against his aiming at his ultimate ideal of framing a complete mental reproduction of the entire course of events from past to future.

Strictly speaking however there is such an inconsistency necessarily involved: it presents itself in the following way. Look at that complex of phenomena which constitutes the logician's world, with all its aggregate of objects which furnish his examples and illustrate his proofs. These are certainly not supposed to be confined to material objects, but must equally include the thoughts, feelings, and actions of human beings; // for every event without exception which we can suppose ourselves observing may become a logical element. It may stand as a subject or predicate, and it may give ground to an inference. This catholicity of application is in strictness true of every system of Logic, for even on the narrowest formal view of the science we may draw our examples freely from the conduct and character of our fellow men. But in a system of Inductive Logic, especially when this embraces the so-called Sociology, we are much more largely concerned with the doings of men, and the inferences we can draw as to their conduct. Now the moment we do this we find ourselves confronted by a troublesome question. The agents whose performances are thus supposed to be a part of the object world of our logician: are they themselves supposed to be logicians? and if so how can they simultaneously occupy the position of observer and that of being the subject of observation? Any strict view of the logician's stand-point,-when, as now, we are defining it with the utmost accuracy,-is certainly inconsistent with such a supposition. He is assumed to take up a contemplative, not an active position. He has to stand aloof from the phenomena in order to observe, judge, and infer. He must not simultaneously try to form a part of his own observations and inferences; for if he does he will almost certainly introduce a disturbance into them which will invalidate the inference.

It must be admitted that so far as the direct and actual performances of the observer are concerned, the inconsistency here indicated produces no serious results. The department of speculation in which it does give rise to real difficulty is that in which Hypothesis, in its widest signification, has to be resorted to. And it is mainly with this reference in view that

I now insist upon the difficulty. In a future chapter, which will be devoted to the discussion of Hypothesis in general, it will be pointed out that,-not merely in Ethics, but in what may be called the Science of Human Conduct in general,—this unauthorized transference of himself, by the observer into the midst of his observations, is very difficult to exclude, and causes serious logical inconsistency. Within the domain of Physical Science, and over the range of the larger part of human action of the ordinary kind, the difficulty in question is admittedly but slight, and it claims attention here rather from the desirability of complete scientific accuracy of definition than for the purpose of avoiding any actual mischief or error.

At the same time, as we are upon the subject, it is just worth pointing out that the complete attainment of the ideal position of the mere observer is nowhere to be secured even in Physics. Take, for example, the most extreme case, where this position may seem to be most completely securable, viz. the science of Astronomy. Here, if anywhere, the observer might conceive himself standing entirely apart from the objects whose motion he calculates; picturing mentally their career without interfering physically with it. He would claim, apparently with good reason, that he merely watches what they do, and that as he cannot possibly experiment, he cannot in the slightest degree interfere with their motions. No more he does, so far as any results are concerned which the utmost attainable refinement of observation, or indeed any refinement vastly beyond what is attainable, could ever detect. But this does not hold if we like to take account of influences which are undeniably real, though so immeasurably minute that it would be absurd to notice them except by way of illustrating a point of theoretic interest. It cannot be denied then that if the Law of Universal Gravitation is rigidly true the calculator does influence the course of the planets themselves, and does so by the fact of observing them. Every motion to or from his instrument, nay the very calculations he writes down on paper or the words he utters by his voice, are motions of matter, and therefore react on the motions of every other material thing in the universe, including the planets themselves.) Accordingly in calculating their motions as a passive spectator he is in perfect strictness disturbing those very paths which he had calculated, and consequently falsifying

his own conclusions. This impossibility then of complete isolation of observation, existing thus as a speculative truth even in the case of those objects which are physically the most remote from us, assumes more and more of a practical aspect as we proceed in the direction of volitional human actions, and it reaches (as we shall see hereafter) a climax when these latter are treated not as actual but as hypothetical.

There seems to me to be only one way of meeting this difficulty so as to make our position speculatively free from inconsistency. We must start with a fiction which may as well be definitely stated as one of the postulates of Logić. No living human being can be spared to occupy that purely speculative position which is wanted for our logician. Each one of us has his own position amongst the objects which compose the world; he has his own little sphere of activity which he may change only by taking up some other. No one of us can be spared to occupy the ideal logician's seat; and if he try to do so he would find that he was perpetually leaving it, and mixing himself up in some way or other in the course of what should have been to him a wholly external world. What therefore we have to do seems to be this. We have to assume a sort of representative mind, distinct from any one of ours, but endowed with the same conceptions (and of course laws of inference) as we at present possess ourselves. For such a mind as this the ideal position of absolute non-interference with the objects before it, which is denied to any of us, could be rigidly preserved. And when the logician claims, as he sometimes explicitly does, that he has no other function than to observe and judge and infer, he must in consistency create for himself such a fictitious post as this.

