Les ateliers de l'éthique The Ethics Forum DISCRIMINATION, EMOTION, AND HEALTH INEQUITIES

In this paper I argue that certain ways in which the relationship among discrimination, emotions and health is presented can undermine equity. I identify a model of this relationship the discrimination-emotion-health model - and claim that while the model is important for understanding the detrimental impact that discrimination and oppression can have on emotions and health, certain implications of the model are troubling. I identify six critiques of the model, and show that equity could be undermined, for example, when stereotypes of the oppressed are reinforced and the experiences of the privileged are normalized. I then assess the implications of my analysis of the model and its critique for a framework of health equity, demonstrating what such a ABSTRACT: In this paper I argue that certain ways in which the relationship among discrimination, emotions and health is presented can undermine equity. I identify a model of this relationship the discrimination-emotion-health model - and claim that while the model is important for understanding the detrimental impact that discrimination and oppression can have on emotions and health, certain implications of the model are troubling. I identify six critiques of the model, and show that equity could be undermined, for example, when stereotypes of the oppressed are reinforced and the experiences of the privileged are normalized.I then assess the implications of my analysis of the model and its critique for a framework of health equity, demonstrating what such a framework would need to look like in order for it to best represent discrimination as a psychosocial determinant of health. represent discrimination as a psychosocial determinant of health. I highlight six implications of the critique for theories of including the recognition that policies have and that oppressive systems can damage the privileged, as well as the oppressed. Thinking about the model and its critique according to such a theory will provide health policymakers and practitioners with a nuanced way of thinking about the role of emotions in health disparities, which can help with the development of suitable policies and guidance for health equity.


ABSTRACT:
In this paper I argue that certain ways in which the relationship among discrimination, emotions and health is presented can undermine equity. I identify a model of this relationship the discrimination-emotion-health model -and claim that while the model is important for understanding the detrimental impact that discrimination and oppression can have on emotions and health, certain implications of the model are troubling. I identify six critiques of the model, and show that equity could be undermined, for example, when stereotypes of the oppressed are reinforced and the experiences of the privileged are normalized. I then assess the implications of my analysis of the model and its critique for a framework of health equity, demonstrating what such a framework would need to look like in order for it to best represent discrimination as a psychosocial determinant of health.
Social epidemiology, medical sociology, and other social sciences are systematically identifying and mapping out the specific causal pathways that demonstrate what oppressed people feel and know as lived experience-that violations of respect, in forms such as discrimination, are bad for one's health. In the US, for example, much research has been conducted on how racial discrimination against people of colour impacts negatively on their health. Included is research that demonstrates evidence for psychosocial determinants of health-not only does discrimination function to skew access to resources such as healthcare and opportunities for healthy living, but it also creates stress and negative emotional states such as psychological distress, which can in and of themselves be manifestations of ill-health (e.g., in the form of psychological ill-health), or which can increase the risk of physiological disease and impairment, or both. A model of the way in which discrimination can affect emotions and in turn health can be called the discrimination-emotion-health model.
Identifying the causal pathways for specific health outcomes provides essential information to better understand the bad consequences of what is already a manifest injustice-wrongful discrimination-as well as to intervene to improve health outcomes. In this paper, however, I identify an unrecognized and troubling problem: that certain ways in which the relationship among discrimination, emotion, and health is presented actually reinforce inequities-for example, by reinforcing stereotypes, by prescribing how the disadvantaged should feel and behave, and by presenting the emotional life of the privileged as "normal." In the first and second sections of the paper, I identify the discriminationemotion-health model and its implications. This model appears to underlie much public-health literature on discrimination as a psychosocial determinant of health but is not made explicit. In the third section I argue that an appreciation of the political dimensions of emotions indicates that the model can have troubling implications. I will provide a critique focusing on six concerns: the reinforcement, agency, respectability, dissidence, fragility of privilege, and object of emotions critiques. In the final section, I will demonstrate how this political critique helps us to develop the features required of a theory of health equity in order for it to best represent discrimination as a psychosocial determinant of health. I highlight six implications of the critique for theories of health equity, including the recognition that health policies have expressive value, and that oppressive systems can damage the privileged, as well as the oppressed. Thinking about the model and its critique according to such a theory will provide health policymakers and practitioners with a nuanced way of thinking about the role of emotions in health disparities, which can help with the development of suitable policies and guidance for health equity. Wilkinson, 1999;Marmot, 2005;LaVeist and Isaac, 2012;Berkman et al., 2014). Social determinants of health, "conditions in the environments in which people are born, live, learn, work, play, worship, and age that affect a wide range of health, functioning, and quality-of-life outcomes and risks," include factors such as discrimination and socioeconomic conditions (Office of Disease Prevention and Health Promotion; see also Marmot and Wilkinson 1999;LaVeist and Isaac, 2012;Berkman et al., 2014). Two of the primary categories of social determinants of health are material and psychosocial (Wilkinson, 1997;Marmot and Wilkinson, 2001).
Material social determinants of health can be described as the resources available to individuals and communities (e.g., healthcare, nutrition, and housing). Psychosocial determinants affect the individual psychologically, and this in turn impacts on health. The impact on health includes psychological health, which may seem obvious, and also physiological health (measured in terms of an increased risk of mortality from cardiovascular disease, for example). Claims that are made that relative position in a social hierarchy affects health, for example, are often claims about psychosocial determinants of health. It is not (only) the actual material circumstances that are impacting health, but (also) something related to how the situation makes the individuals think and feel about their social position that has this impact. Note that the same determinant can be both material and psychosocial-for example, experiencing homelessness can have a material impact (e.g., one is exposed to cold weather) and a psychosocial impact (e.g., one is anxious about experiencing homelessness and all that it entails).
Research is honing in on the various and complex ways in which social determinants impact the body (e.g., Adler and Newman, 2002;Berkman et al., 2014;Goosby et al., 2018). One of the proposed causal pathways for psychosocial determinants of health is via emotions. Research indicates that the experience of "negative emotions" 1 (e.g., psychological distress) can exacerbate poor health outcomes or increase the likelihood that certain poor health outcomes will develop (Kubzansky et al., 2014). For example, anxiety (Roest and et al., 2010) and anger and hostility (Chida and Steptoe, 2009) have been associated with an increased incidence of coronary heart disease. Furthermore, positive emotions appear to have positive, protective effects on health (Kubzansky et al., 2014, p. 324). It is not only one's experience and expression of negative or positive emotions that is at play, but also, and perhaps primarily, the overarching mechanisms by which emotions are regulated. Being able to monitor and manage emotions has an influence on health-a lack of emotional self-regulation appears to contribute to poor health outcomes, while strong self-regulation appears to protect health (Kubzansky et al., 2014, p. 324-325;337-338).
Emotions and their regulation can affect both healthy and unhealthy populations. They can influence the onset of certain diseases in healthy populations as well as exacerbate current diseases or increase the likelihood of additional diseases developing in unhealthy populations. Depression and anxiety have, for example, been identified as increasing the risk of the onset of coronary heart disease (Kubzansky et al., 2014, p. 335-336). 2 While there is a lack of evidence to indicate that emotions are related to the onset of cancers, there is evidence to indicate that psychological distress plays a role in the further development of cancer after its onset (Kubzansky et al., 2014, p. 342).
While genes and individual behaviours are likely to affect emotions and their impact on health, social epidemiology emphasizes the significant role that social factors play in patterning negative emotions, positive emotions, and emotional self-regulation. Numerous social stressors and combinations of those stressors influence emotions and their regulation. A model used to delineate the causal chain linking social stressors, emotions, and health, as described in this section, where stress influences emotion and its regulation, can be called the stressemotion-health model (Kubzansky et al., 2014, p. 326-327). Childhood trauma is a major stressor that impacts emotion-children who experience traumatic relationships with significant adults are at high risk of emotional dysregulation (Villalta et al., 2018). Discrimination and perceived discrimination are also stressors that can influence emotion and its regulation (Zilioli et al., 2017). 3 Before we explore discrimination in more detail, consider, as illustration, how the stress-emotion-health model could be seen to manifest in the lived experience of the socially disadvantaged. Here is how Darren McGarvey (BBC, 2017), a rapper and social commentator from Glasgow, who grew up in what he refers to as "the lower class," describes his first realization of how class differences in the UK manifest: It was… when I took my first trip across to the affluent side of Glasgow where I really got insight. The first thing I noticed was how calm it was... And my first thought … was, "all right, this is how people dress when they aren't afraid they are going to be stabbed." And for me that was a sort of real epiphany because I thought, okay, people here have more money but actually what they really have over me is an advantage, as they have an emotional reserve. They have an ability to absorb stress. They have this in-built resilience which I don't have because I am constantly in a state of fight or flight.
Notice a couple of particularly interesting and insightful points. First, not only does it seem that being disadvantaged means you are likely to be exposed to more social stressors than someone who is advantaged, you also, McGarvey claims, have less effective coping mechanisms-"they have an ability to absorb stress." This sounds similar to what the stress-emotion-health model would refer to as the importance of emotional self-regulation. Second, consider how he rejects the idea that the best way in which to describe his disadvantage is with reference to resources, such as his lack of income and wealth, although clearly these are disadvantages; rather, a primary disadvantage is related to emotions and, in turn, to its impact on health (e.g., being constantly in a physiological state of fight or flight).
There are different ways of understanding McGarvey's claims, however, and, in fact, his claims foreshadow some of the concerns I raise in section three. Consider that there may be an important difference between having "an ability to absorb stress," which sounds like it could overlap with emotional self-regulation, and having "an emotional reserve" because you are not exposed to a lot of stressors in the first place. 4 It is possible that those who are advantaged have both; however, in section three I will explore claims that, at least under certain circumstances, the advantaged break down easily in the face of stress precisely because they are not used to experiencing it and have not developed abilities to cope with it.

