418 96 3MB
English Pages [499] Year 2021
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Applied Epistemology
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
E N G AG I N G P H I L O S O P H Y This series is a new forum for collective philosophical engagement with controversial issues in contemporary society. Disability in Practice Attitudes, Policies, and Relationships Edited by Adam Cureton and Thomas E. Hill, Jr Taxation Philosophical Perspectives Edited by Martin O’Neill and Shepley Orr Bad Words Philosophical Perspectives on Slurs Edited by David Sosa Academic Freedom Edited by Jennifer Lackey Lying Language, Knowledge, Ethics, Politics Edited by Eliot Michaelson and Andreas Stokke Treatment for Crime Philosophical Essays on Neurointerventions in Criminal Justice Edited by David Birks and Thomas Douglas Games, Sport, and Play Philosophical Essays Edited by Thomas Hurka Effective Altruism Philosophical Issues Edited by Hilary Greaves and Theron Pummer Philosophy and Climate Change Edited by Mark Budolfson, Tristram McPherson, and David Plunkett
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Applied Epistemology Edited by
J E N N I F E R L AC K EY
1
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
1 Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © the several contributors 2021 The moral rights of the authors have been asserted First Edition published in 2021 Impression: 1 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2020950719 ISBN 978–0–19–883365–9 DOI: 10.1093/oso/9780198833659.001.0001 Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Contents List of Contributorsvii PA RT 1 : I N T R O D U C T IO N 1. Applied Epistemology Jennifer Lackey
3
PA RT 2 : E P I S T E M O L O G IC A L P E R SP E C T I V E S 2. When Freeing Your Mind Isn’t Enough: Framework Approaches to Social Transformation and Its Discontents Kristie Dotson and Ezgi Sertler
19
3. Situated Knowledge, Purity, and Moral Panic Quill R. Kukla
37
4. Epistemology and the Ethics of Animal Experimentation Mylan Engel Jr.
67
PA RT 3 : E P I S T E M IC A N D D OX A S T IC W R O N G S 5. A Tale of Two Doctrines: Moral Encroachment and Doxastic Wronging99 Rima Basu 6. Predatory Grooming and Epistemic Infringement Lauren Leydon-Hardy
119
PA RT 4 : E P I S T E M O L O G Y A N D I N J U S T IC E 7. Epistemic Degradation and Testimonial Injustice Geoff Pynn
151
8. My Body as a Witness: Bodily Testimony and Epistemic Injustice José Medina and Tempest Henning
171
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
vi Contents
PA RT 5 : E P I S T E M O L O G Y, R AC E , A N D T H E AC A D E M Y 9. The ‘White’ Problem: American Sociology and Epistemic Injustice Charles W. Mills
193
10. A Tale of Two Injustices: Epistemic Injustice in Philosophy Emmalon Davis
215
PA RT 6 : E P I S T E M O L O G Y A N D F E M I N I S T P E R SP E C T I V E S 11. Rape Culture and Epistemology Bianca Crewe and Jonathan Jenkins Ichikawa 12. Feminist Pornography as Feminist Propaganda, and Ideological Catch-22s Aidan McGlynn
253
283
PA RT 7 : E P I S T E M O L O G Y A N D SE X UA L C O N SE N T 13. Epistemic Responsibility in Sexual Coercion and Self-Defense Law Hallie Liberto
305
14. Sexual Consent and Epistemic Agency Jennifer Lackey
321
15. The Epistemology of Consent Alexander A. Guerrero
348
PA RT 8 : E P I S T E M O L O G Y A N D T H E I N T E R N E T 16. The Internet and Epistemic Agency Hanna Gunn and Michael Patrick Lynch
389
17. How Twitter Gamifies Communication C. Thi Nguyen
410
18. The Epistemic Dangers of Context Collapse Online Karen Frost-Arnold
437
19. ‘Yikkity Yak, Who Said That?’ The Epistemology of Anonymous Assertions457 Veronica Ivy Index
481
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
List of Contributors Rima Basu, Claremont McKenna College Bianca Crewe, University of British Columbia Emmalon Davis, University of Michigan Kristie Dotson, Michigan State University Mylan Engel Jr., Northern Illinois University Karen Frost-Arnold, Hobart and William Smith Colleges Alexander A. Guerrero, Rutgers University Hanna Gunn, University of California, Merced Tempest Henning, Vanderbilt University Jonathan Jenkins Ichikawa, University of British Columbia Veronica Ivy, College of Charleston Quill R. Kukla, Georgetown University Jennifer Lackey, Northwestern University Lauren Leydon-Hardy, Amherst College Hallie Liberto, University of Maryland, College Park Michael Patrick Lynch, University of Connecticut Aidan McGlynn, University of Edinburgh José Medina, Northwestern University Charles W. Mills, City University of New York C. Thi Nguyen, University of Utah Geoff Pynn, Elgin Community College Ezgi Sertler, Butler University
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
PART 1
IN T RODU CT ION
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 19/03/21, SPi
1
Applied Epistemology Jennifer Lackey
Applied epistemology brings the tools of contemporary epistemology to bear on particular issues of social concern. While the field of social epistemology has flourished in recent years, there has been far less work on how theories of know ledge, justification, and evidence may be applied to concrete questions, especially those of ethical and political significance. The present volume fills this gap in the current literature by bringing together essays from leading philosophers in a broad range of areas in applied epistemology. The potential topics in applied epis temology are many and diverse, and this volume focuses on seven central issues, some of which are general, while others are far more specific: epistemological perspectives; epistemic and doxastic wrongs; epistemology and injustice; epis temology, race, and the academy; epistemology and feminist perspectives; epis temology and sexual consent; and epistemology and the internet. Some of the chapters in this volume contribute to, and further develop, areas in social epis temology that are already active, and others open up entirely new avenues of research. All of the contributions aim to make clear the relevance and importance of epistemology to some of the most pressing social and political questions facing us an agents in the world. One question that might be taken up in applied epistemology concerns the power of knowledge or understanding to bring about concrete social change in the world. In Chapter 2, “When Freeing Your Mind Isn’t Enough: Framework Approaches to Social Transformation and Its Discontents,” Kristie Dotson and Ezgi Sertler ask what relationships framework shifts have to social transform ation. In particular, they explore whether a framework shift might have a radical causal impact on actual social arrangements. Dotson and Sertler understand social transformation as concrete changes in the structures and realities that are being “framed,” where the actual structures or social arrangements are not redu cible to the frameworks. They then turn to an example of a framework approach to social justice: the framework analysis around “political prisoners,” which is often taken to be a discourse aimed at liberating “political prisoners.” They exam ine whether generating a new understanding of “prisoners” who are “political,” and shifting our understanding of the current social arrangements, has the poten tial to bring about change in how “political prisoners” are managed by the states in question. Dotson and Sertler argue that the conceptual resilience of carceral Jennifer Lackey, Applied Epistemology In: Applied Epistemology. Edited by: Jennifer Lackey, Oxford University Press (2021). © Jennifer Lackey. DOI: 10.1093/oso/9780198833659.003.0001
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
4 Jennifer Lackey logic and its practices reveals the significant limitations of this approach. “Carceral logic,” according to Dotson and Sertler, refers to a dominant and epistemologic ally resilient logic that legitimizes state-centric practices of incarceration, which are exercised through prisons that impose punitive exclusions through disciplin ary containment. They consider a disaggregation approach, which identifies who a “political prisoner” is and thus intends to separate this category from other kinds of “prisoners,” and a comprehensive approach, which holds that all “prisoners” are in some sense “political prisoners” because the nature of imprisonment is political. According to Dotson and Sertler, the disaggregation approach strengthens the assumption that there are correctly imprisoned people by identifying the category of “political prisoners.” In addition, understanding “political prisoners” differently does not guarantee understanding violence, objection, dissidence, and so on, differently. The comprehensive approach, Dotson and Sertler argue, targets the punitive dimension of the current imprisonment system, leaving the prevent ive aspect untouched. In this way, delegitimizing the punitive aspect of a resilient system does not result in delegitimizing the preventive aspect of it and this, in turn, can legitimize the confinement of all “political prisoners.” Dotson and Sertler conclude that bringing about social transformations through framework shifts is dubious, at best, and ought to be rejected, at worst. In the particular case at hand, they claim that intellectual approaches to social transformation change neither the structural realities of imprisonment nor the epistemological resilience of carceral logic. In Chapter 3, Quill R. Kukla looks at contemporary epistemology in “Situated Knowledge, Purity, and Moral Panic,” and argues that much of it is driven by a kind of a moral panic regarding the concern that there are no “pure” epistemic practices, perspectives, or standards detachable from the social situation of know ers. While some epistemologists argue that various traditional epistemological notions, such as knowledge or justification, are ineliminably situated, others respond by carving out smaller spaces of epistemic purity. Kukla argues that this dialectic is driven in large part by fear rather than by intellectual tension, and that while epistemic practices are ineliminably situated in multiple ways, this should not be feared. Given this, Kukla claims that the quest for purity is misguided, and that the collective goal should be to recognize the fear as a product of ideology, accepting situatedness as an everyday phenomenon. Kukla concludes by arguing that an appropriate naturalized, non-ideal epistemology will treat situatedness not as something mysterious or fearful, but as an empirical fact about our epistemic practices. In Chapter 4, “Epistemology and the Ethics of Animal Experimentation,” Mylan Engel Jr. argues that a close examination of the epistemology of animal experimentation shows that such research is neither epistemically nor morally justified and should be abolished. Engel argues that the only serious attempt at
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Applied Epistemology 5 justifying animal experimentation is the benefits argument, according to which animal experiments are justified because the benefits that humans receive from the experiments outweigh the costs imposed on the animal subjects. According to Engel, the benefits we allegedly receive from animal-based biomedical research are primarily epistemic in that experimenting on animal models is supposed to provide us with knowledge of the origin and proper treatment of human disease. However, Engel argues that animal models are extremely unreliable at predicting how drugs will behave in humans, whether candidate drugs will be safe in humans, and whether candidate drugs will be effective in humans. In addition, Engel shows that animal-based biomedical research has a proven track record of unreliability when it comes to determining the origin, pathology, and treatment of human disease. Since methods that are known to be unreliable are not sources of justification, evidence, or knowledge, regardless of whether one espouses externalism or internalism in epistemology, Engel argues that animal- based research fails to provide the epistemic benefits needed to justify its continued use. Given its unreliability, Engel concludes that animal experimentation does not, and cannot, provide the epistemic benefits needed to outweigh the harms inflicted on the animal subjects involved, and thus animal experimentation is neither epi stemically nor morally justified. One of the more fertile areas of inquiry in applied epistemology concerns epi stemic and doxastic wrongs, both questions about how to understand them and examinations of particular instances of such wrongs. In Chapter 5, “A Tale of Two Doctrines: Moral Encroachment and Doxastic Wronging,” Rima Basu argues on behalf of two theses whereby morality bears on belief—moral encroachment and doxastic wronging—and clarifies the relationship that holds between them. According to moral encroachment, moral features make a difference to whether a given belief is epistemically justified, and thus epistemic justification is not purely a matter of the epistemic. Doxastic wronging is the thesis that because beliefs mediate our interpersonal relations to others, they can be the source of moral wrongdoing. Basu argues that doxastic wrongs are (i) directed; (ii) committed by beliefs, rather than the consequences of acting on beliefs; and (iii) wrongs in vir tue of the content of what is believed. Basu concludes that while moral encroach ment and doxastic wronging are conceptually distinct, evidenced at least in part by the fact that some endorse one thesis while denying the other, we should accept both. In Chapter 6, “Predatory Grooming and Epistemic Infringement,” Lauren Leydon-Hardy identifies, develops, and applies a new concept in the epistemo logical literature—what she calls epistemic infringement. To epistemically infringe on another is to violate interpersonal social and epistemic norms so as to encroach upon or undermine that person’s epistemic agency. Epistemically infringing behavior standardly involves complex projects of deceit, manipulation, and
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
6 Jennifer Lackey coercion. Leydon-Hardy focuses on the phenomenon of predatory grooming to provide a powerful case study of epistemic infringement. Grooming behavior is familiar from some very high-profile cases of sexual abuse, such as those of Jerry Sandusky and Larry Nassar. While grooming often involves adults preying on young children, it can also occur between adults, and even among peers. More precisely, predatory grooming is a process whereby targeted individuals are primed, coached, or generally readied for conduct that is exploitative in nature. Drawing on work in forensic psychology, Leydon-Hardy presents a general model of grooming that involves cycling through phases in a non-linear process that is called test–operate–test, where groomers “take the temperature” of their victims, wait for feedback, and then engage in damage control or push boundaries further, depending on the victims’ reactions. Leydon-Hardy then shows that a central and entirely ignored dimension of predatory grooming is the epistemic force of this phenomenon. Indeed, it is only by viewing this exploitative behavior through an epistemic lens that we can gain a full picture of both the wrongness of the groom ing and the most effective ways to respond to it. In particular, Leydon-Hardy argues that grooming aims at masking abuse even by the lights of the abused and thus crucially involves the cultivation of an unknowingness in its victims. In this way, grooming is not just a lie or a form of deception, but a sustained campaign of manipulation and coercion that untethers victims from their own epistemic resources and from their ability to marshal those resources appropriately. This distinctively epistemic explanation helps explain some of the harms that victims suffer, such as feeling complicit in their own victimization because of their relationships with the groomers, or how they struggle with the fact that they were harmed in plain sight, often right under the noses of those who love them most. Leydon-Hardy also shows how epistemic infringement is unlike any other concept at work in the area of social epistemology. Most forms of epistemic wrong discussed in the literature focus on the inability of people to effectively communi cate their own experiences and beliefs because of various kinds of systematic prejudice. For instance, testimonial injustice occurs when a speaker receives a credibility deficit because of a bias that targets his or her social identity; hermen eutical injustice1 involves a gap in the discursive resources, such as the absence of the concept of sexual harassment, that prevents a speaker from articulating aspects of her social experience; and testimonial smothering2 is at work when a testifier engages in self-silencing because she has reason to believe that she will not receive uptake. But the distinctively epistemic wrong suffered by victims of epistemic infringement cannot be captured in any of these already-existing terms in the philosophical literature. For what is epistemically problematic in grooming does not in any way involve a credibility deficit, a conceptual lacuna, or a form of
