The demarcation problem asks: What criteria distinguish science from non-science, pseudo-science, or metaphysics? Different philosophical schools propose different answers: falsifiability (Popper), verifiability (logical positivists), normal puzzle-solving (Kuhn), or research program progressivity (Lakatos). Each criterion captures something important but faces counterexamples.
Examine boundary cases: astrology vs astronomy, alchemy vs chemistry, Intelligent Design vs evolutionary biology. For each, apply different demarcation criteria and note which succeed or fail.
From your introduction to philosophy of science you know that science is not simply "what scientists do" — it involves specific methods of inquiry, distinctive standards of evidence, and a particular relationship between theory and observation. The demarcation problem asks the sharper question: can we specify, in precise terms, what separates scientific claims from non-scientific ones? This matters practically. Courts have had to decide whether Intelligent Design counts as science. Funding agencies allocate billions based on implicit demarcation judgments. The question is not merely academic.
Karl Popper's answer is the most famous: falsifiability. A claim is scientific if it is in principle refutable by observation. "All ravens are black" is scientific because a single white raven would falsify it. "God loves those who suffer" is not scientific because no possible observation could refute it — the believer can always reinterpret any outcome as consistent with divine love. Falsifiability elegantly handles why astrology feels unscientific: astrologers read favorable predictions into any outcome, insulating their claims from disconfirmation. The criterion also explains why psychoanalysis worried Popper — Freudian theory seemed capable of accommodating any patient behavior, which meant it predicted nothing and therefore risked explaining everything.
Falsificationism faces serious objections, however. The Duhem-Quine thesis (which you may encounter in later study) shows that scientific theories are never tested in isolation — when an experiment fails, it could falsify any of the auxiliary assumptions bundled with the theory, not necessarily the core hypothesis. Scientists routinely protect central theories by revising peripheral assumptions. This is not bad science; it is how science actually works. Newton's mechanics predicted a slight precession of Uranus's orbit that didn't match observation. Astronomers didn't abandon Newtonian gravity — they predicted a new planet (Neptune) to account for the discrepancy. The theory was not falsified; an auxiliary assumption about the number of planets was revised. Popper's criterion, applied strictly, would have required abandoning some of the most productive theories in scientific history.
Kuhn and Lakatos offer more historically grounded accounts. Kuhn holds that normal science is defined by commitment to a paradigm — a shared exemplary practice, not a criterion. Scientists solve puzzles within the paradigm; anomalies accumulate until a crisis forces a paradigm shift. On Kuhn's view, demarcation is sociological before it is logical. Lakatos refines this with the concept of research programmes: a progressive programme generates novel predictions that are confirmed; a degenerative one only adds epicycles after the fact to protect a failing core theory. These accounts explain better than Popper why scientists behave as they do — but they arguably push demarcation into degree and judgment rather than sharp criteria.
The lesson of the demarcation debate is not that the question is unanswerable but that no single criterion is sufficient. Falsifiability, testability, explanatory fruitfulness, novel prediction, intersubjective verifiability — these are all markers of scientific quality, and paradigmatic sciences score high on most of them while pseudosciences typically score low on several. Astrology makes predictions but they are vague; it insulates itself from failure; it generates no progressive research programme; its mechanisms invoke no known physical processes. The demarcation problem remains unsolved as a matter of sharp logical definition, but the diagnostic tools it has generated let us assess any candidate discipline on multiple dimensions rather than seeking a single bright line.
Topics in reflective domains aren't scored by quiz answers. Read, reflect, and mark when you've thought it through.