Welcome to our newest PEA Soup Blog Ethics discussion! This time we are discussing Ella Whiteley ‘s ‘A Woman First and a Philosopher Second’. The paper is accessible open access here. We are pleased to have Susanna Siegel kick off the discussion with a critical précis. Over to Susanna:

According to Whiteley, there’s a way of harming a person that consists in “a specific form of problematic attention…: attentional surplus on the wrong property”. Whiteley draws on a range of reports to argue that they are illustrations of this problem. The reports include:

• scientists who describe the burden of doing academic work around colleagues who view them as “women first, scientists second”.
• artists regularly described in public as “Black artists” or “Hispanic artists”,
as opposed to just “artists”. “Happy to be Black,” writes Zoe Kravitz, “Just don’t need it to be put in front of everything.”
• athlete Monica Korra, who was kidnapped and sexually assaulted, “has said she wants to be known as a runner not a rape victim”.
• Nathan Liu, a person who is blind, reports that his blindness is often more salient than practically anything else, in testimony echoed by other people with visible disabilities.

As these examples show, the status of being “the wrong property attended to” is not intrinsic to the properties of being: a woman, black, hispanic, blind, or a rape victim. Instead, Whiteley thinks the locus of wrongness is the distribution of attention. Just as earlier theorists Gurwitsch and Dicey-Jennings wrote about the structure of attention into background and foreground, Whiteley, following Watzl, thinks of attention as a way for the contents of a person’s mind to become prioritized, both relative to one another, and for the person. In the framework Whiteley shares with Watzl, more attention indicates higher priority.

Once we have vocabulary for talking about how a distribution of attention prioritizes one thing over another, we can ask whether the priorities in a given case are what they should be. Whiteley’s paper examines a way that attention’s prioritizing can go wrong.

According to Whiteley’s analysis, the bad-making feature of these cases has the form of a property getting more attention than it should. A full account of the bad-making feature would then elucidate what makes a property count as getting “too much attention” from a person, in a context.

One approach to what makes a property count as getting too much attention would focus on its role in creating an attentional deficit to a different property, and then explain in general terms what makes that other property merit more attention than it gets. The explanation of bad-making features would then bottom out in the account of attentional deficit, not attentional surplus.

Whiteley focuses on attentional surplus, though, for at least two reasons. First, the discussion is anchored by testimonies that foreground the property to which the testifiers do not want so much attention to be paid. Whiteley’s analysis takes at face value the idea that surplus, not deficit, is central. Second, Whiteley allows that in some cases, there is a surplus of attention to a property, with no accompanying attentional deficit to any other property (such as fixating on a visible disability).

Notice, though, that whether there’s an attentional deficit or not, on Whiteley’s analysis the locus of the wrongness is how much attention a property is paid. X’s prioritizing the wrong property of person Y, Whiteley holds, can constitute harming Y. (It can also be merely instrumentally harmful to Y, but there, the locus of harm is not attention itself).

But what could make it the case, on Whiteley’s view, that too much attention is paid to a property? Whiteley’s analysis needs an account of what could make a quantity of attention too much, such that paying that amount of attention to a property can constitute a harm, or be otherwise constitutively morally problematic.

I agree with Whitely that something is going badly in these cases. It’s worth trying to understand what exactly would need repair. One of many things I like about the paper is how Whiteley gathers a range of voices to identify a specific, repeatable, recognizable social phenomenon. I think Whiteley is right to group together cases where there’s a clear attentional deficit and cases where there isn’t, and to look for bad-making features common to both.

I also agree that the same bad-making features can cut across cases of interpersonal interaction, in which X selects “the wrong properties” of Y to pay attention to, and Y thereby experiences unwanted attention from X; and purely discursive cases, in which forms of words like “hispanic artist” or “female scientist” prompt readers (or listeners) to attend to the corresponding combinations of properties – even if no interpersonal interactions are involved, and even if different readers respond differently to the same prompt.

But I’m going to suggest that the proposal that attention to a property can constitute harm doesn’t capture the role of attention in what’s going wrong in these cases.

My main points are these: (i) the bad-making features are not located in attention to a property (or, therefore, in any quantity of such attention), and (ii) the role of attention in Whiteley’s cases is never constituting the bad thing. Instead, when holding fixed many factors that can vary with context even while attention to properties stays the same, attention to a property can predictably signal something bad. This context-dependent predictability explains why the testifiers articulate their objection in terms of unwanted attention. Attention is the means by which something problematic gets signaled, and it’s what objectors most immediately experience. They are understandably more focused on describing their experience, than they are on identifying what’s fundamentally going wrong in it.

Let’s start with the cases in which attention to properties is prompted by descriptors like “female scientist” or “black artist” that may appear in mass communication. These cases bring out the importance of discursive context in determining whether attention to a property becomes problematic.

Consider a pair of reviews of the same art exhibition, where the only differences are in the implied readers of the venue, and in the descriptors used to refer to the artist. Otherwise the reviews are the same. One review appears in a historically white newspaper, say it’s the LA Times, and refers to “a Black artist”. In the other review, the artist is referred to as “the artist”, and, let’s suppose, this review appears in the Bay State Banner, where the implied reader is likely to be black, and the stories often focus on black life. This review can direct attention to an artist’s property of being black, just by using the term “the artist” in that discursive context. So we have two reviews, both of which direct attention to the same two properties of the same person – being black and being an artist.

The LA Times review could easily illustrate the phenomenon Whiteley sets out to analyze, while the Bay State Banner review could just as easily not instantiate any such problem at all. What might make the difference?

Whiteley’s official answer is that there is too much attention to the artist’s blackness in the LA Times, whereas there’s the right amount in the Bay State Banner – even if by other measures, there’s in general a lot more attention directed to blackness in the Bay State Banner, given its mission to cover news about black life around Boston. This is a first clue that quantity of attention to a property is not the problem, and not what ultimately would need repair.

If we want to understand how there could be a difference between these cases, it seems we should look to the roles of attention to those two properties in the minds of readers – something highly sensitive to the outlooks of the individual readers (psychological context), their social position relative to the subject being described (the historical context), and the presumptions by the writers about implied readers (the discursive context).

If it’s easy to imagine a completely unproblematic art review in the Bay State Banner, perhaps that’s because we’re picturing readers for whom attention to blackness is diaphanous with respect to attention to the artist. It does not distract from anything, or implicate that the individual (by virtue of being black) is exceptional, remarkable, or out of place. In this contrast we see different lines of association a person’s mind could follow, upon attending to “black” and “artist” at the same time. Neither is essential to attending together to being black and being an artist, and the difference is not a function of how much attention is involved.

If the bad-making feature in problematic cases isn’t a quantity of attention, what might it be? Perhaps it is those aspects of a context that suggest that the reader is or should be ready to think of “black”, “hispanic”, or “female” as a potential anomaly with respect to “researcher”, “scientist”, or “artist”. Notice that the problematic thing seems to attach to presumptions, feelings, and other reactions to the relationship between these properties, not to the quantity of attention to any one of them. Such reactions vary, even when attention to the properties stays the same.

Turning to the interpersonal interactions, there, too, the bad-making features seem to lie in something only contingently signaled by the pattern of attention. If the physicist wants her colleagues to focus on the fact that she’s a scientist and not on the fact that she is female, that may be because the colleagues who see one another as “scientists first” will go on to interact in ways that correspond to that role, taking an interest in one another’s work, discussing lab arrangements, the replication crisis, and so on. Their problem is that the same thing isn’t true if they focus on the fact that one of their colleagues is female.

But just as “black” can be diaphanous with respect to “artist”, “female” could be diaphanous with respect to “scientist”. The colleagues’ problem is that for them, it isn’t. If we held constant that the colleagues have this problem, then we might identify the amount of their attention on “female” as a target of repair. But isn’t what’s ultimately in need of repair the colleagues’ point of view, in which “female” when placed next to “physicist” operates as a marker of anomaly? That’s what’s ultimately objectionable – their readiness to take on that point of view, treat it as normal, presume that others share it, and so on. When we don’t like the attention put on “female”, perhaps we don’t like it because we sense that it may be too close for comfort to this readiness.

What about the examples with no attention deficit? “It matters to us how we are seen,” writes Whiteley. It does, and perhaps that’s because it matters to us how others are ready to relate to us – readiness to relate in a certain way is already a kind of relationship. When X gawks at Y’s visible disability, they may signal to Y their uncertainty about which interactions are possible, or their reluctance to presume that the two of them can interact with ease and compassion. And what is a person objecting to, when a stranger pats their pregnant belly, or when a someone beguiled by Afros puts their fingers on the hair they find so fascinating? Perhaps they’re objecting to being seen as someone whose hair or belly is touchable by non-intimates. One may also object to the actual touching, but I take one of Whiteley’s insights to be that something objectionable would be going on even without actual touching, only a perception of touchability, made manifest by gawking.

In all these cases, attention is vehicle for indicating a readiness to relate in problematic ways. When recognizing such readiness to relate to us makes us uncomfortable, it’s natural to identify the culprit as too much attention. But the real culprit is the social readiness indicated by attention. Attention never does that all on its own, and that’s why it isn’t what fundamentally needs repair.

24 Replies to “Ella Whiteley: ‘A Woman First and a Philosopher Second’. Précis by Susanna Siegel

  1. This is a brilliant paper that makes an exciting contribution to our understanding of the normative dimensions of attention. Although I’m very sympathetic to the proposal I have three queries:
    1. Are the morally problematic cases constitutively harmful?
    2. If so, is this harm constituted by attentional surplus rather than by some other psychological phenomenon?
    3. If so, is the attentional harm best understood as disrespecting the victim’s identity?
    The first and second questions are ones that I worry about with my own work, so I would be glad if they are easily despatched. The third might mark a point of divergence between us: I think that the harm of attentional surplus is not a matter of disrespecting identity but rather one of unfairness.

  2. 1. Are the morally problematic cases constitutively harmful?
    You say ‘attending to an individual so that their non-personhood-related traits are their most salient feature is a way of disrespecting their personhood — something that is constitutively bad.’ (p.8, my emphasis) The distinction between instrumental and constitutive harm is explored in Section IVE. Although you note the harmful consequences of surplus attention you claim throughout that it also constitutes a harm to its victim.
    For me this raises some tricky questions around the nature of harm. It’s easy to see how the downstream effects of surplus attention harm the victim, but here’s a line of argument against surplus attention itself constituting a harm:
    i) For something to constitute a harm to the subject it must make a difference to that subject
    ii) Surplus attention (in and of itself) doesn’t make a difference to the target subject
    iii) Therefore, surplus attention (in and of itself) doesn’t harm the target subject
    One way to test whether there’s a constitutive harm here is in terms of legitimate demands for apology. Imagine one of the victims calls out a perpetrator on their wrongdoing. The perpetrator takes responsibility for their overt treatment of the victim and apologises but they do not acknowledge any harm caused by the attentional surplus itself. Would the victim have legitimate grounds to ask for a further apology? Could they say, for instance, ‘I’m glad you’ve taken responsibility for giving the male philosophers preferential treatment but I think you owe me an apology for paying such undue attention to my gender in the first place.’ And why couldn’t the perpetrator defend themselves by saying ‘I’ve apologised for everything that had an effect on you but I’m not going to apologise for things in my head that didn’t make any difference to you.’ (This is a variant of an argument offered by Basu and Schroeder (2019) in the related context of doxastic wronging i.e. wronging someone in virtue of a belief you hold about them.)

  3. 2. Is the harm constituted by attentional surplus rather than by some other psychological phenomenon?

    Our mental states our tightly intertwined. Patterns of attention reflect an agent’s values, emotions, beliefs, desires, attitudes, prejudices and more besides. In each of the examples offered, it looks like surplus attention won’t be the only morally problematic thing going on. Someone who attends to the race of an artist first and their status as an artist second is probably going to have other problematic mental states: perhaps a belief that their art isn’t equal to that of white artists; a bias toward negative appraisals of their art; or an inappropriate emotional response to them personally. Assuming these all qualify as constitutive harms do we need to say that the surplus attention constitutes a further harm? Could it not be that the surplus attention is only judged as morally problematic because it’s associated with these constitutively harmful mental states. Again we can apply the apology test here: if a perpetrator took responsibility for all those morally problematic mental states whilst not taking responsibility for their surplus attention as such, would the victim have legitimate grounds for complaint? Or is the surplus attention just a symptom of the real constitutive harm?