1 See, for instance, Fricker (2007).
2 See Dotson (2011).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Applied Epistemology 7 self-silencing. Rather, Leydon-Hardy argues that norms of trust are used to turn victims against themselves as epistemic agents, leaving them untethered to their own experiences and beliefs. This is a unique, and particularly pernicious, epi stemic wrong. One of the most important and widely discussed areas of inquiry in the current philosophical literature is the intersection of epistemology and various kinds of injustice. In Chapter 7, “Epistemic Degradation and Testimonial Injustice,” Geoff Pynn asks what is the nature of the wrong involved in cases of testimonial injust ice. More precisely, when a person’s testimony is given less credibility than it ought to receive, and this is due to a prejudice on the part of the hearer, how should we understand the wrong that the speaker suffers? After raising problems for accounts that explain the wrong in terms of objectification, where speakers are treated as mere sources of information rather than as informants, and in terms of derivatization, where speakers are treated as if their epistemic contributions are solely in support of, and not in tension with, any of our own capacities, Pynn proposes what he calls a degradation account of the wrong of testimonial injust ice. The wrong of testimonial injustice involves epistemic degradation, which consists in a public violation of a speaker’s epistemic status-linked entitlements. Drawing on the view that knowledge is the norm governing epistemically proper assertion, Pynn argues that a knowledgeable speaker whose assertion is rejected on the basis of an identity-prejudicial credibility deficit suffers a violation of her entitlement to acceptance. According to Pynn, any violation of the knowledge norm will tend to represent the testifier as a non-knower, and thus be moderately degrading. However, where the violation is rooted in a systematic negative iden tity prejudice, the rejection will also represent the speaker as a non-knower who is debased in the ways encoded by the stereotype in question. For instance, the diminished representation may depict a speaker as untrustworthy in virtue of his blackness or irrational in virtue of her femininity. In this way, the wrong of testimonial injustice is a distinctive kind of epistemic degradation that excludes victims from the social and epistemic rank shared by other conversational participants. José Medina and Tempest Henning examine the role that bodily testimony can play in social and political epistemology in Chapter 8, “My Body as a Witness: Bodily Testimony and Epistemic Injustice.” They develop an account of how to understand the testimonial force and content of non-verbal communicative acts, such as gestures and facial expressions, that depends on three features: the com municative context, the embodied positionality of the communicator, and the communicative uptake that the audience gives, or fails to give, to the expressive behavior of the body. In particular, Medina and Henning argue that under condi tions of racial oppression, all racialized bodies—non-white as well as white—are epistemically valued in different ways, and thus receive different kinds of com municative uptake. This differential valuing of racialized bodies and their bodily testimonial expressions, according to Medina and Henning, can result in
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
8 Jennifer Lackey testimonial injustice that targets non-white bodily testimony. At the same time, Medina and Henning argue that bodily group testimony is well suited for culti vating in-group communicative solidarity and for giving center-stage to in-group members in testimonial dynamics, and so bodily communication can be used in resistant testimony. In this way, Medina and Henning conclude that bodily testi mony can provide a powerful way to circumvent verbal limitations when people cannot talk openly and safely about certain issues and can provide a powerful critical tool for resisting epistemic oppression and for creating communicative solidarity. In addition to the examination of general epistemic wrongs, epistemological tools can be applied to concrete issues, such as the role that race plays in academic disciplines. Through a careful examination of the treatment of race in American sociology, Charles W. Mills argues for a radical expansion of the concept of epi stemic injustice in Chapter 9, “The ‘White’ Problem: American Sociology and Epistemic Injustice.” Mills focuses on two senses of racism: the mental/psycho logical—or racism as sentiment and/or belief—and the institutional/societal—or racism as structural domination/illicit advantage. According to Mills, a discipline does not begin ex nihilo, but rather from pre-existing beliefs, concepts, frame works, and norms, and the atmosphere of the United States of the late nineteenth century was deeply pervaded with racist assumptions that had originally devel oped in the colonial period to justify and rationalize African slavery. More gener ally, Mills argues that in societies characterized by deep structural oppression, hermeneutical obstacles will be far more extensive and entrenched than a few missing concepts. The main axes of social subordination—in this case, “race”— will act as powerful generators of cognitive distortion in fields. Because of this, Mills argues that “race” not only has to be rethought in sociology in particular, and in academic disciplines more broadly, but its linkages also need to be recon ceived. In addition, Mills shows that the ideological is linked to the material, and thus social structures and institutions constitute the material base of dominant-group ideologies and place restrictions on the class of respected cognizers. Given this, Mills concludes that the academy exemplifies epistemic injustice on a massive scale, shattering the boundaries typically assigned to the concept. In Chapter 10, “A Tale of Two Injustices: Epistemic Injustice in Philosophy,” Emmalon Davis identifies two different kinds of testimonial injustice— identity-based and content-based—and then argues that they can be used to provide an epistemic explanation for the persistent lack of diversity in academic philosophy. Identity-based testimonial injustice involves a prejudice or other unjust assess ment regarding a contributor’s social identity—such as gender, race, ability, and so on—that influences an audience’s assessment of the contributor’s epistemic stand ing, compromising the audience’s willingness to consider or fairly engage the contributor and contribution. Content- based testimonial injustice involves a
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Applied Epistemology 9 prejudice or other unjust assessment regarding social identity-coded content— such as gender-coded, race-coded, ability-coded, and so on—of a contributor’s contribution that influences an audience’s assessment of the contributor’s epistemic standing, compromising the audience’s willingness to consider or fairly engage the contributor and contribution. An example of the former is the rejection of a speaker’s report simply because she is a woman; an example of the latter is the dismissal of a speaker’s research simply because it is regarded as the kind of work women do. Davis argues that both kinds of testimonial injustice are ubiquitous in academic philosophy and that the prevalence of these injustices provides significant barriers to those targeted that can explain philosophy’s lack of diversity. In particular, targets of either form of testimonial injustice may have their contributions dismissed or ignored, be denied opportunities for sharing their ideas or shaping discussions, have their views attributed to someone else, and so on. Given that philosophy is a discipline that is carried out largely through written and spoken discourse, these communicative obstacles can have profound effects. Continuing the application of epistemological tools to concrete issues, several authors in this volume explore the intersection of epistemology and feminist per spectives. In Chapter 11, “Rape Culture and Epistemology,” Bianca Crewe and Jonathan Jenkins Ichikawa focus on the question of how institutions and individ uals should respond to sexual assault allegations that haven’t been established in legal settings. One tempting view, they suggest, is that institutions and individuals should be deferential to law enforcement, allowing the criminal justice system to handle the central epistemological dimensions of sexual assault allegations. However, Crewe and Ichikawa argue that it is crucial to assess this response within the broader social and political context in which sexual assault occurs, including cultural attitudes about sexual assault, women’s credibility, and mis ogynist assumptions about access and entitlement to women’s bodies. In particu lar, Crewe and Ichikawa maintain that deference to the law should be theorized within the context of rape culture, which is a cultural environment in which sex ual assault and sexualized violence is an expected type of interaction. Moreover, the political and ideological affiliations of legal systems themselves impact whether sexual assault is reported and the corresponding responses. For instance, Crewe and Ichikawa argue that when individuals testify that they were sexually assaulted, their testimony is often considered to be less reliable than most other testimony, as they are frequently regarded as either lying or deluded. Turning to the connection between knowledge and action, Crewe and Ichikawa consider whether an argument on behalf of deference to the law might be grounded in the connection between knowledge and action, which can be cap tured in the connection between knowledge and reasons for action. On this view, one’s reasons constitute all and only that which one knows. Accordingly, if there is
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
10 Jennifer Lackey a doubt, even one that is unwarranted, then one does not know, and hence one does not have a reason to act. By way of response, Crewe and Ichikawa argue that in cases of unwarranted doubts, it is still the case that one should know, and thus inaction is epistemically impermissible. Thus, deference to the law cannot be jus tified through the connection between knowledge and action. After considering and rejecting a further response drawing on pragmatic encroachment, Crewe and Ichikawa explore the complex relationship between feminist epistemology and contextualism. Contextualism about knowledge ascriptions is the view that sentences containing “knows” are context sensitive. According to Crewe and Ichikawa, a contextualist can tell a simple and plausible story about differential standards for action in criminal contexts and other con texts, such as in a university setting: one can count as “knowing” that someone has committed a serious offense in a conversation about how the university ought to respond to it, without counting as “knowing” it in a conversation about whether the state ought to incarcerate the perpetrator. They conclude by noting that, given contextualism, the flexibility of knowledge ascriptions comes along with significant social power, and that the decision to employ some standards rather than others is a political one. Furthermore, they argue that this power tends to be wielded in a way that protects the interests of the status quo. In this way, they conclude that their project can be seen as a continuation of the work of feminist philosophers who argue that the “view from nowhere” is in fact often a view from a very particular and situated location. In Chapter 12, Aidan McGlynn explores whether feminist pornography might have a positive epistemic function in effectively countering the propagandic power of mainstream pornography in “Feminist Pornography as Feminist Propaganda, and Ideological Catch-22s.” McGlynn understands pornography as sexually explicit materials that have the primary purpose of sexually arousing their audience and propaganda as media or speech that acts to spread ideology, where ideological attitudes are characteristically insensitive to evidence. McGlynn is specifically interested in feminist pornography that aims to (i) guarantee appro priate standards of health and safety for those involved in its making, and to minimize the chances that they have been victims of human trafficking; and/or (ii) present a more egalitarian picture of gender and sexuality through, for instance, placing an emphasis on explicit consent, female sexual agency, and so on. The central question McGlynn takes up, then, is whether feminist pornog raphy might be an effective vehicle for a more egalitarian, positive view of human sexual relations. McGlynn argues against this use of feminist pornography, particularly the proposals made by A. W. Eaton and Catarina Dutilh Novaes, showing that they leave us in what he calls an “ideological Catch-22 situation”: either feminist pornography has the power to reshape our sexual desires and attitudes, but we are left without an explanation of how to persuade consumers of
