CONTENTS

1. INTRODUCTION
   1. The subject of this study
   2. The issues
   3. Points vital to the approach itself
   4. Function and content of the following chapters
   
   
2. THE TRUTH AND RELEVANCY OF A DISTINCT DIRECTION IN PHILOSOPHY
   1. Introduction
   2. Truth, from sole ultimate value to criterion   1. Some history in the truest sense   2. Theories of truth   3. Truth in speech acts
   3. Relevance, from implicit to explicit criterion or maxim   1. Some relevant recent history   2. Relevancy as an ethical criterion   3. Structure and basic concepts
   4. Relevance and truth as substantive normative principles
   
   
3. TOWARDS AN ETHICAL THEORY OF RELEVANCY
   1. Introduction
   2. Basic concepts   1. Principle versus criterion   2. Criterion, factor or focus   3. Discriminational and moral relevancy
   3. Criterions of disciminational irrelevance   1. Inconsistence as one of five criterions   2. Fake focuses   3. Pseudofactual relevancy   4. Dependence on internal or external nonrelevance
   4. The relevant principle and attitude
   
   
4. PERSONAL RIGHTS -- taking a normative second-order position on people and their beliefs
   1. Introduction
   2. Doctrinal versus metadoctrinal foundations
   3. The metadoctrinal right-duty constellation
   4. Personal rights and property
   5. Personal rights and the state
   6. A comparison with some other approaches
   
   
5. THE MORAL AND IMMORAL APPEAL TO RIGHTS IN ISSUES OF DISCRIMINATION
   1. Introduction
   2. Discriminating between the foundations of rights
   3. The question of group rights
   4. One regular, doctrinal right of nondiscrimination
   5. The personal right to discriminate in politics
   
   
6. THE ETHICS OF CATEGORICAL DISTINCTIONS   [not carried out]

1    THE SUBJECT OF THIS STUDY

"[H]e has no real critical faculty -- of people, at all events," said Gudrun. "I tell you, he treats any little fool as he treats me or you -- and it's such an insult." "Oh, it is," said Ursula. "One must discriminate." "One must discriminate," repeated Gudrun. "But he's a wonderful chap, in other respects ..."

D.H.Lawrence 1921, Women in Love, p.23

In a sense 'discrimination' is an ontological and epistemological prerequisite of all thought, a practical prerequisite of all behavior. Machan & Den Uyl 1987 describe it as 'a judgmental process involved in all thinking and recognition' (p.111). Thus, some people may know how to 'discriminate' between real and pretended cases of concern for their well-being. Other people may show 'fine discrimination' in only picking out those works of art which are genuine, and perhaps, ... in distinguishing little fools from more sophisticated minds.

To discriminate has a positive connotation in the above instances, meaning to show good judgment, to make an appropriate distinction or to distinguish by discerning or exposing differences (especially when distinguishing one object from another). If the connotation is neutral or absent, it merely means to see or make a difference. This is the way Singer 1978 uses discrimination in Is racial discrimination arbitrary?. All those who have always used discrimination to refer to something that is wrong are told in a footnote that their popular usage is 'an abuse of language, for to discriminate is merely to distinguish, or differentiate' (p.202). Lucas 1965 reminds the same people in Against Equality that one should not criticize certain laws for discriminating but for discriminating irrelevantly(p.144).

Of the philosophers writing on issues of discrimination Gross 1978 seems to be more realistic than the aforementioned authors, insofar as Gross accepts that there is not one sole correct sense of discrimination in English. In Discrimination in Reverse we are given not less than four meanings of to discriminate. Three of them are basically covered by the definitions I have mentioned already; the fourth one is to make an adverse distinction with regard to someone or something (p.7). Gross claims that this is the interesting meaning for the problem of social justice. An example of this sense of to discriminate would be: one may discriminate against boors by not inviting them to dinner. However, being rude, ungraceful and/or insensitive might, unlike belonging to a certain sex or race, be relevant to having dinner together (which is a social occasion after all). If so, then people will not speak of "discrimination" even though those who are uncouth in manners and appearance are adversely affected by not being invited. The reason will be then that the distinction between people who are uncouth and people who are not is considered relevant or not unjust in the context concerned. It turns out that after a list of four different definitions Gross, too, ignores the fact that there is a meaning of discrimination in which it unmistakably refers to the making of a distinction which is wrong, and wrong for another reason than that it is to someone's disadvantage. It is therefore not surprising that Gross must implicitly admit defeat --in a footnote again-- when writing: "In what follows the terms 'discriminate' and its cognates should be understood to be restricted to cases of unjust discrimination unless otherwise noted" (p.146, my emphasis).

