The essential characteristics of their scientific community
What are they? Clearly, such characteristics require an enormous amount of research. In this area, only the most tentative generalizations are possible. Nevertheless, a number of essential requirements for becoming a member of a specialized scientific group have already emerged with surprising clarity. For example, a scientist must devote attention to solving problems concerning the behavior of the natural world. In addition, although their interest in nature is general in scope, the problems they deal with are detailed ones. More important is the fact that the answers satisfying him must not be merely private, but must be accepted as solutions by many people. Yet the group that shares them is not drawn at random from society at large; rather, it becomes a well-defined professional community of scientists.
If it has not yet been written, one of the most powerful rules in a scientific career is the prohibition against appealing to heads of state or to the general public on scientific subjects. The recognition of the existence of an exceptionally competent group of experts, and the acceptance of its role as the full arbiter of professional achievement, have even more profound implications. The members of the group, through training and experience shared individually and collectively, will appear to be the only possessors of the rules of the game, or of the corresponding basis for clear judgment. To doubt that they share some fundamental basis necessary for evaluation would be to acknowledge the existence of incompatible standards of scientific achievement. Such an acknowledgment would inevitably raise the question of whether there can be a single truth in science.
Some of these characteristics common to scientific communities have been drawn entirely from the practice of normal science, and indeed had to be. It is the activity toward which the ordinary scientist’s training is directed. Yet despite their small scale, such characteristics are sufficient to distinguish that community from all other professional groups. Moreover, despite their source in normal science, such characteristics explain many of the peculiar qualities of the group’s responses during revolutions, especially during periods of paradigm debate. We have already seen that a group of this type must regard a change of paradigm as progress. Now we shall see that this perception is, in important respects, self-fulfilling. The scientific community may be called a highly efficient device for maximizing both the number and the precision of the problems solved through changes of paradigm.
Because the unit of scientific achievement consists of solved problems, and because the scientific group knows well which problems have already been solved, few scientists will adopt a viewpoint that calls many previously solved problems back into question. Nature itself must first disrupt the stable state of the specialty by making previous achievements appear problematic. Moreover, even when such a situation arises and a new alternative paradigm emerges, scientists will be reluctant to accept it unless two kinds of very important conditions coincide. First, the new paradigm alternative must appear to solve a conspicuous and generally recognized problem that cannot be dealt with in any other way. Second, the new paradigm must promise to preserve a fairly large part of the concrete problem-solving ability that the preceding paradigm had created for science. Innovation for its own sake is not an urgent requirement in the sciences as it is in many other creative fields. Consequently, although new paradigms possess little or none of the capabilities of their predecessors, they usually preserve many of the most concrete portions of past achievement, and always permit the emergence of additional concrete problem-solutions.
The discussion thus far does not imply that problem-solving ability is the unique or clear basis for paradigm choice. We have already seen a number of reasons why there can be no standard of that kind. But it does suggest that the professional community of science will do whatever enables it continually to increase the body of data it can handle accurately and in detail.
In that process, the scientific community will incur losses. Often, some old problems must be eliminated. Moreover, revolutions frequently narrow the domain of professional interest of the scientific community, heighten the degree of its specialization, and hinder communication with other groups, including the general public and other scientific groups. The depth of science certainly grows deeper, but its breadth may not expand in the same way. If breadth does expand, it becomes markedly broader chiefly through the diversification of scientific specialties, not within the scope of any single independent specialty. Yet despite the effects suffered by individual scientific communities, or other losses, the character of such communities provides a substantial guarantee that both the list of problems solved by science and the precision of each solution will continue to increase. At least, if there is any way such a guarantee can be given, the character of the professional community provides it. What other standard could there be higher than the decisions of the scientific group?
The immediately preceding paragraphs have indicated the directions that must be pursued in seeking a more refined solution to the problem of progress in science. Perhaps this direction will suggest that scientific progress is not what we have generally thought it to be. Yet they also show that one type of development will necessarily characterize scientific activity so long as that activity endures. Science need not have any other type of development. To put it more precisely, we may have to abandon the notion—whether explicit or implicit—that changes of paradigm lead scientists and students of science ever closer to truth.
