I am aware, indeed, that, in the consideration of the simple antecedents, and consequents which nature exhibits, it is not the mere complication of these, by the introduction of new intervening substances or events, which obtains from the mind so ready an adoption of hypotheses. On the contrary, there is a sort of false simplification in the introduction of hypotheses, which itself aids the illusion of the mystery. I term the simplification false, because it is not in the phenomena themselves, but in our mode of conceiving them. It is certainly far more simple, in nature, that bodies should have a tendency toward each other, than that there should be oceans of a subtle fluid, circulating around them, in vortices, – or streams of such a fluid, projected continually on them from some unknown source, merely to produce the same exact motions, which would be the result of the reciprocal tendency in the bodies themselves. But the interposition of all this immensity of matter, to account for the fall of a feather or rain-drop, cumbrous as the contrivance must be allowed to be, is yet in one respect, more simple to our conception, because, instead of two classes of phenomena, those of gravitation and of impulse, we have, in referring all to impulse, only one general class. Man loves what is simple much, but he loves what is mysterious more; and a mighty ocean of ether, operating invisibly in all the visible phenomena of the universe, has thus a sort of double charm, by uniting the false simplification, of which I have spoken, with abundance of real mystery. This mixture of the simple and the mysterious, is, in some measure, like the mixture of uniformity with diversity, that is so delightful in works of art. However pleasing objects may separately be, we are soon wearied with wandering over them, when, from their extreme irregularity, we cannot group them in any distinct assemblage, or discover some slight relation of parts to the whole; and we are still sooner, and more painfully fatigued, when every object which we see is in exact symmetry with some other object. In like manner, the mind would be perplexed and oppressed, if it were to conceive a great multitude of objects or circumstances, concurring in the production of one observed event. But it feels a sort of dissatisfaction also, when the sequences of events which it observes, are reduced to the mere antecedents and consequents of which they consist, and must have a little more complication to flatter it with the belief, that it has learned something which it is important to have learned. To know that a withered leaf falls to the ground, is to know, what the very vulgar know, as well as ourselves; but an ocean of ether, whirling it downward, is something of which the vulgar have no conception, and gives a kind of mysterious magnificence to a very simple event, which makes us think, that our knowledge is greater, because we have given, in our imagination, a sort of cumbrous magnitude to the phenomenon itself.
That hypotheses, in that wide sense of the word which implies every thing conjectural, are without use in philosophy, it would be absurd to affirm, since every inquiry may, in that wide sense, be said to pre-suppose them, and must always pre-suppose them if the inquiry have any object. They are of use, however, not as superseding investigation, but as directing investigation to certain objects, – not as telling us, what we are to believe, but as pointing out to us what we are to endeavour to ascertain. An hypothesis, in this view of it, is nothing more than a reason for making one experiment or observation rather than another; and it is evident, that, without some reason of this kind, as experiment and observations are almost infinite, inquiry would be altogether profitless. To make experiments, at random, is not to philosophize; it becomes philosophy, only when the experiments are made with a certain view; and to make them, with any particular view, is to suppose the presence of something, the operation of which they will tend either to prove or disprove. When Torricelli, for example, – proceeding on the observation previously made, by Galileo, with respect to the limited height to which water could be made to rise in a pump, – that memorable observation, which demonstrated, at last, after so many ages of errors, what ought not for a single moment to have required to be demonstrated; the absurdity of the horror of a void ascribed to nature – when, proceeding in this memorable observation, Torricelli made his equally memorable experiment with respect to the height of the column of mercury supported in an inverted tube, and found, on comparison of their specific gravities, the columns of mercury and water to be exactly equiponderant, it is evident that he was led to the experiment with the mercury by the supposition, that the rise of fluids in vacuo was occasioned by some counterpressure, exactly equal to the weight supported, and that the column of mercury, therefore should be less in height than the column of water, in the exact inverse ratio of their specific gravities, by which the counterpressure was to be sustained. To conceive the air, which was then universally regarded as essentially light, to be not light but heavy, so as to press on the fluid beneath, was, at that time, to make as bold a supposition as could be made. It was indeed, a temporary hypothesis, even when it led to that experimental demonstration of the fact, which proved it forever after not to be hypothetical.
An hypothesis, then, in the first stage of inquiry, far from being inconsistent with sound philosophy, may be said to be essential to it. But it is essential only in this first stage, as suggesting what is afterwards to be verified or disproved; and, when the experiments or observations to which it directs us do not verify it, it is no longer to be entertained, even as an hypothesis. If we observe a phenomenon, which we never have observed before, it is absolutely impossible for us, not to think of the analogous cases which we may have seen; since they are suggested by a principal of association, which is as truly a part of our constitution, as the senses with which we perceived the phenomenon itself; and, if any of these analogies strike us as remarkably coincident, it is equally impossible for us not to imagine, that the cause, which we knew in that former instance, may also be present in this analogical instance, and that they may, therefore, both be reduced to the same class. To stop here, and, from this mere analogy, to infer positive identity of the causes, and to follow out the possible consequences in innumerable applications, would be to do, as many great artists in systematizing have done. What a philosopher, of sounder views, however, would do in such a case, is very different. He would assume, indeed, as possible or perhaps as probable, the existence of the supposed cause. But he would assume it,