But the change of vision that enabled astronomers to see Uranus as a planet does not seem to have affected only the perception of that already observed object. Its consequences were more far-reaching.
Though the evidence is ambiguous, the small paradigm shift brought about by Herschel seems to have helped astronomers, after 1801, rapidly discover a number of minor planets (or asteroids). Because they were small, the asteroids did not display the magnification into anomaly that had astonished Herschel. Nevertheless, astronomers prepared to find more planets were able, using standard instruments, to identify twenty planets during the first fifty years of the nineteenth century.5) The history of astronomy contains many other cases of paradigm-induced changes in scientific perception, some of which appear more certain. For example, can it have been mere chance that, for half a century after Copernicus’s new paradigm was first proposed, Western astronomers first witnessed changes in the celestial world, which had previously been regarded as immutable? The Chinese, whose cosmology did not exclude change in the heavens, had long before recorded the appearance of many new stars. Also, without the aid of the telescope, the Chinese had systematically recorded the appearance of sunspots centuries before Galileo and his contemporaries discovered them.6) Nor were sunspots and new stars the only cases of celestial change grasped by Western astronomy in the period immediately after Copernicus. Using traditional astronomical instruments, including some as simple as a thread, astronomers of the late sixteenth century continued to discover comets wandering freely through regions that had previously been permitted only to immutable planets and fixed stars.7) The fact that astronomers, observing old objects with old instruments, so easily and quickly saw new things tells us that post-Copernican astronomers had come to live in a different world. In any case, their research responded as though that were so.
The preceding examples have been chosen from astronomy because records of celestial observation are often written in a vocabulary of relatively pure observational terms. Only in such records is there any possibility of finding something like a complete analogy between the observations of scientists and the observations of psychologists’ experimental subjects. But we need not greatly emphasize such perfect parallelism, and by relaxing our standards we gain a great deal. If we can be satisfied with the ordinary use of the verb “to see,” we shall soon realize that we have already encountered many cases of changes in scientific perception accompanying changes of paradigm. Extending the use of “perception” and “seeing” will immediately require an explicit defense, but let us first examine its application in scientific activity.
Let us briefly reconsider the two earlier examples from the history of electricity. During the seventeenth century, when research proceeded according to one or another effluvium theory, electricians repeatedly observed particles leaping up from and falling back to electrified bodies that attracted them. At least, this is what seventeenth-century observers said they saw, and we have no reason to distrust the records of their perception more than our own. Sitting before the same experimental apparatus, a modern observer would see electrostatic repulsion, rather than a mechanical effect or attraction; but, except for one historically neglected exception, electrostatic repulsion was not observed in itself until Hauksbee’s large-scale apparatus greatly magnified the effect. Yet repulsion after contact electrification was only one of the many new repulsion effects that Hauksbee saw. Much as in a gestalt transformation, through his work repulsion suddenly became established as the most fundamental feature of electrical action, and it was attraction that then had to be explained.8) The electrical phenomena visible in the early eighteenth century were more subtle and varied than those of the seventeenth. Or again, after assimilation to Franklin’s paradigm, the electrician looking at a Leyden jar saw something different from what he had seen before. The apparatus was now a condenser, and as a condenser it required neither the shape of a jar nor glass. Instead, two conducting coatings—one of which had not existed in the original device—became common. As written discussions and illustrated explanations increasingly demonstrated, two metal plates without a nonconductor inserted between them became the prototype of this class.9) At the same time, new explanations were given for other induction effects, and still other effects came to receive attention for the first time.
Transformations of this sort are not confined to astronomy and electricity. We have already considered some similar visual transformations that can be drawn from the history of chemistry. Lavoisier saw oxygen where Priestley had seen dephlogisticated air, and where others had seen nothing at all. But in coming to recognize that he was seeing oxygen, Lavoisier also had to undergo changes in his views of other familiar substances.
For example, where Priestley and his contemporaries saw elemental earths, he had to see compound minerals, and there were many other such changes. As a result of discovering oxygen, at the very least Lavoisier came to see nature differently. And unless we resort to a hypothetical fixed nature that he “saw differently,” the principle of economy in nature will lead us to say that, after discovering oxygen, Lavoisier worked in an entirely different world.
