On Microaggressions: Cumulative Harm and Individual Responsibility

Original Article Free Access Christina Friedlaender First published: 15 January 2018 https://doi.org/10.1111/hypa.12390 Cited by: 1 Sections Abstract Microaggressions are a new moral category that refers to the subtle yet harmful forms of discriminatory behavior experienced by members of oppressed groups. Such behavior often results from implicit bias, leaving individual perpetrators unaware of the harm they have caused. Moreover, microaggressions are often dismissed on the grounds that they do not constitute a real or morally significant harm. My goal is therefore to explain why microaggressions are morally significant and argue that we are responsible for their harms. I offer a conceptual framework for microaggressions, exploring the central mechanisms used for identification and the empirical research concerning their harm. The cumulative harm of microaggressions presents a unique case for understanding disaggregation models for contributed harms, blame allocation, and individual responsibility within structural oppression. Our standard moral model for addressing cumulative harm is to hold all individual contributors blameworthy for their particular contributions. However, if we aim to hold people responsible for their unconscious microaggressions and address cumulative harm holistically, this model is inadequate. Drawing on Iris Marion Young's social connection model, I argue that we, as individual perpetrators of microaggressions, have a responsibility to respond to the cumulative harm to which we have individually contributed. Since 2012, Black Lives Matter has worked to bring the routine and institutionalized violence against Black Americans into the national spotlight (Black Lives Matter 2016). Many white Americans have taken issue with the tagline itself, subverting it in the reply: “all lives matter.” Proponents of this subverted tagline occasionally fail to see why it is a problematic response, regarding it as more morally inclusive. However, this tagline is a racist erasure of the very problems Black Lives Matter is addressing, namely that Black lives are not valued equally compared to white lives, and thus the very use of this subverted tagline constitutes a kind of harm. Given the impact of similar noninstitutional, interpersonal behaviors, such acts fall under a new moral category of harms resulting from oppression: microaggressions. This category fills both a conceptual and lexical gap in moral language (Lehrer 1974, 95; Lawson 1992, 76–77). Microaggressions are a form of discriminatory behavior toward members of a structurally oppressed group. However, individual perpetrators are often unaware of their behavior or fail to see why it is harmful. Moreover, microaggressions are often dismissed on the grounds that they do not constitute a real or morally significant harm. Thus, my goal is to explain why microaggressions are morally significant and argue that we are responsible for their harms. Microaggressions present a unique case for how we understand cumulative harm, blame‐allocation, and responsibility within structural oppression. Section I provides a conceptual framework for microaggressions and reviews the psychological literature detailing their harms. I take the case of the cumulative harm experienced by an individual victim of microaggressions to be of particular concern. Section II discusses two major problems with a disaggregative model of blame allocation for cumulative harm: 1) epistemic ignorance and 2) the relational nature of these harms. Although some agents are blameworthy for their microaggressive acts, the problems of epistemic ignorance and relational harm show that an account of responsibility for microaggressions cannot rest on an account of blame. Section III highlights a further concern in how we, as perpetrators, respond to those experiencing cumulative harm. In section IV, I provide an account of individual responsibility for microaggressions. Broadly stated, we, as perpetrators of individual microaggressions, have a responsibility to respond to the cumulative harm to which we have individually contributed. Finally, I briefly consider the role of shared and collective responsibility for microaggressions. 1 Microaggressions and Their Harm As a category of behavior, microaggressions are an odd phenomenon. Not only are such acts often the product of an agent's implicit bias, but they are also subtle and attributionally ambiguous (Wang, Leu, and Shoda 2011). As anyone who has ever tried to point out a microaggression knows, pinning down a clearly microaggressive act is a rather frustrating endeavor. Thus, my goal is not to provide a comprehensive answer to the question: “What makes an act a microaggression?” Rather, I rely on how this term is defined within the psychological literature and focus on what sorts of harms result from such acts. In order to deal with the problem of attributional ambiguity, I explain the relationship among microaggressions, their meaning, and structural oppression. Broadly stated, microaggressions are subtle acts of bias that reflect a structural form of oppression toward a specific group of people, such as racism, transphobia, or sexism. Chester Pierce first used the term microaggressions to refer to the everyday, subtle insults that Black Americans experience (Pierce et al. 1978). Microaggressions are the behavioral consequence of an agent's implicit bias against a structurally oppressed group. These subtle insults are predominantly unintentional (that is, they are not usually the product of conscious action). Due to their subtlety, it is often ambiguous as to whether an act was in fact microaggressive, making them rather hard to identify or point out to others (Swim and Cohen 1997; Hylton 2005). If microaggressions constitute a moral wrong, then it is not always clear whether such a wrong has occurred. Like the implicit biases from which they flow, microaggressions are particularly insidious because they are generally committed by agents who consciously regard themselves as committed to egalitarian and progressive values (Dovidio and Gaertner 2002; Sue 2003). Microaggressions are a form of what Jean Harvey refers to as civilized oppression (Harvey 1999, 1–18). Oppression is “civilized” when its violence is not overt nor clear to victims, perpetrators, or bystanders, and thus is particularly harmful because the actual impact of harm is obscured. Because of their subtlety, microaggressions are easily ignored or dismissed. Given that microaggressions are hard to define philosophically, I work from a definition echoed through much of the psychological literature on the matter: “brief and commonplace daily verbal, behavioral, and environmental indignities, whether intentional or unintentional, that communicate hostile, derogatory, or negative slights and insults towards members” (Nadal 2008, 23) of socially, culturally, or politically oppressed groups. Derald Wing Sue, the preeminent psychologist on microaggressions, provides a helpful taxonomy of