In stigmatizing an entire profession because of the behavior of a minority, or in engaging in similar attributions in other areas of decisionmaking, we undertake a process of generalization. “Painting with a broad brush,” the slightly pejorative expression puts it. To refer to the practice as stereotyping intensifies the condemnation, and these days to describe it as profiling condemns more strongly still. Yet regardless of the label the structure of the decision is the same. On the basis of a characteristic of some members of a class, we reach conclusions or make decisions about the entire class.
Decisionmaking by generalization is the stock in trade of the insurance industry. To be an actuary is to be a specialist in generalization, and actuaries engage in a form of decisionmaking that is sometimes called actuarial. Actuaries guide insurance companies in making decisions about large categories that have the effect of attributing to the entire category certain characteristics that are probabilistically indicated by membership in the category, but that still may not be possessed by a particular member of the category. Occasionally the actuarial generalizations of the insurance industry become controversial (for example that gay men pay more for health insure because as a class they have a statistically greater likelihood of contracting the HIV virus), but in the vast majority of instances the actuarial behavior of the insurance industry is accepted simply as a fact of life.
Insurance companies, of course, are not the only ones who operate actuarially. We all do, with far more frequency than most people typically acknowledge.
We cannot so easily dismiss the practice of stereotyping or profiling as necessarily morally wrong.
Nonspurious: generalizations that do have a sound statistical basis
Universal: all of the x’s are y (mostly used by philosophers)
Definitional source of the universality: in many cases universal generalizations are universal as a matter of definition (“all bachelors are unmarried”).
Empirical source of the universality: sometimes, however, universal generalizations are universal not for definition al reasons but for empirical ones; some fact about the world and not about the word makes the generalizations universal. For example: until black swans were discovered, “all swans are white” was a universal generalizations; but the possibility remained open, even then, that there could be black swans, unlike the possibility, nonexistent by definition, that there could be married bachelors.
Not universal: “Swiss cheese has holes,” “Philosophers are clever,” “Volvos are reliable,” etc. In using and understanding such statements, we recognize that they apply only to most cases, and although we could make things more precise by ensuring that we always prefaced each of the foregoing with “all,” “most,” “many,” “a disproportionate number of,” or some similar qualification, it is noteworthy that we rarely do so.
The existence of particular cases in which the generalization does not hold rarely causes us to refrain from using the generalization, provided that the generalization holds true for most cases.
We often generalize, and are understood to have generalized appropriately, even when less than a majority of the x’s are y. For example: when on the basis of accident statistics or personal experience we actuarially castigate a group as “bad drivers”, we often mean only that the group has a noticeably higher percentage of bad drivers than some other group, or than the group of all drivers.
It is important to distinguish those nonuniversal generalizations that have sound statistical basis from those that do not.
Generalizations are statistically sound when they are universal, when they accurately portray the traits of a majority of the members of the class and even when the generalization accurately portrays the members of the class as having a greater prevalence of some trait than has some larger class of which the group is ordinarily taken to be a part, even though the trait appears in less than a majority of the members of both groups. When the trait is relevant, we can say that generalizing from that trait meets the minimum threshold of statistical (or actuarial) soundness.
A generalization can be statistically sound despite the fact that the generalization applies only to a minority of members of the class. A key feature of a sound generalization is its comparative dimension. Even if a generalization accurately represents the traits of even a majority of some group, we would think it misleading to use the generalization if the trait was no more present in the group we were talking about than it was in any other group.
Spurious: generalizations that have no statistical or factual basis
Even though generalizations can be statistically sound in a number of different ways, there are statistically unsound generalizations; when the trait around which the generalization is constructed turns out to be completely irrelevant to the characteristic that we may often be trying to predict. When applied to groups of people, statistically unsound generalizations are often castigated as prejudices. In common parlance, to have prejudice is to have an unsubstantiated belief about a person because of his or her membership in a group. But this way of describing what it is to be prejudiced is ambiguous.
Some people might even say that it is “prejudiced” not only to rely on a spurious generalization, but also to apply a statistically sound but nonuniversal generalization to a particular member of the group.