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Applied Epistemology 11 mainstream pornography to watch feminist pornography rather than mainstream pornography prior to that reshaping; or the propagandic force of pornography is seen as lying in the way it exploits and spreads pre-existing sexist ideology, leaving it a mystery how feminist pornography could bring about a shift in sexual ideology, even in principle. McGlynn concludes that mainstream pornography plays a distinctive role in the lives of its audience, and it is unclear how we can replace it with feminist pornography without censorship or some other kind of practically—and possibly morally—problematic intervention into people’s lives. While there is a great deal of work in the philosophical literature on the nature of sexual consent, very little has been done on its epistemic dimensions. Three authors in this volume take up this issue and, in so doing, lay the groundwork for a much-needed expansion of our understanding of sexual consent. In Chapter 13, “Epistemic Responsibility in Sexual Coercion and Self- Defense Law,” Hallie Liberto focuses on a kind of responsibility that is epistemic in nature: the respon sibility to gather information so as to be appropriately epistemically positioned to act on one’s beliefs. She then explores how to assign and respond to the adjudica tion of this kind of epistemic responsibility in cases of sexual coercion and self-defense in criminal law. In particular, Liberto argues that it is problematic to assess whether those accused of sexual coercion and unjustified killing had rea sonable beliefs, or whether they acted as reasonable people would act. For instance, feminists have long noted the problems that arise when juries are asked to gauge the reasonableness of fear when applying a “reasonable man” standard and have suggested that courts appeal to a “reasonable woman” standard in deter mining the actus reus of criminal cases in sexual assault and coercion. However, drawing on the work of Hubin and Healey, Liberto argues that a “reasonable woman” standard will not help in the prosecution of rape cases, as a “reasonable man” standard would still need to be applied in order to establish the mens rea of sexual offenses. Liberto then presents an alternative suggestion by Hubin and Healey, the “reasonable expectation from state” (REfS) standard, which asks: what is it reasonable for the state to expect of a person? Liberto argues that a REfS standard is preferable to the “reasonable person” standard currently used and should be adopted for adjudicating both self-defense and sexual coercion cases. One advantage of the REfS standard is that it goes some way towards preventing racism: instead of being asked what a reasonable man would believe about a woman’s consent, for instance, the jury would be asked if the man had met the reasonable expectations of the state to be sure of the woman’s consent—expect ations that would be the same whether the victim was black or white. Liberto concludes that, in contrast to Hubin and Healey, the expectations of the state need to be outlined ahead of time and be made known to the public in both sex ual coercion and self-defense cases.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
12 Jennifer Lackey In Chapter 14, “Sexual Consent and Epistemic Agency,” Jennifer Lackey exam ines sexual consent in the context of the widely accepted thesis that knowledge is sufficient for epistemically permissible action; that is, the view according to which if someone knows a given proposition, then it is epistemically permissible for that person to act on it. To the extent that this is denied, it is argued that either more or less than knowledge is required, such as certainty or justified belief. In this chapter, Lackey shows that being able to act on knowledge that someone has con sented to sex provides an interesting challenge to this framework. In particular, Lackey argues that someone may know that another consents to sex and yet it may still be epistemically impermissible to act on this knowledge. This is clearest when the knowledge of the consent in question is secondhand, rather than first hand. When this happens, the problem is not that more, or less, than knowledge is needed to warrant action, but, rather, that a particular kind of epistemic sup port is required, one that involves testimony from the consentee herself. This is due to the fact that the consentee is uniquely positioned with respect to the ques tion of her own consent, both agentially and epistemically. Lackey further argues, however, that it doesn’t follow from this that knowing firsthand that another con sents to sex is sufficient for it to be epistemically permissible to act on this know ledge. For someone might also have background beliefs, either in general or about another in particular, that function as defeaters for such action. Thus, a single instance of testimony granting consent needs to be viewed in a broader evidential framework, one where this piece of evidence alone might not be enough to war rant action on this occasion. In this way, Lackey defends the total evidence view of the epistemology of sexual consent. The upshot of these considerations is that determining whether sexual consent has been given, especially in light of how high the stakes can be, requires that agents be epistemically responsible, where this can go beyond what is required in standard cases of action. In Chapter 15, “The Epistemology of Consent,” Alexander A. Guerrero exam ines two different dimensions of consent. In Part 1 of the chapter, Guerrero focuses on the nature of consent—that is, the question of what it is for an agent, A, to consent to some state of affairs, SA. According to the attitudinal view, spe cific mental attitudes are both necessary and sufficient for consent. In particular, A consents to SA if, and only if, A has an attitude of affirmative endorsement toward SA. While Guerrero doesn’t provide a defense of the attitudinal view in this chapter, he argues that reluctance to endorse it is often due to a failure to appreciate that another agent, B, can non-culpably act as if A consents to SA only if B justifiably believes that A has an attitude of affirmative endorsement towards it. This leads to Part 2, where Guerrero takes a close look at the epistemological dimensions of justifiable belief that another person has consented to some state of affairs. He argues that moral stakes matter when the epistemology of consent is concerned in at least one of the following two ways: (i) the moral stakes or con text matter to whether B justifiably believes, or knows, that A consents to SA; or
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Applied Epistemology 13 (ii) the moral stakes or context matter to whether it is morally objectionable for B to act based on justified belief or knowledge that A consents to SA and whether B is non-culpable for acting as if “A consents to SA” is true. Guerrero goes on to show that in some cases where consent is at issue, the moral stakes are high and thus require that B possess more or stronger evidence in order to justifiably believe, or non-culpably act as if, A consents. Since cases for which sexual con sent and consent to medical treatment are at issue are also ones in which there are high moral stakes, Guerrero concludes that B must possess more or stronger evi dence in order to justifiably believe, or non-culpably act as if, A consents in these cases. In the final section of this volume, epistemic dimensions of the internet, both general and specific, are explored. In Chapter 16, “The Internet and Epistemic Agency,” Hanna Gunn and Michael Patrick Lynch examine the connection between epistemic agency and the internet. They begin by identifying two condi tions that are true of responsible epistemic agency: first, responsible epistemic agents aim to develop epistemic virtues, merit, and capacities that help them to responsibly change their epistemic environment, as well as the capacities that enable them to recognize and respect these epistemic traits in others. Second, responsible epistemic agents treat other epistemic agents with a form of respect that demonstrates a willingness to learn from them. Gunn and Lynch then high light three ways in which the internet has led to the “democratization” of know ledge. In particular, it (i) makes bodies of knowledge more widely available; (ii) makes knowledge production more inclusive; and (iii) is used in ways that expand epistemic participation, and thus enhance epistemic agency, through the develop ment of challenge-specific prizes. At the same time, Gunn and Lynch show that the very ways in which the internet makes information and knowledge more widely available can also undermine our ability to be responsible epistemic agents and may even undermine epistemic agency itself. One way this happens is that the accessibility of information itself can lead to increased epistemic arrogance which, in turn, can give rise to testimonial injustice by unjustifiably discounting the credibility of others. Since showing such respect is one of the features of being an epistemically responsible agent, epistemic arrogance can undermine respon sible epistemic agency. In addition, Gunn and Lynch argue that the personaliza tion of online spaces can unwittingly lead users into echo- chambers and filter-bubbles and away from a diverse range of epistemic perspectives, and fake news and information pollution can make for a hostile online epistemic environ ment. This makes it difficult for online users to fulfill their epistemic obligations. Finally, Gunn and Lynch show that online anonymity can undermine both the responsible acquisition and dissemination of knowledge through testimony. C. Thi Nguyen explores our interaction on a specific online platform in Chapter 17, “How Twitter Gamifies Communication,” arguing that by offering immediate, vivid, and quantified evaluations of conversational success, Twitter
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
14 Jennifer Lackey gamifies communication and, in so doing, changes the nature of the activity. A “design for communication” is the designed technology that offers points and scores. On Nguyen’s view, “gamification” occurs when a player interacts with design for gamification and in fact adopts these points and scores as the primary motivators during the activity, that is, when the activity becomes something like a game for them. Nguyen argues that pre-gamification, the values of conversation are complex and many, involving understanding the world, transmitting informa tion, persuading, connecting to one another, and so on. In contrast, Twitter’s scoring mechanism invites us to replace these values with another, much simpler goal: that of maximizing one’s Likes, Retweets, and Follower. As Nguyen argues, “[g]ames offer us a momentary experience of value clarity. They are a balm for the existential pains of real life.” But when we gamify conversation in this way, we are imposing value clarity on a set of complex and nuanced values. Twitter scores, for instance, make salient the number of users with positive reactions, while de-emphasizing the quality of any particular interaction. In this way, gamification instrumentalizes the goals of our real-life activities, which is problematic when the goals themselves are independently valuable. Nguyen concludes by drawing a connection between gamification, echo chambers, and moral outrage porn: all share a willingness to instrumentalize what shouldn’t be instrumentalized. In par ticular, echo chambers instrumentalize our trust, moral outrage porn instrumen talizes our morality, and gamification instrumentalizes our goals. In Chapter 18, “The Epistemic Dangers of Context Collapse Online,” Karen Frost-Arnold provides a close analysis of the epistemological challenges posed by context collapse in online environments and argues that virtue epistemology pro vides a helpful normative framework for addressing some of these problems. “Context collapse” is the blurring or merging of multiple contexts or audiences into one. For instance, Facebook users may collapse multiple contexts of social relations into one audience when they share a post with all of their Facebook friends, which may include family members, close friends, acquaintances, work colleagues, and complete strangers. The first epistemic challenge that Frost-Arnold identifies is that context collapse facilitates online harassment, which causes epistemic harm by decreasing the diversity of epistemic communities. For instance, a surreptitiously recorded lecture in a feminism course on a college campus might be shared on Reddit, leading to the trolling and harassment of the professor. The second epistemic problem with context collapse is that it threatens the integrity of marginalized epistemic communities in which some types of true belief flourish. For instance, marginalized people may need to create their own language in which to describe their oppression. Context collapse can disrupt these epistemic communities and thereby hinder the production of knowledge. The third epistemic challenge with context collapse is that it promotes misunder standing. Understanding relies on background knowledge which, in turn, is often
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Applied Epistemology 15 context sensitive. If testimony in one context is imported to another without shared background knowledge about the setting of the conversation, the speaker and hearer, past conversations, the goals of the conversation, and so on, then the hearer may misinterpret the speaker’s utterance. Frost-Arnold then argues that we can cultivate and promote the epistemic virtues of trustworthiness and discre tion in order to address some of these problems. When we are trustworthy, we are motivated to avoid taking advantage of the vulnerability of those who trust us and when we are discreet, we use good judgment in sharing speech, especially in a way that is attentive to the costs to others. Veronica Ivy takes up the issue of the epistemology of anonymous assertions in Chapter 19, the final chapter in this volume—“Yikkity Yak, Who Said That? The Epistemology of Anonymous Assertions.” According to Sanford Goldberg,3 there are three central problems with the epistemic status of anonymous assertions. First, hearers do not have access to the kind of information on which necessary credibility assessments of speakers are based. Second, and because of this, hearers lack the sort of counterfactual sensitivity to defeaters, or counterevidence, that responsible hearers need. And, finally, since it is clear that hearers can’t hold speakers responsible for making epistemically defective assertions, speakers lose their motivation to avoid making them. Ivy argues that Goldberg’s third concern depends on a “punitive model of assertoric behavior.” According to this model, speakers generally assert properly only because hearers have the ability to hold speakers accountable for their assertions and are able to punish them in some sense when they assert improperly. Ivy argues against this model by showing that we assert truthfully even when there are no punitive disincentives for lying, and we feel badly for asserting falsely even when there are no punitive consequences. In this way, Ivy claims that it is typical for speakers to exhibit pro-social behav iors—speakers tend to experience an internal duty of honesty and truthfulness and this is what drives assertoric practices. In contrast to Goldberg, then, Ivy maintains that speakers should adopt a default attitude of trust toward assertions, including those that are anonymous. This trust is defeasible: if speakers have strong positive reasons to doubt the assertion, then these reasons serve as defeat ers. However, Ivy argues that anonymity itself is not a defeater and, thus, that doubting anonymous assertions is the exception, not the rule. Finally, while using the social media platform Yik Yak as a model—where fully anonymous unknown assertions are made—Ivy takes up Goldberg’s first two concerns and argues that hearers can form justified, testimony-based beliefs from anonymous assertions on either of the main views in the epistemology of testimony.
