science, unanalysed to its true presuppositions, consistently interprets this connexion into the merely regular succession of the past — a sequence merely de facto; but if we thoroughly consider what is logically presupposed in scientific method as actually used by the competent, we shall readily see that it should be interpreted as necessary and irreversible succession, a sequence inevitable forever. For the vital process in scientific method is induction, or generalisation; and the secret of it, as actually employed in scientific practice, lies in taking observed successions in phenomena, and when with the help of the various methods of precision — agreement, difference, joint agreement, concomitant variation — they are brought to represent exactly what occurs, then suddenly giving to these merely historical successions the value of universal laws, having a predictive authority over the future in perpetuum.
If in this process there is always a cautious reserve in the mind of the practised and sedate man of science, — as indeed there is, — the reserve has no reference to the amazing final act of generalisation: all the anxieties of the expert are about the precision of his facts. His instinctive assumption about the generalisation is, that, when once the particulars are settled, this process takes place of itself, is matter-of-course, is resistless and flawless: if there is error anywhere in the scientific proced-
