully reduced to the logical elementary components that might function in its place? These questions will be addressed in Section V, and the answers to them and to similar questions will prove fundamental to understanding the related concepts of normal science and paradigm. But the more abstract discussion will depend on whether one has already encountered examples of normal science, or examples of paradigms at work. In particular, all these related concepts will become clearer by noting the fact that there can be types of scientific research without paradigms, or at least without paradigms possessing the clarity and binding force of those named above. The acquisition of a paradigm, and of the more profound form of research it permits, is a sign of maturity in the development of any given scientific field.