Those who agree that one should adopt a (non-meta-)doctrinal, nonpropositional principle of discriminational, nondoxastic relevance, might still disagree about the adoption of a personal or a universal version of this principle. The personal version only puts an end to the dependence on internal nonrelevance. The relevance of the distinction they make does, then, not depend on a causal connection between this distinction and an attitude or practise of making another nonrelevant distinction by themselves. The universal version goes further. If possible, this version puts an end to the dependence on external nonrelevance, when a distinction made by others is nonrelevant.

The difference between the personal and the universal version of the relevance principle has no bearing on a moral agent`s right of personhood to choose for the one or the other version. In both cases `e exercises `er own personal freedom to make a normative decision, albeit a more demanding one when espousing the universal version. What is also the same in both cases, is that either version has an active and a nonactive component. The nonactive component may also be called "restrictive" or "proscriptive". For P (the personal version) it reads, "one should not make a distinction which is not relevant", while the addition for U (the universal version) is, "or the relevance of which depends on a nonrelevant distinction made or on the not making of a relevant distinction". The active component may also be called "prescriptive", and reads for P, "one should make a distinction which is relevant", while the addition for U is, "unless this relevance depends on a nonrelevant distinction made, or on the not making of a relevant distinction".

As noted in section I.7.4.2 --that is, 7.4.2 of the Book of Instruments-- the choice whether to assume that other people act from the same normative theory (as in the universal version) or not (as in the personal version) cannot be determined by a mathematical or logical calculus. The question of action versus nonaction is more complicated tho, because many consequentialists argue that there is no difference between an action and a nonaction when the effects are the same. But usually the effects are taken to be something like the amount of pleasure or suffering the act or omission or abstention brings about, and not a quality inherent in the action or nonaction itself. When a quality in an action or nonaction itself is a value or disvalue as well, the matter is not so simple anymore, and that is why we have to postpone the discussion of the proscriptive and prescriptive readings of our normative principles until a later chapter. The question of the difference in versions themselves we must tackle now, however, for a personal version of the principle of relevance would practically exhaust all its normative vigor.

Those who suggest that people need or can only adopt the personal version of the relevance principle, do not only accept that people would show less strength in matters of relevance in practise, they must also hold that it is not the case that people ought to display more strength in such matters, even tho they had every ability and opportunity of doing so. On the other hand, those adopting the universal version of the relevance principle do hold that one should display as much strength in such matters as one can. If one cannot possibly do anything more than what is required according to the personal version, this is also what is required according to the universal version. Yet, the underlying attitude is quite different. Those pleading for the personal version tend to reconcile themselves to what they but too easily assume cannot be avoided. Thus, where the majority in a community discriminates against a minority, they have this discrimination determine the 'relevance' of their own distinctions, as tho submitting to the will of a monolithic Exclusionist. All these people embracing the personal version of the relevance principle are therefore themselves part of, or constitute, the very majority which --they claim-- is beyond their control.

To illustrate what it means to accept the universal version of the principle of relevance, or the criterions doing away with assigning relevance to both internal and external nonrelevance, we will give some time and attention to a concrete example at the end of this division.