Because the work “prejudice” is pejorative, it is often used to refer not only to statistically unsound generalizations, but also to the inappropriate use of even statistically sound generalizations. Much the same can be said about the word “stereotype,” which has as even more negative connotation. As with “prejudice,” it seems wrong to use the word “stereotype” to refer to those cases in which a generalization is both empirically and morally appropriate. There are serious and unresolved definitional issues that surround the topic of generalization. That the terms in common usage, such as “prejudice,” “stereotype,” and “profiling,” are ambiguous on the conceptual, empirical, and moral issues is not a solution, but a signal that we have problem in need of analysis.
Many people believe it wrong to make individual decisions on the basis of nonuniversal group characteristics even if the group attributions have a solid statistical grounding. From this perspective it is simply always wrong to make actuarial decisions. But although this belief in the wrongfulness of reliance on even statistically sound but nonuniversal generalizations is widespread, it still may not be correct of even plausible.
Particularist: to make decisions on the basis of the characteristics of particular events or particular individuals, rather than on the basis of the characteristics of the groups of classes of which the particulars may be members.
For the Philosopher Onora O’Neill, for Thomas Hardy, for William Blake and for many others, maximum particularity is a characteristic of both justice and wisdom, and reliance on nonparticular categories or principles is at best a necessary evil, at worst and injustice, and all too often a demonstration of stupidity. But perhaps generalization is more inevitable and even more desirable than Blake and others have thought. On closer analysis it may appear that there is more to be said (and not only by the insurance companies) for actuarial decisionmaking than is commonly supposed.
In most cases involving generalizing on the basis of race, gender, and sexual orientation, the stereotyping I simply statistically unsound. But we have moved too quickly from condemning stereotyping or profiling on the basis of characteristics such as race, gender, and sexual orientation to believing in the moral impermissibility of all stereotyping and all profiling. The use of generalizations in decisionmaking – stereotyping – may sometimes be desirable even when it produces more and not fewer errors in the long run.
These days when profiling is the subject it is race that is the question. Issues of race and ethnicity have become inextricably connected with the issue of profiling, and debates focusing on race and ethnicity have irreversibly transformed the nature of the discussion about profiling.
An example are the customs searches at airports, whereat a disproportionately large number of African-American women were subjected. Many people appeared to believe that the targeting of African-American women was a product of a formula that included race and gender among its elements, it turns out that this was not the case at all. The customs officers at the particular airport did indeed have a formula – a profile, in one sense of the term – but that formula made no mention whatsoever of either race or gender as factors whose presence increased the likelihood that the passenger was transporting unlawful drugs. And this is not surprising, because, according to the Customs Serivce, there was no evidence that either race or gender was statistically relevant to identifying drug couriers. Race having been spurious in the underlying reality and nonexistent in any formal profiles that existed, the incidents at the airport were caused not by the inclusion of race or gender in a list of factors to be used in determining whom to search, but rather by individual customs officials’ deciding on their own, or on the basis of collective but false folk wisdom, that African-American women were to be targeted. When the practice of targeting African-American women was halted, the effectiveness in locating drug smugglers actually increased. The elimination of a spurious factor cannot help but increase the probability of successful identification of those with illegal drugs.
The exercise of individual discretion is often, though of course not always, inferior to decisionmaking based on more systematically compiled actuarial information. People are often inclined to overestimate the proportion of a particularly salient component within a larger population. Because race is salient for most people, they are consequently likely to amplify the extent to which members of a race other than their own are represented in a larger population with negative attributes, just because the observer is likely to focus more on the “out group” member of that population, and consequently take those “out group” members as being more representative of the group than they in fact are.
There are two possibilities: one is simple racial animus, in which hostile acts against people because of their race are based on pure dislike of people of that race, with no pretense of an empirical basis for that dislike. But it can also be (at least partially) the product of mistaken empirical beliefs on the part of the officers that race was a reliable predictor of narcotics trafficking. Both situations are quite different from the reliance on race under circumstances in which race is a nonspurious indicator of likely criminal conduct.
The use of race in airport drug-courier investigation may be spurious, but the intellectually harder and potentially more important issues arise when race turns out to be nonspurious. The most salient example is that of suspected terrorists who would seek to bomb, hijack, or otherwise interfere with the operation of commercial airplanes. Unlike being a drug courier, here it appears that neither race nor gender is spurious. Those who commit acts of airplane terrorism are disproportionately younger Muslim men of Middle Eastern background.
