Welcome to the new PEA Soup Ethics discussion! This post focuses on Robin Zheng’s and Nils-Hennes Stear’s paper ‘Imagining in Oppressive Contexts, or, What’s Wrong with Blackface?’. The paper is available through open access here.  Paul C. Taylor has kindly provided a critical précis to kick off the discussion:

I’m grateful for the chance to introduce this piece. Zheng and Stear undertake a careful and illuminating study of a serious and interesting question, and they do it in edifying and engaging and, ultimately, persuasive ways. Their work leaves me with some questions, but I’ll inquire against a backdrop of broad and enthusiastic agreement.

The authors ask when and why an exercise of the imagination is intrinsically wrong – wrong, that is, not because of its harmful consequences, but wrong in itself. This is a serious question in part, as the authors note, because the prospect of holding people accountable for their imaginings has become, for many of us, a matter of some controversy. (The authors mention controversies over “hipster racism”; I thought of debates about “cancel culture.”)

On one side of this discussion is a worry, as the authors put it, about “oversensitive, sanctimonious” types wrongly condemning off-color jokes (for example) that play with but do not endorse worrisome attitudes. On another side is the worry that simply deploying these worrisome attitudes is an ethical error, active endorsement aside.

The authors think both sides in this debate – the advocates, as they say, for imaginative laxity and imaginative strictness, respectively – have a point. Unfortunately, neither strictness nor laxity leads to the right conclusions, for want of a coherent account of ethically flawed imaginings. The authors use the phenomenon of blacking up – of altering the appearance of people not-racialized-as-black so that they appear to be black – as a testing ground for their escape from this impasse.

The US-based performance tradition of blackface minstrelsy, rooted in slavery-era depictions of backward “darkies,” is perhaps the most famous example of blacking up. But blacking up can involve not just visible alterations using makeup, wigs, and the like, but also vocal performances (say, in voice-over work) and other activities by which someone might imaginatively adopt a black persona. Blackface is a helpful on-ramp to the topic of unethical imaginings because it is widely regarded as deeply problematic even as it remains a weirdly prominent feature of worldwide racial practice.

The authors cite several recent examples of blackface and could have cited many more, from all over the world. The one I find particularly illuminating concerns a white law professor at the University of Oregon who (badly) darkened her face and donned an (absurd) afro wig for a Halloween party. (The parenthetical interruptions are necessary. I’ll explain.) She did this, she said, to present herself as a character in an anti-racist book that she wanted to celebrate. In the furor that predictably followed, no one disputed the professor’s anti-racist intentions. But the furor followed, nevertheless, rooted in worries about things other than the professor’s endorsed racial attitudes.

The Oregon case instructively draws out the relevant intuitions. One can understand the thought, even if one is not strongly drawn to it, that it should count in the professor’s favor, somehow, that she did not endorse racist attitudes. On the other hand, one can also understand the thought that something went wrong at that party, no matter what she intended.

Strictness holds that simply deploying the imagination in ethically problematic ways – in this case, play-acting that is rooted in centuries of anti-black stereotypes – is an ethical error. On this approach, the professor’s imagining was ethically flawed, her intentions notwithstanding. Laxity, by contrast, holds that simple deployment is not enough: that endorsement is a necessary condition for ethical error. On this approach, since the professor did not, as the authors say, “present the imagined attitude as suitable for ‘export’ into the actual world,” she cannot be ethically blameworthy simply for blacking up.

The solution

Zheng and Stear find a middle ground between these extremes by borrowing from speech act theory, from some of its recent applications in philosophical race theory and feminist philosophy, and from complementary developments in sociological race theory. They argue that imaginings are sufficiently like speech to make it worth considering their locutionary and illocutionary force, and that this approach is particularly important in societies shaped by histories of oppression. This approach focuses attention on the contexts for our imaginings and on the social realities they constitute, thereby expanding the field of evaluative maneuver beyond strictness and laxity.

Here’s how the authors describe their middle ground:

“We agree with the lax that imaginings may be intrinsically unethical in virtue of the attitudes they endorse. Yet, we disagree that they must endorse an attitude to suffer such a flaw…. We agree with the strict… that fictively deploying unethical attitudes can generate intrinsic ethical flaws. Yet, we disagree that it must…. We argue that imaginings… exhibit a kind of intrinsic ethical flaw in virtue of… fictively deploying attitudes that realize oppression, normalize oppression, or license oppressive behaviour. ” (p.387)

An imagining realizes oppression in the sense that it expresses or instantiates certain meanings that an oppressive social context makes available to it. This is the locutionary dimension of the speech-act approach: the law professor can’t unilaterally decide what her imaginative act means because meaning doesn’t work like that. Blacking up is an act that has had its meanings baked into it by centuries of anti-black practice. One can’t just choose how one stands with respect to all that, any more than one can just decide that “dog” denotes cats.

In addition to realizing oppression, the law professor’s imagining has the illocutionary force of normalizing oppression. The law professor, as Patricia Hill Collins inspires the authors to say, provides some of the “ideological glue” that links ideas about certain kinds of people and what they deserve to the practices that (mis)allocate social goods. As the authors say: “by imagining a ‘ghetto thug’… one contributes to making black men be seen as dangerous and violent” (p.399). To imagine in this way just is to normalize this way of reading black male bodies, even if one’s imagining has no causal impact on other people.