microaggressions. Microaggressions can be (1) verbal, (2) nonverbal, or (3) environmental (Sue 2010a, 25). Environmental microaggressions are those that cause individuals to feel excluded from a space on the basis of identity. Verbal and nonverbal microaggressions are communicated by a particular individual (for example, a white woman suddenly gripping her purse tighter when a man of color sits next to her on the subway). Further, there are three types of microaggressive acts: (1) micro‐insults, (2) micro‐invalidations, and (3) micro‐assaults (28). Although micro‐assaults are conscious (for example, displaying a swastika), micro‐invalidations and micro‐insults are not. Take the following example of a potential micro‐invalidation: a woman voices a valuable idea in a classroom and the professor later attributes that idea to one of her male peers. This act is attributionally ambiguous because, on the one hand, the professor may simply have been forgetful, but on the other hand, in later attributing the idea to a male student, the professor might have been reflecting an implicit bias against women's intellectual abilities. This ambiguity contributes to our current failure to take microaggressions seriously as moral wrongs. Oftentimes, people are accused of seeing oppression where there is none or making a big deal over a “small” action (Sue 2010a, 2010b). In fact, much of the public discourse around microaggressions, particularly on college campuses, has turned in this direction, decrying that the focus on microaggressions reflects a new moral “culture of victimhood” (Campbell and Manning 2014, 714–16) or “coddling” of people's emotions and intellectual exposure (Haidt and Lukianoff 2015). But the harm of microaggressions is not insignificant. They affect individuals psychologically, behaviorally, socially, and physiologically, and serve to reinforce the very structural oppression that produces them. Thus, I offer the following taxonomy of microaggressive harms. First, a single microaggression can cause negative emotional, behavioral, and cognitive responses in the target (Nadal et al. 2011, 238). When the professor misattributes her idea to a male peer, the female student might suffer from negative cognitive effects (Sue 2010b, 3–25) and stop responding to the present discussion in class. Second, attributional ambiguity itself causes targets harm in their having to second‐guess whether their social group categorization motivated the act. As Jennifer Wang and her colleagues point out, white people rarely, if ever, have the experience of having to question whether a particular slight was racially motivated, whereas people of color disproportionately engage in this second‐guessing (Wang, Leu, and Shoda 2011, 1674). Third, when a microaggression occurs, targets are socially burdened with feeling as if they must suppress their reaction to the slight (1666–67). Because of attributional ambiguity, the reactions of victims are often regarded as disproportionate, socially unacceptable, or cases of false perception (that is, seeing oppression where there is none) (Sue 2010b). Fourth, microaggressions produce material harms and reinforce larger structural problems (for example, race/gender wage gaps). Microaggressions can affect everything from employment (Purdie‐Vaughns et al. 2008; Sue, Lin, and Rivera 2009; Sue 2010b), to education (Bell 2002; Guzman et al. 2010; Kim and Kim 2010; Sue 2010b), to health‐care access and treatment (Sue and Sue 2008; Nadal, Rivera, and Corpus 2010). For example, microaggressions can produce a hostile working environment and negatively affect hiring, retention, and promotion (Pierce 1988; Hinton 2004). Moreover, microaggressions can reinforce stereotypes about oppressed groups (Sue 2010b), putting individuals at risk for stereotype threat. Lastly, the frequency with which individuals experience microaggressions constitutes a cumulative harm. This has two components: (1) the mere addition of the aforementioned harms occurring on numerous occasions over the course of one's life; and (2) the relational nature of individual harm insofar as the harms intensify one another in the process of accumulation. Psychological evidence suggests that this cumulative harm produces negative long‐term emotional, cognitive, behavioral, physiological, and social responses (Sue 2010b, 3–25). These responses can include stress, anxiety, depression, high blood pressure, insomnia, substance abuse, eating disorders, social withdrawal, suicidal ideation, and post‐traumatic stress disorder (Clark et al. 1999; Brondolo et al. 2003; Leland 2008; Sue 2010a; 2010b; Nadal 2011). If these claims are correct, then the harm of microaggressions is not insignificant, especially in the cumulative case, and should not be dismissed. Microaggressions can both reinforce the structural, material harms of oppression and significantly affect an individual's capacity to function well. Given that my project here is ultimately to hold individuals accountable for such acts, I want to briefly explain how the relationship between structural oppression and microaggressions lessens the problem of attributional ambiguity, making it easier to say when such an act has occurred. Microaggressions are contextually defined and thus cannot be understood in isolation because they occur within a broader framework of structural oppression. In order to determine whether an act was microaggressive, one must be able to establish a connection between the act and an existing form of structural oppression. Ann Cudd's account of oppression offers a helpful way to determine whether this background condition is present. For Cudd, structural oppression exists if and only if: (1) there are institutional practices that cause harm; (2) a social group is the target of such harm; (3) a different group benefits from the first group's institutional harm; and (4) there is “unjustified coercion or force that brings about this harm” (Cudd 2006, 25). If these four conditions are met, then there is an objectively existing form of structural oppression. For example, women are objectively oppressed structurally, even if individual women do not subjectively believe that they are oppressed. Thus, an act is a microaggression if and only if we can establish a link between the act and an objectively existing form of structural oppression. As contextually defined acts, microaggressions are communicative in accordance with their background conditions. In order to avoid attributional ambiguity, the determination of a particular act as microaggressive requires knowledge about the form of oppression to which the act is linked (that is, its structural functions and manifestations). Microaggressions are contextually communicative because they enact and reiterate a particular social, cultural, or political history in accordance with a form of structural oppression, such as racism. The background conditions of the act determine whether an act is communicative in this way and what exactly the act is communicating. Furthermore, by linking microaggressions to structural oppression