How Knowledge Breaks and Remakes Itself: Thomas Kuhn and the Secret Life of Scientific Change
For three hundred years we told ourselves a comforting story about science. Knowledge, we said, grows the way a coral reef grows…
How Knowledge Breaks and Remakes Itself: Thomas Kuhn and the Secret Life of Scientific Change
For three hundred years we told ourselves a comforting story about science. Knowledge, we said, grows the way a coral reef grows: patiently, one true fact cemented onto the last, the whole edifice rising steadily toward a complete and final picture of the world. Then in 1962 a soft-spoken physicist who had wandered into history handed us a different story, and the comfortable version never fully recovered.
Thomas Kuhn argued that science does not crawl forward in a straight line at all. It works quietly inside a shared framework he called a paradigm, ignores the cracks for as long as it can, and then, when the cracks become impossible to ignore, it convulses. An old world is abandoned, a new one is adopted, and afterward the textbooks are rewritten so smoothly that no one remembers there was ever a fight. He gave us the phrase paradigm shift, and then watched in dismay as the world used it to sell everything from software to breakfast cereal.
This book tells the secret life of that idea. It follows a reluctant revolutionary from a stunned afternoon reading Aristotle through the most cited book in the history of the social sciences, into the great quarrels it provoked with Karl Popper, Imre Lakatos, Paul Feyerabend, and Michael Polanyi, and onward to the strange war over truth that erupted at the close of the twentieth century. It explains, without a single equation, what a paradigm really is, why rival scientists can seem to live in different worlds, what science quietly forgets every time it advances, and whether any of this means that knowledge is merely a matter of fashion. It closes by watching a brand-new scientific idea fight for its life in our own moment, so that the reader can see Kuhn’s machinery turning in real time rather than safely embalmed in the past.
Keywords: Thomas Kuhn, paradigm shift, philosophy of science, scientific revolutions, history of science, how knowledge changes, incommensurability
[embed]
Contents
Chapter One — The Word That Got Away. 16
Chapter Two — A Physicist Ambushed by Aristotle. 23
Chapter Three — The Tidy Universe Before Kuhn. 30
Chapter Four — Popper and the Cult of the Crucial Test 37
Chapter Five — Ludwik Fleck and the Thought Collective. 44
Chapter Six — Koyré, Butterfield, and the History That Bites Back. 51
Chapter Seven — Rehearsal: The Copernican Revolution. 58
Chapter Eight — What on Earth Is a Paradigm?. 66
Chapter Nine — The Quiet Labor of Normal Science. 73
Chapter Ten — When the Anomaly Will Not Go Away. 80
Chapter Eleven — Crisis, and the Loosening of the Rules. 87
Chapter Twelve — The Gestalt Switch. 94
Chapter Thirteen — Do Scientists Live in Different Worlds?. 101
Chapter Fourteen — Incommensurability, or Talking Past Each Other 109
Chapter Fifteen — How Textbooks Hide the Revolutions. 117
Chapter Sixteen — Choosing Theories Without a Rulebook. 125
Chapter Seventeen — Progress That Is Not Going Anywhere. 134
Chapter Eighteen — The Duel: Kuhn Against Popper 142
Chapter Nineteen — Lakatos and the Defense of Reason. 151
Chapter Twenty — Feyerabend and the Glorious Anarchy. 159
Chapter Twenty-One — Polanyi and the Knowledge We Cannot Speak. 167
Chapter Twenty-Two — The Relativist He Refused to Be. 175
Chapter Twenty-Three — When Sociologists Took the Wheel 182
Chapter Twenty-Four — The Science Wars and the Sokal Bomb. 190
Chapter Twenty-Five — The Historian at the Bench: Black-Body Theory. 198
Chapter Twenty-Six — Dropping the Word, Keeping the Idea. 206
Chapter Twenty-Seven — Revolutions as Speciation. 214
Chapter Twenty-Eight — Kuhn-Loss, or What Science Forgets. 222
Chapter Twenty-Nine — The Empire of the Word “Paradigm”. 230
Chapter Thirty — Paradigms in the Age of Big Science and Machines 239
Chapter Thirty-One — A Living Anomaly: A Constant Drawn from the Nucleus. 247
Chapter Thirty-Two — How a New Paradigm Fights for Its Life. 255
Chapter Thirty-Three — After Kuhn: A Maturer Picture of Change. 262
The Copernican Revolution. 278
Phlogiston and the Chemical Revolution. 279
The Birth of Quantum Theory. 280
Neptune and the Phantom Vulcan. 281
Röntgen and the Discovery of X-rays. 282
Wegener and Continental Drift 283
Semmelweis and the Doctors Who Would Not Wash. 284
The Stomach Ulcer Revolution. 285
Darwin and the Tree of Life. 286
Newton’s Action at a Distance. 287
The Ether and the Michelson-Morley Experiment 288
The Anomalous Playing Cards. 290
The Sokal Hoax and the Science Wars. 291
The Reproducibility Crisis. 292
The Detection of Gravitational Waves. 293
The Keynesian Revolution in Economics. 294
The Double Helix: Myth and Reality. 295
The Cosmological Constant Problem.. 297
The Thinkers Who Shaped Kuhn. 325
The Great Debates: Critics and Rivals. 326
Understanding Kuhn: Secondary Works. 327
The Sociologists and the Science Wars. 328
The Wider Philosophy of Science. 330
Preface
There is a particular kind of sentence that has escaped from the laboratory and now roams freely through the wild, breeding indiscriminately and answering to no one. You have met it. A breathless executive announces that his new mobile application represents a paradigm shift in the way we order sandwiches. A magazine declares a paradigm shift in eyebrow grooming. Somewhere, at this very moment, a consultant is being paid an indecent sum to stand in front of a slide that says PARADIGM SHIFT in letters large enough to be read from low orbit, and not one person in the room could tell you where the phrase came from or what it was originally supposed to mean.
It came from a quiet man named Thomas Kuhn, and he came to regret it almost immediately. Kuhn was a physicist by training and a historian by accident, and in 1962 he published a slim book with the forbidding title The Structure of Scientific Revolutions. He expected it to be read, if at all, by a few dozen specialists in the dusty corner of the academic world where the history of science was then kept. Instead it became one of the most cited books ever written in any field, a permanent fixture on university reading lists, and the unwitting parent of a phrase that would go on to be abused in boardrooms for the next sixty years. Kuhn spent much of the rest of his life trying, with the patient exasperation of a man whose dog has run off with the neighborhood, to explain that he had not meant quite what everyone thought he meant.
This book is about what he did mean, and about why it matters far more than the slogan suggests. Because underneath the buzzword lies one of the genuinely unsettling ideas of the twentieth century: the suggestion that science, the most reliable knowledge-making machine our species has ever built, does not actually work the way we were all taught it works. We were taught that science is a ladder. Each generation climbs a rung higher than the last, adding new truths to the pile, correcting the small mistakes of its predecessors, marching with cheerful inevitability toward a complete and final description of reality. It is a lovely story. It is also, Kuhn argued, mostly false. Science does not climb a ladder. It lurches. It spends long stretches doing meticulous, almost clerical work inside a settled framework, and then it undergoes something closer to a religious conversion or a political revolution, in which one entire way of seeing the world is overthrown and replaced by another that is, in some deep and disturbing sense, not even speaking the same language.
If that is true, a great many comfortable assumptions begin to wobble. We like to think that when scientists disagree, the disagreement can always be settled by an appeal to the facts, to the evidence, to a decisive experiment that any reasonable person would accept. Kuhn suggested that during the deepest disagreements, the rival camps cannot even agree on what counts as a fact, because what counts as a fact is itself determined by the framework you already hold. We like to think that scientific progress means getting closer and closer to the truth. Kuhn suggested, in his most provocative mood, that we might do better to drop the destination altogether and think of science the way we think of evolution: as a process that moves away from where it has been without necessarily moving toward anything in particular. These are not small claims. They reach all the way down to the question of whether there is such a thing as objective knowledge at all, and people have been arguing about them, sometimes with surprising venom, ever since.
I have written this book for the general reader, by which I mean the intelligent, curious person who is not a professional philosopher and has no intention of becoming one, but who would nonetheless like to understand one of the most important intellectual revolutions of modern times without being asked to wade through a swamp of jargon. There is no mathematics in these pages. There is no specialized vocabulary that I do not stop to explain in plain words. What there is, I hope, is a story, because Kuhn’s idea is best understood not as a dry thesis but as a drama, complete with a reluctant hero, a cast of brilliant antagonists, a few genuine villains, and a plot that is still unfolding in the science of our own day.
The drama has a definite shape, and so does this book. We begin with the man himself, and with the strange afternoon in 1947 when, while trying to understand why Aristotle had been so spectacularly wrong about motion, Kuhn experienced a flash of insight that would organize the rest of his life. We will sketch the comfortable picture of science that he inherited, the tidy world of the logical positivists and of Karl Popper, so that we can feel the full force of what he disturbed. We will meet his forgotten predecessors, men like Ludwik Fleck who had glimpsed pieces of the same idea decades earlier and been ignored. Then we will walk slowly through the machinery of his theory itself: the paradigm, that slippery word with at least twenty-one meanings; normal science, the patient puzzle-solving that occupies most working scientists most of the time; the anomaly that refuses to go away; the crisis that follows; and finally the revolution, the gestalt switch in which the world itself seems to change shape.
From there we will follow the idea out into the world, where it promptly started fights. We will referee the famous duel between Kuhn and Popper, a clash of two utterly different temperaments over the soul of science. We will sit in on the responses of Imre Lakatos, who tried to rescue reason from what he saw as Kuhn’s slide into mob psychology, and Paul Feyerabend, who cheerfully went further than Kuhn ever dared and declared that in science, anything goes. We will give Michael Polanyi his due, the chemist-philosopher who had said much of it first and watched Kuhn get the credit. We will examine the charge that haunted Kuhn for the rest of his career, the accusation that he had made science irrational, a mere matter of taste and persuasion, and we will hear his stubborn lifelong insistence that he was nothing of the kind. Freeman Dyson once recorded Kuhn snapping, I am not a Kuhnian, and we will try to understand the frustration behind that strange little protest.
Then the stakes rise. In the closing decades of the twentieth century, Kuhn’s ideas were taken up by a new generation of sociologists and literary scholars who pushed them in directions he found horrifying, and the result was the so-called Science Wars, a genuinely bizarre episode in which physicists and humanists hurled accusations of fraud and stupidity at one another, culminating in a famous hoax that detonated in the pages of a respectable journal. We will pick through the wreckage and try to say what, if anything, was actually settled.
And we will not leave Kuhn frozen as the author of a single famous book, because he was not. We will follow him into his difficult later years, when he quietly abandoned the very word paradigm that had made him famous, turned to the deep puzzles of language and meaning, and began comparing scientific revolutions not to political coups but to the branching of species in biological evolution. This late Kuhn is far less famous than the early one, and far more interesting than his reputation suggests.
Finally, because an idea about how knowledge changes is best understood when you can watch it actually changing knowledge, the book ends in the present tense. I will take a single concrete example of a young scientific idea trying to win acceptance in our own moment, walk it through every stage of Kuhn’s machinery, and let the reader see for themselves whether the old historian got the mechanism right. The example happens to come from my own corner of the scientific world, which gives me the advantage of knowing it from the inside, and the disadvantage of being unable to pretend I am a neutral bystander. I will try to be honest about both.
A word of warning, and a word of encouragement. The warning is that Kuhn is genuinely slippery. He was a careful, anxious, often contradictory thinker who revised his views repeatedly and was the first to admit that key terms in his masterpiece were used in incompatible ways. Anyone who tells you that Kuhn proved science is irrational, or that Kuhn proved all truth is relative, or that Kuhn proved any particular comfortable or uncomfortable conclusion, has almost certainly oversimplified him, and we will spend a fair amount of effort resisting that temptation. The encouragement is that the effort is worth it. To understand Kuhn is to acquire a new set of eyes. Once you have grasped how paradigms work, you begin to see them everywhere, not only in physics and chemistry but in medicine, in economics, in your own profession, in the quiet assumptions that govern how you and everyone around you decides what is true. You begin to notice the anomalies that polite society agrees to overlook. You begin to recognize the early tremors of a crisis. And you acquire a certain useful humility about your own most confident convictions, a suspicion that they too rest on a framework that one day, perhaps sooner than you think, the world will quietly set aside.
It is worth pausing, before we begin, on why any of this should matter to a person who is not a scientist and never intends to become one. The answer is that we live inside the results of science to a degree no previous generation ever did, and we are constantly being asked to trust it, to defer to it, to reorganize our lives around its pronouncements. We are told what to eat and what to fear, which medicines to swallow and which habits to abandon, how old the universe is and how it will end. Most of us accept these claims, sensibly enough, because the alternative is to verify everything ourselves, which is impossible. But trust without understanding is a fragile thing. When the experts change their minds, as they regularly do, the trusting public feels betrayed, mutters that scientists cannot make up their minds, and grows a little more receptive to the cranks and charlatans who promise certainty. Kuhn offers something better than blind trust and something better than cynical dismissal. He offers an understanding of how a healthy science actually behaves, why changing its mind is a sign of strength rather than weakness, and how to tell the difference between a science working through a genuine crisis and a mere fashion dressed up in a lab coat. That is not a luxury for specialists. In an age drowning in confident assertions, it is close to a survival skill.
I should also say plainly what this book is not. It is not a biography, though the man’s life runs through it, because the idea matters more than the man and Kuhn himself would have insisted on the point. It is not a textbook, and it does not pretend to settle the scholarly disputes that still rage among professional Kuhn specialists, who can argue for an afternoon about what he meant by a single paragraph. It is not a debunking, and it is not a celebration; Kuhn was wrong about some things and maddeningly vague about others, and I see no reason to hide either fact from a grown-up reader. What this book tries to be is a clear, honest, and frankly entertaining tour through one of the great ideas of modern thought, conducted by a guide who finds the idea endlessly fascinating and assumes you will too. If by the end you find yourself quietly auditing the paradigms that govern your own field, your own politics, your own certainties, then the book will have done its work, and Kuhn, wherever the dead historians of science go, can rest a little easier knowing that at least one more reader understood him before reaching for the phrase he came to dread.
That is the secret life of scientific change, and it is the subject of this book. We will start, as Kuhn himself did, by accident, with a young physicist, a borrowed office, and a very old book by Aristotle that should have been wrong in the way he expected it to be wrong, and instead turned out to be wrong in a way that changed everything.
Chapter One — The Word That Got Away
In the grammar books of the ancient world, a paradigm was a humble and well-behaved thing. It was a pattern, a model to be copied, the conjugation table a schoolboy memorized to learn his Latin verbs. Amo, amas, amat: there is your paradigm, the exemplary case from which every other verb of its type could be derived by faithful imitation. For roughly two thousand years the word kept this modest profession. It belonged to teachers of language, who used it to mean a standard example, and it bothered no one. The eighteenth-century German aphorist Georg Christoph Lichtenberg gave it a slightly wider airing, and a few others borrowed it now and then, but on the whole it lived a quiet, dignified, grammatical life. Then a historian of science picked it up, and within a generation it had become one of the most overworked words in the English language, summoned to dignify everything from corporate reorganizations to changes in hairstyle, until it now means so many things that it means almost nothing at all.
The historian was Thomas Kuhn, and the story of how his word got away from him is worth telling first, before we get anywhere near his actual argument, because it illustrates something Kuhn himself spent his life worrying about: the way a term can travel far from its birthplace and arrive somewhere its inventor never intended. Kuhn took the old grammatical word and asked it to do enormous new work. For him a paradigm was, at its core, a concrete achievement so impressive that it became a model for an entire scientific community, the way a single brilliantly solved problem can show a whole field how its problems ought to be solved. Newton’s account of the planets, Lavoisier’s chemistry, Darwin’s account of descent: these were paradigms in Kuhn’s original sense, exemplary triumphs that other scientists studied, imitated, and extended. The trouble was that Kuhn could not keep the word still. Almost as soon as he had defined it this way, he began using it to mean something much larger and vaguer: the entire constellation of beliefs, values, instruments, and assumptions shared by the members of a scientific community. A paradigm was a single shining example. A paradigm was also an entire worldview. It was both at once, and Kuhn slid between the two meanings without always noticing.
He was not the only one who noticed the slipperiness. A sharp-eyed linguist named Margaret Masterman, reading The Structure of Scientific Revolutions with the patience of a prosecutor, went through the book and counted the ways Kuhn had used the word. She found at least twenty-one distinct senses. Twenty-one. A word that had spent two millennia meaning roughly one thing was now, in a single short book, meaning more than twenty. Kuhn, to his credit, did not bluster when this was pointed out. He more or less agreed. In a long essay added to the second edition, and in later writings, he tried to clean up the mess by splitting the word in two. There was the broad sense, which he proposed to rename the disciplinary matrix, the whole shared apparatus of a scientific community. And there was the narrow, original sense, the exemplary problem-solution, which he proposed to call the exemplar. The rescue operation was sensible, even elegant. It came far too late. By then the word was loose in the world, breeding, mutating, and entirely beyond the reach of its creator.
What the world seized on, of course, was not the careful distinction between disciplinary matrices and exemplars. The world seized on the phrase paradigm shift, two words that capture a feeling of dramatic, wholesale transformation, of the rug being pulled out and a new rug rolled into place. It is a wonderfully satisfying phrase to say. It sounds momentous. It flatters whoever uses it, implying that they are present at a great turning point in history rather than, say, attending a meeting about quarterly logistics. And so it spread, first through the universities, where scholars in fields that had nothing to do with physics began announcing paradigm shifts in their own disciplines, and then out into the wider culture, where it was adopted by advertisers, motivational speakers, and the authors of business books, that vast literature of confident nonsense in which every modest novelty is a revolution and every revolution is a paradigm shift.
Kuhn watched this happen with the particular dismay of a parent whose child has fallen in with a bad crowd. He had wanted to describe something precise and difficult about the history of physics and chemistry. He had ended up coining a marketing slogan. There is a melancholy comedy in this, and Kuhn felt it. He had, after all, written a whole book about how meaning shifts, how words get reattached to new things, how communities that use the same vocabulary can mean wildly different things by it. And now his own central term had become a living demonstration of his thesis, drifting from community to community, picking up new meanings in each, until the word as used by a software salesman and the word as used by Kuhn shared little more than their spelling.
It is worth pausing to see just how much Kuhn eventually tried to pack into the broad sense of the word, because the richness of what he meant is precisely what the slogan throws away. When he sat down, in that long clarifying essay, to spell out the components of a disciplinary matrix, he found at least four distinct ingredients tangled together. There were the symbolic generalizations, the compact formulas and laws that the members of a field write down and deploy without argument. There were the shared models and analogies, the mental pictures a community agrees to take seriously, such as the picture of a gas as a swarm of tiny billiard balls in motion. There were the values, the largely unspoken agreements about what makes a theory good, the preference for accuracy, for simplicity, for predictions that reach beyond the cases a theory was built to explain. And there were the exemplars, the concrete solved problems that students absorb in their training and carry with them as templates for everything they will later attempt. A paradigm, in the full Kuhnian sense, was all of this at once, an entire culture of knowledge held in common by a community of practitioners. To compress that into a phrase meaning roughly any big change is a little like compressing an entire civilization into the word lunch.
And here lies the deepest irony of the whole affair, the one Kuhn must have savored even through his irritation. His central argument was that scientific communities sharing the same vocabulary can come to mean radically different things by their shared words, and that this drift of meaning is precisely what makes communication across a revolution so treacherous. The fate of his own word is a perfect miniature of that thesis. The grammarians meant one thing by paradigm. Kuhn meant something far larger and more various. The sociologists who borrowed it from Kuhn meant something different again, and the management consultants who borrowed it from the sociologists meant almost nothing at all, just a vague aroma of importance. The word traveled from community to community, and in each one it quietly became a different word wearing the same costume. Kuhn could hardly have designed a better demonstration of his own ideas if he had tried, and there is reason to think that some wry part of him appreciated the joke even as the rest of him winced.
The damage to public understanding, however, was real, and it is worth naming. When a phrase like paradigm shift becomes an all-purpose idiom for any change whatever, it quietly teaches people a false picture of science, the very picture Kuhn had set out to correct. It teaches them that science lurches from revolution to revolution, that today’s truth is tomorrow’s discarded fashion, that nothing in science is ever really settled, and therefore that one expert opinion is about as good as another. This is a caricature, and a dangerous one, because the genuine Kuhn believed nothing of the kind. He thought revolutions were rare and that the long stretches between them, the patient and cumulative work he called normal science, were where most of the real labor and most of the real progress took place. The slogan amputates exactly the part of his theory that gives science its stability, keeps only the part that makes it sound like a fashion show, and then hands the result to anyone looking for a reason to distrust expertise. Recovering the real Kuhn from the slogan is therefore not merely a matter of scholarly tidiness. It is a small act of intellectual repair, and it is one of the quieter purposes of this book.
There is a lesson in this small disaster, and it is the right note on which to begin. The lesson is that we should be careful, throughout this book, to distinguish between Kuhn the cultural phenomenon and Kuhn the actual thinker. The cultural phenomenon is enormous and mostly garbage: a vague licence to call any change a revolution and any revolution a transcendence of the old order. The actual thinker is subtler, stranger, more cautious, and far more interesting than the slogan suggests. He did not believe that every change of mind is a paradigm shift; he believed that genuine paradigm shifts are rare, traumatic, and separated by long stretches of patient, unglamorous work. He did not believe that anything goes; that was someone else’s slogan, as we shall see. He spent the second half of his career trying to recover his own ideas from the people who admired them, which is a strange and rather sad way to spend a career, and it should make us read him with extra care. When a thinker is famous chiefly for a phrase he came to hate, the honest reader’s first job is to set the phrase aside and ask what the man was really trying to say. That is what we will spend the rest of this book doing. But first we must go back to the moment it all began, which was not a moment of triumph or ambition but a moment of sheer confusion, on an ordinary afternoon, in a borrowed office, over a very old and supposedly very wrong book.
Chapter Two — A Physicist Ambushed by Aristotle
The summer of 1947 found a young Harvard physicist in a state of mild professional irritation. Thomas Kuhn had been asked to interrupt his real work, which was physics, in order to prepare a case study in the history of science for a course aimed at non-scientists. The case study was to be about the development of mechanics, the science of motion, and so Kuhn, dutifully, sat down to find out what people had believed about motion before Galileo and Newton came along and got it right. This meant reading Aristotle, whose physics had dominated European thought for the better part of two thousand years. Kuhn opened the text expecting to find the usual thing one finds when a modern scientist reads an ancient one: a brilliant mind, hampered by the absence of good instruments and good mathematics, groping toward truths it could not quite reach, getting some things half right and many things wrong. He expected, in short, to find a primitive and inferior version of the physics he already knew.
What he found instead was a catastrophe, and the catastrophe was the most important thing that ever happened to him. Aristotle, read as a draft of Newton, was not merely incomplete. He was hopeless. His statements about motion were not approximately right; they were not even wrong in interesting ways; they appeared to be the ravings of a man who had never bothered to look at the world around him. This was impossible to believe. Aristotle was, by any measure, one of the most powerful intellects in human history, a man whose work on logic, biology, ethics, and politics had earned the awe of every generation that followed. How could the same mind that produced those masterpieces have written such transparent gibberish about something as basic as a falling stone? Either Aristotle had suffered some bizarre and selective stupidity that afflicted him only when he turned to physics, or, Kuhn slowly realized, the fault lay not with Aristotle but with the reader.
Kuhn sat at his desk, staring out the window, the text open before him, and then it happened. He would describe the experience for the rest of his life as a sudden change, a moment when the fragments of Aristotle’s physics rearranged themselves before his eyes and fell, all at once, into a new pattern. Suddenly Aristotle made sense, not as a failed Newton, but as a coherent thinker who had been asking entirely different questions about an entirely different world. Where Newton asked about quantity of motion and changes in it, Aristotle was concerned with qualities, with the natures of things, with where each kind of substance belonged and how it sought its proper place. A stone fell because earth seeks the center of the cosmos; smoke rose because fire seeks the heavens; this was not a clumsy guess at gravitation but a different conceptual scheme in which the word that we translate as motion meant something far broader, encompassing growth, change, and the ripening of a thing toward what it was meant to become. Read on its own terms, inside its own framework, Aristotle’s physics was not stupid at all. It was internally consistent, deeply reasoned, and a perfectly sensible way of organizing experience, given its starting assumptions.
It is worth dwelling on the specific examples that had baffled Kuhn, because the abstract point becomes vivid only when you see the particular sentences that seemed insane. Aristotle had written that heavier objects fall faster than light ones, in proportion to their weight, a claim that any schoolchild now knows Galileo demolished by, supposedly, dropping weights from a tower. Read as physics in our sense, it is simply false. But Aristotle was not describing objects falling in a vacuum, which he considered an impossibility, a notion as incoherent as a square circle. He was describing objects moving through a resisting medium, through air or water, where heavier things genuinely do tend to descend faster, and where the speed really does depend on the relationship between the moving thing and the stuff it moves through. Within his world, a world that contained no vacuum and treated every motion as the resolution of a tension between a body and its surroundings, the claim was not a blunder but a reasonable generalization from ordinary experience. The same went for his insistence that a projectile keeps moving only because the air rushes around behind it and pushes it along, a claim that sounds ridiculous until you remember that Aristotle had no concept of inertia, no notion that a thing in motion simply continues in motion unless something stops it. Lacking that concept, he needed something to keep the arrow flying, and the air was the only candidate available. Each apparent absurdity dissolved the moment you stopped reading it as a failed attempt at Newton and started reading it as a successful attempt at Aristotle.
Kuhn came to call this kind of experience, with a certain affection, the Aristotle experience, and he believed it was the single most important thing a historian of science could undergo. It was not unlike the moment when a foreign language stops being a string of opaque sounds and suddenly resolves into meaning, or the moment when a puzzle-picture flips and the duck becomes a rabbit. Before the shift, the old text is gibberish; after it, the same words are luminous, and you cannot for the life of you recover how they ever looked like nonsense. This irreversibility fascinated Kuhn. It suggested that understanding a past science was not a matter of degree, of getting gradually closer, but a matter of a sudden reorganization, an all-or-nothing click into a new gestalt. And it suggested, ominously, that the same might be true of the scientists who actually lived through a change of worldview, that they too might experience not a gradual persuasion but a sudden conversion, after which the old way of seeing became literally unrecoverable.
From this single insight Kuhn derived a whole method of reading, which he practiced for the rest of his career and urged on every historian who would listen. When you encounter a passage in an old scientific text that seems plainly absurd, he advised, do not congratulate yourself on your superior knowledge and move on. Stop. Treat the absurdity as a signal that you have misunderstood, that you are mistranslating some key term, that the author meant by a word something other than what you mean by it. Search for a reading on which the apparent nonsense becomes sensible, and when you find it, you will usually find that you have also recovered the meanings of a whole cluster of related terms, and that a large stretch of previously baffling text has come clear all at once. This principle of charity toward the dead, this assumption that great minds of the past were not idiots but inhabitants of different conceptual worlds, became the cornerstone of a more humane and more accurate history of science. It also planted the seed of a far more radical thought, the thought that would eventually flower into the doctrine of incommensurability and bring down upon Kuhn the accusation that he had made science irrational.
The shock of that afternoon was twofold, and both halves would shape everything Kuhn later wrote. The first half was a lesson about reading the past. You cannot understand an old scientific theory by measuring it against the present one and tallying up where it falls short, because to do so is to misread it from the very first sentence, to mistranslate its key terms, to ask it questions it never meant to answer. To understand a past science you must perform an act of imaginative reconstruction, learning its language from the inside, recovering the questions it was actually trying to answer, until the thing that looked like nonsense suddenly clicks into a sensible whole. This is hard, humbling work, and it is the opposite of the smug backward glance that treats every predecessor as a stepping-stone to ourselves. Kuhn had stumbled into the historian’s deepest discipline almost by accident, and he never got over it.
The second half of the shock was more disturbing, and it took longer to bloom. If Aristotle’s physics was a coherent framework, and Newton’s physics was a coherent framework, and the two could not be combined or measured against one another by any neutral yardstick, then what exactly happened when the world moved from one to the other? It was not simply that Newton knew more facts than Aristotle. It was that the two men inhabited different conceptual worlds, in which even the most basic words, motion, force, place, void, carried different meanings and divided up experience along different lines. The transition from one to the other was not a matter of adding new bricks to an existing wall. It was a matter of tearing down the wall and building a different structure on the same ground. And if that was how science had moved from Aristotle to Newton, perhaps it was how science always moved, in convulsions rather than increments, in conversions rather than accumulations.
This is the seed of everything. The mature theory of paradigms, normal science, crisis, and revolution all grew from that single disorienting afternoon, from the experience of watching a dead and discredited physics come suddenly alive and make sense. Kuhn had set out to write a tidy little case study about the rise of mechanics, and had instead caught a glimpse of something that would consume the next four decades of his life and unsettle the philosophy of science for generations. He abandoned physics. He became a historian. And he set himself the task of explaining, to himself and eventually to the world, what it really means for one scientific worldview to give way to another. The journey from that office in 1947 to the publication of The Structure of Scientific Revolutions in 1962 would take fifteen years, and along the way Kuhn would have to confront the prevailing picture of science he had been raised on, the confident and orderly picture that his Aristotle experience had quietly demolished. To feel the force of what he did next, we need to understand that picture, the tidy universe of science as it was understood before Kuhn came along to complicate it.
Chapter Three — The Tidy Universe Before Kuhn
Every revolution needs an ancien régime to overthrow, and the philosophy of science had one ready and waiting in the middle of the twentieth century. It was a confident, well-organized, and in many ways admirable picture of how science works, and for a few decades it enjoyed something close to the status of orthodoxy among the people who thought professionally about such things. To understand why Kuhn landed like a bomb, you have to first appreciate the building he landed on, which was so neat, so reasonable, and so reassuring that disturbing it felt almost like vandalism.
The picture had grown largely out of the work of a group of philosophers and scientists who had gathered in Vienna in the 1920s and 1930s, a circle so closely associated with that city that they are still known as the Vienna Circle. They were brilliant, optimistic, and allergic to nonsense. They had watched philosophy spend centuries spinning elaborate theories about God, the soul, the Absolute, and the ultimate nature of reality, theories that could never be tested, never be settled, and never be cleared away, and they had concluded that most of this was an enormous waste of human intelligence. They wanted to put thought on a sounder footing, and they found their model in science. Science, they believed, was the one human enterprise that genuinely made progress, that actually settled its questions instead of arguing them in circles forever, and they set out to understand its secret and, if possible, to bottle it.
The secret, as they saw it, lay in a strict discipline about meaning. A statement was worth taking seriously, they argued, only if it could be tied down to observation, only if there was some possible experience that would confirm or deny it. Talk that floated free of all possible observation, the grand metaphysical pronouncements of traditional philosophy, was not false so much as empty, a kind of sophisticated noise. Real knowledge had to be anchored in what could be seen, measured, and recorded. And science was the great machine for turning such observations into reliable general truths. You gathered the facts, carefully and without prejudice. You found the patterns in them. You framed laws and theories that captured those patterns. You tested your theories against fresh observations. And in this way, observation by observation and theory by theory, the edifice of scientific knowledge rose steadily upward, each generation building soundly on the verified results of the last.
This is the picture most of us absorbed in school, often without anyone stating it outright. Science as the patient accumulation of established facts. Science as the realm where personal opinion, cultural prejudice, and human messiness are scrubbed away by the rigor of method, leaving only what the evidence compels any rational mind to accept. Science as a ladder climbing toward truth. It is a noble picture, and it has a great deal going for it. It explains why science commands respect, why its results travel across cultures and centuries, why a bridge built with good engineering stands up whether its builders were Christian or Muslim, capitalist or communist. There is something in science that does seem to transcend the local quarrels of human beings, and the tidy picture was an honest attempt to say what that something is.
The Vienna Circle’s great instrument was a principle they wielded like a scalpel, the principle of verification. The meaning of a statement, they proposed, simply is the method by which you would verify it; a statement with no possible method of verification has no meaning at all. Armed with this principle, they set out to clear the field of centuries of accumulated metaphysical undergrowth. Is the Absolute beyond time and space? The question cannot be verified by any conceivable observation, and so, on their view, it is not a deep question at all but a pseudo-question, a grammatically respectable arrangement of words that fails to say anything. This was bracing, liberating, and more than a little arrogant, and it had the great virtue of taking science as the gold standard of meaningful talk. Whatever resembled science was sound; whatever did not was suspect. The trouble, which took some years to surface, was that the verification principle had an awkward habit of cutting its own throat. By what observation could you verify the verification principle itself? None, apparently, which meant that the principle, by its own standard, was meaningless. The orthodoxy was built on a foundation that could not satisfy its own demands, and this was the first small crack in the tidy universe, a crack its own adherents were honest enough to worry about.
There was a further subtlety in the orthodox picture that mattered enormously for what Kuhn would later attack, and it concerned a distinction between two very different questions one can ask about any scientific idea. There is the question of where the idea came from, the messy human story of hunches, dreams, lucky accidents, and flashes of insight that lead a scientist to a hypothesis in the first place. And there is the separate question of whether the idea is any good, whether the evidence supports it, a question to be settled by cool public testing regardless of how the idea was born. The philosophers called these the context of discovery and the context of justification, and they insisted, reasonably enough, that the two be kept rigidly apart. It does not matter that the chemist dreamed of a snake biting its tail; what matters is whether the structure he proposed survives the evidence. Discovery is psychology and biography, irrational and unanalyzable; justification is logic and method, the proper business of the philosopher of science. By drawing this line, the orthodoxy quarantined all the human messiness of real scientific life into the context of discovery, where it could be safely ignored, and reserved the context of justification as a clean, rule-governed space where reason reigned. Kuhn’s eventual heresy was, in large part, to erase this line, to insist that the messy human factors the orthodoxy had banished to the context of discovery were in fact present, and decisive, in the supposedly pure context of justification as well.
Into this scene came a complication, in the formidable person of Karl Popper, whom we will meet properly in his own chapter, but who must be introduced here because he was both part of the orthodoxy and its first great internal critic. Popper noticed a flaw in the simple accumulation story. No matter how many times you observe the sun rising, he pointed out, you can never prove with certainty that it will rise tomorrow; no matter how many white swans you count, a single black one can shatter your confident generalization that all swans are white. You cannot reach a universal law by piling up particular observations, because the next observation might always overturn it. Popper’s ingenious solution was to flip the whole thing around. Science, he argued, does not advance by proving theories true, which is impossible, but by proving them false, which is achievable. A good scientific theory is one that sticks its neck out, that forbids certain things from happening, that makes risky predictions which could fail. Scientists should try their hardest to refute their own best ideas, and a theory earns its keep by surviving such assaults. What separates real science from pretenders like astrology is precisely this willingness to be proven wrong, this exposure to falsification.
Popper’s falsificationism was a genuine improvement, and it remains seductive; many working scientists, asked to describe their method, will still recite a version of it. But notice that Popper, for all his cleverness, left the deeper architecture of the tidy universe standing. He still saw science as a fundamentally rational enterprise governed by clear rules, in which evidence speaks plainly, in which a decisive experiment can settle a dispute, in which the relationship between theory and fact is clean and well-behaved. Science was still a contest conducted under agreed rules, with the evidence serving as impartial referee. The drama Popper described was sharper and more heroic than the placid accumulation of the Vienna Circle, but it was still a drama in which reason and evidence held the commanding heights, and in which the history of science was, broadly, the history of progress toward truth by the steady elimination of error.
This, then, was the tidy universe that prevailed when Kuhn sat down to write: a world of clear methods, impartial evidence, decisive tests, and cumulative progress, presided over by the cool authority of reason. It was a philosophers’ picture, constructed by thinking hard about how science ought to work. And it had one striking feature that would prove to be its undoing. It had been built almost entirely without looking very carefully at how science had actually worked, at the messy, contingent, human record of what scientists had really done across the centuries. The philosophers had reasoned out an ideal. They had not done the history. Kuhn had done the history, or at least had begun to, and the history he found did not match the picture at all. Where the tidy universe promised steady accumulation, the historical record showed convulsions. Where it promised impartial evidence settling disputes, the record showed evidence interpreted differently by rival camps who could not agree on what it meant. Where it promised cumulative progress toward truth, the record showed old frameworks abandoned wholesale, their hard-won achievements sometimes simply forgotten. The gap between the philosophers’ ideal and the historians’ record was the opening Kuhn drove through, and to understand how he did it, we must now examine the most powerful version of the orthodoxy he was about to challenge, the falsificationist vision of Karl Popper, and the strange and revealing duel that resulted when these two great rivals finally met.
Chapter Four — Popper and the Cult of the Crucial Test
On the twenty-ninth of May, 1919, a small party of British astronomers stood on a humid island off the coast of West Africa, watching the moon slide across the face of the sun. They had not come for the spectacle. They had come to weigh a theory. Albert Einstein had predicted that the sun’s gravity would bend the path of starlight passing near it by a precise and tiny amount, twice the amount that Newton’s older theory allowed. During the few minutes of total eclipse, when the sun’s glare was blotted out, the stars near its edge would become visible, and their apparent positions could be measured and compared against where they sat when the sun was elsewhere in the sky. If the stars appeared shifted by Einstein’s amount, his strange new picture of gravity as the bending of space and time would stand. If they did not, it would fall. The measurements came in. The stars had moved. Einstein was right, Newton was dethroned, and the young philosopher Karl Popper, reading about it in Vienna, found in that eclipse the template for everything he would come to believe about science.
What thrilled Popper was not merely that Einstein had won, but the manner of the victory. Einstein had made a bold, risky prediction, one that could easily have been wrong, one that staked his entire theory on a single number that nature was free to refuse. He had, in effect, told the world exactly what observation would destroy his theory, and then invited the world to look. This, Popper decided, was the very signature of genuine science: the willingness to be proven wrong, the courage to expose a theory to a test it might not survive. He contrasted Einstein’s daring with what he saw all around him in the intellectual fashions of the day, the theories of Freud and of Marx and of Adler, which seemed to him able to explain absolutely anything that happened and therefore to forbid nothing. A man pushes a child into a river; the Freudian explains it by repression. A man jumps in to save a child; the Freudian explains it by sublimation. Whatever happens, the theory is confirmed, and a theory that cannot conceivably be refuted, Popper concluded, is not a strength but a fatal weakness. It tells you nothing, because it rules nothing out.
From this contrast Popper built his entire philosophy, and its central idea can be stated in a sentence: what makes a theory scientific is not that it can be proven true, but that it can be proven false. He called this the criterion of demarcation, the line separating real science from its many imitators, and falsifiability was the line. A scientific theory must stick its neck out. It must forbid certain observations, must declare in advance that if such-and-such is seen, the theory is dead. Astrology, by this measure, is not science, because its predictions are so elastic that no outcome can ever embarrass them. Einstein’s relativity is science, gloriously so, because it forbade a great deal and survived the attempt to catch it out. The scientist’s proper attitude, Popper insisted, is not the loving defense of one’s own ideas but their ruthless attempted murder. You propose a bold conjecture, and then you do everything in your power to refute it. The theories that survive this relentless assault are the ones we provisionally keep, never as proven truths, for nothing is ever finally proven, but as the best-tested guesses currently standing.
It is a thrilling vision, and a flattering one. It casts the scientist as a kind of intellectual hero, fearless, self-critical, perpetually willing to slay his own darlings in the service of truth. It explains, cleanly, why science deserves our trust in a way that ideology and superstition do not. And it has lodged itself so deeply in the self-image of scientists that if you stop a working physicist in a corridor and ask her how science works, there is a good chance she will give you a rough version of Popper, all conjectures and refutations and theories that must be testable. Popper’s falsificationism is, in many ways, the official philosophy of science, the one printed in the opening chapters of textbooks and recited at graduation ceremonies.
