It is widely supposed that the scientists in any field use identical standards for evaluating theories. Without such unity of standards, consensus about scientific theories is supposedly unintelligible. However, the hypothesis of uniform standards can explain neither scientific disagreement nor scientific innovation. This paper seeks to show how the presumption of divergent standards (when linked to a hypothesis of dominance) can explain agreement, disagreement and innovation. By way of illustrating how a rational community with divergent standards can encourage innovation and (...) eventually reach consensus, recent developments in geophysics are discussed at some length. (shrink)

Although there are numerous and significant differences between the theories of scientific growth and change proposed by Kuhn, Lakatos, and Laudan, they all hold that specific scientific theories should be viewed as constitutive of more comprehensive theories. Kuhn calls those more general theories ‘paradigms’, Lakatos labels them ‘research programmes’ and Laudan refers to them as ‘research traditions’. They all argue that scientists are much more willing to give up the specific theory within a given research programme rather than the programme (...) itself, and that individual theories should be viewed as attempts to increase the overall explanatory power of the more general theories, since the ultimate concern of the scientist is with the success of the general rather than the specific theory. When a basic theory or research programme is confronted with severe criticism, proponents attempt to protect the hard core or central elements of their programme through the invention of auxiliary hypotheses. Good auxiliary hypotheses adequately answer the objections for which they are designed, and suggest new avenues of research. In 1928, Arthur Holmes provided proponents of continental drift theory with an auxiliary hypothesis which afforded them a badly needed account of the forces responsible for continental drift. Although Holmes' proposal was not ultimately correct, it was the first plausible alternative offered by an exponent of the continental drift research programme. (shrink)