We shall soon consider the possibility of avoiding this unfamiliar expression, but first we need more examples that accord with such usage, and these may be drawn from the most famous part of Galileo’s work. Since time immemorial, most people had seen one heavy object or another, suspended from a cord or chain, swing back and forth until it came to a complete stop.
Because Aristotelians believed that a heavy body, by its own nature, moved from a higher place toward its natural state of rest in a lower place, a swinging body was for them merely something falling with difficulty. Since the body was bound by a chain, it would come to rest in its lower position only after a considerable time, by way of a curved motion. Galileo, on the other hand, looking at a swinging body, thought of a pendulum: a body that repeats almost endlessly the same motion. Seeing this pendular motion, Galileo constructed a great deal around the pendulum’s other properties. From the properties of the pendulum, for example, Galileo derived not only his account of the relation between vertical height and final velocity in motion down an inclined plane, but also his only fully complete and firm theory concerning the irrelevance of weight to the speed of fall.10) He saw all these phenomena of nature differently from the way they had been seen before.
Why did such a transformation of vision occur? Of course, through the personal genius of Galileo. But it is worth noting that here such genius was not manifested in observing the swinging body more accurately or more objectively. As a description, the Aristotelian perception is no less accurate than Galileo’s. When Galileo reported that, for amplitudes up to ninety degrees, the period of the pendulum is independent of amplitude, his view of the pendulum led him to see far greater regularity there than we can now see.11) Rather, what seems to have been involved in his transformation of vision was the genius’s use of the possibilities of perception provided by a medieval paradigm transformation. Galileo was not raised as a perfectly Aristotelian thinker. Rather, he had been trained to analyze motion in terms of impetus theory, a late medieval paradigm that explained the continued motion of a heavy body by saying that an internal force had been implanted in it by the projector that initiated its motion. It was Jean Buridan and Nicole Oresme, fourteenth-century Scholastics, who gave impetus theory its fullest systematization, and they are known as the first figures to have anticipated some of Galileo’s observations on oscillatory motions. Buridan explains the motion of a vibrating string as the first infusion of impetus when the string is struck. The impetus is then expended in changing the string’s position against the resistance of the string’s tension. The tension then snaps the string back, infusing more and more impetus until it reaches the midpoint of its motion. After that, the impetus again displaces the string in the opposite direction against the tension, and so on in a symmetrical process that can continue indefinitely. Later in the fourteenth century, Oresme diagrammed a similar analysis for a swinging stone, and this now appears to be the first discussion of the pendulum.12) His view is clearly very similar to the way Galileo first approached the pendulum. At least in the case of Oresme, and almost certainly in Galileo’s as well, before the scholastic impetus paradigm for motion had been created out of the originally Aristotelian one, scientists did not see pendulums, but only swinging stones. The pendulum took on reality through something very much like a paradigm-induced gestalt switch.
But is it necessary to describe what distinguishes Galileo from Aristotle, or Lavoisier from Priestley, as a transformation of vision? Looking at the same kinds of objects, did those men truly see different things? Is there any valid sense in which we can say that they carried out their research in different worlds? These questions can no longer be postponed, for there is plainly another, far more conventional way of explaining all the historical examples surveyed above. Many readers will surely want to say that what changes with a paradigm is only the scientist’s interpretation of observations, which are themselves permanently fixed by the nature of the environment and the perceptual apparatus. According to this view, Priestley and Lavoisier both saw oxygen, but interpreted their observations differently.
Aristotle and Galileo both saw pendulums, but differed in their interpretations of what they had seen.
First of all, there is nothing wholly wrong, nor can it be called sheer error, in this exceedingly common view of what happens when scientists change their thinking on fundamental subjects. Rather, it is the essence of the philosophical paradigm created by Descartes and at the same time developed as Newtonian mechanics. That paradigm contributed greatly to both science and philosophy. Like mechanics itself, the thorough use of this paradigm made possible fundamental understandings that could not otherwise have been achieved. But as the case of Newtonian mechanics also illustrates, even the most astonishing successes of the past do not guarantee that crisis can be postponed indefinitely. Today, research in philosophy, psychology, linguistics, and even art history all alike suggests that the traditional paradigm was somewhat astray. Such failures of adaptation are becoming increasingly evident as well through the historical study of science, on which our attention here is for the most part necessarily focused.