3 Goldberg (2013, 2015).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
16 Jennifer Lackey
References Dotson, Kristie (2011). “Tracking Epistemic Violence, Tracking Practices of Silencing.” Hypatia 26: 236–57. Fricker, Miranda (2007). Epistemic Injustice: Power & the Ethics of Knowing. Oxford: Oxford University Press. Goldberg, Sanford (2013). “Anonymous Assertions.” Episteme 10: 134–51. Goldberg, Sanford (2015). Assertion: On the Philosophical Significance of Assertoric Speech. Oxford: Oxford University Press.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
PART 2
E PIST E MOLO G IC A L PE R SPE C TI V E S
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 19/03/21, SPi
2
When Freeing Your Mind Isn’t Enough Framework Approaches to Social Transformation and Its Discontents Kristie Dotson and Ezgi Sertler
But enlightenment does not result in real freedom, or even a mental state of pleasure. (Angela Davis 1998, “Unfinished Lecture on Liberation—II,” 58)
1. Introduction As it becomes more and more popular to view social justice issues through epistemological approaches, it has become important to take stock of what such approaches hope to accomplish. We define “epistemological approaches to social justice” as approaches to social and political problems that highlight epistemic features of those problems. Framework approaches to social problems can be defined as a particular intellectual approach where framework transformation is taken to equate changes in social arrangements. In this chapter, we probe the transformative potential of framework approaches to social justice. We have a broad question which we are attempting to pose and respond to in the course of the chapter. Here we are asking, “What relationships do framework shifts have to social transformation?” We understand “social transformation” in modest terms as any significant shift in social arrangements. There are at least two ways to understand this opening question. First, we might be posing the question as to the plausibility of the idea that a framework shift has a radical causal impact on actual social arrangements. Or, second, we might be querying whether it is viable to assume that a framework shift is, itself, a shift in social circumstances. In this chapter, we are concerned primarily with the first of these two interpret ations. Specifically, we probe the basis of the assumption that shifting frameworks have a profound impact on social transformation. And we are asking this ques tion because we have our doubts about the plausibility of such an assumption. Our doubts do not follow from the judgement that framework changes are unim portant for social transformation. In fact, we might concede that, in a certain Kristie Dotson and Ezgi Sertler, When Freeing Your Mind Isn’t Enough: Framework Approaches to Social Transformation and Its Discontents In: Applied Epistemology. Edited by: Jennifer Lackey, Oxford University Press (2021). © Kristie Dotson and Ezgi Sertler. DOI: 10.1093/oso/9780198833659.003.0002
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
20 Kristie Dotson and Ezgi Sertler sense, a framework shift around a given issue constitutes a change in social circumstance. What we doubt is that such a shift is a change in social arrangements. In short, we are concerned about a potential propensity to intellectualize social transformation so that one might assume that enlightenment is real freedom (Davis 1998, 58).1 In what follows, we outline a particular approach to social transformation that implies or assumes a thick causal connection between frameworks and social transformation that, for us, needs to be reconsidered. To execute this analysis, we introduce a range of concepts that aid in illuminating the limitations of a particu lar position towards a thick relationship between malfunctioning frameworks and social arrangements. We challenge the idea that framework shifts at different levels equate to changes in the social arrangements they aim to reconceptualize. Ultimately, we claim that framework approaches to social transformation have two limitations, which include: (i) failing to lead to the epistemological ingenuity they often promise; and even where such ingenuity might be achieved, (ii) leaving untouched the actual social arrangements that facilitate the circumstances under analysis. This chapter proceeds in three further sections. In section 2, we explain what we mean by a framework approach to social transformation. In section 3, we take an example of a framework approach to social justice in the framework approaches to understanding “political prisoners” and its potential aims and aspirations. In this section, we gesture to the limitations of such approaches in articulating social circumstances as modes of social transformation. Finally, in section 4, we conclude by responding to a potential objection for this framework analysis assessing the “work” of framework analyses.
2. A Framework Approach to Social Transformation Before we offer an example of a framework approach to social transformation and what it may or may not accomplish, we will explain what we take to be a frame work approach to social transformation and the terms we will rely upon through out the course of this chapter. As such, this section proceeds in two parts. First, we explain our terminology, i.e. first-, second-, and third-order frameworks, conceptual resilience, social arrangements, social circumstances, and social transformation. Second, we briefly articulate a particular intellectualist approach 1 Here, we refer to an assumption Angela Davis (1998) criticizes in the “Unfinished Lecture on Liberation—II.” In that article, Davis discusses how Frederick Douglass “has arrived at a consciousness of his predicament as a slave” (57). This consciousness for Davis is “at the same time a rejection of his predicament” (57–8). However, this consciousness or this enlightenment “does not result in real freedom, or even a mental state of pleasure” (58). Davis writes: “Referring to his mistress, Douglass says: ‘She aimed to keep me ignorant, and I resolved to know, although knowledge only increased my misery’ ” (58).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Framework Approaches to Social Transformation 21 to social transformation, i.e. changes in social arrangements as simply framework transformations.
2.1 Terminology A framework, here, refers to an interpretative schema that filters the “world out there” with processes of selection that generate a schedule of salience that actively maintains and supports the framework itself. In this, we borrow from Snow and Benford (1992), when they write that a frame is “an interpretative schema that simplifies and condenses the ‘world out there’ by selectively punctuating and encoding objects, situations, events, experiences, and sequences of actions within one’s present or past environment” (e.g. frames to organize the presentations of opinions and facts) (Snow and Benford 1992, as cited in Boykoff 2007, 31). Frameworks are often used to center the points of emphasis that serve to produce understandings of a given phenomenon or a domain of inquiry. As Entman (1993) explains, framing “involves selection and salience. To frame is to select some aspects of a perceived reality and make them more salient . . . in such a way as to promote a particular problem definition, causal interpretation, moral evaluation, and/or treatment recommendation for the item described” (52). As such, high lighting frameworks as a point of departure is to interrogate what features and emphases aid in generating “understandings” or “knowledge” about some given social phenomenon. And frameworks can be said to generate understandings and/ or knowledge at several levels of abstraction. For our purposes, we will introduce three levels of inquiry with respect to frameworks as a kind of engaged epistemol ogy, where one is articulating and interrogating frameworks for comprehension or knowledge production of some domain of inquiry. They are first-, second-, and third-level framework investigations. First-level frameworks represent those frameworks that emerge from what Hortense Spillers (1984) calls first-order discourse, which expresses “the experi ences of a community” in real time (89). In other words, a first-level framework is simply a framework in operation as it is used to generate understandings of one’s circumstances or some domain of inquiry. Second-level frameworks attempt to isolate and make sense of the ways first-level frameworks are constructed so that they render reasonable or “com mon sense” targeted features of those worlds. These frameworks primarily focus on ranges of relevance and schedules of salience that operate like selection criteria and meaning-making devices for narrating a given phenomenon, happening, or structure. That is to say, a second-level framework analysis targets first-level frameworks embedded in first-level discourse for articulation and analysis. Third-level frameworks, which most people identify as, broadly speaking, epi stemic, attempt to identify what kinds of broad orientations facilitate the ways
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
22 Kristie Dotson and Ezgi Sertler both the ranges of relevance and the schedules of salience are organized in second-level framework analysis. These frameworks attempt to organize and otherwise understand how second-level frameworks are designed and outlined. As one can guess, a framework analysis is infinitely regressive. For example, our analysis here might be understood as a third-level framework analysis. We are concerned with the potential of second-level frameworks to disappear actual social arrangements. There is an open question whether the changes that result from framework shifts are merely framework deep. That is to say, changing the way we understand social arrangements, for example, may only change our understanding, not the social arrangements themselves. There are usually one of two assumptions concerning goals that underlie framework analyses when they are executed for the sake of social justice ends. They include: (i) the assumption that framework analyses have the potential to change social arrangements; and/or (ii) the assumption that framework analyses have the potential to alter social circumstances. By social arrangement, we mean the actual structures and situations in our social, political, and institutional envir onments about which frameworks are attempting to produce knowledge and/or understandings. Frameworks with respect to our social worlds, on our account, are, in part, attempting to offer understandings and/or knowledge about social arrangements. As a result, social arrangements are not simply frameworks; they exceed frameworks and are, in many ways, the point of generating frameworks at all. For many, poor frameworks can hinder our understanding and knowledge of social arrangements. Some make the strong assumption, however, that social arrangements are causally tied to prevailing frameworks. For others, however, the aim of a framework analysis for social justice is weaker. They may hope that altering a framework concerning some social justice issue can aid in changing targeted social circumstances. By social circumstances, here, we are referring to some so-called “fact” or condition related to some event, arrangement, or happening. That is to say, a social circumstance is directly related to understandings of the social arrangements themselves or the kinds of entail ments we imagine resulting from existing social arrangements. As a result, some assume that a shift in framework can have the potential of shifting social circum stances with respect to some given social arrangement so that what makes sense about that arrangement is challenged or transformed. What is important to note is that there are strong and weak aims for frame work analyses as modes of social justice engagement. The strong aim includes an assumption that framework shifts are correlated in some strong sense with poten tial transformation in social arrangements. In the weaker sense, one makes an assumption that framework shifts are correlated with potential transformations in social circumstances. In our estimation, it is hardly surprising that one avenue of engaged epistemology takes the form of framework analyses for shifts in either social arrangements or social circumstances. In academic circles, there tends to
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Framework Approaches to Social Transformation 23 be stronger and weaker senses that “writing” and texts change the world. This is one way to make “sense” of that sense. Our principal example of a framework discussion that aims at shifting our understanding of social arrangements, where this shift in understanding equates to a shift in the social arrangements themselves (i.e. the strong assumption) con cerns attempts to understand and outline the definition and situations of “polit ical prisoners.” One of the reasons the framework analyses become so attractive, specifically to the “academic” social justice worker, is that there is an assumption that framework shifts either have the potential to disrupt or discontinue a distor tive framework analysis or that framework shifts can play a large causal role in shifting the social arrangements in question.2 Part of what might be assumed as a “doing” in framework analyses is either directly or potentially compromising the conceptual resilience that frameworks are thought to take on and, in the strong sense, through that disruption having the potential to disrupt the actual state of affairs. To say that a framework (or a set of frameworks) has taken on conceptual resilience is to say that it has a commonly recognized domain of stability that is difficult to disrupt, (i.e. that the framework(s) in question ‘just makes sense’ to people). The hope of some framework analyses is to disrupt the “sense” some frameworks make so as to make the everyday strange and open to critical reflection. That is to say, framework analyses may be thought to either directly disrupt a domain of stability or shrink the conceptual area within which a given domain is stable. Our opening inquiry, restated with these terms in mind, is: “Does a second-order framework analysis have the potential to directly disrupt or limit the domain of stability of a first-level framework, and are these ends also trans formations of actual, social arrangements?” This question is aimed, for us, at querying the presumed constitutive role of frameworks of social arrangements. The broadness and difficulty of understanding social arrangements, for example social and political structures, often prompts one to be careful and con siderate about the frameworks one deploys in understanding them as structures. This carefully considered reflection on frameworks, for us, is often executed in second-level framework analyses. For example, when discussing the ways dissent is framed and criminalized, Jules Boykoff (2007) explains: frames not only overlap and reinforce each other, but also frequently compete . . . On one level, coverage of dissidence can be seen as a framing contest whereby different social actors and groups present their frames in an effort to gain social 2 One could easily imagine an analysis of our analyses that operates at another level of framework analysis: either first, second, third, or a fourth level of analysis. We don’t find this problematic. Part of our concern is a concern over what we imagine framework analysis to be doing (at whatever level), when it is thought to be “doing” anything at all. As such, we will need to navigate what we imagine to be “doing” in our own framework analysis and will do so in our conclusion.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
24 Kristie Dotson and Ezgi Sertler currency, the contested topography of public discourse. However, at the end of the day, the mass media collectively serve as the arbiter of these framing contests by implementing and synthesizing their own frames. (32)
What often happens, in public discourse around social justice issues, like the importance and preservation of dissent, are framework analyses of prevailing frameworks. First- level frameworks are factored through media that then generate second-level frames that ultimately adjudicate a “framing contest”, i.e. which frame will win the day. A second-level framework analysis, in some domains, can be cast as judge and jury and as having the potential to effectively determine the “state” of a discourse. We want to ask, however, what is the goal of a second-level framework approach to transformation for social arrangements themselves? Not for our understanding of them, but for the actual structures and realities we are attempting to understand or produce knowledge about? There seems to be an assumption that such framework analyses can counter problematic first-level frameworks via attempts to disrupt or challenge the potential or actual resilience of a given framework. We suppose that part of the rationale for this understanding of the “work” of second-level discourse analysis concerns the idea that “what you don’t know can hurt you.” But we are uncertain what kind of work, “knowing” differently through different frameworks is actually doing, in terms of transformation of social arrangements, i.e. the strong assumption of the aim of framework analyses. Transformation of social arrangements, or social transformation, here refers to concrete changes in the structures and realities that are being “framed,” where the actual structures or social arrangements are not reducible to the frameworks because the frameworks never fully illuminate the structures in question.