DISCRIMINATION, EMOTION, AND HEALTH
Discrimination, such as racial discrimination or discrimination on the basis of sexuality, has been identified as a significant social determinant of health (Krieger, 2014(Krieger, , 1999. It can have a material or psychosocial influence, or both. Discrimination (e.g., in the form of residential segregation) can influence where people live, leading people to live in areas with poor infrastructure, much pollution, exposure to environmental toxins, food deserts, a lack of safe spaces for exercise, and low-quality healthcare facilities. In this way, discrimination is a cause of material determinants of health and of exposure to toxic environments. However, responses to discrimination, such as psychological distress, mean that it can also be a psychosocial determinant of health. Here the stress-emotionhealth model becomes significant-the specific stress involved is discrimination.
Consider racial discrimination as an example. Exposure to the stressors of everyday racism and microaggressions can make so-called negative emotions more likely to occur. In turn, these emotions, as we have seen, are linked to negative health outcomes. This includes mental ill-health-for example, an anxiety disorder (Levine and et al., 2014) or depressive symptoms (Nadal et al., 2014). Distress can also lead to or impact on poor health beyond constituting mental illhealth-for example, chronic worry about racial discrimination could be one of the factors that explains Black-White disparities in preterm birth (Braveman et al., 2017). While a majority of research on the impact of discrimination on health in the US has been done on racial discrimination, there is also evidence that other forms of discrimination (e.g., those based on sexuality, gender, age, religion, class, disability, and immigrant status) have similar effects on health. There is only limited research, however, on some of these forms of discrimination, such as age and disability, or on combinations of them (Krieger, 2014, p. 61-67, 81-105).
When the stress is discrimination, I will call the stress-emotion-health model the discrimination-emotion-health model. While this model is seldom explicit in the public health literature something like this model often underlies research on the influence of discrimination on emotions and health. 5

