Science Turned Upside Down

The word “revolution” first brings to mind violent upheavals in the state, but ideas of revolution in science, and of political revolution, are almost coeval. The word once meant only a revolving, a circular return to an origin, as when we speak of revolutions per minute or the revolution of the planets about the sun. As a political metaphor, a revolution could, in that sense, mean only a return to better times, or to the true constitution, a ridding of excess or usurpers. The transition to our modern idea was made with the Glorious Revolution of 1688, which made Parliament supreme in England and brought in a Bill of Rights. Yet it could be called a revolution, in those days, partly because of the old sense of the word: the traditional rights of Englishmen had been returned to them. A century later the American and French revolutions were definitive breaks with the past, irreversible establishments of a new order, but even in those cases there was a lingering sense of return. The original rights of man had been restored.

Revolution in science parallels these developments. At the time of the Glorious Revolution there was some rather inconsequential talk of the new science of the seventeenth century being revolutionary. A century later Lavoisier—progenitor of modern chemistry—confidently and unself-consciously proclaimed that he and his peers had effected a revolution. This was a truly irreversible change, and not a return to some older knowledge. It is possible to argue that our present conception of revolution was staked out more securely in science than in political action.

Despite a certain amount of rhetoric, such as “the second American Revolution,” there is a fair consensus about which events in the affairs of a people can rightly be called revolutions. It is also clear that such revolutions are proper objects of study for the historian. In science, however, the label of “revolution” has been so taken up by public relations that any innovative design for an aircraft wing, or technique for splicing genes, is hailed as revolutionary. The merely novel is hyped into a revolution. If we resist such exaggeration we will settle on the work of certain scientists as revolutionary: Copernicus, Darwin, or Einstein, for example.

But it is not evident that revolution should be as interesting in the history of science as in the history of nations. The French Revolution was important because it was a revolution. Darwin was important, but being revolutionary is hardly what makes his contribution to science important. One can even begin to wonder whether to call an event a revolution in science is merely to give an honorific award, like a knighthood or a Nobel prize. Not surprisingly, “revolution” was not a term used in an analytic or theoretical way in the history or philosophy of science until after the Second World War. Since then, however, it has been a central part of a dialogue, or confrontation, between two very different approaches. One is evolutionary and continualist. The other is revolutionary or cataclysmic.

Most scientists and lay people unreflectively have the evolutionary attitude. Certainly at the time of Lavoisier, there were rapid changes in chemistry. Burning had been understood in terms of an imponderable substance called phlogiston; in the 1780s it became well understood that there is no such thing, and that a newly discovered substance, named oxygen, was the key to combustion, breathing, and bleaching. Acids had long been known: now oxygen was (wrongly) taken as the defining component of acids. Even more important, there arose a clear understanding of the difference between chemical compounds (such as iron oxide or carbon dioxide) and mixtures (such as bronze or air). But, the continualist adds, although these changes were rapid enough to merit the praising word “revolution,” they are part of a larger evolution, and differ only in pace, and not in kind, from other developments. Oxygen had been isolated by at least two investigators before Lavoisier. The conceptions of element, compound, and mixture have a long and sometimes painful history.

The chemical revolution of the 1780s, on this view, was simply the rapid fruition of one group of ideas and techniques. Some of it was permanent, while other parts of it needed firm correction (e.g., the account of acidity). It certainly commenced a period of less sensational progess into the inorganic compounds, followed, in the latter part of the nineteenth century, by bold leaps of understanding, analysis, and synthesis of organic compounds, which involve carbon and are the substance of life itself. The dramatic events of the past thirty years in molecular biology are part of the ongoing story—a story of overall progress, of building on the discoveries of our predecessors.