Although none of the subjects fostering these crises has yet provided a valid alternative to the traditional epistemological paradigm, they are beginning to suggest what some features of such a paradigm might be. For example, I am well aware of the difficulties caused by saying that when Aristotle and Galileo looked at a swinging stone, Aristotle saw a constrained fall, while Galileo saw a pendulum. At the beginning of this section, that very difficulty appears in a more fundamental form. Though the world does not change with a change of paradigm, the scientist afterward works in a different world than before. Nevertheless, I am convinced that we must learn to understand the meaning of statements at least similar to these. What occurs during a scientific revolution cannot be wholly reduced to a reinterpretation of individual and stable data. Above all, the data are by no means indisputably stable.
A pendulum is not a falling stone, and oxygen is not dephlogisticated air. Consequently, as we shall soon see, the data that scientists gather from these various objects are themselves different. More importantly, the process by which an individual scientist or a scientific community accomplishes the transition from constrained falling motion to a pendulum, or from dephlogisticated air to oxygen, is not a process resembling interpretation. How could a scientist do so in the absence of fixed data to be interpreted? The scientist who adopts a new paradigm is less an interpreter than a person wearing lenses that reverse what is seen. Confronting the same countless objects as before, and knowing that he is looking at objects that have not changed, the scientist nevertheless comes to realize that, here and there in their details, those objects have been transformed through and through.
This remark is not intended to indicate that scientists do not characteristically interpret observations and data. On the contrary, Galileo interpreted his observations of pendulums, Aristotle his observations of falling stones, Musschenbroek his observations of charged jars, and Franklin his observations of condensers. But each of these interpretations presupposed a certain paradigm. And these, as we have seen above, were parts of normal science—work aimed at refining, extending, and clarifying a paradigm that already existed. In Section III, several cases were presented in which interpretation played a central role. Such cases are typical of the overwhelming majority of scientific research. In each of them, the scientist, on the basis of an accepted paradigm, came to know what the data were, what instruments should be used to obtain them, and what concepts were relevant to their interpretation. Given a paradigm, the interpretation of data becomes the core of the work of exploring that paradigm.
But this work of interpretation—and this was dealt with in the paragraph before last—can only refine a paradigm; it cannot alter it. Paradigms, after all, are not things that can be corrected by normal science. Rather, as we have already seen, normal science ultimately leads only to the recognition of anomalies and to crisis. And these are brought to an end not by deliberation and interpretation, but by relatively sudden and unstructured events like a Gestalt switch. Scientists then often speak of “scales falling from the eyes,” or of a “lightning flash” that “inundates” a previously obscure puzzle, thereby making its components appear, for the first time, in a new way that permits solution. In other cases, the relevant insight is obtained in a dream.13) No ordinary meaning of the term “interpretation” fits these flashes of intuition through which a new paradigm is born.
Such intuitions depend upon anomalous and also integrative experiences obtained within the old paradigm, but the intuitions are not, as in interpretation, connected logically or piecemeal to particular items of such experience. Rather, intuition gathers together many portions of that experience and transforms them into a quite different bundle of experience, which thereafter will be linked one by one not to the old paradigm but to the new one.
To understand in detail what these differences in experience amount to, let us return for a moment to the story of Aristotle, Galileo, and the pendulum. What data did the interaction between their different paradigms and their common environment make accessible to each of them?
Observing constrained fall, an Aristotelian would measure—or at least discuss, for Aristotelians rarely measured—the weight of the stone, the vertical height to which it had been raised, and the time it took to fall and come to rest. Together with the resistance of the medium, these were the conceptual categories used by Aristotelian science in dealing with a falling body.14) Normal research guided by them could not have produced the laws that Galileo discovered. It could only have led to a series of crises that gave rise to Galileo’s way of thinking about the swinging stone, and along another route it actually did so. As a result of those crises and other intellectual changes, Galileo came to see the swinging stone in an entirely different way. Archimedes’ studies of floating bodies made the medium inessential. Impetus theory made motion symmetrical and enduring. And Neoplatonism turned Galileo’s interest toward the circular form of motion.15) Thus he measured only weight, radius, angle of displacement, and period of swing, and these were precisely the data that could be interpreted to yield Galileo’s laws of the pendulum. In fact, interpretation proved almost unnecessary. Given Galileo’s paradigm, regular motions such as that of the pendulum were very easy to approach. How else can one explain Galileo’s discovery that the period of the pendulum is entirely independent of its amplitude? That discovery was something that the normal science originating with Galileo had to uproot, and it is something we today find quite difficult to record in detail. Regular phenomena that could not exist for an Aristotelian—and in fact are nowhere exactly exemplified by nature—were the results of immediate experience for a person who, like Galileo, saw the swinging stone in that way.