THE POLITICAL CRITIQUE OF THE DISCRIMINATION-EMOTION-HEALTH MODEL
The stress-emotion-health model, and more particularly the discriminationemotion-health model (from here on I will refer to the latter as "the model"), can clearly be significant in helping to develop our knowledge about population health and health inequity. Furthermore, the model often dovetails with humanities literature emphasizing the internalization of oppression and its psychological and emotional burdens, 6 by demonstrating some of the likely biological and psychological pathways that connect disadvantage and the risk of poor health outcomes. In this section of the paper, however, I will argue that, seen in isolation from a broader social context, and particularly in light of critical theory and political philosophy on race, class, gender, and the emotions, the model has possible implications that raise some concerns.
Criticisms of aspects related to the model are not unusual-for example, methodological concerns have been raised about measuring emotions primarily through self-report assessments (Kubzansky et al., 2014, p. 330-331), and concerns have also been expressed about how epidemiological research on discrimination and health disparities primarily focuses on interpersonal discrimination, rather than on structural discrimination (Krieger, 2014). The critique I formulate here is different, however, although at times it overlaps with some of the ongoing criticisms; in the final section of this paper, I will discuss how the criticism related to the neglect of structural discrimination is relevant to my critique.
The model could be taken to have the following implications: First, it takes as given that there are negative and positive emotions and it is fairly clear which are which. Second, it understands the relationship among social stressors, emotions, and health as following this pattern: On the one hand, the disadvantaged, due to their disadvantage, have an increased likelihood of experiencing negative emotions and a decreased likelihood of experiencing positive emotions. They are also at risk of emotional dysregulation. On the other hand, then, it follows that at least relatively, the privileged, due to their privilege, have a decreased likelihood of experiencing negative emotions, an increased likelihood of experiencing positive emotions, and an enhanced ability to regulate their emotions. For the disadvantaged the risk of negative emotional states and dysregulation can exacerbate, or increase, the likelihood of the incidence of illhealth, while the privileged receive relatively greater protection from these risks. 7 I am not saying that researchers, practitioners, or health policymakers who implicitly assume such a model are necessarily committed to these implications-my claim is rather that these implications could follow from the model, and, more specifically, as I discuss in this section, where they do, they raise often-unrecognized and troubling concerns.
Before we investigate the critique, I will make four clarifications or qualifications about my claims. First, I consider "discrimination" and "oppression"; second, which oppressions; third, the use of political in "political critique"; and, last, theories of emotions.
While I am assessing the discrimination-emotion-health model I will often refer to "oppression" and to "oppressed people." Social groups on whom this epidemiological research tends to focus are often not only discriminated against, but also oppressed. Discrimination can be understood as differential treatment, and wrongful discrimination then would be wrongful differential treatment (Hellman, 2011). Not hiring someone for a job because they are black is wrongful racial discrimination.
While wrongful discrimination is problematic in and of itself, and while it often overlaps with oppression, discrimination does not fully represent oppression. Concerns with how women and people of colour are unfairly disadvantaged in society are often concerns of more than discrimination but also of gender and racial oppression, respectively. Oppression also includes violence and exploitation, for example (Young, 1990;Cudd, 2006), and neither of these are well represented as discrimination, or as merely discrimination. 8 Exploitation committed against certain social groups is not primarily wrongful because it is wrongful differential treatment even though it is indeed wrongful differential treatment; it is because it is exploitation that it is wrongful-the treatment itself is morally wrong whether or not it is differential, although it becomes a concern of structural group oppression when one group (e.g., immigrants; people of colour) is more likely to suffer exploitation than another (citizens; white people) (see Haslanger, 2012, p. 311-338 on structural group oppression). Even if the epidemiological literature with which I am concerned is mainly focused on discrimination, in this critique I will refer to oppression as well, in an effort to recognize that often the discrimination being identified is part of systematic oppression. Moreover, analysis of the literature on oppression has necessitated this critique-it is when we assess the relationship between oppression and emotions that aspects of the critique become apparent. There is more to be said about the importance of considering oppression, but I will discuss this in the final section of this paper; preliminarily, I have explained my use of terms, foreshadowing my discussion of the implications of the critique later on.
Second, I will focus primarily, but not exclusively, on three forms of discrimination and oppression-racial, gender, and class based, as well as intersections of these. This does not mean that my claims about oppression, emotion, and health necessarily lack application to other axes of oppression, such as disability and sexuality; rather, much of the relevant literature that seems to apply well to my critique tends to centre around socially constructed race, gender, and class, although it may also apply further. I consider the particular oppressions I discuss to be examples of the categories of critique rather than fully representative of them. I do, however, also recognize that particular oppressions and intersections of those oppressions have unique histories and features. The particular point I aim to make in this paper is not, however, about one particular form of oppression but about oppression more generally-it would be fruitful, however, to explore these critiques according to each particular axis of oppression, and their intersections, to consider how applicable they are and where distinctions may lie.
Third, let me note why I refer to this as a political critique. I take political here to emphasize the relationship between the state and its residents (the people living within its territories, no matter their legal status). My use of this term is pragmatic-not all of the aspects of the critique below seem necessarily directly related to this relationship. However, by using the term "political" I want to emphasize that while discrimination and oppression can occur outside of relationships between the state and its residents, the kinds of discrimination and oppression that, as things stand, should cause us most concern morally, and which require the most urgent action, are systematic forms of group oppression. These oppressions, while they exist in everyday and interpersonal relationships, are very much a feature of the state, its agencies, its laws, its policies, and its communications.
A final clarification concerns philosophical theories of emotions. I take no direct stand in this piece on what kind of theory of emotions should be endorsed-this would be beyond the scope of this paper. However, it is worthwhile to acknowledge that some of the claims I make preclude certain theories of emotions and assume others. For example, my claims preclude physiological and sensation theories of emotions, where these claim that "emotion is considered primarily or exclusively a 'feeling'" (Calhoun and Solomon 1984, p. 9). They also preclude the notion that emotions are exclusively inner, private feelings, and that they are exclusively or primarily instinctive, rather than (also) learned (Calhoun and Solomon 1984, p. 14, 33). Instead, my claims assume that emotions have a cognitive basis (e.g., Nussbaum 2003) and that they are strongly socially shaped (e.g., Ahmed, 2015).

a. The Reinforcement Critique
We should be concerned if the discrimination-emotion-health model is used in such a way that problematic stereotypes of oppressed people and problematic connotations of certain emotions are reinforced. First, the stereotyping of social groups often includes stereotypes about their emotional states and their emotional regulation. Across numerous cultures, women have often been characterized as being "emotional" (Jaggar, 1989;Ahmed, 2015, p. 168-172, 195;Niedenthal and Ric, 2017, p. 247-271), which implies that they have an inability to control their emotional responses-for example, in being prone to crying. Black women, more specifically, are often depicted as angry (Moreton-Robinson, 2003, p. 70;Lorde, 2007Lorde, [1984. Working-class men in the UK are also often depicted as angry and hostile (Nayak, 2006, p. 823;Wollaston, 2018). The model could be interpreted as validating these stereotypes, encouraging the idea that, yes, people who are socially disadvantaged are indeed more likely to experience negative emotions or to struggle with emotional self-regulation. The model provides a more sympathetic account of the relationships between emotions and social groups than models that claim that these emotions are determined by innate characteristics-it is precisely because of social injustice that members of certain groups are likely to experience these problems with affective states; however, the problem remains that the model could be described in a way that endorses the stereotypes.
Second, we should also be concerned if the model not only reinforced stereotypes about oppressed people, but also reinforced problematic connotations of emotions and of emotional self-regulation. Consider that the model may take for granted what is a "negative" or "positive" emotion. Being emotional, crying, openly showing distress, and being angry are not only often associated with being a woman or with being a black woman, but they are also often associated with something negative, something that you are doing wrong. The discrimination-emotion-health model seems to reinforce ideas that these are indeed negative emotions, by directly labelling them as such and by emphasizing the health costs associated with them. I am not claiming that we have no reason to be concerned about the possible consequences of certain affective states associated with anger and anxiety. However, it is important to emphasize that notions of what are positive or negative emotions are not politically neutral, and which emotions should be encouraged and which discouraged is already imbued with social value. Particularly of concern is that these values are often likely to reflect the values or perceived characteristics of the privileged-for example, it is the stereotypically feminine mode of emotional being to be distressed or to be lacking in control over emotions that is considered negative. Analyses of oppression and emotion can question and complexify this-these emotions are appropriate (Srinivasan, 2017) and can, at least in certain ways, be "positive." Consider, for example, Audre Lorde on the productivity of anger in the face of oppression (Lorde, 2007) and Darren McGarvey on the justifiability of anger as a norm among the working class in the UK (McGarvey, 2017). Here the stereotypes about the oppressed and their tendencies to certain negative emotions are in fact embraced-yes, many African-American women are angry and, yes, so are working class men in the UK, but they are angry because that is the appropriate, and even productive, reaction to the injustices they suffer.
Endorsing the discrimination-emotion-health model needn't commit one to reinforcing troubling stereotypes, nor to the problems highlighted by the remaining critiques; however, it's important to recognize that there is a danger of this happening when one is endorsing the model, and that when one is discussing, researching, or acting on the literature on discrimination and health, that one take care to avoid the dangers associated with these critiques.