This comfortable view of the growth of knowledge has an iconoclastic and cataclysmic counterpart. Science, it is proposed, grows by fits and starts. There are periods of routine learning that run into some sort of block. The only resolution is a wholesale revision of assumptions and concepts. Lavoisier did not merely advance upon his teachers. He transformed the way in which his group of coworkers (and their successors) think of the world. Earlier descriptions of burning came to be regarded less as false than as unintelligible. Scientific revolutions are just those breaks in knowledge that reorganize our vision of some aspects of the world—reorganize it to such an extent that we can no longer see the world in the old way. Moreover, according to a strong version of this point of view, such mutations in thought are not peculiar to singular heroes such as Lavoisier, Darwin, or Einstein. They are common in the numerous subdisciplines. Radical cuts with the past are perhaps essential to any long-term inquiry that is not to be trapped in scholastic doldrums.

Note, incidentally, that the word “revolution” is not needed to express this point of view. I’ve just spoken of breaks, cuts, and mutations, all words put to such a use in the Thirties by the prolific French philosopher of science and of aesthetics, Gaston Bachelard. They are now the standard terms for French writers, with coupure, in numerous schools, being the word of choice. “Revolution” is, however, the word for debating such issues in English, and there is a history to that which I shall soon describe.

It is perfectly natural to be eclectic between the evolutionary and revolutionary attitude, but a number of vigorous writers align themselves one way or the other. These include both working scientists and observers such as historians or philosophers or science writers. If we were to appoint an arbitrator to help us decide where the truth lies, we would first want a magisterial survey of all the different examples of revolution in science, so that we could think about their frequency, their centrality, and their structure. That is one of the promises of I.Bernard Cohen’s book. It might seem, from a look at its table of contents, that every plausible candidate for a revolution in science since the Renaissance will receive its due. Moreover Cohen presents himself as a historian, uncommitted to this or that philosophical position, but determined to display the historical evidence that is of interest.

Revolution in Science is a cornucopia of information, some familiar, some arcane, part of the knowledge accumulated by one of our most distinguished and most versatile historians of science. He is not one for cataclysms in knowledge, but he is no foe of novelty. He has just reissued a book first published in 1960, The Birth of a New Physics, an important account of the transformation of science from Aristotle to Kepler, and of how that set the stage for Newton.¹ The new edition adds some fifty pages of supplements to the former text of 150 pages. Some of these are technical, while others are elaborations of such themes as “What Galileo ‘saw’ in the heavens.” Cohen’s favorite scientist is Newton, whom he sees as revolutionary; indeed his most recent volume about the man appeared in 1980 as The Newtonian Revolution, with Illustrations of the Transformation of Scientific Ideas.²

Like Cohen’s numerous other books, those two fix on a single limb of science. Revolution in Science addresses the entire body. There is nothing else like it for a panorama of what people have chosen to call revolutionary in science. It is also immensely rich in erudition, pleasantly and casually presented in 472 pages of beautifully printed text; almost one hundred pages of more finely printed supplements, and over fifty pages of notes. Weighty, yes, but the material is not in the least oppressive.

The title of the book is significantly not “Revolutions in Science,” an imagined work that might have chapters on Copernicus, Harvey, Galileo, Freud, Einstein, and so forth. Such a book would chiefly tell us what these men achieved. But no, the title is “Revolution in Science.” It is about the idea of revolution in science. The book is a triple-decker. On the first deck are some scientific revolutions, some associated with the names I’ve just listed, and Lavoisier too. On the second deck there are people who describe those events as revolutions. Thus we learn how Lavoisier himself wrote of a revolution in chemistry, and we learn how subsequent writers of different kinds did likewise. Finally on the third deck, or perhaps the bridge, stands Cohen, describing how people described events in terms of revolution.