2.2 Framework Analysis as Social Transformation Again, what we are probing in this chapter is the notion that a framework analysis has a direct relationship to social transformation, i.e. transformations of actual social arrangements. Both authors have run into academics in extended fashion and in passing who assume that thinking about things differently will auto matically result in changes in social arrangements or the social, political, and institutional realities one is attempting to understand or produce knowledge about. That is to say, we have both run into the strong assumption that framework transformation is simply social transformation. There are several challenges we would like to pose to this assumption. First, we think that proponents of this approach to framework activism need to consider that what one knows does not mean that one knows better. That is to say, feeling like one has a better under standing of social arrangements or structures does not mean that one actually has
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Framework Approaches to Social Transformation 25 a transformative understanding (however we define such a thing). For instance, understanding what the prison industrial complex is, how it functions, and what its problems are (a better understanding) is not adequate for “imagining alterna tives” to prisons or “envisioning” a social order without prisons (a transformative understanding) (Davis 2003, 112, 10; Davis 2005; Heiner and Tyson 2017). Becoming more aware does not equate “epistemological ingenuity” (Heiner and Tyson 2017, 4). Epistemological ingenuity, here, refers to the ability to construct understandings and knowledge beyond what is given with different frameworks that are better than the faulty framework one identifies. Second, we also maintain that what you know can still hurt you because social arrangements are not con structed of frameworks, even if our understandings of them are so constructed. In short, targeting frameworks that legitimate3 unjust structures leaves relatively untouched the material circumstances that may facilitate the social arrangements in question. The second-level framework approach might be necessary for knowing differently (or similarly) than a first-level framework allows, but such an analysis does not necessarily create the conditions necessary for imagining alternative frameworks. Specifically, we wonder what “work” framework analyses do and what such analyses leave undone and undertheorized. We are probing this line of inquiry because we also want to understand whether framework analyses strengthen or weaken the structures they attempt to target. In what follows, we will construct a case concerning the landscape of frame work analysis around “political prisoners” which is often taken to be a discourse aimed at liberating “political prisoners.” What we found is that there are at least two ways in which second-level frameworks are used to imagine liberatory ends, i.e. a disaggregation approach and a comprehensive approach. After outlining these two second-order framework approaches, we articulate how they may fall short of disrupting the kind of conceptual resilience they identify around the con struction of “political prisoners” and, more importantly, how difficult it is to see how the framework transformation as social transformation assumption is sup posed to be borne out given that it is a kind of epistemic intervention in a social, political, and historical situation.
3. Example: “Political Prisoners”/Political Imprisonment Discussion We consider an analysis of political imprisonment to be a second- level framework analysis that targets the framework around imprisonment with 3 This legitimation can occur for already-established structures, as well as structures that are in the process of being established and/or reinforced/stabilized. (We thank our reviewer for clarifying this point.)
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
26 Kristie Dotson and Ezgi Sertler political motivations. This is because an analysis of political imprisonment usually aims to disclose how the current social arrangements are constructed in a way to legitimize imprisonment with political motivations in different ways. While doing so, an analysis of political imprisonment hopes to determine the state of the d iscourse on imprisonment in a way that challenges the legitimation of criminalization with political motivations. In other words, a key part of the political imprisonment discourse lies in a conceptual effort to relocate ‘certain groups of people’ beyond how the current social arrangements have defined them. And this relocation involves generating a new understanding of certain incarcerated people as political. As a framework analysis that hopes to create a first-level framework shift (i.e. generate a new understanding/critique of who “political prisoners” are), the polit ical imprisonment discourse can either have strong or weak aims as a mode of social justice engagement. This is to say that a discussion of political imprison ment can see the framework shift it creates as a potential transformation in either social arrangements or social circumstances. In its effort to generate a new under standing of certain incarcerated people as political, the political imprisonment discourse criticizes the social circumstances these “prisoners” are in by challen ging how the existing social arrangements define, produce, and legitimate these circumstances. What we want to investigate here is the possible strong aim the political imprisonment discourse as a framework analysis can have. We want to question whether these framework analyses have the potential to transform social arrangements through the first-level framework shifts they achieve. In other words, we would like to ask whether understanding “political prisoners” differ ently can result in concrete changes in the structures or social arrangements that are being framed. In what follows, we identify two different approaches within a discussion of political imprisonment. We will call the first approach the “disaggregation approach,” and the second one the “comprehensive approach.” These two approaches suggest two different second-level framework analyses which try to establish different definitions of “political prisoners” and different ways of delegitimizing imprison ment with political motivations. We argue that both approaches, in different ways, demonstrate how second-level framework approaches to social transformation (i.e. framework transformation as a change in social arrangements) have two limitations: (i) they may not lead to epistemological ingenuity at which they aim; and (ii) they leave untouched the actual social arrangements that facilitate the circumstances under analysis.
3.1 Disaggregation Approach As McEvoy et al. (2007) note, one of the central concerns of the political imprisonment discussion is “the ways in which such prisoners are defined” (293).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Framework Approaches to Social Transformation 27 The question of how to define ‘what a political prisoner is’ or ‘what a political crime is’ is crucial for the discussion because it forms the foundation through which the actions of the imprisoning state can be contested. In other words, deciding what constitutes ‘the political’ in front of the moniker “prisoner” forms the basis on which the political imprisonment discussion challenges the ways in which imprisonment of political crime is legitimized. As a second-level frame work approach, then, an analysis of “political prisoners” hopes to determine the state of the discourse for imprisonment by suggesting definitions of certain incar cerated people as “politically motivated” or by underlining the political nature of their acts. We call this effort to decide ‘who counts as a political prisoner’ the disaggregation approach. As the name suggests, a disaggregation approach within an analysis of political imprisonment intends to determine ‘who a political prisoner is’ and therefore intends to separate the category of “political prisoners” from other kinds of “prisoners.” There are many different disaggregation approaches within political imprisonment analyses: some are more detailed than others. In “Political Imprisonment and the ‘War on Terror,’ ” McEvoy et al. (2007) suggest “five broad and sometimes overlapping categories of inmates as ‘political prisoners’ ” that, we think, is adequate for our purposes here: “1. prisoners of war, 2. ‘Prisoners of con science’, 3. Conscientious objectors, 4. Radicalized ‘ordinary’ prisoners, and 5. Politically motivated prisoners” (294). This listing is an attempt to disaggregate different kinds of “prisoners” in order to isolate which sets of “prisoners” are unduly politically motivated.4 Prisoners of war mostly refer to combatants captured as a result of conflict, and they have been an important discussion within international humanitarian law. Even though the Geneva Conventions, while trying to regulate international interstate conflicts, provide a definition of prisoners of war, “who does or does not qualify as a POW [prisoner of war] has remained highly contested” in practice (McEvoy et al. 2007, 295). In addition, McEvoy et al. (2007) highlight that when states deal with internal conflicts, they tend to sidestep the Geneva Conventions because they usually categorize internal conflicts as insurrections and would like to resolve them without any regard to international humanitarian law. As they further note, the notion of “prisoners of war” seems also important for cases such as the War on Terror: “The US administration in particular has invested considerable energy in denying applicability of the Geneva Conventions and redefining those detained under the War on Terror as something other than POWs [prisoners of war]” (2007, 295).