It is now time to note that, until the last few pages remained, the term “Truth” in this essay was mentioned only in the sense quoted from Bacon. And even when it was used in that way, it was used merely as the source of the scientist’s conviction that incompatible rules of scientific activity cannot coexist except in revolutionary periods, when the principal task of the specialty becomes to eliminate all systems of rules but one.
The process of development described in this essay has been a process of evolution from a primitive initial stage—a process characterized by the fact that the successive stages that follow make possible an increasingly detailed and refined understanding of nature. But nothing discussed so far, nor anything to be said later, is meant to say that the development of science is a process of evolution toward something.
Inevitably, this gap will confuse many readers. We are all very accustomed to regarding science as an activity that steadily approaches some goal preset by nature.
But must science necessarily have such a goal? From the level of knowledge of the scientific community at any given point, can we not explain both the existence and the success of science in evolutionary terms? Does it help us to think that science has one account that conforms nature perfectly, objectively, to truth, and that a proper measure of scientific achievement is the degree to which it has brought us closer to that ultimate goal? If we can replace evolution-toward-what-we-wish-to-know with evolution-from-what-we-do-know, a number of confusing problems may disappear. For example, the problem of induction must lie somewhere in this maze.
I cannot yet set out in detail the consequences that this alternative view of scientific
progress would bring about. But it helps us to realize that the conceptual transition proposed here is very similar to the phenomenon that occurred in the West a century ago. It is especially helpful because, in both cases, the chief obstacle to the transition is the same. When Darwin first published his theory of evolution by natural selection in 1850, what most troubled many experts was neither the concept of species change nor the possibility that human beings might have evolved from apes. Evidence indicating evolution, including human evolution, had been accumulating for decades, and the idea of evolution had been proposed and widely disseminated before. Although the concept of evolution itself met with resistance, especially from religious groups, it was by no means the greatest difficulty faced by the Darwinians. The difficulty arose from views very close to Darwin’s own idea. The famous theories of evolution before Darwin—those of Lamarck, Chambers, Spencer, and the German Naturphilosophen—all regarded evolution as a goal-directed process. The “idea” of humanity and of the flora and fauna of the time was believed perhaps to have existed in the mind of God from the first creation of life. Such an idea or plan set the direction of the entire evolutionary process and served as its guide.
Each new stage in evolutionary development was a more complete realization of the plan that had existed from the beginning.4)
For many people, the collapse of this teleological kind of evolution was the most significant and most difficult issue to accept in Darwin’s proposal.5) On the Origin of Species acknowledged no goal set by either God or nature. Instead, the mechanism of natural selection, operating within a given environment and upon the actual organisms provided as material, was posited as the cause of the gradual yet steady emergence of more elaborate, more complex, and far more differentiated organisms. Even astonishingly well-adapted organs such as the human eye or hand were products of a process that began from a primitive origin and proceeded steadily without a goal—organs that had previously served as powerful arguments for the existence of a supreme creator and a preordained plan. The belief that natural selection, the result of simple competition among organisms for survival, could have produced human beings along with higher plants and animals was the most difficult and confusing aspect of Darwin’s theory. In the absence of a specific goal, what could “evolution,” “development,” and “progress” mean? For many people, these terms suddenly came to seem self-contradictory.
The analogy that connects the evolution of organisms with the evolution of scientific concepts can all too easily be stretched too far. But as far as the themes of this final section are concerned, it fits almost perfectly.
The process described in that section as the resolution of revolutions is a process of selection arising from conflict within the scientific community over the most suitable way to conduct the science of the future. The net result of such a series of revolutionary selections, separated by periods of normal research, is the astonishingly well-adapted set of instruments we call modern scientific knowledge. The successive stages in that developmental process are marked by increasing articulation and specialization. And, as we now imagine biological evolution to have been, the entire process of scientific development may have occurred without the benefit of a set goal, a permanently fixed scientific truth, of which each stage in the development of scientific knowledge is a better exemplar.