  4. 3. Is the attentional harm best understood as disrespecting the victim’s identity?
    Let’s assume there are good answers to 1 and 2 and that surplus attention can indeed constitute a harm. This leaves us with the question of what the bad-making feature is of this surplus attention. The answer you offer is that attending to the wrong property disrespects the subject’s identity. I see three kinds of case that are hard to capture on this account.
    A) The same trait can be good to prioritise in one context but bad to prioritise in another. This is discussed in Section VB with examples such as it being fine to attend more to someone’s gender in the context of an event promoting women in philosophy. You deal with this by saying that whether a trait is functioning as an identity trait changes with context, meaning that what attentional patterns constitute disrespect change with context too. I’m not sure whether identity-related traits fluctuate with context like this so wonder if there’s a better way of accounting for the context-relativity.

    B) Some harmful cases of surplus attention don’t seem to have much to do with your identity. If your parent attends to your failures more than they do to your successes then it’s plausible that you’ve suffered a harm of the kind targeted by the paper. But the harm doesn’t seem to depend on those successes being integral to your identity. The successes and failures might even be in domains that you’re totally indifferent to, but the misprioritisation of your failures would still be harmful.

    C) Many cases of surplus attention only seem morally problematic relative to the perpetrator’s wider patterns of attention. Imagine someone for whom gender is always salient – perhaps because they spend their life researching issues of gender equality. They see women at philosophy events as women first and philosophers second, but they also see men at those events as men first and philosophers second. The disrespecting identity account seems to yield the same verdict for each case: if it disrespects identity to prioritise your gender in the first case then it must also do so in the second. But to me the second case really doesn’t seem so morally problematic (I can’t speak from personal experience with this particular example but can think of parallel cases where prioritising traits the same way for everyone makes an attentional structure okay).

    An alternative account that might capture cases A-C is one on which a person is harmed when surplus attention to a trait is unfair. In C, it’s harmful for someone to see you as a woman first and a philosopher second whilst seeing the men as philosophers first and men second. Once you take the inequality out of it the moral harm seems to disappear. In B, the parent paying more attention to your failures than to your successes is unfair regardless of how they relate to your identity. And in A, what counts as a fair distribution of attention is something that changes with context. Attending to your gender at the event for women in philosophy is not part of a discriminatory pattern of attention, whereas the distribution of attention in the original case would be. This is not to say that identity is irrelevant to the moral story here: in each of the testimonies, the unfair distribution of attention is driven by identity-related traits. But maybe the easiest way to accommodate A-C is to say that unfairness, rather than disrespect to identity, is the real moral difference-maker.

    Reference: Basu, Rima & Schroeder, Mark (2019). Doxastic Wronging. In Brian Kim & Matthew McGrath (eds.), Pragmatic Encroachment in Epistemology. Routledge. pp. 181-205.

  5. Thank you so much to Susanna for writing such a vivid, clear, and illuminating précis; it’s such an honour to have your rich, wonderful analysis to engage with. I know I’ll be returning to it plenty after the PEASoup blog closes.

    Siegel’s key point of contention with my argument is that “the proposal that attention to a property can constitute harm doesn’t capture the role of attention in what’s going wrong in [the testimonies]” for two reasons: (i) the bad-making features are not located in (any quantity of) attention to a property, and (ii) the role of attention is never constituting the bad thing. To respond, let’s first consider the examples she develops to make her points.

    Siegel develops a great example pertaining to reviews of the same art exhibition. One is in the historically white LA Times, and refers to ‘a Black artist’. The other is in the historically Black Bay Street Banner, and refers simply to ‘the artist’. Despite the latter simply referring to ‘the artist’, Siegel suggests that features of the specific discursive context around Bay Street Banner – notably, that the implied reader is likely to be Black and the publication focuses routinely on Black life – mean that this phrasing nevertheless directs attention to the artist’s Blackness. The first thing Siegel takes from this example, then, is the contingency with which a given phraseology triggers a certain attentional pattern in readers. Secondly, both reviews direct the same high level of attention to Blackness. This means that my account struggles to account for how one is morally problematic (LA Times) while the other is not (Bay Street Banner), despite both involving the same quantity of attention to Blackness. For Siegel, this provides the ‘first clue’ that attentional quantity to a property is not the issue.

    The first thing I want to say is that I’m not convinced that both reviews direct attention to the artist’s property of being Black to the same extent, contra Siegel’s analysis. I need to think about this more (which goes for all of my replies here…), but here’s an early sketch of my thoughts. First consider a line from Monica Esopi’s testimony, a doctoral candidate in chemical engineering (at the time of the testimony). While I summarised Esopi’s interview in the original paper, here’s more of the original quote: “I now work in a department where diversity-related issues are acknowledged and discussed openly and frequently … I no longer feel like I’m seen as a woman first; I am just a researcher, a scientist, an engineer.” Taking this on face value, it looks like being in a context where diversity-related issues are discussed ‘openly and frequently’ is consistent with decreased attention to any one individual’s marginalised identity.

    This might seem counter-intuitive, but I think we can make sense of this. Think of an entirely different context: certain mindfulness techniques. Here, explicitly and regularly confronting one’s worries – letting them ‘pass through’ one’s mind instead of pushing them into its recesses – can ultimately work to reduce their salience in one’s attention. Perhaps something a little similar is going on here; we can lessen the strikingness (i.e. one form of salience) of a person’s marginalised identity by talking about issues surrounding that identity head-on. As a man, say, gets more familiar with thinking about gender issues in his department thanks to its new EDI initiatives (etc.), he might start to understand more about gender, have more healthy relationships with women, and so on. This could work to lessen his previous (what we might call) ‘exoticisation’ of ‘womanness’, which in turn makes the gender of his colleagues who are women less striking/ salient.

    This isn’t to say that gender, or race to return to Siegel’s review example, falls out of the picture. There’s something importantly right about Siegel’s claim that ‘the artist’, in the discursive context of the Bay Street Banner’s art review, directs attention to Blackness. What I want to suggest, though, is that it is unlikely to be selectively prioritised over the other properties of that artist, which is key to my description of a relative attentional surplus. (Note: this definition is to bring out the contrastive nature of attention, to capture the foreground/background nature of attention. It doesn’t necessarily mean that the properties in the ‘background’ represent an attentional deficit – we’d have to look into whether or not the backgrounded properties are morally relevant.) I suggest that the discursive context of The Bay Street Banner – the predominantly Black readership, the frequency with which Black issues are discussed – does make Blackness salient in the publication overall. But it plausibly does so in a way that, perhaps counterintuitively, helps readers to attend better to other attributes of the specific individuals it writes about and reviews. The Blackness of the individuals written about is not saliently striking for the publication’s readers, plausibly. We might cash this out in terms of Blackness being granted a diachronic lower-level salience across the publication’s articles. (If one doesn’t like the idea of salience being graded, then one would instead talk about a property occupying a high-ish – but not the highest – rung on its authors’ and readers’ priority structures.) That is consistent, I think, with this particular review, which mentions “the artist”, not making this individual’s Blackness their most salient feature. This would preserve the sense in which it is the quantity of attention, contra Siegel, which is the thing making the difference between the two reviews (one selectively prioritises Blackness over everything else, the other doesn’t).

    The problem is that this answer doesn’t get to her real argument. We could amend the example so that Blackness is granted a high level of salience in the Bay Street Banner, but in a way that doesn’t seem intuitively problematic. Then we’d still need to answer Siegel’s question: what makes the difference such that identical attentional patterns either count as a morally problematic surplus in the one context, but not the other? Indeed, we’d also encounter this problem with the low-level salience I describe above: there will be cases where it seems acceptable, others where it doesn’t. Siegel’s primary move here is to suggest that it is in certain of the features of the context where we find the moral difference-making features, not in the attentional pattern.

    She begins by homing in on what she calls psychological context; the claim is roughly that (usually?) identical attentional patterns can be bad vs. good depending on what other mental states or activities they’re accompanied with (presumptions, feelings, other reactions), or more broadly the wider outlook the individual has. For instance, she suggests that what makes the attentional distribution invited by the LA Times piece bad is likely other psychological things, such as the proneness of certain readers to treat ‘Black’ as a potential anomaly with respect to ‘artist’.

    But one thing I want to claim here is this: sometimes attentional patterns can come apart from other standardly associated mental states. Consider the following example (this might seem to conflict with the diagnosis of Esopi’s testimony I give above, but I think they can be made consistent). Imagine someone who is in a predominantly male Philosophy department in the UK, but who has read a good amount of relevant literature on dismantling the patriarchy, has a gender diverse friendship group, etc. Despite this, it remains the case that very few of their colleagues are women, very few philosophers in the media/ ‘conventional’ canon are women, etc. As I discussed in the paper, the simple statistical distribution of the Philosophy world around them means that they will therefore likely have attentional patterns that involve finding the gender of their women colleagues salient (thanks to our attentional patterns being so sensitive to signal-to-noise ratio).

    One of the reasons to think that such a scenario is possible – where one might have problematic attentional patterns but not the usual accompanying bad beliefs, feelings etc. – is that we’re plausibly less good at noticing and reflecting on our attentional patterns, than we are for many of our other mental states. As for perceptual attention, for instance, studies show that participants have a particularly poor awareness of their eye movements (see e.g. Clarke et al., 2017, ‘People Are Unable to Recognize or Report on Their Own Eye Movements’). If this is the case, this could mean that any problematic attentional patterns we have would be harder to correct. The opaqueness of our attentional patterns to ourselves would mean that even those who explicitly abhor e.g. sexism, can unintentionally attend excessively to the gender of their women colleagues.

    Another potential case where attentional patterns come apart from standardly associated mental states is this: aggregate attentional patterns. In some scenarios there may be morally permissible reasons for each individual attender to focus on a e.g. a speaker’s womanness at a talk – and yet if everyone is selectively prioritising the speaker’s womanness, this seems troubling to me. (Georgi Gardiner considers similar cases in their wonderful chapter ‘Attunement: On the Cognitive Virtues of Attention’.) Here there aren’t consistent and/or bad other mental states to find fault with in each token attentional pattern, and yet the attentional patterns on aggregate seem morally problematic.

    The issue is, Siegel would likely explain these cases, if she conceded that other psychological features weren’t to blame, by citing other aspects of context: the ‘historical context’ and ‘discursive context’. For Siegel, it’s always in factors pertaining to these contexts that we find the ‘real’ sources of the moral problem, which the patterns of attention only ‘contingently’ signal. Ultimately, though, I think that Siegel is wrong to suppose that dependence on context means that attentional patterns can’t be constitutively problematic.

    For starters, the idea of x constituting a moral flaw is often treated by others as consistent with this being context-dependent. Consider speech act theory (which admittedly uses a specific notion of ‘constitution’ – more on that in subsequent replies – I see Tom has raised some similar issues!). For Rae Langton, for instance, pornography doesn’t just cause, but can constitute the subordination of women. However, Langton flags that this is contingent on various contextual conditions being met, such as that the speaker has the relevant authority, that the utterance occurs against the background of certain gendered power relations, and so on. This idea is affirmed across many who write about speech acts. Zheng and Stear’s’s recent Ethics article on “What’s Wrong with Blackface says “which action an utterance constitutes is also context dependent.”. They then go on to apply this rule to private imaginings in the mind, not just utterances. Here, I think it would be odd to insist that actually the utterance isn’t the real moral issue, or what needs repair. More natural I think would be to suggest that both the utterance and the contextual features are morally problematic (perhaps co-constituting a particular wrong/ harm, perhaps), and both need repairing. Can’t the same go for the attentional patterns I discuss?

