It has been the tendency of science to confine the domain of deductive methods more and more, and not to be content with arguments from analogy, which are the foundation of most errors of the human mind, and to which may be traced the religious and other ideas of man in a primitive state of culture, and, to a certain degree, even in a state of advanced civilization. Science is constantly encroaching upon the domain of the argument from analogy, and demands inductive methods.

Nevertheless the psychological and scientific value of the argument from analogy cannot be overrated: it is the most effective method of finding problems. The active part it plays in the origin of philosophical systems and grand ideas which sometimes burst upon scientists is proof of this. But, as far as inductive methods can be applied, - and we believe that their domain will continue to increase, - induction must scrutinize the ideas found by deduction.

Boas, Franz. 1887 [1989]. The principles of ethnological classification. In G. Stocking. Ed. A Franz Boas Reader: The Shaping of American Anthropology, 1883-1911. Chicago: University of Chicago Press, p.64-65.