[Sidenote: but is very apt to lead to false beliefs,] 1. While many empirical conclusions are, roughly speaking, correct; while they are exact enough to be of great help in practical life; while the presages of a weatherwise sailor or hunter may be more accurate, within a certain restricted range, than those of a scientist who relies wholly upon scientific observations and tests; while, indeed, empirical observations and records furnish the raw or crude material of scientific knowledge, yet the empirical method affords no way of discriminating between right and wrong conclusions. Hence it is responsible for a multitude of _false_ beliefs. The technical designation for one of the commonest fallacies is _post hoc, ergo propter hoc_; the belief that because one thing comes _after_ another, it comes _because_ of the other. Now this fallacy of method is the animating principle of empirical conclusions, even when correct--the correctness being almost as much a matter of good luck as of method. That potatoes should be planted only during the crescent moon, that near the sea people are born at high tide and die at low tide, that a comet is an omen of danger, that bad luck follows the cracking of a mirror, that a patent medicine cures a disease--these and a thousand like notions are asseverated on the basis of empirical coincidence and conjunction. Moreover, habits of expectation and belief are formed otherwise than by a number of repeated similar cases. [Sidenote: and does not enable us to cope with the novel,] 2. The more numerous the experienced instances and the closer the watch kept upon them, the greater is the trustworthiness of constant conjunction as evidence of connection among the things themselves. Many of our most important beliefs still have only this sort of warrant. No one can yet tell, with certainty, the necessary cause of old age or of death--which are empirically the most certain of all expectations. But even the most reliable beliefs of this type fail when they confront the _novel_. Since they rest upon past uniformities, they are useless when further experience departs in any considerable measure from ancient incident and wonted precedent. Empirical inference follows the grooves and ruts that custom wears, and has no track to follow when the groove disappears. So important is this aspect of the matter that Clifford found the difference between ordinary skill and scientific thought right here. "Skill enables a man to deal with the same circumstances that he has met before, scientific thought enables him to deal with different circumstances that he has never met before." And he goes so far as to define scientific thinking as "the application of old experience to new circumstances." [Sidenote: and leads to laziness and presumption,]