deformed," has a definite meaning, though we may be unable to define the precise meaning of our words, or to decide which statement is true in particular cases. The great difficulty in the case of insanity corresponds to this. The physician tells us that the madman's mind works abnormally, but not abnormally in the sense merely of having some faculties weaker and others stronger than is common; but in such a way as to indicate disease, and, moreover, a particular kind of disease, or one, perhaps, of several particular kinds of diseases. The vagueness of this statement provokes lawyers, who have a natural love of definite external tests to govern their decisions; and it has led to a number of delicate discussions, upon which I need not enter. The legal problem seems essentially to be, what tests should guide us in determining whether a man should be regarded as a normal human being, or as a being so far differing from the normal type that he should be treated exceptionally, and especially put under the guidance of other persons, and excused from legal responsibility, that is, liability to punishment. I have to do with the moral problem alone. It is a still more difficult problem; but it has this advantage, that we do not require so definite an answer. We have not, happily, to decide whether our fellows shall go to heaven or to hell, though we have to decide whether they shall be hanged or locked up; and we must be content as a rule with very vague estimates as to their moral character. What we practically have to take, more or less roughly, into account is simply this: that our inference from conduct to character has often to be modified by the existence of these abnormal cases. A man is drunk on an important occasion; I infer, as a rule, that he has all the qualities which go with low sensuality; but in some cases the inference is wrong; the man may be really a person of most admirable feelings; but one of his instincts has suddenly taken an abnormal development, owing to a set of causes entirely different from the usual causes. Another man suddenly and causelessly kills a friend. The natural inference that he must be a bloodthirsty brute is erroneous, if it turns out that he has acted from impulses not generated by any habitual want of benevolence, but from some special defect in the constitution of his brain. In other words, our moral judgment must vary in the two cases, and may vary so much that the same action may rightly suggest only pity in one case and abhorrence in the other; although, in many cases, where it may be very difficult to say what is the precise implication as to character, the judgment must, if we are properly diffident, remain obscure. The moral problem always depends ultimately upon this: What is the character implied by this conduct? If the moral conduct shows malignity within the normal type, it justifies condemnation; if it shows only a blind instinctive impulse, due to a