We are hence eager to supply, by a little guess-work of fancy, the parts unobserved, and suppose deficiencies in our observation where there may truly have been none; till at length, by this habitual process, every phenomenon becomes, to our imagination, the sign of something _intermediate_ as its cause, the discovery of which is to be an explanation of the phenomenon. The mere succession of one event to another appears, to us, very difficult to be conceived, because it wants that intervening something, which we have learned to consider as a cause; but there seems to be no longer any _mystery_, if we can only suppose something intervening between them, and can thus succeed in doubling the difficulty, which we flatter ourselves with having removed; since, by the insertion of another link, we must now have two sequences of events instead of one simple sequence. This tendency of the imagination to form and rest on hypotheses,--or, in other words, to suppose substances present and operating, of the existence of which we have no direct proof,--we found to be one great source of error in our practice of philosophizing. Another source of error, we found to be the _too great extension_ of what are termed general laws; which though a less error in itself, is yet, in one respect, more dangerous than the former; because it is the error of better understandings,--of understandings that would not readily fall into the extravagant follies of hypotheses, but acknowledge the essential importance of induction, and think they are proceeding on it without the slightest deviation, almost at the very moment when they are abandoning it for conjecture. To observe the regular series of antecedents and consequents, and to class these as similar or dissimilar, are all which philosophers can do with complete certainty. But there is a constant tendency in the mind, to convert a _general_ law into an _universal_ law,--to suppose, after a wide induction, that what is true of many substances that have a very striking analogy, is as certainly true of _all_ that have this striking analogy,--and that what is true of them in _certain_ circumstances, is true of them in _all_ circumstances,--or, at least, in all circumstances which are not remarkably different. The widest induction which we can make, however, is still limited in its nature; and, though we may have observed substances in many situations, there may be some new situations, in which the event may be different, or even, perhaps, the very reverse of that which we should have predicted, by reasoning from the mere analogy of other circumstances. It appeared to me necessary, therefore, in consequence of the very ambiguous manner in which writers on this higher branch of logic speak of reasoning from general laws to particulars, to warn you, that the application to particulars can be made with certainty, only to the very particulars before observed and generalized,--and that, however analogous other