If there is a beginning time point for the Age of Scientific Reversal, it may be 1887—the year when Albert A. Michelson and Edward W. Morley conducted what is often called the world’s most famous failed physics experiment.
For more than two centuries researchers had proposed that light was a wave of some kind traveling through an ineffable material that pervaded everything, even the space between atoms. No evidence of this all-permeating substance—the aether, as it was called—had ever been detected. Still, most scientists firmly believed it must exist. How could a wave be seen to travel unless there were something it was traveling through? Working in Cleveland, Ohio, Michelson and Morley sought to measure the aether’s effects with some of the most sensitive equipment ever built. To their shock, they found absolutely no trace of it.
Baffled and discouraged, the two men gave up plans for follow-up experiments. Other physicists were even more dismayed. The great theoretical physicist Hendrik Lorentz said the results put him “utterly at a loss.”
On supporting science journalism
If you’re enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
Yet they were not a loss for science. The Michelson-Morley tests actually led to a remarkable intellectual 180-degree turn and a forward leap in physics. The aether, scientists had believed, would provide a fixed background—a universal reference for all celestial objects. The discovery that outer space was a featureless, nearly empty vacuum—which stemmed from Michelson and Morley’s work—meant that objects could be located only in reference to one another. And that realization fed into an even bigger 180-degree turn: Albert Einstein’s theories of special and general relativity, which upended previous notions of gravity and turned space and time into a single curvature created by mass and energy.
Or … or … maybe the Age of Scientific Reversal began after 1860, when chemist and microbiologist Louis Pasteur presented a long, bluntly written memoir that proved fermentation was caused by microorganisms, not some self-starting chemical reaction, which was the reigning theory. Pasteur’s work led to a pitched intellectual battle—and the eventual triumph of germ theory, which overturned earlier ideas about infectious disease.
Coming one after another, such volte-faces gave rise to a popular notion of scientific progress as a series of upheavals in which mavericks throw out the entrenched views of the past. In countless stories in movies, television and novels, revolutionary thinkers (or, rather, wannabe revolutionaries) have their ideas dismissed by hidebound colleagues, yet they triumph in the end.
But that’s not how science works. Or, more precisely, it’s not how science works except in two specific, relatively unusual circumstances.
The first is when research disciplines are young, thinly populated and just developing instruments of sufficient power to test their initial beliefs, as was the case with the Michelson-Morley experiment and Pasteur’s fermentation. The second, possibly more consequential situation is when scientific findings lead to so much public interest that they become of concern to political authorities. Contemporary examples, such as the fraught debate over whether women under the age of 50 should be routinely screened with mammograms, have filled recent headlines. But these political issues have influenced science in the U.S. since at least the 19th century, when the country began trying to move immigrants across the Mississippi River into what might or might not have been hostile, uninhabitable land.
The image of scientific rebels forcing other researchers to reverse themselves was codified in philosopher Thomas Kuhn’s 1962 book, The Structure of Scientific Revolutions. In Kuhn’s view, there are periods of “normal science” in which researchers have a shared consensus—a paradigm, in his terms—about how nature works. Then a new theory or experiment shatters the paradigm. Believers in the old paradigm resist furiously, but eventually the old ideas are ejected. From the reversal emerges a new paradigm, which will be thrown over in turn.
Structure was a bombshell. It is one of the few academic tracts to leap outside the classroom and influence the larger culture. Since its publication, stories about “revolutionary” new scientific studies that “overthrow everything we believed” have become staples in journalism, Hollywood and YouTube health-influencer videos.
The lighter side of this trope is embodied in characters such as Doc Brown, the DeLorean-driving inventor in the 1985 movie Back to the Future, whose unconventional ideas about time travel cause his colleagues to dismiss him as a crackpot. The darker side leads to figures such as discredited anti-vaccine researcher Andrew Wakefield and germ-theory denialist and writer Mike Stone, whose followers claim that their findings have been suppressed by the scientific establishment in the name of profit and political ideology.
