Logical Method in the Social Sciences

Abstract

In the view of a more or less characteristic attitude toward “arm-chair stuff,” it is fair to say that this paper is frankly of that variety, it does not present the results, tabulated or otherwise, of any inductive investigation, under controlled conditions and no apology will be made to anybody on that score. Nor will much time be taken in arguing the worth-whileness of abstract thinking on the question of methods, or in settling the place of deduction in relation to induction. One way to study the problem of methods would be to try all possible methods on all possible problems and see which gave useful or satisfying results. In the same way, one way to study any science would be to try all possible experiments and see which ones succeed. (mushrooms and toadstools). I shall simply assume without much discussion that while the only way to settle any question in case of persistent uncertainty of disagreement is to try it, on the other hand two things are true; in the first place much labor can be saved by getting as clear an understanding of any problem as possible in advance of inductive study that blind trial and error methods are wasteful; and moreover, inductive investigation itself rarely proves anything unless it is intelligently planned and carried out with a view to proving or disproving some hypothesis which has seemed in the light of careful previous analysis to be worth testing. And of course I assume that these principles hold in relation to the study of methods as well as in other lines of inquiry. It is true that much a priori discussion is inconclusive, misdirected, and futile; it is often true that it is needlessly prolonged, carried beyond the point necessary to define the issues and serve as a guide in an investigation of facts. It is even true, though very exceptionally, that in some cases it is entirely helpless and the only thing to do is to begin experimentation blindly or wait for some accidental coincidence to suggest a solution of a problem. For example, the discovery of effective catalyzers in chemistry, or perhaps the stains to differentiate a suspected bacterium or the treatment for a new disease. But is also true that an enormous proportion of inductive scientific study is misdirected and futile because the problem was not adequately analyzed to begin with. In the field of social science, with which I am fairly familiar it is triturating the obvious to say that the greater part of the statistical studies either prove only what everybody already knew and which called for no proof or else seem to prove something that everyone knows is false. And when the result of statistics conflicts with the convictions of common sense, which is really well-informed judgment, it is a trite observation that it is usually the statistics which are wrong (Pearson). The man in the street has only too much justification for the common belief that anything can be proved by statistics, and for the well-known division of liars into the three grades: plain liars, damned liars, and statisticians.

Citation

Emmett, R.B. (2011), "Logical Method in the Social Sciences", Emmett, R.B. (Ed.) Frank H. Knight in Iowa City, 1919–1928 (Research in the History of Economic Thought and Methodology, Vol. 29 Part 2), Emerald Group Publishing Limited, Leeds, pp. 103-121. https://doi.org/10.1108/S0743-4154(2011)000029B013