It may seem to some readers little else than a waste of time to have enunciated and discussed two such postulates as those which have just been laid down. I have however adopted this course deliberately; partly in order to secure perfect speculative consistency, and partly also to prepare the reader for the extreme importance of that general view of Logic in pursuance of which it becomes necessary to remove even such apparently far-fetched difficulties as these out of our path. I proceed to add some pages of illustration of this general view under a new head.

V. What we have to take for granted in Logic is, then, a

| duality, external and internal. On the one hand, outside us, there is the world of phenomena pursuing its course; and, on the other hand, within us, there is the observing and thinking mind. Logic is concerned with the judgments of the latter) about the former. The entire omission of either of these two elements, if indeed such were possible,-would involve the destruction of the science, as any undue stress upon either leads to confusion and to inconsistency. The thorough-going retention of this duality is one of the leading characteristics of the whole treatment adopted in this work. Its extreme importance will only gradually be appreciated as one doctrine after another comes up for discussion, and as we find ourselves influenced in our decision about each by the principle in question, but a slight sketch indicative of its significance may conveniently be given

at once.

Logic then as here conceived is neither a purely objective nor a purely subjective science. It involves both elements, consisting essentially in the relation of one to the other, and serious error results from the neglect of either aspect, and even from insufficient recognition of it.

Consider, for instance, what would follow if we were to propose to drop the mental or subjective side. Such a proposal has been made, and has even been incorporated into the definition of the science. Thus Mr H. Spencer lays it down that 1 Logic formulates the most general laws of correlation among existences considered as objective", that "Logic, instead of being a science of certain subjective correlations is a science of certain objective correlations'." Strictly maintained, such a view as this would confine us to a bare statement of those objective laws or regularities which lie at the base of all inductive inferIt would deal with exactly the same subject-matter as that with which each of the special physical sciences is concerned, though it would be more comprehensive in its range than any one of these, covering in fact the ground common to them all. Just as each special science treats the laws distinctive of that group of objects which assigns its unity to the science in

ence.

1 Principles of Psychology, Vol. п. pp. 87, 100. The difference between us however is by no means so sharp as these passages would indicate, since Mr Spencer proceeds to make a distinction between Logic' as objective and the 'Science of Reasoning' as subjective.

question, so, it may be urged, can we treat of the uniformities common to the sciences in general, and regard this as a single objective science, viz. Logic. In both cases alike our faculties of observation and reasoning have to be taken into account, but they are only recognized tacitly or indirectly. Their existence is not expressly noticed except when it is considered that they are likely to become sources of error and confusion.

It appears to me that such a view as this is altogether insufficient, and would, if consistently adhered to, lead to the rejection of most of what has always been regarded as forming a part of the subject-matter of Logic. It seems indeed obvious that any attempt to confine ourselves to a bare statement or analysis of facts of nature must be insufficient when what we are concerned with is inference about those facts for inference turns almost entirely upon the distinction between what is known and what is unknown, and this distinction cannot be sought in the facts but in our appreciation of them. I quite admit that all science involves this element, but it does so indirectly; it does not make this element its express subjectmatter. For instance any treatise on Astronomy must involve certain relations to the current standard of attainment and information at the time. It will not state what is already perfectly familiar to every one, and it cannot state what is unknown to any, so it deals mostly with what has been comparatively recently acquired. To this extent the purely objective treatment is conditioned by subjective considerations; but such a reference I consider as subordinate and indirect.

Now when we turn to Logic we find that our treatment is conditioned by such considerations in a very different sense; for the current, or even personal state, as regards knowledge, is not here an inevitable accident but constitutes a part of our subjectmatter. Look, for instance, at the distinction between the 'essential' and the 'accidental' attributes of anything, upon which the whole significance of Connotation depends, including amongst its consequences the doctrine of Definition which has always formed one of the central parts of every system of Logic. If we objectify too much we simply annihilate this distinction. Beyond a doubt the essential and the accidental attributes are both, as the phrase runs, 'in the facts', but the distinction between them must be sought, not there, but in our estimate

V