4 As far as we understand, disaggregation approach offers a new definitional framework precisely to separate those who can be defined as political prisoners from “regular” prisoners. This approach is (usually) offered with a critical and/or liberatory point of view, where the states’ treatments of these kinds of prisoners are criticized, and their conditions are contested.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
28 Kristie Dotson and Ezgi Sertler Prisoners of conscience, first coined by Amnesty International in 1961, refers to people who have not “used or advocated violence but is imprisoned because of who they are (sexual orientation, ethnic, national or social origin, language, birth, colour, sex or economic status) or what they believe (religious, political or other conscientiously held beliefs)” (www.amnesty.org/en/what- we- do/detention). Conscientious objectors, a category that closely overlaps with prisoners of conscience,5 refers to the cases where people object to mandatory military service on grounds of conscience. How the category of conscientious objectors is characterized and whom the conscientious objector status can be granted to depend on how national criminal codes structure their military service, their alternatives to military service, and their “permissible parameters of conscientious objection” (McEvoy et al. 2007, 297). For instance, in cases where a conscientious objector is defined as “an individual who objects to war per se or the use of violence in any form”, troops who refuse to serve due to “a religious or moral objection to a particular war” are less likely to be categorized as conscientious objectors (McEvoy et al. 2007, 297). Radicalized ‘ordinary’ prisoners refer to “individuals imprisoned for non-political offences but who become radicalized while in prison” (McEvoy et al. 2007, 297). McEvoy et al. (2007) suggest that “the transformation of ordinary black prisoners into political militants” can be used as a prime example of this category (298). This transformation usually occurs when people start seeing their imprisonment as a result of an “oppressive politico-economic order” and become “conscious of the causes underlying their victimization” (Davis 1998, 47).6 The final category McEvoy et al. suggest is politically motivated prisoners. While admitting the fact that this term can apply to all the cases mentioned above, McEvoy et al. (2007) note that it has been recently adopted as a more neutral terminology than ‘terrorists’, considering the baggage the words ‘terrorists’ and ‘terrorism’ carry today (299). As we can see from this brief discussion, a disaggregation approach tries to provide tools to categorize certain incarcerated people as political. And one cen tral aim of this categorization is to question, and perhaps contest, how the state in question manages/imprisons them. In other words, a disaggregation approach, by establishing categories of political prisoners versus others, tries to criticize how imprisonment becomes the way in which states manage and answer to political motivations. Our question here is not “how to define ‘political prisoners.’ ” Instead, we are interested in “what defining ‘political prisoners’ as a separate category” (first-level framework shift articulated through a second-level framework analysis) can do. We wonder whether generating a new understanding of “prisoners” who are “political” while shifting our understanding of the current social arrangements, has the potential to accelerate a shift in the social arrangements themselves 5 McEvoy et al. (2007), for instance, note that “Amnesty International recognized members of the US military who were jailed for refusing to serve in Iraq as prisoners of conscience” (297). 6 Also see Nagel (2015, 44).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Framework Approaches to Social Transformation 29 (in this case, causing concrete changes in how “political prisoners” are managed by the states in question). We want to demonstrate why this potential has limita tions due to the conceptual resilience of carceral logics and its practices. By carceral logics, we refer to a dominant and an epistemologically resilient logic that legitimizes only one intelligible schema of understanding and account ing for how serious harms can be redressed or prevented. This schema legitimizes state-centric practices of incarceration that are exercised through prisons which regulate punitive exclusions by disciplinary containment. This logic is employed and supported by, and supports and strengthens, the practices and policies of states and their entities and apparatuses. It can further establish societies’ reliance on incarceration and confine its capacities for thinking, feeling, imagining, and acting (Heiner and Tyson 2017; Kim 2015). As an epistemologically resilient logic that has conceptually resilient frame works that uphold it, carceral logics are sense-making devices. As an epistemo logically resilient logic/system, it “upholds” and “preserves” our sense-making mechanisms or conceptually resilient frameworks. Dotson (2014) argues that forms of theoretical resilience has two factors: “the scope of the domain for stabil ity and the magnitude of disturbance required to motivate significant change” (132). This is to say that in an attempt aiming to challenge an epistemologically resilient logic via conceptually resilient frameworks one needs to either directly disrupt a domain of stability or shrink the stable area of the given epistemo logical system. What a disaggregation approach tries to achieve is to limit where and to whom carceral logics can apply or who a carceral state can legitimately imprison. As a second-level framework approach, the disaggregation approach tries to shrink the conceptual area of carceral logics. It does so by introducing a category of “uncommon criminal,” through the definition of “political prisoners,” therefore suggesting a subtraction of those “criminals” from where the carceral logic applies to. In other words, by defining “political prisoners” and what constitutes political imprisonment, the disaggregation approach destabilizes a framework where imprisonment systematically applies to political motivations. However, the intro duction of the “uncommon criminal,” or the “political prisoners” as a separate category at the same time tends to restabilize the carceral logics by suggesting a concept of “common criminals.” This is to say that by introducing a category that is “wrongly” managed or “wrongly” imprisoned, the disaggregation approach seems to suggest that the punitive approach of carceral logics should not directly apply to “political prisoners.” This suggestion, at the same time, reinforces or strengthens how the punitive approach is used towards other “prisoners.” This implicit reinforcement of carceral logics suggests that a disaggregation approach is not a direct disruption to carceral logics or the existing social arrange ments embodying it. This situation becomes clearer when we look at how a second-level framework analysis through the disaggregation approach calls for
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
30 Kristie Dotson and Ezgi Sertler other second-level framework analyses for other frameworks that are currently controlled by carceral logics and employed by the existing social arrangements. When we look at McEvoy et al.’s (2007) discussion of separate categories of “political prisoners,” for instance, we can see that arguing for a new understand ing of political does not make it quite clear “who a ‘prisoner of war’ is” in practice, ‘who counts as violent,’ ‘what counts as a violent action,’ ‘what it means to be non-violent as a person but be part of an organization that committed violence,’ ‘what constitutes objection, dissidence, moral objection, etc.’ That is, a framework shift aimed at shrinking the domain of stability for a way of understanding imprisonment and incarceration may only shift the terms of discussion, without prompting any changes to the social arrangements in question. In fact, we can see that an introduction of the category of “political prisoners” does not guarantee a situation where the carceral state will not simply utilize the newly introduced framework to legitimate persisting and unchanged social arrangements. Thus, we can see that the disaggregation approach as a second-level framework analysis targets the imprisonment discourse by suggesting that its legitimation of imprisoning political actors needs to be questioned. However, this suggestion, while shrinking the area where the carceral logics apply, which is a kind of shift in social circumstances, does not create the magnitude of disturbance that is neces sary to disrupt carceral logics, that are entailed by the social arrangement of mass incarceration itself.7 The resilience of carceral logics becomes clear when we real ize first, that the category of “political prisoners” strengthens the assumption of correctly imprisoned people and second, that ‘knowing political prisoners’ differ ently does not guarantee knowing ‘violence, objection, dissidence, etc.’ differently. This means that when one conceptual domain shrinks (i.e. disaggregated “prison ers”), another expands to take up the slack, thereby leaving resilient carceral logics intact. This expansion is non-accidental. It can be predicted as the sense incarceration continues to make of itself as an epistemologically resilient logic that, itself, generates varying conceptual resiliency. That is, the legitimation forces of carceral logics and the social arrangements they “make sense of ” are still at work. That is why, we think, the potential of a second-level framework analysis as a disaggregation approach fails to achieve epistemological resilience. As such, such efforts never quite reach the epistemological ingenuity some imagine as the result of their efforts. That is to say, it is unclear that the shift to identify “uncom mon criminals” is better than general criminalization, where better than is the 7 In other words, this scope-shrinking strategy is at the core of the disaggregation approach we identify here. This is to say that an approach that tries to identify tensions and contradictions in the logic in order to disrupt it and contest the underlying logic itself and not just its scope would not be categorized under a disaggregation approach. Furthermore, we think that identifying tensions and contradictions in a logic does not automatically lead to disrupting that logic. As we argue in this chap ter, disrupting logics is a matter of structures as well. For instance, identifying the tensions and contra dictions in the logics that structure social arrangements does not lead to disrupting those logics precisely because functioning social arrangements prevent those logics from being disrupted.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Framework Approaches to Social Transformation 31 institution of a framework that disrupts epistemological resilience of carceral logics. The social arrangements of incarceration, where accountability for serious harm is largely understood in terms of confinement and deprivation, are undis turbed with the “uncommon criminal” class; after all, there are still plenty of “common criminals” to lock up. It is according to this kind of assessment that we wonder how a framework shift that fails to achieve epistemological ingenuity has a hope of catalyzing a shift in the actual social arrangements that have epistemo logically resilient and widespread logics that coincide with the actual social arrangements, which are capable of employing and refiguring conceptually resili ent frameworks indefinitely without ever changing their material constitution.8
3.2 Comprehensive Approach The comprehensive approach is another second-level framework analysis that tries to establish a different definition of “political prisoners” and therefore a dif ferent way of delegitimizing imprisonment of political motivations. The compre hensive approach refers to the view that all “prisoners” are in some sense “political prisoners” because of the fact that the nature of imprisonment is political. Acknowledging the role imprisonment plays in sustaining and maintaining the social order demonstrates how imprisonment in general is heavily politicized and how the nature of crime is political as well (McEvoy et al. 2007, 294). For instance, when we look at how states approach particular kinds of behavior, as McEvoy et al. (2007) suggest, we realize that they tend to criminalize the behaviors of the weak and the poor, while at the same time “condoning” or even “encouraging” the behaviors of the powerful and the rich (294). As Rodriguez (2006) also argues, we have to see the relationship between imprisonment and prisons and the legal framework they function in as part of a “broader process of social ordering designed to sustain hegemonic definitions of right and wrong as well as maintain the existing social order and dominant forms of class and race relations” (Rodriguez 2006, as cited in McEvoy et al. 2007, 294). While McEvoy et al. (2007) mention that the approach of “All crime is political crime” has been criticized and has been attributed a limited analytical utility,9 the 8 One of the assumptions that may underwrite a second-level framework analysis is that a resilient wide-spread “logic,” e.g. discursive way of understanding current happenings, institutions, and sys tems, is formed by one conceptually stable domain that can be shrunk or expanded. This, it seems to us, can be doubted. Or, at the very least, one has too simple an understanding of conceptual domains so that a definitional intervention looks like it causes more changes than it actually hopes to do. 9 They mention both the critique that viewing criminals as all political in nature overlooks/disre gards the victims of those crimes “who were often themselves the poor, women or other vulnerable groups” and the critique that in cases such as Northern Ireland, “where disputes concerning the political character of inmates were quite literally matters of life and death,” the idea that all crime is political crime was not analytically helpful to understanding the specifics of these situations (McEvoy et al. 2007, 294).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
32 Kristie Dotson and Ezgi Sertler c omprehensive approach is still visible in a context where mass incarceration and prison industrial complexes are pressing realities. As a second-level framework analysis, the comprehensive approach aims for a different shift in our understanding of the existing social arrangements. Unlike the disaggregation approach, the comprehensive approach, as a second-level framework analysis, tries to challenge the epistemological resilience of carceral logics through a direct disruption. Its definition of political imprisonment, or the new understanding of “prisoners” it suggests does not aim to shrink the concep tual area, but rather aims for direct disruption of carceral logics’ domain of stabil ity. This is because claiming that ‘all imprisonment is political’ disputes the legitimacy of imprisonment in general and contests how carceral logics estab lishes that legitimacy. By doing so, the comprehensive approach challenges the punitive capacity of the state by rendering it illegitimate. In other words, if all incarcerated people are political, the way in which a state punishes gets delegitim ized. This is how the comprehensive approach hopes to challenge the moves of legitimation that the carceral logics/state has established. However, we argue that this disruption still might fail to accelerate a shift in the actual social arrange ments due to the material realities of the prison industrial complex. Judith Resnik (2010) argues in “Detention, the War on Terror, and the Federal Courts” that the current carceral logics employed by the United States has both preventive and punitive confinement regimes (673). Resnik, while analyzing the “judicial responses to the central challenges, faced daily by governments trying to maintain peace and security,” highlights how many countries have responded to terror by “detaining individuals preventively” (584, 679). “Around the world,” Resnik notes, “countries authorize incapacitation for ‘public protection’ based on an array of grounds—illegal immigration, sexual predatory behavior, heinous criminal actions, terrorist threats—that undermine the presumption that it is conviction and punishment that is required for incarceration” (Resnik 2010, 679). In other words, this authority to incapacitate, previously exercised in the presence of conviction and punishment, can now be exercised by claiming that even an uncertainty about whether a person might inflict, or will inflict, harm to the social order/national security/peace can form a “reasonable ground” for author ities to license forms of preventive detention. It is this exercise of detaining indi viduals preventively as opposed to punitively that accelerated in the context of 9/11 (also in the context of criminalization of border-crossing, conflation of migrants with terrorists, and criminalizing and silencing political dissent) and that challenges the comprehensive approach’s aim for disruption (Resnik 2010; Chang 2002; Fekete 2004). When the comprehensive approach renders the punitive confinement regime of the current imprisonment system irrelevant (through redefining political imprisonment), it does not quite touch upon the preventive aspect of it. Thus, a move to delegitimize the punitive aspect of a resilient system does not result in
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Framework Approaches to Social Transformation 33 delegitimizing the preventive aspect of it, which in turn can legitimize the confinement of all “political prisoners” or everyone imprisoned currently or future detainees. As Resnik (2010) suggests, the current carceral logics and the actual social arrangements employing it have stepped beyond the idea that confinement is only used for punitive measures. This becomes clear when we look at how the “uncertainty about which persons have done or will do harm” (584) does not pre vent the current system from confining individuals. The carceral logics, as it func tions and makes sense right now, can justify/justifies a preventive confinement/ detention. In other words, the current social arrangements and its carceral logics seem to have the capacity to justify “All prisoners are political prisoners” and “All prisoners are preventatively confined as political prisoners”, while never trans forming its material functioning. This indicates the difficulty of epistemological ingenuity, even when successful, to render change of actual social arrangements. Furthermore, when legitimizing detaining people “who are suspected to do something,” states do not have to categorize them as prisoners. This is because preventive detention seems to allow state structures to use carceral logics in a way that is adapted to manage new groups of people. As a result, even if states accept that all prisoners are political prisoners, they can still continue categorizing these (preventively detained) people as not fully prisoners, and thus evade the connec tion between preventive detention and political imprisonment. Epistemologically resilient systems that support and are supported by the current structures or social arrangements can absorb an extraordinary amount of disturbance while maintaining business-as-usual operations.10 Both disaggregation and comprehensive approaches to political imprisonment can be categorized as second-level framework analyses that aim to shift our understanding of the current social arrangements. In other words, both approaches hope to generate a new understanding of certain “prisoners” as political by dem onstrating and challenging how the current social arrangements legitimate imprisonment of political motivations. These framework analyses can be, and are, used as modes of social justice engagement. When used as modes of social justice engagement, framework analyses can have stronger or weaker claims with respect to what kinds of shifts or transformation they hope to achieve. We have discussed above that it seems quite difficult to translate the shift these framework analyses achieve in our understanding to a shift in the actual social arrangements them selves. This is particularly due to the capacity of epistemologically resilient logics 10 It seems to us that it is not quite clear how the statement “All prisoners are political prisoners” and a framework approach centered around that statement can answer to the introduction of prevent ive detention as a new way of categorizing people. This is because, on the one hand, carceral logics located in the existing state structures seems to have the capacity to justify “All prisoners are political prisoners” and “All prisoners are preventatively confined as political prisoners.” On the other hand, by categorizing preventively detained people as not really prisoners, carceral logics can also allow the existing state structures to evade the connection between political imprisonment and preventive detention.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
34 Kristie Dotson and Ezgi Sertler and the capacity of structures employing them to absorb certain framework shifts without an actual transformation. It seems important to think about this capacity given the increasing criminalization of dissent (in various forms) and political rhetoric of security.
