Scientific method

• 1. Introduction

  ›

  The scientific method is not a single recipe, and one of its most famous victories began as a wrong guess. Linus Pauling proposed that DNA might be a triple helix. James Watson and Francis Crick considered the same idea, then discarded it once they understood from existing data that Pauling was mistaken. That small drama, hypothesis, doubt, and revision, captures something stranger than the tidy diagram taught in classrooms. The scientific method is an empirical way of acquiring knowledge through careful observation, rigorous skepticism, hypothesis testing, and experimental validation. It has characterized science since at least the 17th century. Yet some of the people who study it most closely insist it barely exists. Why would a physicist publish an essay titled "There Is No Scientific Method"? Why did a famous book called Against Method argue that the only universal rule is "anything goes"? And how did a phrase that did not even have a name until the 19th century come to define what counts as real knowledge?

• 2. An Idea With Many Fathers

  ›

  Aristotle, Epicurus, and the ancient Stoics all left early expressions of empiricism, long before anyone used the words. Alhazen, Avicenna, and Al-Biruni carried the work forward, as did Roger Bacon and William of Ockham. The development of rules for scientific reasoning was never straightforward. The method has been the subject of intense and recurring debate, with eminent natural philosophers arguing for the primacy of rival approaches. In the Scientific Revolution of the 16th and 17th centuries, Francis Bacon and Robert Hooke pushed empiricism forward. Rene Descartes described a rationalist approach, while inductivism rose to prominence through Isaac Newton and his followers. Newton postulated four principles that form the basis of modern science. Experiments were advocated by Francis Bacon and performed by Giambattista della Porta, Johannes Kepler, and Galileo Galilei. The skeptic Francisco Sanches contributed theoretical work, joined by John Locke, George Berkeley, and David Hume. The very term scientific method emerged only in the 19th century, alongside the institutional growth of science and new words drawing boundaries, such as scientist and pseudoscience. Through the 1830s to the 1850s, when Baconianism was popular, naturalists like William Whewell, John Herschel, and John Stuart Mill argued over induction and facts. Each was wrestling with the same stubborn problem of how knowledge gets generated.

• 3. The Skeptics Of Method

  ›

  By the 1960s and 1970s, Thomas Kuhn and Paul Feyerabend questioned whether any universal scientific method existed at all. They largely replaced the idea of science as a homogeneous, universal method with science as a heterogeneous, local practice. In the 1975 first edition of Against Method, Feyerabend argued there were no universal rules of science. He suggested that for any specific method or norm, one can find a historic episode where breaking it advanced science. Karl Popper disagreed, as did Gauch in 2003. The physicist Lee Smolin pressed the case further in his 2013 essay "There Is No Scientific Method," offering two ethical principles instead. The historian Daniel Thurs, writing in the 2015 book Newton's Apple and Other Myths about Science, concluded that the method is a myth or, at best, an idealization. Robert Nola and Howard Sankey complicated the rebellion in their 2007 book Theories of Scientific Method. They argued that Feyerabend, despite his title, accepted certain rules and even tried to justify them with a meta methodology. Staddon, writing in 2017, took yet another view, that science is best understood through examples. Even so, algorithmic moves like disproving a theory by experiment reach back to Alhacen in 1027 and his Book of Optics, and to Galileo in 1638 with Two New Sciences.

• 4. Anatomy Of Inquiry

  ›

  Characterizations, hypotheses, predictions, and experiments are the elements the scientific community generally agrees on. They tend to be more characteristic of experimental sciences like physics, chemistry, biology, and psychology than of the social sciences. The process is iterative and cyclical, with information continually revised, and each element is subject to peer review for possible mistakes. A hypothesis is a suggested explanation, often taking the form of a mathematical model, and it must be falsifiable. That means one must be able to identify a possible outcome that conflicts with the hypothesis, or it cannot be meaningfully tested. Scientists are free to draw on creativity, ideas from other fields, inductive reasoning, or Bayesian inference to imagine explanations. Albert Einstein observed that "there is no logical bridge between phenomena and their theoretical principles." Charles Sanders Peirce, borrowing from Aristotle's Prior Analytics, called the incipient stage of inquiry abductive reasoning, instigated by the "irritation of doubt." When several explanations fit equally well, Occam's Razor serves as a rule of thumb for choosing the simplest. To guard against confirmation bias, strong inference urges entertaining multiple alternative hypotheses at once. Not every inquiry uses every element, nor in the same order, and chance has played a role in numerous discoveries that never followed the textbook model.

