ResearchI work on a variety of issues in moral and political philosophy as well as philosophy of language. I'm especially interested in questions relating to responsibility for the public meaning of our actions.
My work on statistical evidence and profiling was recently featured in the podcast Hi-Phi Nation. Listen to the podcast (here) or read the full paper (here) if you're interested! a quick clarificatory note on citation:I'm using "Bolinger" as the last name, "Jorgensen" as middle name, following the convention invoked by Delia Graff Fara (who notes that "Judith Jarvis Thomson"/"Elizabeth Cady Stanton"/"Hillary Rodham Clinton"/"Ruth Barcan Marcus" invoke the same convention).
|
Vertical Divider
|
renee.bolinger@gmail.com
renee.bolinger@anu.edu.au (note: just one 'L' in `Bolinger') |
published work |
forthcoming. Strictly Speaking
Analysis. Co-authored with Alexander Sandgren. (Philpapers link)
Analysis. Co-authored with Alexander Sandgren. (Philpapers link)
- A type of argument occasionally made in metaethics, epistemology and philosophy of science notes that, since most ordinary uses of a given term do not satisfy the strictest interpretation of the term, most ordinary assertions involving it are false. This requires there to be a presumption in favour of a strict interpretation of expressions that admit of interpretations at difference levels of strictness. We argue that this presumption is unmotivated, and thus the arguments fail.
2018. The Rational Impermissibility of Accepting (some) Racial Generalizations
Synthese (2018). doi: 10.1007/s11229-018-1809-5 . (Read online here; no paywall)
Synthese (2018). doi: 10.1007/s11229-018-1809-5 . (Read online here; no paywall)
- I argue that inferences from highly probabilifying racial generalizations are not solely objectionable because acting on such inferences would be problematic, or doing so violates a moral norm, but because they violate a distinctively epistemic norm. They involve accepting a proposition when, given the costs of a mistake, one is not adequately justified in doing so.
2017. Reasonable Mistakes and Regulative Norms: Racial Bias in Defensive Harm
Journal of Political Philosophy, vol. 25 no. 2 (2017): 196–217. doi: 10.1111/jopp.12120 . (Philpapers link)
Journal of Political Philosophy, vol. 25 no. 2 (2017): 196–217. doi: 10.1111/jopp.12120 . (Philpapers link)
- A regulative norm for permissible defense distinguishes the conditions under which we will hold defenders to be innocent of any wrongdoing from those in which we hold them responsible for assault or manslaughter. We have some reason to want a norm that considers a mistake permissible when it was highly likely on the evidence that defense was proportionate and necessary to avert a threat. However, I survey empirical data suggesting that the type and extent of bias prevalent in the US renders such a norm unjust.
2017. Revisiting the Right to do Wrong
Australasian Journal of Philosophy, vol. 95 no. 1 (2017): 43-57. doi: 10.1080/00048402.2016.1179654 . (Philpapers link)
Australasian Journal of Philosophy, vol. 95 no. 1 (2017): 43-57. doi: 10.1080/00048402.2016.1179654 . (Philpapers link)
- Rights to do wrong are not (contra Waldron, Enoch, and Herstein) necessary (even) within the framework of interest-based rights aimed at preserving autonomy. Agents can make morally significant choices and develop their moral character without a right to do wrong, so long as we either acknowledge supererogatory and minimally satisficing acts, or allow that there can be moral variation within the set of actions an agent is obligated to perform. Furthermore, the intuition that in some cases individuals do have a right to do wrong can be explained as stemming from a cautionary principle motivated by the asymmetry between the risk of wrongly interfering and that of refraining from interfering.
2015. The Pragmatics of Slurs
Nous, vol. 51 no. 3 (2017): 439-462. (Published online March 2015) doi: 10.1111/nous.12090 . (Philpapers link)
Nous, vol. 51 no. 3 (2017): 439-462. (Published online March 2015) doi: 10.1111/nous.12090 . (Philpapers link)
- I argue that the offense generation pattern of slurring terms parallels that of impoliteness behaviors, and is best explained by appeal to purely pragmatic mechanisms akin to those at work in impoliteness phenomena. In choosing to use a slurring term rather than its neutral counterpart, the speaker signals that she endorses the term (and its associations). Such an attitude is offensive, and consequently slurs generate offense whenever a speaker’s use demonstrates a contrastive preference for the slurring term. Since such an explanation comes at low theoretical cost and imposes few constraints on an account of the semantics of slurs, this suggests that we should not require semantic accounts to provide an independent explanation of embedding patterns.
