If it be urged, as it may naturally be, that on this showing there is no solid or 'absolute' starting-point at all, the contention may be conceded. The only fixed and steady points in knowledge are points hypothetically fixed,--certified, that is, for the time and in the circumstances we employ them. But in the open field--or rather in the wilderness--of knowledge, where the ground of fact is not staked off, and the unexpected may always turn up, the only test of truth is the corroboration given by the consilience of paths initiated from different points: it is only by an undesigned coincidence in the results of independent operations that you can succeed in orienting yourself. You begin your road at two ends, and you meet: you locate or fix your point by drawing its co-ordinates to two direction-lines taken anyhow at first, and only in formed science diverging at a fixed angle. And in the absolute your direction-lines cannot be supposed fixed: you can only gradually adjust them to each other as you proceed. Intelligence, says Aristotle, is a principle, a beginning; and intelligence, he says again, supplies beginnings[4]. Science, in the technical sense, only comes into operation,--or, in other words, deduction and (in the narrower sense used by Mill, and proceeding by _pure_ observation and experiment) induction only find a way,--where beginnings and principles have been set up, where an approximate order or provisional system has been established. And if logic, in its stricter sense, is the method of sciences already made and in their essentials constituted, then logic can be asked to do no more than to provide a theory of such formal processes. If it traces the path which leads 'from the known to the unknown,' if it always proceeds on the hypothesis of a given knowledge, then such induction or deduction (from certain and approved singular facts, or from certain and approved general truths) fully satisfies the practical need of the scientific reasoner. But if Logic be, as it sometimes is, and may very reasonably be, taken in the wider sense of an epistemology,--a theory of the nature and origin of knowledge as a whole, and not of mere inference or syllogism;--if it does not merely ask how we can satisfactorily get from one piece of knowledge (we are supposed to have) to another (not yet supposed to be), but how we come to have knowledge at all; then its problem must go behind the rudiments of vulgar induction and deduction. It must ask--what, so far as one can see, Mill and his mere followers have never seriously asked at all--what induction is, what are its relations with deduction, and what is the place of either in the process of knowledge. And as the process of knowledge is the path to reality, it must also ask about the nature of this goal,--reality and truth. It is all very well for the narrower Logic to formulate in terms the methods actually employed in sciences: to state in abstract canons what is there seen in life and action. But