What several philosophers try to deny is what can be found in any modern dictionary, namely that there is also a derogatory or condemnatory meaning of to discriminate or to discriminate against/between. It is a meaning of to discriminate which is not restricted to formal usage. Its definition is to make an irrelevant/unjust distinction. This is the meaning discrimination has when someone is accused of, for example, 'sex discrimination'. It does not mean to treat differently --as dictionaries may want us to believe too-- because it is quite well possible to treat two or more things differently without discriminating between them. (Sexual differentiation, for instance, need not be sex discrimination in the contexts of procreation and sexuality, when sex often is relevant.) When having a negative connotation, to discriminate means to treat differently by drawing an irrelevant/unjust distinction.

The present study is a fundamental study in normative philosophy dealing primarily with 'discrimination' in the sense in which it is something wrong or, more specifically, morally wrong. This, of course, is also the meaning discrimination has when we speak of "a non-" or "antidiscrimination principle", for there is no principle which bans all differentiation, that is, all distinctions. The special perspectives from which I will concern myself with the subject of discrimination are that of relevancy and that of ethical rights. It is these two perspectives together which appear to provide the best view of fundamental normative issues of discrimination.

2    THE ISSUES

Why is discrimination wrong? Judging by the definition given above, two possible answers present themselves: (1) because the distinction made is unjust, and (2) because it is irrelevant or (if morally wrong) irrelevant while the context in which it is made is a moral one. But what is (un)just, and what is (ir)relevant? The answer to the former question lies in a theory of justice; to the latter one in a theory of relevancy. It must not be assumed, however, that the one approach is necessarily independent of the other. In ethical literature two basically different suggestions can be found with respect to the relationship between (theory of) justice and (theory of) relevancy. Blackstone 1977, who distinguishes a 'normative' sense of relevant from a 'fact-stating and descriptive' one, says that the former sense of relevant 'presupposes a theory of justice or a set of general moral principles which define or specify criteria of relevance' (p.62). Murphy 1984, on the other hand, says that, since the fundamental principle of justice treat like cases alike is plagued with many problems, we 'need, of course, a theory of relevance' (p.588). I myself will argue, in line with Murphy's suggestion, that ethical notions such as nondiscrimination, justice, equality and universalizability insofar as they concern the making of distinctions depend on relevancy and its criterions, and not the other way around. The ultimate normative principle involved in discrimination as discrimination is not justice, comparative or otherwise, but relevance (in conjunction with a principle of well-being).

A few authors reject relevancy altogether, not only as a separate principle but also as a criterion of justice in issues of discrimination. Fiss 1976, for instance, notices that the antidiscrimination principle (interpreted as a principle of relevance) 'can be used just as comfortably to challenge a statute that draws a distinction between opticians and optometrists or one that draws a distinction between filled milk and margarine as it can be used to challenge a statute that draws a distinction between whites and blacks' (p.105). Against this position I will argue that one must differentiate a fundamental normative principle of relevance and a nonfundamental moral principle of nondiscrimination. (Filled milk and margarine in themselves do, indeed, not constitute a moral context.) Wasserstrom 1980 writes that the chief defect of racial segregation is not its systemic capriciousness, and that the fact that race is an irrelevant characteristic is not what is wrong about racial discrimination (p.64). The fundamental thing which was wrong with the most hideous practice, slavery, says Wasserstrom, is the practice itself 'no matter how the assignment of the victims has been made' (p.65). My objection to this argument is that slavery as such is the violation of a negative right --of what I will call "a personal right"-- not racism as racism. Slavery is also objectionable when race plays no role whatsoever. Racism, racial discrimination and racial segregation as such involve, by definition, a distinction made on the basis of race, and in the context concerned there must be something wrong with drawing this particular distinction.

