Attentional discrimination and victim testimony

ABSTRACT

Sometimes, a form of discrimination is hard to register, understand, and articulate. A rich precedent demonstrates how victim testimonies have been key in uncovering such “hidden” forms of discrimination, from sexual harassment to microaggressions. I reflect on how this plausibly goes too for “attentional discrimination”, referring to cases where the more meaningful attributes of one social group are made salient in attention in contrast to the less meaningful attributes of another. Victim testimonies understandably dominate the “context-of-discovery” stage of research into these initially opaque forms of discrimination; a victim’s encounter with the gap between their experience and dominant conceptual frameworks for understanding it is what provides an initial foothold for analysis to begin. Some object, however, to this methodology continuing to dominate the later “context-of-justification” stage, where the hypothesis is rigorously challenged. I argue that this objection underestimates not just how other methodologies are more likely to inherit the various mechanisms of invisibility hiding the discrimination in question, but also how victim testimonies are distinctively well-suited to recognize and challenge those mechanisms. Victim testimonies, then, ought to continue playing a dominant role into these later stages of research regarding hidden forms of discrimination.

KEYWORDS:

• Attention
• salience
• invisibility
• hermeneutical injustice
• microaggressions
• standpoint epistemology

1. Introduction

Sometimes, a form of discrimination is hard to register, understand, and articulate. This can be down to several factors, which can intersect and overlap. It might be obscured by our conceptual resources and language, which, perhaps due to sheer historical accident, do not provide adequate tools for us to conceptualize and verbalize it. The discrimination might be obscured by ideological narratives, which purposely work to shroud it, to serve certain interests. It might be hidden in virtue of the nature of the discrimination itself, such as when that discrimination is unintentional on behalf of a perpetrator, and therefore prima facie hidden to that perpetrator.

In this paper, I focus on one type of discrimination that I call “multiply hidden” in these and other ways: a subtle form of attentional discrimination, wherein certain social groups have the “wrong” property (such as their skin color) attended to relatively better than the “right” property (such as their accomplishments). I introduce this form of discrimination in §2, mentioning some key testimonies from purported victims, which have shaped its study. Drawing on a rich precedent, I argue in §3 that (purported)¹ victim testimonies are particularly important for the “context-of-discovery” stage of research into forms of discrimination that are initially hidden; when generating an initial hypothesis to explain poorly understood experiences of discrimination, paying close attention to individuals with the relevant experiences is key. Some would suggest that the utility of victim testimony stops there, however. In §4 I consider an argument that this method ought to be significantly less prominent in later, “context-of-justification” stages of the scientific investigation, as it produces evidence that is both unreliable and biased. Consulting independent observers of a purported instance of the discrimination, as well as those from different ideological vantage points, is particularly important; these individuals counterbalance the issues of unreliability and bias in victim testimonies. Against this view, I argue in §5 that victim testimonies should continue to play a primary role in later stages of the research process for phenomena like attentional discrimination. This is because other methodologies are liable to inherit various of the mechanisms that make these forms of discrimination invisible in the first place – mechanisms that victim testimony is distinctively well-suited to recognize and challenge.

2. Attentional discrimination

Here, I briefly sketch the case study for this paper. In previous work (Whiteley, 2022, 2023), I drew on existing work on the topic of salience and attention² to propose a particular form of morally problematic attention, which I studied by canvasing testimonies from individuals who experience marginalization. Representative testimonies included the following:

Jean-Michel Basquiat once said: “I am not a Black Artist. I am an artist” (Basquiat, in Buchhart & Bloom, 2015, p. 20). This statement indicates Basquiat’s frustration with how he was treated in the art world – as othered, fetishized, and so on. Reflecting on Basquiat’s statement, Zoe Kravitz (in Willis, 2017) says this: “Happy to be black. Just don’t need to say it in front of everything”. Here Kravitz seems to be complaining about the relative salience her Blackness receives in language, insofar as it is mentioned before her other traits. This objection to the wrong ordering of one’s traits is in the background of testimonies from some racially marginalized individuals working in the fields of science, technology, engineering, and mathematics (STEM). One participant in LaVar Charleston and colleagues’ study on this topic, says: “My belief is that … I am seen as a Black person first” (Charleston et al., 2014, p. 281, emphasis added).

Similar issues are raised by those from marginalized genders. For instance, one blog by an anonymous philosopher discussed her concerns about being seen primarily as a woman. She says “you [women philosophers] will always end up philosophically on the subject of your gender simply because you will be seen as a woman first and a philosopher second” (Anonymous, 2015). Using similar language, chemical engineering doctoral candidate Monica Esopi (in Science Careers Staff, 2018) talked about the relief of moving into a department that explicitly addressed diversity-related issues, saying, “I no longer feel like I’m seen as a woman first; I am just a researcher, a scientist, an engineer”.

A recurring complaint from those who are disabled is that one is not “first and foremost” someone with a disability, contrary to how those with disabilities are regularly perceived. Kristine Stebler offers this advice to counselors: “I hope that from the time you meet and assess your first and your last client/patient that first and foremost you will remember that I am not a disabled person, but a person who happens to be disabled” (Stebler, in Marini et al., 2012, p. 396).³

While these testimonies might be interpreted in various ways,⁴ I suggested that one overlooked, plausible explanation points to a particular form of morally problematic attention; these individuals feel that the “wrong” parts of their identities are being made more salient than the “right” parts in the attention of others. I sketch that suggestion below.

Attention, as described by Sebastian Watzl (2017), is the activity of structuring an individual’s occurrent mental states (perceptual, cognitive, conative, and so on) so that some are more central and others more peripheral – a type of mental management.⁵ A mental state is more central when it is selectively prioritized, relative to other states (ibid, p. 71). This priority relation, in virtue of which different forms of attention count as a unified phenomenon, is phenomenologically familiar; he suggests that “we are acquainted with the reflection of priority in our conscious experience, as a kind of prominence or centrality in consciousness” (ibid, p. 77). On a street corner, for instance, one might be seeing several cars, feeling the warmth of the sun, hearing the sound of an ambulance siren, and thinking about the loudness of sirens. Some mental states will be dimmer in one’s experience, while others will stand out; perhaps the sound of the siren is most central in your attention.