b. The Agency Critique
One could use the model to encourage social control over oppressed groups, even if unintentionally, thus interfering with their behaviour and undermining their agency. The model's users could be seen to promote certain attitudes and behaviours associated with emotions, for example, encouraging the avoidance of negative emotions, thus seeming to prescribe how oppressed people should feel and how they should be regulating their emotions. This can be called a problem of agency because it implies that the disadvantaged have a diminished capacity for intentional action-at least in terms of the experience and expression of emotions and of the regulation of these-and for this reason, their behaviour needs to be externally influenced so that they are able to act healthfully.
Of course, public policies often intend to influence, even dictate, behaviour (e.g., requiring seatbelt usage). The concern here is not, however, that the behaviour of the population as a whole is being influenced; rather, the concern is related to inequity between or among social groups. When the focus is on how discrimination influences emotion and health, the control that it might imply over behaviour is specifically related to the behaviour of oppressed people-they are the ones, according to the model, suffering a greater risk of health problems due to the experience of negative affective states and emotional dysregulation, as well as to the lack of protection that positive states would provide them. Not only is the model in danger of justifying control of oppressed groups, it also seems to normalize the behaviour of the privileged-it is they who seemingly experience healthy, normal emotional states-and pathologizes the behaviour of the oppressed. The analogy with seatbelt regulations, for example, would seem more relevant if the disadvantaged were the only group who were encouraged to wear seatbelts. A further point to consider is that emotions and emotional regulation are somewhat morally different from behaviours such as seatbelt usage. Which emotional states we experience and how we express and regulate those emotions can be a part of our identity in a way that wearing seatbelts appears not to be. Thus, this is a concern not merely about interfering with our independence as agents, but interfering with our identities. Public health efforts that aim to help individuals regulate their emotions might be not only ignoring that those emotions are apt and justifiable, but also undermining individuals' identities, trying to make them into different, albeit, healthier people. 9 In fact, the causal picture that the model could put forward may be flawed in a further way that does not take the relationship between agents' identities and emotion into account. Here let's return to an example of class. Using the work of Raymond Williams and Annette Kuhn as her references, Beverley Skeggs (1997) claims that when it comes to the relationship between class and emotions such as anxiety, it is not merely that being working class causes anxiety but that class is constituted by certain emotions. Class is, among other things, a "'structure of feeling,'" and being working class means that "doubt, anxiety and fear inform the production of subjectivity" (Skeggs, 1997, p. 6). How should we understand this? What is significant here is that individuals shouldn't be seen (merely) as agents who encounter the world and then whose subjectivity and identity are influenced by the world via emotions; rather, who that agent is and how that agent's identity has been shaped from the start is already impacted, among other ways, emotionally, by how the agent is "classed." 10 Emotions and disadvantage are then deeply entrenched in a person's socially determined identity and in how that identity shapes that person's subjectivity.

132
This can have implications for how we try to resolve the problems caused by social stressors-when we think of a pathway from stress to emotion to health it seems as if intervention at the point between stress and emotion (for example, providing enhanced access to mindfulness training) might be particularly fruitful. However, when we emphasize that the agent's identity has already been partially constituted by expectations of emotions associated with social grouping, then trying to intervene to stop ongoing stressors influencing emotion becomes a less appropriate approach. This does not mean that such interventions should not be implemented-they may well need to be-however, as I discuss in the final section of the paper, they are often second-best solutions to what are major structural injustices. Ideally, it is the structural injustices rather than primarily problems of an individual's health that need to be addressed, and focusing on the individual's health can interfere with that individual's agency and identity, even though it may improve health.
It seems, however, that by describing class in this way we might be making the problem of agency even worse-the implication is not only that oppression can influence behaviour, but also that what it means to be oppressed is constituted partially by particular emotional expectations. Does this critique not have even more troubling consequences for the free agency of the oppressed than the possible implications of the model do? This, however, would miss an important part of the critique. The claim is not that the emotional lives of only those who are disadvantaged-e.g., the working classes-are partially constituted by social structure. Rather, class is, among other things, a structure of feeling, and that means all classes, including the privileged (the upper and middle classes). Which feelings, however, depends on which class. In other words, while the discrimination-emotion-health model could be assumed to imply that the privileged have an increased likelihood of experiencing healthy emotions and strong emotional self-regulation, whereas the disadvantaged have an increased likelihood of experiencing unhealthy emotions and poor emotional self-regulation, as if the privileged are unfettered and their behaviour is "normal," the claim here is that everyone's emotional life is shaped by privilege and disadvantage. The ways in which the emotions and subjectivity of the privileged are shaped is not necessarily "normal," even if in many ways it represents or entrenches their privilege. The last two critiques-fragility of privilege and object of emotions-will explore this point in more detail.

c. The Respectability Critique
A third concern is that the model may unintentionally support respectability politics. Marginalized peoples are often expected to act in a so-called respectable way, which can require them to disassociate themselves from the values and practices of their own communities and cultures and to act according to the norms of the privileged. Besides reinforcing problematic notions of what it means to be a citizen, this could be self-defeating, as the model may then be encouraging frustration, anger, and distress, for example, at having to maintain respectability, and these are some of the very emotional states that, it purports, can negatively impact health.
Critical theories of class as well as race express concern that the exemplar of a citizen demands particular attitudes and behaviours (Young 1990;Skeggs, 1997;Cooper, 2017). Not only do these demands encourage people to act in ways that may not suit their particular personalities, but these attitudes and behaviours are often those typically associated with privilege. The ways in which the privileged tend to act in a particular society are the basis for the norms to which everyone else is subject in order, for example, to be respected, to be taken seriously, and to garner valuable social opportunities, such as jobs and education. These can be called norms of respectability.
Consider, for example, the behaviours, attitudes, and preferences that are often associated with being highly educated and well-to-do in the US-these will differ at least somewhat from one community to the other, but among them are likely to be expectations about emotional states and their regulation, including expectations about being able to control the expression of emotions, particularly any strong emotions. Here the respectability critique expresses a concern independent of the agency critique; it is not only that the agency of the disadvantaged is undermined, but also that how they are expected to behave-respectably-is independently troubling.
The first concern from the respectability critique is that the discriminationemotion-health model can be seen to encourage the oppressed to act respectably-that is, to adopt the particular norms of respectability determined by dominant values. The second concern is that this emphasis on being respectable may in fact create the problem that it seems that the model would want us to rally against-consider that the pressure to act respectably could generate frustration and anxiety, as well as the suppression of emotions, thus seemingly creating greater risks to health. Indeed, there is growing evidence that the need to act respectably has health costs for African-Americans (Lee and Hicken, 2016). 11

d. The Dissidence Critique
As the agency and respectability critiques claim, the model may be used to prescribe feelings and behaviours to the disadvantaged. An overlapping yet independent concern is that the model can be used to encourage them to behave in particular ways in the face of injustice when health becomes prioritized over other values. Indeed, even more specifically, it appears that the model can be used to shape their responses to the discrimination with which it is concerned. If the advice that follows from the model is that negative emotions such as anger should be avoided and that high-intensity emotion should be controlled in the name of health, for example, then it may sound as if the model encourages passivity in the face of injustice and brands anger and the motivation to protest as pathological. Dissidence, except in its most subdued forms, may be undermined.
One could argue in response to this concern that by indicating how emotions are linked to health outcomes, the model could be used to provide important