    I think there could be different ways of defending this sort of picture. One was (tentatively – so we can’t blame him if it doesn’t work, ha) suggested to me recently by the excellent Facundo Rodriquez. He talked about debates in particularism perhaps lending some clarity to this issue. Sometimes, one points to features of the context in a way that debunks the moral relevance of factor x, upon which one initially was focussed. These contextual features are doing the real work, it turns out. Other times, one points to contextual features that simply enable (or disable) the original factor x to become morally relevant/ ‘have their effect on the moral nature of the whole’. One example given by Michael Ridge and Sean McKeever’s Stanford entry on Moral Particularism says this: “the fact that one’s audience will appreciate a (non-offensive) funny remark enables the humor to be a reason. Here the humor is a reason, but only against the background of a receptive audience; the background functions as an “enabler”.” Perhaps we can see the contextual factors I point to in the paper as enabling (or disabling) the moral valence of the attentional patterns. This would allow me to preserve the idea that the attentional patterns really are wrong-making features.

    Siegel ends with a discussion of actionability – readiness to relate to people in certain ways as being the likely source of the problem. I think of her ‘soliciting affordances’ (2014) in this vein. All I’ll say here is that I agree this is an important dimension to consider, but not the only one, as I’ve tried to argue. I think that cases that just involve selectively prioritising the wrong feature of a person or social group can be morally problematic even if they don’t involve a readiness to act to others in morally problematic ways.

  6. I’m so grateful to Tom McClelland for this commentary—both for its kind words and its incisive and important questions. I’ll (try to) answer them in turn. Some responses echo those I have for Susanna Siegel, so I’ll point that out where that happens. (I’m also switching from speaking to the commentator in third-to second-person as I think it will be more readable – sorry Susanna!)

    1. Are the morally problematic cases constitutively harmful?
    The first thing to say here is that I am not committed to the attentional patterns constituting harm in particular; I talk about them being constitutively morally problematic, leaving it open as to what that more specific moral problem amounts to. They might be constitutively wrongful, disrespectful, unjust, and so on. I bring this up in case some agree that harm has to make a material (?) difference to a subject (making them less well-off), but that the same is not required for e.g. wronging or disrespecting a person.

    I do think that there would be legitimate grounds for apology in the case you mention. I suppose in part this might turn on how sympathetic one is to related views in other debates. Do you think you can be wronged by your partner cheating on you even if you don’t find out about it? Do you think your well-being can be hurt by your hopes and desires being frustrated after you die? Those who are sympathetic to positive answers to such questions might be persuaded by the idea that attentional patterns can e.g. wrong/ disrespect you even if you don’t find out about them. You might here wonder how narrowly or widely we are to interpret your phrasing of ‘making a difference’ to the target subject, then.

    I also think there are quite a few moral frameworks in Philosophy that don’t focus so much on measuring the ‘difference’ a given mental state/activity or behaviour makes to its victim, to nevertheless find that the e.g. mental state in question is constitutively morally problematic. Virtue theory, for instance, arguably focuses on whether one is exemplifying a virtue or vice, whether or not one is definitively making a ‘difference’ to a given individual. And many (e.g. Georgi Gardiner, Chappell and Richard Yetter Chappell and Helen Yetter-Chappel, Nicolas Bommarito) find that attentional patterns are constitutive of vices and virtues. I also think that discussions of Strawson’s ‘participant stance’, i.e. seeing and acknowledging others as persons, not things, a notion used in moral discussions from objectification, e.g. Rae Langton’s ‘Sexual Solipsism’, to doxastic wronging, e.g. Rima Basu, ‘What Do We Epistemically Owe to Each Other’), could be analysed in similar ways.

    I do recognise, though, that not everyone will subscribe to this sort of view. For those who are unpersuaded, I think there might be other ways of rescuing a constitutive evaluation of the attentional pattern, albeit not an entirely non-instrumental one. Here, we can turn to tools in speech-act theory. Many experts there, such as Rae Langton, Catharine MacKinnon, and Mary Kate McGowan, have described ways in which an utterance can constitute an act that is itself morally condemnable. As the utterance ‘I do’ in an official marriage ceremony does not just cause two people to become married, but itself constitutes the act of marrying, a ‘Whites Only’ sign, hung in a restaurant in Apartheid South Africa, does not just cause pernicious effects, but itself constitutes subordination. In particular, the latter ranks black people as inferior, deprives them of rights, and legitimates discrimination against them. The notion of constitution being employed here does not attempt to screen off real-world effects of (in this case) the utterance; it focuses instead on drawing a norm-driven (as opposed to simply causal) connection between them. So, one option is to concede that for something to constitute a moral problem for the subject, it must make a difference to them, and therefore to limit my analysis to this subset of attentional patterns (e.g. those that go on to make a difference to the attendee). I’d then say that the connection between the attentional pattern and its negative effects is not just causal but constitutive. Robin Zheng and Nils-Hennes Stear (in their recent Ethics article “What’s Wrong with Blackface”) apply this framework to imaginings, so there’s scope I think to extend this analysis to internal mental states and activities, including attentional patterns.

    2. Is the harm constituted by attentional surplus, rather than some other psychological phenomenon?
    I agree that often our mental states are tightly intertwined, and that a problematic attentional pattern will often sit alongside problematic beliefs, emotional responses, and so on. But, not always. I’ll largely point to my response to Siegel here, where I talk about cases of problematic attentional patterns coming apart from other commonly associated problematic mental states, as well as my comment about aggregate attentional patterns.

    A final point I’ll mention is this: can’t attentional patterns partially constitute another mental state (as per a suggestion considered in Sebastian Watzl’s excellent 2022 chapter ‘The Ethics of Attention’? E.g. an outlook, etc.? One might say that part of what it is to have a racist outlook is to e.g. notice primarily the Blackness of individuals who are Black, finding their distinctive and/or personhood-related traits less salient. If the outlook is morally bad, then the attentional patterns that partially constitute it are also morally bad. So, this response doesn’t deny that other psychological phenomena are (inevitably?) connected to the attentional patterns in question, but it does deny that this connection entails that it’s the other psychological phenomena that therefore soak up all the moral wrongness.

    3. Is the attentional harm best understood as disrespecting the victim’s identity?

    A) You say that you are not sure that identity-related traits (I say personhood-related) fluctuate with context in the way that I assume, so that e.g. a woman’s gender can function as a non-personhood-related trait in a departmental meeting, say, but instead is a personhood-related trait in a Woman in Philosophy event, as per your example.

    I suppose I just don’t share your worry here; contextual fluctuation regarding which properties (/don’t) showcase one’s personhood seems plausible to me. For instance, I think it’s quite plausible to those writing about objectification; there, the body is often conceived of as functioning as a non-personhood related trait in many contexts, which is what generates the distinctive moral problems in objectification. But those same writers could hold that the body can function as a personhood-related trait in other contexts, such as an athletic event, for instance. More needs to be said here, I realise, but I’ll leave it here at different intuitions.

    B) You say that some harmful cases of surplus attention don’t seem to have much to do with your identity(personhood), so my disrespect-to-identity(personhood) diagnosis doesn’t work. The first thing to say here is that I am not claiming that disrespect-to-personhood will be the right diagnosis for all attentional surpluses. It’s just one I find compelling for this particular case. So, I would just agree that the example you give – of your parents attending more to your failures than your successes – looks like it probably requires a different diagnosis. Unless I am missing something (totally plausible…), that is an acceptable outcome (namely, the issue of an attentional surplus being wrongful, harmful, disrespectful, and/or unjust in different ways in different scenarios). Incidentally, the reason I think the disrespect-to-personhood diagnosis is promising for the specific testimonies I surveyed is that the individuals all reoriented people to attend to properties that were quite clearly showcasing their personhood, such as their career status and achievements.

    C) You say that many cases of surplus attention only seem morally problematic relative to the perpetrator’s wider patterns of attention; a gender researcher for whom gender is always salient might see all philosophers – women, men, nonbinary etc. – in terms of their gender before their career. But, you suggest, it doesn’t seem to disrespect the men’s identity(personhood). So what gives? I think that this is a great example. I’m inclined to say that perhaps both the women and men have their personhood disrespected here, but that other factors mean it’s much less morally problematic for the men to have had their gender selectively prioritised, because of their overall privilege. I’m also put in mind of what Rima Basu says in the context of what we epistemically owe each other. She says “There is something more that is epistemically owed when it comes to our attitudes and beliefs towards members of non-dominantly situated groups..” (Basu, 2019: 924 – ‘What we Epistemically Owe to Each Other’). She takes a case of mistaking a white man at a Beyoncé concert for a staff member. She says “A mistake has been made; a wrong has been done. You failed to relate to him as he sees himself … You, in short, observed him in the way a scientist observes the planets. Still, though we can grant that there is wrong in this case, it is much less severe than the wrong done to members of marginalized groups when we form beliefs about them from outside the participant stance and therein treat them as objects.” (ibid.). I’m inclined to go down a similar route for this case; it’s bad both for the women and men, but more bad for the women.

    I realise that there will be some cases that just don’t seem morally problematic, though; a Black person navigating a white institution might find the Blackness of others at that institution salient, but only because they are searching for those who will truly understand their experience as minoritised in the institution, etc. Here, I think features of this specific context mean this individual is likely *not* disrespecting other Black individuals’ personhoods. To borrow something from my response to Siegel, these contextual features might ‘disable’ the standard moral valence of the attentional pattern in question.

    I really like your suggestion about unfair treatment capturing some of the issues regarding attentional surpluses. Indeed, in my recent work, I’ve started to think about attentional surpluses in terms of ‘unequal consideration’, thanks in large part to a discussion with my excellent colleague Adam Lovett. I have been considering how different moral diagnoses might depend on the comparison class one takes for a given attentional pattern; deontic diagnoses, regarding e.g. disrespect-to-personhood, will often be appropriate for comparisons of properties attended to (e.g. the body) versus those neglected (e.g. conversational contributions). Unequal treatment diagnoses will often be appropriate for comparisons of an attentional patterns on one group (e.g. women) versus that on another (e.g. men). So, I do see the potential importance of unequal treatment diagnoses for the testimonies in my paper, given that they also touch upon comparisons of the latter sort too. However, I want to suggest that *both* diagnoses are relevant, to bring out different kinds of moral problem going on in the scenarios. Further, the unequal treatment/ unfairness diagnosis will presumably depend on some further diagnosis that specifies what it is that makes one pattern of treatment worse than another. That won’t always be disrespect-to-personhood, as your example of a parent differentially attending to the successes versus weakness of their children shows. Sometimes it will be the *effects* of the differential treatment. But sometimes (often?) it could be disrespect-to-personhood; what makes the differential treatment of the women and men in the department example unfair, is that the trait that is selectively prioritised is one that showcases personhood for one group (men), but not the other (women).

  7. Thanks, all for the illuminating discussion here, and especial thanks to Ella for the wonderful article. I’ve been thinking about attention from within epistemology rather than from a moral perspective, but I find the exchange between Susanna and Ella really helpful because it brings out two key questions that I think any normative evaluation of attention needs to grapple with…

    The first is: what is the appropriate locus of evaluation here *within attention*. It doesn’t often seem like one individual instance of attention is especially significant either way, it tends to be patterns or dispositions to attend. Is that right, or could one individual instance of this kind of surplus attention be morally problematic all on its own? Ella thinks we can have a surplus of attention *without* that involving a deficit elsewhere… I tend to think we need to evaluate attentional dispositions as a set or network, where attention to anything inevitably comes at the cost of attention elsewhere, and where the normative significance arises at the level of that network as a whole, not at the level of individual nodes or instance of attention within it…

    The second is how we decide what work *attention* is doing, and what work is being done by the inputs and outputs of attention… which of these levels is where the wrongness resides? And is that the same as asking which of these levels would need repair to fix the problem? I think that’s an interesting question partly because it has implications for philosophical methodology more broadly, especially in these sorts of cases. We tend to want to convince ourselves that a certain phenomenon is significant by imagining some situation in which we have it in isolation, and then seeing what follows from that. But that can involve some pretty significant idealising away from any real world scenario, in a way that can alter the target of our analysis in the process. Suppose that Ella is right that attention is where the wrongness arises in these cases. That clearly doesn’t mean that attention would exhaust the problems here. Attention is significant in part because it gives rise to a host of downstream consequences in thought and behaviour. And how we attend is itself the produce of a very complex set of causal factors that are both internal and external (whatever that means) to the individual doing the attending. There are inevitably going to be multiple levels of normative analysis jostling alongside one another. And a fix at any one level is unlikely to fix all other levels… In my work I’ve argued that attentional patterns can be constitutive of prejudice, so I’m inclined to be on board with Ella’s claim that they can be a locus of wrongness in their own right. But how do we determine that? What’s the right methodology to adopt here? And to what extent should we seek a reduction of these levels to one another, and some ultimate locus of wrongness?

