Difference of object leads to difference of method. If an historian proposes to himself to compose exact “annals,” to relate things as they took place, he will begin by the special history of the various peoples, which, in its turn, is founded upon the chronicles of the provinces and the towns. It will be necessary for him to investigate documents in detail, and to neglect no source: the work of combination will only come subsequently. But if our end is the abstract science of history, that is to say the linking together of social phenomena, quite a different course will have to be followed. Indeed all the classes of these phenomena are simultaneously developed, and under the mutual influence of one another. We cannot explain the line of advance followed by anyone among them, without having first conceived in a general way “the progression of the whole.” Before all things then we must set ourselves to conceive the development of the human species in its widest generality, that is to say, to observe and to link together among themselves the most important steps towards progress which it has successively taken in the various fundamental directions. Then we shall subdivide the periods and the classes of the phenomena to be observed.[246] These “various fundamental directions” correspond to what Comte called later the “social series.” By this he indicates the groups of social phenomena arranged for a scientific study. Once these groups are formed, then, according to the totality of historical facts, the sociologist seeks to determine the continuous growth of each, physical, moral, intellectual or political disposition or faculty, combined with the indefinite decrease of the opposite disposition or faculty: for instance, the tendency of human society to pass from the warlike form to the industrial form, from revealed religion to demonstrated religion, etc. From this will be drawn the scientific forecast of the triumph of the one and the fall of the other, provided that this conclusion is also in conformity with the general laws of the evolution of Humanity. Such a forecast could never be founded upon the knowledge of the present alone. For the present exposes us to the danger of confusing the principal with the secondary facts, of “placing noisy passing demonstrations above deep-seated tendencies,” and of regarding institutions or doctrines as growing which are really on the decline. Our statesmen scarcely look back beyond the XVIII. century, our philosophers beyond the XVI. This is too little. It does not even suffice to make us understand the French revolution. The study of the “historical series” alone allows the understanding of the present and the prevision of the future. The sociologist will even exercise himself in predicting the past, that is to say, in acquiring a rational knowledge of it, and in deducing each historical situation from the whole of its antecedents. He will thus become familiar with the spirit of the historical method.