But, on the other hand, as we have already seen, the positive sciences are not self-sufficient. They need to be crowned and ordered by a philosophy. If then a philosophy is indispensable, and if, at the same time, this philosophy must be positive, relative like the sciences themselves, and _homogeneous_ with them, only one solution remains possible. The philosophy of the sciences will consist in substituting the point of view of the whole to that of the parts. It will still be a product of the positive spirit; but in it this spirit from special will have become general; from particular it will have become universal. This universal character remains common to Comte’s philosophy and to that of his predecessors. But Comte did not understand it as they did. For metaphysicians in general, and still for Kant, universality is the distinctive sign of knowledge which does not come from experience, which is therefore necessary and _a priori_. Comte, who does not know of any _a priori_ in the Kantian sense, calls that knowledge universal which remains relative, and which is founded upon induction, but which regulates the other forms of knowledge in the order of generality. Thus the principle of laws is universal. The encyclopædic laws of phenomena are universal. The point of view of humanity is universal, because from this point of view a synthesis of the whole of our knowledge is possible. And, as universality is a relative thing, we conceive universalities of different orders. Henceforth the philosophy of the sciences is easily defined. Are we concerned with a certain science considered by itself? The philosophy of this science consists in embracing at a glance the whole, the object and the method, as opposed to the special point of view of the scientific man who follows the discovery of more or less special laws in a branch of this science, but such a philosophy necessarily remains imperfect and fragmentary. The philosophy of a science is only really established in the general philosophy of the sciences, that is to say by a view at once synthetic and single of all the sciences, in which are co-ordinated the objects which they study, the laws which they discover, the methods which they make use of, and the ends which they should pursue.