V O L U M E 1 3 N U M É R O 3 A U T O M N E / F A L L
guidance on how best to pursue dissidence and protest. The advice that might follow from the model could emphasize that there are healthier and unhealthier responses to injustice, and that in the face of injustice, while dissidence may be necessary, one would do best to regulate one's emotions and to avoid negative emotions, thus protesting and resisting injustice in a calm (and healthy) manner.
While the burdens of dissidence against injustice, which include emotional and health burdens, should be recognized and further researched, 12 this response to the dissidence critique is not ideal. First, it still seems subject to both the agency and respectability critiques-the model appears to be prescribing what dissidents should feel and express if health is overly prioritized, and it does so in ways that seem to reinforce notions of respectability. The anger of dissidents, for example, and especially the anger of women of colour, is often used as an excuse by dominant groups not to listen to the claims made by oppressed groups (Lorde, 2007;Cherry, 2018). The model could be seen to provide a scientific and paternalistic basis for the privileged to continue to ignore dissidence that is expressed with anger or other intense emotions-"calm down, it's not healthy for you to express your claims in this way." I am not implying that the disadvantaged would be wrong to take the potential health consequences of intense emotions into account, nor for epidemiologists to warn about the links between emotions and health-rather, I want to point to the pitfalls associated with expressing the relationship among discrimination, emotions, and health in an overly simplistic way that would neglect the concerns I express in this section.
Second, this response remains vulnerable to the dissidence critique itself when we consider certain understandings of dissidence-that is, when we think of emotions as dissidence. Sara Ahmed (2015) argues, for example, that it is not that oppression is likely to spur emotions such as anger and fear, which in turn can encourage dissent or which can be expressed as part of resistance to oppression, but indeed, the actual experience and expression of these emotions is resistance. She argues that challenging unjust social norms means adopting a new emotional response to those norms, and it is that emotional response that partially constitutes one's resistance (Ahmed, 2015, p. 144-190, 196). What I am emphasizing here is that separating out emotions and ill-health from dissidence-thus, for example, trying to promote calm, healthy dissidence-can be problematized when we consider that dissidence can require the experience of certain so-called negative emotions, and thus the dissidence critique of the model (the claim that this model could be used to discourage dissidence) can apply to even a modified version of the model's implications.

e. Fragility of Privilege Critique
The model's potential implications may also neglect the ways in which privilege can make one particularly vulnerable to negative emotions. Here we can refer to the "fragility of privilege critique" relying on what Robin DiAngelo (2011; 2018) has influentially termed "white fragility." DiAngelo argues that white people in the US suffer a kind of fragility "in which even a minimum amount of racial stress becomes intolerable, triggering a range of defensive moves" (2011, p. 57). Among these defensive reactions, DiAngelo identifies the expression of emotions such as anger, fear, and guilt. An example of a situation that might trigger these defensive moves is when a person of colour describes the US as still severely hampered by racial oppression.
DiAngelo implies that white people thus have-in the terms familiar to the discrimination-emotion-health model-a reduced capacity for emotional self-regulation and a tendency towards negative emotions that stem from their privilege. This disrupts the idea that it is being oppressed that will increase the likelihood of negative emotions or troubling emotional traits. DiAngelo also argues not only that white people are lacking, at least under certain circumstances, in what she refers to as "psychosocial stamina" (p. 56), but also that people who are oppressed do have this kind of stamina.
DiAngelo is referring to a specific axis of oppression-racial oppression-and within a specific context, the US. We can see, however, that this kind of identification of privilege with fragility has also been identified in some other contexts of oppression. For example, think of some feminists' claims about the vulnerabilities of masculinity within patriarchy. Among the harms to men in patriarchal societies are those related to stress and emotions. The pressure to live up to expectations of masculinity, such as the expectation to be the primary provider for a family and to be strong and in control, which includes being very much in control of emotions, for example, has been identified as leading to "repress-ing…emotions, failing to develop emotionally" (Jaggar, 1989) and even promoting violence (Miles, 1992). Public health research also indicates that women may cope better with stress because they tend to employ better coping strategies such as seeking social support (see, e.g., Williams 2003); one of the reasons that this may be the case is related to gendered norms-for example, the norm that men are relatively discouraged from seeking support from others. One can say, then, that men in patriarchy are often likely to be emotionally impaired, despite their overall privilege.
A noteworthy corollary of the notion that privilege can make one fragile is that moving from a position of disadvantage to one of advantage also implies vulnerability. Consider, for example, the way in which Lynsey Hanley (2017) describes moving from the UK working class, in which she was born and raised, to the middle class: "I am… one of 'the uprooted and anxious': at once socially mobile and psychologically stuck, or at least divided, somewhere in between our place of origin and the place we must inhabit in order to 'get on'" (p. xiv). Social mobility, she claims, comes with its own emotional problems. When Hanley became more class privileged, she did not also become emotionally privileged, despite the typical association of being of higher class and being less likely to suffer (at least certain) negative emotions. Note that she is not claiming that her social mobility caused anxiety because, as someone who started off as working class, she does not have the tools to regulate her emotions, tools which someone who started off privileged might be more likely to have, according to our model. Her claim is that in order to be middle class, one has to disassociate one's self from one's working-class values and background, and this causes anxiety; thus, this risk of anxiety becomes a necessary part of climbing the class ladder. According to Hanley, then, social immobility in the form of working-class people remaining working class does psychological and emotional damage (p.xii) as one might expect, considering the discrimination-emotion-health model, but so does social mobility, a more surprising claim.
By including Hanley's claims, we can say that the fragility critique means that being privileged can be associated with emotional problems, and so can becoming privileged. One should not then assume-as the discrimination-emotionhealth model might be taken to imply-that disadvantage should necessarily be associated with an increased likelihood of emotional problems, and privilege with healthy emotions.

