Monday, February 23, 2015
Imre Lakatos and Scientific Research Programmes
"Einstein's results again turned the tables and now very few philosophers or scientists still think that scientific knowledge is, or can be, proven knowledge."
Imre Lakatos (1922 – 1974) was a Jewish-Hungarian philosopher of science and mathematics, and a long-time colleague of Karl Popper at the London School of Economics. For most of the war Hungary was ruled by the nationalist Horthy regime, which cooperated militarily with the Nazis but refused to cooperate in the Holocaust. When the Horthy regime was forced from power in 1944, Hungary’s Jews were exposed. Imre Lipschitz changed his last name to Molnar and joined a unit of Marxist guerillas; his mother and grandmother were murdered at Auschwitz. After the war he changed his last name again (to Lakatos) and accepted a university position in the Soviet-backed Rakosi regime. He was jailed for several years, but released on Stalin’s death in 1953. He participated in the 1956 uprising, and had to flee the country when it was suppressed by the Red Army. He finally settled in Great Britain, where he earned his doctorate and spent the rest of his career.
Like Feyerabend, he worked to reconcile traditional, proscriptive approach (Popperian philosophy of science) with the comparatively recent descriptive approach (Kuhnian history of science.) Unlike Feyerabend, he thought an orderly process of discovery survived the merger – namely, the Methodology of Scientific Research Programs. According to Lakatos, these programs are composed of a “hard core” of assumptions which are not normally open to questioning, and a “protective belt” of claims that can be modified as needed in order to suit new and potentially problematic observations. When observations do not match predictions, the usual assumption is that the fault lies in the peripheral and not the core elements – however, if enough of these mismatches pile up, eventually the core itself could be threatened, and the Research Program jeopardized.
Unlike Kuhn, who held that a single Paradigm dominates all science at once, Lakatos argued that multiple Programs compete within or across fields simultaneously. The value of an individual Program might then be assessed by comparing the results it generates with those of its competitors. A progressive Program modifies its protective belt in a way that incorporates new information and generates new predictions; a stagnant Program incorporates new information but does not generate new predictions. The validity of a Program is a matter of its short term history of (and thus prospects for) generating useful information. However, a stagnating program is not necessarily a dead one – just as progressive theories can break down, worn out theories can be revitalized. Put another way, it is the dynamics, rather than the statics, of a program that determine its validity.
To take an example from the history of science, it was known in the 19th century that light had certain wave-like properties. Because a wave is a disturbance in a medium, and light clearly traveled through space, it was reasonable to suppose that space contained such a medium, which the physicists called “lumineferous ether.” Lakatos would say that the protective belt of the theory were modified in order to reconcile the hard core assumptions with observations. However, as problems began to pile up for the Newtonian view of particle physics, and repeated experiments failed to discover any trace of the supposed ether, the core itself was jeopardized. This (among other things) led to what Kuhn would describe as a period of “revolutionary science” in the first decades of the 20th century, in which the new programs of Relativity and Quantum Mechanics displaced the old Newtonian program.
In the 1930’s, however, physicists was realized that the predicted behavior of galaxies under Einsten’s model of gravity did not square with observations. Reasoning that the observations (not the theory) were in some way faulty, physicists decided that a significant portion of these galaxies must be composed of a non-luminous (and hence invisible, or “dark”) matter. In the 1990’s physicists also realized that galaxies were not slowing down (as one would expect if gravity is acting as a drag on their momentum) but speeding up. A new (and also invisible, or “dark”) form of energy was hypothesized in order to account for this. In both cases, the hard core of the program was preserved by modifying the protective belt. According to Lakatos, there is nothing wrong with this, because the validity of a program is a question of its momentum, not its internal consistency. However, because the confidence which scientists place in a theory is ultimately dependent on the success of its predictions, dark matter and dark energy cannot stay invisible forever. Eventually they will either be observed, or Relativity will be discarded in favor of some new explanatory model. In other words, the protective belt can only be arranged so many times. The intense interest of physicists in the questions of dark energy and dark matter are propelled by this sense of suspense, for it is clearly a question of fundamental importance for their work, whether Relativity will be confirmed by the discovery of dark matter and dark energy, replaced by some new Program, or perhaps challenged by a revived Newtonian Program (MOND.)
Imre Lakatos died unexpectedly in 1974, of a brain hemorrhage. His papers, which had previously been scattered throughout various scholarly journals, were published posthumously in 1978, under the title Philosophical Papers. His discussion of Research Programs appears in the first volume. Perhaps also of interest is his philosophy of mathematics, published in Proofs and Refutations (1976), which held that mathematical proofs are fallible(!), and advocated an experiment-driven approach to their verification.
Part of a series on Science, Technology, and Society (X of XX)