786 THE AMERICAN JOURNAL OF SOCIOLOGY
it must face two ways — inward to the springs of the emotions, outward to the means by which it shows itself before the world. It is, therefore, on the one hand the state of feeling awakened in a man by the consciousness in and around him of mysterious powers, and on the other, those manifestations in words, deeds, customs, and institutions which testify to his belief in these powers, and serve to bring him into relation to them.® Under the wide arch of this definition, all the phenomena of religion find a place, and we shall not have to cross and recross a sacred threshold as we turn from sacrifice to purification, which would be the case if we attempted to frame our definition so as to shut out magic from religion.
We therefore place magic within the sphere of religion — by means of a definition! But a definition never solves our diffi- culties ; it is at best but our working hypothesis ; and the question now is : Does it work ? So with our tables cleared we return to our magic. But we return to it, not as a thing by itself, a science
" This definition is based upon two, neither of which seems to rae wholly satisfactory : that of L. Marillier in his admirable article on "Religion" in La Grande Encyclopedic, and that of C. P. Tiele in Elements of the Science of Re- ligion (I, 4). Marillier says: "Religion is the sum of three things: (i) the state of feeling awakened in a man by the obscure consciousness in and around him of powers, superior and analogous to himself, with whom he can enter into relation ; (2) the representations engendered by these sentiments and which furnish them definite objects; and (3) the ritual acts called forth by the combined action of these emotions and beliefs." "Tide's definition is as follows : "By religion I mean those manifestations of the human mind in words, deeds, customs, and institutions which testify to man's belief in the superhuman, and serve to bring him into relation with it" The difference between either of these and the composite definition above is obvious. The one point at which immediate exception may be taken in Tide's definition is its attitude toward religious beliefs, whether myths or theologies. These do, as Marillier insists, play a great and essential role in religion and should receive clear recognition. But "the consciousness of mysterious powers" however acutely it may be developed, is rather to be viewed as a religious stimulus than as religion itself. The con- templation of the beatitudes of heaven or of the sufferings of Christ lifts or stirs the emotions like those desert silences that find a lyric voice in St. Jerome or the sound of music or any other powerful impulse. Nor are the powers "analogous to himself" as Marillier states, until animism is reached, when the growing consciousness of individuality throws its changing reflex upon the screen of environing reality.
