5. The Evolutionary Method also leads to Reaction. — The law of preservation from which Spencer is deceived into deducing the necessity of the evolutionary process only applies to the quantitative relations of the forces at work in the system; hence it can give us no information as to the direction the changes will take. The qualitative transformation of forces, on the other hand, is subject to the law of degradation,[1] according to which the imperceptible differences, and more especially the inequalities existing in the redistribution of energy in respect to masses, constantly tend to diminish, so that the natural course taken by physical phenomena makes for the greater homogeneity of the system, though this is diametrically opposed to Spencer’s assertion. As far as the principle of conservation is concerned, it is a matter of indifference whether we pass from the homogeneous form of heat to differing forms of energy, or whether the process be reversed, since in either case it remains unchanged in its totality.
As Lalande[2] has well said, this permanency would be equally true even if the progress of the world were suddenly to be reversed, supposing, that is to say, trees were to grow smaller instead of taller, till they returned to the germs from which they had developed, and mankind were to grow towards youth instead of age, reaching the embryonic stage at the end of life instead of at the beginning. Nor is this hypothesis purely fantastical! There are many biological instances of retrogression or involution of organs,[3] yet the law of the persistence of force is in no way affected thereby. For that matter, does not Spencer himself deduce from the law of conservation the necessary dissolution of the system when its cycle has been accomplished? How marvellous is this law, from which we may deduce on the one hand, when it suits us so to do, the necessity of passing to the heterogeneous, and on the other with equal facility the no less necessary return to primitive homogeneity!
The evolutionary process cannot be deduced from a