The Gore Commission: a profile must be based on “reasonable predictions of risk, and not stereotypes or generalizations.” The Commission announced that including race (as well as ethnicity, national origin, and religion) in the profile was improper, but it said nothing about why that was so. It sidestepped the question whether using race was improper because it was statistically irrelevant or improper despite its statistical relevance.
The serious proponents of using race or ethnicity do not claim that either should be the exclusive factor, but only that each is permissibly a factor that should be employed to the full extent of its statistical relevance. Proponents of ethnic profiling maintain, therefore, that when ethnicity is merely one of multiple factors employed in deciding whom to target, then there is, assuming statistical relevance, no good reason not to include ethnicity and, when justified, race, as part of the larger profile.
In order for an algorithm to make statistical sense, each of the factors would have to contribute something to the reliability of the full algorithm. When race would indeed substantially increase the effectiveness of the algorithm, it would appear that there is a strong argument in its favor. Especially when the consequences are catastrophic.
At times we may wish to impose a compensatory underuse of a relevant factor in order to account for an expected overuse.
In a list of factors to employ in determining when to trigger special security investigations, it is quite possible that race and ethnicity would be treated as the most important among many attributes of these factors even if they were not. One of the most important reasons for this is that race, gender, age and often ethnicity, unlike many other attributes, have a visibility and a consequent salience that makes them stand out more than other factors. Such attributes thus have a tendency to be utilized more than their actual predictive contribution would justify. Because these attributes, unlike other personal characteristics and attributes, are “visually accessible, culturally meaningful, and interactionally relevant,” such factors occupy more of the decisionmaking space than their empirical role would support. We might worry, on efficiency grounds as well as on moral ones, about the overuse of race and ethnicity and the consequent underuse of other relevant factors, and we might as a compensatory strategy prohibit the use of race and ethnicity just to ensure that they were not overused. The very visibility of race and, often, ethnicity, may produce, among other things, the conditions of their frequent overuse. If the likely overuse was sufficiently large to suggest a compensatory underuse, the resultant prohibition on the use of race and ethnicity would probably produce a lowered level of effectiveness as compared to allowing race and ethnicity to be used in their proper proportion by ideal airline employees. It could well turn out that, in a world of non-ideal employees inclined to overuse race and ethnicity even when they are statistically relevant, mandatory underuse or nonuse would actually be more effective, not because race and ethnicity are not relevant, but because dramatic overuse of race and ethnicity might detract from the ability to examine closely those passengers with an even larger number of properly suspicion-raising characteristics and behaviors. But under some circumstances race or ethnicity, even though likely to be overused, will be sufficiently statistically contributory that precluding their sue may produce some net decrease in law-enforcement effectiveness. But when the accurate use of race or ethnicity probduced more racial or ethnic separation than was morally or socially desirable, a society might decide that the loss of effectiveness was sufficiently small that it would be a price worth paying.
We need to examine more closely just what the price would be. If excluding the relevant factor of Middle Eastern appearance from the algorithm made it necessary to increase the scrutiny of everyone – if excluding ethnicity while still including everything else increased waiting time at airports an average of thirty minutes per passenger – this might still be a price worth paying. So if the increase in waiting time fully compensated for a suboptimal algorithm, the question is no longer one of ethnicity versus security, or ethnic sensitivity versus increased terrorism, but is the more manageable one of time versus security. The question of racial or ethnic profiling in air travel is not the question of whether racial and ethnic sensitivity must be bought at the price of thousands of lives. Rather, it is most often the question of whether racial and ethnic sensitivity should be bought at the price of arriving thirty minutes earlier at the airport.
The use of ethnicity in the context of airport security is a comparatively new concern, but the use of race and ethnicity in other law-enforcement contexts has been an issue for considerably longer.
“Driving while black” has long been the phrase used sarcastically by African Americans to describe the extent to which they are targeted for traffic stops predominantly because of their race. There was no indication that there were developed written formal profiles for determining which drivers were most likely to be committing offenses other than the ones that were visible from outside their automobiles. The fact that a profile is not written down and promulgated officially, however, does not mean that it does not exist. Supposing that an unwritten norm existed, however, as it almost certainly did, leaves open a range of possibilities about what the substance of this norm actually was. It could have simply been a norm of racial animus. But alternatively, It is well within the realm of possibility that the police officers acted not (or not only) on the basis of racial animus, but also on the basis of a mistaken belief about the role of race in predicting highway drugrelated offences.