And yet there is a worm in the apple, and Kuhn would eventually feed it until it grew large enough to swallow the whole thing. The worm is this: real scientists, the great majority of the time, do not behave like Popperian heroes at all. They do not spend their days trying to murder their own best theories. They spend their days assuming their best theories are correct and using them to solve problems, and when an observation comes in that contradicts the theory, their first response is almost never to declare the theory refuted. Their first response is to assume that something went wrong with the observation, or the equipment, or the assumptions feeding into the calculation, or the competence of the graduate student who ran the experiment. And, crucially, they are usually right to respond this way. The history of science is littered with anomalous results that turned out to be mistakes, and a scientist who abandoned a powerful, productive theory at the first whiff of a contradictory measurement would be not a hero but a fool, throwing away a working tool because of a stray glitch.
Consider a famous case. In the nineteenth century, astronomers noticed that the planet Uranus was not quite where Newton’s theory said it should be. Its orbit wandered, slightly but stubbornly, from the predicted path. Here, by Popper’s rules, was a falsification: an observation contradicting the theory, an opportunity to declare Newton refuted. But that is not what anyone did. Instead, two astronomers reasoned that if Newton was right, there must be an unseen planet out there whose gravity was tugging Uranus off course. They calculated where such a planet would have to be, pointed a telescope at the spot, and found Neptune. The anomaly that might have killed Newtonian theory instead led to a triumphant new discovery, and the theory emerged stronger than ever. Now consider a second case that looks identical. The planet Mercury also wandered from its Newtonian orbit, in a small and persistent way. Astronomers, encouraged by the Neptune triumph, postulated another hidden planet, which they hopefully named Vulcan, and searched for it for decades. They never found it. Mercury’s anomaly was real, and it was eventually explained not by a hidden planet but by Einstein’s new theory of gravity, which dissolved the problem at a stroke. Two anomalies, treated in exactly the same way, with opposite results. The lesson Kuhn drew is that there is no rule, no Popperian algorithm, that tells a scientist in advance which anomalies are mere puzzles to be patched and which are the cracks that signal a coming earthquake. That judgment is not dictated by the evidence. It is made by human beings, working within a framework, drawing on experience and instinct and the shared values of their community.
There is a still deeper trouble lurking beneath the Mercury and Uranus stories, one that philosophers had spotted even before Popper and that gnaws at the very root of his clean picture. When you test a theory against an observation, you are never testing the theory alone. You are testing the theory together with a whole web of supporting assumptions: that your telescope is working properly, that the laws of optics governing its lenses are correct, that the atmosphere is behaving as expected, that the dozens of background theories feeding into your calculation are sound. If the observation comes out wrong, logic alone cannot tell you where the fault lies. Perhaps the theory under test is false. But perhaps, instead, one of the background assumptions is the culprit, and the theory is innocent. The French physicist Pierre Duhem had pressed this point hard around the turn of the century: an experiment can refute a whole bundle of assumptions at once, but it cannot point a finger at the guilty party within the bundle. This means that a clever and determined scientist can almost always save a cherished theory from apparent refutation by adjusting some other part of the web instead, blaming the instrument, questioning the data, modifying an auxiliary assumption. Popper knew this objection and wrestled with it for the rest of his life, insisting that there was an honest way to play the falsification game and a dishonest way, that good scientists do not endlessly wriggle out of refutations even though logic would permit them to. But the wriggle room is real, and it is exactly the room in which Kuhn’s whole account of normal science would later make its home.
This is the gap Kuhn would drive through, and it is why Popper matters so much to our story even though his great clash with Kuhn lies many chapters ahead. Popper had given the cleanest, most heroic statement of the rationalist picture of science, the picture in which clear rules and decisive tests govern the growth of knowledge. By examining that picture against the actual historical record, Kuhn found that it described not the science we have but a science we might wish we had, an idealized contest that real scientists, sensibly, decline to play. The crucial experiment, the clean test that settles everything in an afternoon, is far rarer than the legend of the 1919 eclipse suggests, and even that famous eclipse, when historians later looked closely at the messy data and the judgment calls made in interpreting it, turned out to be a good deal less decisive than the triumphant story let on. Science, Kuhn was coming to believe, does not advance by crucial tests. It advances by something stranger, something that owes as much to the sociology of communities as to the logic of evidence, and to see what that something is, we must turn to a forgotten Polish physician who had glimpsed it a quarter-century before Kuhn, and whose neglected little book Kuhn would one day acknowledge as a crucial influence.
Chapter Five — Ludwik Fleck and the Thought Collective
Every revolutionary likes to imagine he has sprung from nowhere, a lightning bolt out of a clear sky. The truth is almost always more humbling. Ideas have ancestors, and the more original a thinker seems, the more interesting it usually is to discover who got there first. In Kuhn’s case the forgotten predecessor was a man named Ludwik Fleck, a Polish-Jewish microbiologist who, in 1935, published a book in German with a title that promised far less than it delivered: Genesis and Development of a Scientific Fact. Almost no one read it. It appeared in the worst possible circumstances, written by a Jewish scientist in central Europe on the eve of catastrophe, in a language and a discipline where philosophy of science was not expected to come from a working laboratory man. Fleck himself would soon be swept into the horror of the age, surviving the concentration camps of Auschwitz and Buchenwald, where he was put to work, grimly, producing typhus vaccine. His book sank without a trace. And yet tucked inside it was a substantial part of what the world would later celebrate as Kuhn’s great insight.
Fleck’s subject was, on the surface, narrow and technical: the history of a particular medical fact, the test for syphilis developed by August von Wassermann at the turn of the twentieth century. But Fleck used this single example to ask a radical question. We speak of a scientific fact as if it were simply out there, waiting to be discovered, a hard nugget of truth that any competent observer would find. Fleck looked at the actual history of the Wassermann test and saw something quite different. He saw a fact being constructed, slowly and collectively, by a community of researchers who argued, disagreed, refined their methods, changed their minds about what they were even measuring, and gradually settled into a shared way of seeing that eventually hardened into the solid fact taught to medical students. The fact did not drop from the sky. It was built, by a community, over time, and the finished version concealed all the messy labor that had gone into making it, presenting itself as if it had always been obvious.
To describe how this collective construction worked, Fleck coined two phrases that deserve to be far better known. The first was the thought collective, by which he meant a community of people who share ideas and maintain a continuous exchange, a group bound together by a common way of thinking. The second was the thought style, the particular set of assumptions, habits of perception, and shared commitments that the members of a thought collective hold in common, usually without being aware of holding them. A thought style, Fleck argued, does not merely influence what conclusions you reach. It shapes what you are able to perceive in the first place. Two researchers from different thought collectives, looking at the same slide under the same microscope, may genuinely see different things, because each has been trained to notice some features and ignore others, to interpret a smear here as significant and a smudge there as noise. Perception itself, Fleck insisted, is conditioned by the collective. There is no innocent eye, no observer who simply registers what is there without the shaping influence of a community and its style.
Fleck added a further refinement that is worth recovering, because it explains something we all experience and rarely examine: the way a tentative, hedged, deeply uncertain claim made by a handful of specialists gradually hardens, as it travels outward, into a flat and confident fact that everyone simply knows. Fleck pictured a thought collective as a set of concentric circles. At the center sit the esoteric few, the working specialists who know how provisional and contested their results really are, who remember the failed experiments and the unresolved disputes and the awkward exceptions. Around them spreads the exoteric many, the wider circle of educated outsiders, textbook readers, journalists, and the general public, who receive the specialists’ conclusions stripped of all their qualifications. As a claim moves from the inner circle to the outer, it loses its hedges and gains certainty. The cautious maybe of the laboratory becomes the confident fact of the textbook becomes the unquestioned common knowledge of the street. And then, remarkably, this hardened public confidence feeds back inward and pressures the specialists themselves, so that what began as a tentative guess becomes a fixed point that working scientists are reluctant to question because everyone already knows it to be true. The fact acquires a solidity it never earned, manufactured by the social journey it has taken rather than by any new evidence.
It is impossible to read Fleck without feeling the shadow of the catastrophe closing in around him as he wrote. Here was a man arguing that knowledge is a fragile, collective, humane achievement, the patient work of communities thinking together, at the very moment when the most murderous thought collective in history was preparing to engulf his continent. Fleck was deported, and in the camps his expertise in microbiology was turned to the grim task of producing typhus vaccine for his captors, a horror that also, perhaps, kept him alive. He survived, returned to scientific work after the war, and lived long enough to see the first faint stirrings of interest in his neglected book, though he died in 1961, a year before Kuhn’s great success would begin, in time, to send readers back to the obscure Polish physician who had seen so much of it first. There is a terrible irony in a theory of how communities build knowledge being written by a man whom a community was simultaneously trying to erase, and it lends Fleck’s quiet, technical little book a moral weight that Kuhn’s more academic masterpiece never quite carries.
If this sounds familiar, it should. It is, in embryo, Kuhn’s entire theory. The thought collective is the scientific community. The thought style is the paradigm. The idea that perception is shaped by prior commitments is the idea that observation is paradigm-determined, that scientists with different frameworks see different worlds. The notion that facts are collectively constructed and that the finished product hides its messy history is the notion that textbooks conceal the revolutions behind them. Fleck even anticipated the difficulty of communication across thought styles, the trouble that Kuhn would later christen incommensurability. The resemblance is so strong that one might suspect outright borrowing, except that the honest record is more interesting than theft. Kuhn did read Fleck, early, in the 1950s, and he found the little book strange, difficult, and somehow important. He did not fully absorb it, and he was working out many of his own ideas independently, but he never denied the debt. When The Structure of Scientific Revolutions appeared, Kuhn acknowledged Fleck in the preface as a man whose work anticipated many of his own ideas, a gracious admission that did much to rescue Fleck from oblivion and restore him, decades after his death, to his rightful place in the story.
It would be easy, and wrong, to read Fleck as a cynic who thought scientific facts were mere social fictions, invented by communities and answerable to nothing outside themselves. He believed no such thing, and the distinction matters because the same false charge would later be hurled at Kuhn. Fleck was a working microbiologist who spent his days fighting real diseases caused by real microbes, and he never doubted that nature pushed back, that the world resisted some constructions and permitted others. His point was subtler. He held that the resistances nature offers do not come to us raw and uninterpreted; they are always met by a community already equipped with a thought style that decides which resistances to notice, how to describe them, and what to make of them. The microbe is real, and it constrains what the community can say. But the community is also real, and it shapes how the microbe’s reality gets registered, named, and woven into the fabric of accepted knowledge. Fact and framework are partners, not rivals. This middle position, neither naive realism nor giddy relativism, is exactly the territory Kuhn would later try to occupy, and exactly the territory that his careless admirers and his angry critics would both refuse to let him stand on.
Why did Fleck’s ideas languish while Kuhn’s conquered the world? The honest answer has little to do with the merits of the ideas themselves and a great deal to do with the contingencies of history, which is itself a rather Kuhnian moral. Fleck wrote in German, in 1935, as a Jewish scientist whom the coming years would nearly destroy, in a discipline and a moment utterly unprepared to receive his message. Kuhn wrote in English, in 1962, as a respected member of the American academy, riding a wave of postwar interest in the history and philosophy of science, with the institutional backing and cultural timing to make his book travel. The same idea, offered by different people at different moments to different communities, met wildly different fates. There is a paradigm at work even in the reception of theories about paradigms, and Fleck’s sad obscurity and Kuhn’s dazzling fame are a reminder that being right is rarely enough. You must also be heard, and being heard is a matter of community, timing, language, and luck. Fleck supplied the insight. History supplied the megaphone to someone else. But Fleck was not the only ancestor of Kuhn’s revolution. There was also a quieter influence, less philosophical and more historical, coming from a new way of writing the history of science itself, and to that we now turn.
Chapter Six — Koyré, Butterfield, and the History That Bites Back
There is a particular vice that historians fall into so naturally that it has earned its own name, and the name was coined, with some asperity, by an English historian named Herbert Butterfield in 1931. He called it the Whig interpretation of history, after the British political party that liked to tell the national story as a long, triumphant march toward its own enlightened values. The Whig historian, Butterfield observed, reads the past as a kind of rough draft of the present, dividing its actors into the progressives who were helping the world along toward us and the reactionaries who were getting in the way. He judges the past by the standards of the present, hands out praise and blame accordingly, and produces a tidy story of inevitable progress that flatters the present age by making it the goal toward which all of history was secretly striving. It is, Butterfield argued, a deeply distorting way to write history, because it stops you from understanding the past on its own terms, from seeing why people who were neither stupid nor wicked believed and did things that look strange to us now.
Butterfield was writing about political history, but the warning applied with even greater force to the history of science, which had long been the most Whiggish history of all. The standard account of science was a hall of heroes and villains: the brave Galileo against the benighted Church, the clear-eyed Lavoisier against the muddled believers in phlogiston, a steady procession of geniuses who saw the truth battling against the fools and reactionaries who clung to error. Every past scientist was graded by how close he had come to what we now know, awarded marks for his correct anticipations and demerits for his quaint mistakes. This is exactly the attitude that had sent the young Kuhn into Aristotle expecting to find a primitive Newton, and exactly the attitude that his shattering afternoon had taught him to abandon. Butterfield gave the vice a name; he later wrote a history of the scientific revolution himself, and though he remained in some ways a man of his time, he helped legitimize the idea that the history of science deserved to be taken as seriously, and written as carefully, as any other branch of history.
But the figure who did most to transform the field, and who influenced Kuhn most directly, was a Russian-born French philosopher and historian named Alexandre Koyré. Koyré brought to the history of science a conviction that scientific ideas could not be understood in isolation, as a series of technical discoveries, but had to be seen as part of the whole intellectual and even spiritual world of their time, woven together with philosophy, theology, and metaphysics. His great studies of Galileo and the birth of modern physics argued that the scientific revolution was not primarily a matter of new observations or better instruments, but a profound transformation in the very framework of thought, a shift from one entire vision of the cosmos to another. The medievals had inhabited a closed, finite, hierarchical world, a cozy cosmos with the earth at its center and the heavens arranged in nested spheres around it, each thing in its proper place. The moderns came to inhabit an infinite, homogeneous universe, a vast indifferent space in which the same laws held everywhere and no place was privileged over any other. The move from the one world to the other, Koyré argued, was the deepest thing that happened in the scientific revolution, deeper than any single discovery, and it was a change of metaphysical vision, a reorganization of the most basic assumptions about the nature of reality.
Koyré was not entirely alone in this recovery of the past, and one of his most striking forerunners was the very same Pierre Duhem whose logical objection to falsification we met a chapter ago. Duhem was that rare and valuable creature, a working physicist who was also a serious historian, and when he went digging into the medieval period that the standard story dismissed as a long intellectual night, he found something the Whig historians had been too contemptuous to notice. The Middle Ages, far from being a barren waiting room before the dawn of real science, had been alive with sophisticated debate about motion, force, and the structure of the cosmos. Medieval thinkers had anticipated ideas that the textbooks credited entirely to the heroes of the seventeenth century. Duhem’s discovery scandalized the comfortable narrative of a sudden leap from darkness to light, and though later historians would quarrel with the details of his claims, his deeper point stood: when you actually read the past instead of grading it, the tidy story of heroes and villains dissolves into something far richer, slower, and more continuous than the legend allows. The very sharpness of the supposed scientific revolution began to look like an artifact of bad history, a trick of perspective produced by refusing to look closely at what came before.
Behind these individual scholars lay a quieter institutional change that mattered just as much: the history of science was becoming a genuine profession. Where once it had been a hobby for retired scientists and a source of edifying anecdotes for textbook introductions, it was now acquiring its own departments, its own journals, its own standards of rigor. A Belgian-American scholar named George Sarton devoted his life to establishing it as a serious discipline, and though his own vision remained more cumulative and less revolutionary than Kuhn’s would be, the field he helped build created the conditions in which a Kuhn could flourish. There were now people whose full-time job was to read old scientific texts carefully and sympathetically, to reconstruct lost frameworks, to take the past seriously on its own terms. This was the community into which Kuhn, abandoning physics, walked in the 1950s, and it gave him both the tools and the audience he needed. A revolution in the understanding of science required, fittingly, a small revolution in the discipline that studied it, and Kuhn arrived just as that discipline was coming of age.
Here, in Koyré’s work, Kuhn found the historical embodiment of his own dawning conviction. Koyré was describing exactly the kind of wholesale transformation of frameworks that Kuhn had glimpsed in his Aristotle experience, and he was describing it with a richness and seriousness that treated the old framework not as a heap of errors but as a coherent world worth understanding from the inside. From Koyré and the new history he represented, Kuhn took both a method and a vindication. The method was to read past science sympathetically, reconstructing its frameworks rather than grading its results. The vindication was the discovery that, read this way, the history of science really did seem to proceed by great transformations of vision rather than by smooth accumulation, just as his own afternoon with Aristotle had suggested. The philosophers, reasoning from their armchairs about how science ought to work, had given him the tidy universe of cumulative progress and decisive tests. The historians, actually examining how science had worked, gave him something else entirely: a record of frameworks born, frameworks reigning, and frameworks overthrown.
It is worth naming the deep methodological quarrel that all of this set up, because it runs like a fault line through the rest of our story. The philosophers of the tidy universe worked from the top down. They began with an idea of what rational inquiry ought to look like, a logic of evidence and proof, and then held actual science up against that ideal, treating any deviation as a regrettable lapse to be explained away. The new historians worked from the bottom up. They began with what scientists had actually done, the full documentary record of triumphs and blunders and dead ends, and tried to find the patterns in it, letting the ideal emerge from the practice rather than imposing it from above. These are not merely two methods; they are two temperaments, two visions of where authority lies. Does the philosopher tell the scientist how knowledge must grow, or does the historian show the philosopher how it actually grew? Kuhn, trained as a physicist but converted into a historian, planted himself firmly on the historians’ side of this divide, and much of the fury his book provoked among philosophers can be traced to his impertinence in suggesting that their elegant logical reconstructions had simply failed to describe the thing they claimed to explain. He was, in their eyes, letting the grubby facts of history dictate terms to the pure tribunal of reason, and some of them never forgave him for it.
And so the pieces were assembled. From his own experience, Kuhn had the raw shock of watching a dead science come alive. From Fleck, he had the notion of communities and styles of thought that shape perception itself. From Koyré and the new history, he had the evidence that science really does move by transformations of framework, and the method for studying them. What remained was to weld these elements into a single, coherent theory of how science changes, a theory general enough to cover not just the move from Aristotle to Newton or from the closed world to the infinite universe, but the whole pattern of scientific development across every field and every age. That welding took Kuhn the better part of a decade, and the result, when it finally appeared, would be a slim and explosive book. But before we can watch the revolution it caused, we must understand the machinery it described, and that machinery begins with the single most important, most influential, and most maddeningly slippery concept in the whole of Kuhn’s thought: the paradigm itself.
Chapter Seven — Rehearsal: The Copernican Revolution
Five years before he wrote the book that made him famous, Kuhn wrote a different book, and it is the one in which you can watch him teaching himself, on a single great example, the lessons he would later generalize into a theory. The book was called The Copernican Revolution, published in 1957, and on the surface it is a straightforward account of how Western civilization came to believe that the earth goes around the sun rather than the other way around. Read more carefully, it is a dress rehearsal. Every theme of his later masterpiece is already present, tested against the one historical episode Kuhn knew best, and the careful reader of the 1957 book is never surprised by anything in the 1962 one.
To feel the force of the story, you have to first set aside everything you were taught in school, which presents the triumph of Copernicus as the obvious victory of truth over a foolish ancient error. The old earth-centered system, worked out in its mature form by the astronomer Ptolemy in the second century, was not foolish. It was a magnificent intellectual achievement, the product of centuries of patient observation and ingenious mathematics, and it worked. It predicted the positions of the sun, moon, and planets with remarkable accuracy, accurate enough to set the dates of religious festivals, guide navigation, and satisfy nearly everyone for some fourteen hundred years. It fit the plain evidence of the senses, for the earth certainly feels stationary and the heavens certainly appear to wheel around us. And it fit the whole framework of physics and philosophy inherited from Aristotle, in which the heavy earth naturally rests at the center of the cosmos while the lighter heavens revolve around it. The Ptolemaic system was not a superstition. It was normal science of the highest order, a reigning paradigm doing exactly what a good paradigm does.
It did, however, have a complication, and the complication grew slowly over the centuries in a way Kuhn would later recognize as the classic signature of an accumulating strain. To make his earth-centered model match the observed wanderings of the planets, Ptolemy had been forced to introduce a clever device: each planet moved not in a simple circle but in a small circle, an epicycle, whose center itself moved around a larger circle. With enough of these epicycles, circles upon circles, the model could be tuned to fit the data. But as the centuries passed and observations grew more precise, the fit drifted, and astronomers patched it by adding more epicycles, and more, and adjusting the geometry with ever more intricate auxiliary contrivances, until the system became a baroque tangle of wheels within wheels. It still worked, more or less, but it had grown monstrous, and a vague dissatisfaction spread among those who handled it daily, a sense that something so cumbersome could hardly be the true architecture of the heavens. This is precisely what Kuhn would later call the buildup of anomaly and the onset of crisis, the slow loss of confidence that precedes a revolution, and he watched it unfold in the astronomy of the late Middle Ages with the eye of a man recognizing a pattern.
Into this dissatisfaction stepped Nicolaus Copernicus, a cautious Polish churchman who proposed, in a book published as he lay dying in 1543, that the apparent motions of the heavens would be far simpler to explain if the sun, not the earth, sat at the center, and the earth were just another planet wheeling around it along with the rest. Here the schoolbook story declares victory and rings down the curtain. Kuhn, the careful historian, refused to let it. For the most striking and instructive fact about the Copernican proposal is that, when it first appeared, it was not obviously better than the system it sought to replace. It was simpler in its broad conception, yes, but Copernicus had clung to the ancient assumption that heavenly motions must be perfect circles, and to make circles fit the data he too needed epicycles, nearly as many as Ptolemy had used. His system was not dramatically more accurate. It violated the plain evidence of the senses, since the earth manifestly does not feel as if it is hurtling through space. It contradicted the reigning physics, which had no account of how a heavy earth could move or why we are not flung off a spinning globe. And it raised an awkward astronomical objection that Copernicus could not answer: if the earth orbits the sun, the nearby stars should appear to shift against the distant ones over the course of a year, and no such shift could be detected. By the ordinary standards of evidence, a sensible astronomer in 1543 had excellent reasons to reject Copernicus and stick with Ptolemy.
It is worth lingering on that undetectable stellar shift, because it is one of the finest examples in all of science of how an anomaly can sit on either side of a dispute depending on the framework you bring to it. The objection ran as follows. If the earth really swings around the sun in a vast yearly circle, then we are observing the fixed stars from a wildly different vantage point in June than in December, and the nearer stars ought to appear to shift their positions against the farther ones, just as a nearby tree shifts against the distant hills when you walk past it. This shift, called parallax, is exactly what a moving earth predicts, and in the sixteenth and seventeenth centuries no one could detect the slightest trace of it. By the ordinary logic of testing, this was a clean refutation of Copernicus. The defenders of the new system had only one escape, and it required an act of faith that struck their opponents as absurd: they had to claim that the stars were simply so unimaginably far away that the shift, though real, was far too small to measure. They were right. Stellar parallax was finally detected in 1838, nearly three centuries after Copernicus, by which time the heliocentric framework had long since triumphed for entirely other reasons. The anomaly that should have killed the theory was simply carried, unresolved, for three hundred years, tolerated because the framework had won on other grounds. Anomalies, Kuhn learned, do not automatically refute. A confident paradigm can shoulder an unsolved problem indefinitely, and a struggling one can be killed by a problem that later turns out to have a perfectly good answer.
There was also, of course, the matter of God, and Kuhn was careful not to caricature it. The standard tale pits brave science against a reactionary Church, and there is some truth in it, for the Copernican idea did eventually collide with religious authority, most famously in the trial of Galileo. But the deeper resistance was not simply a matter of clergymen defending their turf. The earth-centered cosmos was woven into an entire vision of reality, a vision in which the heavens were the realm of perfection and the earth the realm of change and corruption, in which the whole drama of sin and salvation was staged on a stationary world at the center of creation. To move the earth was not to adjust one technical detail. It was to pull a thread that ran through theology, philosophy, physics, and the ordinary person’s sense of his place in the scheme of things. The resistance to Copernicus was, in part, an entirely intelligible reluctance to unravel a coherent and meaningful world for the sake of a mathematical convenience that did not yet even work better than what it replaced. Kuhn the historian understood this resistance from the inside, as he had understood Aristotle, and he refused to dismiss it as mere stupidity or cowardice.
And there was one final lesson in the Copernican story, the most sobering of all, concerning how a revolution actually gets completed. It is not, in the main, that the defenders of the old paradigm are persuaded and change their minds. It is that they grow old and die, and a new generation, trained from the start in the new way of seeing, takes their place and finds the new framework natural. The physicist Max Planck would later put this with brutal candor, observing that a new scientific truth does not triumph by convincing its opponents but rather because its opponents eventually die and a new generation grows up that is familiar with it. Science, on this view, advances one funeral at a time. Kuhn found exactly this pattern in the slow Copernican victory, and it deepened his unease with the tidy picture of rational conversion. If the old guard never actually changes its mind, if the framework wins simply by outliving the people committed to its rival, then the triumph of a new paradigm starts to look less like a verdict of reason and more like a changing of the generational guard. This troubling thought, planted in the Copernican rehearsal, would grow into one of the most contested features of his mature theory.
And most of them did, for the better part of a century. This is the lesson Kuhn drew, and it is the heart of his mature theory in miniature: the move to a new paradigm is not compelled by the evidence, because at the moment of choice the new paradigm is usually weaker, not stronger, in the things the old one did best. What the new framework offers is not immediate superiority but promise, a different way of organizing the problem that some thinkers find compelling for reasons that are partly aesthetic, partly metaphysical, partly a faith that the new approach will eventually pay off. The Copernican system won not because a decisive experiment settled the matter, but because, over the following century, the work of Kepler, who replaced the perfect circles with ellipses and thereby made the system genuinely accurate, and of Galileo, who looked through his telescope and found new heavens that fit the new picture, and finally of Newton, who supplied a physics in which a moving earth made perfect sense, gradually built out the promise into a triumph. The revolution was not a single event but a long, contested, generational process, and at its outset the revolutionaries were backing a theory that, judged by the cool tribunal of evidence, was losing. Kuhn had found, in the one episode he understood most deeply, that science does not change its mind the way the philosophers said it should. He spent the next five years asking whether the same was true everywhere, and concluding that it was.
Chapter Eight — What on Earth Is a Paradigm?
Ask a physics student how she learned to think like a physicist, and she will not tell you that she memorized a list of the rules of scientific method. She will tell you about problem sets. Night after night, she sat with a textbook full of worked examples and a homework sheet full of unsolved problems, and she struggled to make the second look like the first. She learned by imitation, by seeing how a particular kind of problem had been cracked and then trying to crack a new one the same way, until the method became second nature, a set of reflexes she could deploy without consciously articulating them. If you asked her afterward to state the rules she was following, she would struggle, because she was not really following stated rules at all. She had absorbed a way of seeing problems, a sense of what counts as a legitimate move and a satisfying solution, directly from examples, the way one learns a language or a craft rather than the way one learns a recipe.
This, in its most precise and original form, is what Kuhn meant by a paradigm, and it is worth recovering this narrow meaning from the cloud of vaguer senses that, as we saw at the very start of this book, eventually engulfed the word. We noted earlier that Kuhn used paradigm in at least twenty-one ways, and that he later split it into the broad disciplinary matrix and the narrow exemplar. It is the exemplar, the concrete solved problem that serves as a model for further work, that Kuhn came to regard as the deepest and most important sense, the one he wished he had emphasized from the start. A paradigm, at bottom, is not a theory or a belief or a worldview, though it carries all of those along with it. It is an achievement, a particular triumph so compelling that a community adopts it as the pattern for its own future efforts. Newton’s solution of the planetary orbits, the precise way he set up and solved that specific problem, taught generations of physicists not merely a result but a method, a style, a sense of what a proper physical explanation looks like. They learned the science by mastering the exemplar.
The reason this matters so much, the reason Kuhn thought he had found something genuinely new here, is that it locates the deepest level of a science not in anything that can be written down explicitly but in something that has to be shown, practiced, and absorbed. The philosopher Michael Polanyi, whom we will meet at greater length later, had a fine phrase for this kind of knowledge: he called it tacit knowing, the knowledge we possess but cannot fully put into words, the way we can recognize a face in a crowd without being able to specify the features that let us do it. A great deal of what a scientist knows, Kuhn argued, is tacit in just this way. It lives in the trained judgment, the practiced eye, the acquired sense of which problems are interesting and which approaches are promising, and all of this is transmitted not through explicit instruction but through the shared study of exemplars. This is why you cannot become a scientist merely by reading about science, any more than you can become a carpenter by reading about carpentry. You have to do the problems. You have to be inducted into a community of practice and acquire its tacit skills by apprenticeship.
Why could the deepest level of a science not simply be written down as a set of rules? Kuhn’s answer drew on one of the subtlest ideas in twentieth-century philosophy, the meditation on rule-following that Ludwig Wittgenstein had pursued in his later years. Wittgenstein had noticed that no rule can ever fully specify its own application. Suppose I teach you the rule add two and you produce two, four, six, eight; how do you know that when you reach a thousand you should produce a thousand and two rather than, say, a thousand and four? The rule itself does not say; some further interpretation is always required, and that interpretation cannot be supplied by yet another rule without launching an infinite regress. What actually guides us, Wittgenstein concluded, is not an explicit rule but a shared practice, a trained agreement in how to go on that we absorb from our community and cannot fully articulate. Kuhn seized on this. The scientist confronting a new problem is in exactly the position of Wittgenstein’s pupil: no stated rule tells her how to apply her paradigm to the novel case. What guides her is her trained sense, acquired from exemplars, of what counts as going on in the same way. The paradigm works not as a rulebook but as a model, and the leap from model to new case is a matter of perceived resemblance, not logical deduction.
This is also why Kuhn resisted defining the paradigms of a field by listing the beliefs its members hold in common, for he doubted that any such list could be drawn up. Borrowing again from Wittgenstein, he suggested that the problems a community treats as legitimate are united not by a shared essence but by a family resemblance, the way the members of a human family resemble one another without any single feature being common to all. One problem resembles the founding exemplar in this respect, another in that, a third in some further way, and the whole forms a loose, overlapping network held together by similarity rather than by a defining rule. A practitioner can recognize a legitimate problem of her field when she sees one, just as you can recognize a member of a family by their look, without being able to state the criterion she is using, because there is no single criterion, only a web of resemblances learned from cases.
It helps to see how this plays out across different sciences, for the exemplar takes a different concrete form in each. In physics it may be a mathematical problem solved in a particular way, the pendulum or the inclined plane worked through so cleanly that it becomes the template for a hundred variations. In chemistry it may be a classic determination of a structure, a way of reasoning from evidence to the architecture of a molecule that students reenact until it becomes instinct. In biology it may be a decisive experiment or a model organism, the fruit fly or the particular cross whose analysis shows how the whole discipline ought to proceed. In each case the community is bound together less by a creed than by a shared repertoire of triumphs, a canon of solved problems that defines, by example, what good work in the field looks like. To enter the field is to apprentice yourself to that canon, and to master the field is to be able to extend it to cases its founders never imagined.
Once you grasp this, several puzzling features of scientific life snap into focus. It explains why scientific training is so strikingly uniform and, frankly, so dogmatic. Unlike students of literature or philosophy, who are encouraged to argue with the great thinkers and form their own views, students of physics or chemistry are handed a single authoritative way of doing things and drilled in it without much invitation to question the foundations. To an outsider this looks intellectually stifling, and in a sense it is. But Kuhn argued that this very narrowness is the secret of science’s power. By agreeing, without further argument, on a shared paradigm, a community frees itself from the endless foundational debates that consume other fields and can pour all its energy into the detailed, cumulative work of extending and refining its framework. The dogmatism is not a defect. It is the price of admission to the most productive kind of inquiry, and the productivity is what we will examine next, in the quiet, patient, deeply underrated activity that occupies most scientists for most of their careers.
There is one more consequence of the paradigm-as-exemplar that deserves a place here, though we will return to it in force later, because it is among the most startling things Kuhn ever claimed. A paradigm does not merely tell a scientist how to solve problems; it tells her, in advance, which operations and measurements are even worth performing, and thereby shapes what she will treat as data at all. The measurements a scientist makes in her laboratory are not simply handed to her by experience, lying about waiting to be picked up. They are laboriously selected, out of the infinite things one could in principle measure, precisely because the paradigm has marked them as the ones likely to repay the effort. The scientist does not measure everything; she measures what her framework has taught her is significant. And this means that two communities working under different paradigms will not only interpret the same data differently; they will, to a real degree, collect different data in the first place, performing different operations in their laboratories because their frameworks point them toward different things. The neutral bedrock of pure observation, the same for everyone regardless of theory, on which the tidy universe had built its account of how evidence settles disputes, begins here to look far less solid than the philosophers had assumed. We will see, when we reach the question of whether rival scientists inhabit different worlds, just how far Kuhn was prepared to push this unsettling thought.
It also explains why paradigms are so resistant to overthrow, a fact that will loom large when we come to revolutions. If a paradigm were merely a set of explicit beliefs, you could refute it with a counterargument and a contrary belief would slot neatly into its place. But a paradigm is not a set of explicit beliefs. It is a whole acquired way of practicing, embedded in tacit skills, trained perceptions, shared standards, and a community’s entire culture of work. You cannot refute a way of life with a single experiment, any more than you can refute a language. To abandon a paradigm is to give up not a proposition but a craft, a set of hard-won competences and the community that sustains them, and human beings do not do this lightly or for trivial reasons. Understanding the paradigm as exemplar, as practiced craft rather than stated doctrine, is therefore the key that unlocks everything that follows, and it repays the effort of holding it clearly in mind.
Chapter Nine — The Quiet Labor of Normal Science
The word revolution gets all the attention, but it describes the rarest and least typical thing that happens in science. If you want to know what science actually is, as a daily human activity, you must look not at the dramatic upheavals but at the long, calm stretches between them, the years and decades during which the vast majority of scientists do the vast majority of their work. Kuhn gave this ordinary, undramatic activity a name that perfectly captures its character: normal science. And though the name sounds dismissive, almost a put-down, Kuhn meant it as the highest tribute, for he believed that normal science is where the real cumulative achievement of the scientific enterprise actually happens, and that without it there could be no science worth the name at all.
What does a scientist do during a period of normal science? She solves puzzles. Kuhn chose the word puzzle with great care, and the analogy is worth taking seriously. A puzzle, in his sense, is a problem that has a guaranteed solution, a problem whose very statement assures you that an answer exists and gives you the rules for finding it. A jigsaw puzzle is the obvious model: you know in advance that the pieces fit together into a definite picture, and the challenge is purely one of skill and patience in finding how. A crossword is another: the answers exist, the conventions are fixed, and your task is to be clever enough to recover them. Normal science, Kuhn argued, consists of puzzles in exactly this sense. The paradigm guarantees that the problems it poses have solutions, and it supplies the rules and tools for finding them. The scientist’s job is to apply her trained skill to bring the recalcitrant bit of nature into line with what the paradigm leads her to expect. When she succeeds, she has not discovered some shocking new truth that overturns her field. She has demonstrated her competence and extended the reach of the paradigm a little further, exactly as the solver of a difficult crossword has not revolutionized the dictionary but shown her mastery of it.
This puts the relationship between the scientist and her anomalies in a surprising light, one that flatly contradicts the Popperian heroism we examined earlier. When a normal scientist runs an experiment and the result fails to match the paradigm’s prediction, she does not, as Popper’s philosophy would have her do, joyfully conclude that she may have refuted her field’s reigning theory. She concludes, almost always, that she has failed to solve the puzzle. The fault, she assumes, lies with her: a flaw in her apparatus, an error in her technique, an overlooked complication in her setup. And this assumption is not intellectual cowardice; it is sound professional judgment, vindicated overwhelmingly by experience. Nine hundred and ninety-nine times out of a thousand, the anomalous result really is a mistake, and the scientist who tracks down her own error and gets the expected result has done her job correctly. A paradigm, Kuhn observed, is not on trial during normal science. The scientist is. Her skill is being tested against nature’s puzzles, with the paradigm itself serving as the unquestioned standard of success, and this arrangement, far from being a failure of nerve, is precisely what allows normal science to be so astonishingly productive.
For productive it certainly is, and Kuhn insisted on this with real admiration. Precisely because the normal scientist does not waste her energy questioning fundamentals, precisely because she takes the paradigm for granted and trains her full attention on a narrow, well-defined problem, she can achieve a depth and detail of result that would be impossible for a mind perpetually relitigating its own foundations. The paradigm acts like a pair of blinders, and the blinders, by restricting the field of vision, allow an intensity of focus that produces real discoveries. Whole continents of precise knowledge have been opened up by scientists who never once doubted the framework within which they worked. The measurement of fundamental constants to ever more decimal places, the determination of chemical structures, the mapping of physical regularities to exquisite precision, all of this patient, accumulating, genuinely cumulative work is the fruit of normal science, of communities committed to a paradigm and devoted to working out its consequences. When Kuhn said that science is mostly normal science, he was not belittling it. He was pointing out that the steady, reliable, accumulating progress we rightly associate with science happens almost entirely inside paradigms, not at the moments when paradigms change.
Kuhn took the trouble to sort the daily labor of normal science into three broad kinds, and the taxonomy is illuminating because it shows just how little of ordinary scientific work consists of testing fundamental theories. The first kind is the determination of significant facts: measuring the things the paradigm has marked out as important, pinning down the wavelengths of spectra, the boiling points of compounds, the orbits of bodies, with ever greater precision and across ever more cases. The paradigm tells you which facts are worth the immense trouble of measuring accurately, and normal science measures them. The second kind is the matching of facts with theory: working out the consequences of the paradigm in enough detail that they can be compared with observation, often a formidable labor of calculation and approximation, since the bare theory rarely yields predictions about real, messy situations without enormous interpretive effort. The third kind is the articulation of the paradigm itself: resolving its internal ambiguities, determining its constants more precisely, extending it to new domains, tidying the framework and pushing its boundaries outward. Notice that none of these three is an attempt to test whether the paradigm is true. All three presuppose its truth and labor to work out what follows from it. The paradigm is the unquestioned ground on which all the work stands.
A homely image Kuhn liked was that of mopping up. The revolutionary act of establishing a new paradigm is like a great military breakthrough; what follows is the long, unglamorous campaign of occupying and consolidating the conquered territory, of cleaning up the pockets of resistance, of building the roads and institutions that turn a battlefield into a settled province. Most scientists are mopping-up workers, and Kuhn meant no insult by it. The mopping up is where the territory actually gets developed, where the abstract promise of the breakthrough is converted into the dense, detailed, usable knowledge that constitutes a mature science. It is also, he noted dryly, the part of science that the textbooks and the popular accounts almost entirely ignore, because it makes for a dull story. We tell tales of Newton and Einstein and the great revolutionary leaps; we say almost nothing of the legions of patient workers who spent their careers turning those leaps into a working science, even though they are the overwhelming majority and their cumulative contribution is, in a sense, the science itself.
There is something almost medieval, Kuhn suggested, in the structure of a normal-scientific community, something closer to a craft guild than to the popular image of the free-thinking individual genius. A guild trains its apprentices in a fixed body of technique, certifies their competence by their mastery of that technique, polices the boundary between qualified members and outsiders, and judges work by how well it conforms to the standards of the craft. A scientific specialty does much the same. It trains its students in a shared body of exemplars, certifies them through examinations and degrees, controls entry through a system of credentials, and evaluates contributions by whether they solve the field’s recognized puzzles in the field’s recognized ways. This guild-like character is, again, not a corruption of science but a condition of its power. It is what allows a community to maintain the shared standards and tacit skills without which cumulative work would be impossible. But it has a consequence that will matter enormously when we turn to revolutions: a community organized like a guild, devoted to a single craft and trained to defend its standards, is by its nature deeply conservative, profoundly resistant to anything that would require it to abandon the very techniques that define it. The strength of normal science and its resistance to change are the same trait seen from two sides.