Since the Aristotelians left no record of discussions of the swinging stone, perhaps such an example is too fanciful. Within their paradigm, it was an immensely complex phenomenon. But the Aristotelians did discuss a simpler case, namely a stone falling without special constraint, and here too the same difference of vision is revealed. Observing a falling stone, Aristotle saw a change of state rather than a process. Therefore, for him, the meaningful measures of motion were the total distance moved and the total time elapsed, parameters that provide not what we today call velocity, but what we should call average velocity.16) Likewise, because Aristotle held that the stone was determined by its nature to reach its final position of rest, he saw the meaningful distance parameter at any point during the motion as the distance to the final point of the motion, not the distance from its starting point.17) Such conceptual parameters are embedded in most of his famous “laws of motion” and give them their meaning. But scholastic criticism, partly through the impetus paradigm and partly through the theory known as the latitude of forms, changed this way of thinking about motion. A stone moved by impetus acquired more and more impetus the farther it moved from its starting point. Thus the distance from the starting point, not the distance to the point of rest, became the meaningful parameter. In addition, Aristotle’s notion of velocity was divided by the scholastics into two branches, which soon after Galileo became established as today’s concepts of average velocity and instantaneous velocity. But when these concepts are considered through the paradigm of which they formed a part, the falling stone, like the pendulum, almost at a glance revealed the laws that governed it. Galileo was not the first to assert that stones fall with uniformly accelerated motion.18) Moreover, before he carried out experiments on the inclined plane, he had already developed his theorem on this subject together with many of its consequences. That theorem was another among the patterns of new regularity recognized by genius in a world determined by nature and by the paradigm that had formed Galileo and the scholars of his age. Living in such a world, Galileo could, if he wished, explain why Aristotle had seen motion as he did. Nevertheless, the immediate content of Galileo’s experience of the falling stone was different from Aristotle’s.
Of course, it is by no means certain that we need be so concerned with “immediate experience”__that is, with the perceptual features that highlight the way paradigms reveal their regularities all at once. Those features must surely change according to the scientist’s commitment to a paradigm, but they are far removed from what we generally associate with raw data or blind experience, which are regarded as what advances scientific research. Perhaps immediate experience, being fluid, must be set aside, and instead we should discuss the concrete operations and measurements the scientist carries out in his laboratory. Or perhaps the analysis must proceed more deeply from what is immediately given. For example, it might be conducted in terms of some neutral observation-language, one designed to correspond to the retinal images that convey what the scientist sees. Only through one of these methods could we hope to recover once again a realm in which experience is permanently stabilized__a realm in which the pendulum and constrained fall would not be different acts of perception, but rather different interpretations of the definite data obtained by observing a swinging stone.
But is sensory experience firm and neutral? Is theory nothing more than an interpretation that humans attach to given data? The epistemological viewpoint that has almost dominated Western philosophy for the past three centuries immediately and emphatically answers yes. I believe that, in the absence of a developed alternative, it is impossible to withdraw that viewpoint entirely. Yet that viewpoint no longer functions effectively, and attempts to make it function by introducing a neutral observation-language now seem to me to offer no prospect.
The operations and measurements that a scientist performs in the laboratory are not so much “the given” in experience as “the collected with difficulty.” They are not what the scientist sees__at least not until his research has advanced to a considerable level and his attention has been focused. Rather, they are concrete indices of the meaning of more basic acts of perception, and they are selected as objects of rigorous investigation in normal research only because they promise opportunities for the fruitful refinement of an accepted paradigm. Operations and measurements are, in part, far more paradigm-dependent than the intuitive experience that gives rise to them. Scientists who adhere to different paradigms perform different concrete experimental operations. The measurements to be carried out with respect to a pendulum are bound to differ from those relevant in the case of constrained falling motion. The experiments needed to elucidate the properties of oxygen cannot be the same as the operations required in considering the characteristics of dephlogisticated air.