(1) The spontaneous tendency is to take as a basis of reasoning those "common-sense truths" which form nearly the whole of our knowledge of social life. Now, the greater part of these are to some extent false, for the science of social life is still imperfect. And the chief danger in them lies in the circumstance that we use them unconsciously. The safest precaution will be always to formulate the supposed law on which we propose to base an argument. In every instance where such and such a fact occurs, it is certain that such and such another fact occurs also. If this proposition is obviously false, we shall at once see it to be so; if it is too general, we shall inquire what new conditions may be introduced to make it accurate. (2) A second spontaneous impulse leads us to draw consequences from isolated facts, even of the slightest kind (or rather, the idea of each fact awakens in us, by association, the idea of other facts). This is the natural procedure in the history of literature. Each circumstance in the life of an author supplies material for reasoning; we construct by conjecture all the influences which could have acted upon him, and we assume that they did act upon him. All the branches of history which study a single species of facts, isolated from every other species (language, arts, private law, religion), are exposed to the same danger, because they deal with fragments of human life, not with comprehensive collections of phenomena. But few conclusions are firmly established except those which rest on a comprehensive body of data. We do not make a diagnosis from a single symptom, but from a number of concurrent symptoms. The precaution to be taken will be to avoid working with an isolated detail or an abstract fact. We must have before our minds actual men, as affected by the principal conditions under which they lived. We must be prepared to realise but rarely the conditions of a certain inference; we are too little acquainted with the laws of social life, and too seldom know the precise details of an historical fact. Thus most of our reasonings will only afford presumptions, not certainties. But it is with reasonings as with documents.[202] When several presumptions all point in the same direction they confirm each other, and end by producing a legitimate certitude. History fills up some of its gaps by an accumulation of reasonings. Doubts remain as to the Phoenician origin of various Greek cities, but there is no doubt about the presence of the Phoenicians in Greece. CHAPTER IV THE CONSTRUCTION OF GENERAL FORMULÆ I. Suppose we had methodically arranged all the historical facts established by the analysis of documents, or by reasoning; we should possess a systematised inventory of the whole of history, and the work of construction would be complete. Ought history to stop at this point? The question is warmly debated, and we cannot avoid giving an answer, for it is a question with a practical bearing.