Finally, the illocutionary force of these imaginings also involves licensing oppression. The idea here is that oppressive imaginings set the normative conditions for conversational exchange. The law professor made a move in the game of social exchange that invited, licensed, additional moves in the same spirit, additional moves that would also help constitute (reproduce, reinscribe) the very system of oppression that her putatively anti-racist performance meant to contest.

The punchline, then: laxity rightly declines to assign blame to just any imaginative deployment of problematic ethical attitudes, as strictness demands. But strictness rightly refuses to wait for the agent to endorse problematic attitudes, as laxity demands. The truth is somewhere in the middle, next to the realization that the attempt to examine the imaginative act by itself, in isolation, is what blocks the path to a resolution. Speech act theory reminds us to study social contexts – to see what our imaginings can plausibly claim to be about and to do – while critical social theory reminds us to look closely at specific social contexts to see how their histories motivate ethical criticisms.

I have of course sped past several vital details in attempting to summarize this fascinating argument. There’s no space here to examine those details, or to consider the additional questions to which they might lead. I’ll close instead with some questions less about how this argument is supposed to go than about how it might go further or differently, questions that I could not answer by looking more closely at the details.

  1. In making the point about locutionary force, the authors rely primarily on the fact of semantic constraint. One can’t, they say, simply decide to use “dog” to denote cats. This is why the law professor can’t simply declare that her performance of blackface is in fact anti-racist. But one can decide to buck semantic convention in this way, and one can live with this decision if one is willing to accept a degree of illegibility to one’s fellows. I wonder if this possibility matters in oppressive contexts, especially in divided pluralistic societies, some members of which will wear their illegibility to others as a badge of honor, while treating the illegibility of others to themselves as evidence of perfidy or viciousness. I wonder, put differently, if the law professor’s willingness, if she were willing, to bite the bullet on her illegibility might signal a relationship to oppressive meanings, or to the prospect of meaning change, especially in divided societies, that the authors might want to account for.
  2. I wonder, in addition, if semantic constraint is less important for this account than semantic deference. Most attempts at making meaning involve writing checks against a vast store of semantic resources that we don’t personally command. We manipulate symbols whose meanings we can’t fully cash out, and we assume that some Putnam-style semantic expert can cash them out for us. This leaves us to some degree at the mercy of our semantic resources, making meaning on credit, as it were. All of which is to say that the Oregon law professor may have simply meant more than she intended, because she chose a meaning-bearer that carried with it all the content with which it had been funded by centuries of anti-black ideology. I wonder why Zheng and Stear don’t address this aspect of their topic.
  3. Finally, I’m drawn to the law professor case because it highlights the importance of quintessentially aesthetic considerations like criticism and performance. This was the point of my parenthetical additions earlier: it should matter to the way one reads the case that the law professor imagined herself as black in a very particular way. The photo that accompanies the story (at the link in the authors’ notes) shows the professor’s slapdash makeup job and ridiculous wig, both of which are consistent with the aspect of the blackface tradition that depicts black people as disheveled and messy. This matters because blacking up is a performance tradition, and reading instances of it requires critical acumen. On a generous read of the case, one could say the law professor lost her semantic bearings because she lacked this acumen and couldn’t read the relevant racial meanings. This matters because a core feature of contemporary US racial politics involves a concerted effort to discourage and prevent people from developing the acumen for properly reading racial and racist performances (by, for example, banning responsible pedagogy concerning US racial history). I wonder how this aspect of the situation bears on the argument that Zheng and Stear have developed.


16 Replies to “Robin Zheng & Nils-Hennes Stear: “Imagining in Oppressive Contexts, or, What’s Wrong with Blackface?” Précis by Paul C. Taylor

  1. I really enjoyed this paper, and the above is an excellent critical precis to kick off discussion.\

    I had a very, very basic question, though. The authors claim that they want to show that some imaginings are wrong regardless of harmful consequences (or the risk of harm). But I kept coming back to the thought that by identifying the relationship to *oppression* as the key feature of wrongful imaginings, the authors could not escape harm or the risk of harm – or at least making things bad for others – as central to the wrong.

    After all, what is wrong with oppression? Put simply, it is bad for the oppressed. Attempts to explain the wrongfulness or badness of oppression that do not focus on the lives of the oppressed, the ways in which they are treated, and the ways in which that is bad for them or makes their lives worse than they otherwise would be in some important respect, would surely miss the point.

    So what is wrong with realising, normalising, licensing oppression? If oppression is wrong because of the way it harms its victims, aren’t all these things wrong because of their contributions to that system of harm? That they at least *risk* furthering or sustaining oppression? Is that wrong? If so, how?