f. Object of Emotion Critique
A final concern is the idea that the person for whom or the thing for which we should feel emotion is already imbued with social and political value. Here I am thinking of claims, such as those expressed by Judith Butler (2006) and Sara Ahmed (2015), that emotions are often the products of social norms that aim to uphold or create power relationships. In this sense, the objects of emotions-e.g., the people for whom one should feel emotions-are socially determined and reflect which lives are valued and which are not. According to this critique, then, the privileged often exclude the disadvantaged as the objects of emotions like grief and compassion.
Consider, for example, the journalist Seymour M. Hersh's reaction to the charge sheet brought against one the US soldiers responsible for the My Lai massacre during the Vietnam War, which asserted that he was being charged with the premeditated murder of 109 "Oriental" human beings: "Did the Army mean to suggest that one 'Oriental' life was somewhat worth less than that of a white American? It was an ugly adjective" (Hersh 2018, p. 57). The object of emotion critique claims that the use of the word "Oriental" marks out the victims of the massacre as undeserving of the same emotional response as (white) Americans should have for the murder of (white) Americans. One need not turn to the 1960s for relevant examples-when heads of state, here Donald Trump and David Cameron respectively, describe migrants as "rapists" (Jacobs, 2018) and use the collective noun "swarms" to refer to them (BBC, 2015). Among other effects, they indicate that migrants don't deserve to be objects of emotions such as compassion, love, pity, and grief; the objects of emotions such as these are human beings, implied in this case to be limited to citizens. While migrants are denied status as objects of these emotions, indicating that their lives have less value, this kind of language seems to also encourage particular emotions towards them-e.g., while compassion and grief may be denied them, hatred, outrage, and fear seem appropriate when we are dealing with "rapists" and "swarms." This can be termed a critique of the discrimination-emotion-stress model because it complexifies the idea that if privilege does have an influence on emotion, then it is likely to advantage the privileged. As this critique and the previous onethe fragility of privilege critique-indicate that there is potential for privilege to impair the emotional lives of the privileged. In the case of the object of emotion critique, the privileged appear more likely to have a restricted and troubling emotional range towards the oppressed and disadvantaged. Thus, the privileged do not have so-called normal or healthy emotions in this respect; they have an impaired emotional range. This insight-that the privileged aren't necessarily emotionally privileged-is one of the significant implications of the political critique that I explore in the next and final section of this paper.

IMPLICATIONS OF THE POLITICAL CRITIQUE FOR HEALTH EQUITY
In the previous section I highlighted six critiques of potential implications of the discrimination-emotion-health model. I want to emphasize that my claim is not that the model is wrong-indeed, I believe it provides a highly significant perspective on the relationship among discrimination, emotions, and health. However, I am claiming that, seen in isolation, it provides a very partial picture of these relationships and that the implications of this partial picture can be troubling. The primary aims of this paper were to identify the often-implicit model and its possible implications, and to highlight my critique-accomplished in sections 1-2 and 3, respectively. In this final section, I explore what we can learn from the critique in order to represent the relationship among discrimination, emotion, and health more fully. Particularly, the political critique has implications for how we should understand health equity/inequity. While much of the analysis in this section is (necessarily) theoretical, there are likely to be significant implications for practice-for example, for epidemiologists in conducting research, for public health practitioners in developing guidance for individuals and communities, and for health policymakers.
I will highlight six implications of my critique for an understanding of how a framework of health equity/inequity should be specified and I will demonstrate how positioning our understanding of the relationships among discrimination, emotion, and health within such a framework will help us to avoid troubling implications of the model.
First, health inequity needs to be understood as part of a wider and pluralist notion of social injustice. 13 In this particular case, the inequities associated with relationships of discrimination, emotion, and health are best understood in the context of social injustice more generally and not merely as issues of health inequity. By a pluralist notion of social justice, I refer to a specific kind of pluralism, dimension pluralism-that is, pluralism in the dimensions of an individual's well-being or capabilities that are relevant to equity. 14 According to dimension pluralism, when we aim to determine what is equitable we need to consider each of these dimensions, which could include being healthy, being respected and having self-respect, and being autonomous. 15 Health is seen as only one of a number of significant morally valuable dimensions in an individual's life, not as the only value or as necessarily the most important value. We can see how this is related to the political critique of the discrimination-emotionhealth model by considering as an example the dissidence critique. Where this critique seems particularly relevant is in cases where we use the model combined with health exceptionalism, which means that health is treated as the only or the most important dimension. When we do so, we promote the idea that an individual's health should be prioritized above all else, including the individual's dissidence in the face of injustice. A dimension-pluralist notion of justice, however, helps us to recognize that health is only one value, so we are more likely to recognize that promoting the best health outcome could undermine other values, such as an individual's attempts to act autonomously and with selfrespect by protesting injustice. 16 Second, recognizing the significance of the model and of the critique necessitates consideration of second-order injustices. Second-order injustices occur due to another injustice (the first-order injustice). A concern that could be expressed about my argument is that taking both the model's possible implications and their critique into account might result in inconsistency. For example, if we take both health and self-respect into account, this could interfere with the provision of coherent guidance as to which should be promoted, because each is morally significant. Speaking about anger that is both apt and counterproductive, Amia Srinivasan (2017, p. 134) puts it this way: "We want to say both at once, and yet that will be to offer practically incoherent advice." However, rather than this being a problem with my argument, it is actually an important consideration that should be accommodated by a framework of health equity. It is not that there is only an initial injustice-discrimination-which in turn can have unjust consequences in the form of health inequity, but rather that discrimination can also often lead to second-order injustices in terms of the conflict an individual has to face in being forced to choose between two significant dimensions. It is another injustice, rather than merely a moral conflict, when it is the disadvantaged, those who are already suffering the first-order injustice, who then, in light of this injustice, must face a further burden-making a choice between independent, morally significant values. 17 Using Srinivasan's analysis of anger, we can refer to this kind of second-order injustice, where it is related to apt emotional responses, as an affective injustice: "It forces people, through no fault of their own, into profoundly difficult normative conflicts-an invidious choice between improving one's lot and justified rage" (p. 136).
Frameworks of health equity that can thus incorporate a notion of second-order affective injustice may do a better job of representing the relationships among discrimination, emotion, and health. Such a framework indicates that even when emotions and health are not negatively affected by discrimination, a significant and inequitable loss may still have occurred, such as when individuals have sacrificed their autonomy and self-respect in order to behave more healthfully.

