gathered, not only from the conflicts and rapid changes of theory in this field, but also from the following considerations: 1. Almost every fact of the savage's mentality has to be interpreted. He can directly tell us about it far less than can be directly told by question-list respondents concerning their inner life. Further, the interpreter is generally not the fact-gatherer. The psychologist is obliged to make the best he can of data gathered by other persons who may not have had in mind his own questions, interests, or points of view. 2. Here, as in other evolutionary studies, a Janusfaced difficulty is almost unavoidable. At our end of the evolution stand highly complex processes and a set of preferences, ideals, or valuations that we call higher as distinguished from lower. At the other end of the scale we seek for minimal complexity, and for lower, instinctive preferences, both process and preferences being such as to connect man with the animal series. Evolution implies that we are to think these extremes as somehow one; the complex must be seen to come out of the simple, the high values out of the low. Consequently we tend to overlook contrasts. We do it, on the one hand, by oversimplifying our own culture, or by forced classification of the higher with the lower; or, on the other hand, by attributing too much to particular phenomena of the savage mind. By combining the two-oversimplification on the one hand and overloading on the other—almost every item of savage religion has been made to carry the religious universe on its back. The gods are ghosts of dead men; the gods are naturepowers that smite the attention; the gods are imagina tive projections of some social unity. The psychical root of religion is fear; the root is the experience of sex; the root is the economic interest of some group. Totemism, magic, sacrifice, myth-each has loomed overwhelmingly. The movement from myth to theology, from spell to prayer, from the festivals of fertility-gods to Easter, from mystery-initiations to baptism, from totemistic eating of the god to the eucharist, from taboo to Sunday and other sanctities, from the instinctive sense of tribal solidarity to the ideal of a kingdom of God, from spooks to hope of heaven-this movement, this mass of evolutionary ties between us and our ancestors easily creates an impression that our religion is a vestigial phenomenon, a remainder from savage crudity, whereas religion has evolved away from as well as out of savagery. The nature of this evolution will be gone into at some length in chap. xiii. Meanwhile, we may well remind ourselves that analysis establishes differences as well as similarities, and that the differences between two things are neither wiped out nor explained by placing them in the same evolutionary series. Finally, we have no right to assume that origins are in the past alone. Evolution may be X original at every step, and may be going on with originality in our own experience. We ourselves may conceivably live within the sources; they may even be pouring themselves forth with greater freedom than in the case of early man. For these reasons one may hesitate to follow Wundt in his conviction that an anthropological-genetic account of religion is the psychology of religion. In the fourth place, data may be ascertained by experimental methods. An indirect experimental contribution is made, of course, by laboratory studies of part processes that enter into religious reactions, such as suggestion, emotion, and belief-formation. Theoretically, integral religious experiences, with sincere and complete letting go, might be evoked under laboratory control. But the difficulties are obvious. The whole laboratory spirit of aloofness from all interests except analysis stands in the way. A case is on record of an experiment in which the subjects read or listened to religious sentiments under controlled conditions, and then wrote introspective records of the results. The upshot of the experiment concerns rather the psychology of language and of edifying discourse in general than analysis of religious experience. A very different procedure appears in another report—a study of the elements of certain services of common worship. By employing the order-of-merit method, which has already been referred to in the Appendix to chap. ii, it was found possible to determine the relative values of these elements for a group of fifty persons. The order-ofmerit method consists fundamentally in this: The same set of items is placed by each one of a large number of persons in the order, "first," "second," "third," etc., on the basis simply of more and less without regard to how much. Thus, colors or pictures may be arranged in an order of preference, or stories may be arranged on the more specific basis of "more or less humorous." By 1 W. Stählin, "Experimentelle Untersuchung über Sprachpsychologie und Religionspsychologie," Archiv für Religionspsychologie, I (1914), II7-94. * See a summarized report of a paper read by Mark A. May before the New York Branch of the American Psychological Association in Journal of Philosophy, XII (1915), 691. combining these judgments, and by the use of certain statistical methods of analysis, it is possible to construct a scale of merit for the items concerned, and even to establish some definite quantitative relations. There appears to be no reason why this method may not be used to determine value relations in religion for far larger groups than the one here studied. There is a place, also, for the field use, as distinguished from the laboratory use, of experimental methods. In the religious education of children, in the conduct of worship, in the whole plan and organization of a religious society, particular factors can often be identified, sometimes changed at will. Many a roughand-ready experiment in religion has been made in the interest of religion, and various modes of control have thus evolved in religious communions. There is nothing to prevent the deliberate, scientifically controlled refining of such experiments.1 A final caution must be uttered against taking the notion of method too narrowly. If one should ask, for example, how Höffding ascertained the data of the third division of his Philosophy of Religion, an answer could hardly be given offhand. He makes little use of anthropology, or of sacred literatures, or of religious biographies, or of question-list returns; yet his analysis of the religious experience is among the most noteworthy. The reason is that, though he adduces few new data, he sees far into common facts. Now, this far-sight of his is not an accident; it is rather the ripe fruit of 1 A beginning has been made, for example, in the ascertainment of the reactions of children to common worship. See H. Hartshorne, Worship in the Sunday School (New York, 1913). long experience with psychological facts and problems. We may say, then, that in addition to the digging out of fresh material, research may take the direction of fresh analysis of material that is commonplace. APPENDIX GERMAN VIEWS OF THE "AMERIKANISCHE RELIGIONSPSYCHOLOGIE" Two American productions in this field, Starbuck's Psychology of Religion and James's The Varieties of Religious Experience, have been translated into German, both under the influence of interests that center in theology. Around these works an interesting discussion has arisen that concerns the task and the method of the psychology of religion. Faber, Das Wesen der Religionspsychologie, gives abundant references to books and articles on the subject, besides contributing an extended essay of his own. Various German theologians saw in the two works just referred to methods that seemed to open up vast possibilities of increase in scientific knowledge of religion. In addition, they found support for theological or transcendental presuppositions in this branch of psychology, and some of them attempted to fuse psychology and what may be called metaphysical theology. On the other hand, there have come, partly from theologians also, sharp criticisms of the methods of the two American writers just named, and strenuous opposition to theologized psychology. In his Probleme, Part IV, Wundt regards James's book as not psychology but an extract from a pragmatic philosophy of religion. In this German discussion the phrase "the American psychology of religion" has sprung up. It is used, as far as I have discovered, in a very narrow way, and I regard it as misleading. It is used narrowly because it takes works published from 1899 to 1902 as sufficiently typical. It is misleading because, first, it ignores the fact that the methodological faults in these earlier works were promptly pointed out in this country and have not been repeated, and, secondly, it ignores the great distance in methods and results between the two authors already named and Ames, King, Stratton, and Leuba I shall go into no critique of |