The reality is closer to what happened with Michelson and Morley. Physics as a field of knowledge has existed at least since the times of Greek savant Thales (circa 625–545 B.C.E.). But the professional discipline—practiced by credentialed professors who work in specialized laboratories and belong to learned societies—was in its infancy when the two scientists looked for the aether. The U.K.’s first specialized physics group, the Physical Society of London, had been founded just 13 years earlier.
Physicists in those early years were reexamining ideas that often dated back to the Greeks (Aristotle, in the case of the aether) and had yet to be probed with modern tools. Michelson and Morley bounced light among 16 specially prepared mirrors with positions that had to be adjusted so precisely that the two men had to machine custom-calibration screws with 100 threads per inch—implements that couldn’t have been made in Thales’s time or even Isaac Newton’s. Given that many of the foundational assumptions in physics had never been carefully tested, it seems almost inevitable in retrospect that a considerable number would fall to the first scrutiny.
Consider the long-standing belief that the universe preserves parity—that the mirror reflection of any physical process is identical to its unmirrored counterpart except for being flipped from left to right. This is obviously true in the world we live in: shooting one billiard ball at another will have the same effect no matter what direction the cue ball comes from. But matters are less obvious in the quantum realm.
In 1956 physicists Chen Ning Yang of Princeton University and Tsung-Dao Lee of Columbia University wondered whether anyone had proved that parity was preserved in quantum interactions—and found that nobody had checked the “weak nuclear force,” which is responsible for radioactive decay. The first research team that looked at the weak interactions, led by Columbia’s Chien-Shiung Wu, found that the weak force did not conserve parity. Stunned, Yang sent a telegram to physicist J. Robert Oppenheimer about Wu’s experiment. “Walked through door,” the gobsmacked Oppenheimer cabled back.
Lee and Yang won the 1957 Nobel Prize in Physics for beginning the parity U-turn. But that was arguably the last time particle physics went through such a sweeping reversal. Yes, the field has seen extraordinary discoveries since then—quarks and gluons, neutrino oscillations, gravity waves, you name it. But they were new phenomena, not refutations of prior beliefs.
The lack of 180s partly results from the way that scientific disciplines ground themselves over time. In retrospect, one can’t be surprised that the first experiment to carefully examine the aether failed to find it. But it would be extremely surprising if, after decades of experimental verification, quarks were shown not to exist. In addition, as disciplines grow older and bigger, they end up naturally absorbing people with minority points of view. So instead of entire disciplines executing a U-turn, these minority beliefs shift and twist while becoming acceptable to the majority.
In particle physics, an idea known as S-matrix theory dominated in the 1950s and 1960s, but it always had skeptics. When experiments pointed toward an alternative—a quantum field theory and quark model—the field shifted. But it wasn’t exactly a U-turn, because quantum field theorists had been working on their ideas all along. And S-matrix theory never vanished. It morphed into string theory, a current attempt to unify relativity and quantum mechanics.
Similarly, one of the first vogues in the field of artificial intelligence was the perceptron, a computational system that 1960s-era AI researchers argued would rival human intelligence and ultimately lead to machines with true consciousness. Researchers published thousands of papers extolling and developing perceptrons—an outburst that stopped abruptly after 1969, when computer scientists Marvin Minsky and Seymour Papert took a careful, Michelson-Morley-style look at the idea. They detailed basic tasks that perceptrons could never do, including distinguishing between odd and even numbers. With this embarrassment the perceptron bubble popped. But it didn’t disappear. As AI research slowly grew, perceptrons changed into more sophisticated neural networks, which in turn played a role in the development of today’s “large language model” artificial intelligence.
Youthful fields can take even more dramatic turns. Pasteur’s work on the role of microorganisms in infectious disease inaugurated the modern discipline of microbiology—and led to a host of about-faces in previous medical beliefs. German researcher Robert Koch, often considered microbiology’s co-founder, then discovered the microbes that caused anthrax, cholera and tuberculosis. All cast aside earlier ideas. For instance, many in Koch’s Germany believed tuberculosis was a hereditary disease passed down through families until 1882, when the scientist unveiled Mycobacterium tuberculosis, the bacterium responsible for the disease.
In the researchers’ map of the region drained by the Mississippi was a label: GREAT AMERICAN DESERT.
