Scientific experimentation is not possible in politics, because we cannot use a society purely as a subject of demonstration. Moreover, we are not given the opportunity of studying the in fluenee of any one cause by carefully excluding the operation of all others, as we may do in physical experiments. While. however, the sci entific experiment pure and simple is by the na ture of things excluded, practical experiments are constantly being made in polities, i.e. courses of action are tried without a knowledge of the exact results to he produced by them, but in the hope that certain improvements will be effected ; as, for instance, in changes of the eriminal la w, or in the methods of nominating and electing pub lic officials. Such action, although not under taken purely as au experiment, is experimental in the sense that the nature of its consequences can be definitely ascertained only through experi ence, and that the action is undertaken with a knowledge of the possibility of varying results. Therefore, while politics as a science cannot set aside certain social factors for experimental pur poses, as an art it is constantly making experi ments in practice which in turn furnish the most valuable kind of material to scientific political study.
The importance of the analytical method has already been indicated above. As in the inves tigation of political and legal institutions the juristic bias is naturally very strong, there is constant danger in political science of resting sat isfied with mere analysis of structural forms. A scientific method, while it must make use of the trenchant instrument of analysis, must also extend its range so as to embrace the real motive forces that lie back of institutional forms. The natural law theory has always favored a purply juristic interpretation of institution:. and the scientific study of polities has been much re tarded by the one-sided use of analytical methods which lead to disputes about fine-drawn hut un important distinctions, while apparent definiteness of the results obtained often causes men to over look the powerful forces operating in political life.
The deductive method is thus explained by Pro fessor Sidgwiek in his Elements of Polities: "We assume certain general characteristics of social ninn, ancl we consider what and institutions arc likely to conduce most to the welfare of an aggregate of such beings." While this method is perfectly legitimate and has been fruitfully employed by most of the great political writers, it would be a mistake to suppose that it is the sole or even the principal reliance of sci entific polities. In fact, if employed alone and unassisted by the study of concrete facts, it opens the door to hasty generalization and to the profit less expounding of threadbare theories. In itself it is sterile. It simply analyzes, develops, and applies certain facts assumed to be axiomatic. It takes on the subjective coloring of the writer's mind and it does not in itself furnish a stable basis for the scientific pursuit of political investi gation. Polities is made a science through the use of inductive processes. The careful study of political experience and of the ever-varying forces of political life provides the rich harvest of ob servation upon which general principles and maxims may be founded with a certain amount of warranted assurance; and though it may rot as yet provide a large number of exact generaliza tions and rules for practice, it still is recognized as leading to a fuller mastery than does deductive reasoning, by giving the mind an insight into the wealth of incident and the dramatic action of politics. The deductive method is of value in the matter of applying knowledge to political action, when niter a careful study of political forces and institutions their general nature and operation is understood. Such knowledge may then well be applied by a deductive process to concrete politi enl action.