There is a tendency in the literature to equate the irrelevance of a distinction with its being arbitrary or capricious. Thus Rawls 1958 (and 1971) says in Justice as Fairness that in its usual sense justice is essentially the elimination of arbitrary distinctions (and the establishment of a proper balance between competing claims) (p.77). Bedau 1971 says that the writings of Godwin and Tawney demonstrate that egalitarians have always protested against 'inequalities based on 'arbitrary' or 'capricious' distinctions' (p.170). In the same vein Frankena 1977 agrees with Aristotle that the essence of distributive justice is arbitrary discrimination between relevantly similar cases, even though disagreeing over which characteristics are relevantly similar and which distinctions arbitrary. (See Feinberg & Gross 1977, p.6-7 and 46-54.)

Feinberg 1977 first suggests that 'arbitrary discrimination' is wrong or unfair, because people are hurt and the hurt is in some important way offensive to reason (p.64). Again in a footnote, however, Feinberg draws an interesting distinction between irrelevance and arbitrariness when saying that 'in the case of unjust discrimination, there is an irrelevant criterion employed, or else there is no 'procedure' and no 'rationale', but only arbitrariness through and through' (p.74, my emphasis). And, certainly, an irrelevant factor is something else than an arbitrary factor. An arbitrary factor is one chosen at random or without any (normative) principle in mind. Such an arbitrary factor is not per se irrelevant: it may happen to be relevant in the context concerned. Conversely, an irrelevant factor may be used as the basis of a distinction in a capricious manner, but it may also be, or have been, believed to be relevant in an old tradition or in a long process of deliberation. A 'procedure' or 'rationale' may prevent arbitrariness, but it still does not guarantee relevance.

Also Singer 1978 regards relevancy and (non)arbitrariness as identical notions in Is racial discrimination arbitrary?. In such a case we may have to read "irrelevant" instead of "arbitrary" (just as we may prefer to read "(relevant or irrelevant) differentiation" instead of "discrimination"). The confusion of relevancy and arbitrariness does not affect Singer's criticism, though, against the so-called 'standard objection to racial discrimination', namely the irrelevance of race. Singer says that this account of why racial differentiation is wrong is 'inadequate because there are many situations in which ... the racial factor is by no means irrelevant' (p.187). But what is attacked here is first of all the idea that anyone who makes any kind of racial distinction would be a racist pur sang. This idea is indeed nonsensical, for there definitely are situations in which race is, or would be, neither arbitrary nor irrelevant. Take, for example, the distribution of suntan lotion over friends with a different skin color. Singer's examples, however, are less obvious than this one. The question we must pose with regard to each example adduced by Singer or others is --apart from that of the adequacy of the alternative solution-- whether the 'relevance' of the distinction concerned does not somehow depend on the irrelevance of other distinctions made at the same time or before. This is definitely a challenge to anyone believing and defending that it is the irrelevance of one or more distinctions drawn in a moral context which makes discrimination morally wrong. This issue is part of the more general issue of what are the criterions of relevancy.

Instead of defining to discriminate (in a sense in which it is wrong) as to make an irrelevant distinction (with or without further qualifications), Webster's Third New International Dictionary (1986) defines it as to make a difference in treatment or favor on a class or categorical basis in disregard of individual merit. There are two assumptions implicit in such a definition. The first one is that only merit justifies equal or unequal treatment. This does not deserve our further attention here, because any factor can in principle justify (non)differentiation. One such factor is sex, and it certainly is not a merit to be male or female. The second assumption is that to treat someone on a categorical basis, that is, as a member of a class or group, is never relevant or just. Only an individual distinction can be relevant and just on such a view; that is, a distinction on the grounds of a quality someone personally or individually possesses. Nonetheless, this does not dispose of the question of whether a distinction on a categorical basis is never relevant. Such a categorical distinction is a distinction between categories of which all members are treated as having one or more qualities in common in addition to (and logically independent of) the one(s) which define(s) the category, regardless of whether an individual member has this or these qualities. If people of age, marital status or profession A are on average much safer drivers, for instance, than people of age, marital status or profession B, then every member of A is treated as a safe driver, and every member of B as an unsafe one. Actuarial thinking is a typical example of a kind of thought thriving on categorical distinctions, although it is certainly not the only kind to employ them.