  2. Hey Patrick, thanks for the compliment and the question. Robin may have other things to say but I think what we should say is that the thrust of your point is right. There may be any number of things that are bad about oppression and there may even be isolated benefits, for want of a better word (e.g. positive stereotypes), but chief among the bad-making things is the way oppression harms the oppressed. And if this is a necessary feature of oppressive systems, and it seems to be, then the wrong of engaging in the sorts of imaginings we describe depends on harm in at least one way. It’s an important subtlety your comment brings out nicely.

    We did not mean to deny this point, which seems right. Rather, we meant only to deny that realizing, normalizing, and licensing oppression must themselves be harms, or causes of harms (or potential harms) in order to be morally criticizable. Even if an imagining fails to be or cause a harm, it may be still be ripe for moral criticism because of the way it constitutes part of (or, if you like, contributes constitutively to) an oppressive system—namely, in the ways we describe. Our idea is just that the imagining itself need not be, or causally contribute to, a harm to be morally criticizable.

  3. *content warning: discussion of racist art*

    Hi all, I enjoyed this paper immensely, and am grateful for the opportunity to comment on it. Many thanks to Paul C Taylor too for a really helpful precis, and for the stimulating points of discussion.

    I wanted to ask a few things, the last couple relating to the issue of resisting controlling images and the complexity of black-face critique.

    1. Does your account assume a particular concept of uptake from speech act theory? Maybe something like the more radical constitution theory, where our words can have an illocutionary force that we didn’t intend (or even foresee)? In speech, a more conservative approach (what McDonald 2021 calls ‘ratification’) is adopted, where uptake can determine the success of our speech act, but not its content. I wonder if the constitution approach seems more apt for cases of art and representative imaginings – while similar to speech they are more complicated compared to ordinary conversations.

    Regarding your points about resisting controlling images, I found the analogy with speech really helpful, namely, that much like I can negate a sentiment in speech, this too can be done with satirical imaginings which ‘deploy’ controlling images. I have a few queries about this.

    2. As you show, intention is somewhat beside the point. But does the *identity* of the imaginer matter when determining the illocutionary force of their imagining? If we extend the analogy with speech, slurs for instance can be appropriated or reclaimed when used by the originally oppressed group. In the same way, the identity of the imaginer may be a way to facilitate the negation of the unethical attitude (for instance if the creator of a representation that deploys black-face is black themselves). This may go some way to explaining why Brett Bailey’s ‘Exhibit B’ (2013) installation backfired. If we can understand his performance of the 19th century colonial human zoo as containing controlling images, despite his intention to reveal and critique contemporary racism, the fact he is a white man ended up re-enacting the very power dynamic he was trying to critique. So, perhaps the very identity of the imaginer, like in speech, can affect the illocutionary force of their imaginings.

    3. The problem with the above suggestion though, is that even in cases where the identity of the imaginer could in theory suffice to negate the attitude they are fictively presenting, the controlling image might still ‘leak’ out oppressive content/force. A case I’ve been working on recently is Makode Linde’s ‘Painful Cake’ (2012), where the artist uses the image of a g*lliwog to represent himself as a black woman undergoing FGM as people cut into her. The shocking piece was made by an Afro-Swedish man, but it has divided critics as to whether or not the piece is racist. Taylor’s point about particular aesthetic considerations of the imagining is relevant here. I wonder if the Linde cake as an imagining/representation might be *just too awful*? I’ve recently been working on how controlling images might behave similarly to slurs, which is relevant here: in the same way that the sting of some slurs is not insulated or inoculated by operators like negation or belief reports, some imaginings – especially when manifest in more ‘permanent’ works of art (i.e. not a fleeting performance or attendance at a Halloween party) – may also license oppression, simply because of the extremity of the piece itself.

  4. Thanks to Robin and Nils-Hennes for a fantastic and thought-provoking paper. Thanks also to Paul for starting us off. I so loved reading this article. When I entered the profession (decades ago), I never would have *imagined* that a “tip-top” journal would publish a paper on this topic. In those days, at least in the analytic circles of my training, anything applied – anything relevant even – was regarded as non-serious. It brings me joy to see some of the ways that our profession has changed for the better.

    I suspect that you won’t be surprised to hear that I mainly agree with what you say and how you support it. In addition to praise, though, I have some questions.

    First, I am not entirely confident that I am clear on the difference between realizing and normalizing. Here, I will wear my metaphysician’s hat. Please bear with me. Social constructions ontologically depend on mental states (i.e., a complex form of collective recognition). Money is money only because we regard it as such. This means that mental states partially constitute (that is, make it the case that) social constructions obtain. So, when an individual imagining instantiates a controlling image, that imagining partially constitutes the collection of mental states (partially) constituting the oppressive system. (I say that the collection of mental states partially constitutes the system because other stuff (e.g., actions) is required. In this way, an individual imagining realizes oppression. In cases where enough other people also realize that controlling image, it is normalized. When that’s true, an individual instantiation partially constitutes its normalization so it is also normalizes oppression. So, normalizing and realizing both involve the same thing (namely, the instantiation of a controlling image). Is that correct? I ask because I am now wondering how distinct they really are. Whether an individual instantiation realizes or normalizes depends on features of the context. An individual instantiation of a controlling image realizes so long as the social context is oppressive and that controlling image purports to justify that oppression. I am wondering whether that can that be the case without that controlling image being normalized? If not, then every case of realization is also a case of normalization. Is that correct and are you ok with that?