One can be disposed to have certain priority structures (ibid. 98). This might be relative to a certain stimuli or activity. Unsurprisingly, I may be disposed to make the sound of fire alarms central in my consciousness when I hear them. It might be relative to an activity. When I am thinking about my friend Amir, I might be disposed to prioritize thoughts about his ditsiness over his intelligence.

Contributing to a growing literature uncovering the ethical and political dimensions of attention,⁶ I suggested that the individuals in the testimonies above are plausibly objecting to how others attend to them, insofar as the wrong thing is being made problematically salient in the attention of others – in particular, their gender, racial identity, or disability. More specifically, I suggested that these individuals are suffering from a “relative attentional surplus on the wrong property”. This is a particularly subtle form of morally problematic attention. Instead of objecting to a more extreme form, where others are fixating on the wrong content (such as their gender) and ignoring the right content (such as their professional identity as a philosopher), the issue is more insidious; the relative levels of attention on these respective properties are subtly “off” in some way.

In which way, though? I suggested that attentional patterns like these can be considered morally problematic for a host of reasons. Some of those reasons might be instrumental; having one’s gender more salient than one’s philosophical achievements could plausibly harm one’s career trajectory, or induce lower self-esteem, for instance. Other reasons might be non-instrumental. These attentional patterns, I suggested, can constitute a subtle way of disrespecting these individuals’ personhood. Notice that, across many of the testimonies mentioned, the individuals in question encourage attention on their careers and passions. Basquiat is indicating that he wants others to see him as an artist. The statements from women commenting on a surplus of attention on their gender instead reorient the audience to their careers as philosophers, engineers, and so on. These traits are demonstrative of these individuals’ personhood; they are things that those individuals have chosen, and thus demonstrated their personality in, exercised control over, and so on. Their identity as Black, as a woman, as disabled, is plausibly not functioning, in these specific cases, to demonstrate these things. They may instead be functioning as non-personhood-related traits.

There may be other grounds on which this type of attentional pattern can count as morally problematic. For instance, we might draw on a relational egalitarian framework to suggest that attentional patterns can be an aspect of equality of consideration. A systematic difference in attentional patterns on two groups, group a and group b, whereby it is the positive or more meaningful features of group a that are selectively prioritized, while for group b it is their negative or less meaningful features, counts as a way in which those groups are relationally unequal. I develop this proposal elsewhere (Whiteley, m.s.).

In sum, a kind of attentional discrimination can occur, I suggested, where individuals from certain social groups are attended to in a morally problematic way that does not extend to individuals from certain other social groups.

3. Defending victim testimonies in the context-of-discovery stage of research

In this section, I describe in more detail how I used testimonies to analyze attentional discrimination, before attempting to justify their role.

3.1. A research diary: Victim testimonies and attentional discrimination

It can be a transformative moment when someone gestures toward an experience of theirs that resonates with an experience of one’s own – but which has up until that point remained fuzzy and inchoate in one’s mind. In addition to a sense of reassurance and solidarity with others, this can come with the feeling that one is unlocking a deeper understanding of oneself and the world. As for the topic at hand, this once occurred for me at a Women in Philosophy group in my graduate school when someone articulated their frustration with how they were seen by some of their peers – summed up by a sense of too much “emphasis” being placed on their gender. In a casual conversation not long after, an artist acquaintance complained of something similar, using again the word “emphasis”; they felt that their ethnic heritage was emphasized too much by those writing about or otherwise discussing their work. Encounters like these helped me better to notice and grasp an experience of my own – that something distinctively subtle was going wrong in certain of my interactions.

This made me curious. There were familiar tools to refer to adjacent complaints, of being “defined” or “categorised” as a woman or in terms of one’s ethnic heritage. But the chosen language of “emphasis” indicated for me something a little different – something relating to structural or organizational notions of foregrounding and backgrounding. At this time, I became interested in the philosophy of salience and attention. Part of what it is to have what Elizabeth Camp (2017, p. 80) calls a “perspective” is to find that certain properties in the world are salient, insofar as they “[stick] out relative to the background, like a bright light or a hugely bulbous nose”. Sebastian Watzl’s (2011, p. 849) account of attention focussed on similarly organizational notions, highlighting the way in which one’s consciousness is organized so that “some things are in the foreground relative to others”. These accounts offered a fresh lens to interpret my and others’ experiences, generating a first-cut hypothesis: something related to salience and attention might be at issue. In other words, instead of substantive beliefs, classificatory systems, or behaviors causing the discomfort that some feel in the sorts of interactions described above, the problem is with how these individuals’ attributes are being organised by another’s attention, so that some attributes receive more emphasis than others.⁷

Armed with my personal experience, a collection of anecdotes, and an initial hypothesis built from certain philosophical tools, I began to look through blogs, newspaper articles and books that collated testimonies from marginalized individuals. I sifted through posts on Women in Philosophy blogs and articles that interviewed individuals from social groups that were minoritized in fields like science, technology, engineering and mathematics. I would look for instances of individuals describing how they felt they were perceived by others in a way that seemed relevant to my hypothesis. As not every individual would describe their experience using the word “emphasis”, I expanded the search terms to include a list of synonyms relating to organization and order. These synonyms were informed in both a top-down way, from the theoretical construct of “attention” (e.g., using phrases like “centrally” and “peripherally”, “notice”, “ignore”, which crop up most frequently in theoretical accounts of attention and salience), and in a bottom-up way, through noticing phrases evoking the structuring and organization of consciousness used in the testimonies themselves (e.g., “first” and “second”, “first and foremost”). This was an iterative process, then, of creating and adjusting the search terms through repeated engagement with both testimonies and philosophical theories.

