Science Doesn’t Have One Magic Method—and Here’s Why
The Mystery of Childbed Fever
In 1847, a Hungarian doctor named Ignaz Semmelweis walked the wards of Vienna’s General Hospital, baffled. In one maternity clinic, one out of every ten new mothers died of childbed fever. In the other clinic, run by midwives, fewer than one in fifty did. The patients faced very different fates—but why?
Semmelweis started testing possible explanations. Maybe the fever came from dirty air, or from something in a mother’s diet, or from the way doctors examined patients. He checked each idea by deducing what should happen if it were true, and then watching real patients. When an idea contradicted what he saw, he rejected it. Eventually he tried having doctors wash their hands with a chlorine solution after performing autopsies. The death rate plummeted. Semmelweis had used careful reasoning to find the real cause—even though he didn’t know germs existed yet.
This story is often told as a perfect example of the scientific method: observe, form a hypothesis, test it, and draw a conclusion. But was there one master method at work? Or do scientists use many different methods, depending on the problem, the time, and the field? Philosophers have been arguing about that for over two thousand years.
Aristotle’s Two Paths
Long before Semmelweis, the ancient Greek thinker Aristotle (384–322 BCE) wrote the first systematic guide to investigating nature. He believed science begins with empirical observation—using your senses to gather facts about the world. But just piling up facts wasn’t enough. For Aristotle, real scientific knowledge meant arranging those facts logically, showing why things happen the way they do.
He described two basic moves a thinker could make. One move, later called induction, starts from observed details and climbs toward general principles. The other move, deduction, starts from general principles and works down to new specific conclusions. Induction builds the staircase; deduction walks you back down to reach exact predictions.
This two-directional framework, often called analysis and synthesis, shows up again and again in the history of science. Medieval scholars like Thomas Aquinas (1225–1274) and William of Ockham (1287–1347) refined rules for how observations could justify inductive leaps. They kept alive the idea that science needed both careful looking and careful thinking. Still, no one had pinned down a single recipe that always guaranteed truth.
The Battle over Induction
By the 1600s, natural philosophers demanded more than Aristotle’s balance. Francis Bacon (1561–1626) complained that thinkers jumped too quickly from a few examples to sweeping claims. In his vision, the true scientist collected mountains of specific observations first, cleaned the mind of biases, and only then climbed, step by slow step, to reliable laws. His idea of method was a ladder built entirely out of recorded facts.
Isaac Newton (1642–1727) seemed to put a version of this into practice. His law of gravitation emerged not from first principles dreamed up in an armchair, but from mathematical analysis of astronomical data. Newton famously declared that he framed no hypotheses—meaning he wanted his theories to be drawn from observed phenomena, not invented from pure thought.
During the 1800s, John Stuart Mill (1806–1873) and William Whewell (1794–1866) debated just how ironclad such induction could be. Mill argued that science searches for regularities among events and eventually welds them into laws, like finding that all metals expand when heated. Whewell stressed that the mind contributes its own organizing ideas. For him, real discovery involved a creative leap—what he called the colligation of facts—where a scientist invents a hypothesis that suddenly makes many separate observations click into place, like seeing a picture hidden in a puzzle. Then testing must confirm the leap.
Both sides agreed that the hypothetico-deductive method (often called H‑D) was key: propose a hypothesis, deduce what new observations you should expect if it’s true, check those predictions, and see if the hypothesis survives. But they disagreed about whether that process made knowledge certain or just increasingly probable. That uncertainty set the stage for an even bigger question.
Popper: Falsification, Not Proof
In the 20th century, philosopher Karl Popper (1902–1994) looked at the H‑D method and spotted a logical trap. No matter how many observations support a hypothesis, you can never be absolutely certain it’s true. (One day a single black swan can overthrow a lifetime of seeing only white swans.) But you can be certain that a hypothesis is false if the evidence contradicts it. That asymmetry, Popper argued, is the engine of science.
He proposed falsifiability as the mark of real science: a genuine scientific claim must be testable enough that we can imagine an observation that would prove it wrong. Marxism or psychoanalysis, Popper worried, could explain away any possible fact, so they were not falsifiable. Real science, on the other hand, takes big risks. If Einstein’s theory predicts that starlight will bend a certain amount during an eclipse, and it doesn’t, the theory is in trouble.
Popper admitted that scientists do sometimes rescue a theory by adding an extra, ad hoc hypothesis. For instance, when Uranus’s orbit didn’t match Newton’s predictions, some astronomers suggested an unseen outer planet. That clever guess turned out to be correct—Neptune was discovered. Popper acknowledged that such “immunizations” could be scientifically fruitful. The key was whether the new idea led to further testable predictions. So even falsification wasn’t a simple recipe you could mechanically follow every time.
Kuhn: Science Changes Its Method
In 1962, a historian of science named Thomas Kuhn (1922–1996) lobbed a grenade into the debate. He argued that scientific work normally happens inside a paradigm—a shared set of puzzles, tools, and standards that a community treats as the right way to do science. During these normal periods, textbooks make it look as if there is one stable method. But paradigms aren’t fixed forever.
Over time, anomalies pile up—problems the current paradigm can’t solve. Eventually, a crisis pushes the community to switch to a new paradigm. When that shift happens, the very rules for what counts as a good explanation often change, too. The value that scientists place on simplicity, accuracy, or even what counts as an observation can shift. Kuhn’s point was that method is not a timeless engine bolted onto science; it’s built into each paradigm and transforms with it.
Kuhn’s friend Paul Feyerabend (1924–1994) went further. He claimed that any fixed rule for doing science would slow progress. Looking at history, he argued that Galileo advanced partly through clever persuasion, not just pure logic. Feyerabend concluded that the only rule that always works is “anything goes.” Most philosophers didn’t go that far, but after Kuhn and Feyerabend, far fewer people believed that a single, universal scientific method lay waiting to be discovered.
So What Is Science Really?
If there’s no one grand recipe, what makes science special? Recent thinkers have offered a quieter answer. Science doesn’t use a method that’s totally unique. Instead, it takes ordinary, everyday ways of investigating—looking, guessing, checking, asking others—and makes them much more systematic. Descriptions are more precise, predictions are spelled out in greater detail, alternative explanations are deliberately hunted down, and errors are chased out more stubbornly.
Philosophers like Hoyningen-Huene and Susan Haack (b. 1945) argue that the real difference is a matter of degree, not kind. A scientist testing a new material isn’t running a mysterious ritual. She’s doing exactly what you do when you figure out which route to take to school: collect evidence, form a belief, test it against new experience, and adjust. In science, though, that process is stretched and sharpened—checked by wider communities, recorded transparently, and pushed to cover bigger, stranger questions.
Early 2000s worries about the “replication crisis,” when many published studies couldn’t be repeated, underlined this. Failures to reproduce results don’t mean there’s no method at all; they remind us that method is never a guarantee, only a disciplined effort to be less wrong over time. Science remains our most reliable way of knowing, not because it follows a magic formula, but because it refuses to stop asking, “Am I sure?”
Think about it
- If two different scientists use the same procedure but get opposite results, is the procedure broken—or is something else going on?
- Can you imagine a question so big that no single, step-by-step method could ever answer it? What might that question be?
- Should we still call an activity “science” if it uses methods that look nothing like what you learn in science class? What would matter more—the method itself or the results it produces?