And yet the blinders that make normal science so powerful also set the stage for its eventual crisis, in a way that gives Kuhn’s whole picture its peculiar dramatic shape. The same intense, focused work that extends a paradigm so successfully also drives it, relentlessly, into the corners where it does not quite fit. A community devoted to matching nature against its framework, with ever-increasing precision and ever-widening scope, is a community engaged in the most thorough possible search for the places where the framework fails. Most of the mismatches it turns up are dissolved as ordinary puzzles, mere errors to be corrected. But every so often a mismatch resists every attempt at solution, returns again and again despite the best efforts of the most skilled practitioners, and slowly forces itself upon the community’s attention as something more than a personal failure. The puzzle that will not be solved begins its quiet career as an anomaly, and the anomaly that will not go away is the seed from which crisis and revolution grow. Normal science, in other words, is not the opposite of revolution but its necessary precondition, the patient process that alone can generate the stubborn anomalies capable of bringing a paradigm down. To understand how a settled framework comes apart, we must therefore look closely at the anomaly, the unwelcome visitor that refuses to leave, and at the strange way a community responds when its most trusted tools begin, persistently, to fail.
Chapter Ten — When the Anomaly Will Not Go Away
On a November evening in 1895, a German physicist named Wilhelm Röntgen was working alone in his darkened laboratory with a glass tube through which he was passing an electric current, when he noticed something that should not have been there. A small screen coated with a fluorescent material, sitting on a bench some distance away, had begun to glow faintly. There was no reason for it to glow. The tube was covered in black cardboard; no light could be escaping it; and yet the screen shone whenever the current flowed and went dark when it stopped. Röntgen, instead of assuming he had made a mistake and going home, did the thing that distinguishes the discoverer from the merely competent: he took the anomaly seriously. He spent the following weeks obsessively chasing the strange glow, and discovered that some unknown kind of ray was streaming from his tube, passing through cardboard, wood, and flesh, casting on the screen the shadow of the bones inside his own hand. He had stumbled onto X-rays, and the discovery would transform medicine and physics alike. But notice the shape of the event. A discovery began with a violation of expectation, with something appearing where the reigning framework said nothing should appear. Discovery begins with anomaly.
This is one of Kuhn’s sharpest and least intuitive claims, and it inverts the comfortable picture of science as the steady confirmation of what we expect. New things are found, Kuhn argued, precisely when nature violates the expectations the paradigm has trained into us, when something refuses to behave as it should. And here lies a deep paradox that Kuhn relished. A paradigm, by telling us so precisely what to expect, is what makes anomaly possible in the first place. Only against a background of confident expectation can something show up as surprising, as wrong, as out of place. A vague and flexible framework that expected anything could be surprised by nothing. It is the very rigidity of a strong paradigm, its precise and demanding predictions, that allows the world to announce, by violating them, that something new is afoot. The paradigm is both the thing that anomalies eventually threaten and the precondition that makes them visible at all. Normal science, the patient and conservative activity we examined in the last chapter, is therefore the most effective possible instrument for generating the surprises that will one day overthrow it.
But not every surprise becomes a genuine anomaly, and here we must draw a careful distinction, for it is one that Kuhn’s casual readers often blur. The vast majority of unexpected results are not anomalies at all in Kuhn’s strong sense; they are puzzles, mismatches that the community quickly resolves by correcting an error, refining a technique, or adjusting some minor assumption. These are the daily fare of normal science, and they leave the paradigm untouched, indeed they strengthen it, since each one resolved is another demonstration that the framework can handle whatever nature throws at it. An anomaly in the deep sense is something rarer and more dangerous: a mismatch that resists resolution, that returns again and again despite the best efforts of the most skilled practitioners, that cannot be made to go away by any of the paradigm’s standard moves. Such a thing does not announce itself dramatically at first. It begins as just another puzzle, indistinguishable from the thousand puzzles that get solved. Only gradually, as attempt after attempt fails, does it dawn on the community that this particular puzzle is different, that it touches something the paradigm cannot reach.
Consider the slow death of phlogiston, one of Kuhn’s favorite cases. For most of the eighteenth century, chemists explained burning by means of a substance called phlogiston, supposedly released from materials as they burned. The theory was productive and explained a great deal. But it had a nagging problem: when certain metals were burned, the resulting residue weighed more than the original metal, not less, even though the metal was supposed to be losing phlogiston as it burned. How could a thing weigh more after giving something off? For decades this was treated as a minor puzzle, patched with various contrivances; some chemists even proposed, heroically, that phlogiston had negative weight. The puzzle did not bring the theory down on its own. It sat there, an irritant, a small cloud on an otherwise sunny horizon, tolerated because the framework was otherwise so useful. It was only when Lavoisier, working within a different vision, recognized that burning involves combining with something from the air rather than releasing something into it, that the troublesome weight gain suddenly became not an embarrassment but the central clue, the very thing the new theory was built to explain. The same fact was a minor nuisance under one paradigm and a foundation stone under the next.
This reveals something crucial about the life cycle of an anomaly: its significance is not fixed but depends entirely on the framework that beholds it. The increase in weight on combustion was, for the phlogiston chemist, a marginal puzzle to be explained away; for the oxygen chemist, it was the heart of the matter. The precession of Mercury, which we met earlier, was for the Newtonian a small unsolved problem to be patched with a hypothetical planet; for Einstein, it was a triumphant confirmation of a new theory of gravity. An anomaly, in other words, is not simply a fact that contradicts a theory. It is a fact whose meaning and importance shift dramatically depending on which framework is asking the question, and a mismatch that one paradigm can shrug off may be exactly the lever that pries it loose once a rival appears that can make sense of it. This is why anomalies so rarely topple a paradigm by themselves, in isolation. They become deadly only in conjunction with an alternative, only when there is somewhere else for the community to go.
The oxygen story carries a further lesson that Kuhn drew out with particular delight, because it dismantles one of our most cherished and most naive ideas about science: the idea that discoveries happen at definite moments and have definite discoverers. Ask who discovered oxygen, and you will find there is no clean answer, and the absence of a clean answer is itself the point. A Swedish apothecary named Scheele isolated the gas first but did not publish in time. An English clergyman named Priestley isolated it, described it carefully, and went to his grave convinced it was something he called dephlogisticated air, interpreting his own discovery entirely within the dying framework of phlogiston. It was Lavoisier who recognized that the new gas was an element and grasped its true role in combustion, yet even he came to a full understanding only gradually, and arguably never entirely abandoned certain confusions of his own. So who discovered oxygen, and when? The question has no answer, Kuhn argued, because discovery is not an instantaneous event but an extended process, a matter of both noticing that something new exists and grasping what it is, and these two achievements can be separated by years and distributed across several people who do not even agree on what they have found. The tidy image of the eureka moment, the lone genius crying out at the instant of revelation, is a fiction the textbooks tell. Real discovery is collective, gradual, and bound up at every step with the shifting framework through which the new thing is perceived.
How, then, does a community come to take an anomaly seriously, to promote it from background irritant to foreground crisis? Kuhn observed that this usually requires not one anomaly but a convergence: several stubborn problems accumulating at once, or a single anomaly that strikes at the paradigm’s most central commitments rather than its periphery, or a practical urgency that makes the unsolved problem impossible to ignore. The mere existence of an unsolved problem is never enough; every paradigm at every moment is surrounded by unsolved problems, and a science that abandoned its framework at the first difficulty would never accomplish anything. What matters is whether the anomalies begin to seem not like the ordinary residue of incomplete work but like symptoms of something fundamentally wrong, whether the community starts to lose its confidence that the puzzles will eventually yield to the standard methods. That loss of confidence is the threshold. On one side of it lies normal science, serenely solving its puzzles; on the other lies crisis, the strange and unsettled condition in which a community begins, for the first time, to doubt the framework it had always taken for granted.
The transition across that threshold is gradual, contested, and rarely visible to those living through it until it is well advanced. There is no announcement, no moment when the community formally declares that an anomaly has become a crisis. Instead there is a slow shift of mood, a creeping unease, an increasing tendency to talk about the troublesome problem at conferences and in correspondence, a growing willingness to entertain ideas that would earlier have been dismissed out of hand. The puzzle that would not go away has begun its quiet work of corrosion, eating away at the confidence that holds a paradigm in place. And once that confidence is sufficiently weakened, the whole character of the science begins to change, the rules begin to loosen, and the community enters the peculiar and creative and frightening state that Kuhn called crisis, in which the ordinarily unquestionable suddenly becomes a matter for open debate, and the door swings open, for the first time in perhaps a century, to genuinely new ways of seeing the world.
Chapter Eleven — Crisis, and the Loosening of the Rules
In the spring of 1900, the eminent physicist Lord Kelvin is said to have surveyed the magnificent edifice of classical physics and pronounced it essentially complete, marred only by two small clouds on the horizon. The clouds were a pair of nagging experimental anomalies that the great framework of Newton and Maxwell could not quite accommodate. Kelvin expected them to be cleared up with a little more work. Instead, within a few years, those two small clouds had broken into the two greatest storms in the history of physics: one became the theory of relativity, the other became quantum mechanics, and between them they tore the classical edifice apart and rebuilt physics from its foundations. There is no more vivid illustration of Kuhn’s central warning. The community that feels most complete, most confident that only mopping up remains, may be standing on the very edge of its deepest crisis, and the small clouds it dismisses may be the first signs of an approaching revolution.
What happens to a science when it enters a crisis? The most important change, Kuhn observed, is that the rules begin to loosen. During normal science, the paradigm is unquestioned, the standards are fixed, and the range of acceptable problems and solutions is narrow and well-defined. In a crisis, all of this softens. The community, having lost confidence that the standard methods will resolve its deepest problems, becomes willing to consider approaches it would earlier have rejected as illegitimate, even absurd. The boundaries of the field grow blurry. Practitioners who would never have dreamed of questioning their fundamental assumptions begin to question them. And, strikingly, scientists in crisis often turn to philosophy, to the kind of foundational reflection that a healthy normal science regards as a waste of time. When the framework is working, no one asks what space and time really are, or what it means to measure something; one simply gets on with the work. When the framework is failing, these deep questions come flooding back, and the scientists find themselves doing the kind of fundamental conceptual analysis that, in calmer times, they would have left to the philosophers and the cranks.
It is worth making one of Kelvin’s two clouds concrete, because it shows a crisis in motion and because Kuhn himself would later devote an entire book to it. The cloud concerned the way hot objects glow. Classical physics, applied with perfect rigor to the radiation given off by a heated body, made a prediction that was not merely slightly wrong but catastrophically, absurdly wrong: it implied that any warm object should be pouring out an infinite amount of energy at the short-wavelength end of the spectrum, which would mean, among other consequences, that the inside of your oven should blast you with lethal radiation every time you opened it. This was so plainly false, so violently at odds with the most ordinary experience, that it could not be dismissed as a minor puzzle. It was an anomaly striking at the foundations. And the resolution, when Max Planck reluctantly arrived at it in 1900, required an assumption so strange that Planck himself regarded it for years as a mere mathematical trick rather than a truth about nature: the assumption that energy comes not in a smooth continuous flow but in discrete indivisible lumps. That desperate expedient, adopted to quiet one of Kelvin’s little clouds, was the seed of quantum theory, and it would eventually overturn the entire classical conception of the physical world. A crisis at the periphery had reached, through the patient pressure of an unsolvable anomaly, all the way to the center.
A second symptom of crisis is the proliferation of competing versions of the paradigm. As long as a framework is healthy, it exists in essentially one authoritative form, shared by all competent practitioners. As the crisis deepens, the single paradigm begins to fragment into a multitude of variants, each adjusted in some different way to try to cope with the troublesome anomalies. Where there was once one Ptolemaic astronomy, there came to be many, each tinkering with the epicycles differently. Where there was once one classical physics, there came to be a swarm of modifications and patches as physicists scrambled to save the old framework from the new data. This proliferation is itself a diagnostic sign. When the experts can no longer agree even on the proper form of their own shared theory, when the single voice of a confident discipline dissolves into a babble of competing adjustments, the science has entered the unmistakable condition of crisis, and the stage is set for something more drastic than adjustment.
Kuhn captured the underlying dynamic in a phrase that became the title of one of his later essay collections: the essential tension. Science, he argued, requires two opposed virtues that sit uneasily together. It requires tradition, the deep conservatism of normal science, the willingness to work patiently within an inherited framework and not go chasing every novelty, for without this discipline no cumulative work is possible and the field dissolves into endless foundational squabbling. But it also requires innovation, the capacity, at the right moment, to break with tradition and embrace a radically new framework, for without this a science would ossify and die, forever patching a framework that no longer fits the world. The healthy scientist must be both a staunch traditionalist and a potential revolutionary, and these are not easy companions in a single mind. Most of the time the traditionalist must dominate, or there would be no normal science at all. But in a crisis, the balance must somehow shift, and the community must find within itself the capacity to do the very thing its training has taught it never to do: to abandon the paradigm.
This is psychologically wrenching, and Kuhn did not pretend otherwise. To a scientist who has spent decades mastering a framework, whose entire professional identity and competence are bound up with it, the suggestion that the framework should be discarded is not an abstract intellectual proposal but a threat to everything she is. The tools she has spent a lifetime sharpening would become useless; the problems she knows how to solve would cease to matter; the skills that make her an expert would be devalued overnight. It is no wonder that the response to crisis is so often denial, resistance, and a redoubled effort to save the old framework by ever more elaborate contrivances. The defenders of a paradigm in crisis are not stupid or dishonest. They are doing what their training and their reasonable professional judgment tell them to do, which is to assume that the problems will eventually yield, as so many problems have yielded before. The tragedy, and the drama, of a scientific revolution lies in the fact that they are sometimes right to resist and sometimes wrong, and that there is no rule to tell them which.
A crisis, then, is a peculiar and unstable condition, and it does not last forever; it must be resolved one way or another. Kuhn identified three possible outcomes. Sometimes the old paradigm proves equal to the challenge after all, the anomaly is solved by normal means, and the crisis passes, leaving the framework intact and even strengthened. Sometimes the problem resists all solution, but no acceptable alternative emerges either, and the community simply sets the anomaly aside, labels it intractable, and bequeaths it to a future generation with better tools, returning to normal science with the troublesome problem quarantined. And sometimes, the outcome that interests us most, a new candidate paradigm arises, and the field is plunged into the strange contest between old and new that constitutes a scientific revolution. It is only in this third case that the crisis flowers into genuine upheaval, and even then the upheaval does not proceed as a rational weighing of evidence on neutral scales. It proceeds, as we are about to see, as something far more like a conversion, a transformation of vision in which the world itself appears to change before the scientist’s eyes.
And here we arrive at the threshold of the most radical and most contested idea in all of Kuhn’s work. For if the transition between paradigms were simply a matter of weighing the evidence and choosing the framework that fits the facts better, there would be nothing revolutionary about it, and the tidy universe could absorb the whole story as an unusually dramatic episode of ordinary rational progress. Kuhn insisted that it is not like this at all, that the transition involves something the tidy universe cannot accommodate: a shift in perception so total that the rival scientists, looking at the same world, come to see different things. To understand what he meant, and to feel why it provoked such fury, we must turn to a humble little drawing from the psychology of perception, a drawing that can be seen, at will, as either a duck or a rabbit, and that Kuhn made into the emblem of the most mysterious moment in the life of science.
Chapter Twelve — The Gestalt Switch
There is a famous little figure, a few simple lines on a page, that the psychologists of perception love to show their students. Looked at one way, it is plainly a duck, its bill pointing to the left. Looked at another way, the very same lines become a rabbit, the former bill now a pair of ears pointing right. Nothing on the page changes; the ink is identical from one moment to the next. What changes is something in you, the seer, and the change is sudden and complete. You do not see a blurry intermediate creature, half duck and half rabbit, that gradually resolves; you see a duck, and then, in an instant, you see a rabbit, with no transition in between. And once you can do this, you cannot un-know it; you can flip back and forth at will, but you can never recover the innocent state in which the figure was simply and only a duck. Kuhn seized on this humble drawing and made it the central image of his theory of revolutions, for he believed that what happens to a scientist who converts from one paradigm to another is, at the deepest level, a perceptual shift of exactly this kind.
This is a startling claim, and it is essential to grasp how much stronger it is than the ordinary observation that scientists change their beliefs. Of course scientists change their beliefs; everyone does. Kuhn was saying something more disturbing: that in a genuine paradigm shift, the scientist does not merely come to believe new things about the same world, she comes to see a different world. The astronomer who accepts the Copernican framework does not simply add the belief the earth moves to her existing stock of beliefs while everything else stays put. She comes to see the rising sun differently, for she is now watching the horizon of a spinning earth dip below a stationary sun, where before she watched a sun climbing above a stationary earth. The chemist who accepts oxygen does not merely revise one item in her theory; she comes to see combustion itself differently, to perceive the burning candle as a combining rather than a releasing. The world the scientist works in, the world as it presents itself to her trained perception, is transformed. Kuhn went so far as to write, in one of his most provocative formulations, that after a revolution scientists are responding to a different world.
It is worth pausing to ask how seriously Kuhn meant this, because it is exactly the kind of statement that made his critics apoplectic. Did he really believe that the world physically changes when a scientist changes her mind, that the sun literally alters its behavior because an astronomer has adopted a new theory? Of course not; Kuhn was not insane. He was pointing at something real and elusive about the nature of perception itself, a point that the psychology of his day strongly supported: that what we see is never simply a passive recording of what is out there, but is always shaped, organized, and interpreted by what we already know and expect. The trained radiologist and the layman look at the same X-ray and see different things, because the radiologist’s perception has been educated to pick out patterns the layman’s cannot. The experienced birdwatcher and the novice look at the same flicker in the bushes; one sees a particular warbler, the other sees a brownish blur. Perception is theory-laden, shaped by prior training, and since a paradigm is precisely the deepest layer of a scientist’s training, a change of paradigm reaches all the way down into perception and changes what the scientist sees. That is what Kuhn meant by a different world, and it is a good deal more defensible than the caricature his enemies attacked.
The gestalt analogy also illuminates why revolutionary change is so sudden, so total, and so resistant to gradual persuasion. You cannot ease someone from duck to rabbit by degrees, presenting them with a sequence of slightly more rabbit-like images; the switch, when it comes, comes all at once, a reorganization of the whole rather than an accumulation of parts. So too, Kuhn argued, with conversion to a new paradigm. The scientist does not inch toward the new framework one belief at a time, growing incrementally more convinced; she resists, resists, resists, and then, often quite suddenly, the whole pattern reorganizes and she finds herself on the other side, seeing the new way and unable to recover the old. This is why Kuhn so often reached for the language of religious conversion, of scales falling from the eyes, of a flash of illumination. He was not being mystical or sloppy. He was describing the actual phenomenology of theory change as the historical record and the testimony of scientists themselves reported it, a phenomenology that the tidy picture of gradual rational adjustment simply failed to capture.
Kuhn supported the perceptual reading with a genuinely unsettling experiment from the psychology laboratory, one he cited with relish. In the 1940s, two psychologists named Bruner and Postman flashed playing cards in front of subjects and asked them to identify each one. Most of the cards were normal, but a few had been doctored: a red six of spades, a black four of hearts, cards whose color and suit had been deliberately mismatched. What happened was remarkable. For short exposures, subjects confidently and without hesitation identified the anomalous cards as normal ones, reporting the red spade as either a normal red heart or a normal black spade. Their minds, trained by a lifetime of ordinary cards, simply forced the anomalous stimulus into one of the expected categories; they could not see what was actually in front of them because their perceptual framework had no slot for it. Only when the cards were shown for longer, and shown repeatedly, did the subjects begin to hesitate, to sense that something was wrong, to grow confused and uncomfortable. And then, often quite suddenly, the framework would adjust, and they would see the cards as they really were. Here, in miniature, in a controlled laboratory setting, was the entire drama of Kuhn’s theory: expectation shaping perception, anomaly resisting recognition, mounting unease, and finally the sudden gestalt switch into a new way of seeing. Science, Kuhn was suggesting, behaves like those subjects, and a paradigm is the deck of expected cards.
The irreversibility of the switch deserves emphasis too, for it explains one of the most poignant features of scientific revolutions. Once a scientist has made the gestalt shift into a new paradigm, she generally cannot recover the old way of seeing; like the viewer who can no longer find the innocent duck once the rabbit has appeared, she finds the former framework not merely wrong but almost unintelligible, a way of thinking she can describe but no longer inhabit. This is why communication across a completed revolution is so difficult, and why the victors so often misremember and caricature the vanquished position, reading the old paradigm as a confused anticipation of the new rather than as the coherent alternative it once was. The convert cannot un-convert, cannot climb back inside the discarded vision, and so loses precisely the sympathetic access to the old framework that would let her do it justice. The history of science gets written by the converted, in the language of the new paradigm, and the old world it replaced becomes, in the retelling, a mere collection of errors, its inner coherence forgotten. Recovering that lost coherence, as we saw at the very beginning of this book, is the historian’s hardest and most valuable task.
But the gestalt switch carries a sting in its tail, and the sting is the part of Kuhn’s theory that would cause him the most grief for the rest of his life. If converting from one paradigm to another is like flipping from duck to rabbit, then it is not, on its face, a process governed by argument and evidence at all. You cannot argue someone into seeing the rabbit; either the switch happens or it does not, and no chain of reasoning compels it. If theory change is really like this, then it begins to look as though the choice between paradigms is not a rational choice in the ordinary sense, not something the evidence forces upon any competent observer, but something closer to a leap, a conversion, a change of vision that reasons may encourage but cannot compel. And if that is so, then the comfortable idea that science is the realm where disputes are settled by evidence, where any rational person must in the end bow to the facts, starts to look like a myth. This is the doorway to the charge of irrationalism, the accusation that Kuhn had reduced the grandest achievements of human reason to a matter of mob psychology and fashion. He spent decades insisting that this was a misreading, that he had never meant anything of the kind. Whether he was right to insist this, and whether his theory can really escape the relativism his critics smelled in it, is among the great unresolved questions of his thought, and it is a question we are now, at last, equipped to confront head-on, beginning with the deepest and most difficult of all his ideas: the strange notion that rival scientists, in some sense, live in different and untranslatable worlds.
Chapter Thirteen — Do Scientists Live in Different Worlds?
Imagine two astronomers standing side by side on a hilltop at dawn, watching the sun come up. One of them lived before Copernicus and is certain the earth stands still; the other lives after Copernicus and is certain the earth is spinning. They are looking at precisely the same scene, the same reddening sky, the same brightening disc climbing above the same horizon. Do they see the same thing? The commonsense answer, and the answer the tidy universe insisted upon, is obvious: of course they see the same thing. They see the sun rising. They merely hold different theories about what is really going on behind the appearance. The appearance is shared, neutral, given; the interpretation differs. This clean separation between what we see and what we think about it was the bedrock on which the whole rationalist account of science was built, because it guaranteed a common court of appeal, a shared body of observation to which rival theorists could submit their disputes for impartial judgment.
Kuhn spent a good deal of his most difficult prose trying to kick that bedrock out from under the edifice. The two astronomers, he suggested, do not simply interpret a shared appearance differently; in an important sense they see different things. The first watches the sun itself move, rising of its own accord above a stable earth. The second watches a stationary sun be revealed as her own moving horizon dips away beneath it. These are not the same perception dressed in different theories; they are different perceptions, organized by different frameworks reaching down into the act of seeing itself. We met the foundation of this claim in the last chapter, in the psychology of the gestalt switch and the theory-ladenness of perception. Now we must follow it to its dizzying conclusion, the conclusion that earned Kuhn the lasting enmity of philosophers who accused him of dissolving the one shared world into a multitude of private ones. For if there is no neutral observation, no appearance that all parties share regardless of theory, then the impartial court of appeal vanishes, and with it the comfortable guarantee that scientific disputes can always be settled by looking.
The provocation reaches its peak in a handful of sentences where Kuhn wrote that after a revolution scientists work in a different world. Taken literally, this is either trivial or mad. Trivial, if it means only that scientists think about the world differently after a revolution, which no one would deny. Mad, if it means that the physical universe itself rearranges its furniture to suit the latest theory, which no one could believe. Kuhn was aiming at something between these poles, something genuinely hard to state without falling off one side or the other, and his lifelong difficulty in pinning it down is a measure of how slippery the territory is. What he seems to have meant is that the world a scientist actually engages with, works in, perceives and manipulates, is not the bare physical universe in itself but that universe as carved up, categorized, and made meaningful by a paradigm; and since the paradigm changes in a revolution, the carved-and-categorized world the scientist inhabits changes too, even though the underlying stuff, whatever that even means apart from some framework, does not.
This is where the deepest philosophical anxieties begin to stir, the anxieties that go by the name of the question of realism. Is there a single, fixed, mind-independent reality that science is gradually getting right, a way the world is in itself, independent of any framework, against which our theories can be measured for truth? The tidy universe answered yes, confidently. Kuhn’s answer was, at best, a deeply uncomfortable maybe, and at worst a quiet no, and he was never able to settle it to his own satisfaction. He did not want to say that there is no reality, that scientists simply make it all up; he was not a fool, and he respected science far too much for that. But he could not bring himself to say that across a revolution our theories are converging on a single true description of a framework-independent world, because the historical record showed him frameworks succeeding one another in ways that did not look like convergence on anything, that abandoned old questions and old entities entirely rather than refining the answers to them. He was caught, and he knew it, between a realism he could not fully embrace and an anti-realism he could not fully accept, and he spent his career uncomfortably suspended between them.
It helps to see the role that instruments play in this drama, for they make the abstract point tangible. We like to think of a measuring instrument as a neutral window onto reality, simply reporting what is there. But an instrument embodies a theory; it is built according to a framework’s assumptions and reads out its results in the framework’s categories. A thermometer presupposes a whole theory of heat; a voltmeter presupposes a theory of electricity; the elaborate detectors of modern physics presuppose vast towers of theory, without which their flickering outputs would be meaningless noise. When the framework changes, the meaning of the instrument’s readings can change with it, and sometimes the old instrument becomes simply irrelevant, measuring a quantity the new paradigm no longer recognizes as real. The phlogiston chemist and the oxygen chemist, performing what looks like the same operation in the laboratory, are not gathering the same data, because the very quantities they take themselves to be measuring belong to different and incompatible theoretical worlds. The neutral observation language, the shared bedrock of pure fact, dissolves on inspection into a thousand theory-laden dialects.
A vivid case is the moon as Galileo first saw it through his new telescope. He reported mountains and craters, a rugged, earth-like surface, and to him this was simply what he observed, plain as day. But to many of his learned contemporaries, who inhabited the older framework in which the heavens were the realm of incorruptible perfection, the idea of mountains on the moon was not merely improbable; it was almost unintelligible, and some of them, looking through the very same instrument, could not see the mountains at all, or dismissed them as flaws in the lens, illusions, smudges. Were they simply being stubborn? Partly, no doubt. But Kuhn would insist there was more to it: their trained perception, schooled by a framework that ruled out earthly imperfection in the heavens, genuinely did not organize the play of light and shadow into mountains the way Galileo’s did. The telescope did not hand its users a neutral image that they then interpreted differently. What they saw through it was already shaped by what they expected to be there, and the framework reached down into the seeing itself.
Kuhn offered an even homelier example, one he clearly loved, drawn again from the history of motion. Picture a heavy object swinging back and forth on a string. To an Aristotelian, schooled to see every motion as a body seeking its natural place, this is a case of constrained fall: a heavy thing trying to reach the ground, prevented by the string, laboriously and repeatedly failing, its swinging a kind of frustrated falling that only gradually subsides. To a Galilean, schooled in the new framework, the very same swinging object is a pendulum, an entity that repeats its motion almost indefinitely, that exhibits a beautiful regularity in the relation between its length and the time of its swing, that is a near-perfect example of a conserved periodic motion. Same string, same weight, same swinging. But the Aristotelian sees a failure and the Galilean sees a paradigm of lawful regularity, and because they see different things, they notice different features, ask different questions, and take different measurements. Galileo discovered the laws of the pendulum partly because he was able to see a pendulum where his predecessors had seen only a stone reluctantly falling. The shift in framework did not merely reinterpret a shared observation; it made a new kind of object swim into view.
And yet, and this is the crucial qualification that Kuhn’s harshest critics ignored, none of this amounts to the claim that anything goes, that one framework is as good as another, that science makes no progress and reality places no constraints. The world, whatever it is, does push back. Not every framework survives; not every paradigm can be made to work; the recalcitrance of nature, its stubborn refusal to fit certain pictures, is exactly what generates the anomalies that bring paradigms down. Kuhn never denied this. What he denied was something more specific and more subtle: that the pushing-back of the world comes to us in a pure, framework-free form that can serve as a neutral arbiter between competing paradigms. Reality constrains, but it constrains from within a framework, never from some impossible view from nowhere. This is a defensible and even a sober position, far removed from the giddy relativism his enemies attacked, but it is a hard position to hold steadily, and Kuhn’s own wavering formulations gave his enemies plenty of ammunition.
The question of whether scientists live in different worlds turns out, then, to be less a question about physics than a question about the relationship between mind and reality, the oldest question in philosophy, dressed in the new clothes of the history of science. Kuhn had stumbled, by way of Aristotle and the Copernican revolution and the psychology of perception, into the ancient problem of whether we can ever know reality as it is in itself or only ever reality as it appears through the categories of our own understanding. He was, in this respect, a distant heir of the philosopher Immanuel Kant, who had argued that the mind imposes its own forms on experience and that we can never reach the thing in itself behind the appearances. Kuhn was sometimes called a Kantian with movable categories, a memorable phrase that captures his position well: where Kant thought the categories through which we structure experience were fixed and universal, the same for all rational minds, Kuhn thought they varied from paradigm to paradigm and shifted in revolutions. The forms that structure a scientist’s world are real and powerful, but they are not permanent, and when they change, her world changes with them. To say more precisely what such a change involves, and why it makes communication across a revolution so treacherous, we must now examine the single most notorious term in Kuhn’s entire vocabulary, the word that more than any other was taken to prove him a relativist: incommensurability.
Chapter Fourteen — Incommensurability, or Talking Past Each Other
The word comes from mathematics, where it has an old and precise meaning. Two quantities are incommensurable when they share no common unit of measure, when there is no single ruler, however fine, that can measure both of them in whole numbers. The side of a square and its diagonal are incommensurable in this sense; the ancient Greeks discovered, to their reported horror, that no matter how small a unit you choose, you can never use it to express both lengths exactly. Kuhn borrowed this word and applied it to rival scientific paradigms, and in doing so he detonated a controversy that has not fully subsided to this day. Two paradigms, he claimed, can be incommensurable: there is no common measure, no neutral standard, no shared language in terms of which they can be fully and fairly compared. And if that is so, then the choice between them cannot be the simple matter of measuring each against a common yardstick of evidence that the tidy universe had imagined.
It is essential to be clear about what Kuhn did and did not mean, because few of his terms have been more wildly misunderstood. Incommensurable does not mean incomparable. Kuhn never said that rival paradigms cannot be compared at all, that there is simply no rational basis whatever for preferring one to another. That would indeed be a mad relativism, and he spent years denying it. What he said was subtler: that the comparison cannot be made by means of a single neutral standard that both sides share, point for point, without remainder. The paradigms can still be compared, but the comparison is partial, difficult, and never decisive in the way a measurement against a common ruler would be. There is no algorithm, no mechanical procedure, that takes two paradigms and the evidence as input and delivers the rationally compelled choice as output. The comparison requires judgment, and judgment, unlike measurement, can be exercised well or badly by reasonable people who reach different conclusions.
Kuhn came to locate the heart of incommensurability in language, and specifically in meaning. When a paradigm changes, he argued, the meanings of its central terms change with it, and they change in ways that cannot be captured by simple translation. Consider the word mass. In Newton’s physics, mass is a fixed and conserved quantity, the same whether an object is at rest or in motion, an unchanging measure of how much matter a thing contains. In Einstein’s physics, mass is something different; it varies with velocity, it can be converted into energy and energy into it, it is woven into the very structure of space and time in a way that would have been unintelligible to Newton. The same word, mass, means something genuinely different in the two frameworks, and there is no way to translate Newtonian mass into Einsteinian terms without distortion, because the concepts are knit into different and incompatible webs of meaning. To learn the new physics is not to attach new beliefs to old concepts; it is to acquire new concepts, a new conceptual vocabulary in which the old terms have quietly become false friends, familiar in sound but altered in sense.
This is why, Kuhn observed, the proponents of rival paradigms so often seem to talk past one another, arguing endlessly without ever quite making contact, each unable to understand why the other cannot see what is so obvious. They are, in a real sense, speaking different languages that happen to share much of the same vocabulary, and the shared vocabulary is precisely the trap, for it conceals the depth of the difference and lures each side into thinking the other is simply being stubborn or stupid. When the Newtonian and the Einsteinian both use the word mass, they assume they are talking about the same thing and disagreeing about its properties, when in fact they are using the word to pick out subtly different concepts, and their disagreement is at least partly a failure of translation rather than a clean clash of claims about a shared subject. The communication breakdown across a revolution is not a contingent failure of goodwill or intelligence; it is built into the semantic structure of paradigm change itself.
A simpler and more concrete illustration than mass is the humble word planet, whose meaning has been quietly revolutionized more than once. For Ptolemy, a planet was a wandering light in the sky, one of the bodies that moved against the fixed stars, and by this definition the sun and the moon were planets while the earth, solid and central and stationary, emphatically was not. For Copernicus, the earth became a planet and the sun and moon ceased to be ones, because the category had been redrawn around a new principle: a planet was now a body orbiting the sun. The word survived the revolution, but what it picked out, the very membership of the category, was reshuffled, and a thing that had been a paradigm case of a non-planet, the earth beneath our feet, was suddenly enrolled among the planets while two former members were expelled. To shift from the old usage to the new is not to learn a new fact about planets; it is to learn a new meaning of the word, a new way of carving the heavens into kinds. And, as the recent and much-publicized demotion of Pluto reminds us, these taxonomic boundaries remain live and contestable even now, capable of provoking real passion, because to change a classification is to change, in a small way, the structure of the world one inhabits.
Philosophers of language had been worrying, in Kuhn’s era, about a related and equally vertiginous problem, the problem of what it would take to translate a wholly unknown language from scratch, with no dictionary and no interpreter, working only from observing what its speakers say and do. They had concluded that such radical translation is haunted by a deep indeterminacy: the available evidence never quite pins down a unique correct translation, and different translation schemes, each fitting all the behavior, can disagree about what the foreign words mean. Kuhn’s incommensurability is a cousin of this idea, applied not across human languages but across the conceptual languages of successive paradigms. The historian trying to understand a dead science is in the position of the radical translator, confronting a system of meanings that cannot be smoothly mapped onto her own, forced to reconstruct an alien web of concepts from the inside rather than looking each term up in a ready-made glossary. This is exactly the labor that Kuhn’s own Aristotle experience had demanded of him, and it is why he came to see the history of science as an exercise in translation, and a peculiarly difficult one.
It is here, in the figure of the translator, that Kuhn found his way out of the worst of the relativist trap, though his critics rarely gave him credit for it. If paradigms were strictly untranslatable, sealed off from one another absolutely, then communication across a revolution would be flatly impossible, and the whole history of science would be a series of hermetic worlds that could never make contact. But Kuhn did not believe this, and the existence of historians like himself proved it false. The historian can, with great labor, learn to inhabit a dead paradigm, to think in its categories, to see the world through its eyes; she can become, in effect, bilingual, fluent in both the old framework and the new, able to move between them even though she cannot fuse them into a single neutral super-language. Incommensurability, properly understood, is not the impossibility of understanding the other framework; it is the impossibility of capturing both frameworks at once in a single shared idiom, the impossibility of the neutral point-for-point translation that would let one mechanically compute the better theory. Understanding is possible; it just takes the hard, immersive work of learning a second conceptual language, rather than the easy work of consulting a common dictionary that, Kuhn insisted, does not exist.
This bilingual resolution also clarifies why scientists living through a revolution have such a hard time, harder in a way than the later historian. The historian comes to the old paradigm as a foreign language to be learned afresh, with no stake in it and no rival framework already installed. The working scientist, by contrast, already lives inside one of the two frameworks, has built her whole competence within it, and must somehow learn the other while her native paradigm pulls at her perception and resists the intrusion. She is not a neutral translator but a partisan trying to understand the enemy’s tongue, and the very fluency that makes her expert in her own framework makes the rival one feel, at first, like nonsense. This is why conversion, when it comes, so often comes suddenly and late, and why it so often skips a generation, the young learning the new language natively while their elders struggle and mostly fail to acquire it as a second tongue.
In his later years, as we will see when we follow the older Kuhn into his difficult final phase, he refined this idea considerably, recasting incommensurability as a matter of taxonomy, of the way a science classifies the things in its domain into kinds. Every science sorts the world into categories, draws lines around groups of things and treats them as the same kind for the purposes of theory. A revolution, the later Kuhn argued, is fundamentally a change in this taxonomic structure, a redrawing of the lines that divide the world into kinds, and the trouble is that the new classification cannot always be mapped onto the old one without violating the logical rules that any sane classification must obey. You cannot simply translate the old categories into the new, because the new ones cut the world at different joints, group together things the old scheme held apart and separate things it held together. This is a more precise and more defensible version of incommensurability than the early talk of meaning change, and it shows Kuhn working hard, decade after decade, to articulate exactly what he had glimpsed and to defend it against the charge that it made science irrational.
For that charge was the shadow that incommensurability cast, and it darkened the rest of Kuhn’s life. If rival paradigms are incommensurable, if there is no neutral standard and no common language and no algorithm for choosing between them, then on what basis does a scientist ever rationally prefer one to another? Does the choice not collapse into mere preference, mere taste, mere the persuasion of the crowd? Kuhn insisted, with increasing exasperation, that it does not, that incommensurability makes theory choice difficult and non-algorithmic but not irrational, that scientists choose between incommensurable paradigms by the exercise of trained, communal judgment guided by shared values, and that this is a perfectly rational thing to do even though it cannot be reduced to a rule. Whether this defense succeeds is one of the central questions we will weigh when we turn directly to the accusation of relativism. But first we should look at one of the most practical and least appreciated consequences of the whole theory, the curious way in which the very textbooks that train each new generation of scientists systematically erase the revolutions that produced them, rewriting the past so thoroughly that working scientists come to believe their field has always marched in a straight line toward the present.
Chapter Fifteen — How Textbooks Hide the Revolutions
Open any introductory science textbook and you will find, in its opening pages, a peculiar little ritual: a brief, sanitized history of the field, in which a procession of past geniuses each adds a tidy contribution to a steadily rising structure, building patiently toward the splendid edifice that the rest of the book will lay out. Galileo contributes this, Newton adds that, each error gently corrected by the next advance, the whole forming a smooth and reassuring ascent from ignorance to the present state of knowledge. This little history is almost entirely false, and Kuhn argued that its falseness is not accidental but systematic, indeed nearly inevitable, a structural feature of how science transmits itself to the young. The textbook does not lie out of malice or carelessness. It distorts the past because its purpose is not to teach history but to train practitioners, and the two purposes pull in opposite directions.
The textbook’s job is to induct the student efficiently into the current paradigm, to equip her as quickly as possible with the concepts, techniques, and exemplars she will need to do normal science. For this purpose, the messy actual history of the field is not merely useless but positively harmful. The student does not need to know that the founders of her science believed things now regarded as absurd, that they reached the present framework through a series of convulsions and false starts, that the very concepts she is learning had once to be wrenched into existence against fierce resistance. All of that would only confuse her and slow her down. What she needs is the finished framework, presented as cleanly and authoritatively as possible, with just enough history to give her a sense of belonging to a noble tradition. And so the textbook presents the past through the lens of the present, picking out from the historical record only those bits that look like steps toward the current paradigm, and silently discarding everything that does not fit, the abandoned frameworks, the lost questions, the roads not taken.