research areas
ethics & politicalMoral Rights, Consent, Self-Defense, War & Policing |
epistemologyProfiling & Statistical Evidence, Moral Encroachment |
languageSlurs & Epithets, Disagreement, Semantics |
the papers below are works in progress; feel free to email me to ask for a draft.
ethics & political philosophy
moral rights
Epistemic risks & objective moral rights: beyond subjective permissibility
under review
This paper argues that our understanding of objective rights must be sensitive to agents’ epistemic limitations. On one popular understanding (which I call the `full-information fact-relative' interpretation), considerations about ignorance are relevant only to the `subjective permissibility' of an act, affecting culpability but not whether an act is a rights-violation. Against this view, I argue that subjective permissibility is not an adequate answer to the problems that agent ignorance poses for the deliberative and distributive roles of moral rights. If rights are to fill the theoretical role assigned to them, they must issue fact-relative permissions that are at least somewhat sensitive to agents' evidential and epistemic limitations.
under review
This paper argues that our understanding of objective rights must be sensitive to agents’ epistemic limitations. On one popular understanding (which I call the `full-information fact-relative' interpretation), considerations about ignorance are relevant only to the `subjective permissibility' of an act, affecting culpability but not whether an act is a rights-violation. Against this view, I argue that subjective permissibility is not an adequate answer to the problems that agent ignorance poses for the deliberative and distributive roles of moral rights. If rights are to fill the theoretical role assigned to them, they must issue fact-relative permissions that are at least somewhat sensitive to agents' evidential and epistemic limitations.
consent
Moral risk & communicating consent
under review
In addition to protecting agents’ autonomy, consent plays a crucial social role: it enables agents to secure partners in valuable interactions that would be prohibitively morally risk otherwise. To do this, consent must be observable: agents must be able to track the facts about whether they have received a consent-based permission. I argue that this morally justifies a consent-practice on which communicating that one consents is sufficient for consent, but also generates robust constraints on what sorts of behaviours can be taken as consent-communicating.
under review
In addition to protecting agents’ autonomy, consent plays a crucial social role: it enables agents to secure partners in valuable interactions that would be prohibitively morally risk otherwise. To do this, consent must be observable: agents must be able to track the facts about whether they have received a consent-based permission. I argue that this morally justifies a consent-practice on which communicating that one consents is sufficient for consent, but also generates robust constraints on what sorts of behaviours can be taken as consent-communicating.
self-defense
The Problem of the Mistaken Defender
Work in progress.
There is widespread agreement that agents are justified in imposing defensive harm in at least some cases where they are reasonably mistaken about whether the threat is genuine (perhaps because the apparent aggressor is bluffing). Still, not every evidentially justified mistaken defense is permissible: in some cases (like McMahan's Mistaken Resident) the mistaken defender wrongs the apparent aggressor. I argue that none of the dominant accounts are able without supplementation to secure and explain the intuitive verdicts about the full range of cases of mistaken defense. At minimum we should look for an extension that identifies the moral feature that renders permissible mistakes permissible, links the permission in the right sort of way to the apparent aggressor's actions, and keeps the account from over-generating.
Moral Self-Defense without Omniscience
under review
Sometimes an agent D, acting on her best evidence, imposes defensive harm on an apparent aggressor A, but in fact A posed no genuine threat to D. When such a mistake is induced by A’s acting in a way that foreseeably appears threatening, A shares responsibility for the error, and has at most a diminished complaint of having been wronged. I argue that this is best explained by positing duties grounded in agents’ entitlements to a fair distribution of the risk of suffering unjust harm. I suggest that the content of these duties is filled in by social conventions: when individuals A behave in ways that conventionally signal that they pose a threat to an agent D, A cannot reasonably demand that D refrain from defensive action, so A has at most a diminished complaint against D’s defensive action. Since this account makes essential reference to the socially defined meaning of the marked signaling behaviors, I conclude that this gives us reason to think that the social context can make a difference to agents’ defensive permissions.
The Case for Conventional Defensive Permissions
Work in progress.
