Two methods in search of science

Conclusion

In terms of the criteria for the growth of knowledge formulated by Popper, I have tried to demonstrate the superiority of the methodology of research program over the methodology of induction. Although the argument used Skocpol's and Trotsky's theories of revolution as illustrations, I constructed general claims organized around the contexts of discovery (induction versus deduction), justification (verification versus falsification and prediction), and scientist (external to or part of the object of knowledge). So long as philosophers of science were concerned to discover the scientific method, they could successfully compartmentalize these contexts. However, as soon as they became concerned to explain the development of scientific knowledge, they quickly discovered, as we have, that these contexts are irretrievably intertwined. So we require alternative categories for comparing methodologies.

(a) Grounds of scientific objectivity

I have tried to demonstrate that the method of induction stands on a false objectivity. While it claims to generate explanations that map the empirical world, it actually erects barriers to the comprehension of that world. Not “the facts” but methodological premises and arbitrary explanatory hunches become the hidden anchors for theoretical conclusions. The method is at odds with its aims. Paradoxically, the methodology of the research program, precisely because it is self-consciously anchored in a complex of moral values, a conceptual system, models (analogies and metaphors) and exemplars - what Skocpol refers to as “blinders or heavily tinted lenses,” what Lakatos refers to as negative and positive heuristics - creates a more effective dialogue with those “historical patterns.” Blindness comes not from pre-existing theories but from failing to recognize their necessity and then failing to articulate and defend their content.

(b) Problem versus puzzle oriented science

The method of induction claims to be outside and beyond theoretical traditions. Thus Skocpol reduces the classics of Marx, Weber, and Durkheim to inspirations, sources of hypotheses, and even to variables out of which a true macro sociology can be forged. “Compelling desires to answer historically grounded questions, not classical theoretical paradigms, are the driving force [of historical sociology].”¹ We select a problem that takes our fancy and induce its solutions from the facts. Since, in the final analysis there is only one theory compatible with the facts, there is no need to go through the falsification of alternative theories or put one's own theory through severe tests. The methodology of research programs, on the other hand, is concerned to solve puzzles, that is, anomalies thrown up by its expanding belt of theories, discrepancies between expectations and “facts.”² The health and vitality of a research program depends not on the concealment, obfuscation, denial of anomalies but on their clear articulation and disciplined proliferation. Continual dialogue between theory and data through falsification of the old and the development of new hypotheses with predictions of novel facts is of the essence of a progressive research program. Trotsky's prophetic powers all originate in, even if they are not determined by his commitment to Marxism - a recognition of its anomalies and the need to solve them in an original manner.

(c) Internal versus external history

The method of induction regards the facts as irreducible and given, the problem is to come to an unbiased assessment of them. Science grows by the accumulation of factual propositions and inductive generalizations. This is its internal history. “But the inductivist cannot offer a rational ‘internal’ explanation for why certain facts than others were selected in the first place.”³ Problem choice, as we said above, is part of the “external” history relegated to footnotes, prefaces, or to the “sociology of knowledge.” By contrast, the methodology of research programs incorporates into its internal history what is branded as metaphysical and external by inductivists, namely its hard core postulates and its choice of puzzles. What is reconstructed as scientifically rational in the one appears as scientifically irrational in the other.

Although what is constituted as rational in research programs encompasses much more than the rationality of induction, nevertheless even here external forces necessarily influence the scientific process. This is particularly so in the social sciences where the object of knowledge autonomously generates new anomalies that the positive heuristic has to absorb. External forces can be seized upon as opportunities for the rational growth of knowledge, but they can also be the source of irrationality. Thus, research programs become degenerate when they seal themselves off from the world they study or when that world wrenches the research process from its hard core. Marxism is particularly sensitive to external history. Where it seeks to change the world it is more likely to be sensitive to anomalies than where it is a dominant ideology and thus more vulnerable to the repression of anomalies.

Obviously the methodology of research programs has its own distinctive problems that energize its development. Is it possible to identify a single core to a research program or are there a family of cores and how does the core change over time? What is the relation between positive and negative heuristics? How easy is it to distinguish between progressive and degenerating research programs? How do we know that an apparently degenerating program will not recover its old dynamism? How does one evaluate the relative importance of progressive and degenerating branches of the same program? Is it possible to stipulate the conditions under which it is rational to abandon one research program in favor of another? Such probloems notwithstanding I hope I have made a case for the superiority of the methodology of research programs over the methodology of induction as a mode of advancing social science.