4. Conclusion: An Objection Epistemological resilience, which is underwritten by actual, social arrangements, is not primarily constituted of conceptually resilient frameworks, where if one changes the framework one changes the epistemological resilience. Epistemological resilience, in our estimation, is shaped by material, social arrangements that gen erate domains of stability within which conceptual resilience frameworks either make sense or do not. That is, epistemologically resilient logics are made of struc tural realities. And epistemologically resilience logics often make sense of and promote already existing structural realities that are neither reducible to those structures nor exhausted by them. As a result, epistemic resources, like frame works, can change without the epistemological systems, or the social arrange ments generating them, changing. It is according to these understandings that we find the strong aim of framework analysis for social transformation dubious, at best, and to be rejected, at worst. Intellectual approaches to transformation in social arrangements, alone, neither change the structural realities of confinement nor do they change the epistemological resilience of carceral logics.11 By way of conclusion, we respond to one potential objection to our, primarily, third-order framework discussion in this chapter. If framework analyses do not have the potential to effect changes in social arrangements, then why execute this analysis at all? Here is where we lay our card on the table and admit that we sub scribe to some version of the weak aim for framework analysis. It is entirely likely that though enlightenment is not “real freedom,” it is nonetheless a step towards that goal. A change in conceptual framework is, for us, a change in social circum stances and has been known to effect real changes for individuals within oppres sive systems, such as the prison industrial complex. We think here of the successful 11 In fact, as Joy James argues, intellectual approaches can further “deflect from real structures of oppression” (1996, 52). In discussing structural racism, for instance, she highlights that: racialized identity and speech are endemic to the United States. Yet a focus on these alone deflects from the political and economic aspects of structural racism and white supremacy. Whether or not anything is publicly said—and no matter how one racially self-identifies— policies perpetuate dominance and genocide. Racism has come to be understood as a “form of discourse . . . that can be effectively blocked by means of linguistic taboos;” as racial epithets become taboo, so does antiracist terminology. (James 1996, 49) This is to say that a mere focus on words, discourses, and frameworks in order to build a “critique” of state violence overlooks what is required beyond “literary insurgency or rhetorical resistance” when confronting state violence (James 1996, 23).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Framework Approaches to Social Transformation 35 and important campaign to free Angela Davis as a “political prisoner.” But this change in social circumstances was not also a change in social arrangements. So, though the modest aim of framework analyses (i.e. changes in social circum stances) is difficult for us to contest, the strong aim of framework analyses (i.e. simultaneous changes in social arrangements) is difficult to defend.12
References Boykoff, Jules (2007). Beyond Bullets: The Suppression of Dissent in the United States. Chico, CA: AK Press. Chang, Nancy (2002). Silencing Political Dissent: How Post September 11 Anti-Terrorism Measures Threaten Our Civil Liberties. New York: Seven Stories Press. Davis, Angela (1998). The Angela Y. Davis Reader. Edited by Joy James. Malden, MA: Blackwell Publishers. Davis, Angela (2003). Are Prisons Obsolete? New York: Seven Stories Press. Davis, Angela (2005). Abolition Democracy: Beyond Empire, Prisons, and Torture. New York: Seven Stories Press. Dotson, Kristie (2014). “Conceptualizing Epistemic Oppression.” Social Epistemology: A Journal of Knowledge, Culture and Policy 28(2): 115–38. Entman, Robert M. (1993). “Framing: Toward Clarification of a Fractured Paradigm.” Journal of Communication 43(4): 51–8. Fekete, Liz (2004). “Anti-Muslim Racism and the European Security State.” Race & Class 46(1): 3–29. Heiner, Brady and Sarah Tyson (2017). “Feminism and the Carceral State: GenderResponsive Justice, Community Accountability, and the Epistemology of Antiviolence.” Feminist Philosophical Quarterly 3(1). Article 3, http://ir.lib.uwo.ca/ fpq/vol3/iss1/3.
12 In this chapter, we aimed to discuss framework directed approaches and how weak they seem compared to the social arrangements in place. Different approaches one might identify through Joy James and Angela Davis might seem like stronger versions of these approaches. However, we think that James’s or Davis’s approaches to social transformation are precisely not framework-directed approaches. Their approaches criticize the framework-directed approaches and their prioritization over collective organizing where the confrontation with state policies, resistance, and organizing prac tices are seen as way more valuable than an emphasis on frameworks. In other words, in James’s or Davis’s discussions, framework-directed approaches are only valuable when they are part of an activist framework that ‘materially’ engages with states’ practices and its violence. This is not to say that both James and Davis overlook the importance of building ‘good’ or ‘better’ rhetoric. However, they both seem to underline the importance of how critiquing bad rhetoric, unless it is embedded in practices of collective organizing, does not amount to much, and in fact could deflect from structures of oppres sion and their material consequences (see also Davis 2005). We thank our reviewer for calling our attention to the differences and similarities between approaches to social transformation one might identify in Joy James and Angela Davis and framework-directed approaches.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
36 Kristie Dotson and Ezgi Sertler James, Joy (1996). Resisting State Violence: Radicalism, Gender, and Race in the U.S. Culture. Minneapolis, MN: University of Minnesota Press. Kim, Mimi (2015). “Dancing the Carceral Creep: The Anti-Domestic Violence Movement and the Paradoxical Pursuit of Criminalization, 1973–1986.” Institute for the Study of Societal Issues, University of California Berkeley (October 14, 2015), http://escholarship.org/uc/item/804227k6. McEvoy, Kieran, Kirsten McConnachie, and Ruth Jamieson (2007). “Political Imprisonment and the ‘War on Terror’, ” in Yvonne Jewkes (ed.), Handbook on Prisons. Portland, OR: Willan Publishing, 293–323. Nagel, Mechthild (2015). “Angela Y Davis and Assata Shakur as Women Outlaws: Resisting U.S. State Violence.” Wagadu: A Journal of Transnational Women’s & Gender Studies 13: 43–78. Resnik, Judith (2010). “Detention, the War on Terror, and the Federal Courts.” Yale Law School Faculty Scholarship Series, Paper 678, 579–685, http://digitalcommons. law.yale.edu/fss_papers/678. Rodriguez, Dylan (2006). Forced Passages: Imprisoned Radical Intellectuals and the US Prison Regime. Minneapolis, MN: University of Minnesota Press. Snow, David A. and Robert D. Benford (1992). “Master Frames and Cycles of Protest,” in Aldon D. Morris and Carol McClurg (eds), Frontiers in Social Movement Theory,. New Haven, CT: Yale University Press, 133–55. Spillers, Hortense J. (1984). “Interstices: A Small Drama of Words,” in Carole S. Vance (ed.), Pleasure and Danger: Exploring Female Sexuality. London: Routledge & Kegan Paul, 73–100.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 19/03/21, SPi
3
Situated Knowledge, Purity, and Moral Panic Quill R. Kukla
In this chapter, I will argue that a great deal of contemporary epistemology is driven by a kind of moral panic1 over the worry that there are no ‘pure’ epistemic practices, perspectives, or standards detachable from the social situation of knowers. Epistemology is currently caught in a dialectic: From various quarters, different strands of social epistemology aim to show in different ways that various traditional epistemological notions—including objectivity, justification, warrant, and knowledge—are in fact ineliminably situated. Meanwhile, a host of traditionalists attempt to fend off each encroachment of situatedness, and to mark off a space of epistemic purity. They do so by, on the one hand, carving out an eversmaller and ever-more-elaborately-defined space, within which values, identities, contexts, and social dependencies purportedly do not tread; and, on the other hand, warning of the dire threats to the possibility of knowledge and rationality that we face if we don’t fend off these incursions. I believe that epistemic practices are ineliminably situated in multiple ways, and that this is not to be feared. We cannot do epistemology without fundamental, central attention to social identities, power relations, and the social institutions and structures within which epistemic practices happen. But this result is of no threat to our usable notions of objectivity, justification, and the like. Indeed, I will argue that most of the claims made by defenders of situated knowledge are not only true but commonsensically and uncontentiously true, although they have often been presented as more radical and counterintuitive than they actually are. The dialectic I described above is driven in large part by fear rather than intellectual tension. The quest for purity is unnecessary, and our goal should not be to counter each tactical attempt to protect it with elaborate counterarguments, but rather to recognize the fear as a product of ideology and to become comfortable with situatedness as an everyday phenomenon. 1 The term ‘moral panic’ is generally attributed to Stanley Cohen, from his classic 1973 work, Folk Devils and Moral Panics. A moral panic is a broadly distributed social phenomenon, wherein a specific source of purported risk comes to be seen as a severe threat to our basic social order and security. Moral panics are characterized by simplistic causal stories, risk distortion, and a moralistic, characterological condemnation of groups. Quill R. Kukla, Situated Knowledge, Purity, and Moral Panic In: Applied Epistemology. Edited by: Jennifer Lackey, Oxford University Press (2021). © Quill R. Kukla. DOI: 10.1093/oso/9780198833659.003.0003
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
38 Quill R. Kukla My primary goal here is not to hash out the details of various positions, but rather to explore the rhetorical metanarrative I have just described. I will end the chapter by arguing that a proper naturalized, non-ideal epistemology—one more continuous with the empirical sciences—will treat situatedness not as something spooky or epistemologically threatening, but just as an empirical fact about our epistemic practices. For epistemic practices are, after all, natural practices performed by finite beings; they are part of the material world, just like everything else.
