with children, storms with shipwrecks, passions with tragedies. If he begins in the middle he will still begin at the beginning of something, and perhaps as much at the beginning of things as he could possibly begin.
On the other hand, this whole supposition may be wrong. Things may have had some simpler origin, or may contain simpler elements. In that case it will be incumbent on the philosopher to prove this fact; that is, to find in the complex present objects evidence of their composition out of simples. But in this proof also he would be beginning in the middle; and he would reach origins or elements only at the end of his analysis.
The case is similar with respect to first principles of discourse. They can never be discovered, if discovered at all, until they have been long taken for granted, and employed in the very investigation which reveals them. The more cogent a logic is, the fewer and simpler its first principles will turn out to have been; but in discovering them, and deducing the rest from them, they must first be employed unawares, if they are the principles lending cogency to actual discourse; so that the mind must trust current presumptions no less in discovering that they are logical — that is, justified by more general unquestioned presumptions — than in discovering that they are arbitrary and merely instinctive.
It is true that, quite apart from living discourse, a set of axioms and postulates, as simple as we like, may be posited in the air, and deductions drawn from them ad libitum; but such pure logic is otiose, unless we find or assume that discourse or nature actually follows it; and it is not by deduction from first principles, arbitrarily chosen, that human reasoning actually proceeds, but by loose habits of mental evocation which such principles at best may exhibit afterwards in an idealised form. Moreover, if we could strip our