'Group thought' is also found in philosophers who favor (corrective) preferential treatment, or who claim the existence of group rights. Thus, Fiss 1976 criticizes the antidiscrimination principle (interpreted in such a way that it does not allow of categorical distinctions) for embodying 'a very limited conception of equality' and for being 'highly individualistic' (p.85). Fiss's own group-disadvantaging principle should, in the eyes of its creator, have 'a good ... claim to represent the ideal of equality' and take 'a fuller account of social reality'. We encounter Fiss's recognition and protection of social groups again in Glazer 1978 who argues for a group rights approach because 'the differences between some groups are so great that they cannot achieve satisfaction on the basis of individual rights' (p.102).

"There are few issues of contemporary institutional morality which have engendered more controversy than that of whether programs variously called "programs of affirmative action", "preferential treatment" or "reverse discrimination" are justifiable", writes Wasserstrom 1980 (p.51). Whole bookcases have been filled with books and articles by proponents and opponents of what is also called "positive" or "preferential discrimination". It is not my intention to add yet another writing to that collection. My interest in corrective preferential treatment --'corrective' as distinct from 'favoritist'-- stems from what the debate on this issue may contribute, firstly, to the solution of the fundamental question of the relevancy of distinctions, and secondly, to that of the question of which foundations of rights play a role in issues of discrimination. The fact that corrective preferential treatment is such a controversial topic should, then, only be extra instructive. Thus, against the position of people like Fiss and Glazer, Sher 1977 writes: "[T]he argument ... that groups as opposed to their individual members are the sorts of entities that can be wronged and deserve redress, is itself problematic. ... No one ... has yet produced a powerful argument to this effect, and I am not hopeful about the possibilities" (p.50). Fullinwider 1980 notes that it is not an error to treat a group of individuals as itself an individual when this group is a corporation. But, Fullinwider says, 'we are on very uncertain footing in talking about the "group rights" of blacks even when we purport to be talking about corporations and nations' (p.65).

When we return to 'regular' issues of discrimination, that is, issues in which the correction (including compensation) of wrongs does not (yet) play a part, Sher 1977 and Fullinwider's positions match that of Fishkin. Fishkin has written that 'a determination of qualifications should not rest simply on statistical inferences', that 'an individual in a given group ... is subjected to 'statistical discrimination' in this way. (See Sher 1987, p.190.) In this case, however, Sher 1987 opposes Fishkin's 'principled rejection of statistical predictions of performance' (p.193), now defending the use of statistical information, and therefore of categorical distinctions, on the grounds of values like utility and desert. But even if we were willing to make a distinction on those grounds, the fundamental question remains whether it is discriminatory when the quality of work to be done is the recognized goal defining the context. If so, there is always the next question whether it should be made on other grounds nevertheless. In my analysis statistical relevancy (as used in predicting performance, for instance) is not the same as discriminational relevancy (which, in turn, is not the same as moral relevancy). When deciding about the legitimacy of categorical distinctions these different types of relevancy should not be mixed up.

Discrimination does not only involve a person or a sentient or living being as a 'moral recipient' but an actor too. Talking about individual or group rights against discrimination is taking the position of this recipient. However, can we also speak about the actor's rights besides, perhaps, the actor's moral mandatory (half-)right not to discriminate for or against anyone or anything? Is a person or a group allowed to make a distinction when we or others know or believe it to be discriminatory? (The discriminators themselves will, of course, seldom or never call the distinction in question "irrelevant" or "unjust".) I will first deal with this problem as one of a fundamental ethical, not of a political or legal, nature. It may be rephrased as Does a person have the ethical right to make a distinction which others consider discriminatory and therefore immoral?. The problem does not become political until we ask whether the state and its citizens have that ethical right, or when moral opponents, the ethical right to check discrimination wherever and however it takes place. To answer the fundamental ethical question in the affirmative is to recognize that there is an ethical right to be immoral, granted that it is immoral to discriminate. This standpoint sure is paradoxical. To solve the apparent contradiction I will use a technique reminiscent of Russell 1908's solution of the logical (set-theoretical and semantic) paradoxes, and in line with the idea of ordering in several contemporary philosophical systems (like, for example, that of Raz 1975, with reasons as basic constituents).