These reversals did not always have the revolutionaries and traditionalists one would expect on opposing sides. French physician Alphonse Laveran spotted microscopic living creatures in malaria patients’ blood in 1890. Since the Greeks, doctors had believed that malaria was brought about by “miasma”—misty air polluted with particles from decomposed matter. (The disease’s name comes from mal aria, or “bad air” in early Italian.) Based on his observations, Laveran proclaimed that malaria was caused by protozoans. These microbes are now known to be several species in the genus Plasmodium.
Laveran’s fiercest critics were not miasma theorists, however, but Pasteur’s disciples, who insisted a la Kuhn on yet another paradigm: infectious diseases were caused by bacteria—bacteria floating in mist, in this case. France’s leading malaria authority sneered at Laveran, as did Koch. In reaction, Laveran doubled down, proposing that Plasmodium was carried by mosquitoes, not mist. This, too, was dismissed. But Laveran was subsequently proved right, and within a decade scientists had to reverse themselves again.
Paralleling the rise of institutional particle physics, the microbiology of Pasteur and Koch expanded into an enormous discipline with thousands of researchers, multiple subfields—and ever fewer reversals. Today the International Union of Microbiological Societies has 57 groups from 45 nations; Italy, long a center for this kind of research, has six professional societies of its own. Last year’s annual meeting of the American Society for Virology attracted more than 2,000 attendees from 50 countries.
Then there is the political influence on scientific 180s. In the U.S., politics and science collided right after 1803, the year of the Louisiana Purchase. The U.S. government knew so little about its new possession that it dispatched no fewer than four teams to survey the territory. One, led by U.S. Army officers Meriwether Lewis and William Clark, crossed the continent by a northern route and became a celebrated part of American history. The three other expeditions went into the southern and central plains and were repelled by Spanish troops and Indigenous nations. Not until 1819 did the U.S. try again, sending a team led by engineer Stephen H. Long.
Although Long didn’t know it, the southern plains were beset by a multiyear drought. While surveying the Platte and Canadian Rivers, his team almost starved. Unsurprisingly, the expedition’s report portrayed the southern plains as “presenting the aspect of hopeless and irreclaimable sterility.” The land was “almost wholly unfit for cultivation, and of course uninhabitable by a people depending upon agriculture for their subsistence.” In the center of the team’s map of the “country drained by the Mississippi” was a capitalized label: GREAT AMERICAN DESERT.
Today we know that in the central and southern plains, long-term atmospheric fluxes from the Pacific (the El Niño–Southern Oscillation, for example) and the Atlantic (the Atlantic Multidecadal Oscillation) mix with warm, moist air currents from the Gulf of Mexico and cold, dry air from the Arctic jet stream. These phenomena collide unpredictably, causing tornadoes, blizzards, severe hailstorms, epic heat waves and, notably, lengthy droughts—the 1930s Dust Bowl being the most well known.
In what one pictures as the usual course of events, Long’s report would have been followed by other surveys, some challenging his views, some backing them. Presumably the back-and-forth would slowly have revealed that Long’s belief that aridity was the region’s permanent state was incorrect because arid and wet periods came in irregular cycles. Yet that realization was not what happened, because politicians and wealthy interests, especially new railroad tycoons, wanted people to move to the plains and create communities that both produced and bought goods and crops. These would, of course, be transported by trains.
So not only did critics dispute the existence of the Great American Desert, but they said that rainfall in the area was increasing—because of farming. In his 1880 book Sketches of the Physical Geography and Geology of Nebraska, University of Nebraska scientist Samuel Aughey explained that prior to the arrival of Europeans, the prairie had been “pelted by the elements and trodden by millions of buffalo,” which packed the soil too hard to absorb water. But with settlers’ plows breaking up the hardpan, “the rain as it falls is absorbed by the soil like a huge sponge.” More water retained in the land means more evaporation over it, which “must give increasing moisture and rainfall.” A slogan emerged: “Rain follows the plow.”