The effect, Kuhn argued, is to make every revolution invisible. Because the textbook describes the past entirely in the categories of the present paradigm, it cannot represent the earlier paradigms as the coherent, different worlds they actually were; it can only represent them as imperfect, partial versions of the present one. The phlogiston chemists become merely confused proto-oxygen chemists; the Ptolemaic astronomers become merely inaccurate proto-Copernicans. The genuine discontinuity, the wholesale replacement of one framework by an incommensurable other, is smoothed over into a continuous accumulation, and the student comes away with the firm impression that science has always thought more or less as it thinks now, only with less detail. The revolutions are not denied; they are something subtler, rendered literally unthinkable, because the conceptual apparatus the student is given has no place for them. She inherits a history written by the victors, in the language of the victors, in which the defeated frameworks survive only as the errors that the victors corrected.
Kuhn drew a deliberately provocative comparison that he knew would sting: he likened the textbook’s treatment of the past to the rewriting of history in a totalitarian state, the systematic revision of the record to make the present regime appear as the inevitable and rightful culmination of all that came before. The comparison is arresting, and it contains a real insight, but Kuhn was careful, mostly, to add that the scientific case is different in a crucial respect: the rewriting is not a cynical political imposition but an honest and largely unconscious byproduct of the genuine need to train practitioners efficiently. No one decides to falsify the history of chemistry; the falsification emerges naturally from the textbook’s pedagogical function. Still, the parallel illuminates something important about how communities construct usable pasts, selecting and shaping their history to serve the needs of the present, and it helps explain why working scientists, for all their brilliance, are so often spectacularly unreliable as historians of their own fields. They have absorbed the textbook history, the triumphant straight-line ascent, and they have every professional reason to believe it and none to question it.
Consider how a famous discovery is remembered versus how it happened. The unraveling of the structure of the genetic molecule is told, in the textbooks and the popular legends alike, as a clean and inevitable triumph: two brilliant young men see the double helix and the secret of life is revealed. The actual history, recovered by patient historians, is a tangle of false models, lucky guesses, borrowed and arguably misappropriated data, fierce rivalry, gendered injustice, and a good deal of sheer contingency, in which the winning structure emerged from a messy human process that could easily have gone otherwise. None of that mess survives in the textbook, because the textbook is not trying to tell you what happened; it is trying to teach you the structure and to give you a tidy founding myth to belong to. The myth is more useful for training than the truth would be, and so the myth is what gets transmitted, hardening over the generations into something everyone simply knows. Multiply this single case by every discovery in every field, and you begin to grasp the scale of the systematic forgetting on which a scientific education rests.
The distortion is reinforced from another direction by the memoirs and reminiscences of scientists themselves, who are, after all, human beings constructing a narrative of their own lives. Looking back from the vantage of a completed achievement, a scientist naturally tells the story as a coherent march toward the result she now knows to be correct, smoothing away the wrong turns, the periods of confusion, the ideas she held with conviction and later abandoned. She is not lying; memory itself works this way, reorganizing the past in the light of the present, editing the record into a sensible story with the ending already known. The historian who later examines the actual notebooks and letters often finds a process bearing little resemblance to the scientist’s own later account, full of dead ends and confusions that the memoir has quietly erased. The makers of science are, by and large, no better than its textbooks at preserving the genuine shape of their own past, and for much the same reason: the present reshapes the memory of how it came to be.
It is worth pausing to notice that this forgetting is not confined to revolutions; it swallows the patient labor of normal science just as thoroughly, only more quietly. The thousands of careful workers whose accumulated puzzle-solving actually constitutes the bulk of any mature science vanish almost entirely from the story, which retains only a handful of revolutionary names. The textbook is a hall of heroes, and heroes are few; the vast army of competent practitioners who did the real cumulative work, measuring the constants, determining the structures, extending the framework into a thousand corners, leaves no trace in the popular memory of the field. So the public inherits a doubly distorted picture: revolutions rendered invisible by being smoothed into continuity, and the continuous labor that is most of science rendered invisible by being eclipsed by a few dramatic names. Neither the upheavals nor the patient ordinary work survives the textbook’s editing; what survives is a fairy tale of lone geniuses ascending a ladder, which is almost the opposite of the truth on both counts.
And yet, before we grow too indignant at all this distortion, Kuhn would have us pause and acknowledge that it serves a genuine and even an admirable function. A science that insisted on teaching its full and honest history to every beginner, that required each new student to relive all the confusions and convulsions of the past before being permitted to do any work, would be a science crippled by its own scruples, unable to train practitioners efficiently, condemned to refight its old battles in every generation. The ruthless forgetting, the editing of the past into a clean ascent, is the price a discipline pays for the extraordinary efficiency with which it equips its young and gets them quickly to the frontier. The distortion is, in a sense, the cost of normal science’s power. Kuhn was not denouncing the textbooks; he was diagnosing them, showing how their necessary pedagogical function inevitably produces a false history, and how that false history in turn produces the tidy universe, the smooth cumulative picture of science that he had spent his career trying to correct. The illusion was not a mistake anyone made; it was a structural byproduct of how science reproduces itself, and that, in a way, made it far harder to dislodge than any mere error.
This has a consequence that brings us back to where this whole book began. It explains why Kuhn’s own thesis was so surprising, so resisted, so hard for scientists to accept, even though it was, in a sense, merely an accurate description of what their own history actually showed. The scientists could not see the revolutions in their past because their education had been carefully designed to render those revolutions invisible. They had been trained, quite literally, not to see the thing Kuhn was pointing at. When he told them that their field had been transformed by upheavals as radical as any political revolution, they looked back at the smooth ascending curve of their textbook history and saw no such thing, and they concluded that Kuhn must be exaggerating, or confused, or making mischief. It took the special perspective of the historian, the outsider trained to read the past on its own terms rather than through the categories of the present, to see what the textbooks had hidden. Kuhn had become a historian almost by accident, that long-ago afternoon with Aristotle, and it was precisely his historian’s eyes that let him see what the scientists, for all their genius, had been educated not to notice. The tidy universe was, in part, a creature of the textbooks, an artifact of the very process by which science trains its young, and it took someone standing outside that process to recognize how thoroughly it had reshaped the past in the image of the present. With this, the machinery of Kuhn’s theory is fully before us, and we are ready to confront the great unanswered question it raises: if paradigms are chosen without a neutral standard, by something like conversion, then in what sense, if any, does science make progress, and how does a scientist ever rationally choose at all?
Chapter Sixteen — Choosing Theories Without a Rulebook
Here is a small scandal that the tidy picture of science cannot easily explain. Take two scientists of the first rank, equally brilliant, equally honest, equally well-informed, equally committed to the evidence. Set before them the same data, the same two competing theories, the same body of experimental results. Ask each, independently, which theory the evidence favors. And watch as they reach opposite conclusions, each convinced that the facts plainly support her choice, each baffled that so able a colleague could read the same record so differently. This happens constantly in real science, especially during the contested periods when paradigms compete, and it ought to be impossible if science worked the way the textbooks say, if the evidence simply dictated the rational conclusion to any competent mind. Kuhn took this scandal seriously, and his attempt to explain it produced what may be the most careful and most rescuing part of his whole philosophy, the part that answers the charge of irrationalism most directly.
His answer appeared most clearly in a lecture he gave with the deceptively dry title Objectivity, Value Judgment, and Theory Choice, and it began by conceding to his opponents exactly what they demanded, then turning the concession into something they had not expected. Yes, Kuhn agreed, there are shared standards by which scientists judge theories; he was not claiming that scientists choose arbitrarily or by whim. He could even list the standards, and he did. A good theory should be accurate, agreeing with the results of observation and experiment. It should be consistent, both internally and with other accepted theories. It should have broad scope, reaching beyond the particular things it was first designed to explain. It should be simple, bringing order to phenomena that would otherwise be tangled and confused. And it should be fruitful, disclosing new phenomena and new relationships that no one had anticipated. These five virtues, Kuhn proposed, are shared across the scientific community and across its revolutions; they are the relatively stable values by which scientists everywhere assess their theories.
So far this sounds like a complete vindication of the rationalist picture: here are the objective criteria, here is the rulebook after all. But then Kuhn sprang his trap, and the trap is subtle and devastating. These five virtues, he pointed out, do not function as rules that determine a unique choice. They function as values that influence choice without dictating it, and they fail to dictate for two distinct and equally important reasons. The first reason is that the virtues are imprecise in application. Two scientists who both prize simplicity may disagree about which of two theories is actually simpler, because simplicity can be measured in different ways and a theory simple in one respect may be baroque in another. Accuracy itself is rarely clean; a theory may fit one body of data beautifully and another poorly, and how one weighs these partial successes is a matter of judgment. The values are real, but they do not come with instructions for their own precise application, and reasonable people apply them differently.
The second reason cuts even deeper. The five virtues frequently conflict with one another, and when they do, nothing in the list tells you how to rank them. A new theory may be more fruitful but less accurate than its rival, opening exciting new territory while fitting the existing data slightly worse. It may be simpler in its foundations but broader and messier in its consequences. When the Copernican system first appeared, it arguably scored higher on a certain kind of harmony but no higher, perhaps lower, on raw accuracy than the Ptolemaic system it challenged. Faced with such trade-offs, a scientist must decide how much weight to give each virtue, and there is no further rule, no meta-criterion, that fixes the proper weighting. One scientist, temperamentally drawn to fruitfulness and bold new directions, leans toward the promising newcomer. Another, prizing accuracy and consistency with established results, holds to the proven incumbent. Both are applying the shared values; both are being entirely rational; and they reach opposite conclusions because they weight the values differently.
This is the heart of Kuhn’s rescue operation, and it is worth stating its upshot plainly because so many readers missed it. Theory choice in science is neither arbitrary nor algorithmic. It is not arbitrary, because it is genuinely governed by shared objective values that constrain what any reasonable scientist can think; a scientist who simply ignored accuracy and scope would be drummed out of the profession. But it is not algorithmic either, because those values underdetermine the choice, leaving room for the trained judgment of individuals to tip the balance one way or the other. Theory choice is a matter of reasoned judgment under shared values, the kind of judgment that an experienced practitioner exercises well and a novice exercises badly, the kind that can be argued about and defended with reasons even though it cannot be reduced to a calculation. This is exactly the structure of judgment in many domains we rightly regard as rational, from the verdict of a wise judge to the diagnosis of a skilled physician, none of which follows mechanically from rules and all of which can nonetheless be done better or worse.
Kuhn even argued that this looseness, far from being a defect in science, is a positive necessity, without which science could not function. Suppose, contrary to fact, that there were a rigid algorithm that compelled every competent scientist to the same conclusion the moment a new theory appeared. Then, at the first signs of crisis, either everyone would instantly abandon the old paradigm together, or no one would, and in either case the community would lurch as a single block. But this would be disastrous, because new paradigms are always weak and incomplete when they first appear, riddled with unsolved problems, unable yet to match the achievements of the framework they hope to replace. A community that abandoned the old paradigm the instant a promising rival appeared would be perpetually chasing half-formed novelties, never staying with any framework long enough to develop its power. What the community needs is to hedge its bets, to have some members boldly developing the risky newcomer while others doggedly extend the proven incumbent, so that whichever way the truth lies, someone is exploring it. The very variation in how individuals weight the shared values, the very thing that makes theory choice non-algorithmic, is what distributes the community’s effort sensibly across the old and the new during the dangerous transition.
There is a lovely irony in this, which Kuhn appreciated. The feature of science that looked, to his critics, like a fatal slide into subjectivity, the fact that equally rational scientists reach different conclusions, turns out on his analysis to be a precondition of scientific rationality at the level of the community. The individual variation that seems like noise is actually a vital mechanism, a way of spreading risk that no rulebook could provide. A community of identical reasoners, all applying the same algorithm, would be brittle and foolish; a community of varied reasoners, applying shared values with individually different weightings, is resilient and wise, able to pursue multiple paths at once and to keep faith with a struggling new framework long enough to find out whether it will bloom. Rationality, Kuhn was suggesting, lives not only in the individual scientist but in the community, in the way its internal diversity allows it to do collectively what no individual following rules could do.
History furnishes a vivid case of these virtues pulling in opposite directions. For much of the nineteenth century, light was understood as a wave, and the wave theory enjoyed spectacular successes, explaining the delicate patterns of interference and diffraction that the older particle picture could not touch. Yet when the new century arrived, certain stubborn phenomena, the way light knocks electrons out of metal, refused to fit the wave picture and yielded only to the idea that light comes in discrete packets, like particles after all. For a time, physicists faced the maddening situation of two incompatible pictures, each accurate within its own domain, each failing in the other’s. Which to prefer? The wave theory had breadth and a long record of consistency; the particle idea had a troubling novelty but a fruitfulness that opened the whole new world of the quantum. A physicist’s choice in that unsettled interval depended on which virtues she prized and which phenomena she took as central, and reasonable physicists divided. The resolution, when it eventually came, was stranger than either party imagined, a framework in which light is somehow both at once; but for years the community lived in exactly the condition Kuhn described, applying shared values that pointed, frustratingly, in different directions.
This account also makes sense of something that the tidy picture had treated as faintly disreputable: the role of persuasion in science. If theory choice were algorithmic, persuasion would have no honest place; one would simply present the proof and any rational mind would be compelled. But because the shared values underdetermine the choice, the advocate of a new paradigm must do more than display the evidence; she must persuade her colleagues to weight the virtues as she does, to see the promise she sees, to give the struggling newcomer the benefit of the doubt long enough for its fruitfulness to show. This persuasion is not a corruption of scientific rationality but a legitimate part of it, the means by which a community talks itself, argument by argument and example by example, toward a new consensus. Kuhn insisted that the reasons deployed in such persuasion are good reasons, real appeals to accuracy and scope and simplicity and fruitfulness, not mere rhetoric or trickery. But they are reasons that persuade rather than compel, that move a judgment rather than force a conclusion, and the difference between persuading and compelling is the whole difference between Kuhn’s picture of science and the tidy universe he displaced.
Whether this rescue fully succeeds is a question we will keep returning to, for the suspicion lingers that Kuhn had merely relocated the problem rather than solving it. If the weighting of the values is not itself governed by any standard, if it comes down to individual temperament and the persuasion of one’s peers, then has he really escaped the charge that paradigm choice is, at bottom, a matter of taste and rhetoric rather than truth? Kuhn would say no: temperament and persuasion operate within the constraints of shared values and a recalcitrant world, and that is enough for rationality even if it is not enough for the algorithmic certainty the tidy universe craved. His critics would say he had given the game away, that values without fixed weights are no real constraint at all. The dispute turns on how much rationality requires, on whether reasoned judgment under shared but underdetermining values deserves the name of rationality or only its appearance. It is among the genuinely open questions in the philosophy of science, and we will not pretend to close it here. But it leads directly to an even larger question, the one that Kuhn found hardest of all and that troubled him to the end of his life. If theory choice is reasoned judgment rather than mechanical proof, and if successive paradigms are incommensurable rather than cumulative, then in what sense does all this choosing add up to progress? Is science going anywhere at all, or merely changing?
Chapter Seventeen — Progress That Is Not Going Anywhere
Of all the comfortable assumptions Kuhn disturbed, none was dearer than the assumption that science makes progress, and none did he handle with more evident discomfort. That science progresses seems beyond question; the whole spectacle of modern technology, the conquest of diseases, the reach of our instruments into the atom and out to the edge of the cosmos, all of it testifies that science gets somewhere, that we know vastly more than our ancestors. Kuhn did not deny any of this. What he questioned was the natural gloss we put upon it, the assumption that progress means getting closer and closer to a final truth, to a complete and accurate picture of reality as it is in itself. He suspected that this gloss, however irresistible, might be a confusion, and in trying to articulate why, he reached for an analogy that shocked many readers and that he came to regard as one of his most important contributions: the analogy with biological evolution.
Before Darwin, most people who thought about the development of life assumed it must be going somewhere, ascending toward some goal, some perfect or highest form toward which the whole pageant of living things was striving, with humanity, naturally, at or near the summit. Darwin’s great and disturbing achievement was to explain the entire magnificent history of life without any goal at all. Evolution, in his account, is not a climb toward a predetermined peak; it is a process of variation and selection that pushes life away from where it has been, adapting it ever more intricately to its circumstances, without aiming at anything, without any destination written in advance. The eye was not the goal toward which evolution was working; it is simply what selection produced, again and again, because seeing happened to help. There is progress, in the sense of increasing adaptation and complexity and diversity, but it is progress from a starting point, not progress toward a finish line. It is a process that moves, magnificently, without going anywhere in particular.
This, Kuhn proposed, is exactly how we should understand the progress of science. We should stop thinking of science as moving toward a final true theory of the world, a fixed target it is gradually approaching, and start thinking of it as moving away from less adequate frameworks, becoming ever better adapted to the problems it confronts, ever more refined and powerful and detailed, without any of this implying convergence on a single predetermined truth. Science progresses from primitive beginnings, not toward an ultimate goal. Each paradigm is better than its predecessor at solving the problems the community cares about, just as each evolved form is better adapted than its ancestors to its environment; but just as there is no perfect organism toward which evolution aims, there may be no final true theory toward which science aims. The progress is real, but it is the progress of increasing fitness, not of approaching a destination, and to insist on the destination may be to import a comforting metaphysical assumption that the actual history of science does nothing to support.
Why was Kuhn so reluctant to say the simple, satisfying thing, that science gets closer to the truth? His reluctance came from the historical record he knew so well, which refused to show the smooth convergence the simple story required. When he looked at successive paradigms, he did not always see the later one preserving and extending the achievements of the earlier, the way convergence would demand. He saw, instead, frameworks that abandoned old questions entirely, that ceased to recognize entities the previous framework had taken as fundamental, that sometimes lost the ability to explain things an earlier paradigm had explained well. We will examine this disturbing phenomenon, the actual loss of knowledge across revolutions, in its own chapter. For now the point is that this record does not look like a series of better and better approximations to a single fixed truth; it looks more like a succession of differently adapted frameworks, each excellent for its own purposes, related to one another not as rough and refined versions of the same picture but as genuinely different pictures, incommensurable in the way we have explored. And if that is what the record shows, then the language of convergence on truth may simply not fit it.
There was also a deeper, more philosophical worry beneath Kuhn’s reluctance, and he stated it with disarming honesty. To say that science is getting closer to the truth, he observed, presupposes that there is a definite truth, a way the world really is in itself, independent of all paradigms, that our theories can match more or less well. But this framework-independent reality, the world as it is apart from any way of carving it up, is precisely what we can never get at, for we always encounter the world through some paradigm or other, never raw. The notion of a match between our theories and the unconceptualized world, Kuhn suspected, might be not so much false as meaningless, a comparison we can never actually perform because we have no access to one of its terms. He did not insist on this; he was genuinely uncertain, and he wavered. But the worry was enough to make him recoil from the easy formula that science approaches the truth, and to send him searching for a way to describe scientific progress that did not depend on a comparison he feared might be empty.
A nagging difficulty shadows even Kuhn’s own preferred way of describing progress, and he was honest enough to feel it. He wanted to say that each paradigm solves more problems, or more important problems, than its predecessor, so that progress could be measured in problem-solving power even if not in approach to truth. But here incommensurability returns to bite him. If rival paradigms carve the world differently, ask different questions, and recognize different phenomena as significant, then how do you even count the problems each one solves, let alone compare the tallies? A new paradigm does not simply solve all the old problems plus some new ones; it typically discards some of the old problems as misconceived, declares some of the old questions meaningless, and raises new questions the old framework never dreamed of. There is no neutral ledger, no common currency of problems, in which the achievements of incommensurable paradigms can be added up and compared. Kuhn’s measure of progress threatens to dissolve in the very incommensurability that drove him away from the measure of truth, and he never found a wholly satisfying way around this, which is part of why his account of progress remained, by his own admission, the most unfinished corner of his thought.
The evolutionary analogy, for all its power, carries a buried question that is worth dragging into the light, because answering it shows both the strength and the limit of Kuhn’s picture. Biological evolution has a mechanism: natural selection, the differential survival and reproduction of variants in an environment. If scientific change is evolutionary, what plays the role of selection? What is the environment, and what does the selecting? Kuhn’s answer, developed more explicitly in his later years, was that the selecting is done by the scientific community itself, through its judgments of which puzzle-solutions work, which frameworks prove fruitful, which problems matter. The community, applying its shared values to a recalcitrant world, is the environment in which theories struggle to survive, and the theories that flourish are those best adapted to satisfy the community’s standards while withstanding the world’s resistance. This is an elegant completion of the analogy, but notice what it implies: the direction of scientific progress is set partly by the values and interests of the community doing the selecting, not by the world alone. The world constrains, but it does not dictate; the community’s priorities help steer. This is a long way from the tidy universe’s vision of nature simply imposing the one true theory on any rational observer, and it returns us, by yet another road, to the deep question of how much of science is discovery and how much is, in a sense that need not be cynical, construction.
It must be stressed, because it is so often missed, that none of this made Kuhn a sceptic about science or a denier of its achievements. He was not saying that one theory is as good as another, that science makes no progress, that it is all a matter of fashion. He was saying something far more specific and far more interesting: that the progress is real but its character has been misdescribed, that we have wrapped the genuine and undeniable advance of science in a metaphysical story about converging on absolute truth that the advance itself does not require and the history does not support. You can have all the progress, all the increasing power and precision and scope, without the metaphysics of the final true theory. The evolutionary analogy was his way of showing how: just as life can advance gloriously without aiming at a goal, so science can advance gloriously without approaching a fixed truth. The progress is in the leaving-behind, not in the arriving.
Many found this unbearable, and the resistance is understandable, for the evolutionary picture asks us to give up something we cherish: the sense that our science is not merely useful and powerful but right, that it is latching onto the way things actually are. Kuhn was asking us to hold our theories more lightly, to see even our finest current frameworks as adaptations that future revolutions may replace with others differently adapted, rather than as the final truth at last attained. This is a humbling vision, and a vertiginous one, and it is small wonder that scientists raised on the triumphant narrative of approaching truth found it hard to swallow. Whether Kuhn was right, whether science really does converge on a framework-independent reality or merely adapts itself ever better without any destination, remains one of the great open questions, fiercely debated by philosophers to this day under the heading of scientific realism. Kuhn did not settle it. What he did was to make it impossible, for anyone who had really understood him, to assume the comforting answer without argument. And it was precisely this refusal of the comforting answer, this suggestion that science might be going nowhere in particular even as it advanced, that brought down upon him the wrath of the one contemporary who cared most fiercely about defending the rationality of science, and who became his greatest and most revealing antagonist.
Chapter Eighteen — The Duel: Kuhn Against Popper
In July of 1965, in a lecture hall at Bedford College in London, the philosophy of science staged its most famous confrontation. A large international colloquium had been convened, and at its center sat the two men whose rival visions of science would define the field for a generation: Karl Popper, the formidable Viennese champion of falsification and critical reason, by then a knighted and lionized figure, and Thomas Kuhn, the quieter American historian whose slim book was already reshaping how the world thought about science. The proceedings were later published under a title that captured the stakes with unusual drama for an academic volume: it asked, in effect, whether the growth of knowledge proceeded by criticism or by revolution. Everyone present understood that something important was being contested, that two fundamentally opposed pictures of the scientific enterprise had been brought face to face, and that they could not both be right.
On the surface, Popper and Kuhn agreed about a surprising amount, and it is worth marking the common ground before the chasm. Both rejected the naive picture of science as the simple accumulation of confirmed facts. Both insisted that observation is shaped by theory, that there is no pure innocent seeing. Both saw the history of science as punctuated by dramatic theoretical change rather than smooth growth. Both took the actual practice of science seriously as a subject for philosophy. They were not opposites in every respect; they were near neighbors who disagreed all the more fiercely for the closeness of their positions, in the way that the bitterest quarrels are so often among those who share most. But beneath the shared ground lay a disagreement so deep that it touched the very definition of science, and it crystallized around a single concept: Kuhn’s notion of normal science.
For Popper, normal science was not merely uninteresting; it was a danger, almost a betrayal. Recall that for Popper the essence of the scientific spirit was relentless criticism, the perpetual readiness to challenge and test and try to overthrow even one’s most cherished theories. The scientist, in Popper’s heroic image, is a permanent revolutionary, never resting in any framework, always probing for the weak point, always willing to abandon a theory the moment it fails a genuine test. Now along came Kuhn, describing the ordinary working scientist as someone who does precisely the opposite: who takes her framework for granted, who does not question fundamentals, who treats anomalies not as refutations to be welcomed but as puzzles to be solved or, failing that, set aside, who is, in short, a dogmatist devoted to defending and extending an unquestioned paradigm. And Kuhn was saying not that this was a regrettable failing of lesser scientists but that it was the normal and necessary condition of science, the very thing that made science productive. To Popper this was horrifying. The normal scientist, as Kuhn described her, was for Popper not a good scientist at all but a person who had given up the critical attitude that was the whole point, a technician, an applied puzzle-solver, someone who had stopped doing science in the deepest sense and started merely operating within it.
Popper made the point with a memorable edge. The normal scientist, he suggested, was a person who had been badly taught, taught to accept a framework uncritically rather than to challenge it, and he found something almost morally objectionable in Kuhn’s calm acceptance of this dogmatism as the natural state of science. For Popper, who had lived through the rise of totalitarian ideologies in Vienna and had written passionately against the closed societies and closed minds that sustained them, the uncritical acceptance of a framework was not a neutral technical matter; it carried echoes of the authoritarian habits of mind he had spent his life combating. To describe science, the great exemplar of open critical rationality, as fundamentally dogmatic, as resting on the unquestioning commitment of its practitioners to an unexamined paradigm, struck Popper as not merely mistaken but almost a slander upon the noblest of human enterprises.
Kuhn’s reply was patient and, in its way, devastating, because it rested on the one thing Popper could not easily counter: the historical record. Popper, Kuhn suggested, had described not science as it actually is but science as he wished it to be, an idealized portrait of perpetual revolution that bore little resemblance to what scientists really do day to day. If scientists actually behaved as Popper demanded, abandoning their frameworks at the first whiff of a contradictory result, science would never get anywhere, because every framework faces anomalies from the start, and a discipline in permanent revolution would be a discipline that never settled down long enough to develop the deep, detailed, cumulative knowledge that is science’s glory. Popper had mistaken the rare and dramatic moments of revolution for the whole of science, when in truth they are the exception, made possible only by the long stretches of dogmatic normal science that Popper despised. Far from being a betrayal of science, Kuhn argued, normal-scientific dogmatism is the engine that drives science forward, the disciplined commitment that allows a community to push a framework to its limits and thereby, eventually, to discover the very anomalies that will overthrow it.
There was a further, sharper thrust in Kuhn’s reply, aimed at the heart of Popper’s own criterion. Popper had made falsifiability the mark of science, the line dividing real science from pseudo-science. But Kuhn pointed out that by this criterion, normal science, which is most of science, would scarcely count as science at all, since the normal scientist precisely does not try to falsify her paradigm and does not treat anomalies as falsifications. And worse, the moments Popper most admired, the great revolutionary overthrows, do not actually proceed by clean falsification either, as we have seen at length: paradigms are not abandoned because a crucial experiment refutes them, but through the messy, judgment-laden, generational process of crisis and conversion. Popper’s falsificationism, Kuhn was suggesting, described neither the normal science it dismissed nor the revolutionary science it celebrated; it was a philosopher’s ideal that matched the actual history of science at almost no point. This was a grave charge, and the debate it opened was never fully resolved between the two men.
It is worth knowing that the 1965 colloquium was not a private duel but a crowded stage, and the supporting cast would shape the next act of our story. Gathered in that London hall were several of the sharpest minds in the philosophy of science, among them two younger thinkers who had studied under Popper and would soon stake out their own positions in the space the debate had opened: Imre Lakatos, who would try to build a bridge between Popper’s rationalism and Kuhn’s historicism, and Paul Feyerabend, who would gleefully blow up the bridge and everything around it. The confrontation between Kuhn and Popper was thus less an ending than a beginning, the opening of a broad and many-sided argument about the rationality of science that would occupy the field for the next two decades. The volume that emerged from the colloquium became one of the most influential collections in the history of the discipline, precisely because it captured a moment when the old certainties had cracked and the new questions had not yet found their answers.
Popper, for his part, had his own answer to the problem of progress that so troubled Kuhn, and the contrast between their answers is illuminating. Where Kuhn flinched from saying that science approaches the truth, Popper embraced a version of exactly that idea, under the technical name of verisimilitude, or truthlikeness. Science progresses, Popper held, because successive theories are closer to the truth than their predecessors, even though we can never know with certainty that we have reached the truth itself; each bold conjecture that survives our fiercest attempts at refutation is, we may reasonably hope, a better approximation to reality than the one it replaced. This preserved the comforting picture of convergence that Kuhn had abandoned, and it is one reason Popper’s philosophy felt safer to many scientists. Unfortunately for Popper, the technical notion of verisimilitude ran into serious logical difficulties that philosophers exposed in the years that followed, and it never achieved the rigor he had hoped for. The disagreement over progress thus remained genuinely unresolved: Kuhn could not say clearly how science progresses without converging on truth, and Popper could not say clearly what it would mean to converge on truth. Each had put his finger on a real difficulty in the other’s view, and neither had a fully satisfying account of his own.
There was, finally, an irreducibly personal dimension to the duel, a clash of temperaments that no summary of their arguments should omit. Popper was combative, magisterial, supremely confident, a man who brooked little dissent and was famous, despite his philosophy of openness to criticism, for his impatience with critics; a wit of the period quipped that the author of The Open Society was running a rather closed one in his own seminar. Kuhn was anxious, careful, perpetually qualifying, a man who revised his views, worried at his own formulations, and seemed almost pained by the controversy he had unleashed. The contrast in style shaped the contrast in reception: Popper’s bracing certainties appealed to those who wanted science defended as the citadel of reason, while Kuhn’s troubled subtleties appealed to those who suspected that the truth about science was messier than any citadel. In the end, it was Kuhn’s picture that captured the wider culture, perhaps because it rang truer to the lived experience of science, perhaps merely because paradigm shift was a better slogan than critical rationalism. But Popper’s challenge never went away, and the question he pressed, whether Kuhn had saved a realistic account of science or quietly surrendered it, remained alive in the work of the man who tried hardest to have it both ways.
The duel ended without a victor, as such duels do, and its real significance lies less in who won than in what it revealed about the choice the two men forced upon their field. Popper stood for a vision in which science is defined by a method, a logic of criticism that prescribes how the rational scientist ought to behave, and in which the philosopher’s task is to articulate and defend that method against the messy failings of actual practice. Kuhn stood for a vision in which science is defined by what scientists actually do, in which the philosopher’s task is to understand and describe the real historical process rather than to legislate an ideal, and in which the gap between the idealized logic and the actual practice is evidence against the logic, not against the practice. Behind their quarrel about normal science lay this deeper quarrel about authority: does philosophy tell science how it must reason, or does the history of science tell philosophy how reasoning actually works? It is the same fault line we traced back to the very origins of Kuhn’s thought, between the top-down ideal of the philosophers and the bottom-up record of the historians, and the Kuhn-Popper debate was its sharpest and most public eruption. Popper was not, however, the only one who rose to defend the rationality of science against what he saw as Kuhn’s dangerous historicism. A brilliant student of Popper’s would soon attempt a far more sophisticated rescue, one that tried to keep what was true in Kuhn while saving reason from the abyss, and it is to that ingenious effort, and to its even more radical and unruly counterpart, that we turn next.
Chapter Nineteen — Lakatos and the Defense of Reason
If Kuhn was right that paradigms are chosen without a neutral algorithm, through something like conversion, then a dreadful question presents itself, and one man felt its force more keenly than anyone. The question is this: how, after Kuhn, can we any longer tell good science from bad? If a community can simply commit to a framework and defend it against all anomalies, treating every contrary result as a mere puzzle, then what stops a pseudoscience from doing exactly the same? What distinguishes the stubborn defense of a productive paradigm from the stubborn defense of a crackpot theory, if both consist of clinging to a framework and explaining away the difficulties? Kuhn seemed to have dissolved the very distinction between rational and irrational science that Popper had labored to draw, and to a certain Hungarian philosopher this was intolerable, a surrender of reason that had to be reversed.
The philosopher was Imre Lakatos, and he was one of the most colorful figures in our story, a man whose life had been a paradigm shift of its own. Born in Hungary, he had survived the Nazi occupation, become a committed communist, risen in the postwar regime, then fallen foul of it and spent years in a Stalinist prison before fleeing west after the failed uprising of 1956. He arrived in England with a ferocious intelligence, a taste for combat, and a deep personal knowledge of what it means for a system of thought to defend itself against all evidence, having watched an ideology do exactly that. He became Popper’s colleague and, in a sense, his heir, but an heir determined to rebuild the inheritance to withstand the Kuhnian earthquake. Lakatos set himself the task of saving the rationality of science, of showing that there really is a difference between good and bad science, while absorbing what was undeniably true in Kuhn’s historical picture.
His central innovation was to shift the unit of analysis. Popper had focused on individual theories, asking whether each could be falsified. Kuhn had focused on paradigms, the whole frameworks within which scientists work. Lakatos proposed something in between, which he called the research programme, and the concept is genuinely illuminating. A research programme, in his account, has two parts. At its center sits a hard core, a set of fundamental assumptions that the practitioners treat as beyond question, that they refuse to abandon, that define the programme’s very identity. Around this core they build a protective belt of auxiliary hypotheses, the adjustable assumptions that connect the core to observation and that can be modified, added to, and reshaped as needed. When an anomaly arises, the practitioners do not blame the hard core; they adjust the protective belt, exactly as Kuhn’s normal scientists adjust their assumptions to absorb a recalcitrant result. So far this simply restates, in new vocabulary, the Kuhnian insight that frameworks are defended against anomalies rather than refuted by them.
But here Lakatos made his decisive move, the move designed to rescue rationality. There is, he argued, a crucial difference between two ways of adjusting the protective belt, and this difference is the long-sought line between good science and bad. Sometimes the adjustments are progressive: they not only patch the anomaly but predict new and unexpected phenomena, open new territory, lead to fresh discoveries. The programme grows, reaches beyond its original problems, anticipates facts no one had foreseen. This is a progressive research programme, and it is the mark of healthy, rational science. At other times the adjustments are degenerating: they are mere patches, contrived solely to save the core from refutation, predicting nothing new, explaining only what they were specifically rigged to explain, the programme forever retreating and shoring up its defenses without ever advancing. This is a degenerating research programme, and it is the mark of bad science, of a framework in its death throes, clinging to life by ever more desperate and barren contrivances.
With this distinction, Lakatos believed he had recovered everything that mattered. He could grant Kuhn that scientists defend their frameworks against anomalies, that there is no instant falsification, that commitment to a core is normal and even necessary. Yet he could still draw the line Popper had wanted, between the rational and the irrational, only now it ran not between theories that are falsified and theories that are not, but between programmes that progress and programmes that degenerate. A scientist rationally sticks with a progressive programme even through its anomalies, because it is bearing fruit; she rationally abandons a degenerating one, because it has ceased to advance. The pseudosciences, on this view, are not distinguished by being unfalsifiable in Popper’s simple sense, but by being chronically degenerating, forever patching and never predicting, parasitic on a reality they cannot anticipate. Astrology fails not because it makes no predictions but because, as a programme, it has degenerated, producing no new knowledge for centuries while real sciences bloomed around it.
A pair of contrasting examples makes the distinction vivid. Recall the discovery of Neptune, which we met earlier: when the planet Uranus misbehaved, the Newtonian programme responded not with a mere face-saving patch but with a bold new prediction, the existence and location of an unseen planet, which was then triumphantly found. That is a progressive adjustment in Lakatos’s sense, a modification of the protective belt that reached beyond the original anomaly and delivered a startling new fact. Contrast this with the way certain doctrines respond to every difficulty by absorbing it after the fact without ever sticking their necks out: whatever happens is explained, but nothing new is ever predicted, the theory forever trailing behind events, accounting for them only once they have occurred. Lakatos regarded the grander pretensions of psychoanalysis and of certain versions of Marxism in just this light, not as obviously false but as degenerating programmes, endlessly flexible, endlessly accommodating, and precisely for that reason barren, generating no surprising predictions that might have shown them to be in genuine contact with a reality they did not already know. The difference is not that one camp adjusts its assumptions and the other does not; both adjust. The difference is whether the adjustments lead anywhere new.
There is real beauty in Lakatos’s construction, and real insight; the distinction between progressive and degenerating programmes captures something true about the difference between living and dying science, and it remains a useful tool of judgment to this day. But it purchased its rationality at a price, and the price became apparent when philosophers pressed on the details. The trouble is timing. Lakatos’s criterion can tell you, in retrospect, that a programme was progressing or degenerating over some past stretch. But it cannot tell you, at the moment of decision, what to do now, because a programme that is degenerating today may stage a spectacular recovery tomorrow, and one that is progressing today may stall and rot. The history of science is full of programmes that everyone had written off as degenerate, only to see them roar back to life when some new idea revived them. So how long should a scientist stick with a temporarily degenerating programme before abandoning it? Lakatos’s honest answer was that there is no rule; one may rationally persist for a long time, hoping for a recovery, or rationally jump ship, and only history will tell who was wise.
And with that admission, the rabbit Lakatos had pulled from the hat began, quietly, to climb back in. For if the criterion cannot guide decision at the moment of decision, if it can only render verdicts long after the fact, then it does not actually do the job Popper wanted, which was to distinguish rational from irrational choices as they are being made. Lakatos had given us a magnificent tool for writing the history of science, for narrating in retrospect why some programmes triumphed and others failed, but not a rulebook for the working scientist facing a live decision under uncertainty. At the sharp end, where the choice is actually made, the scientist is back in Kuhn’s position, exercising judgment under shared values without an algorithm to compel her. Critics, Feyerabend foremost among them, pounced on this with glee, charging that Lakatos’s rationality was a sham, a rationality that conveniently appeared only after the dust had settled and could never be caught in the act of actually guiding anyone.
Lakatos felt this thrust keenly, the more so because it came from his closest intellectual friend, for he and Feyerabend, despite their opposite conclusions, were devoted companions who planned to write a book together arguing out their differences, a book Lakatos’s sudden death in 1974 left forever unfinished. The friendship is itself a kind of parable. Here were two men who agreed almost entirely on the facts, on the historical record of how science actually behaves, on the failure of the simple Popperian and positivist pictures, and who drew from these shared premises diametrically opposed conclusions: one straining every nerve to preserve a defensible rationality, the other cheerfully concluding that rationality, in any rule-bound sense, was a myth that science had always quietly ignored. Their quarrel was the quarrel within every honest reader of Kuhn, the quarrel between the hope that science is, at bottom, a rational enterprise and the suspicion that this hope rests on a picture the history will not support.
Lakatos stands, then, as the most determined and most sophisticated of those who tried to have it both ways, to keep Kuhn’s historical realism about how science behaves while preserving Popper’s normative conviction that science is the paragon of rationality. His research programmes remain a permanent contribution, a genuinely better way of thinking about how scientific traditions grow and decay than either Popper’s lonely theories or Kuhn’s holistic paradigms. But whether he succeeded in the deeper task, whether he truly saved rationality or merely postponed its surrender to the verdict of history, is a question on which thoughtful people still divide. What is certain is that he could not lay the Kuhnian ghost to rest, could not restore the clean algorithmic rationality the tidy universe had promised, and that the gap he could not close was precisely the gap into which his brilliant, infuriating friend would now step, not to mourn the loss of method but to dance on its grave.
Chapter Twenty — Feyerabend and the Glorious Anarchy
There is a sentence that has done more to scandalize the respectable philosophy of science than any other, and it belongs to Paul Feyerabend: the only principle that does not inhibit progress, he wrote, is anything goes. With those two words he announced himself as the enfant terrible of the field, the man who would take the unsettling implications that Kuhn had handled so anxiously and Lakatos had labored so hard to contain, and embrace them with wicked delight. Where Kuhn worried that he might have made science seem irrational and spent his life denying it, Feyerabend looked at the same conclusions and said, in effect, yes, exactly, and isn’t it wonderful. He was the philosopher as provocateur, as court jester, as demolition expert, and beneath the showmanship lay a serious and genuinely radical argument that deserves to be taken seriously even by those it appalls.