Yet what about those instances in which using race is not totally spurious? If and when this is so, then we again confront the question of the strength of the factor as a component of a larger profile. In the context of using race to determine which traffic-offending drivers to investigate for possible non-driving offenses, race may be like wearing sunglasses while going through customs – statistically relevant but vastly under- and overinclusive.
There is a vast difference between a “racial profile” and what we might call a “profile that includes race.” But on many of these issues the empirical evidence is spotty. Because what evidence we do have is based upon those who are stopped and not on a less biased array of all passengers or all drivers. Even if the racial or ethnic factors are statistically relevant, it again does not necessarily follow that they ought to be used. Especially when the statistical contribution of race turns out to be small, and when the use of all other statistically relevant factors except race does not involve giving up much in effectiveness, there may be good arguments in favor of the exclusion of race even when the race-excluding profile is somewhat less predictive than the race-including profile.
If it turns out that race, even when it is a nonspurious indicator of criminal conduct, is being used far in excess of what its statistical contribution would justify, then a statistically legitimate “profile that includes race” is being converted at the point of application into something that focuses only on race and ignores the other indicators that would be part of the full array of indicative factors. Even if we define racial profiling g in its narrowest and most justifiable sense, it may often be the case that in actual application race may take on a larger role than the underlying evidence may justify.
Yet even when race is a substantial factor, and thus even when its exclusion would significantly decrease law-enforcement efficiency, the consequence of excluding race from the profile is an increase in crime only if we are holding cost and efficiency constant. But if we exclude race and are willing to suffer a decrease in efficiency but not in crime, but at more cost and annoyance to everyone, there need not be any reduction in the ability to apprehend criminals.
It has become common to describe the great social issues of domestic politics, and in particular the racial issue, as presenting a conflict between the demands of liberty and equality. Dworkin has in mind the traditional definition of liberty as the absence of constraints places by a government upon what a man might do if the wants to (like the liberal Isaiah Berlin). This conception of liberty as license is neutral amongst the various activities a man might pursue, the various roads he might wish to walk. It diminishes a man’s liberty when we prevent him from talking i=or making love as he wishes, but it also diminishes his liberty when we prevent him from murdering or defaming others. These latter constraints may be justifiable, but only because they are compromises necessary to protect the liberty of security of others, and not because they do not, in themselves, infringe the independent value of liberty. Bentham said that any law whatsoever is an ‘infraction’ of liberty, and though some such infractions might be necessary, it is obscurantist to pretend that they are not infractions after all. In this neutral, all embracing sense of liberty as license, liberty and equality are plainly in competition. Laws are needed to protect equality, and laws are inevitably compromises of liberty. It would be an intolerable muddle, on the view of liberals like Berlin, to use the concept of liberty or freedom in such a way that we counted a loss of freedom only when men were prevented from doing something that we thought they ought to do. It would allow totalitarian governments to masquerade as liberal, simply by arguing that they prevent men from doing only what is wrong. For a true liberal, any constraint upon freedom is something that a decent government must regret, and keep to the minimum necessary to accommodate the other rights of its constituents.
If someone has a right to something, then it is wrong for the government to deny it to him even though it would be in the general interest to do so. This sense of a right (which might be called the anti-utilitarian concept of a right) seems to me very close to the sense of right principally used in political and legal writing and argument in recent years.
The vast bulk of the laws which diminish liberty are justified on utilitarian grounds, as being in the general interest or for the general welfare. The right to liberty is not competitive with strong rights, like the right to equality. Every law is, as Bentham said, an infraction of liberty, but we have a right to be protected against only fundamental or serious infractions. If the constraint on liberty is serious or severe enough, then it is indeed true that the government is not entitled to impose that constraint simply because that would be in the general interest. So there is, after all, a general right to liberty as such, provided that that right is restricted to important liberties or serious deprivations.
The idea of a right to liberty is a misconceived concept that does a dis-service to political thought in at least two ways. First, the idea creates a false sense of a necessary conflict between liberty and other values when social regulation is proposed. Second, the idea provides too easy an answer to the question of why we regard certain kinds of restraints as especially unjust.
