Lakatos and Research Programs

Core Idea

Imre Lakatos attempted to reconcile Popper's falsificationism with Kuhn's historical observations. He proposed that science progresses through research programs with a hard core (fundamental assumptions protected from refutation) surrounded by a protective belt of auxiliary hypotheses that face empirical tests. A research program is progressive if it generates new theories with novel predictive content. Scientists rationally continue with a progressive program even as it faces anomalies, but switch to a better program when the old becomes degenerating.

Explainer

You enter this topic knowing two foundational views in philosophy of science. Popper's falsificationism holds that scientific theories are defined by their falsifiability — a good theory makes risky predictions that could be decisively refuted. When a prediction fails, the theory must be abandoned. This gives science its progressive, self-correcting character. Kuhn's paradigm theory (your hard prerequisite) complicates this picture: science doesn't actually operate by Popperian rules. Scientists defend paradigms against anomalies, engage in normal science without questioning core assumptions, and shift paradigms only in dramatic, gestalt-switch-like revolutions. Falsifying evidence, on Kuhn's historical account, rarely leads to immediate rejection of the core theory.

Lakatos aimed to preserve what was right in both views while avoiding their weaknesses. His central concept is the scientific research program — a structured sequence of theories united by a common hard core of fundamental assumptions that scientists agree to treat as non-negotiable. For Newtonian mechanics, the hard core includes Newton's three laws and the law of gravitation. For Darwinian biology, it includes the principles of variation, heredity, and selection. When an anomaly appears — an orbit that doesn't fit predictions — scientists don't abandon the hard core. Instead, they adjust the protective belt: the auxiliary hypotheses, initial conditions, and background assumptions that surround and protect the core. An anomalous orbit can be explained by positing an undetected planet, by revising measurement assumptions, or by complicating initial conditions.

This is where Lakatos adds normative bite that Kuhn lacks. Not all protective belt adjustments are equally rational. An adjustment is ad hoc if it merely patches over the anomaly without generating new testable predictions. But an adjustment is progressive if it leads to new theories that successfully predict novel facts. When the Newtonian program predicted Neptune's existence (and Neptune was subsequently found), the auxiliary hypothesis was progressive — it genuinely expanded the program's empirical content. When the Ptolemaic system kept adding epicycles to save the phenomena, the adjustments were increasingly ad hoc and the program was degenerating — it accommodated past observations but stopped making successful new predictions.

For Lakatos, the rational unit of scientific evaluation is not a single theory at a single moment (Popper's target) or an incommensurable paradigm (Kuhn's unit), but a research program over time. Scientists are rational to continue with a degenerating program for a while — perhaps new data will revive it, perhaps a young theorist will find a progressive variant. But there is a rational point at which switching to a better program is not a gestalt switch driven by sociology but a scientifically motivated decision. This is Lakatos's attempt to restore methodological rationality to science while accommodating Kuhn's historical realism. The key question his critics press: does the progressive/degenerating distinction give real normative guidance, or can every program look progressive in retrospect?