Feyerabend was an Austrian who had come up through the same Popperian circles as Lakatos, a man of dazzling range and restless brilliance who had studied physics and theater and seemed to regard the whole solemn enterprise of academic philosophy as a kind of performance to be subverted from within. His central work, provocatively titled Against Method, set out to demolish the idea that there is any such thing as the scientific method, any set of rules that scientists follow and that account for science’s success. He did not argue this from the armchair. He argued it, as Kuhn had taught the field to argue, from history, by examining the actual episodes that the textbooks held up as triumphs of scientific method and showing that, in every case, the great scientists had succeeded precisely by breaking the rules, by violating the supposed canons of method, by using propaganda, rhetoric, trickery, and sheer stubbornness to push ideas that the evidence of their day did not support.
His favorite example was the one Kuhn too had cherished: Galileo. The textbook legend presents Galileo as the hero of empirical method, the man who looked through his telescope and let the evidence speak. Feyerabend told a different and more mischievous story. Galileo, he argued, had no good reason, by the standards of his time, to trust what he saw through his crude new instrument; the telescope’s reliability for celestial observation was itself unproven and doubtful, and there were genuine, intelligent objections to the Copernican view that Galileo could not actually answer. Galileo won not because the evidence compelled assent but because he was a brilliant propagandist, a master of persuasion who wrote in vivid Italian rather than scholarly Latin, who ridiculed his opponents, who deployed rhetorical tricks and selective arguments and sheer charisma to make the new view irresistible before it had earned the right to be. And, Feyerabend insisted, this was not a scandal but the way science actually advances. Had Galileo confined himself to what the rules of method permitted, the Copernican revolution might have stalled. Progress required breaking the rules.
From such examples Feyerabend drew his anarchist conclusion. There is no method, no set of rules, that has always been followed by good science and that always leads to success; every proposed rule has been violated, productively, by some great scientist at some crucial moment, and a science that had rigidly obeyed any fixed method would have been the poorer for it. The history of science is not the history of method triumphant but the history of method transgressed, of bold thinkers ignoring the canons of their day and getting away with it. And so the only honest methodological principle, the only one that does not at some point strangle a fruitful line of inquiry, is the cheerfully empty anything goes, which is less a rule than the denial that any rule will do. This is what Feyerabend meant by epistemological anarchism: not that all ideas are equally good, but that no fixed method can be trusted to sort the good from the bad, that science is and ought to be an opportunistic, rule-breaking, improvisational affair.
Feyerabend pushed Kuhn’s incommensurability further than Kuhn himself dared, and in a more unsettling direction. If rival frameworks are genuinely incommensurable, Feyerabend reasoned, then the supposed superiority of modern science over other traditions of knowledge, over myth, over magic, over the cosmologies of other cultures, cannot be established by any neutral standard, because there is no neutral standard; modern science judges these traditions inferior only by science’s own internal criteria, which the other traditions do not share and are not bound by. From this he drew a conclusion that outraged his colleagues: that modern science has no rational claim to the privileged authority it enjoys, that it is one tradition among many, enormously powerful within its own terms but not uniquely or objectively true, and that a free society should not grant it the monopoly on knowledge that it has come to hold. He compared the dominance of science in modern life to the dominance of the Church in earlier ages, and called, provocatively, for a separation of science and state, for science to be stripped of its official privileges and made to compete in the open market of traditions like any other.
It is easy to dismiss this as mere mischief, and Feyerabend, who enjoyed outraging people and adopted positions partly for the pleasure of the scandal, made it easy to do so. But there is a serious core that should not be lost in the laughter. Feyerabend was pressing, relentlessly, on the real weak points in the post-Kuhnian picture, refusing to let his more cautious colleagues paper over the difficulties. If there is no neutral method and no neutral standard of comparison, as Kuhn and Lakatos had in their different ways conceded, then by what right do we proclaim science the one true path to knowledge? The cautious thinkers wanted to keep the unsettling premises while avoiding the unsettling conclusion, to say that science is special and authoritative even though there is no neutral standpoint from which to prove it so. Feyerabend simply refused to let them, insisting that if you take the premises seriously you must follow them where they lead, even into the uncomfortable territory where science loses its privileged crown. He was the conscience of the revolution, the one who would not let it stop half-way.
There was, beneath the provocations, a genuine humanism in Feyerabend that his later writings made plainer, and it softens the caricature of him as a mere wrecker. What ultimately moved him was not a love of chaos but a love of human freedom and variety, a horror of any single system, scientific or otherwise, claiming the right to dictate to all the rest. He had seen, in the ideologies of the twentieth century, what happens when one framework is granted total authority and permitted to crush all alternatives in the name of truth, and he came to regard the imperial confidence of science with a wariness born of that history. His defense of astrology, of traditional medicine, of the cosmologies of other cultures, was not really a claim that these are as accurate as modern science; it was a claim that a free and humane society should not let any one tradition monopolize the definition of knowledge and stamp out the others. Reason itself, when it hardens into a single compulsory method, can become a tyranny, and Feyerabend’s anarchism was, at bottom, a plea for an open society of the mind, in which many traditions are allowed to flourish and compete and enrich one another. Read this way, his anything goes is less a counsel of intellectual despair than a defense of pluralism, and it is no accident that he titled one of his later books a farewell not to truth but to reason in its narrow, domineering, rule-bound form.
Was he right? Most philosophers and nearly all scientists think not, and there are powerful replies to be made. One may argue that science’s extraordinary practical success, its unrivaled power to predict and control and heal, does give it a claim to authority that does not depend on any neutral metaphysical standard, that the proof of the pudding is in the eating and science’s pudding is uniquely nourishing. One may argue that Feyerabend’s anything goes confuses the absence of a single fixed method with the absence of all standards, when in fact science has many flexible standards even if it has no one rigid method. One may argue that his picture of incommensurable, equally valid traditions romanticizes forms of knowledge that simply fail to do what science does. These are good replies, and they have largely carried the day. But Feyerabend’s gadfly function endures: he stands as a permanent reminder that the easy confidence in science’s method, the breezy textbook assurance that there is a single royal road to truth, cannot survive a close look at the actual history, and that those who want to defend science’s authority must find some better ground than the existence of a method that, on inspection, evaporates.
Between them, Lakatos and Feyerabend mark the two roads that lead out of Kuhn’s revolution, and the choice between them is, in a sense, the choice every reader of Kuhn must eventually make. Lakatos’s road tries to preserve as much of the old rationalism as the new history will allow, to salvage a defensible standard of good science from the wreckage of the simple method. Feyerabend’s road accepts the wreckage cheerfully and builds nothing in its place, declaring the search for method itself the error. Most working scientists, understandably, prefer Lakatos’s road, or some descendant of it, because it lets them keep believing that what they do is rational in a robust sense. But the nagging presence of Feyerabend’s alternative, the suspicion that the rationalist road may be a comforting fiction, is part of Kuhn’s permanent legacy, and no honest account of his impact can leave it out. There remains, however, one more figure in this constellation, a thinker who had reached some of Kuhn’s deepest insights before Kuhn did, who watched with mixed feelings as the younger man won the fame, and whose ideas about the unspoken foundations of all knowledge give the whole story an unexpected and humane depth.
Chapter Twenty-One — Polanyi and the Knowledge We Cannot Speak
Try to explain, in words alone, how to ride a bicycle. Set down the complete instructions, precise enough that someone who had never ridden could read them and pedal away on the first attempt. You cannot do it. No one can. The knowledge of how to balance on a bicycle is real, demonstrable, and reliable, yet it resists all attempts to capture it fully in explicit statements; it lives in the doing, in trained reflexes and felt adjustments that we cannot articulate even to ourselves. This humble fact, the gap between what we can do and what we can say, was the starting point for one of the most profound and least celebrated thinkers in our story, a man who built upon it an entire philosophy of knowledge and who had glimpsed much of Kuhn’s vision before Kuhn published a word of it. His name was Michael Polanyi, and his great phrase deserves to be as famous as paradigm shift: we know, he wrote, more than we can tell.
Polanyi came to philosophy by a remarkable route, for he had first been a distinguished scientist, a physical chemist of international reputation who had made real contributions to the understanding of chemical reactions before turning, in mid-life, to the deeper question of what scientific knowing actually consists in. This gave his philosophy an authority that the professional philosophers could not match: he was describing scientific knowledge from the inside, as one who had created it, not as an outside observer reconstructing it by logic. And what he found, looking inward at his own knowing, was that the official picture left out almost everything that mattered. The logical positivists had portrayed scientific knowledge as a structure of explicit statements, publicly verifiable, impersonal, with the knowing subject scrubbed out entirely. Polanyi saw that this was a fantasy, that real scientific knowledge is shot through with the personal, the tacit, the unspecifiable, that the scientist’s skilled judgment and committed participation are not contaminants to be removed but the very heart of the knowing.
The cornerstone of his thought was the concept of tacit knowing, which we met briefly when we examined how scientists learn from exemplars rather than rules. Polanyi developed it into a general theory of how all knowledge works. We know things, he argued, by attending from a host of particulars that we cannot specify, toward a coherent whole that we can. When I recognize a familiar face, I attend from countless tiny features, which I could never list, toward the face as a whole; the particulars are real and active in my knowing, but they function tacitly, below the threshold of articulation, in the background of my attention rather than its focus. All knowing, even the most rigorous scientific knowing, has this structure: it rests on a tacit foundation of skills, perceptions, and commitments that cannot themselves be made fully explicit, because the moment we turn to examine them they cease to function as the tacit ground of our attention. There is, Polanyi insisted, no such thing as wholly explicit knowledge; every articulate statement floats upon an iceberg of the unspoken.
From this followed a vision of science radically at odds with the impersonal ideal. Science, for Polanyi, is a personal achievement, sustained by the passionate commitment of individuals who stake themselves on their beliefs, who pursue intuitions they cannot fully justify, who dwell within their theories the way we dwell within our own bodies, attending from them toward the world. He called this indwelling, the way a skilled practitioner inhabits her framework so thoroughly that she sees through it rather than at it, the way a master uses a tool until it becomes a transparent extension of herself. To learn a science is to come to indwell its framework, to internalize its tacit skills until they become the unnoticed ground of one’s perception of the field. This is why science cannot be reduced to explicit rules, why it must be learned by apprenticeship, why the transmission of a scientific tradition depends on the personal contact of master and pupil, the passing on of skills that no textbook can contain. The parallel to Kuhn’s paradigms, learned through exemplars and embodied in tacit practice, is unmistakable.
And here the story takes a poignant turn, for Polanyi had published these ideas, in lectures and then in his major work Personal Knowledge, before Kuhn’s famous book appeared, and the two men had crossed paths; Kuhn had attended some of Polanyi’s lectures, and the older man’s influence on the younger seems undeniable. When The Structure of Scientific Revolutions made Kuhn world-famous, with its account of tacit paradigms and trained perception and the personal commitment of scientific communities, some of Polanyi’s admirers cried foul, even murmuring the ugly word plagiarism. The charge was unfair in its harshest form; Kuhn had reached his views by his own route, through history rather than through Polanyi’s philosophy of tacit knowing, and the resemblance is as much convergence as influence. But there was a genuine debt, and Kuhn, to his credit, acknowledged Polanyi in later editions of his book. Still, there is something melancholy in the spectacle of the older, deeper thinker watching the younger man’s catchier formulation conquer the world while his own subtler and more humane vision remained the property of a devoted few. It is, once again, the lesson of Fleck: being first and being right are not enough; the world rewards the one who is heard.
Yet Polanyi’s vision differs from Kuhn’s in ways that make it, in some respects, the richer of the two, and the differences are worth marking. Where Kuhn’s account tilted toward the unsettling and the relativistic, toward incommensurable worlds and progress without a goal, Polanyi’s tilted the other way, toward a hopeful realism. The tacit, committed, personal character of knowledge was, for Polanyi, not a threat to its objectivity but the very means by which we make contact with reality. The scientist’s passionate commitment, her sense that her theory is beautiful and must be true, her intuition reaching beyond the available evidence, these were for Polanyi the marks of a mind genuinely in touch with a real and intelligible world, responsibly submitting itself to what it takes to be true. He spoke without embarrassment of the scientist’s contact with reality, of discovery as the grasping of a hidden truth that was there all along waiting to be found. Where Kuhn flinched from the word truth, Polanyi embraced it, finding in the very personal and tacit character of knowing not a reason for doubt but a ground for a chastened and committed confidence.
Polanyi made one further contribution that speaks directly to the questions of community and authority running through this book, and it came from a famous essay in which he described what he called the Republic of Science. How, he asked, does the sprawling enterprise of science coordinate itself, with no central planner directing who works on what, no authority assigning problems, and yet achieve a coherent collective advance? His answer was an image of spontaneous order. Each scientist, he proposed, adjusts her work to the results of those around her, building on her neighbors’ findings, steering toward the most promising openings, submitting her contributions to the judgment of the community, exactly as independent agents in a free market coordinate through countless local adjustments into a larger order no one designed. Science is self-governing, a republic of independent minds bound by shared standards and mutual authority, in which the only discipline is the discipline of peers judging peers. This vision of the self-coordinating scientific community, held together not by rules imposed from above but by a shared tradition tacitly transmitted and collectively policed, complements Kuhn’s paradigm-bound communities beautifully, and it anticipates a question the next chapters must face: if science is a self-governing community sustained by shared commitments, then understanding it may require not only philosophy and history but also sociology, the study of how such communities actually work, and it was the sociologists who would push Kuhn’s ideas to their most explosive conclusions.
This difference reflects a deeper divergence of spirit that gives Polanyi a special place in our story. Kuhn was, at heart, a historian, fascinated by change, by the succession of frameworks, by the discontinuities that the tidy picture had hidden; his eye was on the revolutions, the breaks, the moments when one world gives way to another. Polanyi was, at heart, a philosopher of the knowing person, fascinated by the act of discovery itself, by the commitment and skill and contact with reality that make any knowing possible; his eye was on the continuities, on what endures through all the changes, on the personal participation that every framework requires. The two visions are not contradictory; they are complementary, two halves of a fuller picture. Kuhn shows us how the frameworks come and go; Polanyi shows us what the knowing person brings to every framework, the tacit, committed, reality-seeking activity without which no framework could be inhabited at all. To read them together is to be protected against the errors of each: against Kuhn’s drift toward relativism by Polanyi’s realism, and against any naive realism by Kuhn’s sense of how deeply our frameworks shape what we see.
Polanyi also reminds us of something the more famous quarrels can obscure, something that will matter when we weigh, in later chapters, the charge that Kuhn made science irrational. The tacit, personal, committed character of knowledge does not make it arbitrary or subjective in the dismissive sense. A skill is not arbitrary; it can be done well or badly, and the world is the judge. A trained perception is not a mere prejudice; it is an achievement, a hard-won attunement to real features of things. The personal commitment of the scientist is not the same as personal whim; it is a responsible submission to what she takes to be the demands of reality, a staking of herself that she may be called to account for. By showing that the personal and the tacit are woven into the very fabric of the most objective knowledge we have, Polanyi defused the false dichotomy that haunts the whole debate, the assumption that knowledge must be either impersonal and rule-bound or else merely subjective and arbitrary. There is a third thing, the thing Polanyi spent his life describing: personal knowledge, tacit and committed yet genuinely in contact with reality, and it may be the deepest answer of all to the anxieties that Kuhn’s revolution unleashed. With these thinkers before us, the cautious Lakatos, the anarchic Feyerabend, the humane Polanyi, we have seen the main lines of defense and attack that Kuhn’s work provoked among philosophers. We must now follow the controversy out of the seminar room and into the wider world, where Kuhn’s ideas, taken up by a new breed of sociologists, would be pushed toward conclusions that horrified him and would ignite the strangest intellectual war of the late twentieth century.
Chapter Twenty-Two — The Relativist He Refused to Be
The physicist Freeman Dyson once recorded a small outburst that tells us a great deal. Kuhn, exasperated by what people had made of his work, exclaimed to him: I am not a Kuhnian. It is a strange and almost comic thing for a man to say, this disowning of the very adjective formed from his own name, and yet it captures the central predicament of Kuhn’s later life. He had written a book that was taken to prove that science is irrational, that truth is relative, that one paradigm is as good as another and reality is whatever the powerful agree to call it; and he spent the rest of his days insisting, with mounting frustration, that he had meant nothing of the kind, that the Kuhnians who marched under his banner had misunderstood him, that the relativism attributed to him was a doctrine he found as repellent as his critics did. The man and the movement bearing his name had come apart, and he could not put them back together.
To judge whether Kuhn was a relativist, we have to be clear about what the charge actually means, for relativism is a slippery word that covers several quite different claims. In its crudest form, the form his enemies attacked, relativism holds that there is no objective truth at all, that every belief is merely true-for-some-community, that science is just one more system of myths with no greater claim on reality than any other, and that there is no rational basis whatever for preferring one theory to another. If this is relativism, then Kuhn was emphatically not a relativist, and said so repeatedly and with feeling. He believed that science makes genuine progress, that later paradigms are objectively better than earlier ones at solving problems, that the world constrains our theories and rules many of them out, that the choice between paradigms is governed by real shared values and is anything but arbitrary. A man who believes all that is not a relativist in the crude sense, whatever his careless admirers may have supposed.
And yet the charge would not have stuck so persistently to so unwilling a target if there were nothing to it, and honesty requires admitting that Kuhn’s own words gave it real purchase. He had written that rival paradigms are incommensurable, that there is no neutral standard to compare them, that scientists who hold different paradigms see different worlds, that paradigm choice resembles religious conversion more than logical proof, that there may be no sense in which science approaches a framework-independent truth. Each of these claims, taken to its extreme and stripped of his careful qualifications, points toward relativism, and Kuhn’s prose, especially in the famous first edition, was often vivid and provocative where it should have been guarded. He wrote like a man reaching for something he could not quite grasp, and the reaching produced sentences that, lifted out of context, read as bold declarations of the very relativism he later disowned. He had, in a sense, written more radically than he believed, and then spent decades trying to walk it back.
The walking-back began in earnest with a long postscript he added to the second edition of his book in 1969, seven years after the original, and continued through a stream of essays and lectures for the rest of his life. In these later writings he labored to clarify, to qualify, to distinguish the defensible core of his view from the relativist excess that had been read into it. He sharpened the concept of incommensurability, recasting it, as we have seen, from a vague matter of meaning-change into a more precise claim about taxonomic structure, and insisting that incommensurable did not mean incomparable or untranslatable, only that the comparison required judgment rather than algorithm and the translation required learning rather than a dictionary. He spelled out the shared values that govern theory choice, precisely to rebut the charge that the choice was arbitrary. He affirmed, again and again, that science progresses, that the world resists, that he was no enemy of reason. The later Kuhn is a more careful, more defensive, more qualified thinker than the brilliant provocateur of 1962, and reading the two together one feels the strain of a man trying to rein in a horse that had already bolted.
Did the clarifications succeed? Partly, and partly not, and the honest verdict is that Kuhn occupied a genuinely difficult middle position that is hard to hold steady and easy to misread in either direction. He wanted to say that our knowledge is deeply shaped by our frameworks, that there is no neutral standpoint, that successive paradigms do not simply converge on a single truth, and yet that science is rational, progressive, and constrained by a real world. This is not an incoherent position, but it is a delicate one, perpetually in danger of sliding toward the relativism on one side or collapsing back into the naive realism on the other. Kuhn spent his life on that knife-edge, and his many qualifications were the constant adjustments of balance that staying there required. His critics on the realist side thought he had already fallen off toward relativism and was refusing to admit it; his admirers on the relativist side thought he was losing his nerve and betraying his own best insights. He himself believed he was standing exactly where the truth was, on ground too narrow for either camp to see.
It is worth drawing out what this chastened, non-relativist Kuhn actually asks of us, because it is something valuable and easily lost between the warring caricatures. He does not ask us to abandon belief in truth or to treat all opinions as equal; that is the relativism he rejected. Nor does he ask us to return to the naive confidence that our current science is simply the way things are, the final word, the world correctly mirrored at last. He asks for something harder and more grown-up than either: that we hold our best beliefs firmly enough to act on them and build with them, while holding them loosely enough to remember that they rest on a framework which the future may revise. He asks us to take seriously both the genuine authority of science and its genuine fallibility, to trust it without worshipping it, to use it without mistaking it for revelation. This is not relativism; it is a kind of disciplined humility, the recognition that even our finest knowledge is the knowledge of finite beings working within historically conditioned frameworks, and that such knowledge can be both the best we have and not the last word. To live well with that double awareness, neither cynical nor credulous, is perhaps the deepest practical lesson of Kuhn’s thought, and it is precisely the lesson the relativist label obscures.
There is a deeper reason the charge of relativism was so hard for Kuhn to shake, and it has to do with a confusion that runs through the whole debate, a confusion between two questions that ought to be kept apart. One question is whether our knowledge is shaped, conditioned, framework-dependent, perspectival; call this the question of conditioning. The other is whether, despite being so shaped, our knowledge can still be objective, rational, answerable to a real world; call this the question of objectivity. The crude relativist assumes that a yes to the first question forces a no to the second, that if our knowledge is conditioned by our frameworks then it cannot be objective. Kuhn’s whole life’s work, properly understood, was an attempt to deny exactly this inference, to insist that knowledge can be both thoroughly conditioned and genuinely objective, that the perspectival character of science does not cancel its rationality. The relativists claimed him because he affirmed the conditioning; the realists rejected him because they too assumed conditioning meant the loss of objectivity. Both sides accepted the very inference Kuhn was trying to break, and so both misread him in the same way, the one with approval and the other with horror.
It is worth pausing on how lonely a position this made him, because it explains the note of weariness that creeps into his later work. A thinker who breaks a dichotomy that everyone else accepts is doomed to be misunderstood by all parties, claimed by the side he means to oppose and attacked by the side he means to defend, perpetually explaining that no, he does not mean that, and no, he does not mean that either, and watching his explanations themselves get absorbed into the misreadings. Kuhn was such a thinker. He had glimpsed a way of holding together the conditioning and the objectivity of knowledge that the entrenched categories of the debate could not accommodate, and he lacked, perhaps, the philosophical machinery to state it with the precision that would have forced his readers to see it. So he was left protesting, year after year, that he was not the relativist they took him for, a protest that only confirmed, in the minds of his relativist admirers, that the great man had simply lost his nerve, and in the minds of his realist critics, that he was trying to escape the consequences of his own ideas.
The exclamation to Dyson, I am not a Kuhnian, is therefore not a quip but a cry of genuine distress, the protest of a careful thinker watching his subtle vision coarsened into a slogan and his measured claims inflated into a manifesto for the very relativism he abhorred. It belongs beside the other small tragedy we noted at the very beginning of this book, the fate of the word paradigm, which escaped his control and came to mean almost anything. The man and his vocabulary and his very name had all been seized by forces he could not direct, pressed into the service of conclusions he rejected. And the most painful irony was still to come, for in the decades after his great book, a whole movement would arise that took his ideas as its charter, pushed them far past anything he had intended, and built upon them a sociology of knowledge so radical that it would provoke an actual war, and that movement, which did more than any other to fix the relativist Kuhn in the public mind, drew its inspiration from a man who wanted nothing to do with it.
Chapter Twenty-Three — When Sociologists Took the Wheel
For most of its history, the sociology of knowledge had observed a curious and revealing restriction, a kind of unspoken treaty with the philosophers. Sociologists were permitted to explain false or irrational beliefs by social causes; if a community believed something absurd, one could ask what interests, what power structures, what social needs had produced the absurdity. But true and rational beliefs were thought to need no such explanation. When people believe something because it is true, because the evidence rationally compels it, then truth and reason are the explanation, and there is nothing left for the sociologist to do. Society explains error; reason explains truth. This division of labor kept the sociologists safely away from the content of science, which was assumed to be governed by reason and evidence and therefore off-limits to social explanation. The sociologist could study the institutions of science, its reward systems and its politics, but not its actual knowledge, which belonged to the philosophers and the scientists themselves.
In the 1970s, a group of scholars centered at the University of Edinburgh tore up this treaty, and Kuhn’s work was their warrant for doing so. If Kuhn was right, they reasoned, then the old division of labor was untenable. For Kuhn had shown that even the best, most rational science is shaped through and through by the commitments of a community, that observation is theory-laden, that paradigm choice is not compelled by neutral evidence, that the content of science cannot be cleanly separated from the social life of the communities that produce it. And if that is so, then there is no longer any reason to exempt true and rational beliefs from social explanation. The sociologist should study the production of scientific knowledge in exactly the same way whether the knowledge is later judged true or false, rational or irrational, successful or failed. This was the founding principle of what came to be called the Strong Programme in the sociology of scientific knowledge, and its most famous tenet was the principle of symmetry: the same kinds of social cause should be invoked to explain true beliefs and false ones alike.
The symmetry principle was a genuine intellectual advance in one respect, and it is important to grant this before describing how it went wrong. It freed sociologists to study the actual making of scientific knowledge, to go into the laboratories and the controversies and watch how facts are established, how disputes are settled, how a claim moves from tentative conjecture to accepted truth. The resulting studies were often revelatory, showing in fine detail the human processes, the negotiations, the persuasions, the contingencies, behind results that the textbooks present as the simple unveiling of nature. Much of this work was excellent, and it permanently enriched our understanding of how science actually operates as a human activity. The sociologists had taken Kuhn’s insight that science is a communal, framework-bound enterprise and turned it into a powerful empirical research program, studying scientific knowledge as a social achievement rather than a mere reading-off of nature.
The constructionist impulse did not stop at Edinburgh, and two further developments pushed it in directions stranger still. In France, a brilliant and slippery thinker named Bruno Latour, together with various collaborators, developed an approach that came to be called actor-network theory, which proposed to study science by following the scientists around and describing, in deliberately flat and neutral language, how facts get built out of alliances among people, instruments, texts, and things. Latour pressed symmetry to a vertiginous extreme, insisting that the account should be symmetrical not only between true and false beliefs but between human and non-human actors, granting a kind of agency to the microbes and the machines and the very objects of study. The aim, paradoxically, was sometimes to bring reality back into the story that the sociologists had banished, to let the things themselves participate; but the effect, in the hands of his followers and popularizers, was often to dissolve the distinction between the discovery of a fact and its fabrication so thoroughly that nothing remained that could be called simply true. Latour himself, late in life, grew alarmed at how his tools were being used to deny established science altogether, and tried to pull back, in an echo of Kuhn’s own recoil from his disciples.
A second development came from a group of British sociologists, sometimes called the Bath school, who studied scientific controversies up close and uncovered a genuinely tricky phenomenon they called the experimenter’s regress. Suppose you want to know whether some delicate experimental result is real. You check it by seeing whether a properly conducted experiment detects it. But what counts as a properly conducted experiment? One that gets the right result. And what is the right result? The one a proper experiment yields. In a live controversy, where no one yet knows the answer, this circle has real bite: there is no framework-independent way to certify an experiment as competent except by whether it produces the effect whose reality is precisely what is in dispute. The Bath sociologists showed, in careful case studies, that the closure of such controversies often depends on social factors, on who is trusted, who is influential, which laboratories carry authority, and not on some neutral experimental verdict that simply settles the matter. This was a real and important finding, and it gave the constructionist case its strongest empirical support. The question, as always, was whether it showed that reality plays no role, or only that reality does not play the simple, decisive, self-announcing role the tidy picture had imagined, and the careful answer was the latter, though the bolder spirits claimed the former.
But the symmetry principle harbored an ambiguity that would prove explosive, and it is the same ambiguity we traced in the charge of relativism. To say that true and false beliefs should be explained by the same kinds of cause can mean something modest and defensible: that we should not assume in advance which beliefs are true and explain only the others socially, but should look impartially at how all scientific beliefs are produced. Or it can mean something far more radical: that the truth or falsity of a belief plays no role at all in explaining why it is held, that reality drops out of the account entirely, that scientific facts are simply whatever the relevant community is persuaded to accept, manufactured by social processes with no constraint from a world beyond. The modest reading keeps reality in the picture as one factor among the social ones; the radical reading expels reality altogether, leaving only the social. And the more radical sociologists, intoxicated by the program’s success and by the heady atmosphere of the times, increasingly slid toward the radical reading, writing as though scientific facts were pure social constructions, as though nature contributed nothing to what scientists end up believing about it.
This was the moment when the sociologists drove past anything Kuhn had intended, and it is worth being precise about the difference, because Kuhn himself was horrified by it. Kuhn had said that the world does not by itself dictate a unique theory, that frameworks shape what scientists see, that paradigm choice involves more than neutral evidence. He had never said that the world drops out, that nature places no constraint, that facts are mere social fictions. For Kuhn, reality always pushed back; the recalcitrance of nature was the very thing that generated anomalies and drove revolutions; the world was a partner in the production of knowledge, constraining even if not dictating. The radical sociologists kept Kuhn’s emphasis on the social and quietly dropped his insistence on the constraint of reality, and the result was a doctrine of pure social construction that Kuhn regarded as a grotesque caricature of his views. He had wanted to show that science is both social and answerable to reality; they had kept the first half and discarded the second, and then credited the whole to him.
A new vocabulary spread through these studies, and it carried the radical implication in its very grammar. Scientists were described not as discovering facts but as constructing them, not as finding what is there but as fabricating, negotiating, manufacturing knowledge. A famous study of a biology laboratory portrayed the production of a scientific fact as a process of literary and social construction, in which a statement gradually loses its qualifications and hardens into a fact through the social labor of the laboratory, much as Fleck had described decades earlier, but now with the constraint of reality conspicuously absent or bracketed. The language of construction was sometimes meant innocently, as a vivid way of describing the genuine human labor that goes into establishing a result. But it was easily read, and often meant, in the strong sense, as the claim that facts are made rather than found, that reality is the product of the social process rather than its constraint. And as this language spread from the sociology of science into the wider humanities, into literary theory and cultural studies and beyond, it shed even the empirical care of the original laboratory studies and became a fashionable slogan: that all knowledge is socially constructed, that science is just one more discourse, one more system of power, with no special claim on truth.
By the early 1990s, this had hardened into an orthodoxy in large stretches of the academic humanities, an orthodoxy that treated the social construction of scientific knowledge not as a contestable thesis to be argued but as an established starting point, a thing all sophisticated people knew. Science was widely described, in these circles, as a Western, masculine, hegemonic discourse, no more objectively true than the knowledge systems of other cultures, its claims to universal validity exposed as so many expressions of power. The unsettling questions that Kuhn had raised so carefully, and hedged so anxiously, had been turned into confident assertions by people who had often never read him closely and who pushed his tentative suggestions to conclusions he had spent his life rejecting. The relativism that Kuhn refused had become, by way of the sociologists and their popularizers, a campus commonplace. And the scientists, who had mostly ignored all this academic chatter as harmless, were about to notice it, and to react with a fury that would erupt into open intellectual warfare.
Chapter Twenty-Four — The Science Wars and the Sokal Bomb
In the spring of 1996, a respected academic journal of cultural studies published a long and dense article by a physicist, arguing that the findings of modern physics, and especially the theory of quantum gravity, supported the fashionable view that physical reality is at bottom a social and linguistic construct. The article was festooned with the right citations, spoke the right theoretical language, flattered the editors’ deepest convictions, and arrived at the conclusions they most wanted to hear. They published it with evident satisfaction. And then, on the very day of publication, the physicist announced in another magazine that the whole thing had been a hoax, a deliberate parody, a farrago of nonsense dressed up in fashionable jargon, designed to test whether a leading journal in the field would publish an article that sounded good and confirmed their prejudices, however lacking in sense or rigor. It would, and it had. The bomb went off, and the explosion was heard far beyond the academy.
The physicist was Alan Sokal, a professor at New York University with impeccable left-wing credentials, which mattered, because he insisted that his hoax was not an attack from the political right but a protest from within the left, against what he saw as a fashionable nonsense that was corrupting the humanities and discrediting progressive scholarship. He had grown alarmed at the spread of the view that science is merely a social construction with no special claim on truth, a view he regarded as both intellectually bankrupt and politically suicidal, since the left, of all movements, needed reliable knowledge of the real world to ground its critiques of injustice. His parody had a serious purpose: to demonstrate, by the journal’s own acceptance of transparent gibberish, that a certain style of postmodern theorizing had abandoned the standards that distinguish sense from nonsense, that it would credulously embrace any claim, however absurd, so long as it was couched in the approved vocabulary and reached the approved conclusions.
The hoax did not come from nowhere; it was the most spectacular eruption of a conflict that had been building for several years, a conflict that came to be called the Science Wars. On one side stood scientists and their philosophical allies, increasingly angry at what they saw as ignorant and arrogant attacks on the rationality and objectivity of science by humanists who did not understand the science they presumed to deconstruct. A combative book by a biologist and a mathematician had recently appeared, denouncing the academic left’s relativism about science as a betrayal of reason, and it had drawn the battle lines. On the other side stood the sociologists, literary theorists, and cultural critics who had built careers on the social-constructionist view of science and who regarded the scientists’ outrage as a defensive refusal to submit their own enterprise to critical scrutiny, a circling of the wagons by a priesthood unwilling to have its authority questioned. The two camps talked past each other with a completeness that would have made a fine illustration of incommensurability, each convinced the other was either a fool or a knave.
Kuhn’s shadow lay over the whole conflict, for he was, however unwillingly, the intellectual ancestor of the constructionist side, the man whose ideas had been the seed from which the whole social-constructionist program had grown. When the scientists denounced the relativists, they were denouncing, at one remove, the popular Kuhn, the Kuhn of incommensurable worlds and paradigm choice as conversion and science as the consensus of a community. When the constructionists defended themselves, they cited, among their founding authorities, the same Kuhn. And the real Kuhn, the careful, anxious, qualifying Kuhn who had spent decades denying the relativism now being fought over in his name, was almost entirely absent from the brawl, his subtle middle position invisible to both sides, who saw only the slogan version of his thought. The Science Wars were, in a sense, a war between two misreadings of Kuhn, the scientists attacking the relativist caricature and the humanists defending it, while the man himself, had anyone consulted him, would have disowned both.
Indeed Kuhn, by then in the last years of his life and gravely ill, watched the Science Wars with dismay and tried, in his final interviews and lectures, to dissociate himself from the excesses being committed in his name. He was appalled by the strong social-constructionist program, by the deconstructionists who had taken his ideas, he felt, and run them into absurdity. He insisted that he had never meant to deny the objectivity of science or the constraint of reality, that the relativism attributed to him was a perversion of his views. There is something deeply poignant in the spectacle of the dying Kuhn, the supposed founder of the relativist revolution, straining to defend the rationality of science against the relativists who claimed him as their prophet. The man who had been accused of making science irrational spent his last energies arguing that science is rational after all, and that his accusers and his disciples alike had misunderstood him. He died in 1996, the very year of the Sokal hoax, the conflict over his legacy raging around his deathbed.
What did the Science Wars actually settle? Less than either side claimed, and the honest assessment requires separating the genuine issues from the posturing. Sokal’s hoax certainly exposed a real intellectual rot in a certain corner of the humanities, a willingness to embrace impressive-sounding nonsense, a loss of the standards that distinguish argument from mere fashionable assertion. To that extent it landed a fair and useful blow, and the worst excesses of the social-constructionist rhetoric did recede in its aftermath. But the hoax did not, and could not, refute the serious questions that lay beneath the nonsense, the genuine questions about the role of social factors in science, about the limits of objectivity, about the relationship between knowledge and power, that Kuhn and the more careful sociologists had raised. A parody can puncture pretension, but it cannot answer a real question, and the real questions about how science actually works, how reason and society and reality combine in the making of knowledge, remained exactly as open after the explosion as before. The smoke cleared, the combatants retired bruised, and the deep problems Kuhn had opened were still there, waiting for calmer minds.
What made the hoax so devastating was the sheer transparency of the nonsense the editors had swallowed. Sokal’s parody declared that physical reality, no less than social reality, is at bottom a linguistic and social construct; it suggested that the very notion of an external world independent of human observers was an outdated relic of Enlightenment dogma; it strung together genuine technical terms from physics in ways that were, to anyone with the relevant training, gibberish, and then drew from this word-salad sweeping political and philosophical conclusions. To a physicist the article was instantly recognizable as a put-on, its mathematical and physical claims either meaningless or laughably wrong. The editors, lacking the scientific training to see this and dazzled by the article’s fashionable vocabulary and congenial conclusions, took it for a serious contribution from a scientist generously confirming their worldview. That was Sokal’s point in a nutshell: a field that could not tell a parody of itself from the real thing had lost the capacity for critical judgment that any serious intellectual enterprise requires, and was running instead on jargon and ideological affinity.
Sokal followed the hoax with a book, written with the physicist Jean Bricmont, that turned from parody to documentation, assembling passages in which celebrated theorists, mostly French, had borrowed terms and theorems from mathematics and physics and deployed them in ways that ranged from the merely decorative to the flatly nonsensical, invoking chaos theory and relativity and the mathematics of the infinite as metaphors and authorities without understanding them. The book was meticulous and, on its own terms, damning; it showed a pattern of intellectual imposture, the use of scientific terminology to dazzle rather than to illuminate, to borrow the prestige of science while flouting its standards. Defenders replied that the borrowings were playful metaphors unfairly read as literal claims, and that a few abuses did not discredit a whole tradition of thought. There was something to both sides, but the cumulative effect of the documentation was hard to shrug off, and it marked, for many onlookers, the moment when the spell of a certain grandiose theoretical style began to break.
Yet fairness requires acknowledging that the constructionist side was not simply a parade of frauds and fashionable poseurs, and that real intellectual gains lay tangled up with the excesses the hoax exposed. The careful laboratory studies had genuinely deepened our understanding of how science is done. Feminist scholars of science had raised serious and lasting questions about how the social composition of scientific communities, long overwhelmingly male, had shaped which problems got studied and which assumptions went unexamined, questions that improved science rather than undermining it. The insistence that knowledge is produced by located, interested, embodied human beings rather than by disembodied pure intellects was a real correction to a real distortion. The tragedy of the Science Wars was that the polarization swept all of this together with the genuine nonsense, so that defenders of science were tempted to dismiss the legitimate insights along with the absurdities, and defenders of the constructionist program were tempted to protect the absurdities for the sake of the insights. The bomb that cleared away the rot also scorched a good deal that was worth keeping, and sorting the one from the other was a task the war itself made almost impossible.
Perhaps the most important casualty of the Science Wars was nuance itself, and this is the lesson most worth carrying away. The conflict forced everyone to one pole or the other, to be either a defender of science’s pure objectivity or a champion of its social construction, with no room for the delicate middle position that Kuhn had tried to occupy, the position that science is both genuinely answerable to reality and genuinely shaped by the communities that practice it. Both of these are true, and they are not in conflict, but the war made it almost impossible to say both at once without being claimed by one camp and attacked by the other. The casualty was precisely the subtle, balanced understanding of science that we most need, the understanding that neither worships science as an infallible oracle nor dismisses it as mere ideology, but sees it for what it is: a magnificent, fallible, deeply human enterprise, constrained by a real world and shaped by real communities, our best though never perfect way of finding out how things stand. Recovering that balanced view, against the polarizing pressure that the Science Wars left behind, is part of what it means to read Kuhn honestly today, and it returns us to the man himself, who by the time of the great war over his legacy had long since moved on to quieter and in some ways deeper waters, abandoning even the famous word that had made his name.
Chapter Twenty-Five — The Historian at the Bench: Black-Body Theory
It is one thing to write a general theory about how science changes, and quite another to roll up your sleeves and do the painstaking historical work of reconstructing a single episode in all its technical detail. Kuhn, who had become famous for the general theory, never abandoned the particular craft, and in 1978 he published a dense and demanding book that most of his admirers never read and that revealed him as something his celebrity had obscured: a working historian of physics of formidable depth. The book concerned the birth of quantum theory, the very episode we glanced at earlier when we watched one of Kelvin’s little clouds, the puzzle of how hot bodies radiate, swell into a storm. And the conclusion Kuhn reached, after years of immersion in the original papers and correspondence, was startling enough to provoke a controversy of its own and to embarrass, gently, his own most famous ideas.
