CHAPTER II.

Of Processes subsidiary to Induction.

OF the various mental processes subsidiary to Induction proper, it will be sufficient for our purpose to discuss Observation and Experiment, Classification (including Nomenclature and Terminology), and Hypothesis.

§ 1. Of Observation and Experiment.

These words are now so familiar, that they hardly require any explanation. To observe is to watch with attention phenomena as they occur, to experiment (or, to adopt more ordinary language, to perform an experiment) is, not only to observe, but also to place the phenomena under peculiarly favourable circumstances, as a preliminary to observation. Thus, every experiment implies an observation, but it also implies something more. In an experiment, I arrange or create the circumstances under which I wish to make my observation. Thus, if two bodies are falling to the gound, and I attend to the phenomenon, I am said to observe it, but, if I place the bodies under the exhausted receiver of an air-pump, or cause them to be dropped under any special circumstances whatever, I may be said not only to make an observation, but also to perform an experiment. Bacon

has not inaptly compared experiment with the torture of witnesses'. Mr. Mill distinguishes between the two processes, by saying that in observation we find our instance in nature, in experiment we make it, by an artificial arrangement of circumstances. When, as in astronomy, we endeavour to ascertain causes by simply watching their effects, we observe; when, as in our laboratories, we interfere arbitrarily with the causes or circumstances of a phenomenon, we are said to experiment?!

As Observation often involves little or no conscious effort, while Experiment always implies an artificial arrangement of circumstances, it might be expected that the general employment of the former for scientific purposes would long precede that of the latter. And this supposition is confirmed by the History of Science. Though it is false to affirm that Experiment was never employed by the Greeks3, its general neglect was certainly one cause of the little progress made by them in the physical sciences.

In the attempt to ascertain the effect of a given cause, there can be no question of the general superiority of

1. Quemadmodum enim in civilibus ingenium cujusque, et occultus animi affectuumque sensus, melius elicitur, cum quis in perturbatione ponitur, quam alias: simili modo, et occulta naturæ magis se produnt per vexationes artium, quam cum cursu suo meant.' Nov. Org., Bk. I. Aph. xcviii.

2 Thomson and Tait's Natural Philosophy, vol. i. § 369.

* For a refutation of this popular misconception, see Mr. Lewes' work on Aristotle, ch. vi. Mr. Lewes, however, seems to me not sufficiently to recognise the slight extent to which Experiment was employed in ancient as compared with modern times.

Experiment over Observation. To be able to vary the circumstances as we choose, to produce the phenomenon under investigation in the precise degree which is most convenient to us, and as frequently as we wish to combine it with other phenomena or to isolate it altogether, are such obvious advantages that it is not necessary to insist upon them. Without the aid of artificial experiment, it would have been impossible, for instance, to ascertain the laws of falling bodies. To disprove the old theory that bodies fall in times inversely proportional to their weights, it was necessary to try the experiment; to be able to affirm with certainty that all bodies, if moving in a non-resisting medium, would fall to the earth through equal vertical spaces in equal times, it was essential to possess the means of removing altogether the resisting medium by some such contrivance as that of the airpump. In some of the sciences, such as Chemistry, the Sciences of Heat, Light, and Electricity, it is next to impossible, at least in their inductive stage, to advance a single step without the aid of Experiment. No amount of mere Observation would ever have enabled us to detect the chemical elements of which various bodies are composed, or to ascertain the effects of these elements in their pure state. Even when Observation alone reveals to us a fact of nature, Experiment is often necessary in order to give precision to our knowledge. That the metals are fusible, and that some are fusible at a lower temperature than others, is a fact which we can conceive to have been obtruded upon man's observation,

but the precise temperature at which each metal begins to change the solid for the liquid condition could be learned only by artificial experiment.

But, though, in ascertaining the effect of a given cause, Experiment is a far more potent instrument than Observation, the latter process is also available, and is frequently of the greatest service. Thus, the Science of Medicine equally avails itself, for this purpose, both of observations and experiments. The scientific physician will not only try the effects of different medicaments, different modes of diet, and the like, but he will also watch the effects on the organic system of various occupations, habits, and pursuits. In some cases even, as in all astronomical and many physiological phenomena, the only means open to us of ascertaining the effect of a given cause is Observation. If we wish to ascertain the various phenomena attendant on a shower of meteors or a total eclipse of the sun, we must wait till the shower of meteors occurs or the total eclipse takes place. If we wish to learn the effects of the lesion of a particular part of the nervous system, we must generally wait till an instance offers itself; there are many experiments too dangerous and too costly to be made, at least in the case of man.

While, however, both Observation and Experiment are available in ascertaining the effects of a given cause, in the reverse process of ascertaining the cause of a given effect, Observation alone is open to us. 'We can take a cause,' says Mr. Mill, and try what it will

produce; but we cannot take an effect, and try' [that is, experimentally], 'what it will be produced by. We can only watch till we see it produced, or are enabled to produce it by accident.' In those cases, consequently, in which effects alone are patent to us, and the causes are concealed from our view, we are compelled, unless we are able to reverse the problem in the manner noticed in the next paragraph, to have recourse to Observation. A new disease makes its appearance: the mode of its action, and the conditions favourable or unfavourable to its diffusion, can only be learned by a careful observation and comparison of cases.

It should, however, be noticed that the problem of finding the cause of a given effect is, in practice, as, for instance, in many cases of chemical analysis, often reversed, and that, by setting in action a variety of causes, we try to discover whether any one of them will produce the effect in question. Experiment is thus substituted for Observation.

It will readily be seen that those Sciences which depend wholly or mainly on Observation are, as inductive sciences, at a great disadvantage compared with those in which it is possible largely to employ Experiment. Where we wish to ascertain the effect of a given cause, and we cannot make the instances for ourselves, the want of appropriate and definite instances will often completely baffle us. And, though the cause of a given effect can only be learned by Observation, this is generally an enquiry of extreme difficulty, requiring to be