Debrecen, 1922–1974

Imre Lakatos

Popper's most brilliant colleague tried to save falsificationism from the history of science. Scientists do not drop a theory at the first failed prediction, Lakatos said, and they are right not to. What matters is where the whole programme is heading.

The view

Lakatos was born in Hungary, survived the war, was imprisoned under the communist government, and left after the rising of 1956. He wrote a doctorate at Cambridge and joined Popper's department at the London School of Economics in 1960. He died suddenly at fifty-one.

His first major work, Proofs and Refutations, is a dialogue in which a class argues over a theorem about polyhedra. It shows mathematics growing as science does, by conjectures, counterexamples and repaired proofs.

His later work answered a challenge. Kuhn had shown that scientists routinely keep theories that face known counterexamples. Newton's theory had anomalies from the day it was published. If a single refutation obliged scientists to give a theory up, as a naive reading of Popper suggests, there would be no theories left.

The argument

Lakatos's solution was to change the thing being judged. The unit of appraisal is not a single theory but a research programme, a series of theories with a common structure. At its centre is a hard core of basic assumptions that the programme's followers agree not to give up: for Newtonians, the three laws of motion and the law of gravitation. Around it is a protective belt of auxiliary hypotheses about instruments, initial conditions and the make-up of the system studied. When a prediction fails, the belt is adjusted and the core is left alone.

A programme is progressive while its changes predict novel facts and some of the predictions come true. It is degenerating when its changes only explain away the facts that embarrass it. After Lakatos, "Falsification and the Methodology of Scientific Research Programmes" (1970)

The difference shows in two episodes from one programme. In the 1840s Uranus was straying from its Newtonian orbit. Le Verrier and Adams did not abandon Newton. They adjusted the belt by proposing an unseen planet, and calculated where it must be. Neptune was found there in 1846: a novel fact, and a triumph. Some years later Le Verrier tried the same move for an anomaly in Mercury's orbit and proposed a planet inside it, named Vulcan. No such planet was ever found. Mercury had to wait for Einstein.

A consequence Lakatos accepted is that there is no instant rationality. One cannot tell at the moment whether an adjustment is the start of a recovery or of a decline. Programmes may be nursed through bad patches, and the verdict comes with hindsight.

Where he meets Deutsch

Lakatos is a fallibilist of Popper's school. Knowledge is conjectural; theories are judged against rivals and never in isolation; and the worth of a theory lies in what it adds, which he measured by new predictions. Deutsch shares all of this, along with the stress on ad hoc rescue as the mark of decline. A degenerating programme is, in Deutsch's terms, an explanation that is being varied at will to keep up with the facts.

Proofs and Refutations is closer still. Deutsch also holds that mathematical knowledge is fallible and that proof is a physical process which we understand through our best theories of physics.

Where they part

The hard core is protected by decision. Followers of a programme agree to direct every refutation at the belt. For Deutsch nothing is exempt by agreement, and each explanation has to earn its place on its merits.

The larger difference is over timing. Lakatos can call an adjustment good or bad only after waiting to see whether its predictions succeed. Feyerabend made the obvious complaint: a rule that passes judgement only in hindsight gives no advice. Deutsch's criterion can be applied at once. Neptune was a good explanation before anyone looked through a telescope: a planet of a definite mass on a definite orbit, each detail fixed by the anomaly it had to account for. Vulcan had to be reshaped with each failure to observe it, and every reshaping made it easier to vary.

Lakatos also wanted more than Popper would grant. He asked for some assurance that corroborated theories are likelier to be true, and called it a whiff of inductivism. Deutsch grants no such whiff.

See every view compared →

Further reading

• Imre Lakatos, The Methodology of Scientific Research Programmes: Philosophical Papers, Volume 1 (Cambridge, 1978).
• Imre Lakatos, Proofs and Refutations: The Logic of Mathematical Discovery (Cambridge, 1976).
• Imre Lakatos and Alan Musgrave (eds), Criticism and the Growth of Knowledge (Cambridge, 1970).
• Imre Lakatos, Stanford Encyclopedia of Philosophy.