When we say that someone has a right to speak his mind freely, in the relevant political sense, we mean that he is entitled to do so even if this would not be in the general interest. If we want to defend individual rights in the sense in which we claim them, then we must try to discover something beyond utility that argues for these rights.
First, there are a great many men and women and they undoubtedly form the majority even in democracies, who do not exercise political liberties that they have, and who would not count the loss of these liberties as especially grievous. Second, we lack a psychological theory which would justify and explain the idea that the loss of civil liberties, or any particular liberties, involves inevitable or even likely psychological damage. On the contrary, there is now a lively tradition in psychology, led by psychologists like Ronald Laing, who argue that a good deal of mental instability in modern societies may be traced to the demand for too much liberty rather than too little. In their account, the need to choose, which follow from liberty, is an unnecessary source of destructive tension.
The liberal conception of equality: governments must not only treat people with concern and respect, but with equal concern and respect. This is conception of equality, not of liberty as license, that they state.
Citizens governed by the liberal conception of equality each have a right to equal concern and respect. But there are two different rights that might be comprehended by that abstract right:
The right to equal treatment: the same distribution of goods or opportunities as anyone else has or is given.
The right to treatment as an equal: the right to equal concern and respect in the political decision about how these goods and opportunities are to be distributed.
Dworkin proposes that the right to treatment as an equal must be taken to be fundamental under the liberal conception of equality, and that the more restrictive right to equal treatment holds only in those special circumstances in which, for some special reason, it follows from the more fundamental right. He also proposes that individual rights to distinct liberties must be recognized only when the fundamental right to treatment as an equal can be shown to require these rights.
A government that respects the liberal conception of equality may properly constrain liberty only on certain very limited types of justification. Examples of political justifications:
Arguments of principle: the constraint is required to protect the distinct right of some individual who will be injured by the exercise of the liberty
Arguments of policy: constraints are required to reach some overall political goal to realize some state of affairs in which the community as a whole are better off by virtue of the constraint
Utilitarian arguments of policy: the community as a whole will be better off because more
of its citizens will have more of what they want overall, even though some of them will have less.
Ideal arguments of policy: the community will be better off, not because more of its members will have more of what they want, but because the community will be in some way closer to an ideal community, whether its members desire the improvement in question or not.
Utilitarian arguments fix on the fact that a particular constraint on liberty will make more people happier, or satisfy more of their preferences, depending upon whether psychological or preference utilitarianism is in play. It will not respect the right of everyone to be treated with equal concern and respect. If utilitarian arguments of policy are to be used to justify constraints on liberty, then care must be taken to insure that the utilitarian calculations on which the argument is based fix only on personal and ignore external preferences. However, personal and external preferences are sometimes so inextricably combined, that the discrimination is psychologically as well as institutionally impossible.
There is no such thing as any general right to liberty. The argument of any given specific liberty may therefore be entirely independent of the argument for any other. There can be no inconsistency in denying that a particular right to liberty exists while warmly defending a right to other liberties.
Equality is considered a principle of fundamental importance, which should form the basis of each legal system and which should be respected at all times by legislative, executive and judicial authorities. It is equally clear, however, that it is impossible to realize complete equality in practice. It is often contended that complete equality is not even desirable, because there would then be no scope for differences in personal ambitions, abilities or needs. There is thus a clear tension between the search for equality and the need to treat persons and cases differently (described as the ‘paradox of equality’ or the ‘dilemma of difference’).
In the legal discourse, formal inequality means that two cases are treated unequally, whereas formal equality means that two cases are treated equally. Whether or not the concrete cases show any substantive differences or similarities does not seem to be of any significance to the legal notion of formal equality.
The legal principle of substantive inequality does take factual differences into account. In legal argument, substantive inequality means that substantively unequal cases are treated in the same way. Formal equal treatment can sometimes constitute inequality: inequality in result, caused by the fact that insufficient account is being taken of the differences that in fact exist between groups, persons or situations. For this reason substantive equality is often termed equality in result, whereas formal equality might be described as equality in treatment.
A third important legal concept is that of procedural equality, which is closely related to the concept of equal opportunities. Unlike formal equality, procedural equality takes some account of the differences that exist between individuals, more specifically: of their differences in opportunities. Unlike substantive equality, however, complete equality in result is not the goal sought for. Even if people start from the same position and have equal opportunities, the final results that they achieve can differ.
