Discrimination by people against, or in favor of, people or sentient beings has become one of the main themes of normative philosophy in general, and of ethics in particular. An ultimate definition of discrimination cannot be given without referring to the irrelevant distinction made when discriminating, even if one is willing to agree that all cases of making an irrelevant distinction may be called "cases of discrimination". Curiously enough, the same holds for other essential ethical concepts, such as fairness, (distributive) justice, universalizability and equality. The (practical) significance of these concepts does also crucially depend on what we (or our opponents) believe relevant to be, and subsequently, to be relevant or morally relevant. The phrases relevant respect(s), relevantly similar or different and morally relevant appear throughout the ethical literature on the above-mentioned subjects. Yet, while relevancy (or relatedness) has been recognized as a key notion in several other fields of thought, it has not at the same time received any comparable attention in ethics. This applies even to moral relevancy.

A philosopher may complain at one place that relevance is too vague a criterion to be of any use and 'plainly does not work' as a criterion for distinguishing sense from nonsense, while using this very criterion to distinguish the fair from the unfair at another place. There 'e may argue that it would be unfair if one person 'consistently obtained more', or owned more property, than another person 'with the same, or sufficiently similar, relevant characteristics', or than another person 'situated in relevant respects' as the former one. But relevant or irrelevant have to be used somehow to differentiate the fair and the unfair, or the just and the unjust. It is indeed injustice to treat two similar individuals in similar circumstances in a different way, that is, the one better than the other, unless --as already pointed out by others-- the 'agent or group can establish that there is some relevant dissimilarity nonetheless between the individuals concerned and their circumstances'. It has also been said by a philosopher that one ought not to judge cases differently 'which are not relevantly different', that one ought not to make unjustifiable exceptions in favor of oneself. The inherent suggestion in such an argument is that an exception is not justified when the difference made is not relevant. The rules of justice themselves have been described as 'rules of making judicial and other, analogous decisions impartially, by reference to relevant considerations alone'.

When a person makes a moral judgment in a particular situation, 'e implicitly commits 'imself to making the same judgment in any similar situation. This is what ethical theorists call "the principle of universalizability": if one judges that x is good, right or praiseworthy, then one is committed to judging that anything like x in relevant respects is good, right or praiseworthy. In other words: 'moral judgments are universalizable' and anyone 'who says that a certain action is morally right or wrong, ought or ought not to be done, is thereby committed to taking the same view about any other relevantly similar action'. The key phrase in this formulation of universalizability is relevantly similar, as has been argued elsewhere.

That discrimination itself cannot be defined without making use of relevant or irrelevant is realized by most writers on this subject. So --as noticed before-- somebody may complain about sexism, or discrimination on grounds of sex, when people count sex as relevant in contexts where it is not. Sexism has been defined as preference for members of one's own sex simply because they are members of one's own sex, but in such a definition the crux of sexism remains hidden in the simply because, in the kind of preference concerned and in the kind of actions taken on the basis of this preference. Moreover, the same attitude towards the other sex would be equally sexist, altho aggrandizemental instead of abnegational. (The above definition cannot even handle the difference between sexism and homosexuality, or it does in no way clarify in what it lies.)

However sexism may be defined, it is praiseworthy to point out --it has already been done by others-- that there is a close analogy between this attitude and both racism and speciesism. To define racism and speciesism, race and species have only to be substituted for sex, while the rest of the definition can remain the same. Against attempts to justify a different treatment of animals it has been put forward that this attitude is speciesist, because 'animals are biologically similar in the relevant respects'. On this view the fact that human beings use language, for instance, or a more complex form of language, to communicate is 'not relevant to the question of how animals ought to be treated, unless it can be linked to the issue of whether animals suffer'. (This presupposes a utilitarian morality in which the minimization of suffering, or the maximization of happiness, is the sole thing that counts.)

A definition of discrimination need not mention relevant or irrelevant when it makes use of other concepts which have already incorporated the relevance/irrelevance divide themselves, such as those of fairness, justice or justification. When differentiating the fair and the unfair, or the just and the unjust, relevancy has already been employed, or is employed implicitly. This is the case when an ethical theorist claims that a moral system should not allow 'to discriminate between people for reasons which we would in practise judge to be unfair'. It is also the case when discrimination is not defined as making an irrelevant but as making an unjustified distinction or as difference in treatment or favor on a basis other than individual merit. In the last formulation it is presupposed that a difference made on a basis other than individual merit can never be relevant. Also when the reference is just to favor (rather than to difference in favor) it carries implicitly with it that the favor is not justified, as it is founded upon --again-- a nonrelevant distinction.

From the standpoint of normative philosophy relevance seems to refer to some significant connection with a goal, purpose, function, process or institution universally accepted in a deliberated agreement or by tacit convention. As such it is relational and dependent on the goal, purpose or other directional entity which we will henceforth call for short "the focus (of relevancy)". (Of course, the focus with respect to an institution is its maintenance or enhancement or something of that kind.) The theorist who regards relevancy as a relative notion may say that the relevancy of attributes depends on 'the purposes of a given association or enterprise' or that the criterions for determining what are relevant reasons are 'necessarily linked with the very purpose of the activity of reasoning'. With respect to the differences between men and women, or other groups, it has been pointed out that the question is 'whether any such differences could be relevant to the activity or institution in question'. Because of this relational nature of relevancy it should not surprise us that one author has remarked that what may appear 'relevant from one interested point of view', may 'not appear relevant from the point of view of someone whose situation and qualities are different'.

The knowledge that relevancy is a relative notion and the introduction of the concept focus of relevancy will make it easier to show the significant part played by relevancy in ethical theorizing on the question of equality. This role is in no way taken cognizance of in the classical principle that 'equals should be treated equally and unequals unequally'. This principle may guarantee consistence --as we will see in section 5.4.1 of this chapter-- but just as consistence is no proof of truth, so it is no proof of relevance either. It is that 'equals should be treated equally, unless unequal' which better illustrates what is going on.