religion” which is brought into being is a pleasant thought for an idle hour rather than a foundation and starting-point for the study of ancient religion in general.  Altogether aside from the fact that although primitive religion and nationality are in the main identical, language and nationality are by no means so—­we have the great practical difficulty in the case of Greece and Rome that in the earliest period of which we have knowledge these two religions bear so little resemblance that we must either assert for the time of Indo-Germanic unity a religious development much more primitive than that which comparative philology has sketched, or we must suppose the presence of a strong decadent influence in Rome’s case after the separation, which is equally difficult.  If we realise that in a primitive religion the name of the god is usually the same as the name of the thing which he represents, the existence of a Greek god and a Roman god with names which correspond to the same Indo-Germanic word proves linguistically that the thing existed and had a name before the separation, but not at all that the thing was deified or that the name was the name of a god at that time.  We must therefore be content to begin our study of religion much more humbly and at a much later period.

In fact we cannot go back appreciably before the dawn of political history, but there are certain considerations which enable us at least to understand the phenomena of the dawn itself, those survivals in culture which loom up in the twilight and the understanding of which gives us a fair start in our historical development.  For this knowledge we are indebted to the so-called “anthropological” method, which is based on the assumption that mankind is essentially uniform, and that this essential uniformity justifies us in drawing inferences about very ancient thought from the very primitive thought of the barbarous and savage peoples of our own day.  At first sight the weakness of this contention is more apparent than its strength, and it is easy to show that the prehistoric primitive culture of a people destined to civilisation is one thing, and the retarded primitive culture of modern tribes stunted in their growth is quite another thing, so that, as has so often been said, the two bear a relation to each other not unlike that of a healthy young child to a full-grown idiot.  And yet there is a decided resemblance between the child and the idiot, and whether prehistoric or retarded, primitive culture shows everywhere strong likeness, and the method is productive of good if we confine our reasoning backwards to those things in savage life which the two kinds of primitive culture, the prehistoric and the retarded, have in common.  To do this however we must have some knowledge of the prehistoric, and our modern retarded savage must be used merely to illumine certain things which we see only in half-light; he must never be employed as a lay-figure in sketching