    Second, I am wondering whether you think that purely private imaginings can license. Suppose, for example, that a fantasy involves a controlling image and we stipulate that the fantasy in no way manifests outwardly either at the time of the fantasy or any time later. As you argue, this can realize and normalize oppression but can it license? As you characterize it, licensing involves enacting norms via contributing to (norm-governed) social activities. Do you think that a purely private imagining can do so? In other words, would it count as a contribution to a *social* activity? I think this is needed in order to trigger the norms in the broader social context and I am inclined to think that purely private imaginings do not count as social contributions.

    Third, as you say, negating can backfire. This is something that I worry a lot about. I think that these systems are really pernicious and can be especially difficult to undermine or dismantle. One of the things you rightly stress is that what our imaginings do (realize, normalize, license) depends on the broader context. Since that context is not under our control and since we might not realize what the context is like and how it affects what our imaginings do, what our imaginings actually do outstrips both our intentions and our awareness. The same, it seems, goes for attempted negations. Maybe controlling images work a bit like slurs. If I were to say, “I am not a wop; I’m Irish,” my utterance would be offensive. Even though I am negating, I am using the term ‘wop’, which might validate its aptness as a concept, or involve the adoption a problematic perspective, or convey problematic attitudes towards those of Italian ancestry. Attempts to negate controlling images might backfire similarly by, for example, keeping those images in circulation or validating them conceptually or pragmatically. Attempts to undermine implicit associations can backfire similarly. If I were to say, “Hey, women are not by nature submissive,” my utterance might nevertheless reinforce the very association between women and submissiveness it aims to undermine. Are you able to say any more about how to prevent such backfiring? In asking, I realize that this is a huge topic that you have already carefully flagged but I am nevertheless interested in your thoughts.

    One of the jarring things this kind of work stresses (and I include my own work here) is that we are all involved in these systems. The oppressive norms govern our social spaces and – so long as we participate in social interactions – we are contributing to these systems. And, what our individual contributions do depends on things outside of our control and outside of our awareness. We cannot opt out. Doing nothing is a contribution too. This is daunting. What you have here shown is that this is also true of our imaginings – even the purely private ones. It’s not just our social actions but also our private thoughts that are here shown to (sometimes) be part and parcel of oppression. This result seems at once both shocking and obvious. It’s shocking because it’s screams of thought police; it’s also kind of obvious given how socialization gets in all our heads. It’s jolting all the same.

    Thanks again for such a terrific paper. I look forward to reading your responses and learning even more.

  5. Thanks to Robin and Nils for an insightful and interesting paper and to Paul for this wonderful precis.

    I wanted to raise a question about a use of the imagination that, I think, might have similar effects to the one’s you describe, but does not have the same profile. I’m thinking in particular about what we might call resistant imaginings, as illustrated in the controversy over Halle Bailey playing Ariel in The Little Mermaid live-action film recently released. It seems to me rejections of “a Black mermaid” or Santa Claus, Jesus, etc. also play a role in normalizing certain portrayals of social groups in actual life. Rather than deploying things like “ghetto chick” or “ghetto thug” to contribute to making Black people appear dangerous, violent, or inferior, resistant imaginings deploy a companion content that, in the Little Mermaid case, reaffirms white dominance. So, I suppose my question is about the place of these type of resistant imaginings in your account. Do you think they can normalize oppression? Are they explained by your analysis of things like blackface?

  6. Fantastic paper and precis! I have two totally nerdy philosophy of language questions…

    1. You write that: “Realizing oppression concerns the content of an imagining and corresponds,
    on our analogy, to the locutionary dimension of speech; imaginings realize oppression by partially constituting an oppressive system.” In the Langton/Hornsby/McGowan framework, the locutionary is the content (as with yours), but the constituting of oppression is taken to be an illocutionary act. Since locutionary is just the meaning of the words, for them, it’s hard to see how it could constitute oppression (even partially). I would have expected the locutionary analogy for you to be e.g. Justin Trudeau, with face painted brown. Then the illocutionary would be the oppressive bit. Can you explain a bit more how your story works?

    2. Normalizing oppression and licensing oppressive behaviour are the illocutionary acts for you. They are about making oppression seem justified and activating norms that permit it. As political philosophers of language have generally understood it, uptake is required for acts like these. So I started wondering about whether blackface would still be oppressive if people failed to recognize it as blackface—if someone had an excellent makeup artist do a great job, for example. Would the lack of uptake keep this from being oppressive, on your view?

    OK, now one that’s not philosophy of language so much: I’m wondering about the relationship between blackface and controlling images. Is the idea that any instance of blackface in a culture where blackface has the sort of history you describe is thereby an instance of a controlling image? Or could a white person in such a culture blacken their face without there by being an instance of a controlling image– maybe everyone including understudies falls ill, and a white person has to play Katherine Johnson in a stage production of Hidden Figures. (Would this maybe not be blackface?)