V O L U M E 1 3 N U M É R O 3 A U T O M N E / F A L L
Third, by investigating the relationship of discrimination, emotion, and health within a wider context of social injustice, we can see that greater weight should be put on structural oppression in population and public health, with an emphasis here on both the "structural" and the "oppression." As I mentioned at the beginning of section 3, a central criticism of the epidemiological literature on discrimination is that it neglects structural discrimination. As Nancy Krieger (2014) explains, much research focuses on interpersonal discrimination, which refers to "encounters between individuals in which one person acts in an adversely discriminatory way toward another person" (p. 63). In contrast, there is limited research on structural discrimination, which refers to "discrimination enacted by institutions (e.g., laws or rules that impose adverse discrimination, by design, such as legalized racial discrimination, or in effect, such as the racialized impact of the New York Police Department's 'stop-and-frisk' policy"…) (ibid). While I agree with Krieger's concern, research should focus more on structural oppression, and not merely discrimination. As I mentioned in section 3, discrimination does not seem to capture oppression fully, and thus the various disadvantages associated with racial and other forms of injustice, such as exploitation and violence, would not be fully captured either, even if structural discrimination were taken into account.
Consider, for example, that it is by recognizing the structural nature of oppression that the reinforcement and respectability critiques become apparent. These critiques emphasize that emotions and emotional self-regulation are already imbued with social meaning, and, more specifically, that they are imbued with the meaning of systematic power relationships reflected in structures such as social norms (and not merely in instances of individual behaviours or particular policies). The oppressed are stereotyped as being a certain way emotionally (e.g., having a lack of emotional control or a tendency to express certain kinds of emotions), and that way is seen to be supposedly negative, while the stereotypes of the emotional ways of the privileged are considered to be "positive," and the oppressed are encouraged to be more like the privileged. Furthermore, an overemphasis on the interpersonal appears to focus our attention so much on individuals and their behaviours that it may seem that solutions to problems associated with discrimination, emotion, and health should primarily occur at an individual level as well; for example, that healthcare professionals should be encouraging individuals to develop tools of emotional self-regulation and to experience or express fewer negative emotions. An emphasis on the structural helps to show that ultimately structural solutions are required.
The fourth point, following on from this, is that health-focused interventions, while they may be necessary, are likely to be second-best solutions to the inequities associated with the model. The primary normative concern with social determinants of health, such as discrimination, is that they are not issues of health per se but rather issues of social injustice, and solutions to them should primarily be approached as such (Preda and Voigt, 2015). Racial or class discrimination (and oppression) are the normative problems, whether or not they lead to health problems. In other words, the health inequities that may be associated with discrimination are inequitable consequences of what is already an inequity-the wrongful discrimination itself. Ideally, the primary solution to these inequities is reducing or eliminating discrimination, as opposed to trying to solve the emotional and health concerns associated with them. This does not mean that we-as policymakers and public health or healthcare practitioners and ethicists-should not also be concerned about the health inequities themselves, and we may need to address them. But when we do so, we should recognize that we are providing second-best solutions-in other words, we are aiming to make the situation more equitable for the disadvantaged, but we are not thereby fully solving the problem. 18 Furthermore, what a good second-best solution is should be informed by the wider context of social injustice. Consider again the dissidence critique: if we did not take this critique into account, the discrimination-emotion-health model could be seen to imply that individuals should necessarily forego dissidence to promote better health. Ironically, however, we would then be giving them advice that may undermine one of the few means available to them to try to change an unjust system, such as through protest and civil disobedience.
Fifth, health policies, programmes, and guidance and the discourse surrounding them have expressive value that can influence how equitable they are. This means that the equity of a policy should be assessed not only according to its direct impact on health and healthcare, for example, but also according to the message that that policy expresses (for ease, I will refer in the following only to "policy" but my claims are not limited to policies-public health programmes and the guidance given by health agencies, for example, can also be expressive). In considering whether or not the policy should be pursued, what it expresses should be taken into account, and, more particularly, what it expresses in terms of the respect it shows for the disadvantaged is significant (Anderson, 1999;Voigt, 2018). Part of the concern the political critique is highlighting is expressive. It is asking us to consider what the model's implications may be expressing about the oppressed-is it, for example, problematically implying that the emotions and emotional self-regulation of the privileged tend to be normal, while those of the oppressed tend to be pathological and require change so that they resemble those of the privileged?
As an example of how this might apply, consider an explanation that may be provided to justify an increase in mental-health resources in a deprived urban area where African-Americans are a majority. If the model is used as a justification for this reform, expressive problems could follow-for example, according to the reinforcement and agency critiques, this justification can reinforce stereotypes about the disadvantaged being prone to a lack of emotional selfregulation and to pathological behaviour. Furthermore, this justification appears to be conveying a message to the community that the actual primary injustice they are suffering-racial oppression-will not be addressed, but only its consequences will be. In this case, there are reasons to use a justification that conveys a more respectful message-for example, by referring to redressing the likely healthcare resource deprivation and specifically the lack of mental-health resources in these areas as the central reason for the policy.
Last, when considering health inequity within the context of social injustice, we need to consider the influence of injustice on both the privileged and the disadvantaged. The fragility of privilege and object of emotion critiques especially demonstrate that emotional impairment can be associated with privilege in an unjust society. This fits in well with thinking about structural oppression-the entire social system, including its social norms and major political and economic institutions, reflects and maintains injustices, influencing and shaping everyone within it. The privileged, while mostly benefitted by the system, are not somehow outside of it; they too are shaped by it, and that shaping is not always positive and, indeed, can be damaging in different ways, including emotionally. Policy and guidance that promote the idea that only the emotional states and traits of the oppressed are influenced by injustice, and that the privileged are seemingly autonomous and emotionally "normal" are likely to pathologize and stereotype the oppressed. While I emphasize this point here in relation to health inequity and the discrimination-emotion-health model, I think it requires greater attention even within the wider literature on social injustice and oppression. Much more work needs to be done to unpack how the privileged are damaged by the oppression that, in many but not all ways, privileges them. 19

CONCLUSION
The discrimination-emotion-health model helps to capture a significant part of the way in which social stressors like discrimination can influence emotion and health. In this paper, however, I have identified six aspects of a political critique of the potential implications of the model that demonstrate how they can reinforce injustices, including health inequities and the disrespectful social norms and stereotypes underlying discrimination. To improve the health of populations and communities equitably in light of the psychosocial influence of discrimination and oppression on health, both the model and the critique need to be considered. We need to understand this psychosocial influence within a framework of health equity/inequity, which emphasizes the following: health should be understood within a pluralist theory of social injustice; the relationships among discrimination, emotions, and health lead to second-order injustices; structural oppression, not merely interpersonal discrimination, is a major factor in the social determination of emotions and their regulation; health-focused solutions to the problems created by social injustice to emotion and health are often second-best solutions; health policies and guidance have expressive value; and the emotional states and expectations of the privileged are also shaped by oppression.

