support, and is hardly ever honoured with the name of proof. Nevertheless, when the analogies are numerous, and the subject admits not evidence of another kind, it doth not want its efficacy. It must be owned, however, that it is generally more successful in silencing objections than in evincing truth, and on this account may more properly be styled the defensive arms of the orator than the offensive. Though it rarely refutes, it frequently repels refutation, like those weapons which, though they cannot kill the enemy. will ward his blows.*

III. Testimony.

The third tribe is the evidence of testimony, which is either oral or written. This, also, hath been thought by some, but unjustly, to be solely and originally derived from the same source, experience.† The utmost in regard to this that can be affirmed with truth is, that the evidence of testimony is to be considered as strictly logical, no farther than human veracity in general, or the veracity of witnesses of such a character, and in such circumstances in particular, is supported; or, perhaps, more properly, hath not been refuted by experience. But that testimony, antecedently to experience, hath a natural influence on belief, is undeniable. In this it resembles memory; for though the defects and misrepresentations of memory are corrected by experience, yet that this faculty hath an innate evidence of its own, we know from this, that if we had not previously given an implicit faith to memory, we had never been able to acquire experience. This will appear from a revisal of its nature, as explained above. Nay, it must be owned, that in what regards single facts, testimony is more adequate evidence than any conclusion from experience. The immediate conclusions from experience are general, and run thus: "This is the ordinary course of nature." "Such an event may reasonably be expected, when all the attendant circumstances are similar." When we descend to particulars, the conclusion necessarily becomes weaker, being more indirect; for, though all the known circumstances be similar, all the actual circumstances may not be similar; nor is it possible, in any case, to be assured that all the actual circumstances are known to us. Accordingly, experience is the foundation of philosophy, which consists in a collection of general truths, systematically digested. On

* Dr. Butler, in his excellent treatise called The Analogy of Religion, Natural and Revealed, to the Constitution and Course of Nature, hath shown us how useful this mode of reasoning may be rendered, by the application he hath so successfully made of it, for refuting the cavils of infidelity.

t I had occasion to make some reflections on this subject formerly. See Dissertation on Miracles, part i., sect. i. There are several ingenious observations on the same subject in Reid's Inquiry, ch. vi., sect. xxiii.

the contrary, the direct conclusion from testimony is partic nlar, and runs thus: "This is the fact in the instance specified." Testimony, therefore, is the foundation of history, which is occupied about individuals. Hence we derive our acquaintance with past ages, as from experience we derive all that we can discover of the future. But the former is dignified with the name of knowledge, whereas the latter is regarded as matter of conjecture only. When experience is applied to the discovery of the truth in a particular incident, we call the evidence presumptive; ample testimony is accounted a positive proof of the facts. Nay, the strongest conviction built merely on the former is sometimes overturned by the slightest attack of the latter. Testimony is capable of giving us absolute certainty (Mr. Hume himself being judge*) even of the most miraculous fact, or of what is contrary to uniform experience; for, perhaps, in no other instance can experience be applied to individual events with so much certainty as in what relates to the revolutions of the heavenly bodies. Yet, even this evidence, he admits, may not only be counterbalanced, but destroyed, by testimony.

But to return. Testimony is a serious intimation from another of any fact or observation, as being what he remembers to have seen, or heard, or experienced. To this, when we have no positive reasons of mistrust or doubt, we are, by an original principle of our nature (analogous to that which com, pels our faith in memory), led to give an unlimited assent. As on memory alone is founded the merely personal experience of the individual, so on testimony, in concurrence with memory, is founded the much more extensive experience, which is not originally our own, but derived from others.t By the first, I question not, a man might acquire all the knowledge necessary for mere animal support, in that rudest state of human nature (if ever such a state existed) which was without speech, and without society; to the last, in conjunction with the other, we are indebted for everything which distinguishes the man from the brute, for language, arts, and civilization. It hath been observed, that from experience we learn to confine our belief in human testimony within the proper bounds. Hence we are taught to consider many attendant circumstances, which serve either to corroborate or to invalidate its evidence. The reputation of the attester, his manner of address, the nature of the fact attested, the occasion of giving the testimony, the possible or probable design in giving it, the disposition of the hearers to whom it was given, and several other circumstances, have all consid erable influence in fixing the degree of credibility. But of these I shall have occasion to take notice afterward. It de

* Essay on Miracles, p. 2.

† Dissertation on Miracles, part i., sec. ii.

serves, likewise, to be attended to on this subject, that in a number of concurrent testimonies (in cases wherein there could have been no previous concert), there is a probability distinct from that which may be termed the sum of the probabilities resulting from the testimonies of the witnesses, a probability which would remain even though the witnesses were of such a character as to merit no faith at all. This probability ariseth purely from the concurrence itself. That such a concurrence should spring from chance, is as one to infinite; that is, in other words, morally impossible. If, therefore, concert be excluded, there remains no other cause but the reality of the fact.

