jueves, 29 de marzo de 2007

Methodology of Scientific Research Programs

Imre Lakatos, 1978

An adequately rational reconstruction of science has to evaluate as scientific not a single theory but instead of it a group of interrelated theories; Lakatos called them scientific research programs, which are composed of a conventionally accepted ‘hard core’ of interacting theories and a ‘belt’ of protective auxiliary hypotheses, and proposed the sophisticated falsationism as a criterion to select between rival programs. I agree with that methodology and I am going to present some reasons to support this position.

First at all, the statements which are inconsistent with the program can not defeat it. The program has a heuristic that allows us to save the program: we modify the belt without changing the hardcore; in this way, we can always reconcile the theories and the factual propositions. With sufficient resourcefulness and some luck, we can defend any program for a long time, even if it is false. Mere ‘falsifications’ must not imply rejection; they are to be considered –to convert them in corroborating examples- but need not be acted upon. The pattern of trial –by hypothesis- and error –shown by experiment-, is to be abandoned because no experiment is crucial to defeat a program.

Due to the fact that a program can not be defeated by the anomalies, the only way to decide when to abandon a program is the construction of a better one, which means that it has an excess empirical content over its predecessors, some of which is subsequently verified. As a consequence of this we have rival programs. If a program explains more than the rival, the latter will be falsified; therefore, the falsationism –a sophisticated one- requires competing programs. The competition is significant and leads to a rational reconstruction of scientific change in which each rival try to increase its content and predict novel facts. On the other hand, no advantage for one side can ever be regarded as absolutely conclusive; a novel belt of auxiliary hypotheses can strengthen a program. Consequently, the heuristic allows us to continue working in a program, as a positive outcome of that we won’t reject a program if the anomalies take place, giving us the opportunity to look at the potential of it.

Even though we can modify the belt of auxiliary hypotheses, there is a difference between the scientific modifications and the pseudoscientific ones that we make. A research program is said to be progressing –and therefore scientific- as long as its theoretical growth anticipates its empirical growth, that is, as long as it keeps predicting novel facts with some success. Conversely, a research program is said to be degenerating –and consequently, pseudoscientific- if it gives only a reinterpretation –a linguistic one- of facts anticipated, and discovered in a rival program. According to this methodology the greatest scientific achievement are research programs which can be evaluated in terms of progressive and degenerating problem shifts.

Another fundamental point is that the methodology of scientific research programs has a historical character. If we wan to follow this methodology and develop a scientific program we will look in history for rival research programs and if instead of it we want to explain different speeds of development of different research program we may need to invoke external history. Moreover, the scientific revolutions along the history are shifts of one program over another.

We need a lot of time to evaluate a research program, but this evaluation is stricter because it requires that the belt of auxiliary hypotheses –following a positive heuristic- turn into a progressive shift and predicts novel facts. The methodology of research programs, finally, emphasizes long-extended theoretical and empirical competition of major research programs, progressive and degenerating problem-shifts, and the slowly emerging victory of one program over the other.