  7. Hey Daisy,

    Thanks for your kind words and your incisive questions. Some responses:

    To (1): since our analogy between imaginings and speech is no more than an analogy, I don’t think we’d want to commit to any particular position on the role uptake plays in determining illocutionary force. But, just as you suggested, if we were to push the analogy to its limit, we’d need something like the first theory you mention since, as others have noted, the meanings of imaginings can outstrip the intentions, and thus the communicative intentions, of their participants. Certainly this is how we think about the meanings of imaginings.

    On (2), yes, I think Robin and I are both committed (perhaps not on the paper, but in actuality) to the idea that the identity of the imaginings’ author or other participants matters. It can be part of the contextual material that helps settle the imaginings’ meanings and what is done with them. Brett Bailey’s work (caveat: I have only read about, but never seen, it) seems a good example of this. But it doesn’t by itself settle whether, for instance, a work negates or fails to negate a controlling image.

    Does this mean that a White person, say, could successfully criticize controlling images by blacking up? I think that, for all we have argued in the paper, this is a theoretical possibility, with one important qualification. Such an imagining will be prone to stepping on other ethical mines than the one we devote the paper to examining. Some of these we touch on in the paper, relating, for instance, to causing unnecessary and foreseeable offence. However, it seems likely that there will be other sorts of relevant ethical considerations directly related to author/participant identity that we don’t get into in the paper. I’m thinking here broadly in the space of concerns raised in the literatures on cultural appropriation and, as you alluded to, on the reclamation of slurs.

    To (3), I think there are a couple of things that might happen. The first, which is perhaps not relevant to the sorts of case your considering, is that a member of an oppressed group G tries to negate controlling images of G in an imagining and just botches it, reinscribing the images instead. This is a possibility one would expect if, as I said in response to (2), identity factors in but doesn’t settle the meaning of the imagining. But the second, which I think is your real interest, is whether an imagining can be just so odious or extreme that there is just no edifying way of handling it.

    Again, there are at least two ways this might go. The first is that the imagining itself—the content—is odious and extreme. The second is that something about how the imagining is realized—for instance the props used—is the culprit. Your worry definitely applies to the latter case. To take a frivolous but clear example, someone might literally murder someone to invite an imagining criticizing, say, the deadly inaction on climate change. This type of imagining is irredeemable because, uh, murder. Could there be imaginative content that is similarly irredeemable? It seems likely, at least in a given time and place, if not forever and everywhere, though our argument doesn’t need the answer to come out any particular way. I’d be interested to hear what you and others think about this.

  8. Very interesting paper and discussion. Sorry if this is a bit badly thought through – a bit of time pressure.

    I had two thoughts/ questions

    1) I’m sympathetic to the idea that making moves in a normative system can be oppressive without that move being harmful. In fact, I’d be happy with the stronger view that this does not depend on the system being harmful. Consider a society with laws of slave ownership where X sells a slave, S, to Y and S then becomes Y’s legal property. I think that the acts of the buyer and seller are oppressive in this case even if the trade is not harmful. And I think that this is because these are constitutive moves in an oppressive system – S being sold to Y is constituted by X and Y exercising their normative powers. And I think that the oppression of the slave owning legal system does not depend on it being harmful (though, of course, it will be harmful in fact). A legal system that treats people as property is, I think, oppressive just because people are treated as property, and not only because treating them in this way harms them. It is worth noting that this is so even though buying and selling, in this case, are not morally effective – S is not the property of either X or Y morally speaking. So if you are right that blackface involves a similar kind of move, I think that you can get your conclusion without relying on the idea that making a move in a harmful system can be wrong because the system is harmful, even where the move is not harmful.


    2) I have doubts about the view that constitutive moves in a system like this can be complete without intentions. The strong case where we do things with words involves the exercise of normative powers (consent, authority, promise, request, and so on), where one person directly alters the rights, duties or reasons of another directly. And I think that direct alteration requires an intention to alter. Things that are taken conventionally to alter normative material is, I think, insufficient. Here’s a bit of evidence. Suppose that I invite you into my house, convention treats sitting on my sofa as part of the invitation, but I am strongly opposed to your sitting on my sofa, and I intend you not to do that. I just don’t understand the convention. I haven’t permitted you to sit on my sofa. My intentions are decisive; what other people take to be my intentions, what convention suggests, how people understand me, and so on, are not. And we can see this from the fact that if you realised what my intentions were, and that I didn’t understand the convention, it would be wrong of you to sit on my sofa.

    Licensing, I think, can be understood in a stronger or weaker form. The stronger form involves the exercise of a normative power – a bit like permitting – but I think that requires intentions. The weaker form involves causing or encouraging other people to believe that something is permissible. This doesn’t require an intention, but licensing in this weaker sense also doesn’t constitute a move in a normative system of the right kind for your argument to go through (I think a similar thing about Rae Langton’s views about authority). Licensing in the weaker sense could, of course, still be wrong, and perhaps its wrongness doesn’t depend on its effects (is creating the impression that something wrong is permitted wrong even where no one is affected by the conduct? I think it might be). But I doubt that this is the kind of case where oppression is constituted in the way you have in mind. I wonder if you had any more argument why this view is wrong?