Nevertheless, any reader who has followed the discussion thus far will ask why the evolutionary process fits. What, after all, must nature—including man—be like for science to be possible? Why should scientific communities be able to reach the firm consensus that other fields cannot? Why must consensus persist through successive changes of paradigm? Why must paradigm change always produce, in some sense, a more perfect instrument than those known before? From one point of view, these questions, except for the first, have already been answered. From another, however, they remain as unresolved as when this essay began. It is not only the scientific community that must be special. The whole world of which that community is a part must also possess quite special characteristics, and we know no more than we did at the start about what those characteristics must be. But that problem—what the world must be like in order that man may know it—is not newly raised by this essay. Rather, it is as old as science itself, and it remains unanswered. Yet I do not think it need be answered here. Any conception of nature compatible with the growth of science by evidence can be compatible with the evolutionary view of science developed here. Since this view is also compatible with close observation of scientific activity, there is strong reason to apply it as an attempt to resolve the many problems that still remain open.
"Notes"
1) E.H. Gombrich, Art and Illusion: A Study in the Psychology of Pictorial Representation (New York, 1960), pp. 11-12.
2) Ibid., p.97; and Giorgio de Santillana, "The Role of Art in the Scientific Renaissance," in Critical Problems in the History of Science, ed. M. Clagett (Madison, Wis., 1959), pp. 33-65.
3) Historians of science often confront this blindness in an especially striking form. The group of students who come to the history of science from the sciences is always the most teachable audience for them. Yet it is also usually, at first, the most difficult. For, because science students “know the right answers,” it is particularly hard to make them analyze earlier science in the context of its own time.
4) Loren Eiseley, Darwin’s Century: Evolution and the Men Who Discovered It (New York, 1958), chaps. ii, iv-v.
5) A particularly penetrating account of the famous way in which the Darwinian school struggled with this problem is found in A. Hunter Dupree, Asa Gray, 1810-1888 (Cambridge, Mass., 1959), pp. 295-306, 355-83.
XIV. Postscript—1969
Seven years have now passed since this book was first published.1) During that time, through critical responses and my own further study, I have broadened my understanding of the various problems raised by this book.
My views on fundamental matters have changed little, but I now recognize the character of the book’s earlier formulations, which gave rise to needless difficulties and mistaken interpretations. Since those misunderstandings were in part my own, by removing them I have gained grounds that ultimately provide a basis for revising this book.2) At the same time, I welcome the opportunity to outline the revisions required, to address several recurring criticisms, and to indicate the direction in which my own views are now developing.3) I originally
Since several of the principal difficulties in the original text center on the concept of the paradigm, the discussion in this postscript begins with them. In the subsection that follows, I propose that it is desirable to separate that concept from the notion of the scientific community, indicate how that can be done, and discuss the analytic separation and the significant results that follow from it. Next, I consider what happens when paradigms are sought by examining the behavior of members of a scientific community already determined in advance. This procedure soon reveals that, in much of this book, the term “paradigm” is used in two different senses. On the one hand, a paradigm denotes the entire constellation of beliefs, values, techniques, and so on shared by the members of a given scientific community. On the other hand, a paradigm refers to one type of component in that constellation: the concrete puzzle-solution that, used as a model or example, can replace explicit rules as the basis for solving the remaining puzzles of normal science. The first sense of the term paradigm, which may be called sociological, is the subject of subsection 2 below. Subsection 3 concentrates on paradigms as exemplary past achievements.
At least philosophically, the second sense of “paradigm” is the more profound of the two, and the arguments I advanced under that name have been a major source of the controversies and misunderstandings aroused by this book, especially the charge that I have made science into a subjective and irrational activity. These issues are considered in subsections 4 and 5. In subsection 4, I argue that terms such as “subjective” and “intuitive” cannot properly be applied to the components of knowledge that I described as tacitly embodied in shared examples. Although such knowledge cannot be restated intelligibly by means of rules and criteria without essential change, it is systematic, has stood the test of time, and is in some sense corrigible. In subsection 5, I apply that argument to the problem of choice between two mutually incompatible theories, and, to state the conclusion briefly, I require that people with radically different views be regarded as members of different linguistic communities and that their problems of communication be analyzed as problems of interpretation. Three remaining issues are discussed in the concluding subsections 6 and 7. Subsection 6 examines the charge that the view of science developed in this book is through-and-through relativistic. Subsection 6 begins by considering whether my argument has in fact, as critics have said, gone astray through a confusion between descriptive and normative modes. Subsection 7 concludes with brief comments on topics that would each deserve an essay of their own: namely, the extent to which the main theses of this book can properly be applied to fields other than science.