Cohen provides criteria for recognizing an event as a revolution, but his procedure is curious. He does not primarily give first-deck criteria, at the level of the revolutions themselves, but second-deck criteria, based on whether or not various kinds of people called an event a revolution. He calls four kinds of witnesses. These are (1) scientists contemporary with the event—none better than the very creator of the event, such as Lavoisier in the case of chemistry; (2) later scientific documents in the same discipline—and Lavoisier’s heirs do indeed speak of a revolution in chemistry; (3) historians past and present—once again, these concur that there was a chemical revolution in the 1780s; (4) present scientific tradition in the same field, so that a reflective chemist may be asked whether Lavoisier’s work was revolutionary.

When all four kinds of witnesses agree in calling an event a revolution in science, Cohen is sure that he has located a revolution. There is something strange about this. We do not determine whether there has been a national revolution by asking whether various witnesses call it a revolution. It is instructive, for example, that there is talk of a “quiet revolution,” said to have occurred in Quebec during the past quarter century. But if one wanted to argue that there had been a revolution, one would cite, perhaps, the fact that in a single generation the birth rate dropped from virtually the highest in the world to virtually the lowest, thus suggesting a striking change in the life of the family. If enough interconnected items like that can be presented, we may decide in favor of there having been a quiet revolution. Conversely, if we cannot find such phenomena we will conclude that talk of a quiet revolution is just the hyperbole of journalists and speech writers. The event itself determines whether it is worth calling a revolution, or so we think.

It is not quite fair to report that Cohen attends only to what his four kinds of witnesses say. His first criterion—the event should be called a revolution in its own day—is inapplicable to non-Western science and to Western science before the seventeenth century. So he invokes another principle: a revolution should be marked by a change in the state of the science. Practices, techniques, theories, and investigations should be substantially altered. Cohen uses this idea to examine the famous Copernican revolution in which the sun replaced the earth as the center of our system. Nothing has more often been called a revolution in science than that. Yet a number of scholars have noted that astronomy changed little at the time of Copernicus. The notion that we inhabit a solar system came to matter only when it was allied with mechanics, thanks to the work of Kepler and Galileo around 1610. It was they who made the so-called Copernican revolution, not Copernicus. Cohen argues this case with a certain amount of glee.

That case does not depend on what people at the time of Copernicus called a revolution, and it could not. Once the word “revolution” comes to be used in its modern sense, however, the criteria based on his four kinds of witnesses seem paramount, even though most readers would imagine that substantial change in the state of a science was what revolution is all about. Take, for example, an abrupt change in the practice and thinking of physicians that occurred only a few years after Lavoisier’s chemical revolution. Illness had long been primarily a matter of imbalance in vital fluids. Around the 1790s diseases became located in specific organs. This theoretical change was accompanied by institutional ones, such as the development of teaching hospitals and medical clinics. In his book The Birth of the Clinic, Michel Foucault presents this as a dramatic and almost instantaneous mutation. More cautious writers tell a tale of slower evolution. But change there was. Now unlike his country-man and contemporary Lavoisier, Xavier Bichat—or whomever else we choose as a prime medical revolutionary—seems not to have called this transformation in medical practice a revolution. Nor has it since become common to speak of revolution in connection with this event. On Cohen’s criteria this is no revolution in medicine, and the case does not enter his book. Yet one instinctively feels that Bichat participated in a revolution just as much as Lavoisier did. And if that instinct is mistaken, the argument will not rest on any lack of use of a word, “revolution.”

Conversely, consider the next candidate for a revolution in chemistry after Lavoisier. I mentioned that Lavoisier had the view that every acid must contain oxygen. We now know a simple counter-example—hydrochloric acid, composed of hydrogen and chlorine only. You can still buy it in old-fashioned hardware stores under the label of muriatic acid. Lavoisier well knew one could decompose muriatic acid into two gases, but he thought one gas was a compound of oxygen. In 1810 Humphry Davy showed that chlorine is an element, and hence proved that oxygen is not the essence of acidity. In its day this was hailed as revolutionary. In 1817 the new professor of chemistry at Harvard College devoted much of his inaugural lecture to praising what he called this second revolution in chemistry. There was a real change in the state of chemistry, for not only was there a new (and very important) element in the world, but also the key chemical idea of acidity had to be rethought. Yet by now there is some indifference to Davy’s discovery, and a reluctance to call it a revolution—a reluctance that I don’t share, by the way. Clearly this event does not satisfy all Cohen’s criteria, for present-day chemists and historians seem not to call Davy revolutionary. Very well: but Cohen’s analysis tells us little about why scholars now appear to resist the former description of this “second revolution in chemistry.”