In addition to this refinement of search terms, which influenced the range of testimonies that I consulted, this iterative process also meant that the testimonies I engaged with influenced my theoretical account of salience and attention. As an example of the latter direction of influence, I created new distinctions between types of attentional distribution to reflect a theme found in the testimonies, to capture different degrees by which the levels of (in)attention were problematic. Sometimes a testimony would comment on a high level of attention; one testimony from a social worker with cerebral palsy compared his experience of being disabled to being a celebrity: “Because being famous – I assume without being famous myself – means that you receive a high level of attention whether you like it or not and in every possible situation” (Mik-Meyer, 2016, p. 1342, emphasis added). Other times, a testimony gestured at something more subtle, such as where the “wrong” thing was granted relatively more attention than the “right” thing, without the “wrong” thing taking up such an extreme level of attention. Generally, the testimonies mentioned in §2 fit this picture, given that they choose the word “first” instead of “only”. The right thing was plausibly still being attended to, whether that be a person’s status as a scientist, philosopher, artist (etc.). The issue was that the wrong thing – a person’s gender, race, disability (etc.) – was attended to relatively more. Existing tools in the philosophical literature did not seem adequately to capture finer-grained degrees of (in)attention like these. So, these testimonies shaped a new taxonomy for evaluating attention, distinguishing absolute (in)attention (comprising categories of “attentional fixations”, referring to cases of excessive attention, and the converse of “attentional omissions”, when something is entirely ignored) from subtler forms of (in)attentional distributions (comprising categories of “relative attentional surpluses” and “relative attentional deficits”).

Victim testimonies, then – interpreted using tools in discourse analysis –⁸ were key in generating an initial hypothesis. They enabled a rough, first-cut suggestion, that, in certain contexts, emphasis on a person’s demographic properties can be morally problematic.⁹ A more thorough engagement with the testimonies also enabled a more refined, nuanced hypothesis, that there can be something morally problematic about a person’s demographic properties receiving relatively more salience in attention than other, more meaningful properties, even where the latter are not ignored, nor are the former entirely fixated upon.

3.2. A justification: victim testimonies in the context-of-discovery

The use of (purported) victim testimonies is particularly common in the “context-of-discovery” stage of a research process for investigating a form of discrimination. This refers to an early stage of research, where a new hypothesis, methodology, or broader way of thinking is introduced. Consider the utility of victim testimonies in uncovering “microaggressions” (or “micro-inequities”) – referring to subtle and often brief everyday events that denigrate individuals on the basis of their group membership (Sue, 2010, p. xvi).¹⁰ A paradigm example might include being asked “Where are you from?” as a Black person living in a primarily White country – a question that subtly reinforces the idea that the Black individual is a foreigner in their own land. The concept of a microaggression was introduced by Chester Pierce in the 1970s. Pierce’s investigation began with his own testimony, particularly his lived experience as one of few Black professors at an elite university. He used examples from his own life – times his white students would come to advise him on how to teach after his lectures, times his anecdotes of experiencing racism were invalidated by white friends, and so on (Pierce, 1970, pp. 276–77). Indeed, using self-testimony as methodology is common in microaggressions research more generally (see e.g., Fatima, 2020; Sue et al., 2007, p. 275). Outside of self-testimony, a significant range of microaggressions research uses focus groups and interview techniques to extract themes in the lived experience of victims (see Williams et al., 2021, Table 1).

Why was victim testimony a core part of the methodology for this early stage of investigating microaggressions? Well, the term “microaggressions” was constructed as a hypothesis to explain troubling experiences that were common among those from marginalized backgrounds. These experiences, however, lacked a referring concept. Concepts like “racism” and “racially motivated aggression” existed, but these evoked something else – words and actions that were explicitly, intentionally racially prejudicial or discriminatory, such as the use of a racial slur. As the experiences described by Pierce and others did not fit this profile – they were subtle, covert, and often unintentional on behalf of the perpetrator (Pierce, 1970; Sue, 2010) – they were difficult to categorize and understand. In other words, there was what Miranda Fricker (2007, Ch. 7) would call a “hermeneutical gap”, referring to cases where a significant aspect of one’s social experience is obscured from collective understanding. There are different ways of understanding this. Sometimes, victims understand their experiences well, but are prevented from making their experience intelligible to others because the relevant conceptual resources are not shared by the community at large, for instance. In what follows, I focus on hermeneutical gaps that also affect victims’ own self-understanding.¹¹ Such scenarios constitute a hermeneutical injustice if two key conditions are met: firstly, it is strongly in one’s interests for that experience to be made intelligible (to oneself, others); and secondly, the hermeneutical lacuna is due to structural injustices that have led to one’s social group having less control over creating and shaping the conceptual tools in society. As for the case discussed here, the concept of a “microaggression” was forged to bridge this unjust hermeneutical gap.

If victims of microaggressions lacked a clear way to refer to and therefore understand their experiences before the concept “microaggression” had been constructed, then how was consulting their experiences helpful, methodologically? How can we gain any purchase from a person’s experience, when that experience cannot adequately be understood or articulated? Well, the friction between a victim’s lived experience of discrimination (or some other wrong or injustice) on the one hand, and the conceptual tools to understand and explain it on the other, is often felt. Hilkje Hänel (2018, p. 915), writing about subtle forms of sexual harassment and violence, writes this: “Often, when we experience situations of discomfort or violence, we have at least a vague feeling that something is going wrong even if the dominant interpretation tells a different story. That is, while we might not be able to adequately grasp what is going on, we are morally aware that what is happening is wrong.” She calls such cases “moral rupture”, namely, “a moment of discomfort that reveals a gap between our own feeling of violation and the dominant cultural interpretation of what is happening to us” (ibid.).

Here, Hänel is drawing on the rich history of standpoint theory, which, at its broadest, holds that there is a connection between socially marginalized group status and some epistemic advantage in virtue of this marginalization (cf. Collins, 1990; Hartsock, 1998; Solomon, 2009). This epistemic advantage is standardly restricted to topics regarding which the (marginalized) standpoint is relevant, such as the oppression of the social group in question. As Dror (2023) notes, there are stronger and weaker versions of the thesis. Stronger versions state that the socially marginalized have an epistemic advantage in principle from being oppressed, while weaker versions find that the socially marginalized have a contingent epistemic advantage on the grounds that they are more likely to undergo the sorts of informative experiences that grant them this advantage. Hänel (2021, pp. 14–5) and I are committed to a minimal version of this weaker claim: a person who experiences a moral wrong is, because of that experience, more likely to understand that they were morally injured, compared to someone who has not had a subjective experience of such a wrong.

A good method for investigating an as-of-yet unnamed or otherwise poorly understood form of discrimination, then, is consulting those who experience it; the moral rupture often felt by victims provides a foothold for analysis to begin, insofar as it points to something that is going wrong.