  8. First, thanks to Ella for writing this paper, and to Susanna, Tom and Jessie (so far!) for helping me develop my thoughts about this topic. I am so happy to be part of this conversation, and I know I’ll be continuing to think about it for the next several months. I have two main initial thoughts, both in the spirit of friendly amendments.

    I share Ella’s intuition that the structure of attention can be morally problematic in itself, independent of the particular mode of attention (e.g. surprised, pitying, leering…), content (e.g. being Black, female, disabled…), or consequences (e.g. inferring further properties, taking action). I also share her intuition that the wrong often involves disrespecting personhood. But I also share Susanna’s and Tom’s worries that attention is interwoven with cognitive, affective, and practical effects, which often seem like essential contributors to the wrong. And I think Jessie is right to emphasize that moments of attention are essentially embedded in diachronic patterns which can themselves be morally (and epistemically) problematic. Putting these worries together, it might seem like there just couldn’t be anything wrong with attentional structure at a moment, only with whatever caused and follows upon it.

    My first thought, then, is that I want more clarity about what’s constitutive of synchronic attention, as opposed to antecedent or consequent to it; and how that matters for diagnosing wrongs. In order to redeem Ella’s (and my!) intuition about the inherent wrongness of some synchronic attentional structures, I suspect we may need to bake more into moments of attention themselves than I think she has so far.

    Ella follows Sebastian Watzl in treating attention as “a kind of prominence or centrality in consciousness” resulting from “the activity of structuring [one’s] occurrent mental states…a type of mental management.” It might be useful here to distinguish several factors within ‘prominence or centrality’, and how they contribute to ‘mental management.” In particular, I like Amos Tversky (1977)’s analysis of a feature’s relative prominence (which he calls salience) as a combined function of two factors: intensity, or how much it departs from a baseline; and diagnosticity, or how relevant it is for informing us about other features and categories. At a given moment, our attention is drawn to certain features over others insofar as they are intense and or diagnostic for us. But these intuitive, synchronic assignments of attentional structure presuppose a host of background assumptions about which features matter given our interests; how and why features do and don’t cluster together; and how those features and clusters justify certain affective and evaluative responses given those expectations and interests.

    So, I would suggest that in objecting to disproportionate attention at a moment, one is at least in part objecting to a background perspective: a style of ‘mental management’. But I’d also like to say that the momentary attentional structure is itself objectionable insofar as it isn’t merely a causal effect of but a direct instantiation of that perspective, which in turn functions to produce certain patterns of inference, feeling, and action.

    I’m curious, Ella, whether you feel like I can say both that the background perspective is what makes a synchronic attentional structure objectionable and also that because that synchronic attentional structure occupies that functional role, it is constitutively rather than merely instrumentally objectionable.

    My second thought is more inchoate. Like Tom, I find it odd to say that what counts as a (non-) personhood-related trait depends on context (specifically, on the context of ascription). It seems like Zoe Kravitz is saying that she is *always* happy to be Black – that this is an important part of who she is; it’s the relative attentional structure she finds problematic. Likewise, it seems like Basquiat shouldn’t be forced to disavow race as an important aspect of his personal identity, or even of his artistic identity, in order to claim his place within the category of ‘artists simpliciter’. Indeed, I take it that your core goal is to pinpoint the wrong of just this sort of phenomenon: where both features deserve attention because both are personhood-constituting; but the attentional structure is still problematic.

    Not surprisingly given what I said above, I feel like an important part of the wrong here is that the interpreter presumes to impose their background assumptions about what should be surprising, interesting, or praise- or blame-worthy. Sometimes the presumption may be misplaced because it encodes generally inaccurate statistical or causal presuppositions – about the frequency of artists among Black people or Black people among artists, for instance; or about the inevitable consequences of blindness. Sometimes it may be misplaced because it encodes generally misguided interests or attitudes – for instance, pity at being blind or having been raped; or sexual interest in subordinates; or restrictive norms of beauty. But sometimes the presumption may be misplaced not on general grounds, but because it fails to respect the attended-to individual’s right to set their own priorities and attitudes. Maybe being Black was central to Basquiat’s artistic identity; maybe in that sense he was a Black Artist. But if so, that’s for him to decide – it shouldn’t follow as an entailment of his being Black plus his being an artist. And in any case, it shouldn’t affect his belonging to the category of ‘artist simpliciter’.

    So, I’m suggesting that at least in some cases, disrespect involves a failure to accord agents the freedom to structure their various traits as they see fit (or, perhaps, the freedom to earn an interpretation that structures those traits in a certain way) by living in their own particularity, as opposed to having that structure imposed in virtue of general assumptions about how certain traits correlate, constrain, explain, and justify each other. I’m curious, Ella, whether you find that plausible for some cases, and/or think that covers a wide enough range of the cases you’re interested in.

    Reference: Tversky, Amos (1977): “Features of Similarity,” Psychological Review 84, 327–352.

  9. Thank you so much to Jessie Munton for this excellent commentary, asking some hugely important questions. I’ll try to be more brief in my responses from now on (sorry, readers!).

    You ask what the appropriate locus of evaluation is *within* attention – can it be an individual instance, or is it more likely to be patterns or dispositions? Often I think it will be patterns and dispositions. But Georgi Gardiner (‘Attunement: On the Cognitive Virtues of Attention’) provides two persuasive examples of individual instances that seem problematic: “Visually focusing on disfigurement, even fleetingly, can constitute improper attention, for example, regardless of broader attentional habits. And continuing a causal telephone conversation when a nearby stranger has just fallen from a pier constitutes inappropriate disregard. Even if you can’t help them, their falling warrants attention.” As for your suggestion that normative significance will arise at the level of the ‘set or network’ of attentional dispositions as a whole, I’m not so sure if this has to be the case. Consider someone who spends a staff meeting selectively prioritising the bodies of colleagues who are women while selecting the conversational contributions of colleagues who are men. They never repeat this attentional distribution. If this never gets entrenched in the individual’s ‘set or network’ of attentional dispositions, I worry this won’t get counted as problematic for you, when it seems it to me.

    You also comment on my suggestion that we can have an attentional surplus without a corresponding deficit elsewhere. Something I mention briefly in parentheses in my response to Susanna is relevant here; attentional patterns, as I describe them, are contrastive. Some things are foregrounded, others backgrounded (some don’t even enter the conscious field). So yes, attention to one thing inevitably comes at the cost of attention elsewhere. But, whether we call the lack of attention elsewhere a *deficit* depends on whether we think the thing not attended to (or attended to less) matters, morally. I treat ‘surplus’ and ‘deficit’ as thick ethical terms; I want to reserve ‘deficit’ for cases where the things neglected in attention *ought* to be attended to.

    Your second key question asks how we decide what work *attention* is doing vs. its inputs/ outputs. (I wonder if we add ‘vs. contemporaneous contextual features’ to this list.) And whether this is the same as asking which of these levels would need repair to fix the problem. As for the latter I’m inclined to say ‘no’. In part, I wonder if the question of ‘how should we fix the problem’ is capable of admitting a bunch of different answers depending on the parameters we set for it. Plausibly, some parameters might encourage us to say ‘systems of sexism or racism’ are ultimately what needs repair; if we take this to indicate where the moral problem *really* lies, does that mean a bunch of standard stuff we think of as morally problematic, like racist outlooks and behaviours, isn’t ‘the problem’? So, I think I’d need a clearer grip on what it means to ask ‘what needs repair?’. More generally, can’t one location be where badness resides, but it’s pragmatically-speaking more effective to fix that badness by intervening at a different location? Perhaps I’m going wrong thinking the question is meant in a pragmatic sense.

    As for the former question, I think it is profound and vital to answer (not that I do here, unfortunately), and I entirely agree that it impacts what philosophical methodology we use. I know I’ve tried to employ the abstracting/ idealising method in some of my replies, and I also very much see its pitfalls, some of which you mention. Reading your comment led to a bunch of relatively disparate literatures pop into my mind: intersectionality in social ontology, causation in the philosophy of science, and interactionism in the philosophy of biology. These all contain rich discussions of how to understand the (moral, causal, functional) contribution of a given feature while trying to avoid these issues of abstraction and idealisation. I wonder if they might shed light on the methodological issues and potential solutions here. I suppose an obvious alternative methodology examines a certain phenomenon across a wide range of real-world scenarios, arguably as Susanna begins to do in her precis. My worry is that interpreting the results of an analysis/ experiment employing that method would not be straightforward. As I’ve hinted at in earlier responses, I think that finding an attentional pattern which seemed to have a moral valence in one real-world scenario no longer have it in another, doesn’t definitively show that it never had a moral valence in the first.

  10. I really enjoyed reading this beautifully written and insightful paper, Ella. I could only skim the exciting discussion that’s already been posted so far — apologies if any of my comments repeat others’ reactions. I’ll be thinking/reading more about these issues and I’m so thankful for this opportunity to do so!

    1. What about the distribution of attention needs repair?


Susanna claims that you identify “the locus of wrongness [as] the distribution of attention”—not just attending to x, or not attending to x but rather, the wrongness is in the relative surplus or deficit. This interpretation of your view is consistent with your endorsement of the priority structure account of attention and with much of what you say in the paper. However, if the moral wrong you are identifying is about recognition (of someone as a person), as you claim, it seems to me that in all of these cases, the person attending will be failing to attend (adequately) to the particular property that would grant the target recognition. In your discussion of whether someone might manifest surplus attention on a property without manifesting an attentional deficit, you say that “Conversely, an individual might complain about surplus attention on the wrong things, without necessarily identifying other traits that she believes deserve more attention.” Whether the target can identify what property this would be (though in your cases, these properties are helpfully identified by the claimants of disrespect) seems like a separate, epistemic issue. Irrespective of whether the target knows what grants them recognition, we might think that the wronger wrongs by not attending to the property that awards the target recognition. I sense that you want to reject this explanation. What’s not clear to me is what we gain (theoretically) by insisting on the wrongness issuing from the theoretical surplus. 

    2. Convention and publicity:

    The second confusion I have is around the significance of a pattern of attention. In several points in the paper and in your discussion, you appeal to an analogy with speech act theory and the notion of illocutionary force. This is a bit puzzling to me because an utterance is a meaningful, public, linguistic representation whose force is fixed by social convention. Particular utterances function as “actions” against the backdrop of convention, bringing about normative changes. The institution of apartheid makes it so that the ‘Whites only’ sign is an act of subordination, just as the institution of marriage makes it so that saying ‘I do’ marries you to someone. Maybe I’m lost in the details here but I don’t really see how the analogy to speech acts helps your case. For me, it raises further questions about why we would think one’s private patterns of attention would wrong others if we expect the wronging to function like speech does: by being made public in the context of an already existing institution that associates a representation with a normative change. 

    3. Defining the relationship:

    Across the examples you use, we find individuals in particular contexts — in the art world, in the lab, at school — who tell us that they don’t want to be seen “as an x first”. 
My intuition is that being seen “as an x first’ might be denying not something as basic as respecting someone’s personhood in these cases but rather refusing to or failing to enter a relationship that can only be entered if you see the person “as a y first”. Take Basquiat’s case: I don’t think I can be a proper fan of Basquiat without seeing him as an artist first. Or take being a coworker of Esopi. It seems to me that I wouldn’t be a coworker of Esopi’s if I don’t see her as a scientist first. Note that these relationships — being a fan or being a coworker — come with attentional demands. 

If I’m understanding your view correctly, your diagnosis is that these individuals are being disrespected as people because they’re being seen as black or woman or blind first. This view seems to presuppose that there is a basic relationship that we are obligated to enter as moral equals whenever we interact, which is to relate to each other with respect as people. The testimonies show the failure of this, whereas in other contexts individuals can find that attending to these very same traits affirms their personhood such as the black CEO who wants to “foreground their agency” by being seen as black first (?). Why should we think that more attention to traits that give one power (and in the case of the CEO, autocratic power over others) is something that affirms one’s personhood whereas more attention to traits that put one at a structural disadvantage diminish their personhood? What kind of a relationship is “relating to someone is a person” and what’s its relation to power? I think I have a better grasp of what it means to relate to someone as a fan or coworker but I don’t have a clear understanding of the particular moral relationship you invoke in the paper. 