in this way, there is a stronger case for showing how microaggressions occur in patterned ways and with a certain frequency. If perpetrators argue that their acts are not microaggressions, we can point to a pattern of similar acts that have historically and currently manifested in relation to an objectively existing form of structural oppression. It thus follows necessarily that the targets of microaggressions are members of oppressed groups. Because of this necessary connection, victims cannot be members of oppressor groups qua their oppressor identity. Following Cudd's four conditions, those in the oppressor groups are the overall beneficiaries of practices that cause harm to those in the oppressed groups. Microaggressions serve to reinforce structural relations of oppression by harming those in oppressed groups psychologically, socially, physiologically, and materially. Those in oppressor groups ultimately benefit from the harms of microaggressions because microaggressions reinforce their position of privilege and the sort of well‐being afforded to those who are not subject to structural oppression. Although members of oppressor groups might engage in acts that attempt to reinforce in‐group behavior (for example, calling another white person a “race traitor” for refusing to be complicit in white supremacy), these acts are not microaggressions because the harm does not reinforce their position of privilege by harming those in the oppressed group. Rather, these acts reinforce their privileged position by attempting to punish those in oppressor groups who fail to act in the overall interest of the group. Lastly, if we determine that a particular act is a microaggression by linking it to an objectively existing form of structural oppression, then it also follows that microaggressions and their harm exist objectively, not subjectively. This is important to note because research suggests that microaggressions can harm individuals even if they fail to perceive them as such (Sue 2010b; Wang, Leu, and Shoda 2011). For example, when targets fail to perceive microaggressions, they still internalize negative emotional responses, such as shame or anxiety, as targets still believe, albeit mistakenly, that the acts were about them as individuals (Wang, Leu, and Shoda 2011, 1674–75). Thus, one might be tempted to argue that in order to identify whether an act is a microaggression, we require that the victim perceive the act as necessarily connected to a form of structural oppression. However, the harm does not depend on the victim's perception, and subjective identification of microaggressions leaves room for false perception. There could be cases in which victims perceive an act as microaggressive, but they are wrong in this attribution. That said, wrongful attribution is less likely than the failure to perceive a microaggression, as oppressed individuals are more likely than their nonoppressed counterparts to detect microaggressions (for example, people of color are more likely than their white counterparts to accurately perceive that a racist act has occurred) (Sue 2010b, 3–25). Therefore, subjective perception by a victim is not a necessary component for identification. Although these conditions do not rid us of the problem of attributional ambiguity, they do lessen it by enabling us to more easily exclude other sorts of disrespectful or oppressive behavior from the category of microaggressions. 2 Cumulative Harm: Blame Allocation and Disaggregation In cases of cumulative harm, we often start from the presumption that individuals are blameworthy for their particular contributions. However, the cumulative harm of microaggressions presents two challenges for this model: (1) epistemic ignorance violates an ought‐implies‐can principle; and (2) microaggressive harms relationally intensify in their accumulation. Because of these two concerns, I argue that moral responsibility for microaggressions cannot rest on an account of individual blame nor on a disaggregation model. To begin, there are two general points worth noting. First, blame and responsibility are not the same thing. They come apart (for example, if a baby breaks a vase, they are causally responsible, but not blameworthy). Some individuals are blameworthy for their microaggressions, but my general point here is that responsibility for cumulative harm cannot merely reduce to an account of individual blame because not all individuals who contribute to the cumulative harm are blameworthy for doing so. In the final section, I draw upon Iris Marion Young's social connection model (Young 2011, 95–122) of responsibility in order to explain why all individuals are responsible for microaggressions, despite the fact that not all individuals are blameworthy for them. Second, one might object that the following problems of blame‐allocation and disaggregation are owing to my focus on the individual rather than on a social group. Perhaps all members of oppressor groups are blameworthy for microaggressions (for example, all white people are blameworthy for racial microaggressions), and thus individual blame follows from membership in a blameworthy group. However, this intuition raises concerns of social ontology. Unlike corporate groups, social groups are messy entities. You cannot hold individuals blameworthy for their membership in a nonvoluntary social group unless you can ascribe an intentional structure to the group itself. Claims about group blameworthiness must necessarily track claims about intentionality, and ascribing intentionality to a messy social group looks nearly impossible. Furthermore, all people commit microaggressions, not just members of oppressor groups. Thus, I have chosen to focus on individual accountability. The first problem for blame‐allocation is that microaggressions are often unconscious acts. As unconscious acts, they lack intent and agents are unable to do otherwise. Moreover, agents might lack the relevant knowledge to understand the contextual, communicative meaning of their microaggressive acts, making them unable to identify the actions themselves as microaggressions. This is a problem of epistemic ignorance. Individual ignorance is produced by the broader structural conditions under which the agent is acting. For example: patriarchal socialization into a gender binary produces people who truly believe women are physically weaker than men and thus engage in microaggressions that reinforce this stereotype, such as asking men to move a heavy object even if the women present are physically stronger. If microaggressions are the product of implicit bias and/or the communicative meaning is epistemically inaccessible to the individual, then it seems we cannot regard agents as blameworthy because this violates an ought‐implies‐can principle. It is thus helpful to distinguish between two kinds of epistemic ignorance: 1) genuine ignorance; or 2) ought‐to‐have‐known ignorance. Perhaps this is a controversial claim, but there are conditions under which your epistemic ignorance is genuine, such as cases in which one genuinely