  9. Great discussion! Re: the Dixon and MacGowan comments: co-sign. Sadly, all I have time for right now…

  10. Hi Mary Kate. I’m delighted you enjoyed the paper, not least because your own work was so critical to helping us write it. And I second your pleasure in seeing the field take applied issues seriously. Thanks for your questions and comments.

    On your first question, do normalizing and realizing involve the same thing (namely, the instantiating of a controlling image)? Yes. They are, in the context of our discussion, two distinguishable aspects of the same phenomenon. This is, though treat this analogy with some caution, something like the way that a broken promise to another is both a violation of trust and an offense against the moral law. There’s one crime, as it were, but two different ways of accounting for its wrongness. The two aspects may come apart in other contexts not involving the imagination, just as some broken promises might in some contexts violate trust but not the moral law, or vice versa. But I confess I haven’t given that much thought. Again, Robin might have other ideas.

    On the second, I think the easy answer would be to say that purely private imaginings don’t licence, since this involves the social mechanisms you mention. If that’s right, then such imaginings’ intrinsic unethical potential would be restricted to normalizing and realizing. This would be a subtle but important qualification of the account and a small concession to those wary of moralizing purely private mental states.

    The more daring answer would be to say that there is a kind of reflexive licensing that happens in a society of one, as it were. I am more partial to the daring answer because I think one can set norms for oneself by doing things, including imagining them. Also because, you know, being daring is way cooler. But I don’t have any well worked out arguments for this preference.

    On the third point, I agree that there is a danger that even the most successful attempts to negate controlling images might have a certain amount of backfiring built into them. This is a more generalized version of the worry Daisy raised, I think.

    If the backfiring just were realization and normalization of the controlling image, then we would have to accept that controlling images cannot be challenged head on, and our account, indeed all accounts, would have to accommodate this constraint. However, even supposing there were a certain level of inevitable backfiring, the mechanism might be different to realization and normalization. For instance, just speculatively, it could be that one can successfully negate a controlling image at the level of meaning, but that there are unfortunate empirical effects of such a negation. This might be because of the way that human beings are cognitively wired. Perhaps to understand the negated imagining exposes one to the risk of developing certain prejudices, in something like the way that a brewer tasked with identifying batches with a spoiled flavour might, by repeated exposure, develop a perverse taste for it. If this were how things turned out, then the good news would be that we could challenge controlling images head on and not be forced to merely instantiate them. The bad news, however, would be that, in virtue of negating them, we would be bringing about some other pernicious effect—and perhaps one not so practically different from the wrong of realizing and normalizing a controlling image.

  11. Hi Luvell,

    Thanks for the excellent question. I think the answer will depend on what the resistant imaginings amount to. I can see at least two possibilities. I’ll stick with your example of the live action Little Mermaid.

    The first is a case where, watching the film, the imaginer experiences imaginative resistance in one of the senses from the literature on that topic: they imagine everything else prescribed by the film but refuse, because unable or unwilling (either way, because they are too dense), to imagine Ariel as Black. Instead they imagine that she’s White, ignoring the general convention of actor-character race-matching in film.

    The second case is one where there is a refusal to go and see the film at all.

    I think the first case is a good candidate for our account. To use that prop to engage in that imagining in this political context is to realize a controlling image constituting part of a White supremacist ideology.

    Comparable cases involving Jesus and Santa are even clearer; there is an added layer of resistance built in. A case might be made that the original Ariel is White (NB. I am fat from a Hans Christian Andersen scholar), despite being, uh, of a different mythological species. But whatever weight that has, which is probably not much, the historical Jesus and St. Nicholas would not be considered White today. (I don’t think it’s widely known, but the historical Santa came from what is now Turkey and his complexion would certainly have been much darker than is typically depicted. The less said about the ubiquitous blond Jesus, the better).

    The second case is trickier since it doesn’t involve an imagining. It seems oppressive as a refusal but not for the reasons we identify. That said, I can imagine a plausible account suggested by your question that mirrors ours and draws on the same resources. According to this account, certain refusals to engage in imaginings that subvert controlling images (I think the live action Little Mermaid counts among these) are oppressive or acquiesce to oppression or something similar.

    In fact, this looks like a really good case where normalization and realization come apart (cf. Mary Kate McGowan’s question). One who refuses to imagine Ariel as Black in certain contexts doesn’t realize oppression—there’s no controlling image conjured—but they plausibly normalize it in virtue of their refusal.

  12. Please note that this discussion has now closed. As Nils and Robin are keen to respond to the comments they haven’t been able to engage with yet, the comment section remains open for a few days to give them more time to do so. We kindly ask you to refrain from adding new questions and comments. Many thanks to all who have contributed to this great discussion. 

  13. Hi Jenny,

    Thank you for the kind words. Others should know that Jenny was instrumental in getting me to start thinking about the ethics of the imagination in 2006 when I took her excellent course on feminism at the University of Sheffield. So, partial credit (or blame) must go to her. Apologies for the length of my replies.