  11. Thank you, Ella, for the (characteristically) fantastic paper! I always enjoy engaging with your work and this paper is a particularly rich piece, which I’ll be thinking about (and teaching!) for some time.

    My thoughts about the paper concern a number of dissociable topics, so I’ll present them installments. My first set of questions builds on Susanna, Jessie, and Tom: what is the appropriate *object of evaluation* when we pay too much attention to social identity properties. I’m not sure how much new I’ll add, substance wise.
    But I will introduce some machinery that (I hope) makes the space of views a little clearer. Watzl thinks of attention not only in terms of prioritization, but as a process that *results* in prioritization. Specifically, attention is a personal level activity where some guiding state (say, salience or a goal) causes and causally sustains a priority structure. Attention therefore has three components: (1) a guiding state; (2) a priority structure; and (3) a guidance relation, whereby (1) causes and causally sustains (2). (I doubt that this model applies to all cases of attention: when your mind wanders, you have a priority structure––you focus on some things over others––but without the guidance relation. This will be relevant below, but for now we can focus for now on cases that involve all three elements.)

    In light of this model, consider someone whose implicit biases guide him to focus on his colleague’s face over her philosophical ideas. The former philosopher has (1) a guiding state (his biased salience landscape); (2) a priority structure (he prioritizes his colleague’s social identity properties over her philosophical properties); and (3) a guidance relation, whereby (1) causes and causally sustains (2).

    What’s the appropriate object of evaluation in cases like these? Distinguish three views.

    (1) Priority View: The appropriate object of evaluation is *a priority structure* (so 2 in our schema). You wrong someone if you view her as a woman first and a philosopher second, for example, just because you unduly prioritize her social identity properties.

    I think this is close to the view you defend. Or at least, you argue that the Priority View captures what’s wrong about some core cases. Here’s another view:

    (2) Guiding State View: The appropriate object of evaluation is *a guiding state* (so 1 in our schema). You wrong someone if you view her as a woman first and a philosopher second, for example, because this priority structure stems from the real harm: prejudice.

    There’s a way to read Susanna and Tom as supporting a Guiding State View. As Tom noted, “Patterns of attention reflect an agent’s values, emotions, beliefs, desires, attitudes, prejudices” and so on. All of those states (values, emotions, etc.) can be guiding states for attention. You might think it’s *those* states that are the real problem.

    Similarly, Susanna argues that “the real culprit…” in Ella’s cases is not “too much attention” but rather “the social readiness indicated by attention”. That social readiness, you might further argue, is constituted by prejudices that guide you to unduly prioritize social identity properties.

    There’s precident for views like this in the broader literature on responsibility for attitudes other than attention. Angie Smith argues, roughly, that passive attitudes like forgetting are attributable to us because of how those attitudes manifest our values. When we forget a friend’s birthday, for example, this might reflect that we don’t value them. This lack of value, she argues, is why it would be appropriate for our friend to be angry at us.

    I’m attracted to a third view:

    (1) Process View: The appropriate object of evaluation is the whole process, whereby *one state guides a priority structure* (so 1 through 3 in our schema). You wrong someone if you view her as a woman first and a philosopher second, for example, because your prejudice guides you to prioritize her identity characteristics. That whole process is the blameworthy mental action.

    One reason to like the Process View is due to an analogy with bodily action. Say I steal my roommate’s ice cream. What about my theft is an appropriate object of evaluation? It’s the harmful result, since you’d excuse me if I ate her ice cream by accident (e.g. I genuinely thought it was mine). It’s not just the harmful mental states, since merely intending to eat her ice cream is nowhere as bad as following through on my intention. Rather, it’s the whole action: intentionally eating my roommate’s ice cream.

    You could argue that attention is similar. It’s the whole mental action that’s bad: for example, *prejudice guiding you to prioritize social identity characteristics*.

    I think the Process View has advantages over the Guiding State View. Suppose that someone had prejudicial biases, but suppressed them from guiding attention (say, by exercising attentional control or using debiasing techniques to develop conflicting habits or just having good emotion regulation). I think that a lot better than giving prejudice the reins to attention. In an analogy with bodily action, it really matters whether the bad mental state brings about a harm (in this case, the harm is an attentional pattern.)

    I think the Process View may also have advantages over the Priority View. Susanna brought up one kind of case that might support the Process View (though see Ella’s responses).

    Another kind of case is where your mind idly wanders to someone’s social identity characteristics. Compare two people in a talk. One is guided to focus on the social identity characteristics of the speaker (say, their cool accent). He doesn’t obsess on the speaker’s accent, but the accent nonetheless feels relevant to him: he’s drawn to focus on the accent. Another’s mind idly wanders to the speaker’s accent a few times, but always wanders to something else without much ado. I think the guided agent does something worse, since guidance is normative. When you’re guided to do something, you implicitly apply a normative standard: in this case, a standard that says, “that speaker’s accent is relevant.” In contrast, idle wandering implies no such normative standard.

    BUT: Ella’s response to Susanna already contains a really nice case AGAINST the Process View. Here it is:

    “Imagine someone who is in a predominantly male Philosophy department in the UK, but who has read a good amount of relevant literature on dismantling the patriarchy, has a gender diverse friendship group, etc. Despite this, it remains the case that very few of their colleagues are women, very few philosophers in the media/ ‘conventional’ canon are women, etc. As I discussed in the paper, the simple statistical distribution of the Philosophy world around them means that they will therefore likely have attentional patterns that involve finding the gender of their women colleagues salient (thanks to our attentional patterns being so sensitive to signal-to-noise ratio).”

    I think Ella is 100% correct that (a) the department member would likely have prejudicial patterns of prioritization and (b) this is blameworthy. Yet I’m not totally convinced that this is independent of guiding states. Habits can form in spite of our genuinely held values, since habit-learning systems function partly to track statistical patterns in the environment on the basis of lots of data.

    But habits, I think, still guide us (here I’m following Alex Madva, Michael Brownstein, and Peter Railton on bodily habits and neuroscientist Rebecca Todd on the implicit guidance of attention). In spite of their best efforts, for example, a philosopher might find the identity characteristics of their female colleagues highly salient. But crucially, they don’t just idly wander to those characteristics once and a while. Rather, they are *guided* to her characteristics by salience: her characteristics *draw in* their attention, they *feel relevant*. As I said above, I think that this kind of implicit, habitual guidance of attention sets an internal normative standard for what feels relevant. Even if we reflectively disavow that normative standard, it’s still part of how we actually evaluate the world. And we’re culpable for the mental actions––guided attention––that results from that standard.

    OK that’s it on the three views! Two other things I’d like to get to tomorrow are (1) what to say about the badness of repeated attentional dispositions (building on what Jessie and Ella have already said) and (2) cases where we harm people from minority groups by NOT attending to their social identity characteristics. I think cases like (2) are surprisingly common.

  12. Hello Ella – congratulations on a great paper, followed up by a very high quality discussion.

    1/ As you noted in replying to Susanna, it seems many of our perceptual and attentional biases are very ‘low level’, in the sense that we have no reflective access to them. (An example I like a lot: cricket batsman will always say they keep their eye on the ball, but studies have repeatedly shown predictive saccades in which their eyes run ahead of the ball.) This is important for defending the claim that harmful attentional biases can persist even when all of a person’s accessible mental states are unproblematic. But do you agree that we are talking here of a kind of harm for which blame would be inappropriate? For example, we have a clear novelty bias – our attention will be drawn towards novel stimuli, including people with statistically rare properties. I agree this can be a kind of harm. But I am not sure what we can do about it – beyond trying to reduce the novelty. As long as a property continues to be rare, people will attend to it.

    2/ I’m interested in how you see the relationship between the harmful attentional biases you’re talking about and the learning biases that feature in the cultural evolution literature – especially prestige bias. Prestige bias may have an attentional basis a lot of the time. We attend more to people we see as successful influencers. (Social media obviously exacerbates this.) This seems like a harm to those who are ignored, but can it also be a harm to those who get the attention? They may think to themselves: people only attend to me because I’m perceived as influential. No one actually attends to me because of my qualities as a person.

  13. Thank you much, Elisabeth, for these fascinating and thoughtful comments. As someone who has been very influenced by your work on perspectives, I was very much hoping to hear your thoughts!

    Your first key comment suggests that we’ll likely need to bake more into moments of attention than I have done so far. For some background: one thing that led me to try to isolate the attentional pattern more cleanly from its background causes was imagining the subject of the attention, call them S, arguing with a person attending to them, call them R. Say S picks up on the attentional patterns of R which involve attending more to S’s demographic features than to their identity as a philosopher. R insists they weren’t attending a lot to the person’s race for a prejudicial reason; perhaps they’re interested in tracking race in their role as a demographic researcher (to refer back to a modified example of Tom’s). I was imagining the subject of the attention saying ‘OK let’s concede it wasn’t a bad perspective/ mental state guiding this; your attentional pattern still constitutes a way of disrespecting me relative to the standards appropriate to this context (their particular workplace context). Simply making my demographic properties salient in this context, whatever your reason for doing so, is not appropriate. Now, I realise that there’s a lot of detail that we’d still need to provide to fill out this example so it’s persuasive, and it’s open to different interpretations, but it might give some clarity regarding what made me go down this route in the first place.

    I guess you can see my motivation as the related to that driving some speech act theorists to provide a constitutive evaluation of certain utterances. One reason given for the latter was: showing how e.g. pornography constitutes subordination means you don’t now need to prove complex and contentious causal claims about it *causing* subordination (see e.g. McGowan). Similarly, I was hoping to offer a way in which the attentional pattern constitutes a problem without needing to prove complex causal claims going in the other direction—mental states causing those attentional patterns.

    I am reticent, then, to include too much in the picture of what’s constitutive of synchronic attention; I wanted to focus on the structuring of attention alone. However, I think your excellent suggestion about diagnosticity might need to be added for the reasons you give. I see this as going beyond attentional structure, and adding an evaluative/ interpretive assumptions regarding that structure. Namely, not only does the attentional structure prioritise x over y, but this reflects the attender’s interpretation of x as more important/ relevant than y. Perhaps that’s what is always implied by such a structuring. So, maybe I say that it’s the combination of the attentional structure + attender’s assumptions that the properties structured in the way they are reflect the relative importance/ relevance of those properties in a given context, that’s really going wrong. I wonder if this saves me from needing to add in further aspects of one’s background perspective (all which take me further away from focussing on the specific role of *attention* in these scenarios).

    You also ask if we can say both the “background perspective is what makes a synchronic attentional structure objectionable, and also that because that synchronic attentional structure occupies that functional role, it is constitutively rather than merely instrumentally objectionable.”. I suppose that if we see the attentional structure as *partially constituting* the (bad) background perspective, I think we get a constitutive evaluation of the attentional structure. This is where we can follow those writing about attentional patterns in the virtue theory literature, as mentioned in an earlier comment. Reading ahead, I wonder if Zach’s process view would give us something similar.

    Turning to your second comment about personhood-related traits. I thought you put this in such a helpful and illuminating way. It makes me wonder if one way to retain the emphasis on personhood is to suggest that, in the specific contexts in question, certain traits demonstrate personhood better than others. This is not to say that, contrary to how I phrased it in the piece, e.g. race is functioning as a non-personhood related trait in the contexts of a science lab, for instance. Race here *can* still showcase aspects of that individual’s personhood, here. And it’s relevant to how these e.g. scientists, artists, want to be understood. As you say, this is arguably reflected in how many of the testimonies (actually unlike Basiquat’s) indicate that the individuals still wanted to be seen as women, Black, disabled, in those contexts, so long as they are seen as scientists, artists, philosophers, *first*.

    To deal with this, I wonder if I could amend my diagnosis to this: *in these specific contexts* these demographic features aren’t showcasing personhood *as well as* the e.g. ‘scientist’ identity is. And that’s the source of the harm. This would ultimately be to concede that talking about personhood-related traits vs. non-personhood-related traits is too binary.

