We must now return to the higher forms of intermediate abstraction, since we may not content ourselves with any purely theoretical determination. Characteristic examples must be selected; and here we find a certain embarrassment. Does our choice fall on numeration? Yet on leaving the concrete-abstract period, this at once finds its formative law, and introduces us to pure abstraction. Are we to select language? This procedure might seem to be appropriate, seeing that the general ideas with which we are occupied constitute the substrata of our highly civilised modern languages, when, on the other hand, the more developed concepts (of mathematics, metaphysics, etc.) are only found rarely and incidentally. One might even plunder the dictionary, extracting all general terms, with elimination of those that are purely scientific, and classification of the former according to their increasing degree of generality. But this method, besides being very laborious and incapable of reduction to a clear statement for the reader, would suffer the cardinal defect of being arbitrary. How, indeed, could any common measure be established for all these general terms, issuing from the most diverse sources of human activity?[77] But the best method would seem to be that of taking as our basis the classifications of the naturalists, following their development historically. Here we have the advantage of positive documents, since these refer to concrete beings, and are formed according to characters observed empirically. They create, namely, an ascending progress from the individual to the more general notions, by a methodical process of filiation; they operate upon living beings, or objects of the same nature, having consequently a common standard. The history, even in brief, of these classifications is instructive: it shows the progressive passage of concrete-abstract ideas to more and more abstract concepts, from a statement of gross resemblances to the quest after subtle similarities, from the period of assimilation to that in which dissociation predominates. Among these different classifications, we may select those of the zoologists, since they appear to be the most numerous, most complete, and best elaborated. For the rest, the succeeding observations apply equally, _mutatis mutandis_, to the classifications of the botanists. We need scarcely add that our study is strictly psychological, that its object is not the absolute value of classifications, but the determination of the processes followed by the human mind, in proportion as the zoölogical taxonomy has constituted itself.