and in some degree, and every real, ardent temperament will be unstable in some degree, in some direction. But the same is true also of the nervous and phlegmatic. These, too, are at opposite poles, and somewhere between them all existing temperaments must be found. It follows that every sanguine temperament must also be either nervous or phlegmatic, or belong to some intermediate and more indefinite variety; and this must be also true of every ardent temperament. And every nervous temperament must be either sanguine or ardent, or belong to some intermediate and more indefinite variety; and this must be also true of every phlegmatic temperament. It follows, therefore, since each of the pure types is an exaggeration resulting from the assumption that all the mental processes possess that high degree of a certain quality which has been observed in some of them, that all existing temperaments combine in varying proportions all the four so-called pure temperaments. Thus the most unstable temperament we can find, will have-after eliminating the greater degree of instability which has been due to insufficient discipline in childhood and youth or has been acquired by an unregulated life—some qualities in which it is less unstable and approximates nearer to an ardent type, and if it is also quick and nervous, there will be some qualities in which it is not so quick, and approaches nearer to the phlegmatic type. Thus whatever approach to clearness and precision the different temperaments may seem to possess in their artificial exclusion of one another will be lost in the confusion of their real combinations.

There are other defects of the classical doctrine. It does not lay down any tentative laws such as might be available to us in a scientific treatment of character. It is a barren classification, like those classifications of the emotions that were til recently so prevalent. It does not justify our inferring anything of the temperament of a given individual beyond what we have observed it to be; and it induces us to make erroneous inferences, and is therefore an impediment to a scientific interpretation of character. These several assertions we must now examine.

In the ordinary descriptions of men's characters we do not

lay claim to exactitude. We say that a man is silent, reserved and unsociable, or that he has little capacity for affection, is ungrateful and an infliction to those who love him. Such statements are not scientific; but they have a certain practical value. They teach us to anticipate some of the actions of the same individual in the future. To classify a man under one of the four temperaments may be supposed to have the same sort of practical value. After some observation of a given individual we conclude that he is superficial in certain directions; if we class him as sanguine, does that give us grounds for inferring anything about him that we did not know before? It encourages us, indeed, to believe that he will be superficial in all other directions because that is the character of the type. But before making such an inference we ought to have a very wide knowledge of the man. The doctrine of the temperaments, however, does not justify our anticipating or abridging this knowledge, not only because its assumptions are unscientific, but because the temperaments of men are “mixed." doctrine of the temperaments does not then give us grounds for extending our inferences beyond those limits which our observation of the individual in question justifies. And barren of good results, it encourages us to make those rash inferences to which the human mind is already too prone.

The

If we consider another quality of the sanguine temperament, it will reinforce the same conclusion. When we class a man as sanguine we not only inter that he is generally superficial, but that he has a natural fund of good spirits, that he is indisposed to the depressing emotions, and is, as it is said,, a "born optimist." That some empirical connection has been found to exist between these characteristics may be admitted, and it is here, if anywhere, that the classification should help us by extending our knowledge of the man, and furnishing us at least with some tentative laws of character. But we are here in the same difficulty as we were formerly with regard to the man's superficiality. we have already inferred from our own observation that his disposition is cheerful, we do not require the doctrine of the temperaments to inform us of that which we knew already

L

If

But if we have not made this inference, because we have not extended our observations in that direction, this doctrine does not justify our making it, because the temperaments of men are "mixed." A man may have one of these characteristics without the other.

In the remarkable analysis which Hume gave1 of his own character when he felt his death approaching, he states that he was "naturally of a cheerful and sanguine temper," more disposed to see the favourable than the unfavourable side of things," of which he furnishes several eminent instances: first that when his great Treatise on Human Nature "fell dead-born from the press," he "soon recovered from the blow"; secondly, that his temper was never soured notwithstanding his "frequent disappointments"; but thirdly and most significantly that when he was overtaken by disease, and knew that his end was approaching, he "never suffered a moment's abatement" of his spirits, so much so, he says, that "were I to name a period of my life which I should choose to pass over again, I might be tempted to point to this later period. I possess the same ardour as ever in study, and the same gaiety in company." Here then we have a striking instance of a sanguine temperament proof against what is reckoned the greatest misfortune: here, if anywhere, we ought to be able to infer that the individual in question possessed also the other principal quality of this temperament-superficiality of character. But what are the facts? He declares that he had one "ruling passion," the "love of literary fame," which he pursued with ardour through the greater part of his life, and up to its closing scene. If this great writer had then a sanguine temperament, he had also the essential quality of the opposite or ardent temperament. His most recent biographer describes him as "A man of placid and even phlegmatic temperament."2

The doctrine of the temperaments seems then to afford us no valid grounds for extending our inferences concerning a given individual beyond those limits which our particular knowledge of him justifies. It does not lay down any 1 'My own Life.'

2 'Encyclopædia Britannica,' xi. ed. vol. 13, Art., 'Hume.

provisional laws of the conjunction of qualities in a type that that we can rely on. Its classification of individuals is barren of good results; and its doctrine throughout is pervaded by rash and unscientific inferences.

The examination of this classical doctrine, though it has brought us only to negative results, has seemed necessary, because almost all recent works which have attempted to treat of a science of character have been based upon it; and a doctrine which has lasted more than 2,000 years, and been praised by eminent philosophers, physiologists, and physicians, may well seem deserving of careful examination. But this doctrine though so many modern writers have essayed to give it a scientific form, has never been lifted out of that loose and popular status to which its original defects condemned it, and seems to be incapable of fruitful development. In the next chapter we shall return to the consideration of the natural or innate tempers with which we commenced this chapter, though with the knowledge available at present we cannot do more than indicate what has to be accomplished.

CHAPTER XIV

OF THE INFLUENCE OF THE NATURAL TEMPERS ON THE

STABILITY OF SENTIMENTS (I)

1. How far the Temper of one Emotion involves a

corresponding Temper in other Emotions

THE study of the 'tempers,' in the sense given to the term in the last chapter, seems to be the only base on which we can hope to found a scientific theory of the temperaments, because here we are confined at first to a single emotion, and need not assume that the quality of its temper attaches to other emotions in the same individual. And our task will be to call in question such an assumption, and to ask how far and in what directions the quality of one emotional temper can be inferred to attach to other emotions. For instance, there are those who are very sensible to fear; and women are held to be more timorous than men. But does this natural or innate sensibility to fear carry with it a corresponding sensibility to anger? We should not suppose that it does, but rather that it restricts the sensibility to such an opposite emotion. Men, on the other hand, are supposed to be more sensible to anger than to fear. The males among many animals have to fight in their own defence, for the possession of the females, in defence of them and of the young, and to procure food. Thus the sensibility to fear is probably not the same as the sensibility to anger in either sex; and the natural temper of the one emotion cannot be inferred to attach to the other.

The sensibility to an emotion we shall take to mean the