    Why, though, you ask, am I not just going for a simpler route: just go off a person’s wishes/ decisions about how they want to be attended to! In the piece I briefly mention this way of judging attentional patterns; perhaps we have a duty to respect the wishes of others regarding how they want to be attended (citing the social recognition literature, e.g. Axel Honneth). In fact, I do think it’s likely that the most important, overriding obligation we have is to respect the individual’s own decisions regarding how they’re attended.

    Why offer the disrespect-to-personhood diagnosis, then? Well, firstly, I wanted to explain why so many people’s desires and wishes were for a similar ordering. Why are so many of these testimonies pushing back against demographic features being made salient in those contexts? The disrespect-to-personhood diagnosis was there to offer a morally weighty reason that contextualises those wishes. Sometimes we need to explain why a chosen pattern of attention is so important to respect.

    Secondly, I imagine some people who are being attended to in a way I think is problematic, but they haven’t thought about their wishes yet, don’t have the conceptual (etc.) tools to do so, or perhaps have internalised some very damaging ideologies that involve leaning into being an object of fetishization for others. As for the latter, the idea is not to say ‘sorry, we’ve decided to override your wishes because actually that attentional pattern disrespects you’, not least because of the ironic disrespect to that individual’s personhood this would constitute. It’s more to pull out the tensions and complexities of such a situation, and show that even respecting someone’s wishes, which might be the all-things-considered right thing to do, may also mean hurting (etc.) them in some sense.

  14. Thank you very much, Ella, for your excellent piece of work. I am grateful for the opportunity to deepen my intellectual background in the topics you are presenting here. I also appreciate the insightful analysis from Susana, Tom, Elizabeth, Jessie, and the other contributors to this collective reflection. Although I am not a specialist in this field, I wonder what source of identity you are working with in this paper. A second piece of thought about this work comes from the relationship between moral wrongs and strugles for recognition. I will explain this last question bellow:

    I noticed that you quoted Axel Honneth’s book, ‘The Struggle for Recognition.’ In this book, the author aims to develop an empirically anchored theory of social conflict based on Hegel’s theory of recognition. Honneth argues for an intersubjective view of identity and a moral interpretation of social conflict. According to him, social struggles can be normatively evaluated based on their ability to provide the preconditions for self-realization through three distinct types of recognition: love, respect, and social esteem.

    Honneth’s perspective suggests that the struggle plays a pivotal role in the realization of every practical identity. The moral standpoint seems to offer a framework to assess the concrete disrespect suffered by disadvantaged groups. However, this experience is often tied to the specific structures of social conflicts explained by Honneth, namely, love, respect, and social esteem. As a result, there exists a strong relationship between moral and social theory.

    In your paper, you pointed out the following idea:

    ‘…we cannot experience self-realization, or the development of our capabilities unless others around us respect our identity. We might be able to expand on Honneth’s ideas, then, to demonstrate that one subtle way in which our chosen identities can be unjustly disrespected is by others not attending to that identity in the right way—namely, by failing to make our chosen identity our most salient one’ (p. 514).

    This quotation helped me understand your perspective on the specific advancement your paper offers in relation to Honneth’s thesis. However, I do have a concern that this criticism, in terms of the general consideration about disrespect against our identity, did not fully consider the threefold structure explained earlier. Have you taken into account this theoretical difference in the development of your ideas?

    Thank you once again.

  15. Thank you Ege for these wonderful questions, they’re so clear and interesting. And challenging! I really appreciate hearing your take on the paper.

    1. You’re absolutely right that the testimonies all involve a property that is receiving an attentional deficit. And that one way of explaining what’s going wrong here is to say that “the wronger wrongs by not attending to the property that awards the target recognition”. Let’s start by explaining why I want to suggest that some cases involve a surplus without a correlate deficit, and then say how that could still matter for diagnosing those original testimonies. I offered the taxonomy to provide a way of evaluating attentional patterns more generally, beyond being able to explain the specific phenomenon illuminated by the testimonies I considered. So, consider a case I didn’t discuss in the paper: a woman at a swimming pool might complain about surplus attention to her body from a stranger at the other side of the pool. They aren’t otherwise interacting. The swimmer is not wanting for something *else* about her to receive the attention instead, like her swimming ability, or conversational contributions. She’s just concerned about a surplus of attention on her body. As I say in a response to Jessie, of course there’s things the attender isn’t paying (as much) attention to, but unless these are morally important, this ‘lack’ of attention doesn’t become a ‘deficit’, for me (I treat ‘surplus’ and ‘deficit’ as thick evaluative terms). In the swimmer’s case, I find it hard to pick out something the attender *ought* to be attending to.

    Now, how does this all relate to the testimonies, when those individuals there *are* also pointing to attentional deficits, too? Because surpluses can come apart from deficits, I actually wonder if there’s two problems going on – a surplus on the wrong property which is problematic in its own right (indeed, it comes with its own phenomenology I think – consider the prickly, hot sensation that comes from feeling one’s being stared at – something like this), and a deficit on the right one (perhaps diagnosed in the way you suggest), which only adds to the problem. That might be the ‘more’ we get, theoretically, from focussing on the surplus. I realise I’d need to say more on this, but I’ll leave it there for now.

    2. Convention and publicity: you suggest that the analogy to speech act theory doesn’t work for private attentional patterns. Ultimately, you’re absolutely right to put pressure on this. I was thinking of attentional patterns leading to certain behavioural proxies that mean they’re no longer strictly ‘private’, but hoping to maintain that the attentional pattern is what has the normative link to the consequences it brings about. I wonder if this is an acceptable reading of what Robin Zheng and Nils-Hennes Stear do in their recent piece on mental imaginings (‘What’s Wrong with Blackface?’), insofar as they also deploy related tools in speech act theory.

    If this strategy doesn’t work, though I think I’d need to concede it’s just the behavioural proxies of the attentional patterns themselves that do this. If so I think Austin Baker’s fascinating work on ‘nonverbal marginalisaton’ speaks to this claim. Crucially this would mean losing the speech act theory defence of the constitution claim for *attentional patterns* themselves. In which case I’d rely on other arguments about constitution (e.g. partially constituting a vice, etc.).

    3. You ask what I mean by ‘relating to someone as a person’, as well as its relation to power. You suggest that a more intuitive relationship is something narrower, e.g. relating to someone as a co-worker. Take first the ‘relating to someone as a person’. I guess I’m thinking of something like ‘the participant stance’, invoked by Strawson to describe the ways we see and acknowledge others as persons and not things (something invoked across a range of discussions, including objectification and doxastic wronging). I think there’s a bunch of relationships we can have, as you suggest: to our co-worker, friend, lover, etc. These all involve different bundles of demands, including, as you say, attentional ones. A ‘lover’ or ‘friend’ relationship might make more demands on you than the ‘coworker’ relationship. However, I see something like the ‘participant stance’ as there to pick out a broad theme in what can go right/ wrong across a huge range of those relationships. Perhaps we could think of it as a general need to, in my case, attentionally prioritise personhood-related traits (those showcasing one’s agency, rationality, individuality, etc.). Crucially, this will look different, though, in different settings/ according to different relationships; what is a relevant personhood-related trait in one context/relationship may not be so in another. This might account for the sense of finer-grained relationship-specificness that you’re alluding to.

    You’re right to highlight how my example of the CEO perhaps isn’t very helpful for making it look like (autocratic!) power is important to agency. I’m also thinking of someone wanting to be seen as Black first when explaining their experiences relating to structural injustice; that individual’s Blackness in that case might be foregrounding their personhood. It is showcasing their understanding of their experiences/ of structures injustice (rationality), their ability to craft their own narrative and speak about those experiences (agency), etc. Admittedly a lot more would be needed to be said about this without the concept becoming too loose. I do have a lot of lingering concerns about the ‘personhood’ notion… So, I’ll finish by saying thank you for raising this so articulately and vividly – I’ll continue to think about it.

  16. Wow, this is such a fantastic paper, which I feel makes a really important contribution to the literature on the ethics of attention! Reflecting on the paper and others’ comments, I wanted to flag a couple points. [Also, many apologies for not chiming in earlier—I’m having to do a bit of unexpected traveling so have been in transit!].

    First, I found myself coming back to the question (brought up by other folks here): from a normative perspective, how should we understand the right amount of attention to be paid to various features of a person? Ella has presented us with cases of salience orderings which certainly feel to us to be unjust. But what would the right salience ordering look like in these cases (e.g., thinking about the black artist and woman philosopher cases in particular). Ella argues that morally problematic varieties of attention can hijack salience orderings resulting in deficits and surpluses of attention to the ‘wrong’ properties. As I understand her view, Ella is claiming that context is playing a significant role in determining the morally appropriate degree of attention various social features are be owed. However, there seems to be an assumption that in most cases social features like race and gender should *not* be attended to (because these features are irrelevant to the context of assessment and/or the attended-to agent’s personhood).

    While I agree with Ella that in the cases she presents, agents’ social features are receiving inappropriate attention surpluses, I’m not convinced these social features (e.g., race, gender, etc.) should be ignored altogether and/or should not feature prominently in the salience orderings. In fact, I’m not sure it’s possible to attend to someone as an artist or philosopher without at least minimally attending to aspects of their social identities (like their race and their gender). I found myself thinking of Linda Alcoff’s book Visible Identities, in which she argues (for standpoint theoretic-type reasons) that identities like race and gender fundamentally impact how one understands oneself and others (as I recall, she claims these identities end up constituting the self in virtue of playing this type of role). Thus, if Alcoff is right and social identities like race and gender shape are shaping agents’ epistemic standings in these fundamental ways, then we might consider social features like race and gender relevant for attending to someone’s identity as an artist or a philosopher. However, this won’t only be true for people who occupy marginalized identities. For example, it seems plausible to me that we should be more thoroughly considering identities like whiteness and maleness when we attend to white male philosophers and artists (as these identities are importantly shaping their work and how we engage with it). So perhaps, we shouldn’t be focusing on attending less to the social identities of marginalized groups but attending more to the social identities of nonmarginalized groups (which I think fits nicely within Ella’s attention surplus/deficit structure). So maybe all this is to say that I’m convinced that social identities like ‘philosopher’ and ‘artist’ are entirely separable from identities like gender and race.

    While I agree with Ella that in the cases she presents, agents’ social features are receiving inappropriate attention surpluses, I’m not convinced these social features (e.g., race, gender, etc.) should be ignored altogether and/or should not feature prominently in the salience orderings. In fact, I’m not sure it’s possible to attend to someone as an artist or philosopher without at least minimally attending to aspects of their social identities (like their race and their gender). I found myself thinking of Linda Alcoff’s book Visible Identities, in which she argues (for standpoint theoretic-type reasons) that identities like race and gender fundamentally impact how one understands oneself and others (as I recall, she claims these identities end up constituting the self in virtue of playing this type of role). Thus, if Alcoff is right and social identities like race and gender shape are shaping agents’ epistemic standings in these fundamental ways, then we might consider social features like race and gender relevant for attending to someone’s identity as an artist or a philosopher. However, this won’t only be true for people who occupy marginalized identities. For example, it seems plausible to me that we should be more thoroughly considering identities like whiteness and maleness when we attend to white male philosophers and artists (as these identities are importantly shaping their work and how we engage with it). So perhaps, we shouldn’t be focusing on attending less to the social identities of marginalized groups but attending more to the social identities of nonmarginalized groups (which I think fits nicely within Ella’s attention surplus/deficit structure). So maybe all this is to say that I’m convinced that social identities like ‘philosopher’ and ‘artist’ are entirely separable from identities like gender and race.