On most accounts, whether it is morally permissible to harm someone in self-defense depends exclusively on whether doing so is necessary to avert a genuine threat for which they are responsible, or else is justified by lesser-evil considerations. Problematically, because they can't distinguish between genuine and merely apparent threats, these accounts leave agents prone to unintentional rights-transgressions, and fail to fairly distribute the risk of suffering unjust harm. I argue that this mischaracterizes agents' defensive permissions in a social context. Taking promissory obligations as a model, I show that there is moral reason to introduce a convention to mediate and extend agents' defensive permissions. Given the uncertainty that agents must act under, there is reason to permit defense against agents to avoidably act in ways that are conventional signals of aggression, even if (unknown to the defender) the threat is merely apparent.
Work in progress.
There is widespread agreement that agents are justified in imposing defensive harm in at least some cases where they are reasonably mistaken about whether the threat is genuine (perhaps because the apparent aggressor is bluffing). Still, not every evidentially justified mistaken defense is permissible: in some cases (like McMahan's Mistaken Resident) the mistaken defender wrongs the apparent aggressor. I argue that none of the dominant accounts are able without supplementation to secure and explain the intuitive verdicts about the full range of cases of mistaken defense. At minimum we should look for an extension that identifies the moral feature that renders permissible mistakes permissible, links the permission in the right sort of way to the apparent aggressor's actions, and keeps the account from over-generating.
Moral Self-Defense without Omniscience
under review
Sometimes an agent D, acting on her best evidence, imposes defensive harm on an apparent aggressor A, but in fact A posed no genuine threat to D. When such a mistake is induced by A’s acting in a way that foreseeably appears threatening, A shares responsibility for the error, and has at most a diminished complaint of having been wronged. I argue that this is best explained by positing duties grounded in agents’ entitlements to a fair distribution of the risk of suffering unjust harm. I suggest that the content of these duties is filled in by social conventions: when individuals A behave in ways that conventionally signal that they pose a threat to an agent D, A cannot reasonably demand that D refrain from defensive action, so A has at most a diminished complaint against D’s defensive action. Since this account makes essential reference to the socially defined meaning of the marked signaling behaviors, I conclude that this gives us reason to think that the social context can make a difference to agents’ defensive permissions.
The Case for Conventional Defensive Permissions
Work in progress.
On most accounts, whether it is morally permissible to harm someone in self-defense depends exclusively on whether doing so is necessary to avert a genuine threat for which they are responsible, or else is justified by lesser-evil considerations. Problematically, because they can't distinguish between genuine and merely apparent threats, these accounts leave agents prone to unintentional rights-transgressions, and fail to fairly distribute the risk of suffering unjust harm. I argue that this mischaracterizes agents' defensive permissions in a social context. Taking promissory obligations as a model, I show that there is moral reason to introduce a convention to mediate and extend agents' defensive permissions. Given the uncertainty that agents must act under, there is reason to permit defense against agents to avoidably act in ways that are conventional signals of aggression, even if (unknown to the defender) the threat is merely apparent.
epistemology
moral encroachment, profiling, & statistical evidence
Explaining Justificatory Asymmetries between Statistical and Individualized Evidence
work in progress
In some cases, there appears to be an asymmetry in the evidential value of statistical and more individualized evidence. For example, while I may accept that Alex is guilty based on eyewitness testimony that is 80% likely to be accurate, it does not seem permissible to do so based on the fact that 80% of a group that Alex is a member of are guilty. In this paper I suggest that rather than reflecting a deep fact about the content of types of evidence, this asymmetry might arise from the moral features of the relation between the source of evidence and the target’s agency. While relying on statistical evidence plausibly raises the stakes of error by introducing new kinds of risk to members of the reference class, paradigmatically “individualized” evidence—evidence tracing back to A’s voluntary behaviour—lowers the stakes of error. Plausibly the degree of evidential support needed to justify accepting a proposition as true depends on the stakes of error. If so, then these facts explain the apparent evidential asymmetry without positing a deep difference in the brute justificatory power of different types of evidence.
work in progress
In some cases, there appears to be an asymmetry in the evidential value of statistical and more individualized evidence. For example, while I may accept that Alex is guilty based on eyewitness testimony that is 80% likely to be accurate, it does not seem permissible to do so based on the fact that 80% of a group that Alex is a member of are guilty. In this paper I suggest that rather than reflecting a deep fact about the content of types of evidence, this asymmetry might arise from the moral features of the relation between the source of evidence and the target’s agency. While relying on statistical evidence plausibly raises the stakes of error by introducing new kinds of risk to members of the reference class, paradigmatically “individualized” evidence—evidence tracing back to A’s voluntary behaviour—lowers the stakes of error. Plausibly the degree of evidential support needed to justify accepting a proposition as true depends on the stakes of error. If so, then these facts explain the apparent evidential asymmetry without positing a deep difference in the brute justificatory power of different types of evidence.