could not have either known or done otherwise. Genuine epistemic ignorance arises within what Miranda Fricker refers to as epistemic hermeneutical injustice (Fricker 2007, 147). Such injustices occur when the social conditions in which one exists place epistemic limitations that cannot be individually overcome. In this hermeneutical case, there is a correlation between epistemic ignorance and epistemic injustice. Fricker defines hermeneutical injustice as “having some significant area of one's social experience obscured from collective understanding owing to a structural identity prejudice in the collective hermeneutical resource” (155). This is a nonagential form of injustice because it occurs without reliance on individual agents. Structural oppression produces situated hermeneutical inequality, which affects both the content and the form of individuals’ statements about their circumstances (162). First, it obscures what individuals are able to articulate about their experience. Second, it grants privilege to a particular form of expressing, rendering an oppressed individual's expression unintelligible. As Fricker notes, prior to the conceptual and linguistic category of sexual harassment, women with such experiences were sometimes unable to articulate their experiences in an intelligible way to themselves or others, especially to those without such experiences. Genuine epistemic ignorance necessarily arises in conjunction with hermeneutical injustice. The presence of such injustice is what creates the epistemic gap. If you are raised in a society that defines gender as a male–female binary, and the conditions of epistemic injustice have rendered articulations of gender outside of this binary as inexpressible or unintelligible, then your failure to use the correct gender pronoun for someone is a microaggression, but you are not blameworthy for it. Critical reflection by an individual agent might be necessary in order to avoid future microaggressions of this sort once they are brought to the agent's attention, but this is not sufficient to prevent microaggressions from happening prior to increased knowledge or self‐awareness. If you do not know that more than two genders exist because such existence is inexpressible or rendered unintelligible to you, no amount of individual critical reflection will prevent you from mis‐gendering people who identify outside this binary. The existence of their gender is necessarily outside of your personal epistemic scope and the epistemic scope of the larger circumstances in which you are embedded. Furthermore, if microaggressions are a form of implicit bias and you do not know you hold an implicit bias (or even know what an implicit bias is), then your microaggressive acts are genuinely outside of your control. When genuine epistemic ignorance is present, you are responsible for having contributed to cumulative harm, but are not blameworthy. However, there are conditions under which you ought to have known because there is no hermeneutical gap that individually limits your ability to overcome your epistemic ignorance. In this case, epistemic ignorance is purely a function of being in the position of privilege or willful ignorance. Although privileged positioning can prevent one from accurately perceiving the realities of those who exist in the oppressed social location, people can individually overcome this ignorance. If the microaggression is unconscious but it follows from an implicit bias you know you have, then you ought to have known about your potential to contribute to cumulative harm, but have simply failed to address it. In this case, you are liable for your epistemic failure and thus blameworthy for the microaggression. In the first case people can perhaps shift the blame to the structural conditions that produced their ignorance, but they cannot do so in this second case. Given the availability of such expressible and intelligible knowledge, they ought to have known better, or they were, at least, in the position to widen their epistemic scope individually or address the implicit bias from which the microaggressive acts flowed. In this case, they have contributed to the cumulative harm and they are blameworthy. The purpose of distinguishing between genuine and ought‐to‐have‐known ignorance is not to set up a false dichotomy, but rather to map out two ends of a spectrum. My concern here is fundamentally about not violating an ought‐implies‐can principle when ascribing blameworthiness. Cases of repressed or partial knowledge make this even more complicated, but they do not undermine the distinction. If agents’ knowledge is repressed, it seems that regarding agents as blameworthy for their correlating actions violates ought‐implies‐can. If knowledge is partial, it is still unclear whether we can say they are blameworthy. Concerns about partial or repressed knowledge further highlight the need for care in discharging blame for microaggressions. One might object that there is no such thing as genuine epistemic ignorance of the sort described here. This objection takes two forms: 1) all cases of ignorance are of the ought‐to‐have‐known variety, and 2) people feign ignorance when they do know better but refuse to acknowledge their complicity in oppression (which is why such agents become defensive if someone points out their microaggressive act). I take Charles Mills's discussion of white ignorance to be the force behind this concern. Mills argues that Black people and white people “are not cognizers linked by a reciprocal ignorance but rather groups whose respective privilege and subordination tend to produce self‐deception, bad faith, evasion, and misrepresentation” (Mills 2007, 17). White ignorance is an agreement to misread states of affairs insofar as “the Racial Contract prescribes for its signatories an inverted epistemology, an epistemology of ignorance” (Mills 1997, 18). Further, white ignorance functions in relation to white normativity, which determines the boundaries of race by (1) centering whiteness and (2) denying the mechanisms of oppression created and maintained by such boundaries through cognitive practices such as knowledge, belief, and memory (Mills 2007, 25). Such practices of ignorance are reinforced by white interest in maintaining a system of domination (34). Thus the objection is that genuine epistemic ignorance does not exist because this ignorance is an intentional, sustained resistance to acknowledging the realities of racial oppression. If whites become defensive when called out for their racial microaggressions, then it is because they do know better. On this objection, either people ought‐to‐have‐known or did know and are thus blameworthy. In replying, I start with Cheshire Calhoun's argument regarding normal and abnormal moral contexts (Calhoun 1989, 394–98), reframing her distinction as “dominant” and “nondominant.” In dominant moral contexts, there is a shared moral knowledge, reasoning, and language (394–98). It is rare for people to be excused from moral blameworthiness due to ignorance because of this