I’m not intending to take issue with Cohen’s favored (and fascinating) list of revolutions. I write with the eye of a logician, to whom Cohen’s criteria are very odd indeed. Why did Cohen write this way? It is no use to say, “Logician, be off, let the historian record the facts.” Why these particular facts? Cohen is not just writing history. He is writing a curiously positivist history. The positivist seeks the bare particulars, the thoroughly untheoretical facts. Nothing is less theoretical than the brute occurrence of the word “revolution” on a number of printed pages. Cohen collected hundreds of such occurrences in his book, and doubtless has thousands more in his filing cabinets. But what, one might ask, is he trying to prove? There is an answer, but it leads me to recall a little history of the use of the idea of revolution in science by historians. This historiography is provided by Cohen himself, although I shall end with a twist, by putting Cohen into Cohen’s own story.

“Scientific revolution” as a historian’s term of art and analysis, as I earlier said, is a postwar phenomenon. It came into use almost by accident. Cambridge University decided to allow undergraduates to learn a little history of science; in due course the subject became attached to the Natural Sciences Tripos. But there were no experts to teach it. The late Sir Herbert Butterfield, who was to become Regius Professor of Modern History, but was then an up-and-coming scholar of distinction, agreed to work up some lectures for 1948. They were published the next year as The Origins of Modern Science, 1300–1800. On the first page we read of “the scientific revolution,” and the phrase is liberally sprinkled throughout the book. His text became the primer for fledgling courses all over the English-speaking world. He did not invent the phrase “scientific revolution,” but it was he who got English-speaking historians into the revolution-in-science business.

Butterfield was not a historian of science, but, as he put it, “an historian as such.” He devoured some classical works of science and the relatively little then written about them. He was much taken with new ideas stemming from the finest historian of science of the day, Alexandre Koyré, the Russian-born scholar who worked in Paris. Koyré deeply believed that truly great transitions in ideas are the result of new ways to see and think about the world. They do not come from the observation of new phenomena. Thus Galileo had long been admired as a keen observer. Koyré audaciously alleged that Galileo did not perform the experiments he described. They were thought experiments, and that sufficed. (Koyré was mistaken. Stillman Drake and others have worked through Galileo’s notebooks, reconstructed his apparatus, and shown that long sequences of numbers in the notebooks correspond to measurements that you obtain from the apparatus.) Koyré believed that at the time of Galileo the Western vision of the world was radically transformed. It was an affair of the mind, and new technologies (such as the telescope and some machines) were its mere accompaniment. Koyré did use the word “revolution” to describe this event, but also words that he took from Gaston Bachelard, such as “mutation.” It is perhaps just an accident (or rather a consequence of Butterfield’s sense of English prose) that picked “revolution” as the word that historians of science were to bandy about after Butterfield published his book.

Butterfield wrote not about revolutions but about what he called the scientific revolution, one event spanning five hundred years. He called it the most important event in the history of the West after the invention of Christianity. In the introduction to a later edition of Origins he was to say that it “changed the character of men’s habitual mental operations…and the very texture of life itself.” Such statements are remarkable from a historian who disparaged generalization and often discouraged theorizing, urging his pupils to immerse themselves in the minutiae of documents. It is worth remembering in what follows that there is something of Butterfield’s distrust of generalization in Cohen. It is no accident that when in England Cohen held a Visiting Fellowship at Peterhouse, the Cambridge college of which Butterfield was Master.

