Chapter 12: Lakatos & Research Programmes

From Hungary to Cambridge to LSE

Imre Lakatos was born Imre Lipschütz in Debrecen, Hungary, in 1922. His intellectual biography reads like a picaresque novel. During the Nazi occupation, he changed his name to Imre Molnár to avoid persecution as a Jew, and later adopted the name Lakatos (Hungarian for “locksmith”). He was active in the Hungarian communist resistance, and after the war served in the Ministry of Education, where he was involved in the reform of Hungarian higher education.

In 1950, Lakatos was arrested by the Stalinist government and imprisoned for over three years on charges of “revisionism.” This experience had a profound effect on his intellectual development, turning him against authoritarianism in all its forms and predisposing him to Popper's philosophy of critical rationalism and the open society.

After the Hungarian Revolution of 1956, Lakatos fled to the West, eventually settling in England. He completed a second Ph.D. at Cambridge under R.B. Braithwaite, producing a brilliant thesis on the methodology of mathematics that was later published as Proofs and Refutations (1976). In 1960, he joined the London School of Economics at Popper's invitation, where he remained until his untimely death in 1974.

At LSE, Lakatos became the key mediator between the Popperian and Kuhnian traditions. He was deeply committed to Popper's rationalism but equally impressed by Kuhn's historical arguments. His methodology of scientific research programmes was his attempt to save Popper's key insights — the centrality of criticism, the importance of empirical content, the progressiveness of science — while acknowledging the historical complexities that Kuhn had revealed.

Before developing his philosophy of empirical science, Lakatos made a major contribution to the philosophy of mathematics with Proofs and Refutations, a work that applied Popperian ideas to mathematical knowledge. Written as a Socratic dialogue set in a mathematics classroom, the book traces the historical development of Euler's conjecture about polyhedra (V - E + F = 2) through a series of proofs, counterexamples, and revisions.

Lakatos showed that mathematical knowledge, like empirical knowledge, develops through a process of conjecture and refutation. When a counterexample threatens a theorem, mathematicians do not simply abandon the theorem; they may modify the proof, restrict the domain, redefine key terms, or incorporate the counterexample into a more sophisticated version of the theorem. This process of “proof-analysis” bears a striking resemblance to the way empirical scientists respond to anomalies — a resemblance that foreshadowed Lakatos's later work on research programmes.

“Mathematics does not grow through a monotonous increase of the number of indubitably established theorems but through the incessant improvement of guesses by speculation and criticism, by the logic of proofs and refutations.”— Imre Lakatos, Proofs and Refutations (1976), p. 5

The key insight of Proofs and Refutations — that knowledge grows not by accumulating certainties but by improving fallible conjectures through criticism — became the philosophical foundation for Lakatos's methodology of scientific research programmes. The MSRP can be seen as an attempt to apply the same dialectical model to empirical science.

The MSRP was first presented in Lakatos's landmark paper “Falsification and the Methodology of Scientific Research Programmes” (1970), written for the proceedings of the 1965 London colloquium on Kuhn's work. It offers a framework for understanding scientific development that is more sophisticated than either Popper's falsificationism or Kuhn's paradigm theory.

A scientific research programme is a series of theories linked by a common set of fundamental assumptions and a shared methodology for developing and modifying those theories. Each research programme consists of three elements:

The Hard Core

The hard core consists of the fundamental assumptions that define the research programme and are held to be irrefutable by methodological decision. Scientists working within the programme treat the hard core as sacrosanct: any anomaly must be explained without modifying it. The hard core of Newtonian mechanics, for example, consists of Newton's three laws of motion and his law of universal gravitation.

The concept of the hard core captures Kuhn's insight that scientists do not abandon their fundamental commitments at the first sign of trouble, while giving this insight a rational justification that Kuhn's framework lacked. Protecting the hard core is not mere dogmatism; it is a methodological strategy that allows the programme to be developed in sufficient detail to reveal its potential.

