The statistical analyses that dominate the biological, behavioral, and social sciences tend to test the null hypothesis that the empirical results — whether they be mean differences or correlation coefficients — can be attributed to mere chance. By chance alone, it is possible to obtain sample means that look different, even when in the larger population the means are strictly identical. By mere chance, one can obtain correlation coefficients that appear much larger than zero, even when the correlations in the general population are exactly zero. In an analogous fashion, the theoretical interpretations provided in the previous chapter can be said to represent the null hypothesis that chance can account for all the central empirical features of two key behavioral phenomena — scientific productivity and multiple discovery. In this case, however, chance is not sampling error but rather some unspecified combinatorial processes operating within and among creative scientists. Beginning with the postulate that scientific creativity involves the virtually random combination of the phenomena, facts, concepts, variables, constants, techniques, theories, laws, questions, goals, and criteria that define a domain, a host of detailed features of scientific careers and communities can be explained and predicted. There is no need to hypothesize anything more mysterious, whether genius or zeitgeist. Even logic has no part in the explanation. According to Ockham's razor — the law of parsimony — scientists should not make explanations more complicated than required to fit the facts. By this standard, logic, genius, and zeitgeist appear mostly superfluous.