Where Koyré had written of “the scientific revolution of the seventeenth century,” Butterfield more cautiously wrote that “the 1670s must represent one of the greatest decades in the scientific revolution.” Butterfield saw the entire sequence of scientific work between 1300 and 1800 as a definitive part of modern European history. Writers more strictly concerned with science have tended to truncate this revolution. Thus by 1954 we had A. Rupert Hall’s The Scientific Revolution, 1500–1800 (Copernicus plays a major role in Hall’s work; reviewing Cohen’s book in Nature he defends “the Copernican revolution” against Cohen). By now most scholars are back with Koyré’s “the scientific revolution of the seventeenth century.” The central point is that there is exactly one scientific revolution, at least when the term “revolution” became a technical term for historians of science. A few hardy souls have accepted the invitation to conjure up a second or even a third event of comparable magnitude, but we’re still counting on one hand.

Butterfield may have suspected what was to happen next. In the 1957 introduction to a revised version of Origins he speaks of “the so-called ‘scientific revolution,”‘ an odd phrase from the man who put the term into serious circulation. What happened next was Thomas Kuhn’s widely influential book The Structure of Scientific Revolutions, published in 1962. This book is not at all about the scientific revolution, an event embracing the very texture of Western civilization. It is about hundreds of revolutions, which are supposed to occur in many disciplines, and which typically involve the research work (in the first instance) of at most a hundred or so investigators. Lavoisier’s chemical revolution counts as one, but so does Roentgen’s discovery of X-rays, the voltaic cell or battery of 1800, the first quantization of energy, and numerous developments in the history of thermodynamics.

Kuhn learned from Koyré, but he was an entirely different kind of pupil from Butterfield. He was not a historian “as such” but a former physicist. Often Butterfield and Kuhn will use almost the same words to represent what they got from Koyré. Thus in a recent essay aptly called “What Are Scientific Revolutions?”³ Kuhn gives as his first characteristic of a scientific revolution the experience of “the pieces suddenly sorting themselves out and coming together in a new way.” On the first page of Butterfield’s Origins, in a passage to which Kuhn himself drew attention, we read about placing data “in a new system of relations with one another by giving them a different framework.” Kuhn was fascinated by switches in perspective, and admired a phrase of Butterfield’s to the effect that an intellectual transformation involves putting on “a different kind of thinking-cap.” But for Kuhn new thinking caps are common in the numerous subdisciplines of several sciences.

Incidentally, in this respect he is at one with Bachelard, even though he is influenced by Bachelard, if at all, only through the mediation of Koyré. Bachelard thought that the growth of knowledge proceeds through a succession of breaks, in each of which one overcomes an “epistemological block” that requires a radical reorganization of fundamental concepts—the coupure, or cut away from the past. Evidently the projects of Kuhn and of Bachelard are entirely different from Butterfield’s. Happily Bachelard’s nomenclature itself is different from Butterfield’s; had this been true of Kuhn too we might have been spared some confusion.

Kuhn rightly called his book The Structure of Scientific Revolutions. The structure most important in the book is historical. He argues that the following pattern is regularly revealed in the sciences. There is a period of “normal science,” in which there is a widely held understanding of certain phenomena that interest scientists, and an agreed list of questions and problems to investigate. Teachers, textbooks, peer review panels, and editors know roughly speaking what matters and what counts as a solution to a problem. But gradually there arises a small number of problems that are intractable. Research goes into a state of crisis. That is resolved only by changing the way in which one conceives of the phenomena, a break that may involve junking much former work, not as false but as part of a quickly forgotten example of how not to think. This revolution solidifies and forms the new standards to be pursued by a new normal science.

Such a “structure” invites all sorts of inquiries, some of which are sociological and some of which are conceptual. For example, Kuhn’s notion has prompted an ever-increasing amount of what is sometimes called microsociology among historians and philosophers of science: how, in the relevant research groups, are such abrupt transformations effected? How does a new way of seeing the world actually win approval among a group of researchers?

