The question of whether politics can in fact be studied like the natural sciences is bewildering; it involves questions of free will and determinism, the nature of human action as distinct from animal behavior, the limits to predictability in the field of human affairs, and the limits to the possible understanding of events and institutions remote in time or belonging to a radically alien culture. Dahl is admirably cautious in his claims for political science as, literally speaking, a science. But he tends toward the belief that what is not science is necessarily deficient. To approach politics as a science requires more than objectivity and a scrupulous regard for evidence; it also involves the search for mathematically exact knowledge culminating in the ability to predict political behavior on a large scale. Dahl is aware of the dangers of such a search, and of intruding quantification where it has no place, but the ideal set by natural science has a strong hold on him and immediately leads him to certain errors of emphasis. For example, he devotes a whole chapter to the concepts of power and influence in order to make more exact our use of these words and thus equip us for accurate comparisons between the relative power or influence of various men. If we can do this we are being scientific: we are using our terms precisely and are enjoying an approximation of quantified knowledge. But when political concepts—like power or freedom or the common good—are troublesome, the goal of explication must be philosophical correctness rather than making such concepts serviceable instruments for measurement. Explication will then proceed by sorting out the many uses of the concept and by answering—perhaps tentatively—a host of kindred questions in which the concept figures.
–George Kateb, “Politics for What?” (1963), New York Review of Books, a review of Modern Political Analysis by Robert A. Dahl