A hedonistic method, indeed, that would dispense altogether with direct estimates of the pleasurable and painful consequences of actions is almost as inconceivable as a method of astronomy that would dispense with observations of the stars. It is, however, conceivable that by induction from cases in which empirical measurement is easy we may obtain generalisations that will give us more trustworthy guidance than such measurement can do in complicated cases; we may be able to ascertain some general psychical or physical concomitant or antecedent of pleasure and pain, more easy to recognise, foresee, measure, and produce or avert in such cases, than pleasure and pain themselves. I am willing to hope that this refuge from the difficulties of Empirical Hedonism may some time or other be open to us: but I cannot perceive that it is at present available. There is at present, so far as I can judge, no satisfactorily established general theory of the causes of pleasure and pain; and such theories as have gained a certain degree of acceptance, as partially true or probable, are manifestly not adapted for the practical application that we here require. The chief difficulty of finding a universally applicable theory of the causes of pleasures and pains is easily explained. Pleasures and pains may be assumed to have universally--like other psychical facts--certain cerebral nerve-processes, specifically unknown, as their inseparable concomitants: accordingly, we may seek their causes either in antecedent physical or antecedent psychical facts. But in one important class of cases the chief cognisable antecedents are obviously of the former kind, while in another important class they are obviously of the latter kind: the difficulty is to establish any theory equally applicable to both classes, or to bring the results of the two lines of inquiry under a single generalisation without palpably unsupported hypotheses. In the case of pleasures and pains--especially pains--connected with sensation the most important cognisable antecedents are clearly physical. I do not deny that, when the pain is foreseen, the attitude of mind in which it is met may materially influence its magnitude: indeed, in the hypnotic condition of the brain, the feeling of pain may be apparently altogether prevented by an antecedent belief that it will not be felt. Still in the main, under ordinary conditions, the pains of sensation--probably the intensest in the experience of most persons--invade and interrupt our psychical life from without; and it would be idle to look for the chief causes of their intensity or quality among antecedent psychical facts. This is not equally true of the most prominent pleasures of sense: since antecedent desire, if not an absolutely indispensable condition of such pleasures, seems at any rate necessary to their attaining a high degree of intensity. Still the chief causes of these desires themselves