Now to this species of evidence, testimony, we are first immediately indebted for all the branches of philology, such as history, civil, ecclesiastic, and literary; grammar, languages, jurisprudence, and criticism; to which I may add revealed religion, as far as it is to be considered as a subject of historical and critical inquiry, and so discoverable by natural means and, secondly, to the same source we owe, as was hinted above, a great part of that light which is commonly known under the name of experience, but which is, in fact, not founded on our own personal observations, or the notices originally given by our own senses, but on the attested experiences and observations of others. So that as hence we derive entirely our knowledge of the actions and productions of men, especially in other regions, and in former ages; hence also we derive, in a much greater measure than is commonly imagined, our acquaintance with Nature and her works. Logic, rhetoric, ethics, economics, and politics, are properly branches of pneumatology, though very closely connected with the philological studies above enumerated.

IV. Calculations of Chances.

The last kind of evidence I proposed to consider was that resulting from calculations of chances. Chance is not commonly understood, either in philosophic or in vulgar language, to imply the exclusion of a cause, but our ignorance of the cause. It is often employed to denote a bare possibility of an event, when nothing is known either to produce or to hinder it. But in this meaning it can never be made the subject of calculation. It then only affords scope to the calculator, when a cause is known for the production of an effect, and when that effect must necessarily be attended with this, or that, or the other circumstance; but no cause is known to determine us to regard one particular circumstance, in preference to the rest, as that which shall accompany the supposed effect. The effect is then considered as necessary, but the circumstance as only casual or contingent. When a die

is thrown out of the hand, we know that its gravity will make it fall; we know, also, that this, together with its cubical figure, will make it lie so, when intercepted by the table, as to have one side facing upward. Thus far we proceed on the certain principles of a uniform experience; but there is no principle which can lead me to conclude that one side rather than another will be turned up. I know that this circumstance is not without a cause; but is, on the contrary, as really effected by the previous tossing which it receives in the hand or in the box, as its fall and the manner of its lying are by its gravity and figure. But the various turns or motions given it, in this manner, do inevitably escape my notice, and so are held for nothing. I say, therefore, that the chance is equal for every one of the six sides. Now if five of these were marked with the same figure, suppose a dagger (†), and only one with an asterisk (*), I should, in that case, say, there were five chances that the die would turn up the dagger, for one that it would turn up the asterisk; for the turning up each of the six sides being equally possible, there are five cases in which the dagger, and only one in which the asterisk, would be uppermost.

This differs from experience, inasmuch as I reckon the probability here, not from numbering and comparing the events after repeated trials, but without any trial, from balancing the possibilities on both sides. But, though different from experience, it is so similar, that we cannot wonder that it should produce a similar effect upon the mind. These different positions being considered as equal, if any of five shall produce one effect, and but the sixth another, the mind weighing the different events, resteth in an expectation of that in which the greater number of chances concur; but still accompanied with a degree of hesitancy, which appears proportioned to the number of chances on the opposite side. It is much after the same manner that the mind, on comparing its own experiences, when five instances favour one side, to one that favours the contrary, determines the greater credibility of the former. Hence, in all complicated cases, the very degree of probability may be arithmetically ascertained. That two dice marked in the common way will turn up seven, is thrice as probable as that they will turn up eleven, and six times as probable as that they will turn up twelve.* The * Call one die A, the other B. The chances for 7 are,

The only chance for 12 is A 6, B 6. The 1st is to the 2d, as 6 to 2; to the 3d, as 6 to 1.

degree of probability is here determined demonstratively. It is indeed true, that such mathematical calculations may be founded on experience, as well as upon chances. Examples of this we have in the computations that have been made of the value of annuities, ensurances, and several other commercial articles. In such cases, a great number of instances is necessary, the greatest exactness in collecting them on each side, and due care that there be no discoverable peculiarity in any of them, which would render them unfit for supporting a general conclusion.

PART IV. The Superiority of Scientific Evidence re-examined.

After the enumeration made in the first part of this section of the principal differences between scientific evidence and moral, I signified my intention of resuming the subject afterward, as far, at least, as might be necessary to show that the prerogatives of demonstration are not so considerable as, on a cursory view, one is apt to imagine. It will be proper now to execute this intention. I could not attempt it sooner, as the right apprehension of what is to be advanced will depend on a just conception of those things which have lately been explained. In the comparison referred to, I contrasted the two sorts of evidence, as they are in themselves, without considering the influence which the necessary application of our faculties in using both has, and ought to have, on the effect. The observations then made in that abstracted view of the subject appear to be well founded. But that view, I acknowledge, doth not comprehend the whole with which we are concerned.

It was observed of memory, that as it instantly succeeds sensation, it is the repository of all the stores from which our experience is collected, and that without an implicit faith in the clear representations of that faculty, we could not advance a step in the acquisition of experimental knowledge. Yet we know that memory is not infallible; nor can we pretend that in any case there is not a physical possibility of her making a false report. Here, it may be said, is an irremediable imbecility in the very foundation of moral reasoning. But is it less so in demonstrative reasoning? This point deserves a careful examination.

It was remarked concerning the latter, that it is a proof consisting of an uninterrupted series of axioms. The truth of each is intuitively perceived as we proceed. But this process is of necessity gradual, and these axioms are all brought in succession. It must, then, be solely by the aid of memory that they are capable of producing conviction in the mind. Nor by this do I mean to affirm that we can remember the preceding steps, with their connexions, so as to have them all present to our view at one instant; for then we should, in