Researchers led by geologist John Wesley Powell, director of the U.S. Geological Survey, countered that the region was too drought-prone to sustain agriculture. Early rangeland scientists mocked the idea that the precolonial grasslands couldn’t retain moisture. But their assertions were buried underneath floods of flyers, leaflets and advertisements from railroads that extolled Long’s Great American Desert as a Great American Garden. When a multiyear drought overtook the region in the 1890s, it was a shock—a complete reversal of the expectations of migrants who had relied on the railroads’ descriptions. They fled the area in droves. After the rains returned, new migrants poured in. The 1930s Dust Bowl, when it came, was just as much of a shocking 180 to them as the previous drought had been to their predecessors.
The argument over the climate in the plains was an early example of an increasingly common phenomenon: the mismatch between the slow, unsteady movement of scientific understanding and the immediate, short-term imperatives of politics and economics, which can lead to what seem like vertiginous scientific reversals.
Examples are as near to hand as the COVID pandemic. Early in the epidemic, in March 2020, the World Health Organization (WHO) avowed that COVID could not be transmitted through the air—people picked up the SARS-CoV-2 virus from surfaces. (“FACT”—the agency tweeted—“#COVID19 is NOT airborne.”) Other public health outfits followed suit. Air-pollution specialists, including those within these outfits, were astounded by the claims. In their discipline, it was well known that large particles of soot could travel through the air for miles.
Lidia Morawska, an aerosol specialist at the Queensland University of Technology in Brisbane, Australia, led a group of aerosol researchers and ventilation engineers that contacted WHO about the lengthy travel distances days after the “FACT” tweet. Dismissing this evidence as weak, a WHO advisory group insisted in August 2020 that “SARS-CoV-2 is not spread by the airborne route to any significant extent.”
In part, WHO’s reluctance was a legacy of the previous battle over miasma theory. The fight to eliminate the fear of vapors led infectious disease experts to take as given that nearly all infectious pathogens were spread by “droplets,” generally defined as more than five microns in diameter. Droplets fly out of sick people’s mouths and noses when they cough, shout, sing or sneeze. The particles then land directly on other people or on nearby surfaces that people later touch. Implicit in the definition of droplets was that their relatively large size limited their ability to travel. Thus, WHO focused on getting people to wash surfaces and hands to stop the spread of the virus. Aerosol transmission, in which smaller organisms travel farther in vapor clouds, was thought to occur only for a few well-known diseases, mainly tuberculosis and measles.
WHO tenaciously stuck to its paradigm despite a tsunami of reports of aerosol transmission. Only gradually did the agency admit that such transmission was possible in specific “crowded and inadequately ventilated [indoor] spaces” (July 2020), that the virus could travel in the air “farther than 1 metre” in specific settings (April 2021), and, finally, that “airborne” transmission could occur in some places (December 2021)—a move that was greeted as a long-overdue 180.
The reversal was Kuhnian in the sense that WHO’s scientific paradigm was overturned after resistance. But the scientists who rejected Michelson-Morley were motivated mainly by adherence to scientific orthodoxy, whereas WHO researchers were also responding to an intensely political environment. Agencies such as WHO are supposed to provide guidance for others to act on. Under public pressure to be definitive, they often end up digging in their heels on research questions that are poorly understood. What would in other circumstances be ordinary back-and-forth as researchers resolved questions is transformed into a series of stark, headline-grabbing reversals.
Perhaps nothing better illustrates this type of politically driven reversal than the five-decade controversy over mammograms for women between 40 and 50 years of age. In the early 1970s the National Cancer Institute (NCI) and the American Cancer Society launched the Breast Cancer Detection Demonstration Project (BCDDP) to test the potential of large-scale mammography. Some cancer researchers protested that repeatedly exposing women under 50 to x-rays would do more harm than good, so the BCDDP restricted enrollment of younger women to those at “high risk.”
The results were released in the 1980s. Although the BCDDP design had weaknesses, the study authors said the results showed that mammograms detected breast tumors that would not otherwise have been spotted. And screening did not produce excessive false positives, which can lead to needless biopsies and surgeries.