    In response to (1): I’m not sure I can add much to what’s already in the paper. Part of the difficulty is that the analogy is intended more as a rhetorical crutch than an integral part of the argument’s anatomy; the mapping from imagining to speech act is always approximate at best. That said, maybe this helps. Rather than thinking of Justin Trudeau with his face painted brown as corresponding to locution, it might be better to think of it as corresponding to the utterance. That the imagining Trudeau’s costume invites us to engage with deploys particular attitudes towards certain people of colour isn’t settled by the mere donning of the costume. Rather, it’s settled by doing so within a certain tradition of uses of such costumes, as well as Trudeau’s intentions, aesthetic considerations about how the costume has been executed, the micro-context in which it is donned, and so on and so forth. True, all of these things can be relevant to settling the “illocutionary force” of an imagining, as with an utterance. But they can also settle the locutionary content, by disambiguating homophones in speech, for instance.

    The guiding thought is that we take the imagining’s content to form part of the hegemonic domain. Since this is a contribution one can point out independently of what the imagining does, so to speak, we are at the level of locution rather than illocution. But, for reasons that Mary Kate touched on, one cannot cleave the former quite so neatly from the latter.

    Maybe one way to capture your intuition that the locution couldn’t be the oppressive bit would be to say that realizing oppression, in our sense, isn’t itself oppressive (contrary to our claim), even if normalizing oppression is. I think that view would run into problems, but it’s an option in logical space and not a totally batty one.

    On (2): on one reading, your question gets into the difficult issue of how social meanings get set. On the one hand is the thought that, if no-one on Earth understands practice x as having meaning m any longer, then this just means that x doesn’t have meaning m. On the other hand, if Orwell’s 1984 has a meaning, then it has this meaning for all eternity, regardless of how it is later understood. Perhaps the first intuition pulls hardest when we think of imagining types (e.g., blackface) and the second when we think of imagining tokens (e.g., that time when Justin Trudeau donned blackface).
    But on a different reading of your question, which I think you intended, we encounter a couple of other complications. So, take a case like the one you gave where, perhaps because the execution is so masterful, the token imagining (person A blacking up) is not recognized as belonging to the oppressive type (blackface). This is a really interesting case (Rachel Dolezal springs to mind, though how her adopted identity intersects with the imagination is probably pretty complicated, so leave that to one side). A couple of thoughts.

    One, we deliberately put our account of social meanings, at least of individual imaginings, in a normative rather than causal register to accommodate the second intuition above (the Orwell one). As such, the simple thing to say is: yes, A is engaged in blackface, whether or not others recognize it as such and as such realizes and, in the right context, licenses oppressive behaviour. It’s true that, ex hypothesi, no-one (ignoring A herself) reads A as engaged in blackface. But the point is that they ought to read her actions in that way, since she is in fact in blackface. A’s act of blacking up is a little like the ‘Whites Only’ sign hung ineptly in the window. Perhaps no-one sees it, but we think of it as normalizing and licensing discriminatory practices all the same. No uptake required.

    The more complicated thing to say is that cases like these could push against the conceptual boundaries of the problematic practice. Setting aside blackface, consider Anjelah Johnson’s “Vietnamese Nail Salon” comedy routine. In it she impersonates the workers of the salon, strong Vietnamese accents and all. Johnson’s adopted accent is, from what I gather, nigh on perfect. I’m tempted by the thought that, because of this, what Johnson is doing is different in kind to what, for instance, Rosie O’Donnell did on ‘The View’ in 2006, when she imitated Chinese speech by saying “Ching”, “Chong”, and Chang” in various rapid, stomach-turning combinations. What O’Donnell did is yellowface, is the thought, while what Johnson did falls outside the category in virtue of its excellence. Alternatively, we might want to say that both are forms of yellowface and morally criticizable, but O’Donnell’s imagining is just morally (not to mention aesthetically) worse.

    On the third question. As I mentioned in a previous answer, and subject to the caveats there, I think it’s in principle possible for a White person to engage in blackface without reinforcing a controlling image—namely, in virtue of successfully negating it.

    The understudy case admits of at least two possibilities. Case one, the person playing Katherine Johnson is White and plays her without adopting stereotypical Black mannerisms or costume. This doesn’t look like blackface to me at all any more than if, in re-enacting an anecdote about Paul Robeson, say, I simply repeat something he said. It might be criticizable for other reasons (excluding Black actors, say), although the situation as you described it looks like it’s absolving on that score. In case two, the person does adopt the face-paint, costume, mannerisms, etc. Then this is blackface; there’s no getting around the significance of darkening the face in such a case in our present historical moment. As we mention in the paper, it’s a testament to the recognition of what we’re arguing for that performances of Othello appear to decreasingly look like case two and increasingly like one, even where White actors play the leading part.

  14. Hi Victor,

    Thank you kindly for the suggestions and criticisms.

    I’m happy to take the friendly proposal for not committing ourselves to the dependency of normalizing etc. on harm. I suppose whether the proposal work—whether, for instance, a benign legal system in which one can own people is ipso facto harmful—will depend on how broadly one understands ‘harm’. Do the legal restrictions count as setbacks to interests induced by a wrong, to borrow Feinberg’s analysis? I think I could be persuaded either way. If they do, then it looks like we’re back where we started. If they don’t, or if some other better analysis of harm gives us the less committed result, great!

