Page:Popular Science Monthly Volume 28.djvu/261

Rh, and that therefore the unexplained errors of Uranus were due to the action of an exterior planet. But this assumption was as different as possible from a postulate: it was only applying in a new way a law which had already been verified in so many and such diverse cases that there was scarcely the shadow of a doubt in the mind of any astronomer that it was, as its ordinary name professes it to be, universal throughout the material cosmos.

I am confirmed in this belief by finding the subsequent statement that "the uniformity of Nature is a working hypothesis, and it never can be more"; which agrees very much with the view propounded by Professor Huxley at the meeting of the Metaphysical Society. But I am not quite sure that this is consistent with a previous passage in the lecture, which runs thus:

This, then, is the answer to the question. "Why do we believe in the uniformity of Nature? We believe in it because we find it so. Millions and millions of observations concur in exhibiting this uniformity. And, the longer our observation of Nature goes on, the greater do we find the extent of it. Things which once seepied irregular are now known to be regular. Things that seemed inexplicable on tills hypothesis are now explained. Every day seems to add not merely to the instances, but to the wide-ranging classes of phenomena that come under the rule.

The truth of which I am not concerned to dispute; but the paragraph gives a very different complexion to the principle of the uniformity of Nature from that which belongs to it, when regarded as a postulate upon which all scientific knowledge depends.

The truth which I think is postulated in the case of Nature is that which is involved in the idea of cause and effect. The Bishop of London refers to Hume's famous discussion of this question, and his conclusion that there is nothing more in cause and effect than the notion of invariable sequence. This conclusion has often been controverted, and the Bishop of London refers to the arguments of Kant and of J. S. Mill: it seems to admit of a very simple and irresistible contradiction from the following consideration: It is easy to give instances in which an invariable sequence takes place, and yet the two events which follow each other are obviously not connected as cause and effect. Take the case of lightning and thunder: the thunder follows the lightning with invariable sequence, whether we chance to hear it or not, but the two are separate effects of the same cause acting under different conditions; and no rightly instructed person could imagine that one was the effect of the other. Or suppose that you shout, and produce two echoes from two rocks at different distances; these echoes will satisfy the condition of invariable sequence, and yet will manifestly not be related as cause and effect. Or, to put the case more generally, it is quite possible that a cause may produce more than one