The Protective Belt

Surrounding the hard core is the protective belt of auxiliary hypotheses, which can be modified, adjusted, or even replaced in response to anomalies. When a prediction fails, the scientist modifies the protective belt rather than the hard core. The protective belt includes specific models, initial conditions, observational theories, and mathematical approximations.

For example, when Newtonian predictions about Uranus's orbit proved wrong, scientists modified the protective belt by adding an auxiliary hypothesis (the existence of Neptune) while leaving the hard core (Newton's laws) untouched. This modification was progressive because it led to the prediction and discovery of a new planet. By contrast, when Newtonian predictions about Mercury's perihelion proved wrong and no modification of the protective belt could fix the problem, the anomaly became evidence of the programme's limitations.

The Heuristic

Each research programme also has a heuristic — a set of methodological guidelines that tell scientists how to develop the programme. Lakatos distinguished the negative heuristic (which forbids modifications to the hard core) from the positive heuristic (which provides a general plan for constructing the protective belt and developing the programme's predictive potential).

“The positive heuristic sets out a programme which lists a chain of ever more complicated models simulating reality: the scientist's attention is riveted on building his models following instructions which are laid down in the positive part of his programme. He ignores the actual counterexamples, the available ‘data’... The positive heuristic of the programme saves the scientist from becoming confused by the ocean of anomalies.”— Imre Lakatos, “Falsification and the Methodology of Scientific Research Programmes” (1970), p. 135

The concept of the positive heuristic is perhaps Lakatos's most original contribution. It captures the idea that research programmes have an internal dynamic — a built-in plan for their own development — that is relatively independent of the anomalies they face. Scientists do not simply respond to refutations; they follow a systematic programme of model-building guided by the positive heuristic. This explains why, as Kuhn had observed, scientists often ignore anomalies: they are following the positive heuristic of their programme rather than responding to empirical difficulties.

The central evaluative concept in Lakatos's methodology is the distinction between progressive and degenerating problem shifts (or research programmes). This distinction replaces Popper's simple notion of falsification with a more nuanced account of how research programmes succeed or fail over time.

A research programme is theoretically progressive if each new theory in the series has excess empirical content over its predecessor — that is, if it predicts novel facts that the old theory did not predict. It is empirically progressive if some of this excess content is confirmed — if at least some of the novel predictions are corroborated. A programme that is both theoretically and empirically progressive constitutes a progressive problem shift.

A research programme is degenerating if it fails to predict novel facts, or if the novel facts it predicts are not confirmed. A degenerating programme grows only by ad hoc accommodations to anomalies — it explains facts after the event but never anticipates them. The modifications to its protective belt are not driven by the positive heuristic but by the need to patch up failures.

“A research programme is said to be progressing as long as its theoretical growth anticipates its empirical growth, that is, as long as it keeps predicting novel facts with some success (‘progressive problemshift’); it is stagnating if its theoretical growth lags behind its empirical growth, that is, as long as it gives only post-hoc explanations either of chance discoveries or of facts anticipated by, and discovered in, a rival programme (‘degenerating problemshift’).”— Imre Lakatos, “History of Science and Its Rational Reconstructions” (1971), p. 104

On Lakatos's view, a research programme should be abandoned when it becomes degenerating and a progressive rival is available. This is more lenient than Popper's falsificationism (a single refutation does not condemn a programme) but more demanding than Kuhn's paradigm theory (a programme must eventually produce novel predictions or be judged degenerating). The criterion of progressiveness provides a rational standard for theory choice without requiring the unrealistic Popperian ideal of instant falsification.

Newtonian Mechanics: A Progressive Programme

Lakatos's paradigm example of a progressive research programme is Newtonian mechanics. Its hard core (Newton's laws of motion and gravity) was held fixed while its protective belt was continually developed to accommodate new phenomena. The programme was spectacularly progressive: it predicted the return of Halley's comet, the existence of Neptune, the oblateness of the Earth, and countless other novel facts.

