The application of the genetic method is directed to the dis covery of two things, namely : (1 ) the main stages through which the evolution of the facts under investigation has probably passed, and (2) the causes which produced the various changes constitut ing the several phases in the suggested course of evolution. In dealing with the first of these problems, namely the discovery of the main stages or gradations of an evolutionary series, the cor rect procedure, according to Darwin, is to seek them, in the first instance, among kindred classes of objects. But it is rarely possible to secure sufficient evidence from a study of the nearest kindred only. The investigator is, consequently, often compelled to seek farther afield for missing links in the probable chain of evolution. Thus, e.g., Darwin himself, when investigating the evolution of the honeycomb of the hive-bee, kept closely to kindred species but when he traced the evolution of the eye of verte brates he went far afield, and invoked the help of facetted eyes, eyes without a lens, and eyes that are mere aggregates of pigment cells. Similarly, the comparative psychologist, when tracing the evolution of the human mind, does not confine his attention to the higher apes (the primates), but seeks light also from the lower animals, going even to the ant to learn of its ways. The second of the aforementioned tasks of the genetic method, namely, the de termination of the causes which may have produced the various changes constituting the series of evolutionary stages, is mostly very difficult. A purely mechanistic explanation seems out of the question. In opposition to it there has been put forward the view of emergence (q.v.) or of creative evolution (q.v.). These views may be a healthy reaction against mechanistic pretensions but they run the risk of confusing the christening of a problem with the solution of it.
The term "comparative method" is sometimes used as the equivalent of the evolutionary or genetic method just explained; but sometimes, perhaps more frequently, it is used in the rather ambiguous and indiscriminate sense of "the method of the com parative sciences," such as comparative philology, comparative anatomy, sociology, etc. These sciences, however, employ a variety of methods sometimes including the evolutionary method, and sometimes not. All they seem to have in common is a process of comparison. But, as already explained, comparison is no distinctive method, but rather an activity common to every method and want of method.
Methods for Discovering Causal Connections.—The method of classification and the evolutionary method satisfy to some extent the human quest for order, which is prompted mainly by the need of anticipation as a means to adaptation. Successful classification means a knowledge of the more or less uniform coex istence of certain attributes in the things concerned, so that it is possible to infer the presence of certain unobserved class-charac teristics from certain others which are more accessible to observa tion—say the possession of a ruminant stomach in a quadruped having hoofs and frontal horns. The successful application of the genetic method means the possibility of inferring later stages from the present stage of an evolutionary series. But such uniformities are not the most reliable or most important for science or for life. Another and more important kind of uniform connection, perhaps the most important kind of uniform connection discoverable, is the causal connection between things or events. The general char acters of causal connection is discussed in the article CAUSALITY, and an idea of it will be assumed here. To discover causal con
nection is to discover uniform connections between certain physi cal antecedents and certain consequents. Now, natural events do not disclose their causal connections to mere inspection. These connections can only be inferred or guessed from the observation of certain kinds of sequences or concurrences. The general char acter of the kind of observations which suggest causal connections, or confirm such suggestions, have been formulated in certain canons of induction, which are commonly known as the five simpler inductive methods. But there are also certain other methods of tracing causal connections, notably the statistical method, and certain combinations of the simpler inductive methods and deductive reasoning which may be called collectively the deductive-inductive method. We shall now proceed to deal with these methods in turn.
These are based on our assumption that things and events are not a mere matter of chance, but are, in the main at least, the results determined by certain operative conditions, and occur only when these conditions operate. If, then, while other things remain essentially the same, a certain factor or circumstance cannot be omitted, or changed in quantity, without affecting the phenomenon under investigation, it may be concluded that the factor or cir cumstance in question is an indispensable condition of that phenomenon, that is to say, is causally connected with it. On the other hand, any circumstance the absence, or quantitative varia tion, of which appears to make no difference to the phenomenon under investigation may be regarded as having no causal con nection with it. It is necessary, therefore, to observe instances of the phenomenon concerned under circumstances as varied as possible, in order to find out from such observations what can and what cannot be removed or altered without affecting it. The pro cedure may assume one of two principal forms—a direct and an indirect form. In the direct form it is shown by observation or experiment : (I) that the elimination of a certain factor or ante cedent is followed by the non-appearance of the phenomenon con cerned, or (2) that a quantitative change in a certain factor or antecedent is followed by a quantitative change in that phe nomenon, although, in either case, all other relevant factors re mained the same. The first of these direct forms (I) is known as the Method of Difference, the second (2) is known as the Method of Concomitant Variations. The indirect form of this procedure consists in showing that so long as a certain antecedent or factor is present, a certain result seems to follow, and is not materially affected by changes in any of the other factors or cir cumstances that might appear to be relevant ; and from this it is concluded that the phenomenon is not causally connected with the variable antecedents, but, since it must be causally connected with something, that it is so connected with the factor or antecedent that was present in all the observations made. This indirect method is known as the Method of Agreement. Of the two other methods belonging to this group, one, known as the Method of Residues, is only a modification of the Method of Difference, and the other known as the Joint Method of Agreement and Difference, is really the Method of Agreement supplemented by the study of certain negative instances so as to approximate to some extent to the Method of Difference.