NOTES
1 Preliminarily, a negative emotion is one where its experience is usually considered undesirable by the individual experiencing it, while a positive emotion is one that is usually considered desirable. In section 3 of this paper, I raise concern about what is often an unquestioned presumption in the epidemiological research about which are negative and which are positive emotions. 2 Depression and anxiety, although at times described in the epidemiological literature as "negative emotions" (Kubzansky et al., 2014, p. 330), are not merely emotions. In the psychological literature they are also considered to reflect "complex constellations of chronic elevations of maladaptive cognitions and behaviors" (Kubzansky et al., 2014, p. 330). When I refer to research focused on depression and anxiety, this can be understood as taking an interest in the emotional states with which they are often associated, although I recognize that they are not merely, or even at times necessarily, represented by emotional states. Underlying this discussion are important philosophical questions about emotions, beyond the scope of this paper to address-e.g., what counts as an emotion and thus which are emotions (Calhoun and Solomon 1984, p. 23-26). 3 Health, of course, can influence emotions, and could also be a basis for discrimination so the causal relationship can function in different directions. This paper, however, is focused on discrimination as a stressor and thus as a social determinant of health. The name of the model-stress-emotion-health-indicates the direction of the causality on which we are focused. 4 I thank Kristin Voigt for pushing me to clarify this point. 5 Consider, for example, how abstracting a model from research on discrimination as a psychosocial determinant of health (usefully summarized by Krieger 2014) combined with research on the influence of social stressors on emotions and health (usefully summarized by Kubzansky et al. 2014) is likely to result in the discrimination-emotion-health model. 6 For a classic account of the internalization of oppression, see Frantz Fanon (2008Fanon ( [1952).
For more recent work, see, for example, Ann Cudd (2006) on the psychological mechanisms of oppression and Nora Berenstain (2016) on epistemic exploitation and its emotional burdens. 7 The privileged can also be referred to as the advantaged or the dominant group. The disadvantaged and the privileged are those disadvantaged or privileged by social injustice-e.g., as represented by inequities in the distribution of social goods or by social-relational inequalities-in a particular society. This can occur along a number of axes often represented by membership in a social group. So, for example, merely being a man in many societies will mean that you experience some privilege as a man, even if you are often disadvantaged according to other axes-e.g., class, sexuality, and race. A particular individual then could be disadvantaged when our focus is on one axis (e.g., race), but that same individual could be privileged according to another axis (e.g., class), and whether or not that individual should be considered disadvantaged or privileged will depend on which form of discrimination or oppression we are focused on. It is worthwhile to acknowledge here that people who experience multiple disadvantages are uniquely disadvantaged in ways that cannot be represented by "adding up" the disadvantages of the single axes along which they are disadvantagede.g., black women are disadvantaged not merely as black and as women but also as black women specifically (see, for example, Crenshaw, 1989). 8 The relationship between oppression and discrimination is itself often not explored enough in the humanities literature, with some theorists, for example, even downplaying or ignoring discrimination in their analyses of oppression (Young, 1990). While the relationship requires greater theoretical exploration, I take it here that oppression and discrimination are not identical, and each is significant. I also take it that discrimination is often one of the constituent parts of oppression, but that oppression is not necessary for wrongful discrimination to occur. For example, it is reasonable to claim that it is pro tanto wrongful discrimination when a landlord in Seattle refuses to rent an apartment to anyone who has recently moved there from California, for the reason that she dislikes Californians, but that does not mean that Californians are oppressed in the process. 9 I thank Sara Goering for helping me to formulate this point. 10 See also Sara Ahmed's (2015) theory of the sociality of emotions. She claims that "emotions create the very effect of the surfaces and boundaries that allow us to distinguish an inside and outside in the first place … the 'I' and the 'we' are shaped by, and even take the shape of, contact with others" (Ahmed, 2015, p. 10). 11 Lee and Hicken consider the health costs of "vigilance" as a manifestation of respectability and do not focus on emotions per se; however, it seems reasonable to claim that vigilance is often related to the suppression of emotions such as anger and could increase the likelihood of experiencing emotions such as psychological distress. In fact, one of the primary results of their study is that vigilant behaviours are associated, in a dose-response fashion, with depressive symptoms (Lee and Hicken, 2016, p. 433-435), and thus we can understand their research as partially measuring stress, emotion, and health, where vigilance is the stress (the proximal stress; the cause of this stress is racial oppression), and where the emotion and health aspects are combined in the measurement of depressive symptoms. Vigilance was measured via selfassessment according to the frequency of the following experiences: trying "to prepare for possible insults from other people before leaving home," feeling "that you always have to be very careful about your appearance to get good service or avoid being harassed," trying "to avoid certain social situations and places," and watching "what you say and how you say it" (Lee and Hicken, 2016, p. 429). 12 Political resistance is burdensome in many ways -e.g. taking time away from pursuing other opportunities, demanding emotional and psychological effort, running the risk of arrest and violence, and creating moral dilemmas. See, for example, Lisa Tessman (2005, p. 107-131) on the burdens of political virtue, especially the traits one requires to be a hardened dissident. 13 See, for example, Fabienne Peter (2006) and Madison Powers and Ruth Faden (2006) for their particular arguments to justify the same claim-that health inequity needs to be seen within a framework of social injustice. 14 A theory can be pluralist in numerous ways-consider, for example, the distinction between a dimension-pluralist theory of social justice and a pattern-pluralist theory of social justice (Fourie, 2016, p. 191-192). 15 Examples of dimension-pluralist theories would be Martha Nussbaum's capabilities approach (2000) and Powers and Faden's well-being approach (Powers and Faden, 2006). 16 Admittedly, there are significant philosophical challenges here. Should some of the dimensions be prioritized over others, at least in certain circumstances? If we want to design just policies, how do we adjudicate between conflicts? The only stand I take on these issues in this paper is that health should not necessarily be prioritized over other values. However, a more comprehensive theory of health equity and social justice would have to provide much more detail on this topic; see, for example, the debates about the priorities of social goods in the political philosophy literature (Nussbaum 2000, p. 81-86). 17 See also Killmister (2015) on the "double binds" created by oppression, which often force the oppressed "to either make trade-offs within autonomy or make trade-offs between autonomy and another core personal value. Agents under oppression are thus faced with a particularly tragic dilemma" (p. 162).
18 While we need, and should indeed aim for, a radical restructuring of unjust societies in order to achieve justice, particular norms, policies, and institutions can be more or less just, and where we can make them more just, even though we cannot change the system in entirety, pro tanto, we should do so. Consider, for example, Amartya Sen's (2009, p. 1-27, 87-112) claims that we require comparative justice, and not (or not only) transcendental justice. A comparative framework means that we can compare policies or institutions in terms of how just they are, and advance justice by choosing the more just alternative; on the other hand, transcendental justice requires that we identify and pursue perfect or ideal justice. My claims about second-best solutions can be seen to be aiming for something like comparative justice (although admittedly they don't map precisely onto Sen's framework). 19 There is significant work that has been done on this-however, it remains underdeveloped.
Work in critical race and feminist theory on epistemologies of ignorance, for example, demonstrates how the oppressed are epistemologically privileged, while the privileged are ignorant (e.g., Alcoff, 2007;Mills, 2007). Notice that, on the one hand, this can be said to demonstrate an impairment of privilege and yet, on the other, it serves an important purpose in keeping the privileged in positions of power by helping them, for example, to maintain the myth that their privilege is deserved. See also, for example, Fourie (2012)