Scientific method

The scientific method is not a single algorithm. It is a family of epistemic practices — observation, hypothesis formation, prediction, testing, and revision — that share a common functional logic: beliefs are held accountable to consequences they did not anticipate. The method is not unique to modern science; it appears, in fragmentary form, in Aristotelian biology, medieval Islamic optics, and Galilean mechanics. What changed was not the method but its institutionalization: the construction of social systems — journals, laboratories, funding agencies, peer review — that made the feedback loop systematic rather than occasional.

The standard textbook version — hypothesis, experiment, conclusion — is a pedagogical fiction. Real scientific practice is messier: hypotheses are rarely crisp, experiments are rarely decisive, and conclusions are rarely final. Kuhn's concept of normal science captures this better: most research is puzzle-solving within a paradigm, not revolutionary overturning of one. The method works not because scientists are especially rational but because the institutional architecture selects for results that survive repeated testing.

The deepest question about the scientific method is not whether it works (it does) but whether it is the only method that works. Feminist critics have shown that the supposedly neutral method encodes specific values — objectivity-as-distance, quantification-as-truth, replication-as-gold-standard — that privilege certain kinds of knowledge and marginalize others. Social epistemology adds that the method's reliability is not a property of individual cognition but of distributed social processes. The scientific method is best understood not as a cognitive algorithm but as a social technology for converting disagreement into convergence — a technology that works well in some domains and poorly in others.