Even the programme's failures were instructive. The anomalous precession of Mercury's perihelion was an anomaly that the Newtonian programme could not resolve within its protective belt. But this failure did not immediately condemn the programme; it became a serious problem only when Einstein's general relativity provided a progressive rival that could explain the anomaly while making additional novel predictions.

Marxism: A Degenerating Programme

Lakatos, drawing on his own bitter experience in Hungary, argued that Marxism provides an example of a degenerating research programme. In the nineteenth century, Marxism made bold predictions: the immiseration of the working class, the concentration of capital, the inevitability of proletarian revolution in the most advanced industrial countries. None of these predictions came true. Rather than acknowledging the failures, Marxists continually modified their auxiliary hypotheses to accommodate unfavorable developments: imperialism explained why revolution had not occurred in advanced countries; false consciousness explained why the working class did not recognize its interests; the “uneven development” of capitalism explained why revolution had occurred in Russia rather than England.

Each of these modifications was ad hoc: it was introduced solely to explain away a failure, without generating new testable predictions. The Marxist programme, Lakatos argued, had become degenerating — it could only explain facts after the event, never anticipating them.

Freudian Psychoanalysis: Another Degenerating Programme?

Like Popper, Lakatos regarded Freudian psychoanalysis as scientifically problematic, but for a different reason. Where Popper argued that psychoanalysis was unfalsifiable (and therefore not science at all), Lakatos argued that it was a degenerating research programme. Psychoanalysis did make predictions — for example, predictions about the effects of certain childhood experiences on adult personality — but these predictions were consistently vague, and when they failed, psychoanalysts modified their auxiliary hypotheses without generating new, independently testable predictions. The programme was degenerating not because it was unfalsifiable but because it consistently failed to produce progressive problem shifts.

Lakatos positioned his MSRP as the “best of both worlds” — combining Popper's rationalism with Kuhn's historical realism. Understanding the comparisons is essential for grasping Lakatos's contribution.

Lakatos agreed with Kuhn against Popper that scientists do not and should not abandon their theories at the first sign of falsification. The protection of the hard core is not dogmatism but sound methodology. But he agreed with Popper against Kuhn that there are objective, rational standards for evaluating research programmes. The distinction between progressive and degenerating programmes provides a criterion of appraisal that is neither as rigid as Popper's falsificationism nor as sociological as Kuhn's paradigm theory.

Lakatos also agreed with Kuhn that science involves the coexistence of competing frameworks, but he insisted (against Kuhn) that this competition can be rationally adjudicated. Rival research programmes can and should coexist, and the rational scientist will switch allegiance from a degenerating programme to a progressive rival — not at any particular moment (as Popper would require) but when the evidence of degeneration becomes sufficiently clear.

One of the most significant difficulties with Lakatos's methodology is the problem of temporality. When should a scientist conclude that a research programme is degenerating? Lakatos acknowledged that his methodology cannot provide “instant rationality” — that is, a decision procedure that tells the scientist what to do at any given moment. Whether a programme is progressive or degenerating can only be judged in retrospect, after sufficient time has passed to assess its trajectory.

This creates a dilemma. If we can only judge programmes in hindsight, then Lakatos's methodology cannot provide real-time guidance to working scientists. A programme that appears to be degenerating today might stage a dramatic comeback tomorrow. Lakatos gave the example of Bohr's atomic theory, which went through a period of apparent stagnation before being revitalized by the development of quantum mechanics.

“One may rationally stick to a degenerating programme until it is overtaken by a rival and even after. What one must not do is to deny its poor public record... It is perfectly rational to play a risky game: what is irrational is to deceive oneself about the risk.”— Imre Lakatos, “Falsification and the Methodology of Scientific Research Programmes” (1970), p. 117

This response has struck many critics as insufficient. If scientists may “rationally stick to a degenerating programme,” then what practical guidance does the MSRP provide? Larry Laudan accused Lakatos of offering a methodology that is “toothless” — that cannot be used to criticize any actual scientific decision. Others have argued that Lakatos's concession undermines the very distinction between progressive and degenerating programmes on which his methodology depends.