Demographic Statistics in Defensive Decisions
work in progress
Evidence-relative accounts of moral rights (such as those advocated by Zimmerman [2008], van der Vossen [2016]) imply, among other things, that whether A has a right against being harmed depends on whether B's evidence suggests that he is liable to such harm. One could worry that such views allow B to reason from statistics about a high rate of aggression in A's demographic group to sufficient confidence that A is liable to effectively strip A of his rights. After constructing the strongest form of this objection, I offer several replies: (i) most demographic statistics face an unmet challenge from reference class problems, (ii) even those that meet it fail to ground non-negligible conditional probabilities, (iii) even if they did, they introduce new risks likely to cancel out any justificatory contribution of the statistic, but (iv) even if they didn't, demographic facts are the wrong sort to make a moral difference to agents' negative rights. The first three points can be used by an evidence-relative rights theorist to argue that demographic statistics simply will not make a problematic contribution to determinations of agents' rights. The last point, however, militates against certain evidence-relative accounts and in favor of a family of perspectival fact-relative accounts.
work in progress
Evidence-relative accounts of moral rights (such as those advocated by Zimmerman [2008], van der Vossen [2016]) imply, among other things, that whether A has a right against being harmed depends on whether B's evidence suggests that he is liable to such harm. One could worry that such views allow B to reason from statistics about a high rate of aggression in A's demographic group to sufficient confidence that A is liable to effectively strip A of his rights. After constructing the strongest form of this objection, I offer several replies: (i) most demographic statistics face an unmet challenge from reference class problems, (ii) even those that meet it fail to ground non-negligible conditional probabilities, (iii) even if they did, they introduce new risks likely to cancel out any justificatory contribution of the statistic, but (iv) even if they didn't, demographic facts are the wrong sort to make a moral difference to agents' negative rights. The first three points can be used by an evidence-relative rights theorist to argue that demographic statistics simply will not make a problematic contribution to determinations of agents' rights. The last point, however, militates against certain evidence-relative accounts and in favor of a family of perspectival fact-relative accounts.
philosophy of language
slurs & epithets
Naive and Non-Derogatory Slurring: Delimiting the Linguistic Community
Under review.
It isn't always obvious whether a term is a derogatory slur: some slurring terms are eventually reclaimed; some fade from use and memory. Others are active slurs in some communities, but operate at a remove from that history in the speaker's linguistic community. Sometimes speakers within a linguistic community use a term that they do not conceptualize as a slur, but which other members of that community do. Sometimes these speakers are ignorant or naive, but not always. This paper explores some puzzles raised by these `contested slurs', and what stubborn disagreements over whether a term is derogatory might teach us about the semantics of slurs.
The Language of Mental Illness
Under contract for Khoo and Sterken (eds.) Routledge Handbook of Social and Political Philosophy of Language
This paper surveys some philosophical issues with the language surrounding mental illness, but is especially focused on pejoratives relating to mental illness. I argue that though 'crazy' and similar mental illness-based epithets (MI-epithets) are not best understood as slurs, they do function to isolate, exclude, and marginalize members of the targeted group in ways similar to the harmfulness of slurs more generally. While they do not generally express the hate/contempt characteristic of weaponized uses of slurs, MI-epithets perpetuate epistemic injustice by portraying sufferers of mental illness as deserving minimal credibility. After outlining the ways in which these epithets can cause harm, I examine available legal and social remedies, and suggest that the best path going forward is to pursue a reclamation project rather than aiming to censure the use of MI-epithets.
Reporting Bad Beliefs: Solutions for Pragmatic Accounts
Work in progress.
