Yet although Newton’s work was directed in large part toward concrete standards and problems derived from the mechanical-corpuscular worldview, the influence of the paradigm that emerged from his work produced a profound and, in part, destructive change in the problems and standards deemed proper to science. Interpreted as an innate attraction between every pair of particles of matter, gravity was a mystical quality in the same sense as the scholastic term “tendency to fall” had been. Therefore, while the norms of corpuscularianism exerted their influence, the search for a mechanical explanation of gravity became one of the most serious challenges facing those who accepted the “Principia” as a paradigm. Newton devoted much attention to this problem, and so did his eighteenth-century successors. The only obvious alternative was to reject Newtonian theory because of its failure to explain gravity, and that alternative too was widely adopted. But neither of these views ultimately succeeded. In a situation where it was impossible to do science without the “Principia,” or to make that work conform to the norms of seventeenth-century corpuscularianism, scientists gradually came to accept the view that gravity was indeed “innate.” By the middle of the eighteenth century, such an interpretation had gained almost universal acceptance, and the result was a genuine return—not exactly a regression—to the norms of scholastic philosophy. Attractive and repulsive forces inherent in matter came to take their place alongside size, shape, position, and motion as primary properties of matter that were physically irreducible.7)
The resulting change in the norms and problem-field of the physical sciences was once again inevitable. By the 1740s, for example, electricians could speak of the attractive “virtue” of an electric fluid without being subjected to the ridicule that Molière’s physician, as his opponent, had suffered a century earlier. In doing so, electrical phenomena began to display regularities different from those assumed when they had been regarded solely as the effects of a mechanical effluvium, which could act only by contact. In particular, when electrical action-at-a-distance itself became a subject of research because of its own importance, the phenomenon now called charging by induction could be recognized as one of its effects. Previously, when such a phenomenon happened to be observed, it had been attributed either to the direct action of electrical “atmospheres” or to the leakage current inevitably present in any electrical laboratory. The new view of inductive-current effects then became crucial to Franklin’s analysis of the Leyden jar, and thus to the emergence of a new Newtonian paradigm for electricity. Mechanics and electricity were not the only sciences stimulated by the legitimation of research into the innate forces of matter. Much of the eighteenth-century literature on chemical affinities and replacement series also derives from this supramechanical aspect of Newtonianism. Chemists who believed in these distinctive attractions among different chemical species devised experiments that had previously been unimaginable and sought out new kinds of reactions. Without the data and the chemical concepts developed in that process, Lavoisier’s later work—and, even more, Dalton’s—could not have been understood.8) Changes in the criteria governing permissible problems, concepts, and explanations can transform science. In the next section I shall suggest in what sense such criteria transform the world.
Other examples showing such minor differences between successive paradigms can be found in the history of almost any science in almost any period of scientific development. Here, for the moment, two brief examples will suffice.
Before the chemical revolution, one of the tasks confronting chemistry was to explain the properties of chemical substances and the changes in those properties that occurred through chemical reactions. With the aid of a small number of basic “principles”—of which the phlogiston theory was one—chemists attempted to explain why some substances were acidic, while others were metallic, combustible, and so on. There was a measure of success in this direction.
As we saw earlier in discussing the phlogiston theory, it was possible to understand why metals so closely resembled one another. A similar argument could have been developed for acids. But Lavoisier’s reform ultimately moved away from chemical “principles,” and thus ended by depriving chemistry of some actual and considerable potential explanatory power. To compensate for this loss, a change in standards was required. For a considerable portion of the nineteenth century, failure to explain the properties of compounds did not count as a condemnation of chemical theory.9)
Again, Clerk Maxwell, sharing the views of the nineteenth-century supporters of the wave theory of light, was convinced that light waves must be propagated through a material ether. Devising a mechanical medium capable of supporting the wave nature of light became a standard problem for many of the finest scholars of the age. But Maxwell’s own theory—the electromagnetic theory of light—took no account whatever of a medium that could support the wave nature of light, and it clearly made explanation more difficult than it had previously been thought to be. For this reason, Maxwell’s theory was widely rejected at first. Like Newton’s theory, however, Maxwell’s proved difficult to abandon, and as it rose to the status of a paradigm, the attitude of the scientific community toward it changed. During the first decades of the twentieth century, Maxwell’s claim that a mechanical ether existed, though of course not made dogmatically, came increasingly to sound like a merely verbal fiction, and efforts to devise such an ethereal medium were abandoned. Scientists no longer regarded it as unscientific to speak of electrical “displacement” without specifying what was being displaced. The result was again the emergence of new problems and standards, and this in the end played a major role in the birth of the theory of relativity.10) If these characteristic shifts in the scientific community’s conception of legitimate problems and standards could be assumed always to occur from a methodologically lower level to some higher form, they would have little significance for the theme of this essay. In that case, their results would also appear cumulative. It is no wonder that some historians of science argue that the history of science is the continuous record of the steady maturation and refinement of human conceptions of the nature of science.11)
Nevertheless, the case for the cumulative development of problems and standards in science is even harder to establish than the case for the accumulation of theories. The attempt to explain gravity, abandoned by most scientists of the eighteenth century, was not directed toward an essentially illegitimate problem. Objections to innate forces were not particularly unscientific, nor were they metaphysical in any contemptuous sense. There exist no external norms that permit judgments of that kind.