shared moral framework. On the other hand, in nondominant moral contexts, developments in moral knowledge are confined to a specific sub‐group in society, such as feminist theorists. In this case, genuine moral ignorance is not a rare case, but is the norm for that society (394–98). The fact that moral developments reshape moral language is important because microaggressions are a relatively new linguistic category. It is only in the last few years that the concept has gained widespread attention. Addressing microaggressions and their harm requires a moral re‐education. Calhoun also addresses the worry that ignorance resulting from structural oppression is active resistance, that ignorance is culpable self‐deception about one's complicity. She argues that this is not self‐deception or an intentional state of resistance to moral self‐criticism. Rather, there is no motive to self‐reflect, as “a motive to be morally reflective is exactly what people will lack when moral ignorance is the norm” (399). In cases where perpetrators of microaggressions get defensive when their harmful behavior is pointed out, it is sometimes because they genuinely view their behavior as being in accordance with dominant moral norms. From the perpetrators’ perspective, they are not being introduced to new moral knowledge; rather, they are being unjustifiably reprimanded for morally acceptable behavior. While moral reproach constitutes an ought‐implies‐can violation in such cases, we run the risk of condoning the behavior if we fail to reprimand the offending individual (400–406). Thus, moral reproach might be the only practical solution to addressing moral failures in nondominant moral contexts. My intuition is similar, but, unlike Calhoun, I distinguish between blame and responsibility. If people commit microaggressions unknowingly, we are justified in pointing it out to them and expecting them to take responsibility for their harmful actions going forward. We are not required to blame them, nor, on this account, are we justified in blaming them if their ignorance is genuine. Therefore, I maintain that some individuals display genuine moral ignorance (even if this ignorance is not innocent structurally) regarding their harmful actions. However, saying that some individuals are not blameworthy does not detract from the harm itself. As James Baldwin argues, this sort of genuine ignorance is a horrific sort of innocence—one that is perhaps unforgivable and for which people ought to be accountable, as “it is the innocence [of whites] which constitutes the crime” (Baldwin 1962, 6). My claim here is that though blaming certain individuals might violate an ought‐implies‐can principle, their ignorance is an “innocent” crime for which they need to be held accountable. One need not assume that a lack of blame requires us to forgive the harm. Baldwin further emphasizes how this innocence overlaps with structurally produced ignorance, capturing the fact that even whites who know better still act harmfully: “[The innocent people] are, in effect, still trapped in a history which they do not understand; and until they understand it, they cannot be released from it…. Many of them, indeed, know better, but, as you will discover, people find it very difficult to act on what they know” (8–9). Furthermore, the question of whether a particular form of oppression constitutes a nondominant moral context is quite important. Given that the mainstream prevalence of discussion concerning anti‐Black racism in America began during the institution of chattel slavery itself, we would be wrong to call it a nondominant moral context, and therefore white ignorance is neither excusable nor genuine. On the other hand, moral knowledge regarding the exclusionary nature of the gender binary has only recently become more mainstream in certain moral contexts, and perhaps such knowledge is only now transitioning from nondominant to dominant. However, the very concept of a microaggression itself still seems to constitute nondominant moral knowledge. Most people have no idea what a microaggression is, let alone that they are engaging in them. Therefore, because some contributors to cumulative harm are genuinely ignorant, not all contributors are morally blameworthy. Thus, a responsibility model for microaggressive harm must be able to hold nonblameworthy individuals accountable as well. The second problem for cumulative harm is the temptation to disaggregate to individual contributors. Suppose, for a moment, that all contributors are blameworthy: individuals qua individuals cannot be held blameworthy for cumulative harm, so it seems we ought to disaggregate and blame people for their particular contributions. However, this move fails to address the nature of the cumulative harm of microaggressions because the cumulative effect is not the mere addition of harm, but rather that harms affect one another in accumulation. Microaggressions create a certain amount of harm as individual events, but the quantity of the harm is not specifiable in isolation because microaggressions are not isolated events. The problem of the cumulative harm of microaggressions is similar to that arising in discussions of cumulative harm for CO2 emissions. If I am the only one emitting CO2, then it is easy to point to a quantifiable amount of harm for an isolated emitter. However, if my emissions take place in the context of billions of other emitters, the cumulative effect of emissions is no longer a case of mere addition. If emitters push past an environmental tipping point, the cumulative effect rapidly intensifies in ways that no longer correlate to a quantitative disaggregation of harms to individual contributions. Similarly, the cumulative nature of the harm of microaggressions is about a psychological tipping point in individuals. The tipping point for such relational intensification happens after an unspecifiable number of individual microaggressions. Perhaps the second time a female student experiences her idea being attributed to a male peer causes her to reevaluate the first instance, understand it as more harmful in retrospect, and potentially experience both instances as more harmful overall. Thus, we cannot treat the cumulative harm of microaggressions as the sum of all individual instances. This fails to capture the holistic nature of cumulative harm and the holistic effect it has on the individual. These two concerns—genuine epistemic ignorance and relational harm—call into question the standard model for addressing cumulative harms, namely the practice of holding individual contributors blameworthy for their particular contributions. Thus, moral accountability for microaggressions ought to be addressed as a claim about responsibility. This model for responsibility must account for the fact that all agents have committed microaggressions (regardless of whether they are blameworthy for doing so) and cannot merely disaggregate responsibility for the harm by making individual contributors responsible only for their particular contribution. We cannot claim: for cumulative harm Y, person X is responsible for Z amount of Y. Thus, our account of individual responsibility for cumulative harm cannot rest on an account of individual blame nor on a disaggregation model. 