In the first place, the roots are general; that is to say that they do not express particular things and individuals, but types, and even types of an extreme generality. They represent the most general themes of thought; one finds there, as though fixed and crystallized, those fundamental categories of the intellect which at every moment in history dominate the entire mental life, the arrangement of which philosophers have many times attempted to reconstruct.153
Secondly, the types to which they correspond are types of action, and not of objects. They translate the most general manners of acting which are to be observed among living beings and especially among men; they are such actions as striking, pushing, rubbing, lying down, getting up, pressing, mounting, descending, walking, etc. In other words, men generalized and named their principal ways of acting before generalizing and naming the phenomena of nature.154
Owing to their extreme generality, these words could easily be extended to all sorts of objects which they did not originally include; it is even this extreme suppleness which has permitted them to give birth to the numerous words which are derived from them. Then when men, turning towards things, undertook to name them, that they might be able to think about them, they applied these words to them, though they were in no way designed for them. But, owing to their origin, these were able to designate the forces of nature only by means of their manifestations which seemed the nearest to human actions: a thunderbolt was called something that tears up the soil or that spreads fire; the wind, something that sighs or whistles; the sun, something that throws golden arrows across space; a river, something that flows, etc. But since natural phenomena were thus compared to human acts, this something to which they were attached was necessarily conceived under the form of personal agents, more or less like men. It was only a metaphor, but it was taken literally; the error was inevitable, for science, which alone could dispel the illusion, did not yet exist. In a word, since language was made of human elements, translating human states, it could not be applied to nature without transforming it.155 Even to-day, remarks M. Bréal, it forces us in a certain measure to represent things from this angle. "We do not express an idea, even one designating a simple quality, without giving it a gender, that is to say, a sex; we cannot speak of an object, even though it be considered in a most general fashion, without determining it by an article; every subject of a sentence is presented as an active being, every idea as an action, and every action, be it transitory or permanent, is limited in its duration by the tense in which we put the verb."156 Our scientific training enables us to rectify the errors which language might thus suggest to us; but the influence of the word ought to be all-powerful when it has no check. Language thus superimposes upon the material world, such as it is revealed to our senses, a new world, composed wholly of spiritual beings which it has created out of nothing and which have been considered as the causes determining physical phenomena ever since.
But its action does not stop there. When words were once forged to represent these personalities which the popular imagination had placed behind things, a reaction affected these words themselves: they raised all sorts of questions, and it was to resolve these problems that myths were invented. It happened that one object received a plurality of names, corresponding to the plurality of aspects under which it was presented in experience; thus there are more than twenty words in the Vedas for the sky. Since these words were different, it was believed that they corresponded to so many distinct personalities. But at the same time, it was strongly felt that these same personalities had an air of relationship. To account for that, it was imagined that they formed a single family; genealogies, a civil condition and a history were invented for them. In other cases, different things were designated by the same term: to explain these homonyms, it was believed that the corresponding things were transformations of each other, and new fictions were invented to make these metamorphoses intelligible. Or again, a word which had ceased to be understood, was the origin of fables designed to give it a meaning. The creative work of language continued then, making constructions ever more and more complex, and then mythology came to endow each god with a biography, ever more and more extended and complete, the result of all of which was that the divine personalities, at first confounded with things, finally distinguished and determined themselves.
This is how the notion of the divine is said to have been constructed. As for the religion of ancestors, it was only a reflection of this other.157 The idea of the soul is said to have been first formed for reasons somewhat analogous to those given by Tylor, except that according to Max Müller, they were designed to account for death, rather than for dreams.158 Then, under the influence of diverse, partially accidental, circumstances,159 the souls of men, being once disengaged from the body, were drawn little by little within the circle of divine beings, and were thus finally deified themselves. But this new cult was the product of only a secondary formation. This is proven by the fact that deified men have generally been imperfect gods or demi-gods, whom the people have always been able to distinguish from the genuine deities.160
II
This doctrine rests, in part, upon a certain number of linguistic postulates which have been and still are very much questioned. Some have contested the reality of many of the similarities which Max Müller claimed to have found between the names of the gods in the various European languages. The interpretation which he gave them has been especially doubted: it has been asked if these names, far from being the mark of a very primitive religion, are not the slow product, either of direct borrowings or of natural intercourse with others.161 Also, it is no longer admitted that the roots once existed in an isolated state as autonomous realities, nor that they allow us to reconstruct, even hypothetically, the original language of the Indo-Europeans.162 Finally, recent researches would tend to show that the Vedic divinities did not all have the exclusively naturistic character attributed to them by Max Müller and his school.163 But we shall leave aside those questions, the discussion of which requires a special competence as a philologist, and address ourselves directly to the general principles of the system. It will be important here not to confound the naturistic theory with these controverted postulates; for this is held by numbers of scholars who do not make language play the predominating rôle attributed to it by Max Müller.
That men have an interest in knowing the world which surrounds them, and consequently that their reflection should have been applied to it at an early date, is something that everyone will readily admit. Co-operation with the things with which they were in immediate connection was so necessary for them that they could not fail to seek a knowledge of their nature. But if, as naturism pretends, it is of these reflections that religious thought was born, it is impossible to explain how it was able to survive the first attempts made, and the persistence with which it has maintained itself becomes unintelligible. If we have need of knowing the nature of things, it is in order to act upon them in an appropriate manner. But the conception of the universe given us by religion, especially in its early forms, is too greatly mutilated to lead to temporarily useful practices. Things become nothing less than living and thinking