Now when we ask for the reasons in favour of the actions which moralists recommend, these reasons are, usually, that the consequences of the action are likely to be good, or if not wholly good, at least the best possible under the circumstances. Hence all questions of conduct presuppose the decision as to what things other than conduct are good and what bad. What is called good conduct is conduct which is a means to other things which are good on their own account; and hence the study of what is good on its own account is necessary before we can decide upon rules of conduct. And the study of what is good or bad on its own account must be included in ethics, which thus ceases to be concerned only with human conduct. (§ 3 ¶ 1)

The first step in ethics, therefore, is to be quite clear as to what we mean by good and bad. Only then can we return to conduct, and ask how right conduct is related to the production of goods and the avoidance of evils. In this, as in all philosophical inquiries, after a preliminary analysis of complex data we proceed again to build up complex things from their simpler constituents, starting from ideas which we understand though we cannot define them, and from premisses which we know though we cannot prove them. The appearance of dogmatism in this procedure is deceptive, for the premisses are such as ordinary reasoning unconsciously assumes, and there is less real dogmatism in believing them after a critical scrutiny than in employing them implicitly without examination. (§ 3 ¶ 2)