3 Responses to Cumulative Harm Before turning to my positive account of individual responsibility, it is worth emphasizing the additional harms that result when we fail to understand how cumulative harm affects individuals holistically. This is most clearly seen in how we treat targets who respond to their cumulative harm. To use a nonmicroaggression example: suppose I accidentally step on your foot and cause you a minor harm. Your reaction to my action ought to be proportionate. Perhaps you are justified in being a little angry because I ought to have been more careful, but it would be odd to say that your screaming at me would be justified. This tracks rather uncontroversial intuitions about proportionality and punishment. However, let's change the example: people have been accidentally stepping on your foot all day and you are now in a constant state of pain. Now I come along and accidentally step on your foot, causing you far more harm than I did in the first example because I have added to this cumulative harm. From my perspective, I have caused you the same amount of harm as I did in the first example and thus believe that your justified reaction ought to track that amount of harm. However, I am wrong because I have just added to your cumulative harm. Suppose you do decide to get very angry with me. You are justified in responding to that cumulative harm, but I do not know that. From my incorrect perspective, I have caused you a minor amount of harm and your response is wholly unjustified and thus “disproportionate.” Thus, I respond by telling you that your response is unjustified because I do not realize that I have added to your cumulative harm. For this reason, individual perpetrators often regard the targets as overreacting or exaggerating the harm that has happened. Although such responses are often treated as disproportionate, they are proportionate to the cumulative harm. One might argue that such reactions are unjustified because the response should be proportional to the individual contribution rather than to the cumulative harm to which the perpetrator has just contributed. However, such reactions are justified despite not tracking the individual's particular contribution. This is a question of moral luck. Suppose a woman lashes out at a male colleague, A, for committing a sexist microaggression after encountering, without any response on her part, such acts from other male colleagues (B, C, D) throughout the day. A is not blameworthy for the microaggressions of B, C, and D, but the woman's response is justified on grounds of bad moral luck. A's microaggression was the tipping point, which caused her to finally react to the cumulative harm, and thus her response to A specifically is justified, even though A is not blameworthy for the other contributions. Due to a distorted perception about proportionality of victims’ responses, we create further conditions for harm. By treating such responses as unjustified, one treats the cumulative harm as nonexistent. This is a form of epistemic gaslighting. It reinforces the cumulative harm because it incorrectly minimizes the amount of harm (due to distorted perception) and thus treats victims as if they are over‐reacting to harms that are not there. Gaslighting is especially evident in research on the relationship between microaggressions and sanity checks. Victims must often engage in a sanity check, in which they must find other members of their oppressed group in order to check that their experience was in fact a micro experience of oppression (Sue 2010a, 74–75; Nadal 2011, 332). By gaslighting victims in this manner, we further compound the cumulative harm by making victims feel as if they are experiencing something that no one else is. This gaslighting compounds the negative mental health effects of microaggressions. On Fricker's account, this constitutes both: (1) a testimonial injustice, as it is the failure to give the oppressed epistemic credibility regarding their own experiences (Fricker 2007, 9–29); and (2) a hermeneutical injustice because it makes it harder for oppressed individuals to name their experiences and render them intelligible to both themselves and others. Victims might believe that they have no recourse for the harm they are experiencing, as such recourse requires others to acknowledge that the cumulative harm exists in the first place. 4 Cumulative Harm and Responsibility We have seen that microaggressions cause cumulative harm in individuals. However, a disaggregative model of blame cannot adequately hold all contributors accountable for their microaggressions, nor can it address the relational nature of cumulative harm. Some people are blameworthy due to certain epistemic background conditions; others are not. Disaggregation fails to capture the holistic impact of cumulative harm on an individual. In order to both adequately address the cumulative harm of microaggressions and avoid compounding such harms by individualizing our responses, we require an account of forward‐looking responsibility: one detached from individual blame and that responds to the cumulative harm. Broadly stated: we, as individual perpetrators, have a responsibility to respond to the cumulative harm to which we have individually contributed. In order to motivate this forward‐looking account, I draw upon Iris Marion Young's distinction between two models of responsibility: the liability model and the social connection model (Young 2011, 95–122). As a backward‐looking model, liability claims find fault primarily in cases where agents voluntarily and consciously committed a harm. It is ineffective for dealing with structural injustice as it places too much emphasis on causal responsibility and blameworthiness. This model cannot account for agents who unknowingly commit harms or who do so through indirect causal chains. On the other hand, the social connection model is a forward‐looking account of responsibility, which avoids backward‐looking accounts of blame, guilt, or punishment and the emphasis on voluntary and self‐aware actions. Young's social connection model has five conditions: (1) the focus of responsibility is the cumulative harm of actions rather than the single actions of individuals; (2) the wrongness of an action is determined by the morally laden background conditions (that is, the functions of structural oppression); (3) the model allows a backward‐looking component in order to explain how the cumulative harm generally came about; (4) individuals share responsibility for the outcomes; and (5) responsibility is discharged through collective action (104–13). The first three conditions are helpful in understanding our responsibility for microaggressions. This responsibility model does not require that all contributors are blameworthy, nor does it require us to disaggregate harms to individual contributors. Because the focus