I Nearly everyone agrees that discrimination is sometimes morally wrong. But despite the breadth of consensus on this point, there is remarkably little agreement about what makes something an instance of discrimination, or about what makes an instance of discrimination wrong when it is. I propose that discrimination is differential treatment that has a particular kind of explanatory connection to the agent's differential ascription of some property to the discriminatees, and I identify two things that are sometimes wrong with it. First, in paradigm cases of wrongful discrimination, the act manifests disrespect for the standing as persons of those who are discriminated against. It does this because it stems from a failure to recognize either their equal value, and the presumptive claim to equal consideration this entails, or their standing as autonomous individuals, that is, as part authors of their own lives. In such cases, discrimination is wrong irrespective of both its effects and its social meaning. Second, much discrimination is morally objectionable simply on account of its contingent consequences. It is bad, that is, because it does unjustified harm, or brings about distributions that are unfair. I offer an account of the moral case against racial profiling in law enforcement as an example of how we should think about this second sort of wrongful discrimination.

Les études françaises sur la discrimination en milieu médical s’intéressent aux refus de soins des libéraux selon le système d’assurance des patients. Les recherches anglo-saxonnes examinent les différences de prises en charge sous l’angle de l’ethnie du patient. Burgess (2010) suggère que les discriminations en milieu médical peuvent être plus fréquentes lorsque le soignant est placé sous un niveau élevé de charge cognitive. Les recherches dans le domaine de la psychologie du travail et l’ergonomie indiquent que dans les cas de surcharge de travail, l’opérateur modifie son activité. Les comportements qualifiés de discriminatoires pourraient alors être le résultat des mécanismes de régulation de la charge de travail. Deux vagues d’entretiens auprès de soignants (N = 22 pour la première étude et N = 6 pour la seconde) indiquent un lien de leur part entre discrimination et charge de travail. Les observations menées en milieu hospitalier et couplées à un questionnaire et le NASA-TLX (N = 121) ont permis de constater que les discriminations en milieu médical sont faibles. Elles concernent uniquement la dimension relationnelle, les actes médicaux (examens, soins) sont réalisés normalement. Les variations constatées dans les comportements des soignants adviennent à l’encontre de patients stigmatisés et difficiles à gérer. / The French studies about discrimination in medical settings are focused on the general practitioners’ refusal to provide care depending on the patient’ insurance cover. Anglo-Saxon research examines the differences in care according to the patient’s ethnicity. Burgess (2010) suggests that discriminations in healthcare settings may happen more frequently when health care providers are put under high levels of cognitive load. Research in the fields of occupational psychology and ergonomics indicates that in cases of work overload, workers change their activity. Behavior described as discriminatory might result from regulation mechanisms of work overload regulation mechanisms. Two waves of interviews with healthcare providers (N = 22 for the first time and N = 6 for the second time) point to a link between discrimination and workload. Observations conducted in hospital settings and associated with a questionnaire as well as the NASA-TLX (N = 121) have shown that discriminations in a medical environment are limited. They are only related to the interpersonal aspect of care: medical acts (examination, care) are properly completed. The variations in the healthcare providers’ behavior occur when dealing with stigmatized patients and patients who are difficult to manage.