• 5. The DNA Case File

  ›

  From 1944 to 1953, the discovery of the structure of DNA played out as a working illustration of these elements. By 1950, genetic inheritance was known to have a mathematical description, starting with Gregor Mendel, and DNA was known to carry genetic information through Oswald Avery's transforming principle. But how DNA stored that information remained unclear. In Bragg's laboratory at Cambridge University, researchers made X-ray diffraction pictures of molecules, beginning with crystals of salt. Watson, Crick, and others hypothesized that DNA had a helical structure, which implied its diffraction pattern would be x shaped. That prediction followed from the Cochran-Crick-Vand-Stokes theorem, which explained why diffraction from helical structures produces x-shaped patterns. Watson and Crick first showed an incorrect proposal to a King's College London team of Rosalind Franklin, Maurice Wilkins, and Raymond Gosling. Franklin immediately spotted flaws concerning the water content. Later Watson saw Franklin's photo 51, a detailed X-ray image showing an X-shape, confirming the helix. Guided by bond lengths deduced by Linus Pauling and by Franklin's images, and using Chargaff's rules of base pairing, the pair inferred the structure by modeling the physical shapes of the nucleotides. In their first paper they noted, "It has not escaped our notice that the specific pairing we have postulated immediately suggests a possible copying mechanism for the genetic material."

• 6. When Two People See One Sunrise

  ›

  Johannes Kepler used Tycho Brahe's method of projecting the Sun's image through a pinhole onto paper, rather than looking directly at it. Observing the same sunrise, the two scientists reached different conclusions. Brahe held that total solar eclipses were impossible, but Kepler knew of historical accounts of them and disagreed. He deduced that images grow more accurate as the aperture widens, a fact now fundamental to optical system design. Norwood Russell Hanson used this episode to illustrate a deeper claim, that there is no pure observation. Theory is required to interpret empirical data, so observation is influenced by the observer's conceptual framework. Hanson introduced this idea in 1958, using the concept of gestalt to show how preconceptions shape both observation and description. He pointed to the initial rejection of Golgi bodies as a mere artefact of staining technique. Pre-existing beliefs can distort results through confirmation bias, since people tend to observe what they expect to observe. One vivid case was the belief that a galloping horse splays all four legs while none touch the ground, an image painted by its supporters. The first stop-action photographs by Eadweard Muybridge proved this false, showing the legs gathered together instead.

• 7. Two Engines Of Reason

  ›

  Deductive reasoning derives specific conclusions from general principles, and if the premises are true, the conclusion must be true. Inductive reasoning runs the other way, building general principles from observations, so its conclusions are probable but not guaranteed. Scientific inquiry uses both. Gravitational theory shows the pairing at work across centuries. It took thousands of years of measurements from Chaldean, Indian, Persian, Greek, Arabic, and European astronomers to record the motion of Earth. Kepler and others generalized that data inductively, and Newton unified prior theory and measurement into his laws of motion in 1727. Then came the counterexample. In 1859, Le Verrier pointed out problems with the perihelion of Mercury that showed Newton's theory to be at least incomplete. That discrepancy occurred to Einstein as a possible early test of relativity, and his relativistic calculations matched observation far more closely. The hypothetico-deductive method tests a hypothesis through its implications rather than directly. A positive test cannot definitively prove a hypothesis, since only the inference from not-B to not-A is valid logic. As Gillies put it, "successful theories are those that survive elimination through falsification." In fields like biology, where general laws are few, deductive reasoning is sometimes replaced by abductive reasoning, the search for the most plausible explanation.

• 8. What Makes A Theory Good

  ›

  Good theories are accurate, internally consistent, explanatory beyond the required data, unifying of disparate phenomena, and fruitful for further research. When empirical evidence is limited, scientists favor parsimony, the principle that the simplest explanation is recommended when data supports several. Newton expressed it directly: "We are to admit no more causes of natural things than such as are both true and sufficient to explain their appearances." Yet these criteria contain subjective elements and should be treated as heuristics. Quoting Bird, such criteria "cannot determine scientific choice," because which features satisfy them may be disputable, the criteria are imprecise, and people disagree on how to weigh them. Principles of invariance offer another guide, the idea that good structures are independent of perspective. They gained real weight only after Einstein's relativity, which reduced everything to relations. David Deutsch put it in 2009: "the search for hard-to-vary explanations is the origin of all progress." Einstein's thought experiment of a lab suspended in empty space showed the power of an invariant observation. An experimenter floating in zero gravity would feel an upward acceleration as gravity itself, letting Einstein equate gravitational and inertial mass. The opposite of hard-to-vary is the theory that resists falsification, which Wolfgang Pauli dismissed as being "not even wrong."