Further precedent of studying conceptual lacunas regarding other forms of discrimination corroborates the important methodological role of purported victim testimony. The key example given by in Miranda Fricker (2007, pp. 149–50) to explicate her notion of hermeneutical injustice was “sexual harassment”. Fricker suggests that this was coined in the mid-1970s to bring clarity to experiences of unwelcome sexual behavior, commonly suffered by women – behavior that hitherto had been categorized inadequately as “office banter”, “flirting”, and the like. Berenstain (2020) contests this specific origin story in her paper “White Feminists Gaslighting”, highlighting the much earlier discussions of and activism around the issue (referred to by various terms) occurring in Black, Brown and Indigenous communities. Still, Berenstain, like Fricker, notes the importance of victims sharing their testimonies among themselves, and, in Angela Davies’ words (in Berenstain, 2020, p. 739) “manifest[ing] a collective consciousness of their sexual victimisation”, in helping to shift dominant frameworks for thinking about these experiences.

Another example still concerns the concept of “invisible labour”, initially coined to capture a theme found in testimonies from (predominantly white, middle class) housewives (Daniels, 1987). Vividly demonstrating the hermeneutical lacuna that existed for these individuals, Frye (1963) had earlier described the frustration, dissatisfaction and sense of fatigue reported by housewives in the 1950s as “the problem with no name”. Housewives’ testimonies were critical in opening up the study of “invisible labour”, referring initially¹² to a particular type of gender injustice wherein women are expected to perform a large range of hidden, un(der)valued and un(der)remunerated activities in domestic life. Testimonies from victims have been critical in unearthing various other forms of invisible labor in the proceeding decades, from “aesthetic labour” to “identity work” (see e.g., Brown, 2017; Willis, 2011). (Indeed, this includes the invisible labor performed predominantly by women from marginalized racial, ethnic and class groups who were – and continue to be – regularly brought into white, middle class women’s homes in an outsourcing of the aforementioned domestic labor.)¹³

Not only is there a strong precedent for relying on testimonial methods in the context-of-discovery stage of research into hidden discrimination, then, but there is a clear rationale for this precedent. Hermeneutical gaps are often felt by victims of that gap, through a moral rupture, meaning that their testimonies provide valuable initial seeds for analysis, facilitating the discovery of an (at the time) unnamed, or otherwise poorly understood, form of discrimination. This provides one way of explaining and justifying the focus on victim testimonies in research on attentional discrimination. The brief research diary in §3.1 provides a simplified example of how this method could be applied in practice, when one is attempting to name or give more precise shape to a form of discrimination that currently feels inadequately understood.

4. An objection to the prominence of victim testimonies in the context-of-justification stage of research

So far I have defended the importance of victim testimonial methods only at one stage of the research process: that of discovery. Here, I consider an argument for its utility stopping there.

Let’s return to the topic of microaggressions, as this gives us a good example of this argument. In 2017, Scott Lilienfeld wrote a paper that became an influential critique of the microaggressions research program—one defended by others (e.g., Haidt, 2017). Lilienfeld’s argument has a number of strands, many of which have been analyzed, and in some cases countered, in the proceeding literature (see e.g., Freeman & Stewart, 2020; Thompson, 2020; Williams, 2020). I focus here on key objections he has to the subjectivity of victim testimonies in later stages of the research program into microaggressions, as well as his proposed methodological solution to the limitations he proposes that they have.

4.1. Subjectivity in victim testimonies

Suggesting that the microaggression research program’s hypotheses are “plausible and worthy of further inquiry”, Lilienfeld’s concern is that the program, dominated as it is by methods relating to victim testimony, remains tethered to a “relatively embryonic stage of development”, where its hypotheses “have yet to be subjected to adequate scientific scrutiny” (Lilienfeld, 2017, p. 158). I take his concern with victim testimonies not to be that they are inappropriate for early, hypothesis-generation stages of research, then, but that they inappropriately dominate later, context-of-justification stages, where the hypothesis is tested and challenged. I focus on two issues that Lilienfeld identifies with the subjectivity intrinsic to this methodology.

Firstly, Lilienfeld argues that purported victim testimony is unreliable, as it represents a merely subjective reading of an ambiguous situation. Consider how microaggressions, according to popular definitions, involve negative implicit messages; as discussed earlier, a white person asking, “Where are you from?” to an Asian American individual in the US can be understood to be communicating, in an under-the-radar manner, the message that Asian Americans are not “real” Americans – a racist message (Sue, 2010, p. 32). One problem, Lilienfeld suggests (Lilienfeld, 2017, p. 147), is that victim interpretation of such implicit messages reflects “the cognitive distortion of mind-reading, in which individuals assume – without attempts at verification – that others are reacting negatively to them”.¹⁴ Perhaps the utterance above reflects the speaker’s “genuine and sincere curiosity regarding an individual‘s culture of origin“ (ibid.). In other words, purported victims of microaggressions are interpreting what are usually difficult-to-interpret situations, meaning that there is a significant chance that they can get it wrong;¹⁵ their reports, unless subjected to further scrutiny – including, as I return to in §4.2 below, comparing them alongside the subjective interpretations of others who perceive the incident in question – are unreliable. We can take Lilienfeld’s critique at this stage not to be suggesting that the purported victim’s interpretation is less reliable than anyone else’s (though see the second critique below), but that, insofar as it represents one party’s subjective interpretation of what is routinely an ambiguous situation, it alone should not be accorded much evidentiary weight.

This issue would arguably apply to attentional discrimination cases too; such cases involve subtle, inherently ambiguous cues that purported victims are interpreting as morally problematic. There may be a range of such cues on which the individual is picking up, such as: eye movements that linger on women’s bodies more than men’s bodies; a boss’s ability to better remember the achievements of male colleagues than they can those of female colleagues; or a reviewer mentioning an artist’s race before their status as an artist. In fact, these cases are even more ambiguous than paradigmatic examples of microaggressions; there is not always a clear proposition that can be dissected for interpretation; it may instead simply be the movements of an eye, or longer-term trends in what a person seems better to remember and therefore articulate about one social group over another.