What actually occurred was neither the collapse of standards nor their elevation, but simply the change required by the adoption of a new paradigm. Moreover, such a change has since been reversed again, and can be reversed again.
In the twentieth century, Einstein succeeded in explaining gravitational attraction, and from this particular point of view, that explanation returned science to standards and problems more like those of Newton’s predecessors than those of Einstein’s successors.
And the subsequent development of quantum mechanics overturned the methodological taboo that had originated in the chemical revolution. Chemists now try, with considerable success, to explain the color, state, and other properties of the substances they use and produce in their laboratories. A similar reversal may occur even in electromagnetic theory.
In modern physics, space is not the inactive and homogeneous substratum introduced in both Newtonian and Maxwellian theory.
Some of the new properties of space are not unlike those once attributed to the ether. We may one day come to know what electrical displacement is.
If emphasis is shifted from the cognitive function of paradigms to their normative function, the preceding examples broaden our understanding of the way paradigms give form to scientific activity. Earlier, we examined chiefly the role of paradigms as vehicles for scientific theory. In that role, a paradigm tells the scientist about the entities that nature does and does not contain, and about the ways in which those entities behave. Such information provides a map whose details are disclosed by mature scientific research. And because nature is too complex and diverse to be unveiled at random, such a map becomes as essential to the continuing development of science as observation or experiment. Through the theories they embody, paradigms prove to be constitutive elements shaping research activity. Yet they also become constitutive parts of science from another perspective, and this is the point here. In particular, the examples just given show that paradigms provide scientists not only with a map but also, to some degree, with the directions essential for making maps. In learning a paradigm, the scientist acquires theory, methods, and standards all at once, usually as an inextricably intertwined mixture. Therefore, when paradigms change, there is normally a significant shift in the criteria determining the legitimacy of both problems and proposed solutions.
This observation brings us back to the point from which Section IX began, for it provides the first explicit evidence of why the choice between competing paradigms regularly raises questions that the criteria of normal science cannot resolve. To the extent that two schools of science differ, meaningfully as well as incompletely, over what counts as a problem and what counts as a solution, they will inevitably talk past one another when debating the relative merits of their respective paradigms. In the partially circular arguments that regularly arise, each paradigm will be shown to satisfy, to some extent, the criteria it itself dictates, while also lacking some of the criteria dictated by its rival. In any case, because no paradigm has ever solved all the problems it set out to solve, and because the problems left unsolved by two paradigms are not all the same, paradigm debates always involve the following question: Which problems is it more significant to have solved? Like the issue of competing standards, this question of values can be answered only by criteria that lie wholly outside normal science, and it is this reliance on external criteria that most unmistakably makes paradigm debates revolutionary. But something even more fundamental than standards and values is also at stake. Until now, I have argued only that paradigms constitute science. From this point on, I wish to clarify the sense in which paradigms also constitute nature.
“Notes”
1) Silvanus P. Thompson, Life of Wklliam Thomson Baron Kelvin of Largs(London, 1910).
2) See, for example, P.P. Wiener’s commentary in Philosophy of Science XXV(1958), 298.
3) Jzmes B.Conant, Overthrow of the Phlogiston Therory(Cambridge, 1950), pp.13-16;J.R.Partington, A Short History of Chemistry(2d ed;London,1951), pp.85-88. The most complete and sympathetic account of the contributions to phlogiston theory is H.Metzger, Newton, Stahl, Boerhaave et la dectrine chimique(Paris, 1930), part II.
4) R..B.Braithwaite, Scientific Explanation(Cambridge, 1953), pp.50-87, a conclusion reached through an entirely different type of analysis. Compare especially p.76.