    Switching gears entirely to my second point, I was thinking the normative picture can be further strengthened by emphasizing certain features of the cognitive architecture of attention. According to the picture defended by Watzl (and Ella), attention is directing our limited cognitive resources to the things which are important to us – as represented by orderings within our salience structures. I was thinking that the normative significance of these orderings becomes clearer when one considers the cognitive limitations at play in attention (e.g., working memory capacity limits, speed/accuracy tradeoffs, etc.). Since you cannot attend to everything, the act of attention involves negotiating various tradeoffs (e.g., attending to one thing leaves fewer resources to cognitively process another). Perhaps this framing around resource scarcity provides a response to the question of harm in cases where the person doesn’t know their identity (say, their gender in the female philosopher case) is being excessively attended to, but where we’d still want to say that they are nonetheless being harmed. I know that Ella explains the harm in these cases by reference to personhood, but I think it’s also plausible to frame the harm in terms of monopolization of limited processing resources. Prioritizing the female philosopher’s gender in your attentive behaviors, involves directing limited cognitive resources to their gender over other features that could be attended to, like their philosophical ability. Hence, emphasizing the resource scarcity implicit within the cognitive architecture of attention seems to make the normative case all that much stronger — we cannot attend to everything so we really need an ethical framework to understand our allocation of attention.

    (An aside: though this is on a slightly different point, I’m also not perhaps as convinced by the personhood arguments because – thinking back to Alcoff – I think in most cases features like gender and race will be personhood related… But, even if these features are personhood related, we can still say that attending to the female philosopher’s gender monopolizes our limited cognitive resources, making it more difficult to focus on features of them that are more relevant to the context of assessment — like their identity as a philosopher! So, maybe this is a way to make Ella’s point that isn’t resting on any particular conception of personhood? Not sure, just thinking aloud!)

    Again, thank you Ella for such a wonderful paper and thanks PEA Soup for this fantastic discussion!

  17. Thank you Zac for the lovely kind comments and for the fascinating suggestions/ discussion points, which I’m finding so helpful. I’m also really looking forward to your further questions today. Because of that, this reply therefore will be more brief.

    The machinery you introduce is hugely helpful, and articulated super clearly. Thank you for taking the time to spell it out this way! One key question I have is about how broadly you’re understanding the guiding state. I ask this in part to figure out where Jessie’s view fits. You discuss how sometimes the guiding state might be standard values, attitudes, emotions, etc., hence as you say capturing some of the phenomena described by Susanna and Tom. At one point you introduce the guiding state as a ‘biased salience landscape’. I wonder if we can see this as capturing the ‘network or set’ of dispositions (to find salient/accessible/ to attend) that I take Jessie to be describing. Unlike standard values, attitudes, emotions, etc., this is more of an attentional phenomenon already. It’s just that it’s homing in on the *dispositions*/ habits side of attention.

    If so, what makes this salience landscape/ attentional habits *biased*? Jessie gives one answer to this, of course, which preserves the sense in which certain of these dispositions can be prejudiced in and of themselves. But I wonder if critics will suggest we can’t stop at the network of attentional dispositions – that we need to go back until we find a more conventional belief, judgment, outlook, intention, etc., that will let us judge between different cases of habit/ disposition-driven attentional patterns to find a person’s e.g. demographic features particularly salient. I suppose I’m thinking again of another example I mentioned earlier, of a Black person navigating a white institution/ country, who finds the Blackness of others salient, but only because they are searching for those who will truly understand their experience as minoritised in the institution, etc. (Incidentally, I’ve just remembered that I think I need to thank Lidal Dror for first presenting this particular example to me). Maybe this person’s habits have set an ‘implicit normative standard for what feels relevant’ – namely, demographic features of Black people – but presumably this standard is morally acceptable given the *role* it is playing for this individual. I still hope I can deal with such cases by appealing to features of the context as opposed to a prior mental state, but I realise a) I need to say more about this and b) this won’t persuade everyone. What is your view?

    Anyway that’s not particularly clear – my apologies! – but I’ll stop there to wait for your questions – thank you again Zac!

  18. Thank you so much Jonathan! I really appreciate these kind words and excellent questions.

    1. You talk about many of our attentional patterns being such that we have no reflective access to them (love the cricket example – topical!). You ask if we can therefore be blamed for such attentional patterns. A great question. The first thing to say is that I don’t think most of these attentional patterns are low-level in the sense that we have no reflective access to them, and/or in the sense that we cannot change them. Many might be unnoticeable and uncontrollable in a given moment, but as per Zac’s recent highlighting of attentional habits, we might have (a degree of) long-range control over them. So, we can grant that lots of our implicit biases are inherited from the environment around us. And that, as you’ve rightly picked up on in my response to Susanna, we might be especially *bad* at noticing the attentional biases we inherit. But that’s consistent with us being able to *learn* to notice those patterns, and to change them.

    I believe we’d really need to be guided by what psychologists are finding are successful routes for achieving this, though. Jenny Saul (e.g. ‘Scepticism and Implicit Bias’) and Louise Antony (‘Bias: Friend or Foe?’), for instance, discuss some studies looking at overcoming implicit associations, e.g. between certain types of faces and fear. According to those studies (admittedly these are pre-replication crisis and I need to check what’s happened since…), one does best mechanically to repeat certain intentions ‘e.g. ‘when I see faces like x, I will think ‘safe’. Perhaps similar mechanical repetitions are required for changing attentional patterns. There’s already huge cultural traditions – and industries… – out there that offer guidance on how to change attentional patterns; Buddhist meditation practices, CBT, mindfulness, etc. (Sifting through this for good quality evidence-based advice might be tricky given that the industry-side of this currently answers to a huge range of questionable capitalist demands, of course).

    Finally, as you point out, even if it turns out that we have no control over some of our attentional dispositions, this doesn’t mean they can’t be morally criticisable. In those cases, we might just decide not to assign blame.

    2. You also ask about how attentional biases relate to the sorts of learning biases that feature in the cultural evolution literature, particularly prestige bias. This is a v interesting question. The harm this does to those who are ignored makes me think of Leonie Smith and Alfred Archer’s work on ‘epistemic attention deficits’ (in ‘Epistemic Injustice and the Attention Economy’), which focus on deficits not to individuals’ properties but the individuals (or social groups) themselves. They articulate beautifully the sort of harm/ injustice that can occur in such cases. But you’re interested in what happens to the person who gets all that attention. There’s different ways of analysing this, I think. We might say that there is still a problem with the prioritisation of others’ attention on this individual, if we think the wrong things are being prioritised over the right things. But two things come to mind. Firstly, it’s not clear exactly what metric ‘success’ counts as the wrong thing to focus on. Secondly, even if we can say the attentional pattern is morally problematic, we might still want to stress that this issue is outweighed by the various benefits that can accrue to this individual through this attentional pattern on them.

  19. Hi all. Unfortunately, I need to pause responding to this blog. Because of the time of the day here in the UK, there’s a chance I might not get to respond to the *wonderful* comments from Ignacio and Austin (and Zac’s impending ones, among others) before the blog officially closes tonight. I will ask the organisers of the blog if they’re able to post some of my replies tomorrow on my behalf.

    For now, I want to say a HUGE thank you, both to the wonderful commentators, and to the excellent organisers at PEAsoup for organising and facilitating the discussion. Who knew that getting a bunch of your philosophical heroes together to carefully reflect on a paper you’ve written would be such a rewarding experience…? And also challenging, in case that wasn’t abundantly clear, ha! I will meticulously be going through all of the comments long after this blog closes, mining them for their many gems. I am so grateful.


  20. Ella, this is an excellent paper! I want to focus on Susanna’s worries about the nature of the harm.

    Susanna argues that, “the problematic thing seems to attach to presumptions, feelings, and other reactions to the relationship between these properties, not to the quantity of attention to any one of them. Such reactions vary, even when attention to the properties stays the same.” So, the argument seems to be something like this:

    1. A has attitudes X (the presumptions, feelings, etc to the properties)
    2. A has attentional patterns Y (attending in certain ways to those properties)
    3. X causes Y
    4. So, the wrongness of X subsumes the wrongness of Y.
    5. So, patterns of attention are the wrong focus.
    In other words, having gotten to one of the precursors of the pattern of attention (the attitudes), the wrongness of those attitudes screens off wrongness of the pattern of attention.

    In some cases, I think this screening effect does happen. For example, suppose I’m an inattentive driver. One afternoon, while looking at my phone, I crash into another car and cause some (let’s say minor) injury to the other driver. In this case, I think it’s right that the wrongness of the inattention screens off the wrongness of hitting. But, that’s because it doesn’t seem like there’s much room between those two: having been looking down at my phone while driving, it just doesn’t seem like there’s much I could have done in between my inattention and the accident. The inattention is a proximate cause without much room for intervention on my part. So, hitting the other car isn’t a *separate* wronging—the wrongness of the inattention subsumes the wrongness of hitting the car.

    But, that’s not the case for patterns of attention. One of the things that makes attention such an important site of normative analysis is that we ***do*** have a measure of conscious control over it. Right now, you’re choosing to direct your attention to this sentence rather than the dog begging to go to the park or the text message that just pinged on your phone. Of course, we don’t have total control over how we attend and being aware of our patterns of attention is difficult. Nevertheless, we *can* wrest conscious attention away from our wandering minds. (My “Rumination and Wronging” piece on attention in doxastic wronging focuses on this (https://philpapers.org/rec/SAIRAW).)

    Given that we can intervene in this way and we know, as Whitely points out, that others care about how they are perceived by us, we can make conscious efforts to see people as they want to be seen—even where we have attitudes that dispose us to do otherwise. This is especially important where people make their preferred identities known. Robin Dembroff and I discuss these cases in our “Yep, I’m gay: Understanding Agential Identity” (https://philpapers.org/rec/DEMYIG). There, we argue that there is a prima facie obligation to allow and respect others’ self-determination of their social identities. The construct we define there, agential identity, is the means by which people make that self-determination known. Almost all of Ella’s examples make use of these very direct statements of agential identity (e.g. “I am not a Black Artist. I am an artist.”), and this suggests to me that one way to understand Ella’s article is as pointing out that patterns of attention are a significant element of the uptake of those agential identities. This is especially so on Watzl’s priority structure account of attention, because the gap between what *is* socially important and what we *treat as* socially important is narrow.

    Given this, I think there actually ***is*** an invariant (though subject-sensitive) way to think of the role of attention in Whitely’s paper, which is this: the **match** between the pattern of attention and the subject’s preferred pattern of attention. (Modulo Watzl’s account of attention: between the priority structure others actually exhibit toward the subject and the priority structures the subject would prefer.)

    This match itself is an appropriate target for moral analysis because it’s something that we can act on. (Not to say this is a necessary condition, but maybe a sufficient one.) It will be influenced by the attitudes that Susanna points to, and we won’t be able to match others’ agential identities perfectly, but it’s something that we can intentionally care for, ignore, or violate. This, I think, is a very good explanation for what’s going on when transphobic people pick apart the appearances of trans* people and argue that they don’t pass. It’s not just that they have the background attitudes—it’s that they’re directing their own and others’ attention to things that emphasize and reinforce those attitudes. They are drawing attention aspects of someone’s appearance that they ***know*** work against their preferred social identities and doing so *because* of the resulting attentional mismatch (e.g., if A weren’t a trans woman, B wouldn’t focus their attention on A’s stubble; if A weren’t nonbinary, B wouldn’t notice or focus on the fact that they’re dressed in a stereotypically feminine manner).

  21. (Oi, it looks like the PEA Soup blog doesn’t carry over formatting! The double- and triple- asterisks are just italics and bolding from copy & pasting out of Notion, haha.)

  22. Hi Ella,
    this is an amazing paper! It was extremely rich and I found myself pausing after every other paragraph, thinking about further applications, generalizations, extensions, and so on. I just read that you might not have time to respond to further comments, so I’ll keep my thoughts super short—maybe you’ll still have time to skim them, and if not, we can just talk about them more outside the blog!

