Kant argues that more principles must be valid for all rational beings in all circumstances. He further contends that actions are moral if and only if they are performed without ulterior motives, with no attention to consequences, and out of pure respect for morality. (Initially these propositions are based on general assumptions about morality, though in Chapter 3 Kant shows that they may be based on the concept of free will.) According to Kant, a priori concepts are the only possible basis for a formula that would meet these criteria. Any particular event or decision that we undergo will take place under particular circumstances; any tradition will depend on a particular history; any need or desire will depend on our particular personality. Only a priori concepts apply universally to all experiences of rational beings. Thus Kant concludes that the moral law must be derived a priori. Hegel and other philosophers have pointed out, however, that there are problems with this conclusion. In practice, moral decisions cannot be made a priori. They always take place within a particular society at a particular time, and moral intuitions must be based on social institutions and expectations. Without knowledge of the society in which we are living, we would not be able to tell whether our actions would help or hurt other people. (For more detail on this issue, see the Commentary on Chapter 1.)

Kant's first formulation of the categorical imperative is that we should act only on principles that we would want as universal laws. He writes that this formula could also be stated as a requirement that we act as though our action would make the principle of our action into a universal law of nature. Kant arrives at these formulations in seeking for some moral formula that could apply in all situations and circumstances. Only reason, he argues, can supply principles that are universally valid. When people act according to a principle that they would not want as a universal law, they contradict themselves, for they behave in a way that they would not want others to emulate. Self- contradiction is illogical, and so violates principles of reason. Kant's initial formulation of the categorical imperative provides a moral law based on this principle. From this initial formulation Kant derives his next formulation of the categorical imperative as a requirement that we never treat other people as mere means to our own ends. Kant argues that rational beings are "ends in themselves": they cannot view themselves as mere means to other purposes; rather, they always view themselves as the purpose of their actions. When we fail to respect the fact that other rational beings are ends in themselves just as we are, we advance principles that we would not want as universal laws, and we therefore contradict ourselves. Kant's final formulation of the categorical imperative follows easily from the earlier formulations. Kant's "kingdom of ends" is an ideal community in which all citizens are at once the authors and subjects of all laws. In this community, the only possible laws are laws that could apply to all rational beings. Thus the categorical imperative may be formulated as a requirement that we follow only those principles that could be laws in the kingdom of ends.

Kant defines freedom as the ability to give yourself your own law. Whenever we obey the demands of physical needs, desires, or circumstances, or whenever we make a decision that considers the probable consequences of our action, we receive our motivation from something other than ourselves, and we are not free according to Kant's definition. Freedom, he argues, is possible only in a condition of "autonomy"--that is, of depending only on reason for our motives and principles. The categorical imperative is Kant's litmus test for determining whether our moral principles conform to reason. Thus, according to Kant, we are free only if we obey the categorical imperative. This account of freedom makes sense within Kant's system of concepts, but it seems to exclude certain possibilities. Kant argues that we are in a state of "heteronomy" (and therefore not free) whenever we follow some impulse that does not come from reason. Yet if we are genuinely free, then we should be able to choose options other than reason.