Science is often presented as a steady pile of facts growing taller each year. Its real history is more dramatic. Established ideas have been overturned, outsiders have been ignored and later vindicated, and whole fields have changed their basic picture of the world. This level looks at how that happens, and why the ability to change its mind is science's greatest strength.

Kuhn's revolutions

In 1962 the American historian and philosopher Thomas Kuhn published The Structure of Scientific Revolutions, one of the most cited academic books of the century. Kuhn argued that most science is “normal science”: researchers work within a shared framework, which he called a paradigm, solving puzzles it defines. Results that don't fit, called anomalies, are usually set aside or explained away. But when anomalies pile up, a field can enter a crisis, and a new paradigm may replace the old one. Examples include the shift from an Earth-centered to a Sun-centered universe, and from Newtonian physics to relativity and quantum mechanics.

Kuhn's critics worried that he made science sound like a series of fashions. Kuhn himself rejected that reading. New paradigms win, in the long run, because they solve more problems and make better predictions. Old theories also often survive as useful approximations: engineers still use Newton's laws to launch spacecraft.

The drifting continents

In 1912 the German scientist Alfred Wegener proposed that the continents had once been joined and had slowly drifted apart. He pointed to the jigsaw fit of South America and Africa, matching fossils and rock formations on opposite sides of oceans, and evidence of ancient climates in the wrong places. Most geologists rejected the idea for decades, partly because Wegener could not explain what force could move continents. That was a reasonable objection. In the 1950s and 1960s, new evidence from the ocean floor, including stripes of magnetized rock recording reversals of the Earth's magnetic field, showed that the seafloor was spreading. By the late 1960s plate tectonics had become the foundation of modern geology. Wegener, who died on a Greenland expedition in 1930, did not live to see it.

Washing hands

In 1847, at Vienna General Hospital, the Hungarian doctor Ignaz Semmelweis noticed that mothers in the clinic run by doctors died of childbed fever far more often than mothers in the clinic run by midwives. Doctors often came straight from performing autopsies. Semmelweis ordered them to wash their hands in a chlorinated lime solution, and deaths fell sharply, from around 10 percent in some periods to about 1 to 2 percent. Yet many colleagues rejected his conclusion. He had no theory of why it worked, he published slowly and he attacked critics bitterly. He died in an asylum in 1865, just before germ theory, developed by Louis Pasteur, Joseph Lister and others, supplied the explanation.

The stomach bug

For most of the twentieth century, doctors believed stomach ulcers were caused by stress and acid. In the early 1980s the Australian researchers Barry Marshall and Robin Warren argued that a bacterium, Helicobacter pylori, was responsible. Skeptics said bacteria could not survive in stomach acid. In 1984 Marshall drank a culture of the bacteria, developed stomach inflammation, and treated it with antibiotics. Larger trials followed, and ulcers became curable with antibiotics. In 2005 Marshall and Warren received the Nobel Prize in Physiology or Medicine.

The lesson of the rebels

These stories are inspiring, but they can be misread. For every Wegener there were many people with bold ideas who were simply wrong. What vindicated the rebels was evidence, not boldness or persecution. Being rejected by experts is not a sign you are right. Being rejected by experts and then producing the evidence that changes their minds is how science moves.

Worked example: two planets, two anomalies

Nineteenth-century astronomers found two planets that did not quite follow Newton's laws. Both cases tested how science handles anomalies.

• Uranus: its orbit drifted from Newton's predictions. Rather than abandon Newton, the French mathematician Urbain Le Verrier proposed an auxiliary hypothesis: an unseen planet was pulling on it. He calculated where it should be. In 1846 the astronomer Johann Galle in Berlin pointed his telescope there and found Neptune. Newton's theory triumphed.
• Mercury: its orbit also drifted, by a tiny amount, 43 arcseconds per century beyond what known influences explained. Le Verrier tried the same move and proposed a small inner planet, named Vulcan. Astronomers searched for decades. Vulcan was never found, because it does not exist.
• The resolution: in 1915 Einstein's general relativity predicted Mercury's extra drift almost exactly, with no new planet needed.

The same reasonable strategy, adding a testable auxiliary hypothesis, succeeded once and failed once. What decided each case was whether the new prediction was confirmed. Science does not know in advance which anomalies are puzzles to solve and which are signs that a theory must go. It finds out by testing.