writers who can pretend to no exemption from the failings and feebleness of our common nature, pour forth their confessedly wretched attempts at attaining truth by inference or deduction, into the ears of those who partake of the same logical incapacity?

If they are to be taken at their word, they are at the best engaged in nothing better than

Of dropping buckets into empty wells,
And growing old in drawing nothing up."

The course of the writers here alluded to, well exemplifies the evils arising from the personification of mental operations. If reasoning had not been spoken of as a faculty and if that faculty had not subsequently been personified, or erected into a separate entity, such futile invectives could not have been put forth. No reasoner would have been weak enough to directly disparage the process itself.

Further it is important to have a full and clear apprehension, while considering the two classes of facts above described, primitive facts and derivative facts, that the former do not admit of evidence at all; they are not susceptible of proof; or, in other words, they are not the subjects of inference. If direct evidence consists in known or primitive facts adduced to prove facts not actually known, the former class, it is plain, cannot themselves be the subjects of evidence.

Thus all our own mental states and conditions,

operations and affections, our sensations, emotions, acts of perceiving, of reasoning, of recollecting, of willing, are things felt or done or experienced by us; they are facts known to us, primitive facts ; and they neither require nor admit of evidence or proof to us who experience them.

This truth holds good of, or more correctly speaking comprehends, our perceiving external objects. That I see the trees and the grass and the flowers in the landscape before me, is to me a primitive fact which is not susceptible of proof or evidence. I know it.

The employment of evidence has thus nothing to do with primitive facts; it is legitimately confined to showing either that some event has happened or is happening although unperceived by us, or that some event will happen, which therefore must be equally unperceived.

When any one attempts to prove the existence of objects actually present to our senses, or more precisely speaking perceived through the organs of sense, he falls (on the most favorable supposition) into the absurdity of adducing known facts to prove others equally known; and when he attempts on the other hand to prove the nonexistence of such objects the self-contradiction, as I have shown elsewhere, if not equally manifest is not less real.*

* See Letters on the Philosophy of the Human Mind, First Series, p. 140, and Second Series, p. 132.

These conclusions seem almost too simple to require pointing out, yet they have been so utterly overlooked that philosophers in crowds have endeavoured to prove and others to disprove that we perceive external objects.

If they had duly reflected on the nature of evidence, they would have discerned the absolute futility of all such attempts.

Evidence having been thus shown to consist in known or already proved facts adduced to prove other facts, the next step is to consider in what property of the facts or the evidence the force of the proof lies; or, in different language, how some facts become or are fitted to become the proofs of other facts.

All facts it is plain cannot prove other facts indiscriminately. For one fact to have the capability of proving another there must be a connexion between the facts themselves or the classes of facts to which they belong.

This connexion is that of either causation or concomitance.

All facts adduced as evidence are and must be either causes or effects or concomitant circumstances.

When Robinson Crusoe (if I may resort to an example from fiction) was startled at seeing a foot-print on the sand in his solitary island, he instantly inferred that it had been made by a human being passing over the beach.

This was reasoning from the effect to the cause, or, to put the matter the other way, the print on the sand was evidence to him of the cause which had produced it — of the recent transit of a man.

On the other hand, should any one seeing a boat's crew about to land on the beach, predict from a previous knowledge of the consistence of the sand that they would leave their foot-prints upon it, he would reason from the cause to the effect, and the evidence to his mind would be facts before experienced of an analogous character.

In regard to the second kind of connexion, when concomitant facts like those here referred to are such as are always found together (for some facts are, I scarcely need to say, only casually conjoined) one is the evidence of the other just as in the case of causal succession, but to make them equal to the latter as proofs, they must be joint results of the same cause. Thus the fall of the mercury in a thermometer to zero, or to a certain point above it, and the freezing of the neighbouring pond, are concomitant effects of a great abstraction of heat from the atmosphere; and on seeing the state of the thermometer we may infer the state of the pond without taking the trouble of going to look at it. The state of the thermometer is not the cause of the state of the pond, nor, conversely, is the latter the cause of the former; they are concomitant results of one cause and serve to prove each other.

These are doubtless trite and simple facts and explanations, but since they lie in the course of the argument, their triteness and simplicity do not diminish their importance.

Having thus shown the nature of the connexion between facts by which they become, or may become, evidence of each other, I will proceed to point out the limitations of this function.

At the first glance it is obvious that any given cause does not produce all sorts of effects but one precise effect, and from that particular cause, consequently, it is only that precise effect which can be inferred. The converse equally holds.

Thus we are limited in our inferences to similar cases of causation.

If we have seen A produce B at one time we can infer when we meet with it again, that it will again produce B: we cannot infer that it will produce C.

From no cause can we infer legitimately a future effect unless we have known directly or indirectly a similar cause produce a similar effect to that which we infer; and the same truth holds mutatis mutandis when from effects we infer past causes and past concomitant events the results of the

same cause.

To put the matter the other way, we can have no evidence that an effect will happen from an assigned cause unless we have known similar effects to have happened from causes similar to