But Kuhn has always shied away from that question, strongly preferring, in his purely historical writing, to describe the scientific content of changes rather than their social setting. Perhaps his real love, as speculator about the human condition, is trying to understand what it is to put on “a different kind of thinking-cap.” This is a psychological or conceptual issue. Thus in the recent paper alluded to above, he characterizes the essence of a scientific revolution in a way entirely different from Structure. Using three cases as examples—the move from the Aristotelian to the Galilean idea of motion, the change worked by the invention of the voltaic cell, or electric battery, and the first stages in the emergence of quantum mechanics—he advances three characteristic marks of what he thinks of as scientific revolution. These all have little to do with the historical process of acceptance and much to do with the way in which the world is understood, how its parts are resegmented and reclassified. We could call this a second “structure” connected with what Kuhn calls scientific revolutions, a minor theme—though always apparent—in Structure, but more recently a major research project among psychologists interested in science.

I have some doubts that Kuhn’s two “structures” (as I call them) pick out the same class of historical events. I suspect they merely overlap. No matter. Kuhn has given us perfectly good “first-deck” criteria. We can ask, Is it the case that many events in the sciences display the historical structure that Kuhn describes in his book? Quite a few historians doubt it, but at any rate that is a fair question. Similarly we can turn, as several scholars have already done, to a contemporary event that is called a revolution.

Cohen devotes a chapter to the revolution in geophysics during the last twenty-five years called plate tectonics—the theory that the hard outer layer of the earth is divided into seven major and perhaps a dozen smaller plates, each about 100 kilometers thick and resting on a softer layer. The continents are embedded in some of the plates and passively drift as the plates move about. Nothing fits Cohen’s criteria better than plate tectonics, for nothing has ever been more unambiguously hailed as a revolution in science. It can be argued that this revolution does not exhibit the structure of Kuhn’s Structure since, among other reasons, there was no marked sense of a crisis in geological sciences that the new theory resolved. I don’t think such a conclusion would refute Kuhn. Imagine, as could perfectly well have happened, that Butterfield had not made “scientific revolution” common coin for historians, and Kuhn, drawing on Koyré, had hit upon Bachelard’s word “mutation” instead. He writes The Structure of Scientific Mutations. We decide that the geophysical revolution is not a mutation. So what? Kuhn’s book would lose interest only if we concluded that nothing much has Kuhn’s structure.

But Kuhn did write about scientific revolutions. His book has been so widely read (or skimmed?) that there is no way to discuss scientific revolution without calling Kuhn to mind. It is not obvious why the book caught on so quickly. It was just a short book printed in a moribund academic series, well written, with some radical theses. In no time it was famous enough to figure in a New Yorker cartoon spoofing Kuhn’s new vogue word, “paradigm”—to which I shall briefly return. Undoubtedly one element in the book’s success was its challenge to the time-honored view of science as a paragon of rationality. Normal science was presented as routine puzzle solving, and a revolutionary change was compared more to a switch from one gestalt to another than to ratiocination.

When last I looked in the Stanford University textbook store, Kuhn’s Structure appeared to be an assigned text in six different departments. Yet it was at that time not assigned either in history or in any of the physical sciences, despite the fact that Kuhn wrote solely about the history of the physical sciences. That is the characteristic irony of Kuhn’s reception. His book has had little influence on the way in which most historians of science write history. In contrast it is the work that has been most often cited by philosophers of science in the past two decades. Kuhn once vehemently insisted in print that he was a member of the American Historical Association, not of the American Philosophical Association. Yet he is now housed with the philosophers at MIT, even if his most recent book was a finely detailed study of the early moments of the quantum theory. But no matter whether he has had much impact on the broad mass of historians, the fact is that he dominates the discussion of “scientific revolutions,” with the exception of that entirely different kind of beast, the once and only scientific revolution so called by Butterfield.