The standard story, the one in every textbook, runs as follows. In December of 1900, Max Planck, struggling to explain the radiation of hot bodies, was forced to assume that energy is not continuous but comes in discrete packets, or quanta, and with that revolutionary assumption the quantum age was born, the first crack in the edifice of classical physics. Planck is the hero, the year 1900 is the birthday, and the introduction of the quantum is the decisive revolutionary act. It is a tidy story, with a clear discoverer and a clear moment, exactly the kind of clean origin myth that, as Kuhn himself had taught us, the history of science loves to manufacture and that careful historical work usually dissolves. And dissolve it Kuhn did, turning his own skeptical method upon the founding legend of the very physics he had been trained in.
What Kuhn argued, on the basis of a meticulous reading of Planck’s actual papers, was that Planck in 1900 did not mean what the legend says he meant. When Planck introduced his famous quantities of energy, Kuhn contended, he did not understand himself to be claiming that energy is really, physically discontinuous, that radiation actually comes in indivisible lumps. He was performing a mathematical procedure, a way of counting and dividing that he regarded, in the manner of his classical training, as a formal device rather than a statement about the discontinuous nature of reality. Planck remained, in his own mind, a classical physicist; the radical idea that energy is genuinely quantized, that the discontinuity is a real feature of the world, was not Planck’s achievement in 1900 but emerged only several years later, and it emerged chiefly through the work of others, above all Albert Einstein and the physicist Paul Ehrenfest, who saw in Planck’s formulas a physical discontinuity that Planck himself had neither intended nor at first accepted.
If Kuhn was right, the implications were considerable and more than a little ironic. The quantum revolution did not begin with a single revolutionary act by a single revolutionary in 1900; it crept into being over the better part of a decade, through a process in which the man credited as its founder did not grasp, and indeed initially resisted, the very idea that bears his name. The clean origin story was, like so many clean origin stories, a retrospective construction, a tidy myth imposed by later physics upon a messy and gradual reality. Here was Kuhn’s own teaching about the textbook falsification of history, applied with devastating effect to one of the founding episodes of modern physics, and applied not in the broad strokes of his general theory but in the fine grain of detailed technical scholarship, paper by paper, letter by letter. The historian had caught the physicists in the act of mythologizing their own past, exactly as his theory predicted they would.
Kuhn’s qualifications for this work were extraordinary and worth remarking, for they explain how he could presume to revise the founding legend of quantum theory. He had not merely read the papers; he had, years earlier, conducted an enormous oral-history project, interviewing the surviving founders of quantum mechanics, recording their recollections, gathering their correspondence, building an archive of the revolution from the testimony of those who had made it. In one haunting instance he interviewed the great Niels Bohr, one of the chief architects of the quantum world, on the very day before Bohr died, capturing the old man’s final reflections on the revolution he had led. Few historians of science have ever been so close to their subject, so steeped in both the technical physics and the human testimony, and this intimacy gave Kuhn the standing to challenge the textbook story. He was not an outsider second-guessing the physicists; he was a trained physicist who had spoken with the founders and read every relevant scrap, and when such a man says the legend is wrong, the claim carries weight even where it provokes resistance.
The deeper lesson of the black-body study reaches beyond the particular question of who started the quantum revolution, and it is a lesson this book has been circling from its first pages. Founding myths are manufactured, always and everywhere, by the disciplines that need them. A science requires a clean origin story, a hero and a moment and a decisive act, partly for the same pedagogical reasons that make textbooks distort the past, and partly for the more human reason that communities, like nations, need their founding legends, their heroic ancestors, their decisive birthdays. The messy truth, that the quantum crept in gradually through the half-understanding and mutual misunderstanding of several men over a decade, makes a poor founding legend, and so it is replaced by the tidy myth of Planck and the year 1900. Kuhn’s achievement was to recover the messy truth beneath the tidy myth, and in doing so to demonstrate, on the most prestigious possible terrain, the very thesis about the textbook falsification of history that his famous book had advanced. The black-body study is his general theory of history made flesh in a single, exhaustively documented case.
The book provoked sharp disagreement, and the disagreement is itself instructive. Many physicists and some historians resisted Kuhn’s reading, insisting that Planck had after all introduced the quantum, that the revolutionary content was there in 1900 whatever Planck’s private understanding, that Kuhn was splitting hairs over what Planck meant when the formulas spoke for themselves. The dispute turned on subtle questions of interpretation, on how to read a man’s intentions from his equations, on whether the meaning of a scientific innovation resides in the mind of its author or in the formulas he leaves behind, and it was never fully resolved. But notice that the very shape of the controversy vindicated Kuhn’s larger point. Here were competent scholars, looking at the same documents, reaching incommensurable readings, unable to settle their dispute by any neutral appeal to the evidence, because what the evidence meant depended on the framework each brought to it. The argument about whether Planck started the quantum revolution was itself a small enactment of Kuhn’s thesis about how such arguments work.
There was a further and more delicate irony, one that Kuhn’s critics were quick to press. His detailed historical study of the black-body episode did not obviously fit the dramatic schema of his own most famous book. Where was the paradigm, the crisis, the gestalt switch, the sudden revolutionary conversion? What Kuhn actually found, when he looked closely at the birth of quantum theory, was something slower, messier, more continuous and more gradual than the schema of The Structure of Scientific Revolutions would lead one to expect, a transformation that crept in over years rather than erupting in a moment of conversion. Some took this as evidence that Kuhn’s own grand theory did not survive contact with the detailed history it claimed to explain, that the working historian had quietly refuted the famous theorist. It is a fair challenge, and Kuhn felt its force; the tidy drama of paradigm and revolution sits uneasily beside the gradual creep he documented in the quantum case.
But the tension can be read another way, and reading it generously tells us something important about the later Kuhn. Perhaps the black-body study did not refute his theory so much as mature it, forcing him to see that the broad schema of his famous book had been too clean, too dramatic, too fond of the sharp revolutionary break, and that the reality of scientific change was more continuous, more gradual, more various than his early formulations allowed. The detailed history was teaching the theorist to complicate his theory, to soften the hard edges of the revolutionary model, to make room for the slow transformations alongside the sudden ones. And indeed, as we will see, the later Kuhn did move in exactly this direction, away from the dramatic language of gestalt switches and toward a subtler, more evolutionary picture of change. The black-body book can be seen as a turning point, the moment when Kuhn’s own historical practice began to outgrow and reshape the theory that had made him famous.
Above all, the black-body study reminds us of something easy to forget amid the philosophical controversies: that Kuhn was, first and last, a historian, a man who believed that the way to understand science was to study, in loving and exhausting detail, what scientists had actually done. The grand theory of paradigms was never meant to float free of this detailed labor; it grew out of it and was meant to be tested against it. When Kuhn returned to the bench, to the close reading of Planck’s papers, he was not abandoning his philosophy but practicing the discipline from which it had sprung, and submitting it, as a good empiricist should, to the recalcitrance of the historical evidence. That the evidence pushed back, that it complicated and chastened the theory, is not a scandal but a vindication of his method, a sign that he was doing real history and not merely illustrating a thesis. The historian at the bench was the truest Kuhn, and it was the discipline of the bench that would lead him, in his final decades, to abandon even the word that had conquered the world.
Chapter Twenty-Six — Dropping the Word, Keeping the Idea
Imagine a writer so identified with a single word that the word has become his trademark, his brand, the thing every reader associates with his name, and now imagine that writer quietly deciding to stop using it. This is what Kuhn did with paradigm. The word that had made him famous, that had escaped into the wild and conquered the language, the word on which his whole reputation rested, he gradually set aside in his later work, almost as if disowning it. To casual observers this was baffling, even perverse. Why would a man abandon his most successful idea? But Kuhn was not abandoning the idea. He was trying to rescue it from the word, which had grown so bloated with meanings, so abused and overworked, that it could no longer do any precise intellectual labor. The word had become a liability, and Kuhn, ever the careful craftsman, decided he was better off without it.
We have already seen the seed of this decision. When Margaret Masterman counted twenty-one senses of paradigm in his famous book, she exposed a genuine problem that Kuhn could not deny: the word meant too many things. In response he had split it, as we noted, into the broad disciplinary matrix and the narrow exemplar. But this surgical division, however sensible, could not undo the damage already done in the wider world, where paradigm had become a all-purpose synonym for any framework, viewpoint, or fashionable change. A word so promiscuous was useless for the precise analysis Kuhn now wanted to pursue. And so, in his later writings, he increasingly avoided it, reaching instead for new and more carefully defined terms, accepting the loss of his famous brand in exchange for the precision he prized more highly. It was a quietly heroic act of intellectual hygiene, the willingness to give up fame’s favorite word for the sake of clear thought.
What did he reach for instead? Increasingly, the later Kuhn turned to the language of the lexicon and of taxonomy, to the structures of classification by which a science sorts the world into kinds. We touched on this when we examined incommensurability, but it deserves fuller treatment now, because it represents the mature form of Kuhn’s whole project, the destination toward which his long intellectual journey had been tending. A scientific community, the later Kuhn proposed, shares a lexicon, a structured vocabulary of kind-terms that carves the world into categories and embodies the community’s deepest assumptions about how things are sorted and related. To learn a science is to acquire its lexicon, to internalize its particular way of dividing the world into kinds. And a scientific revolution, in this mature formulation, is fundamentally a change in the lexicon, a redrawing of the taxonomic boundaries, a restructuring of the categories through which the science apprehends its domain.
This taxonomic turn gave Kuhn a more precise and more defensible account of incommensurability than his early talk of meaning-change and different worlds. The trouble with comparing rival paradigms, he could now say, is that they employ different lexicons, different taxonomic structures, and these cannot always be mapped onto one another without violating the basic logical rules that any coherent classification must obey. You cannot simply translate the kind-terms of one lexicon into those of another, because the categories cut the world at different joints, and a term that picks out a natural kind in one taxonomy may straddle the boundaries of two or more kinds in the other. This is a sharper and more technical claim than the vivid early formulations, and it lends itself to careful analysis in a way that the metaphor of different worlds never could. The later Kuhn was trying to give his great insight a rigorous logical form, to convert the brilliant but slippery intuitions of 1962 into something a philosopher could state precisely and defend against objection.
There was a price for this precision, and Kuhn paid it knowingly. The taxonomic, lexical Kuhn is far less thrilling than the Kuhn of paradigms and revolutions and gestalt switches. The early book had blazed with vivid images and bold declarations; the late work is technical, careful, hedged, concerned with the logical structure of kind-terms and the precise conditions of translatability. The early Kuhn captured the imagination of the world; the late Kuhn captured the attention of a few specialists in the philosophy of language. In trading the resonant word paradigm for the dry machinery of lexical taxonomy, Kuhn traded fame for rigor, influence for precision. He became, in his later decades, a less famous and in some ways a better thinker, working out the careful details of a position the world had already misunderstood and moved on from. There is a kind of integrity in this, the integrity of a man who cared more about getting it right than about being celebrated, and who was willing to follow his ideas into territory where his audience would not follow him.
The shift also reflected a change in what Kuhn took the central problem to be. In 1962 the central problem had been historical and dramatic: how does science change, how do whole frameworks rise and fall? By the end the central problem had become semantic and almost philosophical: what is meaning, how do the kind-terms of a science acquire their content, and what exactly happens to that content across a revolution? This was, in a sense, a deepening, a movement from the surface drama of scientific change to the underlying machinery of language and concept that made the drama possible. Kuhn had come to believe that the key to everything lay in understanding how scientific language hooks onto the world, how the categories of a lexicon are learned and how they can shift, and he devoted his last decades to this difficult and unglamorous inquiry, convinced that here, in the structure of scientific kind-terms, lay the real explanation of the phenomena his famous book had only described.
Much of this late development survives only in scattered essays and in works published after his death, which is part of why the late Kuhn is so much less known than the early one. A collection of his later philosophical essays appeared a few years after he died, gathering the lectures and papers in which he had worked out the taxonomic turn, and decades later a further volume assembled his last writings on incommensurability, the fragments of the very theory he had been unable to finish. Reading these, one encounters a Kuhn most of his admirers never met: not the prophet of paradigms but a meticulous philosopher of language, worrying at the precise conditions under which terms can and cannot be translated across a revolution, building careful little arguments about the logic of classification. The famous book was the public Kuhn; these late and posthumous writings are the private one, the thinker still at work on the problem long after the world had decided it knew what he thought.
One strand of the late work is especially surprising and shows how far Kuhn was willing to follow his questions. Having concluded that the heart of a science lies in its lexicon, its structured set of kind-terms, he became fascinated by how such terms are actually learned, how a child or a student comes to acquire the categories through which a community sorts the world. This drew him toward developmental and experimental psychology, toward the study of concept acquisition, as he sought a naturalistic foundation for his theory of meaning, an account grounded in how human beings really come to possess concepts rather than in abstract philosophical analysis. It was a striking move for a man whose reputation rested on grand historical narrative, this descent into the fine detail of how minds acquire categories, and it confirms that the late Kuhn was not coasting on his fame but genuinely pushing into new and difficult territory, trying to ground his whole vision of scientific change in the concrete machinery of human learning.
The taxonomic turn also matured an idea we met earlier, the comparison of Kuhn to Kant. The early Kuhn had seemed a Kantian with movable categories, holding that the mind structures experience through frameworks that, unlike Kant’s fixed forms, change from paradigm to paradigm. The late Kuhn made this more precise and more palatable. The lexicon, the taxonomic structure of a science, became his version of the Kantian categories: a set of conditions that must be in place before experience of a certain kind is possible at all, that shape and make possible the scientist’s perception of her domain. But unlike Kant’s categories, which were supposed to be universal and permanent, the same for all rational minds, Kuhn’s lexical structures are local and historical, shared by a particular community at a particular time and subject to revolutionary change. This gave Kuhn a respectable philosophical pedigree and a precise way of stating his core claim: that there is a kind of a priori, a structure that must be in place for experience to be possible, but that this a priori is not fixed once and for all; it is, in a phrase that captures the whole late project, a historically conditioned, relativized a priori, a framework that makes experience possible and yet can itself change.
It would be easy to see all this as a sad decline, the great man retreating from his bold youthful vision into cautious technical hairsplitting. But that would be to misread the trajectory. Kuhn was not retreating; he was excavating, digging beneath the vivid surface of his early ideas to find the solid conceptual foundations they needed and had lacked. The early book had been a brilliant sketch, a powerful but imprecise vision that had electrified the world precisely because it was bold and a little vague. The late work was the attempt to convert that sketch into a finished structure, to supply the rigor the sketch had lacked, to answer the objections the sketch had provoked. That the finished structure was less exciting than the sketch is in the nature of finished structures; the scaffolding is always more dramatic than the building. Kuhn spent his last years building, patiently and without much applause, and the building he was working on at his death was meant to be his masterpiece, the full and rigorous statement of the vision he had first glimpsed, all those years before, in an afternoon with Aristotle. And the keystone of that building, the idea he was reaching for at the end, was a new and deeper version of the evolutionary analogy, one that would recast the whole notion of a scientific revolution.
Chapter Twenty-Seven — Revolutions as Speciation
When Darwin looked at the finches of the Galápagos, he saw not a fixed set of types but a process of branching, a single ancestral stock that had spread across the islands and, in adapting to different conditions, split into distinct species that could no longer interbreed. Speciation, this splitting of one lineage into two that go their separate ways, is the engine by which the single tree of life has branched into its millions of distinct forms. It is a slow, often imperceptible process, recognizable usually only long after it has occurred, and it is irreversible: once two populations have diverged into separate species, they cannot merge back into one. In the last phase of his life, Kuhn became convinced that this biological process of speciation was the truest model for what happens in a scientific revolution, and the conviction led him to revise, in a deep way, the dramatic picture of his youth.
The picture in his famous book had been essentially one of replacement. A reigning paradigm enters crisis, a new paradigm arises, and the new replaces the old, the way one regime replaces another in a political revolution. The community passes from one framework to the next; the old world is abandoned, the new world adopted. This is a vision of succession, of one thing giving way to another along a single line. But when Kuhn looked harder at the actual history of science, especially with the eyes of his late evolutionary thinking, he came to suspect that this picture of simple replacement missed something crucial about how scientific change usually works. Revolutions, he came to think, are less often a matter of one paradigm replacing another within a single community, and more often a matter of one field splitting into two, of a scientific specialty branching, like a species, into distinct descendant specialties that go their separate ways and can no longer fully communicate.
Consider how the sciences have actually multiplied over the centuries. Where once there was natural philosophy, a single broad inquiry into nature, there are now physics and chemistry and biology, and within physics there are dozens of distinct specialties, and within those, sub-specialties, each with its own journals, its own conferences, its own training, its own lexicon, its own community of practitioners who can barely talk to their cousins in the neighboring fields. This proliferation of specialties is the dominant fact about the development of modern science, far more pervasive than the dramatic wholesale replacements that the revolutionary model emphasized. And it looks, Kuhn realized, exactly like speciation: a single ancestral discipline, adapting to different problems and growing in different directions, splitting into descendant disciplines that diverge until they can no longer interbreed, until a practitioner of one cannot fully understand or contribute to another. The tree of the sciences branches like the tree of life, and by the same kind of process.
This reframing changed the meaning of incommensurability in an illuminating way. In the early picture, incommensurability was the gulf between an old paradigm and the new one that had replaced it, a gulf in time, between successive frameworks of a single community. In the late picture, incommensurability becomes the gulf between contemporary specialties, the inability of practitioners in different branches of science to fully understand one another’s lexicons, a gulf in space, as it were, between coexisting communities that have descended from a common ancestor and diverged. This is a far less threatening kind of incommensurability, and a far more obviously real one. No one is much disturbed by the observation that a particle physicist and a molecular biologist cannot readily understand each other’s technical work; it is a familiar fact of modern science, and it carries none of the vertiginous implications that the early talk of incommensurable successive paradigms had seemed to carry. By relocating incommensurability from the temporal gap between old and new to the lateral gaps between coexisting specialties, the late Kuhn defused much of its threat while preserving its core insight: that distinct scientific communities can have genuinely different and not fully intertranslatable ways of carving up the world.
The speciation model also dissolved one of the most troubling features of the early picture, the apparent loss involved in revolutions. If a revolution is the replacement of one paradigm by another, then whatever the old paradigm could do but the new one cannot is simply lost, a real subtraction from the sum of knowledge, and this Kuhn-loss, as it came to be called, was deeply puzzling, for how can science progress if each advance also destroys? But if a revolution is a speciation, a branching rather than a replacement, then the picture changes. The parent discipline does not vanish; it gives rise to descendants that, between them, may cover more ground than the parent did, each specializing in what it does best. There need be no net loss, only a division of labor, the way the branching of life has produced not a loss of biological capacity but an explosion of it, each species mastering its own niche. The proliferation of specialties is not a story of repeated loss but a story of cumulative diversification, the sciences collectively growing richer and more capable even as individual practitioners grow more narrowly specialized and more cut off from one another.
There is something poignant in Kuhn arriving, at the end, at a vision of science as branching diversification rather than revolutionary replacement, for it is a gentler, less dramatic, more hopeful picture than the one that made him famous. The early Kuhn had given the world a vision of science lurching through violent upheavals, each revolution a trauma, each paradigm shift a kind of collective conversion experience with its winners and its broken losers. The late Kuhn offered something calmer and in a way more profound: science as a great branching tree, growing and dividing and diversifying, its revolutions not catastrophic ruptures but the natural splittings by which a living thing grows more various and more capable over time. The drama is muted, but the wonder is, if anything, deepened, for there is a particular grandeur, as Darwin himself observed at the close of his great book, in this view of a process that, from so simple a beginning, has elaborated itself into endless forms most beautiful and most wonderful.
The speciation model carries over one feature of biological branching that Kuhn found especially apt: its irreversibility. Once two species have diverged, they cannot remerge; the genetic barriers that separate them are permanent, and the tree of life branches but never rejoins. The same, Kuhn observed, holds of scientific specialties. Once a field has split into distinct descendant disciplines, each with its own lexicon and its own community, they do not flow back together into a single undifferentiated parent. Physics and chemistry, having diverged, remain distinct, however much they may exchange ideas at their borders; the branches do not fuse. This irreversibility explains something about the felt direction of scientific history, its sense of moving onward and never back, without requiring the troublesome notion of convergence on a final truth. Science is irreversible not because it climbs toward a fixed destination but because, like life, it branches, and branches do not unbranch. The arrow of scientific time is the arrow of diversification, not the arrow of approach to a goal.
There is also a question of identity buried in the speciation model, and it is one of those quiet puzzles that repays a moment’s thought. When a field splits into two specialties, which of them, if either, is the continuation of the original, and which is the new thing? The question often has no clean answer, just as it has no clean answer in biology, where it is frequently impossible to say which of two daughter species is the continuation of the ancestor and which the offshoot, or indeed to identify the exact moment when one species became two. The boundaries are vague, the transitions gradual, the act of speciation recognizable only well after it has occurred. Kuhn relished this parallel, for it matched what he had found in the actual history of science, where the moment of revolution is rarely sharp, where contemporaries often cannot tell that a revolution is occurring, where the identification of a paradigm shift is a retrospective act performed by historians long after the dust has settled. The vagueness that had seemed a weakness in his early dramatic model became, in the evolutionary reframing, a faithful reflection of how scientific change actually feels to those living through it: not a clean break announced by trumpets, but a gradual divergence noticed only in hindsight.
Kuhn did not live to complete the book in which he meant to work all this out. He died in 1996 with the manuscript unfinished, the full statement of his mature, evolutionary, taxonomic vision left as fragments and drafts that others would later assemble and publish. There is a melancholy fittingness in this incompleteness. The man who had taught the world that science never reaches a final, finished state, that every framework is provisional and every theory open to future revision, left his own life’s work unfinished, a structure still under construction at the end, its keystone never quite set in place. The vision he was reaching for, of scientific change as a branching evolutionary process driven by the splitting of lexicons and the speciation of communities, remains a sketch toward a theory rather than a theory complete. But it points, unmistakably, beyond the famous early picture toward something richer, and it suggests that Kuhn, had he lived, might have given us a maturer and more balanced account of scientific change than the one that conquered the world. That maturer account is part of his legacy too, less famous than the paradigm but in some ways wiser, and it belongs in any honest reckoning of what this remarkable thinker finally came to believe. It also leaves us with unfinished business of our own, for we have not yet faced squarely the most disturbing of all his ideas, the one the speciation model was meant to soften: the claim that science, in advancing, also forgets.
Chapter Twenty-Eight — Kuhn-Loss, or What Science Forgets
We assume that science only ever gains. Each advance adds to the storehouse of knowledge; nothing of value is ever lost; the sum total of what we understand grows monotonically, generation after generation, like a bank account that only ever receives deposits. This assumption is so deep that to question it sounds almost perverse. Yet Kuhn questioned it, and the phenomenon he pointed to is real, well-documented, and genuinely unsettling. Sometimes, when science moves from one paradigm to the next, it does not merely add; it also subtracts. Questions that the old framework answered, the new framework can no longer answer, or no longer even asks. Explanations that the old paradigm provided are simply abandoned, and the things they explained become, for a time or forever, mysteries again. Science forgets. The phenomenon acquired a name, coined by a later scholar in Kuhn’s honor: Kuhn-loss.
The cleanest example comes from the very triumph we usually cite as the model of scientific progress: Newton’s theory of gravity. Before Newton, the dominant approach to the heavens, associated with Descartes, sought to explain the motions of the planets mechanically, by means of invisible swirling vortices of subtle matter that carried the planets around in their orbits the way a whirlpool carries a leaf. This was, by the standards of its day, a satisfying explanation, because it explained gravity in terms of direct physical contact, of pushing and shoving, of matter acting on matter through touch. It answered the question why do the planets move as they do with a mechanism one could picture. Newton swept all this away, and replaced it with a law of universal gravitation of breathtaking power and precision. But Newton’s law came at a cost that troubled even Newton himself. It described how gravity acts, with unprecedented accuracy, but it offered no mechanism, no explanation of how one body could pull on another across the empty gulf of space without any contact between them. Action at a distance, the great Cartesian thinkers complained, was occult, mysterious, a relapse into the very magical thinking that mechanical philosophy had labored to banish.
Here was a genuine loss. The Cartesian vortex theory had answered a question, how does gravity work mechanically, that the Newtonian theory simply could not answer and, in the end, declined to ask. Newton himself was uncomfortable; pressed on how his gravity could reach across empty space, he famously refused to offer a hypothesis, declining to explain the mechanism and insisting only that his mathematical description worked. The new paradigm was vastly superior in accuracy, scope, and predictive power, and no one sensible would trade it back for the vortices. But it had abandoned a question the old paradigm had taken seriously and had thought it answered, and to the partisans of the old view this looked not like progress but like a retreat into mystery, a willingness to describe rather than explain, to calculate rather than understand. The community gained enormously, but it also gave something up, and what it gave up was real.
The history of science offers many such examples once you learn to look for them, and the pattern is always the same: the new paradigm, for all its superiority, drops some explanatory ambition, abandons some question, loses some capacity that the old one had. When chemistry was revolutionized, certain qualitative explanations of why substances behave as they do were set aside in favor of quantitative laws that, for a time, explained less about the inner nature of things even as they predicted more about their measurable behavior. The shift to a new framework characteristically involves a change in what counts as a legitimate question, and questions that fall outside the new framework’s scope are not so much answered as dismissed, ruled out of order, declared not really scientific questions at all. The vortex theorist’s demand for a mechanism of gravity did not get answered by Newton; it got reclassified as a bad question, a confusion, the kind of thing a properly trained scientist learns not to ask.
This is why the comfortable picture of science as pure accumulation cannot be quite right, and why Kuhn-loss matters for the deepest questions about progress. If progress were simple accumulation, if each paradigm contained all the achievements of its predecessors plus more, then we could measure progress straightforwardly, by the growing pile of what we know, and we could speak confidently of science converging on the complete truth. But Kuhn-loss shows that the relation between successive paradigms is not one of simple containment. The new does not contain the old; it replaces it, and in replacing it both gains and loses, keeps some things and drops others, answers new questions while abandoning old ones. The ledger of scientific change has entries in both columns, debits as well as credits, and this is precisely why Kuhn doubted that the net movement could be described as approach to a fixed truth. How do you sum a column that contains both gains and losses, in incommensurable currencies, and declare the total a steady advance toward a destination?
It must be stressed that Kuhn-loss is usually temporary and partial, and acknowledging this guards against overstating the case. Often what is lost in a revolution is recovered later, in a new form, once the new paradigm has matured enough to circle back and address the questions it had initially abandoned. Newton could not explain the mechanism of gravity, but Einstein, centuries later, in a sense restored a kind of mechanism, recasting gravity as the curvature of space and time, answering after a fashion the very question Newton had set aside. The loss was real but not permanent; the question went underground for two hundred years and then resurfaced, answered in a way neither Descartes nor Newton could have imagined. So Kuhn-loss does not show that science marches backward, or that the losses outweigh the gains; the gains are usually overwhelming, and the losses are often temporary. What it shows is subtler: that the path of science is not a straight ascent but a winding road, that progress is real but not simple, that the relationship between old and new paradigms is too complex to capture in the tidy image of an ever-growing pile.
Biology furnishes another instructive case, this time of a loss that most scientists regard as pure gain and yet that fits the pattern exactly. Before Darwin, living things were widely understood in terms of purpose: the eye was for seeing, the heart for pumping, each organ and each creature designed to fulfill a function in a grand scheme, and to explain a feature of an organism was to say what it was for. This teleological way of thinking answered a real question, the question of purpose, with a real and satisfying kind of answer. Darwin’s theory of natural selection swept it away, replacing purpose with the blind interplay of variation and selection, in which nothing is for anything in the old sense, and the appearance of design is a byproduct of survival rather than evidence of a designer. Most biologists count this an immense advance, and so it is. But notice that a kind of question, the question of ultimate purpose, was not so much answered as abolished; the new framework declared it a confused question, a relic of pre-scientific thinking, and trained biologists to stop asking it. To the believer in purpose, this looked like a loss, a refusal to address something the old view had taken seriously, and in a strict sense it was one, however much the gains outweighed it.
The scholar who coined the term Kuhn-loss, a physicist and historian named Heinz Post, did so in the course of defending a principle that stands as a kind of counterweight to it, and the tension between the two is illuminating. Post argued for what he called a correspondence principle: that a good new theory should, in general, preserve the well-confirmed successes of the theory it replaces, recovering the old results as a special or limiting case even as it goes beyond them. Einstein’s relativity, for instance, does not discard Newton’s mechanics but contains it, yielding Newton’s familiar laws as an excellent approximation for objects moving slowly compared with light. By this principle, the gains of a revolution should normally include, rather than abandon, the genuine achievements of the past, and Kuhn-loss should be the exception rather than the rule. There is real force in this, and it tempers the more alarming readings of Kuhn: most revolutions do preserve most of what mattered in the old paradigm, recovering it in new form. But Post coined the term Kuhn-loss precisely to mark the cases that escape the correspondence principle, the genuine losses that do occur, and the existence of even a few such cases is enough to refute the comfortable picture of science as pure, lossless accumulation. The truth lies between the extremes: revolutions mostly preserve, but not always, and the exceptions are real.
There is a humility in the idea of Kuhn-loss that is worth carrying away, and it bears directly on how we should regard our own scientific moment. If past revolutions involved real losses, then it is entirely possible that our current paradigms, for all their power, have quietly abandoned questions that earlier frameworks took seriously and that future frameworks may take seriously again. There may be phenomena our present science cannot explain, not because they are inexplicable, but because our particular framework has ruled them out of consideration, declared them not real questions, trained us not to see them. The vortex theorist’s question about the mechanism of gravity seemed, to confident Newtonians for two centuries, a confused non-question, until Einstein showed it was a good question after all. What questions are we, in our confidence, dismissing today that some future Einstein will redeem? Kuhn-loss teaches us that the boundaries our paradigm draws around the askable and the answerable are not the permanent boundaries of reason but the local and provisional boundaries of our own historical framework, and that the future may redraw them in ways we cannot now foresee.
This unsettling idea returns us to a theme that has run through the whole book and that the speciation model of the last chapter was partly designed to soften. If revolutions involve genuine loss, then the history of science is not the serene accumulation the textbooks describe but a more dramatic affair of gain and loss, advance and retreat, questions raised and questions abandoned. The branching model offered one way to reconcile this with the undeniable overall growth of scientific knowledge: perhaps the losses suffered by any single line are compensated by the gains of the proliferating branches, so that the tree of science as a whole grows richer even as individual paradigms gain and lose. Whether that reconciliation fully works remains, like so much in Kuhn, an open question. What is certain is that the simple story of pure accumulation, the story most of us absorbed without ever examining it, cannot survive a close look at what actually happens when science changes its mind. And it is a measure of how thoroughly Kuhn reshaped our thinking that the very phenomenon of scientific forgetting, invisible before him, now bears his name, one more way in which a single quiet historian rewrote how an entire civilization understands its own most powerful form of knowledge, a conquest of the mind whose astonishing breadth is itself our next subject.
Chapter Twenty-Nine — The Empire of the Word “Paradigm”
A curious thing happened to Kuhn’s ideas after they left the narrow world of the history and philosophy of science: they conquered nearly everything else. We have already traced how the word paradigm escaped into the wild and lost its precision. But the deeper phenomenon is not merely linguistic. Kuhn’s whole way of thinking about knowledge, his vision of communities bound by shared frameworks that periodically undergo revolutionary change, swept through one discipline after another, reshaping how scholars in fields far removed from physics understood their own enterprises. Economists, political scientists, sociologists, theologians, historians, literary critics, management theorists, all began to speak of their paradigms, to analyze their fields in Kuhnian terms, to interpret their own intellectual upheavals as paradigm shifts. The reluctant revolutionary of physics had become, without intending it, the prophet of an entire age’s self-understanding across the whole landscape of thought.
Consider economics, where the language of paradigms found especially fertile ground. The great upheaval associated with John Maynard Keynes, which transformed economic thought in the mid-twentieth century, came to be widely described as the Keynesian Revolution, and economists found in Kuhn a ready framework for understanding how their discipline had been transformed and how it might be transformed again. The successive dominance of different schools of economic thought, the classical, the Keynesian, the monetarist, and their rivals, lent itself naturally to the Kuhnian picture of reigning paradigms challenged by anomalies, plunged into crisis, and overthrown by revolutionary alternatives. When the prevailing economic framework failed to predict or explain some great disturbance, economists reached instinctively for Kuhn’s vocabulary, speaking of anomalies and crises and the need for a new paradigm. Whether economics really works the way physics does, whether its schools are paradigms in Kuhn’s strict sense, is another matter, and a doubtful one. But the borrowing was irresistible, because Kuhn had given the discipline a compelling story to tell about its own turbulent history.
The borrowing went furthest, and grew most problematic, in the social sciences generally, and here we must note an irony that Kuhn himself observed. He had developed his theory by studying the mature physical sciences, fields with a single dominant paradigm shared by the whole community. The social sciences, he thought, were different; they were characteristically pre-paradigmatic, marked not by a single reigning framework but by competing schools, each with its own assumptions, none commanding the assent of the whole field. In this respect the social sciences resembled the early, immature stage of a science, before a paradigm has emerged to unify it. Kuhn’s point was almost the opposite of the use to which his ideas were put: he was suggesting that the social sciences lacked paradigms in his sense, while the social scientists eagerly adopted his vocabulary to describe themselves as having them. There arose what some wryly called paradigm envy, the longing of the softer disciplines to be as unified, as cumulative, as authoritative as physics, and the hope that by adopting the language of paradigms they might somehow acquire the substance.
Even theology and religious studies felt the pull. Scholars of religion borrowed Kuhn’s framework to analyze the great transformations of religious thought, the shifts from one dominant theological vision to another, the way a religious community can be bound by shared assumptions that periodically undergo revolutionary change. The notion of a paradigm proved surprisingly adaptable to the study of how worldviews rise and fall, how communities of belief are organized around shared frameworks, how the unthinkable becomes thinkable and the orthodox becomes obsolete. Some theologians spoke explicitly of paradigm shifts in the history of Christianity, of successive epochs each organized around a different fundamental vision. The model migrated effortlessly from the laboratory to the cathedral, which should perhaps not surprise us, given how often Kuhn himself had reached for the language of conversion and faith to describe what happens when scientists change their paradigms. The boundary between scientific and religious transformation, which the tidy universe had policed so strictly, turned out, in Kuhn’s hands, to be more porous than anyone had supposed.
The migration did not stop at the borders of the academy; it spilled out into the wider commercial culture, where it underwent its final and most thorough degradation. In the literature of business and management, paradigm shift became one of the most reliable pieces of inspirational jargon, deployed to lend an air of intellectual heft to the most ordinary proposals for change. Every reorganization, every new product, every fashionable management technique could be dignified as a paradigm shift, and consultants built careers on promising to help corporations achieve them. In the self-help and motivational literature the word migrated further still, becoming a name for any personal change of perspective, any decision to see one’s life differently, so that a reader might be urged to undergo a paradigm shift in her attitude toward her diet or her relationships. At this furthest remove from its origin, the word had been emptied of nearly all content, retaining only a vague flavor of transformation and importance. Kuhn, who had reached for the word to capture something precise and difficult about the history of physics, lived to see it become a synonym for the contents of a motivational poster, and the descent from the one to the other is a small parable of what happens to ideas when they become too popular for their own good.
Behind the social sciences’ eager embrace of Kuhn lay an old and revealing anxiety, sometimes called physics envy, and it is worth naming because it shaped the borrowing. The physical sciences, with their precise laws, their decisive experiments, their cumulative and authoritative knowledge, had long been the envy of the softer disciplines, which struggled with messy data, contested methods, and a chronic inability to settle their disputes. To the practitioners of these fields, Kuhn seemed to offer both a diagnosis and a hope. The diagnosis was that their disunity reflected a pre-paradigmatic condition, the immature stage before a unifying framework emerges. The hope was that they too might one day achieve a paradigm, might cross the threshold into mature, unified, cumulative science, might become, at last, as respectable as physics. This was a misreading of Kuhn in one respect, since he doubted the social sciences would or even should become paradigmatic in the manner of physics, their subject matter being human and historical in ways that may resist such unification. But the misreading was understandable and even poignant, an expression of a real longing for the authority that the physical sciences command, and it drove much of the enthusiasm with which the social sciences adopted a vocabulary that, in Kuhn’s strict sense, may not have described them at all.
What should we make of this vast migration of Kuhn’s ideas across the whole map of human knowledge? Two things, pointing in opposite directions, and the honest reader must hold both. On the one hand, the migration testifies to the genuine depth and generality of what Kuhn had found. He had not merely described a quirk of physics; he had identified something about how communities of inquiry work in general, how shared frameworks structure thought, how the unthinkable becomes thinkable through collective transformation. That his ideas illuminated so many fields suggests they had captured something real and deep about the social nature of knowledge itself, something that transcends the particular case of the physical sciences. The empire of the word reflects the empire of the insight, and an insight that fertile is no small thing.
On the other hand, the migration involved a great deal of loose and careless borrowing, in which Kuhn’s careful concepts were stretched, diluted, and abused, until paradigm shift meant little more than any significant change of mind, and the rigorous machinery of his theory dissolved into a vague gesture toward the idea that frameworks matter and sometimes change. Much of the social-scientific and popular use of Kuhn was of this debased kind, invoking his prestige without his precision, borrowing the resonant vocabulary while ignoring the careful distinctions that gave it meaning. The empire of the word was, in large part, an empire of misunderstanding, a vast territory in which Kuhn’s name was invoked to license conclusions he had never drawn and to dignify changes he would never have called revolutions. The breadth of his influence and the shallowness of much of it are two sides of the same coin: an idea powerful and resonant enough to conquer a culture is, almost necessarily, an idea coarsened and simplified in the conquering.
There is a lesson here about the fate of ideas in the world that goes beyond Kuhn’s particular case, and it is one of the quiet themes of his own thought turned back upon itself. An idea, once released, takes on a life of its own, travels into communities its author never imagined, acquires meanings he never intended, gets put to uses he would deplore. The author loses control of his creation the moment it leaves his hands, and the more successful the idea, the more completely it escapes him. Kuhn experienced this with unusual intensity, watching his careful historical thesis become a universal slogan, his subtle vision become a buzzword, his name become an adjective he disowned. But the phenomenon is general, and it is, in a sense, the very phenomenon Kuhn had studied: the way meaning shifts as ideas move between communities, the way the same words come to mean different things in different hands. The empire of the word paradigm is one more demonstration of Kuhn’s own central insight, played out on the grandest possible scale, with his own most famous idea as the specimen.
And yet, when all the abuses are tallied, something genuine remains, and it is worth ending this chapter by naming it, because it is part of why Kuhn matters beyond the academy. The widespread adoption of his framework, for all its looseness, did teach a whole culture to think historically and critically about its own forms of knowledge, to recognize that even our most confident frameworks are historical achievements that may one day be superseded, to look for the hidden assumptions that structure our thinking and to imagine that they might be otherwise. This is a valuable habit of mind, a kind of intellectual self-awareness that Kuhn did more than anyone to spread, and it survives even in the debased popular versions of his thought. To have taught an entire civilization to ask, of its own most cherished certainties, what unexamined paradigm lies beneath them and whether it too might someday shift, is no small legacy, even if the asking is often crude. The empire of the word, whatever its excesses, rests on a real conquest of the collective imagination, and the conquest has made us, on the whole, more thoughtful about the contingency of what we know. That conquest, however, was achieved in a world of science very different from our own, and we must ask, before we are done, whether Kuhn’s picture still fits the strange new science of our own century.
Chapter Thirty — Paradigms in the Age of Big Science and Machines
In 2016, the announcement that scientists had directly detected gravitational waves, ripples in the fabric of space and time predicted by Einstein a century before, was reported in a scientific paper. The paper had over a thousand authors. Let that number sink in. A thousand and more human beings, spread across continents and institutions, had collaborated to build and operate the instruments and analyze the data that yielded a single discovery, and all of them shared the credit. This is a kind of science that would have been almost unimaginable when Kuhn wrote his famous book in 1962, a science of vast collaborations, billion-dollar machines, and authorship lists longer than some of the papers themselves. And it raises a pointed question for anyone who has followed Kuhn this far: does his picture of science, drawn largely from the heroic age of individual discoverers, still describe the science of our own time?
