he would estimate the chance bordering upon certainty that he was not deceived in the particular instance before him. It is by such a product, in fact, of separate unlikelihoods, that a man can identify on the instant, not merely his own friend, but his own hat, his own staff, his own umbrella. It is the number of independent characteristics which meet in any of these articles, that leads to so swift and sound a conclusion in regard to them; and precisely on the same principle, it is the number of the independent characteristics of truth meeting in one testimony which gives a power of conviction to it, that surmounts the most violent improbability. This surely is a possible explanation of the fact that so many enlightened minds, on a specific examination of the actual evidence for the miracles of Christianity, have deferred to them-long before Hume framed his metaphysical argument, or La Place revived it in a mathematical form. It is well however that it admits both of a metaphysical and mathematical refutation the strength of which however becomes more palpable, when we pass as in the following section from the consideration of a single testimony, to that of a concurrence of testimonies.

SECTION IV.—On the Power which lies in the Concurrence of distinct Testimonies.

1. In throwing a pair of dice, the chance that aces shall cast up, is as one to thirty-six. There being

six sides on each of the dice, the chance that an ace shall cast up with the first of them is as one to six, and with the second is also one to six: and one to thirty-six expresses the probability that both shall concur the chance of both conditions being realized at once, being just equal to the product of the separate chances into each other.

2. It is precisely at such a rate that the probability for the falsehood of any definition decreases, or the probability for its truth increases, with the concurrence of testimonies in its favour. If one of these testimonies be of such a kind, as, taking all the ostensible circumstances together, to have proved false once in six times, this single testimony gives the probability of six to one on the side of the thing deponed to. The addition of just such another testimony would make out the probability of thirty-six, and of a third two hundred and sixteen-or the probability on the whole, arising from testimony or the truth of any occurrence, may be represented, by the product of the separate probabilities for the truth of each individual testimony.

3. Of course, the testimonies must be supposed independent of each other. And then, we are not to wonder, at the speedy and perfect assurance which, by their means, we obtain of many events, although they should have no other evidence to rest upon. Such, after all, is the majority of truth to falsehood in the world that, on the strength even of one of its every-day testimonies, we place implicit reliance on the truth of an event, whereof previously, we had no expectation. Of

how many of our familiars may it be said, that the chance is at least as a thousand to one, of his speaking truth rather than falsehood. Let two such concur, then, in any deposition and, in as far as the probability of an event depends on the integrity of the witnesses, there are the chances of no less than a million to one in its favour. There must, then, be the inherent improbability of a million to one in the event itself, ere, with such a support from testimony, it can be dismissed as unworthy of credit. It will be seen, by what an immense superiority of evidence, on the addition of a third or a fourth or a certain number of witnesses, even this or indeed any definite improbability might be overcome an evidence, which grows and gathers in rapid multiple progression by the addition of every new witness-provided always, that each depones on his own independent knowledge, and that they have no collusion with one another.

4. It is thus that, had we good enough separate testimonies, we might obtain by their conjunction, an evidence in behalf of a miracle that would outweigh to any amount the improbability which is inherent in the miracle itself. It is quite true that the establishment of a miracle requires stronger testimony than an ordinary event does_ yet let that stronger testimony only be multiplied as much as the weaker, and the result would be, that the miracle should not only be as credible, but indefinitely and to any extent more credible, than the ordinary event. For example let the improbability of a miracle be estimated at a million, and

attested by three witnesses for each of whose separate integrities there is the probability of a million-then from the testimony of any one of these witnesses we obtain an equivalent or equipoise to the improbability of the miracle-leaving the product of the remaining two integrities, or a million of millions to represent the strength of our reason for believing the alleged miracle to be true. Should the ordinary event, on the other hand, have, in certain given circumstances, the improbability of a thousand attached to it, and be attested by three witnesses for each of whose integrities there is the probability of a thousandthen, as before, would the deposition of one of these witnesses neutralize the improbability of the event; but the joint testimony of the remaining two witnesses would only afford the probability of a thousand times a thousand, or of one million to represent the strength of our reason for believing the alleged event to be true. In other words, we should, in the respective circumstances now stated, have a million times better reason for believing in the truth of the miracle than in that of the ordinary event. Sceptics complain of the tax on their credulity, when they are called upon to put faith in miracles. Let them have a care, lest they, all the time, should, in reference to the miracles of the gospel, be resisting a claim upon their belief, many million-fold greater, than is possessed even by the commonest events in the history of past ages.

5. But, to obtain the requisite strength of evidence for overcoming the native improbability of a miracle,

it is not necessary that all the separate testimonies should be of the best and highest description. By one such testimony we might effect an equipoise. And by the addition of another, though of a very inferior sort, we might gain a preponderancy. One testimony of a superlative order, and whose falsehood would be miraculous, is of force enough, at least to countervail the improbability of an event whose truth would be miraculous. The superaddition of another testimony, of so low a character as to have deceived or misgiven once in six times, would of itself establish a proof of six times greater strength than the improbability that had to be overcome by it. It is thus that subsidiary, though inferior testimonies, are not without effect on the general result. They could, even of themselves overbalance the unlikelihood of a miracle-and when compounded, as they are in gospel history, with so many testimonies of the highest kind, the effect exceeds all conception, if not all computation. Nothing therefore can be juster than the reflection of Dr. Paley, when, in coming to a conclusive reckoning with Mr. Hume, he practically disposes of his argument by showing how it fails in a specific

case.

6. This deliverance of Paley, proceeded from the force of the evidence being felt, not from its being calculated. For in order to be felt and felt rightly too, or in some sort of general proportion to its strength, it is not necessary that the calculation should precede the feeling. There is nothing

* See Chap. I. § 20.