5) For general corpuscularism, see Marie Boas, "The Establishment of the Mechanical Philosopy", Osiris, X(1952), 412-431. On the effect of particle shape on taste, see p.493 of the above work.
6) R.Dugs, Ls mecanique au XVIP siecle(Neuchatel, 1954), pp.177-85, 284-98,345-56.
7) I.B.C, Franklin and Newton:An Inquiry into Speculative Newtonian Expermenlal Science and Franklin`s Work in Electricity as an Example There of(Philadelphia, 1956), chaps. vi-vii.
8) For electricity, see chapters viii-ix of note 7). For chemistry, see Metzger, op. cit, part I.
9) E.Meyerson, Identity and Reality (New York, 1930), chp. X.
10) E.T.Whittaker, A Historu of the Theories of Aether and Electricity, Ⅱ(London, 1953), 28-39.
11) For an ingenious and thoroughly up-to-date attempt to fit the development of science to this Procrustean bed, see C.C.Gillispie, The Edge of Objectivity: And Essay in the History Scientific Ideas (Princeton, 1960).
X. Revolutions as Changes of World View
Revolutions as changes of World View
Looking over the record of past research from the standpoint of modern historical interpretation, historians of science may be tempted to assert that when paradigms change, the world itself changes with them. Guided by a new paradigm, scientists adopt new instruments and look into new areas. More important still is the fact that, during revolutions, scientists see new and different things while observing with familiar instruments in places they had previously studied. It is rather as if the professional community were suddenly transported to another planet, where formerly familiar objects appear differently and are mixed with unknown ones. Of course, nothing of quite that sort actually occurs; there is no geographical relocation. Everyday life outside the laboratory continues as before. Nevertheless, paradigm changes cause scientists to see the world of their research activity differently. Insofar as scientists deal with such a world only through what they see and do, we are led to say that after a revolution scientists are responding to a new world.
As a fundamental prototype for these transformations in the scientist’s world, the evidence of familiar shifts in visual gestalt proves highly suggestive. What was a duck in the scientist’s world before the revolution becomes a rabbit afterward. The person who at first looked down from above upon the outside of the box later looks up from below into its interior. Though transformations of this kind are generally more gradual and almost invariably irreversible, they are a common accompaniment of scientific training. Looking at a contour map, the student sees lines drawn on paper, but the cartographer sees a picture of terrain. Looking at a bubble-chamber photograph, the student sees confused and broken lines, but the physicist reads a record of familiar events inside an atomic nucleus. Only after undergoing many such visual transformations does the student become a member of the scientist’s world, seeing what the scientist sees and responding as the scientist responds. Yet the world the student thus enters is not fixed once and for all, on the one hand by the nature of the environment and on the other by the nature of science. It is jointly determined by the environment and by the particular tradition of normal science he has been trained to pursue. Therefore, in periods of revolution, when the tradition of normal science changes, the scientist’s own perception of his environment must be reeducated—the scientist must learn to see a new gestalt in some familiar situation. After he has done so, his world of research will, in various ways, appear unable to be compared by the same standards as the world in which he had previously lived; it will appear
incommensurable. This is another reason why schools guided by different paradigms are always bound to be somewhat at cross-purposes.
Of course, in their most general form, gestalt experiments explain only the nature of perceptual transformations. They tell us nothing about the role of paradigms, or about the role of experience already assimilated in the process of perception. On this point, however, the psychological literature is rich, much of it deriving from the pioneering work of the Hanover Institute. A subject wearing spectacles with lenses that invert the image initially sees the entire world upside down. At first his sensory organs function as they did when they had been trained to function without the spectacles; consequently he suffers extreme disorientation and is confronted with a serious crisis. But only after the subject passes through a transitional period, usually one in which vision is severely confused, and begins to know how to deal with the new world, does his entire field of vision turn upside down.
Thereafter objects are again seen as they were before he put on the spectacles.
The previously anomalous field of vision has been assimilated into the visual field and has transformed the field itself.1) Not only metaphorically but literally, the person accustomed to these inverting lenses has undergone a revolutionary transformation in vision.
The subjects in the anomalous-card experiment discussed in Section VI experienced a very similar transformation. Until they had looked at the cards long enough and realized that there were strange cards in that world, they saw only the card forms that their previous experience had provided for them. But once they had experienced an additional category, the subjects could recognize all the strange cards at a glance, provided they were given enough time to identify them. Many other experiments have also shown that the size and colors of objects presented experimentally are perceived differently according to the subject’s prior training and experience.2) Looking over the rich experimental literature containing such cases, one is led to feel that something like a paradigm is a prerequisite for perception itself. What a person sees depends not only on the object he is looking at, but also on what his previous visual-conceptual
experience has taught him to see. In the absence of such training, there can be only what William James called “a blooming, buzzing confusion.”
