(2) Historical facts present themselves in very different degrees of generality, from the highly general facts which apply to a whole people and which lasted for centuries (institutions, customs, beliefs), down to the most transient actions of a single man (a word, a movement). Here again history differs from the sciences of direct observation, which regularly start from particular facts and labour methodically to condense them into general facts. In order to form groups the facts must be reduced to a common degree of generality, which makes it necessary to inquire to what degree of generality we can and ought to reduce the different species of facts. And this is what historians do not agree about among themselves. (3) Historical facts are localised; each belongs to a given time and a given country. If we suppress the time and place to which they belong, they lose their historical character; they now contribute only to the knowledge of universal humanity, as is the case with facts of folk-lore whose origin is unknown. This necessity of localisation is also foreign to the general sciences; it is confined to the descriptive sciences, which deal with the geographical distribution and with the evolution of phenomena. It obliges the historian to study separately the facts belonging to different countries and different epochs. (4) The facts which have been extracted from documents by critical analysis present themselves accompanied by a critical estimate of their probability.[178] In every case where we have not reached complete certainty, whenever the fact is merely probable--still more when it is open to suspicion--criticism supplies the fact to the historian accompanied by a label which he has no right to remove, and which prevents the fact from being definitively admitted into the science. Even those facts which, after comparison with others, end by being established, are subject to temporary exclusion, like the clinical cases which accumulate in the medical reviews before they are considered sufficiently proved to be received as scientific facts. Historical construction has thus to be performed with an incoherent mass of minute facts, with detail-knowledge reduced as it were to a powder. It must utilise a heterogeneous medley of materials, relating to different subjects and places, differing in their degree of generality and certainty. No method of classifying them is provided by the practice of historians; history, which began by being a form of literature, has remained the least methodical of the sciences.