Pragmatic accounts of slurs are characterized by commitment to the equivalence thesis: that slurs are semantically equivalent to their neutral counterparts. I present the three most common objections to pragmatic approaches stemming from this thesis: the equivalence problem, that it makes “all (NCs) are (slurs)” a conceptual truth; the substitution problem, that the view falsely predicts slurs may be substituted into intensional contexts salva veritae; and the conversational problem, that reports omitting slurs are intuitively incomplete and thus indicate that slurs differ from NCs in semantic content. I show that pragmatic theorists have adequate responses to each objection. The substitution problem relies on principles she is free to deny, while the equivalence objection is question begging. Finally, the conversational problem relies on delicate data that does not clearly support a semantic interpretation.
Under review.
It isn't always obvious whether a term is a derogatory slur: some slurring terms are eventually reclaimed; some fade from use and memory. Others are active slurs in some communities, but operate at a remove from that history in the speaker's linguistic community. Sometimes speakers within a linguistic community use a term that they do not conceptualize as a slur, but which other members of that community do. Sometimes these speakers are ignorant or naive, but not always. This paper explores some puzzles raised by these `contested slurs', and what stubborn disagreements over whether a term is derogatory might teach us about the semantics of slurs.
The Language of Mental Illness
Under contract for Khoo and Sterken (eds.) Routledge Handbook of Social and Political Philosophy of Language
This paper surveys some philosophical issues with the language surrounding mental illness, but is especially focused on pejoratives relating to mental illness. I argue that though 'crazy' and similar mental illness-based epithets (MI-epithets) are not best understood as slurs, they do function to isolate, exclude, and marginalize members of the targeted group in ways similar to the harmfulness of slurs more generally. While they do not generally express the hate/contempt characteristic of weaponized uses of slurs, MI-epithets perpetuate epistemic injustice by portraying sufferers of mental illness as deserving minimal credibility. After outlining the ways in which these epithets can cause harm, I examine available legal and social remedies, and suggest that the best path going forward is to pursue a reclamation project rather than aiming to censure the use of MI-epithets.
Reporting Bad Beliefs: Solutions for Pragmatic Accounts
Work in progress.
Pragmatic accounts of slurs are characterized by commitment to the equivalence thesis: that slurs are semantically equivalent to their neutral counterparts. I present the three most common objections to pragmatic approaches stemming from this thesis: the equivalence problem, that it makes “all (NCs) are (slurs)” a conceptual truth; the substitution problem, that the view falsely predicts slurs may be substituted into intensional contexts salva veritae; and the conversational problem, that reports omitting slurs are intuitively incomplete and thus indicate that slurs differ from NCs in semantic content. I show that pragmatic theorists have adequate responses to each objection. The substitution problem relies on principles she is free to deny, while the equivalence objection is question begging. Finally, the conversational problem relies on delicate data that does not clearly support a semantic interpretation.
semantics & disagreement
Freedom from Free Contrastivism
With Maegan Fairchild, work in progress
Walter Sinnott-Armstrong has recently suggested that the best analysis of free is contrastive. To evaluate this proposal, we work out in more detail the defining features of a contrastive account. Drawing on the structure of contrastive analyses of knows and ought, we argue that contrast classes must display continued relevance and comparative dependence. We then show that while more free than may be contrastive, no contrastive analysis can respect the intuitive concept of free. Nevertheless, Sinnott-Armstrong’s proposed picture may yet be the best analysis of free: his account respects the contours of the intuitive concept, partly in virtue of failing to display features necessary to count as a contrastive analysis.
With Maegan Fairchild, work in progress
Walter Sinnott-Armstrong has recently suggested that the best analysis of free is contrastive. To evaluate this proposal, we work out in more detail the defining features of a contrastive account. Drawing on the structure of contrastive analyses of knows and ought, we argue that contrast classes must display continued relevance and comparative dependence. We then show that while more free than may be contrastive, no contrastive analysis can respect the intuitive concept of free. Nevertheless, Sinnott-Armstrong’s proposed picture may yet be the best analysis of free: his account respects the contours of the intuitive concept, partly in virtue of failing to display features necessary to count as a contrastive analysis.
Metalinguistic Moral Disagreements
Under review
The problem of moral disagreement has been presented as an objection to contextualist semantics for ought, since it is not clear that contextualism can accommodate or give a convincing gloss of such disagreement. I argue that independently of our semantics, disagreements over oughts in non-cooperative contexts are best understood in the framework of metalinguistic negation, which is easily accommodated by contextualism. If this is correct, then rather than posing a problem for contextualism, the data from moral disagreements provides some reason to adopt a semantics that allows variance in the meanings of oughts.