The NCI and almost 20 other medical organizations met to establish guidelines for mammography. A reanalysis of another, earlier, smaller trial, the Health Insurance Program of Greater New York study, also showed that mammography for younger women had positive effects. The combined result was nationwide recommendations, issued in 1989, that women should begin screening for cancer at age 40—and spurred a big advertising campaign by advocacy groups to convince women to do it.
But then, in 1992, the Canadian National Breast Screening Study of Cancer—the first randomized clinical trial designed specifically to examine the effectiveness of under-50 mammography—released a contradictory result: testing younger women did not reduce death rates. Big randomized clinical trials are generally considered the best way to understand the efficacy of medical treatments. Nevertheless, this one was furiously attacked by cancer advocacy groups, clinicians and radiologists, who asserted something must be wrong with the way the trial was done. Oddly, after the NCI convened a workshop on the issue that concluded “there is no reduction in mortality from breast cancer that can be attributed to screening,” the institute also insisted there was no need to change the recommendation for earlier screening. It cited vague “inferential” benefits.
Troubled by the idea of basing nationwide recommendations on what experts judged as low-quality evidence, Samuel Broder, then director of the NCI, announced the institute would not promote screening for women in their 40s. In his view, the potential good effects (possibly catching a few relatively rare cancers early) were far outweighed by the potential bad effects (those false alarms that scare women and can lead to many painful and unnecessary surgeries).
The National Cancer Advisory Board—an NCI advisory group of federal-agency officials, representatives from cancer associations, and cancer researchers—asked Broder not to pull back right away. He and the NCI stuck to their guns. Then U.S. Congress members erupted, calling the institute callous and sexist.
The American Cancer Society, the American College of Radiology, and other medical groups conceded that there weren’t good data to support under-50 mammography. But they felt obligated to do something to address younger women’s fear of breast cancer—a fear that was inflamed, in part, by the organizations’ own public-relations campaigns promoting mammograms and breast self-exams.
Both sides continued their standoff until 1997, when the NIH convened a consensus conference to try to resolve the issue. It concluded that the current data didn’t support under-50 screening. But the hoped-for consensus collapsed when critics, such as a mammography director at a private practice in New Mexico, charged that the agency statement was “tantamount to a death sentence for thousands of women in their forties.” Congress voted 98–0 to order the NCI to back screening for younger women. The institute caved. The American Cancer Society joined it to state that screening for women in their 40s was “beneficial and supportable with current evidence.”
Little of this controversy was visible in doctors’ offices, where women were being told that screening that begins at 40 saves lives. Outside of those offices, advocacy groups were saying the same thing. So many patients were shocked by headlines in 2009, when the U.S. Preventive Services Task Force (USPSTF), an independent and influential expert board advising the federal Department of Health and Human Services (HHS), went in the opposite direction. It said that almost 2,000 younger women would have to be screened to save one life. The other 1,900-plus women would be exposed to the risks of radiation and surgery.
The White House denounced the USPSTF’s stance. The task force backed down, saying instead that women should consult their doctors—an embarrassing break with its mission, which was to assess the state of evidence for entire fields rather than telling patients to rely on the opinions of individual practitioners. Congress passed a law explicitly telling HHS to disregard “the current recommendations of the United States Preventive Services Task Force regarding breast cancer screening.”
Then, in 2024, there was an actual reversal. The USPSTF issued another set of recommendations—but this time it came out in favor of routine mammograms for women in their 40s.
Throughout this time, the data had changed little. When put together, the eight big randomized controlled trials of mammography for women under 50 have shown that the tests produce very high, specific benefits for a small number of women and impose other costs on a much larger number of women. Yet the glare of publicity transformed a slow but fairly typical research debate into a huge controversy culminating in a big 180.
This kind of politically charged reversal shows little sign of declining. Likely future reversals may include causes and treatment of obesity or of Alzheimer’s disease. All are the subjects of intense lobbying by commercial and public-interest groups.
As for reversals in fields where scientific ideas compete in disciplines that lack adequate investigatory tools, who knows? But hints may come from cosmology, where grand ideas about the nature of the universe jostle for prominence. These notions are constrained by the difficulty of gathering data but still driven forward by scientists seeking the thrill of causing yet one more scientific twist.