Another critique of Lilienfeld’s can be read as identifying a second issue with the subjectivity in purported victim testimony; instead simply of giving us unreliable evidence, it produces biased results. One key part of research into microaggressions concerns the generation and refinement of the microaggression construct, which involves listing microaggression items. Lilienfeld notes the emphasis on using focus groups of minoritized individuals in generating these lists. He suggests that “A potentially serious concern with this methodology is that most focus groups have been drawn from highly selected samples, many or all of whom are already predisposed to endorse the concept of microaggressions” (ibid. 149). One example he gives of this is a study that “selected African American faculty in counseling psychology and counseling programs who ‘acknowledge(d) that subtle racism continues to exist in U.S. society’ and reported ‘personal experiences with subtle forms of racism in America’” (ibid.). Such individuals are already predisposed to confirm the hypothesis (in this case, that microaggressions exist, and that the concept captures what is happening in those individuals’ experiences), and thus are biased. They are drawn from a homogenous ideological viewpoint.

I take it that this critique naturally extends to victim testimony. Purported victims of microaggressions are likely to be from minoritized groups (definitionally so, for many accounts) and thereby display some degree of homogeneity in their viewpoint; giving prominence to their testimony, then, is liable to bias the research insofar as this viewpoint is privileged over others.

This issue would apply to my study of attentional discrimination if I continue to consult the testimonies of the sort of marginalized individuals I described in §2. These individuals plausibly represent a homogenous viewpoint; they are drawn from blogs and articles designed as platforms for sharing personal experiences with subtle forms of discrimination – subtle forms that these platforms assume are prevalent in society today. While this research is in the context-of-discovery stage currently, Lilienfeld’s criticisms warn against continuing to give primacy to purported victim testimonies as the research progresses.

4.2. The methodological solution

The overall thrust of Lilienfeld’s proposed solution to these problems is “critical multiplism” (ibid, 151). He says, “By examining a research question from diverse methodological vantage points [i.e., critical multiplism], one can obtain a more complete picture of the robustness of one’s research program” (ibid.). The extent to which comparable results emerge across different methodologies, “the greater the confidence that one can place in one’s research programme” (ibid.).

While critical multiplism is Lilienfeld’s general solution, he places particular emphasis on one key method that he sees as best placed to combat the reliability issues regarding the subjectivity in victim testimonies. This is to “Examine the interrater reliability of judgments of microaggressions” by consulting reports from external, independent observers (ibid. 161). He says, “without evidence that external observers can agree on the presence or absence of microaggressions, item ambiguity [i.e., the ambiguous nature of the message that a purported instance of a microaggression has] raises concerns regarding the extent to which microaggressions can be independently verified. How can we know whether a given microaggression occurred or was merely imagined?” (ibid. 145). Consulting independent observers would help to reduce the unreliable nature of subjective data in microaggression research; if external observers agree that a microaggression occurred, this limits the possibility of erroneous mindreading on behalf of the victim.

To combat issues of bias, Lilienfeld recommends, when generating lists of microaggression items, focussing on individuals who have different (including opposed) ideological viewpoints and experiences from the marginalized individuals who dominate in the methods currently favored by researchers. He recommends researchers “use focus-group members and other individuals drawn from a wide variety of ideological perspectives, including individuals who do not necessarily perceive subtle prejudice as a serious problem in society” (ibid. 161), and individuals who have “few or no marked a priori beliefs regarding the existence and nature of subtle racism” (ibid. 149). This would help to counterbalance the bias that emerges from a sample consisting exclusively of marginalized individuals.

In sum, issues of unreliability and bias mean that purported victim testimonies can no longer take center-stage in the context-of-justification stage of research for microaggressions and, I assume Lilienfeld would argue, attentional discrimination. One must widen the pool to incorporate independent observers of purported microaggression incidents, as well as individuals who are less likely to share the ideological viewpoint that (for instance) sees subtle prejudice as a serious social problem, in order to counterbalance these issues.

5. Defending victim testimonies in the context-of-justification’ stage of research

Here, I argue that victim testimonies ought to continue playing a central role in the context-of-justification stage of research, at least when it comes to what I call “multiply hidden phenomena”. While there are arguably important moral reasons for this,¹⁶ I focus on epistemic justifications. I will first set the groundwork for my argument, by explaining how attentional discrimination, like microaggressions, plausibly constitutes a “multiply hidden phenomenon”. Secondly, I explain the methodological significance of this.

5.1. Attentional discrimination as a multiply hidden phenomenon

In §3, we heard about how the concepts “microaggression”, “sexual harassment” and various concepts relating to “invisible labour”, required what David Chalmers (2020, p. 7) calls “de novo” conceptually engineering, in the sense that new concepts needed forging to capture their existence; before their introduction as concepts, there were plausibly hermeneutical gaps in their place, preventing victims from properly understanding, articulating, and/or successfully communicating to others their experiences of these forms of discrimination. In this sense, these phenomena were hidden. Here, I develop this idea further. In particular, I expand on the multiple sociological and other mechanisms – which can overlap and intersect – that plausibly facilitate the hiddenness of attentional discrimination in particular.¹⁷ This is in contrast to the project of expanding the types of conceptual resource that can be hidden. In recent work, Paulina Sliwa (2023), for instance, has compellingly argued that we must go beyond thinking of hermeneutical injustice always as involving a hidden concept; sometimes, she suggests, it might be a certain “perspective” – conceived as a complex set of interconnected cognitive, affective, and motivational dispositions – that is lacking. Instead, I focus on the mechanisms that do the hiding.

My aim is not decisively to defend any particular purported mechanism of invisibility for attentional discrimination. Instead I hope to illustrate the claim that certain forms of discrimination can be hidden in multiple ways. My broader aim in §5 is to show that if a form of discrimination is multiply hidden, then this would have important implications for the methodologies we choose for studying it in later stages of the research process.