In recent years, many people interested in the history of science have come to recognize how suggestive experiments of the sort described above can be. In particular, H.R. Hanson, by using concrete demonstrations concerning form, brought out some of the same consequences for scientific belief that are my concern here.3) Other colleagues have repeatedly noted that if we can suppose that scientists frequently experience perceptual shifts like those described above, the history of science may take on a better and more coherent meaning. Yet suggestive though psychological experiments may be, by the nature of the situation they can be no more than that. They reveal characteristics of perception that may be central to the development of science, but they do not prove that such characteristics are present in the careful and controlled observations made by scientific researchers. Moreover, the very nature of these experiments makes direct demonstration of that point impossible. If historical examples are to appear connected with these psychological experiments, we must first attend to the types of evidence that history does or does not provide.
Because the subject in a gestalt demonstration can repeatedly switch his perception while holding the same book or paper in his hand, he knows that it is changing. Knowing that there has been no change in the environment, he gradually focuses his attention not on the figure—the duck or the rabbit—but on the lines on the paper he is seeing. Eventually he may become able to see only the lines without seeing any figure, and he will say that what he really sees are these lines, but that he sees them alternately as a duck and as a rabbit. (This is something he certainly could not have said before.) Similarly, the subject in the anomalous-card experiment comes to feel—or, more precisely, is persuaded—that his perception must have changed, because the experimenter, as an external authority, assures him that regardless of what he perceived, he had been looking all along at a black five of hearts. As in all psychological experiments of a similar kind, in both experiments the effect of the demonstration depends on its being analyzable in this way. If there were no external criterion by which a shift of vision could be demonstrated, no conclusion could be drawn about the possibility of an alternative perception.
In the case of scientific observation, however, the situation is completely reversed. The scientist can rely on nothing except what he has seen through his own eyes and instruments. If there existed grounds of higher authority that could show his vision had changed, that authority itself would at once become a source of data for the scientist, and the scientist’s act of seeing would become a source of problems, just as the behavior of experimental subjects was for the psychologist. If a scientist could repeatedly shift his perception, as the subject in a Gestalt experiment can, problems of exactly the same kind would arise. The era when light was “sometimes a wave and sometimes a particle” was an era of crisis__an era in which something was wrong__and that crisis came to an end only when wave mechanics was developed and it became known that light was an entity in its own right, different from both wave and particle. Therefore, if a transformation of perception in science accompanies a paradigm change, we need not expect scientists themselves to attest directly to such changes. A convert to Copernicanism does not look at the moon and say, “Until now I was seeing a planet, but now I am seeing a satellite.” Such a manner of speaking might imply that the Ptolemaic system had once been correct. Instead, the convert to the new astronomy says, “I once thought the moon was a planet—or saw it that way—but I was mistaken.” Statements of this sort appear repeatedly after scientific revolutions have occurred. If such statements ordinarily disguise a transformation of scientific vision, or various mental transformations having the same effect, then we cannot expect direct evidence for that transformation. Rather, we must seek indirect behavioral evidence that the scientist with a new paradigm is seeing in a way different from what he saw before.
Let us now return to the data and see what kinds of transformations a historian of science who believes in such changes can discover in the scientist’s world. Sir William Herschel’s discovery of Uranus is the first example, a case very similar to the anomalous-card experiment. Between 1690 and 1781, on at least seventeen occasions, many astronomers, including some of Europe’s finest observers, saw a star in the region of what is now the orbit of Uranus. The most outstanding observer among this group actually saw the star on four consecutive nights in 1769, but failed to notice its motion, which might have revealed its identity. Twelve years later, when Herschel first observed that very object, he used a far more improved telescope that he had made with his own hands. As a result, he was able to discern a distinct disk, quite unusual at least for something star-shaped. Something was wrong, and so he postponed judgment and investigated further. His investigation revealed the motion of Uranus among the stars, and Herschel therefore announced that he had seen a new comet! After futile attempts to fit the observed motion to a cometary orbit, it was not long before Lexell proposed that its orbit seemed to be that of a planet.4) Once that claim was accepted, the astronomer’s world contained several fewer fixed stars and one additional planet. A celestial body that had been observed off and on for nearly a century came to be seen differently after 1781. The reason, as in the anomalous-card experiment, was that it could no longer fit into the categories of perception provided by the old paradigm—star or comet.