Obviously, it is also possible to claim that there is no ethical right to be immoral. And even if there is one, it may be thought strategically unwise to argue that it exists, or can consistently be adhered to. (Let sleeping dogs make no distinctions.) But the matter is not that simple. First, those who refuse to accept a right to be immoral necessarily tend to restrict their conception of what is immoral. This may explain why someone like Loury 1987 is not afraid of writing: "[W]e all basically accept the legitimacy of the practice of racial discrimination in the intimate, personal sphere" (p.259). According to Loury 'the antidiscrimination principle does not extend into the most intimate of private, associational choices' (p.270). In my analysis the nondiscrimination principle holds everywhere, but this does not mean that individual, private taste cannot be relevant to a purely private purpose, or that a person would not have any right to make a discriminatory distinction. Even in those cases in which a person is allowed to make a distinction which is, but may not be believed to be, discriminatory, it remains immoral to do so. We can and should continue to point this out and to take those steps we are allowed to take on the grounds of the same principle. It is a psychosocial absurdity to morally condemn discrimination in public and to 'equally morally' sanction it when committed by private individuals or organizations, apart from the question of what would be the fundamental criterions for differentiating the public and the private spheres.

There is a second important point involved in the recognition and nonrecognition of ethical rights. It is that the institution of property, which includes private and public funds, or money, is a system of rights, or of rights and duties, as well. The same principle which is the foundation of certain liberty rights (or duties of noninterference) may be the foundation of a certain type of ethical property rights. To reject this principle is, then, to reject both the liberal rights and the property rights, and to accept it is, again, to accept both the liberal rights, such as the right to be immoral, and the property rights. It is these ethical property rights which in my view limit the sphere within which someone (or a group) has the personal right to be immoral. This is most clearly so with respect to the ethical ownership of one's body. I will not, like Macpherson 1973, claim that 'all roads lead to property' (p.121), but there are normative issues of discrimination which cannot be adequately dealt with without paying a minimum of attention to the ethical property rights involved.

Until now I have not differentiated between the factors on the basis of which a distinction is made, except in the choice and discussion of examples of discrimination. In this study, however, I will also consider the question of why, and when, it may be justifiable not to treat every factor of distinction in the same way, even though ultimately the nondiscrimination principle is, as I will defend, equally valid for every factor on the basis of which a distinction can be drawn. What does not seem to support this universal validity is that the antidiscrimination principle is usually restricted to 'suspect classifications' or to 'factors which appear in isms': racism, ethnocentrism, nationalism, sexism and agism. (Speciesism is a special case.) But why would factors like race, ethnicity, nationality, and sex be suspect and other factors not? And what about factors like someone's political convictions, or someone's religious or nonreligious outlook on, and way of, life, or other factors for which there is no 'ism' at present? It goes without saying that distinctions made in the vocabulary of ordinary, traditional language need not be relevant at all, apart from the fact that even ordinary terms are not used by everyone in the same sense.

Baier 1978, for instance, complains that 'recently 'racism' has come to be used ... to mean something that is necessarily morally objectionable' (p.127) and uses it in such a way that it is not morally wrong, the fact that racism is wrong in any normal sense of the word notwithstanding. Racist has even become, as Singer 1978 puts it, a 'bludgeon word', very powerful, effective and expressive. So expressive that Bracken 1978 feels a witness of a world tainted with innumerable 'racisms' each with its own factor of distinction: color ('color racism'), religion ('religious racism'), language ('linguistic racism'), nationality, and so on (p.242). Perhaps, racist and to a lesser extent sexist have acquired such denunciatory power that they solely refer to something that is egregiously wrong, bad or immoral. This may explain why these words are also used when racial or sexual discrimination is, strictly speaking, not at issue. (Who uses ethnocentrism when the factor of distinction is not race but ethnicity?) Perhaps, it is not acceptable (anymore) to speak of "a petty form of racism" or "sexism which is a little bit wrong" to distinguish it from "a serious form of racism" or "sexism which is extremely wrong". But while present-day ordinary language might rule out this usage, the wrongness of racial, sexual or any other type of discrimination itself may vary from that of a harmless peccadillo to that of a cold-blooded murder. This, too, is something to be taken into account in this study.

