we make the transition from cause to effect; and 'tis from some present impression we borrow that vivacity, which we diffuse over the correlative idea. But when we have not observ'd a sufficient number of instances, to produce a strong habit; or when these instances are contrary to each other; or when the resemblance is not exact; or the present impression is faint and obscure; or the experience in some measure obliterated from the memory; or the connexion dependent on a long chain of objects; or the inference deriv'd from general rules, and yet not conformable to them: In all these cases the evidence diminishes by the diminution of the force and in tenseness of the idea. This therefore is the nature of the judgment and probability.
What principally gives authority to this system is, beside the undoubted arguments, upon which each part is founded, the agreement of these parts, and the necessity of one to explain another. The belief, which attends our memory, is of the same nature with that, which is deriv'd from our judgments: Nor is there any difference betwixt that judgment, which is deriv'd from a constant and uniform connexion of causes and effects, and that which depends upon an interrupted and uncertain. 'Tis indeed evident, that in all determinations, where the mind decides from contrary experiments, 'tis first divided within itself, and has an inclination to either side in proportion to the number of experiments we have seen and remember. This contest is at last determin'd to the advantage of that side, where we observe a superior number of these experiments; but still with a diminution of force in the evidence correspondent to the number of the opposite experiments. Each possibility, of which the probability is compos'd, operates separately upon the imagination; and 'tis the larger collection of possibilities, which at last prevails, and that with a force proportionable to its superiority. All these phænomena lead directly to the precedent system; nor will it ever be possible upon any other principles
to give a satisfactory and consistent explication of them.