1. Paradigms and the Structure of Scientific Communities “Paradigm” is a term that appears early in the preceding parts of this book, and the manner of its introduction is essentially circular. A paradigm is something shared by the members of a scientific community; conversely, a scientific community consists of people who share a paradigm. Not all circularities are bad—I shall defend an argument of similar structure later in this postscript—but the circularity here is truly a source of difficulty. Scientific communities can, and should, be formed without prior reliance on paradigms. Paradigms can then be discovered by closely examining the behavior of the members of a given community. Therefore, if I were rewriting this book, I would begin with a discussion of the structure of scientific communities, a subject that has recently emerged as an important topic of sociological research and has also begun to be treated as important by historians of science. Preliminary results, most of which have not yet been published, suggest that the empirical techniques required for such inquiry are by no means trivial; some results have already been obtained, and the rest, I believe, will certainly progress.5) Most scholars engaged in scientific activity, taking it for granted that trust in the various specialties of their day is distributed among at least roughly determinate groups of members, respond immediately to questions about their social affiliations. I shall therefore assume here that more systematic means of identifying their affiliations will be found. Instead of presenting preliminary research results, I will briefly set forth the intuitive concept of the scientific community that underlies much of the earlier chapters of this book. It is a concept now widely shared among scientists, sociologists, and many historians of science.
According to this view, a scientific community consists of the practitioners of a scientific specialty. To a degree incomparable with most other fields, they undergo similar education and professional initiation.
In that process, they have absorbed the contents of the same technical literature and drawn many of the same lessons from it. The range of such standard literature usually marks the boundaries of a scientific subject matter, and each group ordinarily has its own subject. There are schools in science—that is, groups that approach the same subject from incompatible viewpoints. Compared with other fields, however, this occurs far less often in science. They are always in competition. Their competition generally ends quickly. Thus the members of a scientific community regard themselves, and are regarded by others, as those who bear a unique responsibility to pursue a set of shared goals, including the training of their successors. Within such a group, communication is relatively complete, and professional judgment shows relatively close agreement. On the other hand, because the attention of different scientific communities is focused on different subjects, professional exchange across group lines sometimes becomes difficult, often produces misunderstanding, and, in the process, may give rise to considerable disagreements that had not been anticipated.
Scientific communities in this sense, of course, exist at various levels. At the highest, global scale is the community of all natural scientists. At only a slightly lower level are the major professional scientific groups: the communities of physicists, chemists, astronomers, zoologists, and so on.
Grouped in this way, membership in a scientific community is readily established, except at the margins. The subject of one’s highest degree, membership in professional societies, and the journals one subscribes to are generally quite sufficient. In the same way, a single group will be divided into several major subgroups: organic chemists, and also solid-state physicists, high-energy physicists, radio astronomers, and so forth. Empirical problems arise only at the next lower level. To take a contemporary example, how was the phage group separated out before it received public recognition as a group? For such purposes, a scholar must rely on attendance at specialized conferences, on the circulation of drafts or galley proofs before publication, and, above all, on the formal and informal channels of communication discovered in correspondence and in the links among cited references.6) In my view, at least with respect to the present situation and the more recent historical record, that work can be done and will be done.
Typically, it may produce a group of perhaps a hundred members, and in some cases a number far smaller than that. Individual scientists, especially the most capable ones, will ordinarily belong to several such groups, either simultaneously or successively.
Scientific communities of this type are the basic units that this book has presented as the chief agents in the production and verification of scientific knowledge. A paradigm is what the members of such groups share. Thus, without relating them to the nature of the fundamental elements thus shared, the various characteristics of science described earlier in this book cannot possibly be understood. Yet other characteristics of science, though they did not appear separately in my original manuscript, can be readily understood. Therefore, before turning directly to the topic of paradigms, it is worth examining a series of issues that need be related only to the structure of scientific communities.
Among these characteristics, the most striking is what I earlier called the transition in the development of a scientific field from a pre-paradigm period to a post-paradigm period; this transition is precisely the transformation briefly described in Section II. Before the transition occurs, various schools compete for dominance in a given field. Thereafter, in the wake of certain notable scientific achievements, the many schools are greatly reduced in number, usually converging into one, and a more efficient mode of scientific activity begins. By an effective mode of scientific activity I mean the research of a group that generally has the character of a tradition and is oriented toward puzzle-solving—something that becomes possible only when its members take the foundations of the field for granted.