That is where Cohen comes in. His book is in part an attempt to reclaim revolution for the historian. Butterfield wrote that reflection on the scientific revolution was “in need of the intervention of the historian as such.” Cohen holds that the idea of revolution in science is, since Kuhn, in dire need of the intervention of a historian-of-science as such. There is, for him, a tradition since the seventeenth century by which people have spoken of revolutions in science. It is this tradition that should define our concept. Cohen resembles a champion of the common law and precedent, opposed to formal codes, charters, and constitutions. The idea of a revolution, like that of a right, is embodied in our past and evolves in our present, or so Cohen seems to say.

Cohen provides his own structure of revolution in science. He says that “he has found four major and clearly distinguishable and successive stages in all revolutions in science.” To the eye of the logician, this is one of the least successful parts of the book. Here are his four stages: (1) An intellectual revolution takes place when one or several scientists devise a radical solution to a problem, propose a new conceptual frame, or whatever. This “creative act,” Cohen writes, “is apt to be a private or individual experience.” (2) Then the creators write down this new thought in a diary, notebook, or even draft of a lecture. This is a revolution of commitment, which is “still private.” (3) The material is made public between colleagues at a conference, or gets into a book or journal. This is a revolution on paper. (4) Finally there is acceptance; “a sufficient number of other scientists become convinced of the theories or findings and begin to do their science in a revolutionary new way.”

I have qualms about the distinction between (1) and (2). It smacks too much of the light bulb going on over the inventor’s head, followed by writing the thought down in the log. New ideas are just not, in general, like that. But that is a minor issue. My problem is that Cohen’s stages have nothing special to do with revolution. Inventing a better mousetrap has Cohen’s four stages. First I have an idea, then make a sketch, then test a prototype, and finally the world is beating a path to my door. Any successful discovery presumably goes through Cohen’s stages. The science page of today’s paper tells me that a colleague has figured out how some weeds evolve so as to mimic rice plants, thereby evading the vigilance of peasant weeders in Asia. That is good science and beneficial science. There is nothing in the least revolutionary about it, but assuredly we shall be able to trace Cohen’s four stages here too. The same holds for a successful new rock group or an advertising campaign.

Contrast Kuhn’s historical structure in Structure. It is true that some readers have said that some events in the history of art exemplify the same structure. But only some events do so, and only some events in the history of science do so. It is significant that some cynics have urged that no events in science exemplify Kuhn’s structure. That shows that there is real content to Kuhn’s claims. He could be completely wrong. Cohen’s stages have, in my opinion, almost no content. They don’t teach us anything about what is peculiar to revolution, because they seem to be cast in such general terms that they apply to any successful discovery whatsoever.

It is for reasons such as this that Cohens does not succeed in reclaiming revolution for the historian. The historian too must provide analysis and understanding. Cohen is both gracious and accurate in describing Kuhn’s place in the history of the idea of revolution in science. But it would be disingenuous to imagine that Kuhn is not also a target of this work. Even the argument that there was no Copernican revolution in the time of Copernicus is not unconnected with the fact that before Kuhn wrote Structure he had published a book with the title, The Copernican Revolution. It is just possible that had Cohen adopted a less statesman–like approach to Kuhn’s contributions, he might have been more effective. He might have echoed a remark made by Butterfield, perhaps twenty years ago, to a young fellow of his college, “Do you think there is anything in what this chap Kuhn says?”

Cohen’s book is not for the analytic mind. Its qualities are quite other. Its pleasures lie in its rich collection of information. I’ve harped on Cohen’s writing about other people writing down the word “revolution.” But one can also learn a certain amount about the revolutions themselves, whether they were carried out by Freud, or Harvey, or Darwin, or by Tuzo Wilson and the other pioneers of plate tectonics. I also profited from the informative and often entertaining asides, which are legion.