In defining the legal principle of equality, a further distinction is often made between equality before the law and equality of the law:
Equality before the law (formal equality): requires that a rule be applied to all cases which satisfy the conditions laid down by the rule.
Equality of the law (substantive equality): it is unavoidable that each legislative act distinguishes between groups or cases. The principle of equality of the law concerns the reasonableness of such legislative distinctions and classifications.
The legal concept of equality is a comparative concept that presupposes plurality. The element of plurality is relevant at two different levels:
Equality presupposes plurality of subject, which means that a law, decision or act must always concern at least two different persons, groups of cases.
Equality presupposes plurality of object, which means that the relevant persons, groups or cases must be given a different treatment.
The comparative element is closely related to this: the equality principle requires a comparison to be made, either between the treatment that is given to the relevant groups, persons or cases, or between the groups, persons and cases themselves.
When can cases be considered ‘equal’? It is evident that ‘equality’ does not require cases to be identical. Cases are always both comparable and non-comparable, or both equal and unequal. In order to make a meaningful comparison between cases, persons and groups, and to give a meaningful answer to the question whether two or more cases are ‘equal’ or ‘comparable’, it is therefore necessary to establish what characteristics or features of the cases, persons or groups are relevant for the comparison. When choosing a standard of comparison, one should let oneself be guided by the direct context and aims of the difference in treatment. It can be taken as a rule that the proper standard of comparison can be detected by formulating the reason for making a distinction. If a suitable standard of comparison is chosen, it is relatively easy to determine whether cases are comparable or ‘equal’. The standard need only be applied to the factual situation to see whether or not the relevant characteristics are present. The result of this operation is a descriptive or factual determination of equality: by pointing to the standard of comparison and to the facts of the case, it can be determined that the cases (or the treatment) are factually or descriptively equal or unequal.
A conclusion as to how these cases should be treated (differently or similarly) is still to be reached. In order to do so, it is necessary to reason from an ‘is’ to an ‘ought’ – from the fact that two cases are relevantly equal or unequal (descriptive or factual equality), it must be reasoned that they ought to be treated in a certain way (prescriptive, normative or evaluative equality).
In legal theoretical writing, four approaches can be distinguished:
Descriptively equal cases ought to be treated equally, descriptively unequal cases ought to be treated unequally.
This approach is purely formal in character. Whether or not a case of formal treatment is reasonable in substance, is a question that is left unanswered by this approach.
Descriptively equal cases ought to be treated equally, descriptively unequal cases ought to be treated unequally, according to their degree of inequality.
This approach seems to rely on an element of formal justice, just like the first approach.
Descriptively equal cases ought to be treated equally, unless unequal treatment can be justified on the basis of objective and reasonable arguments. Descriptively unequal cases ought to be treated unequally.
Descriptively equal cases ought to be treated equally, unless unequal treatment can be justified on the basis of objective and reasonable arguments. Descriptively unequal cases ought to be treated unequally, according to their degree of inequality.
This approach accepts a fusion between the principle of equality and a substantive conception of justice. According to this approach, an element of substantive justice is always needed to reason from descriptive to prescriptive equality, or from an ‘is’ to an ‘ought’. Regardless of whether the cases are descriptively equal or unequal, the approach always requires an assessment of the justice of treating them in a certain way. This approach is often used in legal practice, although the concept of substantive justice is then usually replaced by the terminology of ‘reasonableness’. It is regularly stated in case-law and legal texts that equal or unequal treatment should be ‘reasonable’ in order to be compatible with the equality principle, or that a difference in treatment is only allowable if it is ‘objectively and reasonably justified’.
It is suggested that the legal issue in Kostantin Markin v. Russia is broader than formal discrimination on the grounds of sex, and encompasses the use of gendered stereotypes of military servicemen and military servicewomen. One stereotype that plays a role in this case is the traditional idea that women are responsible for household and children and the main task of men is to work outside of the home. Another stereotype that plays a role is the idea that fighting and military service is something for men and not for women.
Stereotypes are beliefs about the characteristics of groups of individuals; group members are not seen as individuals, but are judged on the basis of the group membership.