shifts from individual contributions to the cumulative harm itself, individuals are now responsible for the cumulative harm to which they have contributed, not merely for their particular contribution. Further, the shift from blame to responsibility holds all contributors accountable, even when they are “innocent” or genuinely ignorant. Once agents are made aware of their microaggressions and contributions to cumulative harm in another, their responsibility is in how those agents respond going forward. If person B angrily corrects person A for incorrectly assuming B's gender pronoun, and A continues to mis‐gender B or tells B that they are overreacting, A has failed to act in accordance with their responsibility. Since Young's model attempts to capture responsibility for structural harms, its use for microaggressions differs in two ways. First, the liability model's focus on causal relationships between perpetrator and harms is still useful. For micro‐invalidations and micro‐insults, each contributor is causally responsible for the individual contribution(s), even if they qualify as cases of genuine epistemic ignorance. One cannot indirectly engage in these types of microaggressions. However, causal responsibility does not always capture environmental microaggressions, which are often attributable to multiple people, and are thus perhaps better captured by shared responsibility. Second, Young's model posits responsibility as shared, not individual, and thus discharged only via collective action. Perhaps this is the best model for environmental microaggressions, but it does not capture individually performed microaggressions: micro‐insults, micro‐invalidations, and micro‐assaults. Thus, I emphasize individual responsibility here for two reasons. First, we ought to treat individual responsibility, shared responsibility, and collective responsibility as separate but related issues. Determining moral responsibility in the collective case, for instance, does not necessarily determine responsibility for the individual case (Isaacs 2011, 8–12). We ought to deal with each tier of responsibility on its own terms. Second, the majority of microaggressions are committed by single individuals, and thus tackling them requires work at the individual level. We do not have a shared responsibility to change our behavior once we have knowledge of a particular act as a microaggression. Rather, we have an individual responsibility not to engage in that action again. Individuals ought to take preventative measures to reduce the likelihood that they will engage in such acts. Such measures include: (1) self‐education about different forms of microaggressions (for example, racist, sexist, transphobic, ableist, and so on) and how these acts are communicative in accordance with a form of structural oppression; (2) critical self‐reflection as to how one's thoughts and actions contribute to and reinforce structural oppression in these “minor” ways; and (3) reducing one's implicit biases against members of oppressed groups. This part of the responsibility is not shared. Individuals must understand, reflect, and renegotiate their thoughts and actions in a forward‐looking manner. Some individuals might be blameworthy, but some are not. Thus, this forward‐looking responsibility hinges on what we ought to do given that microaggressions are commonplace within structural oppression, that we have all committed such acts, and with the understanding that even the most informed and self‐reflective among us will, in all likelihood, continue to engage in microaggressions unknowingly. That said, it is worth gesturing at what shared and collective accounts of responsibility for microaggressions might look like. First, shared responsibility not only better captures environmental microaggressions, but it also better addresses the obligation to call out microaggressive acts. Becoming more aware of our unconscious behavior often requires others calling our attention to it. As both Larry May and Tracy Isaacs have argued, if we are in a group in which other members are acting immorally or unjustly, then we have a duty to prevent or stop the action, if possible (May 1992, 152–55; Isaacs 2011, 76–79). Similarly, there is a responsibly to call people's attention to their unconscious microaggressions. Of course, who, in particular, shares in this responsibility is a complicated question. On the one hand, it seems that oppressors ought to have a stronger responsibility to call out other members of their identity group, such as men calling out other men for unconscious sexism. However, this can lead to the problem of speaking for or over members of the oppressed group in question. Furthermore, members of oppressor groups are not always in the best epistemic position to realize whether a microaggression has occurred or, even if they are aware, to explain why, exactly, the act was sexist or racist. On the other hand, it would be odd to say that oppressed members have such a responsibility because this creates an additional burden of having to educate one's oppressor, which, given how exhausting and time‐consuming this can be, might constitute an additional harm on top of the microaggression itself. Thus, I will say for now that there is a shared responsibly for calling out one another's unconscious microaggressions, but the specifics are for another paper. Collective, or group, responsibility for microaggressions is more complicated than shared responsibility because collective responsibility tracks the intentional structure of the group itself. Highly organized corporate entities have an intentional structure, but nonvoluntary, social groups do not. One might argue, for instance, that all men are collectively responsible for sexist microaggressions because of their socialization into this privileged social group (May and Strikwerda 1994). But this fails to capture that women and gender‐nonbinary individuals also engage in sexist microaggressions and ought to be held accountable for doing so. Further, all the perpetrators of microaggressions do not constitute a group in the relevant sense here. Collective responsibility cannot literally mean “everyone is responsible,” unless “everyone” refers to a specific and highly cohesive group. Therefore, it seems that responsibility for microaggressions is primarily individual or shared. As I have argued, microaggressions, as a new moral category, ought to be taken seriously as a form of harm. The cumulative effects of microaggressions constitute serious harms for individuals. Failure to address such harms can have larger, negative effects for addressing structural oppression more generally. A disaggregation‐based model of blame does not hold all contributors accountable nor does it address the cumulative harm holistically. Thus, we require forward‐looking, individual responsibility to respond to the cumulative harm to which we have just contributed, not merely our individual contributions. Without this framing of responsibility, we cannot adequately address the wrongful harm of microaggressions and risk engaging in further cumulative harms. Note Thank you to Sharon Crasnow, Anja Karnein, Bill Lawson, Adam Thompson, Deborah Tollefsen, two anonymous referees at Hypatia, and audiences at the University of Memphis Philosophy Graduate Brown Bag (2015), the Central States Philosophy Association (2015), and the American Philosophical Association's Pacific Division (2016) for their helpful feedback on drafts of this article. References Baldwin, James. 