Which mechanisms might hide attentional discrimination? One relates to dominant ideological tropes, referring to ubiquitous narratives in culture that are intentionally deployed to serve certain group interests.¹⁸ As research into microaggressions has demonstrated, there are a range of tropes used both to deflect attention from subtle forms of discrimination and prejudice, as well as downplay their significance. One such narrative centers on the insult “snowflake”. This paints victims of such forms of discrimination as having, in Derald Wing Sue and Lisa Spanierman’s (Sue & Spanierman, 2020, p. 20) summary of the narrative, “weak character, who crumble in the face of the most trivial slights”.¹⁹ The victim might be told that they are “making a mountain out of a molehill” (ibid.). Elsewhere (Whiteley, 2023), I have argued that these responses would be likely for someone who complains about attentional discrimination. Suggesting that one’s manager better notices and attends to the achievements of colleagues who are men over colleagues who are women might well be met with incredulity: “there are women getting sexually harassed in the workplace and you’re worried that your boss slightly better notices the achievements of male colleagues? Get a grip!”. The snowflake trope often paints individuals who complain of subtle forms of discrimination as “paranoid” or “hypersensitive”. This trope would hide attentional discrimination in two ways. Either it shuts down those who complain, or it prevents victims from speaking up in the first place – the latter occurring when victims self-silence, attempting to protect themselves from these sorts of hearer response (Dotson, 2011; Johnson et al., 2021).

A second mechanism concerns dominant institutional and procedural frameworks for dealing with discrimination. Workplace disciplinary processes, for instance, generally require that complaints of discrimination are substantiated with evidence of clear, unambiguously problematic behavior (Jones et al., 2017). Attention-related complaints will rarely meet the burden of proof for standard institutional disciplinary processes; as discussed directly below, attentional patterns might not generate unambiguously problematic behavioral cues, if they generate behavioral cues at all.

A third mechanism relates to the limits of human physiology and psychology in relation to the ontology of the discrimination itself, and intersects with many of the other mechanisms above. In particular, the nature of attentional patterns can make them hard for us to notice. Consider first assessing the attentional patterns of another person. The first obstacle here is that attentional patterns consist in the structuring of an individual’s mental states, and thus are not themselves perceptible to an observer. Instead, an observer will generally infer an attentional pattern from various behavioral cues (such as those mentioned in §4.1). As mentioned, however, these will, as in the case of microaggressions, likely consist in ambiguous behaviors that are consistent with a range of interpretations—²⁰regarding whether they are definitively discriminatory, and whether they are the result of attentional patterns in particular (as opposed to something else, like a problematic belief).

Further, attentional patterns are often hidden even from the attender themselves. There is some empirical evidence for us being poor at recognizing our own perceptual attentional patterns, for instance. Taking eye movements to be reliable behavioral proxies for certain of our perceptual attentional patterns, Yasuo Terao et al. (2017) suggest that “We are virtually unaware of our own eye movements”. Multiple studies show for instance that participant ability to identify their own scan path after viewing a scene was close to chance, leading Alasdair Clarke et al. (2017) to conclude that “awareness of one’s own eye movements is extremely limited”. I would suggest that many of us are not in the habit of reflecting upon which of our mental states are more salient than other states more generally. We might wonder whether attention, as a phenomenon so often controlled not by intention, but instead by automatic processing of the distributions of properties in our environment, or by habits that we have accumulated over time (Wu, 2011), would often fall under conscious awareness.²¹ In this way, discriminatory attentional patterns are often hidden to the perpetrator, in virtue of their ontology and our human limitations.

In sum, cases of hermeneutical gaps rarely involve one simple layer of invisibility, relating to the lack of an adequate concept (or perspective) to refer to a given experience. There are often a variety of sociological and other mechanisms that explain and compound its invisibility. There are reasons to think that attentional discrimination would count as such a “multiply hidden” phenomenon.

5.2. The importance of victim testimonies for multiply hidden phenomena

If attentional discrimination does constitute a “multiply hidden” phenomenon, what are the implications of this for the use of purported victim testimonies in studying it? I suggest that they ought to continue playing a central role in the context-of-justification stage of research.

As suggested in §3, victims of a hidden discrimination tend to experience a moral rupture—’a moment of discomfort that reveals a gap between our own feeling of violation and the dominant cultural interpretation of what is happening to us’. This enables the victim to have a clearer understanding of the form of discrimination they have experienced; for one thing, it allows the victim to realize that an experience that dominant interpretations might see as innocuous is in fact morally problematic. This can be beneficial in the “context-of-discovery” stage of research, enabling a light to be shone on an as-yet-inadequately understood phenomenon.

The effects of moral rupture are not isolated to the “moment” in which one notices the gap between one’s experience and dominant interpretations, however. They tend to inspire a particular kind of critical reevaluation of the broader intellectual and sociological landscape. Focussing on the redefinition of sexual violence inspired by the #MeToo movement, Hänel (2021, p. 15) suggests that “[victims’] experiences triggered a thought process helping them to interpret their experiences and [construct] counter-narratives to the distorted dominant narrative of what sexual violence is”. Fricker (2006, p. 107) makes a similar point, saying that “The sense of dissonance [between one’s experience and dominant frameworks for interpreting it], then, is the starting point for both the critical thinking and the moral‐intellectual courage that rebellion [i.e., the critique of received, internalized frameworks] requires.”

This hermeneutical “rebellion” makes victims more sensitive to the mechanisms of invisibility in those frameworks. For instance, Hänel’s investigation into victim testimony during the #MeToo movement highlights how those individuals vividly see how dominant cultural tropes painting sexual violence as “aggravated stranger rape” constitute one such mechanism of invisibility, insofar as they obscure more subtle instantiations of the phenomenon, such as cases of unwanted sexual advancements that do not include being physically overpowered (Hänel, 2021, pp. 7–8). Consider too Carmita Wood, whose legal proceedings against her workplace played an integral role in the coining (or rather refining or bringing to a wider audience, as per Berenstain’s findings) of the phrase “sexual harassment” in the 1970s (Fricker, 2006, pp. 96–7); Emily Martin (in Crockett, 2016) describes how Wood was able vividly to see the ways in which workplace disciplinary processes hid the harassment she endured from a colleague, conceptualizing it as an “interpersonal” problem, to be worked out privately, as opposed to a legal issue. Saba Fatima (2017, p. 151) talks about how her experience of microaggressions gave her insight into the credibility deficits afforded to women of color in culture – a strong ideological mechanism of invisibility for masking microaggressions, which are targeted against such marginalized groups.

