Lakatos's Methodology of Scientific Research Programmes
Imre Lakatos was once a high ranking minister in the government of Hungary. After the Soviet crackdown in Hungary of 1957, Lakatos found his way to London where he allied himself with his fellow central European refugee, Karl Popper. Lakatos presented his image as a modification and improvement of Popper's basic falisificationist view.
Lakatos presented what many considered to be the Popperian side's response to Kuhn whose defense of dogmatism in normal science was already seen as "heretical" from a falsificationist point of view. Popper had claimed that his account of the history of science in terms of "conjectures and refutations" exhibited the actual historical "rationality" driving the growth of knowledge over time. For Kuhn the decision to switch paradigms is not determined by any such "rationality"; Lakatos explicitly attacked Kuhn as making scientific belief subject to non-rational methods of mass persuasion, as fickle as matters of taste and style. Thus he sets out to build a theory of the rationality of the growth of scientific belief over time which remained true to Popper's falsificationist views but admitted the hisotrical evidence that Kuhn had presented to show that scientists do not abadon theories when confronted by so-called "counterinstances."
Vocabulary Note: The central analytical concept which Lakatos uses, replacing Kuhn's "paradigms," is designated as a "research programme." While the term "research program" is common in science, Lakatos gives this expression a very particular meaning in his philosophical image of the growth of scientific knowledge. For this reason it is customary to use the British spelling "programme" to distinguish Lakatos's expression from its ordinary use.
A research programme is essentially a sequence of theorieswithin a domain of scientific inquiry. Each later, or successor, theory, is held to mark an advance over its predecessor. The move from one theory to its successor within a research programme is called a "problem shift." The question of the rationality of changing one's beliefs in science, or how does scientific knowledge "progress" over time, is thus transformed -in Lakatosian language- into the question of asking "When is a problem shift progressive?"
Problem shifts may be "progressive" in two ways: theoretically or empirically. Theoretically progressive problem shifts are in effect moves to new theories which enable one to predict more than a predecessor theory allowed one to predict. A problem shift is empirically progressive if in addition to predicting new observable evidence, actual observation does indeed confirm this new prediction. In order for a research programme, as a whole, to be progressive, each problem shift must be at least theoretically progressive, and at least occasionally empirically progressive. In other words in a progressive programme, each move from an old theory to a new one must enable us to predict more, and at least sometimes these predictions must be confirmed. If a programme fails to display this characteristic, it is no longer progressive but has become "degenerating." A rational scientist should (note the normativity) stick with a progressive programme but abandon a degenerating programme.
In designing new theories to replace old, the scientist in a research programme adheres to a constellation of beliefs which Lakatos calls a "heuristic". This heuristic includes both positive and negative aspects.
The negative heuristic specifies certain claims of the research programme as not revisable: "tinkering" with these clams is not permitted as long as one adheres to the programme. They thus cardon off a "hard core" which cannot change from one theory to the next. Revising these beliefs is "off limits." [This is Lakatos's analogue to Kuhn's contention that the normal scientist accepts a paradigm "dogmatically."]
The positive heuristic represents a body of beliefs which are allied to the hard core as well as suggestions regarding how these beliefs can be revised. These beliefs can be tinkered with; indeed the life of the research programme essentially consists of learning how to reshape these beliefs in the light of potentially refuting observational evidence so as to protect the "hard core" from being refuted. Thus they form a "protective belt" surrounding the hard core.
As a research programme progresses, scientists will attempt to refute or falisify the then accepted theory, in good falsificationist fashion. This is Lakatos's Popperian heritage. But when refuting evidence is encountered, according to the Lakatosian picture, the scientist will not consider the programme as "refuted." Instead he/she will begin to alter the assumptions of the "protective belt" in ways premitted or suggested by the positive heuristic, such that the "hard core" of the programme can be retained unscathed. As long as such moves enable scientists to predict more new phenomena (i.e. it is theoretically progressive), and at least some of those predictions get confirmed by observation (i.e., it is intermittantly empirically progressive), the programme is progressing and it is rational to pursue it.
However, when modifications to the theory only protect the hard core from refutation, but do not predict new phenomena, and/or none of those new predictions get confirmed by observation, then the programme is degenerating and the rational scientist abandons it.
Unfortunately, Lakatos is forced -by the very historical evidence he seeks to use to illustrate his image of science- to admit that a programme can go through a "bad patch"; i.e., a rather long period in which no empirical progress is made. Historically, looking back on the develoment of a science, it might be easy enough to tell that a programme is beginning to degenerate at a certain point in time. But this is a matter of hindsight and thus of no use to determinig rationality. What about the scientist in the programme itself at that historical moment? How can he/she tell if the programme is truly now beginning to degenerate or if perhaps it is only undergoing a "bad patch"? In order to know whether it's rational to stick with the programme or switch to another different programme, which rejects the old one's "hard core," such a question must be answerable.
But Lakatos's model offers no help in answering this crucial question. On the Lakatosian picture the individual scientist has to make a "subjective" judgment call on whether his programme is merely experiencing a temporary lull in its progressive shifts or if it has truly begun to degenerate. Thus such a scientist's "theory choice" is no more dictated by a scientific "rationality" than the decision to change paradigms is for the scientist on the Kuhnian image of science. But this was the very problem which Lakatos regarded as making Kuhn's analysis unacceptable and which he set out to remedy. For this reason, many philosophers have concluded that while Lakatos's image of science does indeed have much to teach us, as a model for scientific rationality -designed to avert the "rationality crisis"- Lakatos's methodology of scientific research programmes has failed.
Unfortuantely Lakatos himself did not live long enough to develop his view further and/or answer these criticisms.