There is, for example, a magnificent scholarly put-down that is funny in a bookish way. It arises from the fact that in the preface to the second edition of his Critique of Pure Reason, Kant speaks both of Copernicus and of a “single and sudden revolution” in mathematics and natural science. It has become standard to say that Kant wrote that he was bringing a Copernican revolution to metaphysics. He said no such thing. Cohen has a glorious list of quotations from celebrated authors who assert that Kant said what he never said: Bertrand Russell, Gilles Deleuze, John Dewey, Georg Lukács, karl Popper, and the Macropaedia of the new Encyclopaedia Britannica. Nor is it only the mighty who are caught: dozens of lesser and more scholarly fry are remorselessly reeled in. Cohen is able to cite three learned demonstrations (1937, 1959, 1963) that Kant never said what he is said to have said. We are now well placed for a little experiment in the history of ideas. Will even Cohen’s ribald scoffing have any effect? Or is a demonstrable falsehood, when well enough planted, beyond correction?

Everyone, no matter how well informed, will be able to learn some new things from this book, especially from its supplements. Here is something I ought to have known but did not, about Kuhn’s word “paradigm.” I knew of course that when Kuhn wrote in 1962, the word was chiefly a dusty term from old-fashioned grammar, defined in the dictionary as “Example, pattern; an example of a conjugation or declension showing a word in all its inflectional forms.” It also had a little currency at the time among a group of philosophers who used a particularly debased form of reasoning that they called “the paradigm case argument.”

Kuhn innocently used the word as a metaphor for what is standard about normal science, and has to be shattered or replaced by revolution. He used the word to denote both the established and admired solutions that serve as models of how to practice the science, and also for the local social structure that keeps those standards in place by teaching, rewards, and the like. The word was mysteriously launched or rather catapulted into prominence, and now seems a standard item in the vocabulary of everyone who writes about science—except Kuhn himself, who has disavowed it. What I did not know was that this was not the first time that the grammatical metaphor had been applied to science. Nor would I have thought it at all likely—who would compare scientific inquiry to declining a noun or conjugating a verb? Yet late in the eighteenth century the German physicist Georg Christoph Lichtenberg did just that.

I don’t know that I am the wiser for this bit of information (which Cohen was, as it happens, not the first to notice) but it is no vice to acquire trivia like this. Possibly it is not quite trivial.⁴ But when we hear that long ago part of science had been compared, by the metaphor of “paradigm,” to grammar, Kuhn’s metaphor comes to life. It is at present a dead metaphor. I imagine that a very few of those science writers who freely use the word even know its grammatical origins. The casual aside about Lichtenberg can make one rethink the justice of comparing normal science to grammar.

Had I regarded Cohen’s book simply as treasure trove, this would have been a glowing review, and not a carping one. But I have taken the book more seriously. It is an attempt to reclaim the idea of revolution in science for the historian. It fails because in its positivist attempt to record the bare facts, we lose a sense of why the idea should interest anyone. The work of the revolutionaries, of Darwin and Freud, is supremely important. But saying that it is called revolutionary, and was so called, neither adds to nor explains the importance.

The interest in scientific revolution arises from a theory about revolution. Butterfield made real use of his idea on “the” five-hundred-year-long scientific revolution in order to characterize the most important event of modern European history. Likewise Kuhn used his concepts of scientific revolution, so different form Butterfield’s, to challenge complacent orthodoxies about rational, progressive, and cumulative science. It is true that before Butterfield and Kuhn, and their French predecessors, there were people who called this or that event in science revolutionary. But the term was chiefly used in a vein of congratulation or self-congratulation. It had no analytic power until, after World War II, some speculative writers made it powerful. Historians who would like to reclaim the idea of revolution in science are not well advised to restore the prewar status quo. To do so would of course be to effect a revolution, in the original sense of the word: it would be a return, a ridding of usurpers and excesses. But this would be curiously self-defeating. Revolution in science would cease to be an interesting idea.