3    POINTS VITAL TO THE APPROACH ITSELF

4    FUNCTION AND CONTENT OF THE FOLLOWING CHAPTERS

The main function of Chapter Two is to demonstrate that it makes sense not only to regard truth but also relevance as a substantive (rather than purely formal) normative principle. So far as truth is concerned this is a relatively easy task, for the idea that truth is an ultimate value, even the sole ultimate value, has always pervaded both philosophical and nonphilosophical human thought. Unlike truth, relevance, however, has only recently begun to draw the attention of philosophers. The first article about it (by Schiller) appeared in 1912; the first person to raise it to a principle (in philosophy of language) was Grice, in 1975. Also in ethics relevance has now emerged at least as a criterion in analyses of justice, universalizability, equality and (non)discrimination: in a moral context, or with respect to a moral goal, distinctions made have to be relevant and relevant distinctions have to be made.

There are several interesting parallelisms between the role of truth and theory of truth on the one hand, and the role of relevancy and theory of relevancy on the other. To treat both truth and relevance as ultimate values in normative philosophy is to exploit these parallelisms. And to recognize these values is to go down to the (in itself nonevaluative) basis of normative considerations instead of stopping somewhere on the way at a level of moral or ethical considerations which are normative but not ultimate. Thus, the moral principle of nondiscrimination encompasses the ultimate normative principle of (discriminational) relevance in conjunction with considerations of well-being or the minimization of unhappiness (pain or suffering). In themselves, neither relevance nor truth are moral values or disvalues. But, conversely, this means that moral disvalues like discrimination and lying are not fundamental in the field of normative philosophy in general.

After having pointed out the importance of relevancy in discussions about discrimination (and other ethical issues), the key question in Chapter Three is: How do we determine what is relevant or irrelevant?. Other questions are: What does relevant mean? and What are the types of relevancy (if any)?. Is, for example, the relevancy which plays such a crucial role in issues of discrimination the same one as other types of relevancy? It will turn out that the former type of relevancy (discriminational relevancy) is not the same as moral relevancy, but the two are compatible.

What is, or is not, relevant in issues of discrimination is something like age, marital status, sex, sexual orientation, (dis)ability, ethnicity, nationality, language or ideological affiliation, that is, a factor of distinction. Such a factor is, or is not, relevant with respect to a certain goal, aim, purpose or other directional entity, that is, with respect to what I will call "a focus of relevancy". Now, to determine whether a factor of distinction is relevant to a given focus of relevancy or not we will have to study the criterions of relevance and/or irrelevance. Although I will not give a criterion to prove the relevance of a judgment, I will propose four or five criterions which may show up the irrelevance of judgments.

By selecting relevance as an ultimate normative principle (rather than merely a criterion of something else) it is possible to explain why racial and other types of segregationism are wrong, even if the parties involved are or were treated equally well. The fact that irrelevant distinctions on the basis of sex or other such factors are also wrong in what are, perhaps, nonmoral contexts (such as that of the language we speak) is not without significance. For taking this fundamental normative principle seriously is a question of attitude. And such an attitude is not a device to be turned on and off at will, that is, on in a moral context and off in a nonmoral one.

Chapter Four does not specifically deal with matters of discrimination or relevancy. Its function is clearly to prepare the reader for what is argued for and against in the next chapter. The subject of that chapter is rights in issues of discrimination. Also with regard to rights it is important, I believe, to consider the ultimate value(s) on which they rest, and the status of the normative considerations in which they occur. The foundations of rights usually adduced by theorists, such as equality, liberty, utility, interest and autonomy, are often ambiguous or unacceptable for a variety of reasons. Moreover, in my eyes, there is one ultimate (nonteleological) value which is missing. It belongs, not to a first-order doctrine like the other ones, but to a doctrine in which a second-order position is taken on people and their (first-order) beliefs. This value I will term "personhood", and the rights of which it is the foundation "personal rights".