Kuhn himself was aware that science was changing, that the small communities of his historical studies were giving way to something larger and more industrial, and he worried, late in life, that his model might not fit the emerging world. The science he had studied was, for the most part, the work of relatively small communities of individuals, in which a single mind could grasp a whole field and a single discovery could be attributed to a single person. The science of today is increasingly the work of enormous teams, of international collaborations numbering in the thousands, of machines so large and expensive that only a few exist in the world and whole nations must pool their resources to build them. The great particle colliders, the space telescopes, the gravitational wave detectors, these are not the instruments of a lone investigator but the cathedrals of a collective enterprise, and the knowledge they produce is collective in a way that strains the very notion of individual discovery on which so much of the traditional picture rested.
Does this new scale break Kuhn’s model, or merely rescale it? There is a case for saying it merely rescales it, that the fundamentals remain. A thousand-author collaboration is still a community bound by a shared paradigm, still engaged in normal science, still solving puzzles within an accepted framework, still capable, in principle, of encountering anomalies and entering crisis. The detection of gravitational waves was, after all, a magnificent confirmation of an existing paradigm, Einstein’s general relativity, the very triumph of normal science writ large; thousands of people pooling their efforts to extend and confirm a reigning framework is Kuhnian normal science on an industrial scale, not a refutation of the Kuhnian picture. The paradigm still rules; the puzzle still gets solved; the community still shares the framework. Only the size of the community and the scale of the instruments have changed, and Kuhn’s concepts, suitably enlarged, may apply as well to the thousand-author collaboration as to the lone investigator in his laboratory.
Yet there is also a case for saying that something deeper has shifted, that the new scale changes not just the size but the character of scientific change. When a paradigm is embodied in billion-dollar machines, in instruments that take decades to build and the careers of thousands to operate, the costs of revolution rise enormously. A lone investigator can abandon a framework and strike out in a new direction at little cost but to herself. A collaboration of thousands, with billions of dollars of equipment designed around a particular paradigm, cannot pivot so easily; the sheer institutional and financial inertia of big science may make paradigm shifts slower, harder, more resistant than they were in the age of small communities. The very investment that makes modern science so powerful may also make it more conservative, more locked into its existing frameworks, less able to undergo the kind of revolutionary transformation that Kuhn described. The machines that extend a paradigm’s reach also entrench it, and an entrenched paradigm is harder to overthrow.
And then there is the newest and strangest development of all, the rise of the thinking machine as a participant in science. Increasingly, scientific discovery is being done not only by human beings but with the aid of artificial intelligence, by systems that can sift through oceans of data far too vast for any human mind, that can detect patterns no human would notice, that can even, in some domains, propose hypotheses and design experiments. What happens to Kuhn’s picture when the perceiving, puzzle-solving, paradigm-bound scientist is partly replaced or augmented by a machine? Kuhn’s whole theory rested on human psychology, on trained perception, on the gestalt switch, on the tacit skills and shared commitments of human communities. A machine does not undergo a gestalt switch; it does not perceive the world through a paradigm in the way a human does; it does not convert. Does the entry of artificial intelligence into science mean that the human, psychological dimension Kuhn emphasized is becoming less central, and with it his whole account of how science changes? Or will the machines themselves, in some fashion we do not yet understand, come to embody paradigms of their own, frameworks that shape what they can and cannot see?
These are genuinely open questions, and it would be dishonest to pretend that Kuhn, or anyone, has settled them. The science of the twenty-first century, with its vast collaborations, its monstrous machines, and its artificial minds, is in some respects a new thing under the sun, and the model Kuhn drew from the history of small communities of human investigators may need substantial revision to fit it. Perhaps the basic Kuhnian insights survive, suitably enlarged; perhaps they require fundamental rethinking; perhaps the very notion of a paradigm shift will itself shift, become something we cannot yet foresee, in the science of machines and megaprojects. What is certain is that the questions Kuhn raised, about how communities of inquiry are bound by shared frameworks and how those frameworks change, remain as pressing as ever, even if the communities and the frameworks now include silicon as well as carbon, and number their members in the thousands rather than the dozens.
There is a further wrinkle in the new science that some have hailed, prematurely, as a paradigm shift in the very nature of inquiry: the rise of data-driven science, in which discovery proceeds less by framing and testing theories than by trawling enormous datasets for patterns and correlations. Some enthusiasts have proclaimed that this spells the end of theory altogether, that with enough data and enough computing power we no longer need the explanatory frameworks that science has always built, that the numbers will speak for themselves. A Kuhnian should be deeply skeptical of this claim, and the grounds for skepticism are precisely Kuhn’s. Data never speak for themselves; they are always gathered, selected, and interpreted within some framework, however hidden, that determines which patterns count as meaningful and which as noise. The dream of a pure, theory-free, data-driven science is the old dream of neutral observation in a new technological costume, and it founders on the same rock: there is no innocent eye, not even an electronic one, for the machine too must be told what to look for, and the telling embeds a framework. The end of theory is not at hand; theory has merely gone underground, into the assumptions built into the algorithms, where it is harder to see and therefore more dangerous.
Could a machine, then, ever trigger a genuine paradigm shift, see what its human masters cannot, and force a revolution in some field? It is conceivable, and the possibility is genuinely novel. A system that detects, in a flood of data, a persistent pattern that no human framework predicts or accommodates might function as an inexhaustible generator of anomalies, surfacing the very mismatches that, in Kuhn’s account, drive a paradigm toward crisis. But here the limits of the machine reassert themselves. To recognize an anomaly as an anomaly, as something that matters, that threatens the framework, that demands a response, requires the judgment of a community that holds the framework and feels the threat. A machine can flag a deviation; it takes a human community to decide that the deviation is a crisis rather than a glitch, to lose confidence in the old paradigm, to imagine and embrace a new one. The gestalt switch, the conversion, the reweighting of values under uncertainty, these remain, for now, irreducibly human acts. The machine may hand us anomalies by the thousand; what we do with them is still up to us, and still proceeds, when it proceeds at all, by the human and communal process Kuhn described.
It is worth noting, too, that some fields of contemporary science have lately shown signs of exactly the kind of crisis Kuhn described, though not always over a grand theory. A number of disciplines, especially in the study of human behavior and biology, have been shaken by what is called a reproducibility crisis, the alarming discovery that many published results, when others try to repeat the experiments, cannot be reproduced. This is not a crisis of a single paradigm in the strict sense, but it has the unmistakable Kuhnian flavor of a community losing confidence in its own established methods, questioning practices it had long taken for granted, turning, in its distress, to foundational and almost philosophical reflection on what counts as a sound result. The response has involved a loosening of old certainties, a proliferation of proposed reforms, and a degree of soul-searching that the placid days of normal science never required, all of it recognizably the syndrome Kuhn anatomized. Whether these fields emerge with their old frameworks repaired or transformed remains to be seen, but the episode shows that the dynamics Kuhn identified are alive and at work in the science of our own moment, even where no single grand paradigm is at stake.
There is a deeper continuity, though, that survives all these changes, and it is worth ending on, for it is the heart of why Kuhn still matters in an age he never saw. Whatever the scale, whatever the machines, whatever the role of artificial intelligence, science remains a human enterprise in its essential purpose: it is our species’ great collective effort to understand the world, organized through communities that share frameworks and that must, somehow, change those frameworks when they no longer fit. The thousand authors of the gravitational-wave paper, the builders of the great colliders, the designers of the scientific machines, are still doing what Aristotle and Newton and Einstein did, still trying to make sense of nature, still working within inherited frameworks and still, when the time comes, transforming them. The forms have changed beyond recognition; the deep activity has not. And so Kuhn’s questions endure, transposed into a new key, because they were never really questions about the science of any particular era but about the permanent predicament of finite minds trying, together, to understand a world that always exceeds their current grasp. It is to that permanent predicament, and to a living example of it unfolding in our own moment, that we turn in the final chapters of this book.
Chapter Thirty-One — A Living Anomaly: A Constant Drawn from the Nucleus
I have spent this book describing how other people, in other centuries, lived through the transformations of science, and there is a certain safety in that distance. The revolutions are over; the winners are known; we can admire the courage of the revolutionaries and shake our heads at the blindness of the resisters, secure in the knowledge of how it all turned out. But Kuhn’s machinery is not a museum piece, and to feel its real force one ought to watch it turning on a case where the outcome is genuinely unknown, where no one yet knows who is the visionary and who the crank. I am going to take that risk now, and offer as my example a piece of science from my own working life, fully aware that in doing so I forfeit the historian’s comfortable neutrality and expose myself to the reader’s reasonable suspicion. I cannot tell you how this story ends, because it has not ended. I can only show you a living anomaly, and a candidate framework that proposes to resolve it, and let you watch Kuhn’s drama unfold in the present tense.
The anomaly is one of the most notorious in all of modern physics, sometimes called, with rueful affection, the worst prediction in the history of science. It concerns the energy of empty space. According to our best current theories of the very small, the vacuum, the seemingly empty nothing between the stars and between the particles, is not empty at all but seethes with restless activity, a froth of fleeting energy that ought, by the most straightforward calculation, to have an enormous density. And according to our best observations of the very large, the actual energy of empty space, revealed by the way the expansion of the universe is gently accelerating, is not enormous at all but almost inconceivably tiny. The gap between what the theory of the small predicts and what the observation of the large reveals is not a matter of being off by a little, or even by a lot. It is a discrepancy so vast that to write it out would require a number followed by some hundred and twenty zeros. The prediction and the measurement disagree more violently than any prediction and measurement have ever disagreed about anything, and no one knows why.
This is an anomaly in the fullest Kuhnian sense, and it is worth pausing to appreciate how perfectly it fits the pattern we have traced throughout this book. It is not a minor puzzle at the periphery of physics; it sits at the very heart of the framework, at the junction where our theory of the smallest things meets our theory of the largest. It has resisted resolution for decades, despite the efforts of many of the most brilliant minds of the age. It will not go away. And, most tellingly, the community has largely learned to live with it, to set it aside, to get on with normal science while this monstrous discrepancy looms unresolved in the background, exactly as the phlogiston chemists once lived with the awkward weight of their calcined metals and the Ptolemaic astronomers lived with their proliferating epicycles. The cosmological constant problem, as it is called, is the kind of deep, central, persistent anomaly that, in Kuhn’s account, marks a science ripe for crisis, a framework whose confident surface conceals a fault line running all the way down.
Now, an anomaly of this kind invites candidate resolutions, and the physics community has produced many, most of which try to explain why the enormous predicted energy is somehow cancelled or suppressed down to the tiny observed value. The framework I have spent years developing takes a different road, and I describe it here not to persuade you of its truth, which is not mine to assert and not the business of this book, but to use it as a specimen, a concrete example of what a candidate new framework looks like from the inside while it is still fighting for a hearing. The central idea can be stated without a single equation, in the language of metaphor that this book has used throughout. It proposes that the tiny energy of the vacuum that drives the cosmos apart is not a separate mystery to be cancelled, but is connected to, and may even be derivable from, the very same physics that gives ordinary matter its mass deep inside the atomic nucleus.
Let me unfold that metaphor a little, for it is the heart of the matter. Most of the mass of the ordinary world, the mass of every proton and neutron, of every atom in your body and every star in the sky, does not come from the tiny fundamental particles that make up these things. It comes, astonishingly, from the energy of their binding, from the furious activity of the forces that hold the quarks together inside each particle. The proton is heavy not because its ingredients are heavy but because of the seething energy of the strong force that confines them, a kind of internal storm whose energy, by the famous equivalence of energy and mass, is most of what we weigh. My programme proposes that this same internal storm, this energy of confinement that pervades the vacuum itself wherever the strong force operates, leaves a residue, a faint and almost perfectly cancelled remainder, and that this residue is precisely the tiny energy of empty space that we observe driving the cosmos apart. The largest fact about the universe, its accelerating expansion, would then be a distant echo of the smallest, the confinement of quarks within the nucleus. The constant that governs the cosmos would be drawn, quite literally, from the nucleus.
It is worth saying why a framework of this kind, whatever its ultimate fate, has the particular shape that tends to attract the partisans of a new paradigm, for it illustrates the values we examined in the chapter on theory choice. What such a programme offers, above all, is scope and a certain austere simplicity: the promise of connecting domains long held apart, of explaining with one stroke a quantity that the standard picture can only measure, of reaching, as the best frameworks do, beyond the problem it was first built to solve. A theory that ties the largest fact about the cosmos to the smallest fact about the nucleus, that claims to derive a stubborn constant rather than merely to insert it by hand, is the kind of theory that scores high on exactly those values, breadth and unifying power, that move scientists toward a bold newcomer. My own programme makes its wagers in this spirit, staking itself on predictions about the shape and history of the cosmos that observation may one day confirm or refute, and it is precisely by such risky wagers, as Popper rightly insisted, that a framework earns the right to be taken seriously. Whether the wagers will be won is not for me to declare; what matters here is that the framework has the form of a serious candidate, scoring on the values that draw scientists toward revolution, even as it remains, like every newcomer, weak where the incumbent is strong.
I will not pretend to the reader that this idea is established, accepted, or even widely known; it is none of these things, and to claim otherwise would be to betray everything this book has said about the difference between a candidate framework and an accepted one. What I want the reader to notice is not whether the idea is correct, which the future will decide, but how exactly it occupies the structural position that Kuhn taught us to recognize. It is a proposed new framework that addresses a deep and persistent anomaly. It promises a kind of unification, a bridge between the very small and the very large, that the reigning approach does not offer. It claims, if its proponents are right, to derive a quantity that the standard framework can only measure and cannot explain. And it asks the community to consider a reorganization of its picture, a new way of connecting domains that have long been kept apart. Whether such a framework deserves to win is exactly the kind of question that, as we have seen at length, cannot be settled by any neutral algorithm, but only by the long, contested, judgment-laden process that Kuhn spent his life describing.
There is something vertiginous, I confess, in applying Kuhn’s lessons to one’s own work, for they cut both ways and offer no comfort to the partisan. On the one hand, Kuhn teaches the proponent of a new framework to expect resistance, to understand that the community’s reluctance is not stupidity or malice but the normal and even healthy conservatism of a science protecting its hard-won paradigm, and to take heart that great revolutions have always faced exactly such resistance at their outset. On the other hand, Kuhn offers no assurance whatever that resistance means one is right. For every Copernicus dismissed and later vindicated, there are countless genuine cranks who were dismissed and stayed dismissed, who mistook the ordinary resistance to error for the resistance that precedes a revolution, who flattered themselves that their rejection by the establishment was the mark of a misunderstood visionary when it was simply the mark of being wrong. Kuhn gives the revolutionary a framework for understanding her situation; he gives her no grounds for confidence about which kind of revolutionary she is. That verdict belongs to a future she cannot command.
And so I offer my own programme to the reader not as a truth to be believed but as a window, a chance to see, from the inside and in the present tense, what it actually feels like to hold a framework that the community has not accepted, to live in the structural position of the would-be revolutionary while the outcome is still unknown. It is a position of hope and doubt intertwined, of conviction shadowed by the knowledge that conviction is no guarantee, of the daily experience of seeing the world through a framework that one’s colleagues do not share and cannot, at first, even quite understand. This is the human reality behind the dry word incommensurability, and the human reality behind the textbook story of revolutions, and I can describe it with a vividness no historian of dead controversies can quite achieve, because I am living it. How a framework in this position actually fights for its life, the concrete struggle for a hearing in the contemporary world of science, is the subject of the chapter that follows, and it is perhaps the most practical lesson that Kuhn’s thought has to teach.
Chapter Thirty-Two — How a New Paradigm Fights for Its Life
The textbooks make scientific revolutions sound clean and quick, a matter of a decisive insight that, once published, sweeps the field. The reality, as anyone who has tried to introduce a genuinely new framework can attest, is something else entirely: slow, grinding, often discouraging, a long campaign waged against the immense and largely justified inertia of an established science. To understand how a new paradigm actually fights for its life, one must set aside the heroic legend and look at the unglamorous machinery of persuasion, attention, and communication through which a new idea must pass if it is ever to be heard at all. Kuhn understood this machinery better than anyone, and his insights illuminate the struggle from the inside with a clarity that, having lived the struggle myself, I can vouch for.
The first and most basic obstacle is not disagreement but inattention. A working scientist is besieged by claims, drowning in papers, with far more demands on her attention than she could ever meet, and her sensible default, the default that normal science trains into her, is to ignore almost everything that does not bear directly on the puzzles she is currently solving within her accepted framework. A new framework, especially one that comes from outside the central institutions, that crosses the boundaries between specialties, that asks her to reconsider assumptions she has never thought to question, faces not a reasoned rejection but something harder to overcome: a simple failure to register, a sliding-off of attention, the fate of being not refuted but unread. Before a new paradigm can be resisted it must first be noticed, and being noticed, in the vast and fragmented world of contemporary science, is itself an enormous achievement, requiring a sustained labor of communication that has little to do with the merits of the idea and everything to do with the sociology of attention.
This is why the proponent of a new framework must become, of necessity, something the heroic legend never mentions: a tireless communicator, a writer of letters, a giver of talks, a builder of the slow personal relationships through which, in the end, new ideas actually spread. Kuhn’s account of science as a communal enterprise, bound together by personal contact and shared training, has a practical corollary that the would-be revolutionary learns quickly: ideas do not spread by their merits alone, broadcast impersonally into the void, but through the human channels of a community, from person to person, in the patient cultivation of colleagues willing to give the new framework the attention it needs to be understood. I have spent a great deal of my own effort on exactly this unglamorous work, reaching out to physicists and cosmologists across the world, not because the science is settled by such outreach but because, without it, even the best framework would simply never be heard, would die not refuted but ignored, a tree falling in an empty forest.
When attention is finally won, the second obstacle appears, and it is the one Kuhn named most famously: incommensurability, the genuine difficulty of communication across the boundary between frameworks. A scientist who has spent her career within the standard approach to a problem does not merely disagree with a new framework; she often cannot, at first, fully understand it, because its key terms carry meanings, and its concepts presuppose connections, that her own training has not prepared her to grasp. She reads the new proposal through the lexicon of her own paradigm, and through that lexicon it looks confused, or trivial, or simply wrong, in the way that Aristotle looked wrong to the young Kuhn before his afternoon of conversion. The proponent of the new framework finds herself in the position of the translator we met earlier, trying to teach a second conceptual language to listeners who already speak a first and who keep hearing the new words as garbled versions of the old. The labor of being understood, of bridging the incommensurable gap, is far harder than the labor of being right, and many a sound idea has foundered not because it was refuted but because it could not be made intelligible across the divide.
And here Kuhn’s analysis of theory choice, which we examined in an earlier chapter, becomes intensely practical. Because the shared values of science, accuracy, scope, simplicity, consistency, fruitfulness, do not by themselves dictate a unique choice, the proponent of a new framework cannot simply present a proof and compel assent. She must persuade, must argue that her framework, weighed across all the values, deserves to be preferred, or at least deserves the patient development that might one day reveal its power. And she must do this knowing that a new framework is always weakest at its birth, riddled with unsolved problems, unable yet to match the achievements that the established paradigm has accumulated over decades of normal science. She is asking the community to back a newcomer that cannot yet do what the incumbent does, on the promise of what it might one day do, and this is a hard sell, made entirely on the currency of fruitfulness and promise against the established currency of proven accuracy and scope. The new paradigm fights for its life on the most unfavorable terms, with its weaknesses fully visible and its strengths still mostly potential.
Kuhn’s sobering observation about generations hangs over this whole struggle, and the honest proponent must face it squarely. Recall the bleak wisdom, which Kuhn drew from the history of science and from Max Planck before him, that new frameworks often triumph not by converting the established practitioners but by outliving them, winning over the young who have not yet hardened into the old paradigm while the old guard goes to its grave unconvinced. This is a hard truth for anyone who hopes to see her framework vindicated within her own lifetime, for it suggests that the relevant audience may not be one’s contemporaries at all but the next generation, the students not yet committed, the young minds still open to a new way of seeing. There is a peculiar loneliness in this, the loneliness of working for a verdict one may not live to hear, of casting one’s ideas forward to a future court. And yet there is also a strange consolation in it, for it means that the apparent indifference of one’s contemporaries, however discouraging, is not the final word, that the real trial may be held in a courtroom that has not yet convened.
Kuhn’s observation about generations has a constructive corollary that the discouraged revolutionary easily overlooks, and I have come to regard it as one of the most important. If new frameworks win chiefly by being adopted by the young, then the cultivation of the young is not a distraction from the real work but a part of it, perhaps the most consequential part. To take seriously a student not yet hardened into the reigning paradigm, to teach a young mind a new way of seeing before the old way has become second nature, is to plant a seed in the only soil where it can truly take root. I have found, in my own work, that some of the most genuine engagement with a new framework comes not from established experts, for whom it would mean unlearning a lifetime, but from the young, who come to it fresh and can inhabit it without the painful labor of conversion. There is a long view in this, a willingness to work not for the verdict of one’s own moment but for the understanding of those who will still be working when one is gone, and it transforms the apparent indifference of one’s contemporaries from a defeat into something more like a relay, in which the task is less to win the present race than to pass the framework, intact and intelligible, into hands that will carry it further than one’s own ever could.
I want to be scrupulously honest with the reader about what all this does and does not imply, because it would be easy to misuse Kuhn’s lessons as a kind of self-flattery, and the temptation is real. The fact that a new framework faces resistance, inattention, incommensurability, and the slow grind of generational change does not, in itself, count in the slightest toward the framework being correct. These are the obstacles that every new framework faces, the sound ones and the worthless ones alike, the future Copernican revolutions and the endless parade of well-meaning errors. Kuhn’s analysis explains the structure of the struggle; it does not predict the winner. To take comfort in the mere fact of resistance, to reason that because revolutionaries are resisted and I am resisted I must be a revolutionary, is a fallacy so seductive and so common that it has become the special vanity of cranks. The honest proponent holds two things together: the courage to persist against resistance that may be merely the normal friction a true revolution must overcome, and the humility to know that the same resistance may simply mean she is wrong. Which it is, she cannot know, and pretending to know is the first step toward the very dogmatism that Kuhn taught us to see through.
So I offer my own struggle, like my own framework, as an illustration rather than a plea, a chance to see Kuhn’s machinery at work in a living case. Whether the particular framework I have described will win its long campaign, whether the constant drawn from the nucleus will one day be a chapter in the textbooks or a footnote among the discarded conjectures, I do not know and cannot know, and the not-knowing is the whole point. What I do know is that the shape of the struggle, the inattention and the incommensurability and the persuasion and the waiting on a future verdict, is exactly the shape that Kuhn described, and that to have lived it is to understand, in a way no second-hand account can convey, how profoundly right he was about the human reality of scientific change. The drama he anatomized is not safely in the past. It is happening now, in a thousand quiet struggles over a thousand candidate frameworks, most of which will fail and a few of which will remake our picture of the world, and no one living can yet tell which are which. That is the predicament of science in the present tense, and it is the predicament Kuhn taught us to understand.
Chapter Thirty-Three — After Kuhn: A Maturer Picture of Change
We have come a long way from the tidy universe with which this book began, the comfortable picture of science as the patient accumulation of established facts, marching by clear rules and decisive tests toward a complete and final truth. Kuhn shattered that picture, and we have followed the shattering through all its consequences: the paradigms and their normal science, the anomalies and crises and revolutions, the incommensurable worlds and the gestalt switches, the controversies with Popper and Lakatos and Feyerabend, the excesses of the sociologists and the explosion of the Science Wars, the late evolutionary turn and the strange phenomenon of scientific forgetting. The question that now presses, as we prepare to take our leave of this remarkable thinker, is what we are to put in the place of the tidy universe he destroyed. If the old comfortable picture is false, and if the crude relativism that some drew from Kuhn is also false, what is the mature, balanced, honest picture of scientific change that we should carry away?
The first element of that mature picture is the recognition that science is genuinely shaped by the frameworks within which it works, and that there is no neutral, framework-free standpoint from which to do science or to judge between frameworks. This is Kuhn’s deepest and most secure insight, and it survives all the controversies. Observation is theory-laden; perception is trained; the questions a science asks and the facts it gathers and the standards it applies are all conditioned by the paradigm it holds. There is no innocent eye, no pure given, no view from nowhere. We always encounter the world through some framework, and our frameworks are historical achievements that have been otherwise in the past and may be otherwise in the future. To understand science is to understand it as a perspectival enterprise, conducted always from within some particular way of carving up and interrogating the world.
The second element, and it must be held together with the first on pain of sliding into relativism, is that this perspectival, framework-bound character of science does not cancel its objectivity, its rationality, or its genuine answerability to a real world. The world pushes back. Not every framework survives; the recalcitrance of nature generates the anomalies that bring paradigms down; reality constrains, even if it does not dictate, what we can successfully believe. And the choice between frameworks, though it is not governed by any neutral algorithm, is governed by shared values and by reasoned judgment, the same kind of judgment we recognize as rational in the wise judge or the skilled physician, judgment that can be exercised well or badly and defended with reasons even where it cannot be reduced to a rule. Science is both conditioned and objective, both perspectival and answerable to reality, and the whole art of understanding it lies in holding these two truths together without letting either swallow the other. This is the delicate balance that Kuhn himself struggled to maintain, the knife-edge between the tidy universe and the abyss of relativism, and it is the balance that a mature picture of science must learn to keep.
The third element is a chastened and complicated understanding of progress. Science progresses; this is not in doubt, and no honest account denies it. But the progress is not the simple accumulation of an ever-growing pile of truths, nor is it obviously a steady approach toward a single final theory of everything. It is a progress that involves real losses as well as gains, that abandons old questions as it answers new ones, that branches and diversifies more than it converges, that moves, perhaps, more in the manner of biological evolution, away from what came before and toward ever-greater adaptation, than in the manner of a climber approaching a fixed summit. Whether, behind all this branching and adapting, our science is also converging on a framework-independent reality remains one of the great open questions, and a mature picture does not pretend to have settled it. What a mature picture does is hold the genuine reality of progress together with an honest acknowledgment of its complexity, refusing both the naive triumphalism that sees only accumulation and the corrosive cynicism that denies progress altogether.
The fourth element is an appreciation of the essential role of the scientific community, and of the human and social dimension of knowledge that the tidy universe had scrubbed away. Science is not the work of isolated rational intellects confronting nature alone; it is the work of communities, bound together by shared training, shared exemplars, shared values, and shared tacit skills, transmitted from master to pupil through the personal channels of a living tradition. The rationality of science lives not only, perhaps not even chiefly, in the individual scientist, but in the community, in the way its shared standards and its internal diversity allow it to do collectively what no individual could do alone, to hedge its bets, to spread its risks, to maintain a framework long enough to develop its power and yet to transform that framework when the time comes. To understand science is to understand it as a profoundly social achievement, and this social understanding, which Kuhn did more than anyone to establish, is among the most important and most lasting elements of the picture he bequeathed us.
It is worth noting that this mature picture quietly absorbs the best of Kuhn’s great antagonists rather than simply defeating them, for the truest reading of the whole controversy is that each of the major parties had hold of a real piece of the truth. From Popper it keeps the insistence that science must take risks, must make bold conjectures that expose themselves to refutation, must never collapse into the mere defense of dogma; the demand for testable, falsifiable commitments remains a genuine mark of scientific seriousness even if it does not describe the whole of scientific life. From Lakatos it keeps the crucial distinction between frameworks that progress, generating new predictions and opening new territory, and those that merely degenerate, patching their defenses without ever advancing; this remains our best practical tool for telling living science from dying. From Feyerabend it keeps the salutary warning against any rigid method, the recognition that science has advanced by many roads and that no single algorithm captures its creativity, along with the humane suspicion of any framework that claims the right to silence all others. And from Polanyi it keeps the deepest insight of all, that knowledge is personal, tacit, committed, and reality-seeking, that the knowing scientist is not a defect to be scrubbed from the picture but the very means by which we make contact with the world. The mature picture is not Kuhn against the others but Kuhn with the others, each correcting the excesses of the rest, together composing a fuller understanding than any one of them could give alone.
The fifth element is a certain humility about our own present frameworks, a humility that is perhaps the deepest practical fruit of the whole Kuhnian revolution. If past scientists, no less brilliant and no less honest than ourselves, held frameworks that have since been overthrown, frameworks that seemed to them as solid and obvious as ours seem to us, then intellectual honesty requires us to suppose that our own frameworks, too, are provisional, that future revolutions may overturn what we now take for granted, that there are questions we are dismissing and phenomena we are failing to see because our particular paradigm has trained us not to. This is not a counsel of despair or doubt; we are entitled, indeed obliged, to work confidently within our best current frameworks, to build on them and trust them and extend them. But it is a counsel of humility, a reminder that even our finest knowledge is the knowledge of finite beings working within a historical framework, and that the last word, in science, is never actually the last word. To hold our convictions firmly and yet provisionally, confidently and yet humbly, is the difficult wisdom that Kuhn’s history of science teaches, and it is a wisdom we need in every domain of life where we are tempted to mistake our current framework for the final truth.
These five elements, the framework-bound character of science, its nonetheless genuine objectivity, the chastened complexity of its progress, the essential role of the community, and the humility about our own paradigms, together compose the mature picture of scientific change that emerges when one reads Kuhn carefully, holds him at his best, and protects him from both his careless admirers and his uncharitable critics. It is a picture neither as comforting as the tidy universe nor as alarming as the relativist caricature, a picture that asks us to live with more complexity and more uncertainty than either of those simpler stories, but that has the great merit of being true to the actual history and the actual practice of the most powerful form of knowledge our species has devised. To have given us this picture, against the resistance of entrenched assumptions and at the cost of being perpetually misunderstood, is the achievement of the quiet, anxious, careful historian whose afternoon with Aristotle set the whole thing in motion, and it is an achievement that has permanently changed how thoughtful people understand the enterprise of knowing. What that change means for us, for how we live and think and hold our certainties in a world that science has made and is forever remaking, is the question on which this book will close.
Conclusion
A book about how knowledge breaks and remakes itself ought, in honesty, to turn its lesson upon itself at the end, and so I will begin this conclusion with a confession. Everything I have told you about Thomas Kuhn is itself an interpretation, offered from within a framework, shaped by the questions I thought worth asking and the connections I thought worth drawing. Another writer, standing within a different framework, would have given you a different Kuhn, emphasizing what I have slighted and slighting what I have emphasized. There is no neutral, framework-free account of Kuhn any more than there is a neutral, framework-free account of nature, and the reader who has truly absorbed his lesson will hold even this book a little loosely, as one perspective among the possible ones. That is not a weakness to be apologized for; it is the human condition of all understanding, the very condition that Kuhn spent his life teaching us to recognize. To read him well is to read everything, including him, with that recognition firmly in mind.
What, then, has this strange and anxious man left us, when all the controversies are tallied and all the misreadings set aside? He has left us, above all, a new way of seeing the most powerful thing our species does. Before Kuhn, the educated person understood science as a ladder, a steady climb toward truth governed by clear rules and impartial evidence. After Kuhn, that person cannot help but see science differently: as a human enterprise conducted by communities, bound by frameworks they mostly cannot see, advancing through long stretches of patient work punctuated by wrenching transformations, shaped at every level by the trained perceptions and shared commitments of its practitioners. This is not a smaller or a meaner vision of science than the one it replaced; it is a larger and a truer one, a vision that does justice to the actual messy magnificent reality of how knowledge is made, and that takes the human beings who make it seriously rather than scrubbing them out of the picture in the name of a false objectivity.
It is worth being clear, at the close, about what Kuhn did not do, because the misunderstandings on this point have done real damage. He did not show that science is irrational, that truth is whatever the powerful decree, that one theory is as good as another, that reality is a mere social construction with no constraint from the world. These conclusions were drawn in his name by others, and he spent the second half of his life rejecting them, with a vehemence that his relativist admirers found inexplicable and that his realist critics refused to credit. The real Kuhn believed that science is rational, that it makes genuine progress, that the world constrains our theories and rules many of them out, that the choice between frameworks is governed by real and shared values. What he denied was only that this rationality takes the form of a neutral algorithm, that this progress is a simple accumulation, that this constraint comes to us in a pure framework-free form. He complicated the picture of scientific rationality; he did not abolish it. To confuse the complication with the abolition is the single most common and most damaging error in the reception of his thought, and if this book has helped a few readers avoid it, that alone will have justified the writing.
There is a particular relevance to all this in our own moment, which I would be remiss not to name. We live in an age of loud and confident assault on the authority of science, an age in which the failures and revisions of expert opinion are seized upon as proof that the experts know nothing and that one opinion is as good as another. In such an age, a careless version of Kuhn can be, and has been, pressed into the service of a corrosive anti-intellectualism, cited to suggest that since science is just a succession of paradigms, no paradigm has any greater claim on us than any other, and we may believe whatever we like. This is a grotesque misuse, and the careful Kuhn is its antidote rather than its ally. For what Kuhn actually teaches is how to understand the revisions of science not as failures but as the very signature of a healthy enterprise, how to distinguish the genuine self-correction of a living science from the mere thrashing of a pseudoscience, how to trust the scientific community precisely because it changes its mind in the disciplined way that he described. To understand Kuhn rightly is to be inoculated against both the naive worship of science and the cynical dismissal of it, and in an age torn between these two errors, that balanced understanding is something close to a civic necessity.
The deepest gift of Kuhn’s thought, though, is not a doctrine about science at all but a habit of mind, one that extends far beyond the laboratory into every corner of life where human beings hold convictions. It is the habit of asking, of any settled certainty, what framework lies beneath it, what assumptions it rests upon that could be otherwise, what anomalies it is quietly ignoring, what alternative way of seeing it has trained us not to consider. Once you have learned to see paradigms in science, you begin to see them everywhere: in medicine, in economics, in politics, in your own profession, in the unspoken assumptions that govern how you and everyone around you decides what is true and what is possible. You begin to notice the things that polite consensus agrees to overlook, the early tremors that may signal a coming crisis, the quiet possibility that the framework everyone takes for granted may one day be set aside. This is an uncomfortable gift, for it robs us of the easy certainty that our own way of seeing is simply the way things are. But it is also a liberating one, for it opens the mind to the possibility of genuine novelty, to the recognition that the boundaries of the thinkable are not fixed, and that the future may hold ways of understanding the world that we cannot now even imagine.
I have tried, in the closing chapters, to make this abstract lesson concrete by offering a living example from my own work, a candidate framework still fighting for the hearing that will determine whether it is a contribution or a curiosity. I did this not to persuade you of its merits, which are not the business of a book about Kuhn, but to let you feel, from the inside, what it is to inhabit the structural position that Kuhn anatomized, to hold a framework the community has not accepted while the verdict remains unknown. There is no comfort in that position, only the strange double awareness of conviction shadowed by doubt, of working for a judgment one may not live to hear, of seeing a world that one’s colleagues do not yet see and may never see. But there is also a peculiar dignity in it, the dignity of participating, however modestly, in the great ongoing human project of remaking our understanding of the world, the same project that runs from Aristotle through Copernicus and Newton and Einstein and onward into a future none of us can foresee. Whether my own small contribution to that project will endure or vanish, I cannot say. What I can say is that to have understood Kuhn is to understand the project itself, and to find one’s place within it with both the requisite courage and the requisite humility.
There is a question that haunts the whole of Kuhn’s thought, and that he never finally answered, and it is fitting to leave it open rather than to pretend a resolution he could not reach. The question is whether, behind all the changing frameworks, all the incommensurable paradigms succeeding one another down the centuries, there is a single reality that our science is gradually, imperfectly, learning to know, or whether there are only the frameworks themselves, each adapting better than the last to the problems it confronts, with no fixed truth toward which the whole procession tends. Kuhn flinched from the first answer and could not quite embrace the second, and spent his life suspended uneasily between them. I do not think the question can be settled from the armchair, and perhaps it cannot be settled at all; perhaps it is one of those deep questions that each generation must confront afresh, in the light of its own science and its own philosophy. But I am inclined to believe, with Polanyi against the later Kuhn, that the very passion of the search, the conviction that drives the scientist to seek a truth she has not yet found, is itself a kind of evidence that there is something there to be found, that we are not merely spinning frameworks in the void but reaching, however imperfectly, toward a reality that exceeds and resists and instructs us. But I hold this belief, as Kuhn taught me to hold all beliefs, firmly and provisionally at once, ready to be revised by a future I cannot command.
It is striking, looking back over the whole story, how much of it turns on a single quality of mind that Kuhn possessed in abundance and that we would do well to cultivate in ourselves: the capacity to take seriously what seems, at first, to be nonsense. The young Kuhn, faced with an Aristotle who appeared to be talking gibberish about motion, did not dismiss him but stopped, and assumed the fault might lie with the reader rather than the read, and was rewarded with a vision that changed his life and, through him, our understanding of science. This is the historian’s deepest discipline, the willingness to extend to the strange and the foreign and the apparently absurd a charity that assumes there may be sense there that we have not yet learned to see. It is a discipline desperately needed in our own polarized and impatient age, in which the reflexive dismissal of whatever does not fit our framework has become a kind of intellectual reflex. Kuhn teaches us to pause before that dismissal, to ask whether the apparent nonsense might be sense in a language we have not learned, to grant to those who see the world differently the charity we would wish for our own strange ideas. That charity is the moral heart of his method, and it is perhaps the most valuable thing this anxious, careful, endlessly qualifying man has to teach us.
In the end, Thomas Kuhn was a man who looked closely at how human beings come to know the world, and who had the courage to report what he saw even when it overturned the comfortable assumptions of his age and brought down upon him decades of misunderstanding. He saw that knowledge is harder and stranger and more human than we had supposed, that it is made by communities working within frameworks they mostly cannot see, that it advances through transformations as much as accumulations, that it is shaped through and through by the perspectives of those who make it and yet remains answerable to a world that pushes back. He could not tie all of this into a finished system, and he died with his great late work unfinished, the keystone never quite set. But he gave us something more valuable than a finished system: a new and truer way of seeing the most powerful thing we do, and a habit of humility about our own certainties that we need now more than ever.
The story of how knowledge breaks and remakes itself has no ending, because the breaking and remaking go on, in our own laboratories and our own minds, in this very moment as you read these words. Somewhere a young scientist is staring at an anomaly that will not go away, beginning the long uneasy descent into crisis. Somewhere a framework that everyone takes for granted is quietly accumulating the strains that will one day bring it down. Somewhere a new way of seeing is fighting for the hearing that will determine whether it is a revolution or a footnote. We are all of us inside the story Kuhn told, living through transformations we are too close to recognize, holding frameworks that the future will revise, reaching toward a understanding of the world that we will never finally possess. To know this, to hold our certainties firmly and humbly, to extend charity to the strange and courage to the new, to participate with both confidence and humility in the endless human project of understanding, is the wisdom that Thomas Kuhn, against all his own anxieties and all his readers’ misunderstandings, finally has to give. It is a wisdom worth carrying out of these pages and into the world, where the great work of knowing goes restlessly, gloriously, on.
The Copernican Revolution
The shift from an earth-centered to a sun-centered cosmos is the founding example of a paradigm change, and it overturns the schoolbook legend at every point. When Copernicus published in 1543, his system was not obviously more accurate than the fourteen-hundred-year-old Ptolemaic astronomy it challenged; it still used circular orbits and epicycles, it contradicted the plain evidence of the senses and the reigning physics, and it faced an apparently fatal objection in the absence of any detectable shift of the stars. Most competent astronomers sensibly rejected it for the better part of a century. It triumphed only gradually, as Kepler supplied accurate ellipses, Galileo found supporting heavens through his telescope, and Newton provided a physics in which a moving earth made sense. The case shows that a new paradigm is typically weaker, not stronger, at its birth, and that revolutions are won over generations through the development of promise rather than by a single decisive proof.
