The great philosophers have tended to rely on one of two mutually exclusive approaches to the study of ethics – the first entirely empirical, the second entirely a priori. Epicurus, Mill, and Spencer took an empirical approach, while Kant took an a priori approach. The first begins with observation and proceeds by way of generalization and induction, reaching the pinnacle of its development with Mill. It consists in observing man, either when he's alone or with others; noting the circumstances in which he's happy; and then deriving the moral law by generalizing from these findings. Kant, by contrast, begins with the abstract concept of pure morality, assumes that the will is capable of acting independently of sensibility, and then asks what the law of this will must be.

But empiricism, no matter what degree of generalization it's able to attain, can never achieve the universality that characterizes the moral law. All empiricism can formulate are local and provisional rules, good only for a certain time and a certain number of individuals. Conversely, despite the fact that Kant made many concessions and scaled back the rigor of his initial formulas, his ethics remained imaginary, providing us with rules that an ideal and hypothetical being should follow – not man as he is.

The method we've followed, in contrast to empiricism and apriorism, is both deductive and experimental. We began by postulating a fact of experience – moral responsibility.