Consequently, while in the _a priori_ doctrines the possibility of all science rests upon the principle of laws, in Comte’s doctrine, on the contrary, it is the progress of positive science which by degrees founds the principle, and which finally brings it to the universal form in which we find it to-day. Until the creation of sociology, this principle did not yet possess an effective universality, since the moral and social phenomena were not conceived as subject to invariable laws. But when the last conquest of the positive spirit is once accomplished, “this great principle at once acquires a decisive fulness, and may be formulated as applying universally to all phenomena.” Undoubtedly, in each order, we have only established for a few what henceforth we affirm for all phenomena without previous verification. But we think that laws, unknown to us, nevertheless exist. In this we yield to an “irresistible analogy,” which has never been proved to be false. Thus, “the most fundamental dogma of the whole of positive philosophy, that is to say, the subjection of all real phenomena to invariable laws, only results with certainty from an immense induction, without really being deducible from any notion whatever.”[53] This immense induction is a progressive sum of inductions which have taken place successively in each category of phenomena. It would not be absurd, strictly speaking, that a certain category should not be submitted, like the others to invariable laws. But, since sociology has been founded, we know that all are in fact so subjected. The laws are known to us, sometimes by experience, sometimes by reasoning. This diversity of origin in no way influences either the certainty or the philosophical dignity of the laws. Each of the six fundamental sciences gives examples of these two distinct modes of advance which mutually complete each other. “There is not less genius in the discovery of Kepler than in that of Newton. The initial laws of mechanics and even of geometry rest solely upon observation. The logical perfection consists in confirming by one of these ways what must have been found by the other. But one of the two suffices when all the conditions required by the method are fulfilled.”[54] How should the laws obtained by induction be regarded as less certain than the laws obtained by deduction, since the principle of laws itself rests upon an induction? III. In proportion as the several orders of phenomena are conceived as governed by invariable laws, the belief in final causes becomes weaker and tends to disappear. The final causes are imagined by the mind to explain certain combinations of natural phenomena. When the laws of these phenomena are known, this explanation becomes useless, it ceases to have currency. It shares the fate of the whole of theological and metaphysical philosophy, of which it is a part.