Under review
The problem of moral disagreement has been presented as an objection to contextualist semantics for ought, since it is not clear that contextualism can accommodate or give a convincing gloss of such disagreement. I argue that independently of our semantics, disagreements over oughts in non-cooperative contexts are best understood in the framework of metalinguistic negation, which is easily accommodated by contextualism. If this is correct, then rather than posing a problem for contextualism, the data from moral disagreements provides some reason to adopt a semantics that allows variance in the meanings of oughts.
longer work
Book Project - Reasonable Mistakes: Moral Rights Under Uncertainty
manuscript in progress
Uncertainty shapes the moral demands agents can make on each other. Many of our interactions involve actions that are permissible---perhaps even called for---if our assumptions about the moral landscape are correct, but violate someone’s rights if we are mistaken. The need to fairly distribute the risk and costs of these mistakes in a society gives us moral reason to adopt ``normative conventions'': social practices that mediate, and in some cases alter the contours of, agents’ rights. I’m especially interested in the conditions under which such conventions can attach morally significant public meanings to behaviors (e.g. consenting by nodding, promising by shaking hands, etc.), which make it morally permissible for others to treat the agent as having consented, promised, etc. Identifying constraints on normative conventions is theoretically fruitful, allowing us not only to better understand agents’ rights and permissions under uncertainty, but also to make progress on thorny questions about implicit consent, evidential strength, and the ethics of war.
Uncertainty shapes the moral demands agents can make on each other. Many of our interactions involve actions that are permissible---perhaps even called for---if our assumptions about the moral landscape are correct, but violate someone’s rights if we are mistaken. The need to fairly distribute the risk and costs of these mistakes in a society gives us moral reason to adopt ``normative conventions'': social practices that mediate, and in some cases alter the contours of, agents’ rights. I’m especially interested in the conditions under which such conventions can attach morally significant public meanings to behaviors (e.g. consenting by nodding, promising by shaking hands, etc.), which make it morally permissible for others to treat the agent as having consented, promised, etc. Identifying constraints on normative conventions is theoretically fruitful, allowing us not only to better understand agents’ rights and permissions under uncertainty, but also to make progress on thorny questions about implicit consent, evidential strength, and the ethics of war.
Dissertation - Mistaken Self-Defense and Normative Conventions
submitted to the University of Southern California December 2017. Supervised by Jonathan Quong (chair), Mark Schroeder, Robin Jeshion, Seana Shiffrin, Richard John, and John Hawthorne.
My dissertation concerned how non-culpable ignorance affects agents' defensive permissions. I argued that agents' entitlements to a fair distribution of the risk of suffering unjust harm morally justifies a convention which, once adopted, extends defensive permissions to allow reasonable mistakes. Such a convention marks certain avoidable behaviors as signals of aggression, licensing agents to respond defensively. I developed an account of the necessary features and constraints on such `normative conventions', and leverage this to distinguish between defensive errors that are rational (given the defender's evidence) but nevertheless wrong merely apparent aggressors, and errors that are reasonable and morally permitted. The account I offered is descriptively promising, so insofar as it also has normative purchase, it yields prescriptions for reforming actual Anglo-American legal practices. If determinations of the reasonableness of mistakes rely on signals that are not adequately avoidable (e.g., because keyed to racial identities), the practice is morally bankrupt, and must be reformed before it can be used to identify permissible mistakes.
My dissertation concerned how non-culpable ignorance affects agents' defensive permissions. I argued that agents' entitlements to a fair distribution of the risk of suffering unjust harm morally justifies a convention which, once adopted, extends defensive permissions to allow reasonable mistakes. Such a convention marks certain avoidable behaviors as signals of aggression, licensing agents to respond defensively. I developed an account of the necessary features and constraints on such `normative conventions', and leverage this to distinguish between defensive errors that are rational (given the defender's evidence) but nevertheless wrong merely apparent aggressors, and errors that are reasonable and morally permitted. The account I offered is descriptively promising, so insofar as it also has normative purchase, it yields prescriptions for reforming actual Anglo-American legal practices. If determinations of the reasonableness of mistakes rely on signals that are not adequately avoidable (e.g., because keyed to racial identities), the practice is morally bankrupt, and must be reformed before it can be used to identify permissible mistakes.
