Reading passage
Few ideas command as much prestige in modern societies as science, yet the question of what distinguishes genuine science from mere pretension remains surprisingly unsettled. Philosophers call this the demarcation problem, and answers to it have shifted dramatically over the past century. Each proposed solution has captured something real about scientific practice while collapsing under objections, and the debate continues to shape how courts, governments and publics decide which claims deserve trust.
The first influential modern answer came from the Vienna Circle, a group of philosophers and scientists who met in the nineteen twenties. Their doctrine, logical positivism, held that a statement is meaningful only if it can be verified by observation or is true by definition. On this verification principle, scientific claims earn their status by being anchored in experience, while the sentences of metaphysics, theology and much traditional philosophy were dismissed as literally meaningless. The ambition was austere and exhilarating: to rebuild knowledge on a foundation of indisputable fact.
Karl Popper objected that verification was both impossible and misdirected. No finite set of observations can verify a universal law, he noted, since the next swan may always be black. What distinguishes science, he proposed, is falsifiability: a genuinely scientific theory forbids certain outcomes and thereby risks being wrong. Theories that can accommodate every possible result, explaining away all contrary evidence, may be psychologically satisfying but are not scientific. On this view, science advances through bold conjectures followed by severe attempts at refutation, and a theory that survives such tests is not proven but merely corroborated, provisionally retained until a better rival appears.
Historians soon noticed that actual scientists rarely behave as Popper recommended. Thomas Kuhn, studying past episodes of scientific change, argued that researchers normally work within a paradigm, a shared framework of assumptions, methods and exemplary solutions that defines what counts as a legitimate problem. During periods of normal science, anomalies are treated as puzzles to be solved rather than as refutations. Only when anomalies accumulate and a rival framework emerges does a scientific revolution occur, after which the community adopts a new paradigm. Kuhn's most controversial claim was that successive paradigms are incommensurable, meaning they cannot be measured against one another by any neutral standard.
Imre Lakatos attempted to reconcile Popper's rigour with Kuhn's historical realism. In his account, scientists commit themselves to research programmes built around a hard core of assumptions that is protected from direct falsification. Surrounding the core is a protective belt of auxiliary hypotheses, which can be modified to absorb awkward evidence. A programme is progressive, Lakatos argued, when its modifications predict novel facts; it is degenerating when it merely patches itself after the fact. Rationality thus applies to the trajectory of a programme over time, not to isolated theories at a single moment.
Paul Feyerabend drew a more radical moral from the same history. Examining episodes such as Galileo's defence of heliocentrism, he concluded that progress often required violating every rule philosophers had proposed. His slogan that anything goes was not a celebration of chaos but a warning: any fixed method, enforced consistently, would have blocked some of science's greatest achievements. Feyerabend's epistemological anarchism remains a standing provocation, cited less as a programme than as a rebuke to methodological complacency.
A further complication, associated with Pierre Duhem and W. V. Quine, concerns the logic of testing itself. Theories never face evidence alone, because every prediction depends on background assumptions about instruments, initial conditions and auxiliary laws. When a prediction fails, logic does not dictate which assumption to abandon; the blame can always be shifted elsewhere. This underdetermination of theory by evidence suggests that scientific choice involves judgements that go beyond strict deduction, a conclusion that opened the door to sociological studies of how such judgements are actually made.
Contemporary philosophy of science has largely abandoned the search for a single criterion. Bayesian approaches model confirmation as the updating of degrees of belief, while other scholars study the role of values, noting that decisions about evidence thresholds inevitably carry ethical consequences, for instance in drug approval or climate policy. The demarcation problem, once posed as a request for a definition, is now treated as a family of practical questions about reliability, and the failure to find one bright line is regarded less as a defeat than as a discovery about what science really is.