Of course, there is much similarity between personhood as a value and 'equality' (in the sense of treatment as an equal or (equality of) respect) and negative liberty as values; and also between personal rights and human rights. What counts, though, is which value is more plausible or appealing where the presuppositions and/or implications are different. One asset of personal rights is definitely that its foundation is not speciesist (anthropocentric). Another asset is that personal rights are not an arbitrary mixture of negative (political, civil) rights and positive (social, economic) rights: the 'socioeconomic' rights they comprise are a certain sort of property rights which are immediately connected to the duty of noninterference in their interpretation. Furthermore, the domain of personal rights may be ethical, it is not 'moral' in a narrow sense of moral. The ethical, personal right to be immoral loses its paradoxical character when explicitly differentiating the type of considerations which underly personal rights and duties and the type of considerations which underly other, moral rights and duties.

Ethical rights, including personal rights, theoretically hold in a stateless environment of only two people or moral entities as well. Given the existence of states and larger societies, however, I will discuss some of the political implications of recognizing personal rights.

Of the rights appealed to in issues of discrimination the typical ones are, obviously, not personal rights but rights against discrimination. In Chapter Five I will first pay attention to such rights with a first-order doctrinal foundation. One question to be dealt with then is whether there is only one overall right of nondiscrimination, or as many as there are factors on the basis of which one ought not to discriminate. In order to settle this question, regular normative considerations should be distinguished from corrective ones. I will argue that the foundation of the one right of nondiscrimination is regular, whereas the (possible) foundation of group rights in issues of discrimination is corrective.

From the perspective of an adequate first-order normative doctrine the appeal to a right to discriminate is immoral when this refers to the making of an unjustifiable distinction. It is only from a 'meta' perspective that one may recognize an ethical right (not) to discriminate, that is, to make a distinction, to make an irrelevant distinction, or to make an irrelevant distinction in a moral context. However, the person or group discriminating will seldom or never call its own action or attitude "discriminatory" or "unjust", and certainly never "racist", "sexist" or something of that ilk. Yet that they believe their action or attitude not to be discriminatory does not make it nondiscriminatory. How do we combine the right first-order perspective, the 'meta' perspective and a wrong first-order perspective? This question concerns the interplay between the personal right to discriminate and the (first-order) doctrinal right and duty not to discriminate.

Although the principle of nondiscrimination applies equally to all factors of distinction there are some factors which nevertheless may have to be considered separately for one or two reasons: (1) because of actual discrimination in the past, or a (still-)existing practice of discrimination on the basis of such a factor; and (2) because allowing discrimination on the basis of such a factor from a 'meta' perspective runs counter to a basic assumption underlying personal rights. In particular the rights of nondiscrimination relating to these factors deserve special attention.

In Chapter Six [a part of the original plan not carried out] I will first return to the question of what discrimination is. I take it then that not much more need to be said here about the ethics of individual distinctions; the issue is now the ethics of categorical distinctions. The following points will come up for discussion:

• literature which may shed light on the subject, e.g., Simon 1977, Braybrooke 1983, Beider 1987 and Sher 1987
• the categorical nature of group rights
• discriminational relevancy versus statistical relevancy
• the fact that discriminational relevancy requires an explanatory link (cp. Beider 1987; contra Fullinwider 1980)
• the interplay between nondiscrimination and utility, administrative convenience, etc., or the well-being of the persons or groups concerned
• the difference between discriminating (regardless of intention or motive) and intending to discriminate (or having a discriminatory motive)
• the question of information available and the rationality of decision-taking or human behavior under conditions of uncertainty
• if the availability of all relevant information were an absolute requirement, would this not worsen the situation of the worst-off?
• where strict nondiscrimination has to be sacrificed to make categorical distinctions possible (permitting, for example, price- or premium-differentiation), what is the role of the factor(s) of distinction involved?
• the question of so-called 'suspect classifications' or 'ascriptive characteristics'
• what is the role of the rights involved?