Lakatos’ goal is worthwhile: steering the evolution of hypotheses toward those that have greater explanatory power. His method is feasible, if a bit awkward. Usually the hypothesis revision occurs after a project has obtained its research results, so the actual test of the new prediction is deferred for a later paper by the same or different authors. Lakatos’ criterion is virtually unknown and unused among scientists, however. Its problem is the same as that of falsificationism: it is an outside judgment of what scientists should do (according to the proponent), rather than a description of what they actually do, and we scientists are not persuaded of the need to change. We can, however, be alert for ad hoc hypotheses, and we do expect a modified hypothesis to explain more than its predecessor.
I and many other scientists are close to this conventionalist view. We are, perhaps, even closer to Thomas Kuhn’s perspective, described in the next section. ***
Paradigm and Scientific Revolution
Thomas Kuhn’s 1963 (and 1970) book The Structure of Scientific Revolutions overthrew our perception of scientific change. We had imagined scientific change as a gradual process, involving incremental advancement in techniques, evidence, and hypotheses, which resulted in a steady increase in scientific knowledge.
Our textbooks reinforced this view by portraying the history of scientific thought from our present perspective. Early ideas are judged to be important and relevant only to the extent that they contribute to the continuous evolution toward the current ideas. Textbooks express the outcomes of scientific revolutions as discoveries of new ideas; they avoid confusing this picture with discussion of the process of scientific upheavals and of the ideas that have been superseded. Because most science students read textbooks rather than scientific articles prior to initiating their own graduate research, their perception of scientific change is fossilized even before they have a chance to contribute to that change.
Kuhn said that we must consider scientific results in the context of the sociological factors and scientific perspectives of their time. He saw the advance of science more as a staircase than a ramp. Within each scientific field, long periods of stability and consolidation are followed by short periods of major conceptual revision, or paradigm change. I think that this view of science is progressive: not only is it a more realistic perspective, but also it offers insights into which scientific methods are most appropriate at different points in the evolution of a science.
A paradigm is a suite of “universally recognized scientific achievements that for a time provide model problems and solutions to a community of practitioners” [Kuhn, 1970]. Kuhn realized that this definition is vague and sloppy. To me, a paradigm is a coherent suite of theories or concepts that guide interpretations, choice of relevant experiments, and development of additional theories in a field or discipline. Physics paradigms, for example, included Newtonian dynamics, general relativity, and quantum mechanics. We can understand paradigms better by considering a field in its pre-paradigm state. Data collection is unfocused, a fishing expedition rather than a hunter’s selection of prey. Facts are plentiful, but the overall patterns and organizing principles are unclear. Several schools of thought compete, none agreeing on what phenomena warrant study and none providing broad-scope hypotheses. Research is overwhelmed by the apparent complexity of the subject.
***