    The main thought I had when I read your great paper just expands on a point that has been made by several commentators: the question of whether what’s morally problematic in the relevant cases isn’t an outgrowth of attention or salience, or rather something else. Many candidates have been proposed in this discussion, but one I haven’t yet seen, I think, is that what’s in fact problematic is that we attribute causal relevance to a property in a way that’s off. Steven Sloman, a cognitive scientist that has made important contributions to causal cognition, often points out that our attention is often drawn to properties that we take to have causal relevance, and, similarly, that salience is often shaped by assumptions we make about features’ causal importance (see, e.g., his 2009 book, or his response to Del Pinal’s and Spaulding’s paper on bias and causal centrality: https://www.dropbox.com/s/lkqui009cuc6l85/Sloman.pdf?dl=0). If that’s the case, we would have another plausible candidate of a mental state that seems to be independently morally problematic, but that can explain, at the same time, why we feel that attention to certain aspects of social identity etc. is off, insofar as attention is a reliable proxy for morally problematic attributions of causal relevance. Mental states of causal-relevance attributions seem, in some cases, morally problematic insofar as they seem highly constraining of someone’s personhood, as happens in extreme cases of causal-relevance attributions of social identities, such as in cases of psychological essentialism. In these very strong cases, someone’s properties, prospects, are seen as a causal outgrowth, and therefore strongly constrained, by some intrinsic essence. Of course, even less strong cases, in which we don’t fully essentialize someone, but in which we still see someone as strongly causally constrained by this or that social attribute, can still be morally problematic. This sort of explanation would fit in nicely with your intuition that in the relevant cases of wrongful attention, we disrespect the subject in question. It also would dovetail nicely with your realization that whether or not attention is problematic will ultimately depend on context. If we accept that attention is a reliable proxy, but not itself what’s problematic, it is possible to override the feeling that something off is going on if we have more information—e.g., if we know that the reason for attention is *not* that someone sees a feature as causally constraining in a way that’s problematic. So my question would be what you think about this kind of picture. I don’t really see it as a competitor, but rather a nice way to complement your important observation that we often feel something is off when a property is given (relatively) too much attention, and would also dovetail nicely with the growing research of how attention and causal attributions interact. (I realize this isn’t really a novel comment, but as I said, an extension of the thoughts many others had in this thread!)

  23. … I was also curious about what you think is the mechanism is through which attention results in disrespect, or in virtue of what it is that attention constitutes disrespect. I know you have more work on this, so maybe this question is already answered in your other work! (Or maybe you already addressed this in your answers to the other comments! Sorry if anything I’m saying is just repeating others’ stuff!) For example, many feminist philosophers have given analyses of the concrete aspects of pornography that `satisfy’ some criterion for objectification and disrespect: e.g., that it treats or depicts women as a means to an end. In the parts of the paper discussing wrongful attention, it seemed that what you think amounts to disrespect is the fact that we attend more to properties over which the target of attention doesn’t have control, instead of on properties over which people do have control. But as such, this doesn’t yet obviously seem to fall under any of the common ways in which we usually flesh out lack of respect (as treating someone as means to an end, failing to see someone as an individual, etc.). Or maybe it does! So I just wondered what you think the specific ‘disrespect criterion’ is that failing to pay attention to the right properties satisfies. Many of the commentators have flagged the possibility that the aspect that amounts to disrespect is that we fail to respect someone’s right for self-determination of how they *want* to be seen and attended to. I find this pretty plausible. Would you agree with this diagnosis of (one of the) ways in which failing to attend to the right properties in the appropriate way would amount to disrespect? Importantly, if we accept this kind of analysis, some interesting questions open up for social navigation in our practical lives. After all, in many cases, we won’t (and can’t) know exactly how someone wants to be attended to: what about their identity they find most important, how they want to see in certain contexts but not others, etc. In some cases we can—when people explicitly tell us, or when they make it obvious in other ways. In these cases, when we ignore these signals—either through neglect or willfully—it seems obvious that we are epistemically and morally blameworthy. What should we say about the other cases?

  24. Hi all, we’ll finish off this thread with some final comments from Ella, thanks everyone for participating.
    — — —
    I’m back! For one, slightly disappointing encore. Time reasons mean I will inadequately address the remaining wonderful comments. But let it be known that I will be returning to them a lot after this blog officially closes.

    Ignacio: Thank you Ignacio for your kind words. So pleased to hear you’ve been enjoying the blog, and I’m grateful for your excellent questions.

    You ask first what notion of identity I am using. I am trying to use a broad, intuitive notion of identity properties, which leaves it open as to what more precisely they consist in. For the purpose of the paper ‘identity’ is simply there to point to something like a categorisation that describes us in a socially-relevant way. For a more substantive theory of social identity, I really like Linda Alcoff’s work (e.g. ‘Visible Identities’). But, as Austin points out in a comment below, to more fully integrate Alcoff’s theory into my framework would require considering things like how our social identities shape each other to become new intersectional categories. For simplicity’s sake I’ve treated identities as separable traits in this paper, but of course even for illustrative purposes this might obscure things too much / mislead the inquiry. (I say a little more on this in my reply to Austin)

    You then ask specifically whether I had considered how the disrespect-to-identity diagnosis fits with Honneth’s tripartite structure focussing on love, respect and self-esteem. Unfortunately I have not considered this in much depth! I knew that it would require quite a bit of work to explain how attention fits into the different spheres of recognition Honneth describes, and the correlate levels of relationship an individual can have with those spheres. I didn’t want to get into that analysis in this paper. But I felt it was important to flag the social recognition literature as a fruitful avenue to consider on the road to getting to the diagnosis I *did* want to focus on: disrespect-to-personhood. I also mentioned Christine Korsgaard’s notion of ‘practical identities’ in the social recognition vein, but again I have not explored this in depth. I would very much like to turn to this soon, though, so your comment provides a great kick start to that project, as well as some guidance regarding the potential issues that might arise. So, thank you! Incidentally, I’d also like to think of social recognition issues in relation to how Cat Saint-Croix and Robin Dembroff think about agents’ self-determination of their identities (fortunately there’s already a link to that paper in a comment below).

    Austin: Thank you so much Austin for such warm comments, and for your stimulating discussion points.

    1. Am I implying that race/gender should *not* be attended to because they’re irrelevant to the context of assessment/ are functioning in that context as non-personhood-related traits? Should we ignore them?

    Great and important question. One thing to say is that I’m realising I probably ought to amend my gloss of e.g. race/ gender functioning as ‘non-personhood-related traits’ in the e.g. workplace contexts described in the testimonies. I’ll refer back to my response to Elisabeth here. There I talk about changing my account to say that e.g. gender/race showcase personhood *less* than the property of ‘scientist’ in the contexts in question, as opposed to *not at all*. As you and Elisabeth are teasing out, that would also better reflect the testimonies; many said they weren’t seeking *not* to be seen as e.g. women; just not *first and foremost* as women. So, great point!

    1a. You suggest that identities like ‘philosopher’ and ‘artist’ aren’t entirely separable from identities like ‘gender’ and ‘race’.

    This seems like a very important point to bring up, and I’m particularly pleased you brought up Alcoff as I love their work on identity. So, part of the reason I treated these identities as separable was to reflect what I saw in the testimonies; ‘first x and second y’ to me spoke to at least some separation. Now, I would have thought there’s scope for attention on a singular intersectional category like ‘woman philosopher’. But I also take it that there’s a great many complaints in the real world that e.g. women have, insofar as they feel wrongly categorised in terms of this intersectional category. (‘I’m not a woman philosopher! I’m just a philosopher!’) We could read this as a plea for more separation. I suppose there’s a possible distinction here to be drawn between metaphysically how identity works – plausibly everything is complexly intersectional meaning that metaphysically-speaking we cannot prise apart identity categories in the ways I’ve implied we can – and a more political/ ethical/ practical question about how we ought to train our attentional patterns in certain contexts—perhaps politically/pragmatically etc. we sometimes can and should introduce some separation between them.

    I love your point about how perhaps we ought instead be paying more attention to the demographic features of privileged folk – that’s a great suggestion I will be reflecting on more!

    2. You suggest that another way of framing the problem in the cases I mention is monopolisation of scarce processing resources. I really like this. To me though I wonder if it needs a further account to get to the ethical/ moral significance of this resource allocation. What makes certain allocations worse than others? What is it about diverting resources to gender instead of philosophical ability that’s *bad*? To me, that’s where something like disrespect-to-personhood can step in to fill what I see as a gap. (Particularly if we’re after a *constitutive* critique of that attentional pattern that goes beyond the instrumental costs of this differential resource allocation.)

    2a. You wonder about gender/race in fact being personhood-related traits in many contexts. I do agree with this, which is why I suggest that what counts as a personhood-related trait will change with context; sometimes e.g. gender will be functioning as a personhood-related trait. But your point further emphasises to me how I need to amend my discussion of gender/race/ disability *in the contexts described in the testimonies*, in the way that I flag in my reply to your first point. Basically, I wonder if we better spell out the *moral* dimension to the attentional distribution by highlighting how these demographic traits, in these particular contexts, are functioning to showcase these individuals’ personhood *less* than these individuals’ identities as e.g. a *scientist*. While I’m persuaded it will be part of diagnosing the problem in the testimonies, I’m currently not convinced that the diagnosis ‘the trait is not relevant to the context of assessment’ will necessarily capture the *moral weight* of the complaints. I’ll need to think about this more…

    Cat: Thank you so much Cat for such a fascinating comment! Also I am *here* for someone else figuring out a defence of the constitutive claim, ha. I find it very persuasive, and not just because it’s so helpful to me…. Even if we need to rely on a prior outlook (etc.) to tell us when a given attentional pattern is wrong (something I’ve tried to push back against in my replies above), that doesn’t mean the prior outlook *subsumes* the wrongness of the attentional pattern. And that part of the reason for this is the control we have over our attentional patterns.

    One thing I’m still trying to grapple with is this: I take Susanna to be suggesting that it’s the prior state (e.g. the outlook) that gives us a *principled way to judge between which of various identical attentional patterns are morally problematic vs. ones that aren’t*. I wonder if that gives her an extra kind of ammunition against finding the issue to lie in the attentional pattern itself.

    As I mentioned to Ignacio, I’d like very much to start thinking more about the moral diagnosis of attentional problems that couches things in terms of the agent’s self-determination of their social identities. I’ve tried my best in various comments on this blog to justify why I focussed on the more ‘objective’ moral grounds of disrespect-to-personhood in the paper, but I also realise that the self-determination diagnosis is critical (as Liz and others also draw out in their comments). Your and Dembroff’s piece on agential identity is such a wonderful road into that project. I really love how you’ve described this route in terms of the match (or lack thereof) between a subject’s preferred attentional pattern and the one actually exhibited towards them by others. This feels totally right to me.

    P.s. gonna start triple asterisking everything now

    Elli: I’m so grateful, Elli, for your hugely kind words and for your fascinating discussion points.

    1. You wonder if the right diagnosis for the testimonies isn’t an attentional surplus/ deficit but ‘attributing causal relevance to a property in a way that is off’. The way that you describe this sounds very plausible to me. It makes me think of what Elisabeth in her comment above called ‘diagnosticity’, namely, “how relevant it is for informing us about other features and categories” (a part of what Amos Tversky calls ‘salience’, the other being ‘intensity’, capturing the signal-to-noise ratio idea). I’m not sure if this notion of diagnosticity implies that the relevance has to be *causal* or whether it can be instead *explanatory*. I wonder if you have thoughts on this.

    Critically, I did start to concede in my comment to Liz that diagnosticity might be something I need to build into what counts as making something salient in one’s attention. So you’ve picked up on something that I’m really starting to think might be more important to my account than I’d initially thought. A key question I have at this stage is this: does having x being more salient in attention than y essentially mean treating x as more important or relevant (causally, explanatorily))? Does every salience structure imply correlate beliefs (?) about the causal/ explanatory relevance of the properties it structures over? The bottom line is I’m still wondering if I can maintain the idea that a salience structure can be morally problematic simply in virtue of its structuring, even notwithstanding further assumptions (/beliefs?) about the respective properties’ causal relevance.

    2. You ask about the mechanism through which I see attention as resulting in disrespect. This is a great question and I apologise for such a brief and unsatisfying answer (I’ve run out of time, frustratingly): So, I see a range of things constituting disrespect, e.g. ignoring a person’s desires, believing them to lack rationality, etc. These are all things that the literature on objectification has covered as constituting ways of disrespecting personhood/ treating a person as a thing. I’m trying to add the attentional patterns I describe in the piece to that list. One way of failing adequately to see someone as an individual with rationality, integrity, agency etc. is selectively to prioritise traits that do not showcase their rationality, integrity, agency etc., over traits that do.

    You also ask a very important question about others’ self-determination to be attended to in the way in which they wish. I’ve said a few things (particularly in response to Liz) about this that might answer some of your questions here, to contextualise why I wanted to focus on a less desire-centred account.

Comments are closed.