Paul Feyerabend, who had been a close friend and intellectual sparring partner of Lakatos (their planned joint book For and Against Method was never completed due to Lakatos's death), raised the most penetrating objections to the MSRP. Feyerabend argued that Lakatos's methodology, for all its sophistication, ultimately faces the same problems as Popper's.

The central challenge is: who decides when a research programme is degenerating, and on what basis? If the answer is “the scientific community,” then Lakatos's methodology collapses into Kuhn's sociology of science — the judgment is a collective social decision, not a logical one. If the answer is “the methodology itself,” then we need precise criteria for distinguishing progressive from degenerating programmes, and Lakatos's own concession that programmes may be “rationally” retained even when degenerating undermines the force of these criteria.

“Lakatos's methodology is... nothing but words that SOUND like the elements of a methodology. It gives no advice to the scientist, it does not tell him how to proceed, and it does not tell others how to evaluate his research. It is applicable, it produces judgments — but only long after the event.”— Paul Feyerabend, “On the Critique of Scientific Reason” (1976), p. 312

Feyerabend pressed the point with a thought experiment: suppose a programme is currently degenerating but a brilliant young scientist has an idea that might revitalize it. Should she be discouraged from pursuing this idea on the grounds that the programme is degenerating? If so, the methodology stifles creativity. If not, then the methodology provides no guidance at all.

Lakatos never fully answered Feyerabend's challenge. His death in 1974 at the age of 51 cut short what would undoubtedly have been a further development of his ideas. The planned correspondence between Lakatos and Feyerabend, published posthumously as For and Against Method (1999), reveals the depth of their disagreement but also the mutual respect and intellectual excitement that characterized their debate.

One of Lakatos's most distinctive contributions was his account of the relationship between the philosophy and the history of science. He proposed that the philosophy of science provides normative “rational reconstructions” of the history of science, and that these reconstructions should be tested against the actual historical record. The philosophy of science that produces the most faithful rational reconstruction of the greatest achievements of science is the best philosophy of science.

Lakatos captured this idea in a famous paraphrase of Kant: “Philosophy of science without history of science is empty; history of science without philosophy of science is blind.” No philosophical methodology can be accepted if it condemns the greatest achievements of science (such as Newton's mechanics or Einstein's relativity) as methodologically deficient. Conversely, no historical account is adequate unless it is guided by a philosophical understanding of what counts as rational scientific behavior.

This methodological principle provides a criterion for comparing rival philosophies of science. Lakatos argued that his MSRP produces better rational reconstructions than either Popper's falsificationism (which cannot account for the rational tenacity of scientists in the face of anomalies) or Kuhn's paradigm theory (which cannot account for the rationality of paradigm change). Whether this claim is vindicated by the historical evidence remains a matter of scholarly debate.

Lakatos's methodology of scientific research programmes remains one of the most important contributions to twentieth-century philosophy of science. Its strengths include:

• A more realistic account of scientific practice than naive falsificationism
• A rational framework for understanding theory persistence and change
• The concept of the positive heuristic, which captures the internal dynamic of research programmes
• A principled way of distinguishing progressive from degenerating programmes
• A sophisticated account of the relationship between philosophy and history of science

Its weaknesses are also well-documented:

• The inability to provide real-time guidance to scientists (the hindsight problem)
• The difficulty of specifying when a programme has been degenerating “long enough” to be abandoned
• The concession that scientists may rationally stick to degenerating programmes, which weakens the evaluative force of the distinction
• The lack of clear criteria for identifying the hard core and protective belt of a programme

Despite these difficulties, Lakatos's work has had a lasting influence on the philosophy, history, and sociology of science. His concepts of hard core, protective belt, and progressive/degenerating problem shifts are widely used by historians and philosophers of science, and his insistence on the interplay between normative philosophy and descriptive history has become a standard methodological commitment in the field.