Phlogiston and the Chemical Revolution
For most of the eighteenth century, chemists explained burning by the release of a substance called phlogiston. The theory was productive but carried a nagging anomaly: when certain metals burned, the residue weighed more, not less, though phlogiston was supposedly being lost. For decades this was treated as a minor puzzle, patched with ingenious contrivances, even the desperate suggestion that phlogiston had negative weight. The same fact that was a marginal embarrassment under the old framework became the central clue under the new one, when Lavoisier recognized that burning involves combining with something from the air. The case illustrates how the significance of an anomaly depends entirely on the framework that beholds it: a nuisance to one paradigm, a foundation stone to the next.
The Birth of Quantum Theory
The textbook says quantum theory began in a single revolutionary act when Max Planck introduced energy quanta in 1900. Kuhn’s detailed historical study argued that Planck did not at first understand himself to be claiming a real physical discontinuity; he remained, in his own mind, a classical physicist, and the radical idea that energy is genuinely quantized emerged only years later, chiefly through Einstein and Ehrenfest. The case is a perfect specimen of how disciplines manufacture clean founding myths, complete with a single hero and a single birthday, over a reality that was gradual, collective, and shot through with mutual misunderstanding. It also gently embarrasses the dramatic revolutionary model, for what Kuhn found was a creeping transformation rather than a sudden conversion.
Neptune and the Phantom Vulcan
When the planet Uranus wandered from its Newtonian orbit, astronomers postulated an unseen planet, calculated its position, and triumphantly found Neptune; the anomaly that might have refuted Newton instead confirmed him. When the planet Mercury also wandered, astronomers applied the same reasoning and searched for decades for a planet they named Vulcan, which does not exist; Mercury’s anomaly was finally dissolved only by Einstein’s new theory of gravity. Two anomalies, treated identically, with opposite outcomes. The pair demonstrates that no rule tells a scientist in advance which anomaly is a mere puzzle to be patched and which is the crack that signals an earthquake; that judgment is human, fallible, and not dictated by the evidence.
Röntgen and the Discovery of X-rays
In 1895 Wilhelm Röntgen noticed a screen glowing faintly across his darkened laboratory when, by every expectation, it should have stayed dark. Instead of dismissing the glow as a mistake, he pursued the anomaly and discovered a new kind of ray that passed through flesh to cast the shadow of his own bones. The case captures Kuhn’s counterintuitive claim that discovery begins with the violation of expectation, and that it is the precise, demanding expectations of a strong paradigm that make such violations visible in the first place. A vaguer framework, expecting anything, could have been surprised by nothing.
Wegener and Continental Drift
When Alfred Wegener proposed in the early twentieth century that the continents had drifted across the globe, the geological establishment dismissed him, partly because he could offer no acceptable mechanism for how continents could plow through solid ocean floor. For decades the idea languished as the eccentric notion of an outsider. Only in the 1960s, with the discovery of sea-floor spreading and a workable mechanism, did the framework triumph as plate tectonics, transforming the science after most of its original opponents had passed from the scene. The case illustrates both the rational core of resistance, the legitimate demand for a mechanism, and Kuhn’s sobering observation that frameworks often win by outliving their opponents rather than converting them.
Semmelweis and the Doctors Who Would Not Wash
In the 1840s, Ignaz Semmelweis discovered that deaths from childbed fever plummeted when physicians washed their hands before deliveries, yet his findings were resisted and ridiculed by a medical establishment that had no framework to make sense of invisible contamination and was affronted by the suggestion that doctors themselves carried death. Semmelweis died disgraced, his insight vindicated only later when germ theory supplied the framework that made his observation intelligible. The case shows how an observation can be correct and even life-saving yet fail to be accepted because the community lacks the paradigm needed to understand it, and how the want of a conceptual home, not the want of evidence, can doom a true idea.
The Stomach Ulcer Revolution
For most of the twentieth century, medicine confidently held that stomach ulcers were caused by stress and acid, and the idea that they might be caused by a bacterium was dismissed as absurd, since everyone knew bacteria could not survive the stomach’s acid. When Barry Marshall and Robin Warren proposed exactly this in the 1980s, they met fierce resistance, and Marshall resorted to drinking a culture of the bacterium himself to prove the point by giving himself the disease. The framework eventually shifted, and the two later received a Nobel Prize. The case is a modern, fully documented paradigm shift, complete with entrenched resistance, an anomaly long explained away, and the dramatic gesture required to force a complacent community to look.
Darwin and the Tree of Life
Before Darwin, the development of life was widely assumed to be heading somewhere, ascending toward higher forms with humanity at the summit. Darwin explained the entire pageant without any goal at all, as a process of variation and selection that adapts life to its circumstances and branches it into endless forms, moving away from its origins without aiming at any destination. Kuhn seized on this as the model for scientific progress itself: science, too, he proposed, advances from primitive beginnings rather than toward a final truth, becoming ever better adapted without converging on a fixed goal. The case supplies the central analogy of Kuhn’s mature thought and his most radical revision of what progress means.
Newton’s Action at a Distance
Newton’s law of gravitation was a triumph of unprecedented power, yet it abandoned something the older Cartesian vortex theory had provided: a mechanical explanation of how gravity works through direct contact. Newton could describe gravity’s action across empty space with stunning accuracy but could offer no mechanism for it, and famously refused even to hypothesize one. To his critics this looked like a relapse into the occult, a willingness to describe rather than explain. The case is the classic example of Kuhn-loss: a revolution that, for all its gains, genuinely abandoned a question the previous framework had taken seriously, showing that science does not only ever add but sometimes also forgets.
The Ether and the Michelson-Morley Experiment
Nineteenth-century physics assumed that light waves traveled through an invisible medium called the ether, and in the 1880s Michelson and Morley designed an exquisitely sensitive experiment to detect the earth’s motion through it. They found nothing: no trace of the expected effect. This stubborn null result became one of the great anomalies that strained classical physics, and its eventual resolution required Einstein’s special relativity, which dispensed with the ether altogether and recast space and time themselves. The case shows an anomaly at work at the foundations of a framework, and a resolution that came not by patching the old picture but by reorganizing the most basic concepts of physics.
The Ptolemaic Epicycles
To keep the earth-centered cosmos in agreement with observation, astronomers added circles upon circles, epicycles upon epicycles, adjusting and complicating the system over the centuries until it grew into a baroque tangle of wheels within wheels. It still worked, more or less, but its mounting complexity bred a vague dissatisfaction, a sense that the true architecture of the heavens could hardly be so cumbersome. The case illustrates the slow accumulation of strain that precedes a crisis, and the way a framework can be progressively patched into monstrous complexity, technically functional yet quietly losing the confidence of those who handle it daily.
The Anomalous Playing Cards
In a classic psychology experiment of the 1940s, Bruner and Postman briefly showed subjects playing cards, some of them doctored so that suits and colors were reversed, such as a red six of spades. For short exposures, subjects confidently saw the anomalous cards as normal ones, their minds forcing the unexpected stimulus into familiar categories. Only with longer viewing did they grow uneasy, then suddenly perceive the cards as they truly were. The experiment gave Kuhn a laboratory demonstration of his whole theory in miniature: expectation shaping perception, anomaly resisting recognition, mounting unease, and a sudden gestalt switch into a new way of seeing.
The Sokal Hoax and the Science Wars
In 1996 the physicist Alan Sokal submitted a deliberately nonsensical article, dressed in fashionable theoretical jargon and arguing that physical reality is a social construct, to a leading journal of cultural studies, which published it; he then revealed the hoax. The bomb detonated at the height of the Science Wars, a bitter conflict between scientists defending the objectivity of science and humanists championing its social construction. The case shows how Kuhn’s carefully hedged ideas, pushed by others into crude relativism, provoked a polarized war in which his own subtle middle position was invisible to both sides, and in which the chief casualty was nuance itself.
The Reproducibility Crisis
In recent years several fields, especially in the study of human behavior and biology, have been shaken by the discovery that many published findings cannot be reproduced when others repeat the experiments. This is not a crisis over a single grand theory, but it has the unmistakable Kuhnian flavor of a community losing confidence in its own established methods, questioning practices long taken for granted, and turning in its distress to foundational, almost philosophical reflection on what counts as a sound result. The case demonstrates that the dynamics Kuhn anatomized, the onset of crisis and the loosening of settled rules, are alive in the science of our own moment.
The Detection of Gravitational Waves
In 2016 scientists announced the direct detection of gravitational waves, ripples in space and time predicted by Einstein a century earlier, in a paper bearing more than a thousand authors. The achievement was a magnificent confirmation of an existing paradigm, accomplished by a vast international collaboration operating instruments of extraordinary scale. The case raises the question of whether Kuhn’s model, drawn from the age of individual discoverers, still fits the era of Big Science, and suggests an answer: the fundamentals endure, for this is normal science extending and confirming a reigning framework, but the enormous scale and cost of modern instruments may make future paradigm shifts slower and harder than they once were.
The Keynesian Revolution in Economics
The transformation of economic thought associated with John Maynard Keynes in the mid-twentieth century was widely described as a revolution, and economists found in Kuhn a ready framework for understanding the succession of dominant schools, classical, Keynesian, monetarist, and their rivals, as reigning paradigms challenged by anomalies and overthrown. The case illustrates both the genuine reach of Kuhn’s ideas beyond physics and the perils of the migration, since the social sciences, marked by competing schools rather than a single unifying framework, may be pre-paradigmatic in Kuhn’s strict sense even as they eagerly adopt his vocabulary out of a longing for the authority that the physical sciences command.
The Double Helix: Myth and Reality
The discovery of the structure of the genetic molecule is remembered as a clean and inevitable triumph in which two young men saw the double helix and revealed the secret of life. The actual history is a tangle of false models, lucky guesses, borrowed and arguably misappropriated data, fierce rivalry, gendered injustice, and sheer contingency. None of this mess survives in the textbook, which trades the messy truth for a tidy founding myth more useful for training the young. The case exemplifies Kuhn’s thesis that science systematically rewrites its own past, smoothing the human reality of discovery into a legend of lone genius ascending a ladder.
Mendel’s Delayed Harvest
Gregor Mendel worked out the basic laws of heredity in the 1860s, publishing results of real importance, and the scientific world essentially ignored them for some thirty-five years, until they were rediscovered around 1900 by researchers who had independently reached similar conclusions. Mendel had been right and had been heard by almost no one, his work appearing in the wrong place, in the wrong form, before a community prepared to grasp its significance. The case echoes the fate of Ludwik Fleck and underscores one of the quiet lessons running through the whole story: being first and being right are not enough, for an idea must also be heard, and being heard is a matter of community, timing, and luck.
The Cosmological Constant Problem
Our best theory of the very small predicts that empty space should hold an enormous energy; our best observations of the very large reveal that its actual energy is almost inconceivably tiny, the two disagreeing by a factor written as one followed by some hundred and twenty zeros, the worst quantitative prediction in the history of physics. The discrepancy has resisted resolution for decades, sits at the foundational junction where the physics of the small meets the physics of the large, and has largely been set aside while normal science proceeds around it. The case is a living anomaly in the fullest Kuhnian sense, the kind of deep, central, persistent mismatch that marks a science ripe for crisis, and it serves as the contemporary specimen on which this book watches Kuhn’s machinery turning in the present tense.
Glossary
A priori Knowledge or structure that comes before experience and makes experience possible. Kuhn suggested that a paradigm acts like a changeable a priori: a framework that must be in place for a scientist to perceive her world, but which can itself shift in a revolution.
Accuracy One of the shared values by which scientists judge theories: how well a theory agrees with observation and experiment. Like the other values, it guides choice but does not by itself dictate it.
Actor-network theory A later, radical approach in the study of science that describes facts as built from alliances among people, instruments, and things, treating human and non-human participants with deliberate even-handedness.
Algorithm A fixed, mechanical procedure that yields a definite result. Kuhn denied that theory choice follows any algorithm: there is no formula that takes the evidence and outputs the correct paradigm.
Anomaly A result that does not fit the expectations of the reigning paradigm. Most anomalies are minor puzzles, solved by ordinary means; a deep anomaly that resists all solution can drive a science toward crisis.
Anti-realism The view that science does not, or cannot, give us a true description of a mind-independent reality, but only useful frameworks for organizing experience.
Articulation The patient work of normal science that develops a paradigm: determining its constants more precisely, resolving its ambiguities, and extending it to new cases.
Auxiliary hypothesis A supporting assumption, separate from a theory’s core, that connects it to observation. When a prediction fails, scientists often adjust auxiliary hypotheses rather than abandon the core.
Bath school A group of British sociologists who studied scientific controversies closely and described the experimenter’s regress, the circular difficulty of certifying an experiment as competent during a live dispute.
Big Science Modern science conducted by vast collaborations using enormous, costly instruments, such as particle colliders and gravitational-wave detectors, raising the question of whether Kuhn’s model still fits.
Black-body radiation The light given off by a heated object. Classical physics made an absurd prediction about it, an anomaly whose resolution by Planck seeded quantum theory.
Conceptual scheme The structured set of categories and assumptions through which a community organizes its experience of the world; closely related to what Kuhn meant by a paradigm or lexicon.
Confirmation The support that evidence lends to a theory. The logical positivists prized it; Popper argued that theories can never be confirmed, only tested and provisionally survived.
Conjecture A bold, testable guess. In Popper’s philosophy, science advances through conjectures exposed to attempted refutation.
Consistency A shared value in theory choice: a good theory should be free of internal contradiction and should fit reasonably with other accepted theories.
Context of discovery The messy human story of how a scientist arrives at an idea, full of hunches and accidents. The orthodoxy held it irrelevant to whether the idea is any good.
Context of justification The public testing by which an idea’s worth is judged, supposedly governed by logic alone. Kuhn’s heresy was to argue that the human factors banished to discovery also operate here.
Convergence The idea that successive theories approach ever closer to a single final truth. Kuhn doubted that the history of science shows such convergence.
Copernican Revolution The shift from an earth-centered to a sun-centered cosmos, Kuhn’s favorite worked example of a paradigm change and the subject of his first book.
Correspondence principle The idea that a good new theory should preserve the well-confirmed successes of the one it replaces, recovering them as a special case. It is the counterweight to Kuhn-loss.
Cosmological constant A quantity describing the energy of empty space, whose tiny observed value disagrees enormously with theoretical prediction, producing one of physics’ deepest anomalies.
Crisis The unsettled condition of a science that has lost confidence in its paradigm’s ability to resolve a persistent anomaly. In crisis the rules loosen and radical alternatives become thinkable.
Degenerating research programme In Lakatos’s scheme, a programme whose adjustments merely patch anomalies without predicting anything new; the mark of failing science.
Demarcation problem The question of what distinguishes genuine science from pseudo-science. Popper proposed falsifiability as the criterion; Kuhn and others complicated the picture.
Disciplinary matrix Kuhn’s later term for the broad sense of paradigm: the whole constellation of shared symbols, models, values, and exemplars held in common by a scientific community.
Discovery The finding of something new. Kuhn argued that discovery is usually a gradual process, not an instantaneous event, and often cannot be cleanly attributed to one person or moment.
Dogmatism The uncritical commitment to a framework. For Popper a vice; for Kuhn the normal and necessary condition of productive normal science.
Empiricism The view that knowledge is grounded in observation and experience. The logical positivists were strict empiricists.
Epicycle A small circle added to a planet’s path in the earth-centered system to fit observation. Their accumulation made the Ptolemaic system increasingly cumbersome.
Epistemological anarchism Feyerabend’s position that no fixed method governs good science, captured in his slogan that, in science, anything goes.
Epistemology The branch of philosophy concerned with knowledge: what it is, how we get it, and how far it can reach.
Essential tension Kuhn’s phrase for the opposed virtues science requires: the conservatism that sustains normal science and the openness that permits revolution.
Ether The invisible medium nineteenth-century physics supposed light to travel through. The failure to detect it became an anomaly resolved by relativity.
Evolutionary epistemology The view, which the late Kuhn embraced, that scientific change resembles biological evolution: it adapts away from past forms rather than toward a fixed goal.
Exemplar Kuhn’s term for the narrow, original sense of paradigm: a concrete solved problem that serves as a model for further work and is learned by imitation rather than by rule.
Experimenter’s regress The circular difficulty that, in a live controversy, a competent experiment is defined as one that gets the right result, while the right result is what a competent experiment yields.
Externalism In the historiography of science, the view that scientific change is shaped by factors outside the pure logic of evidence, including social and cultural ones.
Fallibilism The recognition that any of our beliefs, however well supported, might turn out to be wrong, and that no scientific result is ever final.
Falsifiability Popper’s criterion: a theory is scientific only if it forbids certain observations and so could in principle be proven false. A theory compatible with everything is unscientific.
Family resemblance The idea, borrowed from Wittgenstein, that members of a category may share no single defining feature but only overlapping similarities, like the members of a family.
Framework A general term for the structure of assumptions, concepts, and methods within which a community works. Roughly synonymous, in this book, with paradigm in its broad sense.
Fruitfulness A shared value in theory choice: a good theory should disclose new phenomena and new relationships that no one had anticipated.
Geocentrism The earth-centered model of the cosmos, dominant for some fourteen centuries before the Copernican Revolution.
Gestalt psychology The study of how the mind organizes perception into wholes, from which Kuhn drew the image of the sudden perceptual switch.
Gestalt switch A sudden reorganization of perception, like seeing a drawing flip from a duck to a rabbit. Kuhn used it as the model for the conversion from one paradigm to another.
Hard core In Lakatos’s scheme, the set of fundamental assumptions a research programme treats as beyond question and refuses to abandon.
Heliocentrism The sun-centered model of the cosmos proposed by Copernicus and eventually triumphant.
Heuristic A rule of thumb or guiding strategy for inquiry, as opposed to a strict method guaranteed to yield results.
Historicism The approach that understands ideas and practices in terms of their historical development and context rather than by timeless logical standards.
Historiography of science The study of how the history of science is and should be written; Kuhn helped shift it away from grading the past by present standards.
Hypothesis A proposed explanation put forward for testing.
Incommensurability Kuhn’s term for the lack of a common measure between rival paradigms: there is no neutral standard or shared language to compare them fully, though comparison and understanding remain possible with effort.
Indeterminacy of translation The thesis that the evidence never quite pins down a single correct translation of a language, a cousin of Kuhn’s incommensurability.
Induction Reasoning from particular observations to general laws. Popper argued it can never establish a universal truth, since the next observation might overturn it.
Indwelling Polanyi’s term for the way a skilled practitioner inhabits her framework so thoroughly that she sees through it rather than at it, as one sees through a familiar tool.
Internalism The view that scientific change is driven by the internal logic of evidence and argument, opposed to externalism.
Kuhn-loss The phenomenon, named in Kuhn’s honor, by which a revolution can lose explanatory content: questions the old paradigm answered may be abandoned, not merely improved upon.
Kuhnian Pertaining to Kuhn’s ideas. Kuhn himself once protested I am not a Kuhnian, disowning the relativism attributed to him.
Lexicon The structured vocabulary of kind-terms through which a science classifies the world. The later Kuhn redefined a revolution as a change in the lexicon.
Logical positivism The early-twentieth-century movement, centered on the Vienna Circle, that sought to ground all meaningful knowledge in observation and to purge metaphysics from thought.
Mature science A field unified by a single shared paradigm, as opposed to a pre-paradigmatic field marked by competing schools.
Mechanism An account of how something works through direct physical causes. Newton’s gravity famously lacked a mechanism, an example of Kuhn-loss.
Metaphysics Inquiry into the most general nature of reality. The logical positivists dismissed it as meaningless; Kuhn quietly readmitted such questions in altered form.
Methodology The study of the methods of inquiry. Feyerabend argued there is no single method that science has always followed.
Mopping-up Kuhn’s image for normal science: the long, unglamorous work of consolidating and developing the territory won by a revolutionary breakthrough.
Natural kind A category that is thought to carve nature at its joints, grouping things that genuinely belong together. Revolutions can redraw the boundaries of natural kinds.
Naturalism The approach that studies knowledge using the methods of the natural and human sciences, as the late Kuhn did when he turned to the psychology of concept learning.
Normal science The patient, puzzle-solving activity that occupies most scientists most of the time, conducted within an unquestioned paradigm; for Kuhn, the source of science’s cumulative power.
Objectivity The quality of being answerable to a real world rather than to mere personal preference. Kuhn insisted that science is objective even though it is framework-bound.
Observation The gathering of data. Kuhn and others argued that observation is theory-laden: what we see is shaped by what we already expect and know.
Ontology The branch of philosophy concerned with what exists. Different paradigms can recognize different entities as real.
Paradigm Kuhn’s central and famously slippery term, used in at least twenty-one senses, ranging from a concrete model problem (the exemplar) to a whole community’s shared framework (the disciplinary matrix).
Paradigm shift The popular phrase for a scientific revolution, the replacement of one paradigm by another. The phrase escaped into general use and lost most of its precision.
Parallax The apparent shift of nearby objects against distant ones as the observer moves. The undetected stellar parallax was long a serious objection to a moving earth.
Perception The act of seeing. On Kuhn’s view, perception is trained and theory-laden, so a change of paradigm can change what a scientist literally perceives.
Philosophy of science The branch of philosophy that studies the methods, aims, and foundations of science; Kuhn transformed it by insisting on the relevance of actual history.
Phlogiston A supposed substance, released in burning, central to eighteenth-century chemistry before the oxygen theory replaced it.
Pre-paradigmatic science An immature field that lacks a single dominant framework and is marked instead by competing schools, a condition Kuhn ascribed to much of the social sciences.
Progress The advance of science. Kuhn argued it is real but is better understood as adaptation away from past frameworks than as approach toward a final truth.
Progressive research programme In Lakatos’s scheme, a programme whose adjustments predict new phenomena and open new territory; the mark of healthy science.
Protective belt In Lakatos’s scheme, the adjustable auxiliary assumptions surrounding a programme’s hard core, modified to absorb anomalies.
Pseudoscience A body of claims that imitates science without meeting its standards. The criterion distinguishing it from real science is contested.
Ptolemaic system The mature earth-centered astronomy that reigned for some fourteen centuries, accurate and sophisticated, before Copernicus.
Puzzle-solving Kuhn’s term for the activity of normal science: tackling problems the paradigm guarantees have solutions, with the scientist’s skill, not the paradigm, on trial.
Quantization The idea that some physical quantity comes in discrete units rather than a continuous flow, the seed of quantum theory.
Radical translation The attempt to translate a wholly unknown language from scratch, used as an analogy for the historian’s effort to understand a dead paradigm.
Rationalism In this context, the conviction that science is governed by reason and clear method. Popper and Lakatos defended versions of it against Kuhn’s historicism.
Realism The view that science describes a real, mind-independent world and that its theories can be true or false of that world. Kuhn was famously ambivalent about it.
Received view The orthodox account of science associated with logical positivism, emphasizing observation, accumulation, and a sharp line between theory and fact.
Reductionism The attempt to explain complex phenomena in terms of simpler, more fundamental ones; not central to Kuhn but common in debates about scientific explanation.
Relativism The view that there is no objective truth, only what is true for a given community. Kuhn was widely accused of it and spent his life denying it.
Reproducibility crisis The recent discovery that many published scientific results cannot be reproduced, displaying a Kuhnian loss of confidence in established methods.
Research programme Lakatos’s unit of analysis: a sequence of theories sharing a hard core and a protective belt, judged progressive or degenerating over time.
Revolution A scientific upheaval in which one paradigm is replaced by another, involving a transformation of perception, standards, and the very questions a science asks.
Science Wars The bitter conflict of the 1990s between defenders of science’s objectivity and proponents of its social construction, culminating in the Sokal hoax.
Scope A shared value in theory choice: a good theory’s consequences should reach beyond the particular cases it was designed to explain.
Simplicity A shared value in theory choice: a good theory should bring order to phenomena that would otherwise be tangled, akin to Occam’s razor.
Skepticism The questioning of whether and how much we can know. Kuhn’s critics accused him of fostering it about science; he resisted the charge.
Social construction The thesis that knowledge is produced by social processes. In its modest form it studies how science is done; in its strong form it denies that reality constrains belief.
Sociology of scientific knowledge The field that studies the social production of scientific knowledge, including its actual content, taking Kuhn as a chief inspiration.
Sokal hoax The 1996 episode in which a physicist published a deliberately nonsensical article in a cultural-studies journal, then exposed it, detonating the Science Wars.
Specialization The proliferation of distinct scientific fields and sub-fields, which the late Kuhn saw as the dominant pattern of scientific development.
Speciation The biological branching of one species into two. The late Kuhn used it as the model for how scientific specialties split and multiply.
Strong Programme The Edinburgh school’s approach to the sociology of scientific knowledge, founded on the principle of symmetry.
Subjectivity The contribution of the individual knower. Kuhn was accused of making science subjective; he argued that personal judgment is compatible with objectivity.
Symmetry principle The Strong Programme’s tenet that the same kinds of social cause should explain true and false scientific beliefs alike.
Tacit knowledge Polanyi’s term for what we know but cannot fully put into words, such as how to ride a bicycle; central to how paradigms are learned.
Taxonomy A system of classification into kinds. The later Kuhn located the heart of a paradigm, and of incommensurability, in taxonomic structure.
Teleology Explanation in terms of purposes or goals. Darwin’s removal of teleology from biology is an example of a productive Kuhn-loss.
Theory A systematic explanation of some range of phenomena. For Kuhn, theories are embedded in larger paradigms that give them meaning.
Theory choice The decision between competing theories. Kuhn argued it is governed by shared values but not by any algorithm, so reasonable scientists can differ.
Theory-ladenness The thesis that observation is never pure but is always shaped by the theories and expectations the observer brings to it.
Thought collective Fleck’s term for a community bound by a shared way of thinking, anticipating Kuhn’s scientific community.
Thought style Fleck’s term for the shared assumptions and habits of perception of a thought collective, anticipating Kuhn’s paradigm.
Truth Correspondence to reality, on the traditional view. Kuhn flinched from saying that science approaches a final truth, doubting that the comparison can be made.
Underdetermination The condition in which the available evidence does not by itself settle which of several theories to accept, leaving room for judgment and other factors.
Vacuum energy The energy that quantum theory attributes to empty space, whose predicted enormity clashes with its tiny observed value in the cosmological constant problem.
Values The shared standards, such as accuracy, scope, simplicity, consistency, and fruitfulness, that guide theory choice without mechanically dictating it.
Verification principle The logical positivists’ rule that a statement’s meaning is its method of verification; statements with no possible verification were dismissed as meaningless.
Verisimilitude Popper’s notion of truthlikeness, his attempt to say that science progresses by approaching the truth even though certainty is never reached.
Vienna Circle The group of philosophers and scientists in 1920s and 1930s Vienna who founded logical positivism.
Vortex theory Descartes’s mechanical account of the heavens, in which swirling matter carries the planets, displaced by Newton’s gravitation.
Whig history The vice, named by Herbert Butterfield, of reading the past as a march toward the present, judging earlier thinkers by how close they came to current views.
Worldview A comprehensive way of seeing reality. A paradigm carries a worldview with it, which is part of why paradigm change is so wrenching.
Timeline
c. 350 BCE Aristotle sets out a physics of motion in terms of natural places and purposes, a coherent framework that will dominate Western thought for nearly two thousand years and that the young Kuhn will one day struggle, and finally manage, to understand on its own terms.
c. 150 CE Ptolemy gives the earth-centered cosmos its mature mathematical form, accurate enough to reign as normal science for some fourteen centuries.
1543 Copernicus publishes his sun-centered system as he lies dying; it is not at first more accurate than Ptolemy’s and will take a century to prevail.
1609 onward Kepler replaces circular orbits with ellipses and Galileo turns his telescope on the heavens, building out the promise of the Copernican framework.
1687 Newton publishes his law of universal gravitation, a triumph of prediction that nonetheless abandons the mechanical explanation of gravity, an early example of what will later be called Kuhn-loss.
Late 1700s Lavoisier’s oxygen theory overthrows phlogiston, turning the anomalous weight gain on combustion from an embarrassment into a foundation stone.
1830s Stellar parallax is finally detected, resolving, three centuries late, an objection that had once seemed fatal to the moving earth.
1859 Darwin publishes his theory of natural selection, explaining the history of life without any goal, the analogy that will anchor Kuhn’s mature account of progress.
1880s The Michelson-Morley experiment fails to detect the ether, producing an anomaly that classical physics cannot absorb.
1895 Röntgen discovers X-rays by taking seriously an anomalous glow, an exemplary case of discovery beginning with the violation of expectation.
1900 Planck introduces energy quanta to solve the black-body problem, an act later mythologized as the sudden birth of quantum theory, though Kuhn will argue the revolution was gradual.
1905 to 1916 Einstein’s relativity recasts space, time, and gravity, dissolving the ether and the anomaly of Mercury’s orbit, and in a sense restoring a mechanism for gravity.
1919 An eclipse expedition confirms Einstein’s bending of starlight; the young Popper sees in this bold, risky prediction the template for his philosophy of falsification.
1920s to 1930s The Vienna Circle develops logical positivism, seeking to ground all meaningful knowledge in observation and to banish metaphysics.
1935 Ludwik Fleck publishes his neglected account of the social construction of a scientific fact, anticipating much of Kuhn’s theory and coining the thought collective and thought style.
1934 onward Karl Popper advances falsificationism, proposing that science is marked not by provable truth but by exposure to refutation.
Late 1930s onward Alexandre Koyré transforms the history of science, treating the scientific revolution as a profound change of metaphysical vision rather than a mere accumulation of facts.
1931 Herbert Butterfield names the Whig interpretation of history, the vice of judging the past by the standards of the present.
1947 Kuhn, preparing a case study on mechanics, has his transforming encounter with Aristotle, the seed of his entire later thought.
1957 Kuhn publishes The Copernican Revolution, a dress rehearsal in which the themes of his later masterpiece are tested on a single great example.
1958 Michael Polanyi publishes Personal Knowledge, developing the idea of tacit knowing and the personal, committed character of all knowledge.
1962 Kuhn publishes The Structure of Scientific Revolutions, introducing paradigms, normal science, anomaly, crisis, revolution, and incommensurability, and popularizing the phrase paradigm shift.
1965 At a London colloquium, Kuhn and Popper clash directly over the nature of normal science, with Lakatos and Feyerabend looking on and preparing their own responses.
1969 Kuhn adds a long postscript to the second edition, clarifying the disciplinary matrix and the exemplar and beginning his lifelong effort to rebut the charge of relativism.
1970 Lakatos publishes his methodology of scientific research programmes, distinguishing progressive from degenerating programmes in an attempt to rescue rationality after Kuhn.
1975 Feyerabend publishes Against Method, pushing Kuhn’s premises to the anarchist conclusion that, in science, anything goes.
1970s The Edinburgh Strong Programme launches the sociology of scientific knowledge on the principle of symmetry, taking Kuhn as inspiration and pushing toward social construction.
1977 Kuhn collects his essays as The Essential Tension, naming the balance between tradition and innovation that science requires.
1978 Kuhn publishes Black-Body Theory and the Quantum Discontinuity, arguing as a working historian that Planck did not at first intend the quantum revolution credited to him.
1980s Marshall and Warren overturn the reigning theory of stomach ulcers, a fully documented modern paradigm shift complete with fierce resistance.
1980s onward The late Kuhn drops the word paradigm, turns to the lexicon and taxonomy, and recasts revolutions as the speciation of scientific specialties.
1990s The Science Wars erupt between defenders of scientific objectivity and proponents of social construction.
1996 Alan Sokal’s hoax detonates in a cultural-studies journal; in the same year, Thomas Kuhn dies, the controversy over his legacy raging around him.
2000 Kuhn’s later philosophical essays appear posthumously as The Road Since Structure, documenting his mature, taxonomic turn.
2016 The detection of gravitational waves, announced in a paper with over a thousand authors, exemplifies the Big Science whose scale tests the limits of Kuhn’s model.
2022 Kuhn’s unfinished late work on incommensurability is edited and published, offering a fuller view of the evolutionary, language-centered vision he was reaching for at his death.
The present Deep anomalies such as the cosmological constant problem, a reproducibility crisis in several fields, and the rise of artificial intelligence in research keep Kuhn’s questions about how science changes alive in the present tense.
Further Reading
This book is a work of popular synthesis, and the reader who wishes to go deeper should turn to the sources below. The list begins with Kuhn’s own writings, which remain the indispensable starting point and are more readable than their reputation suggests, then moves to the thinkers who shaped him and those who argued with him, and finally to the best of the secondary literature and to a few works that press the case against him, for no honest reading of Kuhn is complete without the strongest objections of his critics. Dates given are those of first publication in English where relevant; many of these works have appeared in several editions.
The Works of Thomas Kuhn
Kuhn, Thomas S. The Copernican Revolution: Planetary Astronomy in the Development of Western Thought. Cambridge, Massachusetts: Harvard University Press, 1957. The dress rehearsal, in which the great themes are tested on a single example.
Kuhn, Thomas S. The Structure of Scientific Revolutions. Chicago: University of Chicago Press, 1962. The masterpiece. Later editions add the important 1969 postscript; any edition will serve, and it should be read in full before any commentary.
Kuhn, Thomas S. The Essential Tension: Selected Studies in Scientific Tradition and Change. Chicago: University of Chicago Press, 1977. Essays that clarify and extend the main book, including the title piece on the balance between tradition and innovation.
Kuhn, Thomas S. Black-Body Theory and the Quantum Discontinuity, 1894 to 1912. Chicago: University of Chicago Press, 1978. The demanding historical study of the birth of quantum theory; for serious readers willing to follow technical history.
Kuhn, Thomas S. The Road Since Structure: Philosophical Essays, 1970 to 1993, with an Autobiographical Interview. Edited by James Conant and John Haugeland. Chicago: University of Chicago Press, 2000. The essential collection of the late, taxonomic Kuhn, with a revealing interview.
Kuhn, Thomas S. The Last Writings of Thomas S. Kuhn: Incommensurability in Science. Edited by Bojana Mladenovic. Chicago: University of Chicago Press, 2022. The unfinished late project, offering the fullest view of the vision he was reaching for at his death.
The Thinkers Who Shaped Kuhn
Fleck, Ludwik. Genesis and Development of a Scientific Fact. Translated by Fred Bradley and Thaddeus Trenn. Chicago: University of Chicago Press, 1979. Originally published in German in 1935. The neglected forerunner who anticipated much of Kuhn.
Polanyi, Michael. Personal Knowledge: Towards a Post-Critical Philosophy. Chicago: University of Chicago Press, 1958. The great account of tacit knowing and the personal, committed character of knowledge.
Koyre, Alexandre. From the Closed World to the Infinite Universe. Baltimore: Johns Hopkins University Press, 1957. The historian who treated scientific revolution as a transformation of metaphysical vision.
Butterfield, Herbert. The Whig Interpretation of History. London: G. Bell and Sons, 1931. The short, sharp essay that named the vice of reading the past as a march toward the present.
Butterfield, Herbert. The Origins of Modern Science, 1300 to 1800. London: Bell, 1949. An influential narrative history of the scientific revolution.
The Great Debates: Critics and Rivals
Popper, Karl. The Logic of Scientific Discovery. London: Hutchinson, 1959. Originally published in German in 1934. The founding statement of falsificationism, Kuhn’s chief antagonist.
Popper, Karl. Conjectures and Refutations: The Growth of Scientific Knowledge. London: Routledge and Kegan Paul, 1963. The accessible essays in which Popper’s philosophy is most vividly stated.
Lakatos, Imre, and Alan Musgrave, editors. Criticism and the Growth of Knowledge. Cambridge: Cambridge University Press, 1970. The famous volume containing the direct confrontation among Kuhn, Popper, Lakatos, and Feyerabend; the single best document of the whole controversy.
Lakatos, Imre. The Methodology of Scientific Research Programmes: Philosophical Papers, Volume 1. Cambridge: Cambridge University Press, 1978. The mature statement of the progressive-versus-degenerating distinction.
Feyerabend, Paul. Against Method: Outline of an Anarchistic Theory of Knowledge. London: New Left Books, 1975. The exuberant assault on the idea of a fixed scientific method.
Understanding Kuhn: Secondary Works
Bird, Alexander. Thomas Kuhn. Princeton: Princeton University Press, 2000. A clear and philosophically careful introduction, strong on the later Kuhn.
Hoyningen-Huene, Paul. Reconstructing Scientific Revolutions: Thomas S. Kuhn’s Philosophy of Science. Chicago: University of Chicago Press, 1993. The most thorough systematic reconstruction, endorsed by Kuhn himself.
Sharrock, Wes, and Rupert Read. Kuhn: Philosopher of Scientific Revolution. Cambridge: Polity Press, 2002. A sympathetic account that defends Kuhn against common misreadings.
Wray, K. Brad. Kuhn’s Evolutionary Social Epistemology. Cambridge: Cambridge University Press, 2011. A study emphasizing the late, evolutionary, community-centered Kuhn.
Marcum, James A. Thomas Kuhn’s Revolution: An Historical Philosophy of Science. London: Continuum, 2005. A helpful overview of the development of Kuhn’s thought.
The Sociologists and the Science Wars
Bloor, David. Knowledge and Social Imagery. London: Routledge and Kegan Paul, 1976. The founding statement of the Strong Programme and its principle of symmetry.
Latour, Bruno, and Steve Woolgar. Laboratory Life: The Construction of Scientific Facts. Princeton: Princeton University Press, 1979. The influential close study of fact-making in a working laboratory.
Collins, Harry. Changing Order: Replication and Induction in Scientific Practice. London: Sage, 1985. The source of the experimenter’s regress.
Gross, Paul R., and Norman Levitt. Higher Superstition: The Academic Left and Its Quarrels with Science. Baltimore: Johns Hopkins University Press, 1994. The combative book that helped ignite the Science Wars.
Sokal, Alan, and Jean Bricmont. Fashionable Nonsense: Postmodern Intellectuals’ Abuse of Science. New York: Picador, 1998. Published in Britain as Intellectual Impostures. The documentation that followed the Sokal hoax.
The Case Against Kuhn
Weinberg, Steven. The Revolution That Didn’t Happen. In The New York Review of Books, October 1998. A distinguished physicist’s sharp and readable dissent from Kuhn’s picture of science; essential for balance.
Fuller, Steve. Thomas Kuhn: A Philosophical History for Our Times. Chicago: University of Chicago Press, 2000. A provocative and unsympathetic reappraisal that questions Kuhn’s influence.
Morris, Errol. The Ashtray (Or the Man Who Denied Reality). Chicago: University of Chicago Press, 2018. A hostile and personal memoir-cum-polemic by a former student; one-sided but a vivid counterweight, to be read with care.
The Wider Philosophy of Science
Godfrey-Smith, Peter. Theory and Reality: An Introduction to the Philosophy of Science. Chicago: University of Chicago Press, 2003. The best single survey of the field, placing Kuhn in his full context.
Okasha, Samir. Philosophy of Science: A Very Short Introduction. Oxford: Oxford University Press, 2002. A brief, lucid entry point for the general reader.
Chalmers, A. F. What Is This Thing Called Science? Indianapolis: Hackett, 1976 and later editions. A widely used and accessible introduction to the central debates.
Hacking, Ian. Representing and Intervening: Introductory Topics in the Philosophy of Natural Science. Cambridge: Cambridge University Press, 1983. A landmark that shifts attention from theory to experiment and intervention.
Losee, John. A Historical Introduction to the Philosophy of Science. Oxford: Oxford University Press, several editions. A clear chronological survey of the tradition Kuhn entered and transformed.
메타데이터
- post_id
- e422ce9dbf4b
- slug
- how-knowledge-breaks-and-remakes-itself-thomas-kuhn-and-the-secret-life-of-scientific-change-e422ce9dbf4b
- url
- https://medium.com/@krigerbruce/how-knowledge-breaks-and-remakes-itself-thomas-kuhn-and-the-secret-life-of-scientific-change-e422ce9dbf4b
- canonical_url
- https://medium.com/@krigerbruce/how-knowledge-breaks-and-remakes-itself-thomas-kuhn-and-the-secret-life-of-scientific-change-e422ce9dbf4b
- author_url
- https://medium.com/@krigerbruce
- status
- ok
- fetched_at
- 2026-08-08 06:34:36