These examples indicate the substantial potential epistemic benefits of moral rupture. In Hänel’s (2021, p. 15) words, victims tend to gain a superior epistemic grasp of the phenomenon and its background conditions “not because of their victimhood per se but because their experiences triggered a thought process” that illuminated what I have called “the mechanisms of invisibility” obscuring those experiences. The claim is not that victims necessarily go through these knowledge-enhancing thought processes; the strength of dominant frameworks means that victims might fail critically to capitalize on the moral rupture they experience, for instance.²² Instead, the claim is that victims are more likely to enter a privileged epistemic position regarding their interpretation of the phenomenon, given their “distinct starting point” of moral rupture (ibid. 17).

Contra Lilienfeld, then, we have reason to think that purported victim testimony is not merely a “subjective” interpretation of an ambiguous situation, where the victim is reduced to helplessly unreliable mindreading of the perpetrator. Victims are more likely to make reliable judgments about their experience; thanks to being more likely to have undergone a critical thought process initiated by their experience of moral rupture, victims are in a better position to spot various mechanisms that otherwise might obscure the phenomenon they experience. This makes victims “experienced observers”, or “trained experts”; as McClure and Rini (2020, pp. 7–8) points out, psychology has a long history of relying on such individuals because of their reliability.²³ This also gives us reason not to see victims, insofar as they are likely to be from marginalized groups, as ideologically biased. Instead, they are more likely to have a privileged epistemic vantage point on the matter in question; their experience likely makes them more expert at seeing past the mechanisms of invisibility that can be hiding the discrimination in question.

Compare this to independent observers, as well as to those who are less likely to have experienced the sorts of discrimination in question because they are not marginalized. Without the epistemic benefits of moral rupture, they are less likely to engage in a thought process whereby they gain critical distance from dominant frameworks and are therefore less likely to spot the mechanisms of invisibility that may be part of those frameworks. This makes “independent” observers, and those without experience of marginalization, less reliable – they constitute untrained observers – and more biased – they are more likely uncritically to inherit biased frameworks that hide the phenomenon under investigation.

This might also cause us to worry about other methods. As part of his emphasis on critical multiplism, Lilienfeld suggests that the results from victim testimonies ought to be tested against existing methods for studying prejudice. He says researchers must conduct “correlational or factor-analytic work to demonstrate that microaggressions cohere with other indicators of deliverer prejudice, whether they be implicit, explicit, or both”, suggesting that “it is incumbent on MPR [Microaggression Research Programme] proponents to demonstrate that ostensible microaggressions are statistically associated with at least some other well-validated indicators of deliverer prejudice” (ibid.). What we have seen about mechanisms of invisibility in dominant frameworks should make us cautious here; instead of reflecting poorly on the microaggressions hypothesis, negative results from these methods might reflect that they have inherited certain mechanisms of invisibility. Because of this, they may fail to make visible unconventional and currently poorly understood forms of prejudice.

What I suggest, then, is that we continue to reap the benefits of critical multiplism, but with stronger principles guiding the triangulatation of results that emerge from these multiple sources. For the reasons above, victim testimonies ought to be accorded special weight,²⁴ while results collected from other methodologies that do not support the hypothesis emerging from victim testimonies, ought to be subjected to increased scrutiny. That scrutiny ought to involve a reflexive consideration of the potential ways that alternative methodologies may have inherited certain mechanisms of invisibility that could affect their ability adequately to capture the phenomenon in question. This ought to be guided by the insights of victim testimonies, which are well-placed to illuminate those mechanisms.

What does this tell us about the study of attentional discrimination in particular? Well, if it is a multiply hidden phenomenon, victims are likely to be “experienced observers”, who are in a privileged epistemic position that allows them better to see the mechanisms of invisibility described in §5.1. Other methodologies do look promising in studying attentional discrimination; the eye-tracking studies mentioned in §5.1 might be able to get around certain aspects of the physiological/psychological mechanism of invisibility. One thing to note is that victim testimonies look well-placed to guide the selection of these other methodologies. For instance, consider Austin Baker’s (2022) research into subtle behavioral cues taken to indicate a discriminatory attentional pattern, including (lack of) eye gaze, gestures like nodding, bodily orientation, facial expressions, and so on. This is informed by victims’ testimonies of picking up on such cues, demonstrating how victims can help to clarify which methods are likely to be apt (such as those that capture eye gaze, body positioning, and so on).

6. Conclusion

In this paper, I have reflected on my research process into studying a particularly subtle form of morally problematic attention, noting the significant role that victim testimonies have played in that process. To justify the focus on this methodology, I have drawn on the hermeneutical injustice and standpoint literature to argue that victim testimonies are particularly important for studying unnamed and poorly understood forms of discrimination; the “moral rupture” that victims of such discrimination tend to experience provides the initial seeds of analysis. Some argue that for later, context-of-justification stages of research, however, victim testimonies ought to be supplanted by other methods that produce less subjective – less unreliable and biased – evidence. Instead, I have argued that they should continue to be given special evidential weight. When it comes to hidden forms of discrimination, victims’ experiences of moral rupture likely make them expert, trained observers, who are distinctively well-suited to uncovering the mechanisms of invisibility that shroud the forms of discrimination they experience. The insights they gain from their moral rupture, then, ought to be used to interrogate the adequacy of other methodologies – methodologies that are more likely uncritically to inherit the blind spots and inadequacies in dominant frameworks.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. As will be discussed in this piece, a key point of contention in this paper is whether these individuals are indeed ‘victims’. For instance, perhaps an individual complaining of being discriminated against has misread the encounter in question, thinking that there were racist undertones to an interaction when in fact there were none. To make things more readable, however, I will generally drop the qualifier ‘purported’ unless it is directly relevant to the point being made.

2. See e.g., Camp (2017), Watzl (2022), Siegel (2017).

3. This does not represent a universal experience for those who are disabled (the same goes for the other vectors of marginalization discussed here). Some are happy to be seen as “disabled first”; instead, complaints might focus on problems with the public perception and treatment of (and accommodations for) disability. My aim was to pick up on one trend in responses to social marginalization. Further it is important to note that even for those who do not want to be seen primarily as e.g., disabled, there can be multiple reasons for this, from finding one’s disability irrelevant (or simply not of primary relevance) to many contexts in which one finds onself, to wanting to deflect attention from one’s disability because of the negative social perceptions of it. See Whiteley (2023) for more discussion.