Gender stereotypes become problematic when they operate to deny individuals their human rights. This is the case when gender stereotypes reflect and/or induce sex discrimination. Gender stereotypes are related to discrimination in two ways: they can cause discrimination and they can be discriminatory in themselves. Often gender stereotypes do both.
Stereotypes act as control mechanisms: they limit individuals in their options.
The harm of stereotyping men and women in traditional gender roles is that this leads to a lack of support for people who do not fulfill traditional roles. Gender-role stereotypes construct men or women as inferior in certain spheres of life; they induce and justify gender inequality. This is no different in cases of so-called positive discrimination.
Stereotypes are often translated into laws and regulations and/or used as rationalization for these regulations. As a result, there are two things to be especially alert to:
Blanket restrictions: when stereotypes are transformed into law, this often takes the form of blanket restrictions on what groups of individuals can do. What stereotypes and blanket restrictions have in common is that both are based on certain assumptions about groups of individuals.
Justifications: another way stereotypes often play a role in the legal process is when they are invoked as justification for an act of disparate treatment.
There are a number of things that courts can do to ensure that harmful gender stereotypes are not validated, in addition to avoiding stereotyping in their own reasoning. In order to be able to address the wrong of stereotyping, it is important that judgments name the problem of stereotyping and recognize the harm that is done through it.
Gender stereotypes are not permitted to function as justification for a difference in treatment between men and women.
Stereotypes of men can harm both men and women and vice versa.
The Convention on the Elimination of all Forms of Discrimination Against Women (CEDAW) contains an obligation for States Parties to address harmful gender stereotypes. The CEDAW Committee has observed: “The principle of equality between men and women, or gender equality, entails the concept that all human beings, regardless of sex, are free to develop their personal abilities, pursue their professional careers and make choices without the limitations set by stereotypes, rigid gender roles and prejudices.
The CEDAW Committee has explained that three State parties’ obligations are central to achieving the object and purpose of the Convention:
To ensure that there is no direct or indirect discrimination against women,
To improve the de facto position of women through concrete and effective policies and programmes,
To address prevailing gender relations and the persistence of gender-based stereotypes that affect women not only through individual acts by individuals but also in law, and legal and societal structures and institutions. This obligation has been called “transformative equality”: equality as transformation aims to eliminate the root causes of gender discrimination and gendered disadvantage.
The Committee is of the opinion “that a shift from a focus on women primarily as wives and mothers to individuals and actors equal to men in society is required for the full implementation of the Convention and the achievement of equality of women and men.
There is a rich body of materials that is developed by both international organizations and international courts acknowledging and addressing the link between gender stereotypes and gender inequality.
Intersectional discrimination: discrimination based on several grounds that interact with each other, and produce specific types of discrimination.
In the present case, the difference in treatment that the applicant complains of cannot be reduced to either military status or sex, but is rather the result of a mix of these two grounds. The problem is that when discrimination on the basis of sex and discrimination on the basis of military status are disaggregated, the stereotypes concerning military servicewomen will recede to the background.
The CEDAW Committee approaches discrimination holistically. Such an approach aims at more than purely formal equality and seeks to combat the structural causes of discrimination.
By creating men’s and women’s rooms of the same size, society guarantees that individual women will be worse off than individual men. Part of women’s demand for bathrooms cannot exist for men because only women menstruate. Women make trips to the rest room so secure hygienic and socially appropriate adaptations to this physical fact. And because men’s physiology suits them for the use of urinals, a large number of men can be serviced by a relatively small physical space. Women in our society use toilets to urinate, and toilets require a larger area than urinals. By distributing a resource equally, an unequal result is structurally guaranteed. Equality of square feet to the genders delivers women special burdens of physical discomfort, social disadvantage, psychological anxiety, compromised access to the full product and public ridicule.
An obvious solution, the “liberal” policy, is to make women’s rooms larger than men’s. An alternative solution (“conservative”), would be for women to change the way they do things, rather than for society to change the structuring of rest room space.
As the issues become subtle, reasonable people come to disagree on who should do what and what community expense should be incurred to achieve parity. Besides careful analysis, equality also involves a decision as to who is going to change and in what way.
Je vertrek voorbereiden of je verzekering afsluiten bij studie, stage of onderzoek in het buitenland
Study or work abroad? check your insurance options with The JoHo Foundation
Add new contribution