In trying to find some unity in the morassy spawn of unconnected right-making characteristics and deontic rules, the 'categorical imperative' and 'principle of universalizability' have been proposed. According to the categorical imperative one should 'only act on a maxim which one can at the same time will to be a universal law'. This imperative was designed to establish a monistic kind of rule-deontology. (Rule-deontology holds that moral judgments should be at least implicitly based on nonconsequentialist rules or maxims.) According to the principle of universalizability (already mentioned in 5.1.2) one should be able to universalize one's maxim, that is, if x is right, then anything exactly like x in relevant respects must also be right. This principle tho is nothing else than a principle of relevance turned upside down: the moral agent starts with the distinctions made in the maxim and is asked to inquire subsequently whether these distinctions are perhaps irrelevant, in which case the maxim is not universalizable, or is lost when universalizing it.

To illustrate the impracticability (and naivity) of an ethical guideline such as the categorical imperative, let us consider a biracial community with races R and S, and let us take two individuals in this community: A and B. A loves B, and expresses this love in a physical way towards B. How to describe this action so that it might fall under the categorical imperative or some other rule? It might be said that 'A has physical contact with somebody of the opposite race' if A belongs to race R, and B to race S, or vice versa. (Opposite is then used in the simple antonymical sense, when A and B have never heard and thought of other races than R and S.) Secondly, it might be said that 'A has physical contact with somebody of race S' if B belongs to race S. But thirdly, it could also be claimed that 'A has physical contact with somebody 'e loves' (or 'just likes to have physical contact with'). Now, what maxim does the categorical imperative want A to will to be a universal law? Should A wonder whether 'e can will that everybody is allowed to have physical contacts with person or body B; with a human being of the opposite race; with somebody of race S; or with somebody 'e loves or likes? And if A rejects the idea that all members of 'er community should exclusively express their love towards B, or towards members of race S (so that no-one expresses 'er love towards members of race R), should 'e then consider it 'er duty not to love B, or not to express 'er love in a physical way? Naturally, the right answer is somehow that A is allowed to express 'er love towards B in a physical way if B does (at least) not mind, whether 'e be of the same or of the 'opposite' race, of race R or of race S. The categorical imperative, however, is useless in ascertaining that this is the correct description of the act or maxim in question. What is worse, it has been demonstrated that a 'suitable' description can always be found of any act. (It will be fun to apply the categorical imperative to someone who has decided to practise philosophy as a profession. And what if such a philosopher tells a conscientious objector that if everybody in 'er country refused to join the armed forces nobody could defend 'er country, and the conscientious objector replies that if every moral agent in the world refused to join the armed forces no foreign power could and would attack 'er country to start with?)

In the first instance it seems plausible to infer from the categorical imperative or the principle of universalizability that one should not lie, and that the injunction not to lie is the rule with respect to the truth of one's statements. For if everyone lied, or lied when it suited 'im, lying would not be possible anymore, since a lie can only work in a social environment in which most people expect someone to tell the truth. Yet, if a liar acts on a maxim at all, that maxim need not be that 'e lies or ought to lie. It might be, for example, --as has been argued-- lie when it is the sole way to avoid harming someone or lie when it is entertaining or harmless. Not only is it possible to get around the categorical imperative by employing other or more detailed descriptions of the act, in everyday life implications are also easily avoided by a change in meanings of the terms employed: if telling a (natural or supernatural) falsehood is entertaining and harmless, it is no deceit, and 'may' (or 'must') not even be called "a lie"; and if no-one expects someone to do something, altho 'e has said so, it may not even be called "a promise".

It should have become clear that deontological theories, like all ethical theories, need at least an implicit moral decision-theory, an implicit principle of truth and an implicit principle of relevance. It should have become clear, too, that deontology is doomed to remain a very pluralistic form of ethics with all the ensuing difficulties (or conveniences?) of conflicting ultimate duties. Conflicting duties or values are inherent in every normative doctrine with truth as a principle besides other ones, but when they can be reduced to the smallest number of ultimate duties or values, they can, at least in principle, be solved or avoided. The deontological agent who promises the authorities to kill the first living being 'e will meet on 'er return home, but who simultaneously has the duty not to kill, for example, somebody belonging to 'er own in-group, has no way whatsoever to determine what to do when the first living being 'e meets on 'er return home happens to be somebody of 'er own in-group. To assert that killing somebody of one's own in-group must be worse than breaking a promise, presupposes a higher or more general standard of appraisal (happiness maybe?), and presupposes that there is a higher-level duty to do what gives a better result, or not to do what gives a worse one. To say that killing somebody is worse than breaking a promise is to go by some higher-level principle. If so, then it ought to be revealed, even when it is merely that of adherence to the factual morality of the past and present (a 'principle' all lexical orderings of ultimate duties, values or rights seem to be subjected to). In the event that the deontologist is not capable of doing this, it is probably because there is no normative one on the deontological reckoning. But then it should be underscored that no conflict of duties may ever be created or given rise to, however many ultimate duties may be believed to exist.