4. One plausible explanation concerns the relevance implicatures likely generated in certain of these examples; those involving utterances, such as the phrase “Black artist” used to refer to Basquiat, plausibly implicate various racist and/or othering messages on behalf of the speaker. Thank you to an anonymous reviewer for raising this.

5. For a fuller articulation of his account of attention, see Watzl (2017).

6. In addition to the citations in fn. 3., other recent work has been clarifying the ethical and political role of attention and salience, see e.g., Watzl (2022), Munton (2022) and Gardiner (2022).

7. There is already important work addressing how marginalized individuals are routinely made ‘hypervisible’ in virtue of their marginalisation (see e.g. Settles et al., 2019). The hypothesis offered here was an attempt to spell out in more detail - using tools in the philosophy of attention and salience - one way of understanding what this hypervisibility might consist in. The taxonomy of attentional patterns developed in Whiteley (2023) demonstrated that the issue might be with both a particular kind of attentional content (the properties of a person or social group, as opposed to the person or social group themselves), as well as particular quantitative manner of attention (a “relative attentional surplus” on that content, as opposed to a more extreme “attentional fixation”).

8. Discourse analysis refers to a method for interpreting specific communicative acts not in the abstract, but in the context in which they are communicated (e.g., interpreting a speaker’s utterance, taking into account their social position, the society in which they live, and so on). See Johnstone (2018, p. 2) for a definition along these lines.

9. As suggested in Whiteley (2023), there are contexts in which the prioritization of demographic properties is not a moral problem – indeed sometimes it might be morally required, such as when acknowledging a person’s epistemic authority on a topic related to their demographic properties. Here, I assume that prioritizing a victim’s demographic properties in the context of explaining how these relate to the epistemic insights of a moral rupture (as I discuss in §3.2), then, is not morally problematic.

10. There are disagreements regarding the best term to refer to, as well as characterization of, this phenomenon. As for the former, some prefer the term “micro-inequity”, for instance, as this conveys how the phenomenon needn’t involve any actual aggression (as plausibly demonstrated by the examples I give in the main body). I retain the term “microaggressions” because it remains the most widely-used term, and because of my focus in this section, which is the initial stages of developing new concepts. There are also disagreements about the existence conditions for microaggressions. Some suggest that they are constituted by implicit messages determined by the motives of the perpetrator, others focus on the negative impact on the victim, and others offer alternative accounts still (see an overview in McClure & Rini, 2020). Differences here would affect the methods required to study microaggressions. I do not attempt to settle these debates here, but I do presume that microaggressions have the sort of nature that makes it possible for victims to be able to pick up on their occurrence.

11. Fricker’s account of hermeneutical injustice has been precisfied and modified in various ways since her 2007 account, owing to the need to account for and fix various problems with it (see e.g., Berenstain, 2020; Medina, 2012; Sliwa, 2023). Outside of what I mention in the main text, I mostly leave these nuances aside; a coarse-grained analysis of the concept is sufficient for my purposes.

12. Later, this term was expanded to capture a range of other forms of undervalued, under-remunerated, and under-protected forms of work, affecting a wide number of groups outside of just (white, middle class) women.

13. Rollins (1985).

14. Here I assume that microaggressions are indeed characterized as implicit messages of this sort, though see fn. 9 for a brief discussion of alternative characterizations of microaggressions; in particular, some characterizations would mean that whether an encounter is microaggressive does not depend on ascertaining the perpetrator’s mental states.

15. A certain level of error is consistent with reliability. I take Lilienfeld to be suggesting, though, that the potential for error in victim reports is high given how ambiguous microaggressions tend to be.

16. Many have highlighted the harms and wrongs involved in giving insufficient space and weight to victim testimonies in the context of microaggressions (e.g., Freeman & Stewart, 2020; Thompson, 2020, p. 91). See Hänel (2021) for an illuminating discussion of these moral issues in the context of sexual harassment debates, as well as a discussion regarding how those moral issues intersect with epistemic ones.

17. I borrow talk of “mechanisms of invisibility” from Hatton (2017), whose focus is on sociological mechanisms that hide certain forms of work in particular. They focus on three such mechanisms: cultural, legal, and spatial mechanisms; this framework has influenced what I say here, though I categorize things a little differently.

18. In addition to Hatton (2017), see too Haslanger’s discussion of ideology and “narrative tropes” (Haslanger, forthcoming).

19. This tends to go hand-in-hand with another ideological trope, which paints victim’s complaints as representing “political correctness gone mad” (Sue & Spanierman, 2020, p. 20).

20. We can see here one that mechanisms of invisibility can intersect. The barriers to visibility generated by the ontology of attentional discrimination can be co-opted by ideological tropes, particularly those enabling the discrediting of the victim. This includes the “snowflake” trope described in this section. See Schroer and Bain (2020, p. 226) for a similar point regarding microaggressions.

21. Outside of issues of conscious awareness, this raises questions of control. Considering attentional habits in particular, Jules Holroyd’s work on implicit bias is relevant. They say that while an implicit bias is hard to control in the moment, insofar as it is a habit, we plausibly possess “long-range control” over it, insofar as we can intentionally cultivate and break habits over time (Holroyd, 2012, p. 284).

22. Hänel (2021, pp. 16–17) cites two studies of victims of rape, finding that those victims were unwilling to apply the term “rape” to what they experienced, including at times reciting the rape myths in dominant culture that obscure how those acts constitute “rape”.

23. Note that McClure and Rini do not straightforwardly endorse the claim that (purported) victims are trained experts; instead, they offer this claim tentatively, as a potential way of rescuing psychological accounts of microaggressions (namely, those holding that the existence conditions for a microaggression are contained in the psychological states of the alleged perpetrator).

24. This echoes Hänel’s (2021, p. 17) recommendation that the testimonial practices of victims of sexual violence be given “epistemic priority”. See Thompson (2020) for an alternative proposal regarding how to capitalize on the epistemic insights of victims – insights that help to reveal the limitations of dominant frameworks (including what I have called the mechanisms of invisibility that they encompass). She persuasively suggests using a “community science” approach, which includes non-scientist stakeholders in scientific research programmes, such as by training community members to collect data (ibid. 92). I take this proposal to be compatible with what I suggest here.