    On the issue of intention being necessary for setting social norms. I agree that we want licensing of the stronger sort. I also agree that the sort of case you describe with the sofa is one where intentions set permissions. I don’t think, however, that the case is representative of how social norms get set generally. That is, there will be cases where intentions are decisive and others where they aren’t. The case is one where someone exercises use rights and has a certain authority to set the terms. But many cases aren’t like this.

    In fact, even cases where someone has a privileged say about what the norms are need not hinge on intentions in this way. Suppose, for instance, that I write a novel in which I intend the protagonist to be an admirable hero. Sadly, due to my psychological obtuseness and inept writing, my protagonist is just a complete tool. Now, I might reproach readers who settle on this stipulatively best interpretation because I didn’t give them permission to imagine the protagonist as an absolute tool. Or I may concede that, fine, people can imagine whatever they want, but only because they are not engaging with work properly. They’re merely playing what Kendall Walton calls an ‘unauthorized game’ with the novel. After all, I intended the protagonist to be super cool. Yet, my reproach should fall on deaf ears. Even as the author, I just don’t get to settle all the content and meaning facts of my work (this view is held almost unanimously by professional aestheticians, for what it’s worth).

    Since norms for interpreting a novel are paradigmatic social norms, I put it to you that this is the sort of case you’re looking for.

    So, artworks’ meanings can outstrip their authors’ intentions. This means that imaginings’ meanings can too, since many such artworks just are imaginings in our sense. But even cases more like the one you introduced can enact permissibility facts independently of intentions. Suppose the invitation to visit were jointly authored. Perhaps the board members of the local rotary club invite me to their clubhouse. And suppose that some board members intended the invitation to include the couch, others to exclude it, and still others had no couch-directed intentions. Then, barring some other overriding consideration, whether I am permitted to sit on the couch, I think, will be settled by convention, typical scripts for invitations, etc. I might not sit on the couch to spare the weirder board members’ feelings, but I’m definitely permitted to sit on the couch.

  15. Last but not certainly not least, a huge thank you to Paul for the wonderful précis of the paper. It manages repeatedly to say what we tried to but much more elegantly. And thank you for your questions, Paul.

    I’ll take the first two questions together. If I’ve understood you right, I think semantic deference is much closer to what we had in mind when writing the paper. The point about not being able to use ‘dog’ to mean cat was not meant to deny that we can stipulatively use symbols in atypical ways. Surely, we can take any symbol and chart a new course with it. The point was rather that we can’t stipulatively scrape off the semiotic barnacles that have accrued on its hull—not without some work anyway.

    This is probably another case where the analogy with language should be treated with caution, even though we lean on it repeatedly. The meaning that most words have is entirely arbitrary and conventional. But the meaning that imaginings and imaginative practices have is generally much richer and less straightforwardly conventional. All things equal, I think this makes stipulative counter-moves in the former more promising than the latter.

    As for accommodating a plurality of practice meanings in virtue of people, such as the law professor, deliberately bucking semantic trends, I’m not sure what to say. I’m not totally confident I fully understand the proposal. I’d like to hear more, basically.

    If the idea is that people or sub-communities can have a sort of dialect or idiolect for what practices mean, this seems to be not only possible but actually the case (take flying the stars and bars in the southern US to signify regional pride, for instance). If the thought is that this plurality of meanings can cleanse such sub-communities’ practices from oppressive meanings, it will depend on the case. A Tibetan Buddhist exhibiting a swastika, for instance (to use a case suggested to me by Ken Walton) might do so faultlessly. Someone flying the stars and bars in their front garden, however, has not escaped the racist meaning of that act, however they and their community understand it. Obviously, these aren’t cases that involve the imagination but I assume it’s clear how the thought would carry over.

    On the third comment: hard agree. Something that I think aesthetics can bring to topics like the one we tackle is a sensitivity to the material that informs interpretation. It’s for these sorts of reasons that I suggested above that the pitch perfect impersonation of a Vietnamese (or other) accent reads very differently to a sloppy and hackneyed one, however we ought to judge either morally. You are absolutely spot on in noting the sloppiness of the Oregon law professor’s costume. It makes the ethical problems attending what she did so much plainer to see for all the reasons you state.

    One last thought to address your point concerning the worrying rise in (or at least prominence of) political opposition to teaching race critically in the US. I suspect this will have the sad effect of making oppressive imaginings of a certain type more ubiquitous in the US. To the extent that these developments increase non-culpable ignorance, many of those participating in these imaginings will, in the worst way, become less blameworthy too.

  16. Thanks to everyone who has participated here; it’s been challenging but great fun. I think I speak for both Robin and I in saying how flattered we are to be in such illustrious company. I hope I managed to address your questions and perhaps settle some doubts. Feel free to follow up with Robin or I by email if you’d like to continue the conversation.

    Thanks too to Carline and everyone at PEA Soup for their work in making this happen.

Comments are closed.