1962/1993. The fire next time. New York: Vintage. Google Scholar Bell, Lee Anne. 2002. Sincere fictions: The pedagogical challenges of preparing white teachers for multicultural classrooms. Equity and Excellence in Education 35 (1): 236–44. Google Scholar Black Lives Matter. 2016. About the Black Lives Matter network. http://blacklivesmatter.com/about/ (accessed September 9, 2016). Google Scholar Brondolo, Elizabeth, Ricardo Rieppi, Kim P. Kelly, and William Gerin. 2003. Perceived racism and blood pressure: A review of the literature and conceptual and methodological critique. Annuals of Behavioral Medicine 25: 55–65. Crossref PubMed Web of Science®Google Scholar Calhoun, Cheshire. 1989. Responsibility and reproach. Ethics 99 (2): 389–406. Crossref Web of Science®Google Scholar Campbell, Bradley, and Jason Manning. 2014. Microaggression and moral cultures. Comparative Sociology 13 (1): 692–726. Google Scholar Clark, Rodney, Norma B. Anderson, Vanessa R. Clark, and David R. Williams. 1999. Racism as stressor for African Americans. American Psychologist 54 (1): 805–16. Google Scholar Cudd, Ann E. 2006. Analyzing oppression. Oxford: Oxford University Press. CrossrefGoogle Scholar Dovidio, John F., and Samuel L. Gaertner. 2002. Aversive racism and selective decisions: 1989–1999. Psychology Science 11 (1): 315–19. Google Scholar Fricker, Miranda. 2007. Epistemic injustice: Power and the ethics of knowing. New York: Oxford University Press. CrossrefGoogle Scholar Guzman, Fernando, Jesus Trevino, Fernand Lubuguin, and Bushra Aryan. 2010. Microaggressions and the pipeline for scholars of color. In Microaggressions and marginality: Manifestation, dynamics, and impact, ed. Derald W. Sue. Hoboken, NJ: Wiley. Google Scholar Haidt, Johnathan, and Greg Lukianoff. 2015. The coddling of the American mind. The Atlantic, September. Google Scholar Harvey, Jean. 1999. Civilized oppression. Lanham, MD: Rowman & Littlefield. Google Scholar Hinton, Eric L. 2004. Microinequalities: When small slights lead to huge problems in the workplace. Diversity Inc., March/April. Google Scholar Hylton, Mary E. 2005. Heteronormativity and the experiences of lesbian and bisexual women as social work students. Journal of Social Work Education 41 (1): 67–82. Crossref Web of Science®Google Scholar Isaacs, Tracy. 2011. Moral responsibility in collective contexts. New York: Oxford University Press. CrossrefGoogle Scholar Kim, Suah, and Rachel H. Kim. 2010. Microaggressions experienced by international students attending U.S. institutions of higher education. In Microaggressions and marginality: Manifestation, dynamics, and impact, ed. Derald W. Sue. Hoboken, NJ: Wiley. Google Scholar Lawson, Bill E. 1992. Moral discourse and slavery. In Between slavery and freedom: Philosophy and American slavery, ed. Howard McGary and Bill E. Lawson. Bloomington: Indiana University Press. Google Scholar Lehrer, Adrienne. 1974. Semantic fields and lexical structure. London: North‐Holland. Google Scholar Leland, John. 2008. In “sweetie” and “dear,” a hurt for the elderly. New York Times, October 6. Google Scholar May, Larry. 1992. Sharing responsibility. Chicago: University of Chicago Press. Google Scholar May, Larry, and Robert Strikwerda. 1994. Men in groups: Collective responsibility for rape. Hypatia 9 (2): 134–51. Wiley Online LibraryGoogle Scholar Mills, Charles. 1997. The racial contract. Ithaca, NY: Cornell University Press. Google Scholar Mills, Charles. 2007. White ignorance. In Race and epistemologies of ignorance, ed. Shannon Sullivan and Nancy Tuana. Albany: SUNY Press. Google Scholar Nadal, Kevin L. 2008. Preventing racial, ethnic, gender, sexual minority, disability, and religious microaggressions: Recommendations for promoting positive mental health. Prevention in Counseling Psychology: Theory, Research, Practice, and Training 2 (1): 22–27. Google Scholar Nadal, Kevin L. 2011. The racial and ethnic microaggressions scale (REMS): Construction, reliability, and validity. Journal of Counseling Psychology 58 (4): 470–80. Crossref PubMed Web of Science®Google Scholar Nadal, Kevin L., Marie‐Anne Issa, Jayleen Leon, Vanessa Meterko, Michelle Widerman, and Yinglee Wong. 2011. Sexual orientation microaggressions: “Death by a thousand cuts” for lesbian, gay, and bisexual youth. Journal for LGBT Youth 8 (1): 234–59. Google Scholar Nadal, Kevin L., David P. Rivera, and Melissa J. H. Corpus. 2010. Sexual orientation and transgender microaggressions: Implications for mental health and counseling. In Microaggressions and marginality: Manifestation, dynamics, and impact, ed. Derald W. Sue. Hoboken, NJ: John Wiley & Sons. Google Scholar Pierce, Chester. 1988. Stress in the workplace. In Black families in crisis: The middle class, ed. Alice F. Coner‐Edwards and Jeanne Spurlock. New York: Brunner/Mazel. Google Scholar Pierce, Chester, Jean Carew, Diane Pierce‐Gonzalez, and Deborah Wills. 1978. An experiment in racism. In Television and education, ed. Chester Pierce. Beverly Hills, CA: Sage. Google Scholar Purdie‐Vaughns, Valerie, Paul G. Davis, Claude M. Steele, and Ruth Ditlmann. 2008. Social identity contingencies: How diversity cues signal threat or safety for African Americans in mainstream institutions. Journal of Personality and Social Psychology 94 (1): 615–30. Google Scholar Sue, Derald Wing. 2003. Overcoming our racism: The journey to liberation. San Francisco: Jossey‐Bass. Wiley Online LibraryGoogle Scholar Sue, Derald Wing. 2010a. Microaggressions in everyday life: Race, gender, and sexual orientation. Hoboken, NJ: John Wiley & Sons. Google Scholar Sue, Derald Wing, ed. 2010b. Microaggressions and marginality: Manifestation, dynamics, and impact. Hoboken, NJ: John Wiley & Sons. Google Scholar Sue, Derald W., Annie I. Lin, and David P. Rivera. 2009. Racial microaggressions in the workplace: Manifestations and impact. In Diversity matters: Education and employment, ed. Jean L. Chin. Westport, CT: Praeger. Google Scholar Sue, Derald W., and David Sue. 2008. Counseling the culturally diverse: Theory and practice, 4th ed. Hoboken, NJ: John Wiley & Sons. Google Scholar Swim, Janet K., and Laurie L. Cohen. 1997. Overt, covert, and subtle sexism. Psychology of Women Quarterly 21 (1): 103–18. Wiley Online Library Web of Science®Google Scholar Wang, Jennifer, Janxin Leu, and Yuichi Shoda. 2011. When the seemingly innocuous “stings”: Racial microaggressions and their emotional consequences. Personality and Social Psychology Bulletin 37 (12): 1666–78. Crossref PubMed Web of Science®Google Scholar Young, Iris Marion. 2011. Responsibility for justice. New York: Oxford University Press. CrossrefGoogle Scholar