1. The Quest for Certainty: A Cautionary Tale I begin with an historical analogy. For centuries, philosophers obsessed over certainty. It seemed obvious that unless they could find a domain of absolutely unshakeable knowledge, all our epistemic practices would be hopelessly insecure and unmoored. Hence the central project was to find risk-free, error-proof standards of knowledge. Kant’s response to these assaults on certainty was to focus on discovering the boundaries within which knowledge could be secure, and to carefully remain within those boundaries. Beyond them lay chaos and within them lay purity and safety. Kant argues that we should use no methods or principles that cross those boundaries, as tempting as it will always be to venture beyond them: We have now not only traveled through the land of pure understanding, and carefully inspected each part of it, but we have also surveyed it, and determined the place for each thing in it. This land, however, is an island, and enclosed in unalterable boundaries by nature itself. It is the land of truth (a charming name) surrounded by a broad and stormy ocean, the true seat of illusion, where many a fog bank and rapidly melting iceberg pretend to be new lands and, ceaselessly deceiving with empty hopes the voyager looking around for new discoveries, entwine him in adventures from which he can never escape and yet also never bring to an end . . . [We must remember] by what title we occupy even this land, and can hold it securely against all hostile claims. (1998, Critique of Pure Reason B294–5)
Kant here describes the project of critical epistemology as a project rooted in anxiety and the need for homeland security. Hegel makes this fear of error the centerpiece of his critique of Kant. He argues that Kant presents his critical project as a neutral exercise in maximal rigor, one that assumes nothing about which methods are epistemologically safe, whereas in fact it is driven by pathological risk aversion.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Situated Knowledge, Purity, and Moral Panic 39 It is a natural assumption that in philosophy, before we start to deal with its proper subject matter, viz. the actual cognition of what truly is, one must first of all come to an understanding about cognition . . . A certain uneasiness seems justified, partly because there are different types of cognition, and one of them might be more appropriate than another for the attainment of this goal, so that we make a bad choice of means; and partly because cognition is a faculty of def inite kind and scope, and thus, without a more precise definition of its nature and limits, we might grasp clouds of error instead of the heaven of truth. (Hegel 1973, Phenomenology of Spirit, para. 73)
Hegel argues that the Kantian fear of error and insecurity reduces epistemological practice to boundary policing and method-checking, with no space left for building positive knowledge. He claims that the critical method is not only unproductive and grounded in neurosis; it is also question-begging. In its claim to neutrally and rigorously trust no methods without critical examination, it presupposes that the fear-based, boundary-policing method itself needs no examination. In Hegel’s view, this is just one more method that is fallible and open to critique like any other, and furthermore it is one that prevents us from obtaining knowledge, since doing so requires that we accept epistemic risk: If fear of falling into error sets up a mistrust of Science, which in the absence of such scruples gets on with the work itself and actually cognizes something, it is hard to see why we should not turn around and mistrust this very mistrust. Should we not be concerned as to whether this fear of error is not just the error itself? Indeed this fear takes something—a great deal in fact—for granted as truth, supporting its scruples and inferences on what is itself in need of prior scrutiny to see if it is true . . . [This is] an assumption whereby what calls itself fear of error reveals itself rather as fear of the truth. (Hegel 1973, Phenomenology of Spirit, para. 74)
Eventually Hegel argues that it is in fact mistakes that are the engine of epistemic progress, so a risk-free approach will lead us to a dead end. Thus, Kantian risk aversion is self-defeating. Hegel did not end the obsession with infallibility and fear of error in epistem ology; Nietzsche and others continued the critique. But this dialectic was not won by way of definitive arguments. Eventually, we just kind of got over the fear. There are still infalliblists, of course, but they are a somewhat peculiar minority. Most of the major debates in epistemology now just take it for granted that knowledge is generally fallible. Epistemology stopped worrying about certainty and moved on, taking risky epistemic practices as its assumed topic. Currently vibrant debates in
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
40 Quill R. Kukla epistemology over the status of epistemic intuitions, over lottery problems, the metaphysics of probabilistic knowledge, defeaters, and so forth, all take falliblism for granted without exhibiting undue stress. I claim that our fear of situatedness is closely analogous to our former fear of error. Our obsession with carving out a safe domain of aperspectival, value-fee knowledge is a different kind of critical epistemology focused on boundary policing, and it is similarly a neurosis that we need to just overcome, rather than continuing to try to beat it down with arguments. Situatedness does not undermine the point of our epistemic practices any more than did fallibility. And our reasons for insisting on carving out a safe domain of this sort are analogously question begging. We need to just get on with the project of doing epistemology, under the assumption that it will matter who is doing the knowing, with what interests and investments, and from what social position.
2. Three Types of Situatedness Several strands of contemporary epistemology push against the possibility of pure, unsituated knowledge. These can be divided into three categories, which I will consider in turn. One set of arguments aims to establish the social dependence of knowledge; these arguments purport to show that knowledge is not merely a product of disciplined individual rationality, but also of good-quality social organization and social epistemic luck. A second set of arguments is designed to reveal the ineliminable value-laden character of our epistemic standards and practices; according to such arguments, there are no objective epistemic standards and methods that don’t embed personal and social values and interests. A third set of arguments is directed at showing the essentially perspectival nature of our epistemic notions, such as justification and warrant; these aim to show that who an epistemic agent is and how she is socially positioned irreducibly help constitute her access to evidence, justifications, and knowledge.
2.1 Social Dependence and Epistemic Ecology According to this set of arguments, a great deal of our knowledge is based on our having the right sorts of connections to others and being located properly within an effective organizational system. For instance, much of science is massively co-authored and multidisciplinary at this point. Individual scientists are not in a position to understand or track the contributions of all of their collaborators, who often have very different skill sets and expertise and between whom contact may be minimal (Huebner et al. 2017; Kukla 2012; Goodchild 2006; Yong 2017).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Situated Knowledge, Purity, and Moral Panic 41 In turn, consumers of scientific results must trust the organizational and institutional system within which the results were produced, rather than particular individuals or their own reconstructions of the work. One strand of philosophy of science and social epistemology focuses on understanding teams or communities as the loci and agents of knowledge, rather than individuals (Galison 2003; Tollefsen 2007; Wray 2006, 2007). At the same time, we consume more and more of our information from online sources generated by unknown others. We each live in an online ‘filter bubble’, as the results of our searches and the feeds on our social media pages are algorithmic ally tailored to show us information and interpretations to which we will be receptive, and this has epistemic consequences (Miller and Record 2013; Lynch 2016; Nguyen 2018). Sorting out what to believe is not just a matter of exercising good judgment; to some extent our access to evidence is out of our control and a matter of epistemic luck. One might think that these are contingent features of contemporary science and technology, rather than deep challenges to our epistemic self-sufficiency, and hence that they don’t undermine the possibility of pure epistemic practices unsullied by social dependencies. But first, this isn’t a very useful form of purity, given that many of the things we care about knowing we can know only in these ways. Second, the radical dependencies involved in scientific knowledge and online knowledge are arguably only vivid intensifications of an inevitable kind of mundane epistemic dependency on other people and on social organizations. To repurpose an example from John Haugeland (1997), when, for instance, I say that I know how to get from Washington DC to New York, I mean that I know how to marshal the knowledge of the road makers, or the track layers, along with the train engineers, or the traffic light designers, and so forth. Almost all my abilities to cope with the world and to interpret information involve my dependence on the epistemic skills and contributions of others. All these arguments point to the view that having good-quality knowledge can’t be reduced to a matter of individual epistemic self- discipline—it also requires being fortuitously positioned within what we might think of as an epi stemic ecosystem,2 in which people have complementary and situationally appropriate expertise, are trustworthy, are collaborating in the right way, and are embedded in an infrastructure and organization conducive to accuracy, in which information flows appropriately. This is true both in the formal epistemic domain of science and the informal domain of everyday epistemic competence. Of course, few epistemologists, if any, would deny the role of social embeddedness and moral luck in people’s actual ability to acquire knowledge. But the dominant 2 The first epistemologist to analyze explicitly this notion of epistemic ecologies was Lorraine Code (2006).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
42 Quill R. Kukla tradition in epistemology, inherited at least from Descartes, is methodologically individualist, and focuses on how individuals can have the right kinds of cognitive states and make the right kinds of inferences to hook them to the world in the right way. The arguments from social dependency and ecological embeddedness suggest that we won’t learn much about how good epistemic practices actually work from these individualist analyses.
2.2 Values and Interests in Justification Lively conversations in epistemology and in philosophy of science have converged on the idea that standards of justification (or sometimes of knowledge, warrant, or hypothesis acceptance) are ineliminably shaped by interests and values; accordingly, there are no such things as ‘pure’ canons of rationality and inference, which are legitimate independently of the normative situation of the epistemic agent. In epistemology, pragmatic encroachment theorists and contextualists have argued that whether an agent has (fallible) empirical knowledge depends on more than the evidence available to her; it also depends on her stakes in being right. The two groups locate the essential role of interests or values in slightly different places, but both agree that there is generally no answer to whether a maximally rational person with access to a body of evidence for P knows that P, independent of how much it matters to her that she be right that P (DeRose 1992; Fantl and McGrath 2009; Hawthorne and Stanley 2008; Ichikawa 2017). The general form of the argument is that higher the stake someone has in a proposition being true, the higher the justification bar will be for knowledge of that proposition. So, for example, the theory goes, on an average day I may count as knowing that today, as per usual, there is a 4:00 pm Acela Express train leaving New York City for Washington DC, just because I have the weekly schedule memorized. But if it suddenly becomes crucial that I make it to DC today before 8:00 pm, say, because my son has life-threatening surgery scheduled there at that time, then I will need better evidence than my general knowledge of the schedule in order to count as having knowledge. In this circumstance, I no longer know that this is when the train is going, and I will need to check online or ask the station manager or take other steps in order to have this knowledge. The point is that our empir ical epistemic standards (of justification, knowledge, etc.) are indelibly infected by values and interests that vary from knower to knower. Another passenger’s identical memory of the weekly schedule may count as knowledge for her, even while I don’t count as knowing, because justification is interest relative. The more it matters to me that I be right, and correspondingly the worse it would be for me if I were wrong, the higher the evidence bar for knowledge will be. Thus, which beliefs count as knowledge is sensitive to the practical environment. There are no perspective-free standards of evidence.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Situated Knowledge, Purity, and Moral Panic 43 A parallel discussion in philosophy of science concerns ‘inductive risk.’ According to the inductive risk theorist, the puzzle runs this way: Inferences from evidence to hypothesis acceptance or non-acceptance are always uncertain, and hence any inference comes along with a specific sort of risk—that is, the risk of a false positive (accepting a hypothesis that is actually false), which goes up as our evidence bar goes down, and the risk of a false negative (failing to accept a hypothesis that is actually true), which goes up as our evidence bar goes up. No evidence bar can ever be the ‘safe’ choice, because the risk of false negatives and false positives are constitutively inversely correlated.3 Hence the only thing that can settle where our evidence bar should be is the badness of various kinds of mistakes. This, in turn, can only be settled with reference to values and interests. So, for instance, if mosquito netting has the prospect of saving millions from malaria at low cost and with no physical risk, the risk of a false negative outweighs the risk of a false positive, and our evidence bar for its effectiveness should be relatively low. If a new drug for an already treatable condition promises only modest benefits over the current treatment, the evidence bar for its safety and efficacy should be very high; our primary concern should be to avoid false positives. Inductive risk theorists conclude that the standards for hypothesis acceptance depend ineliminably on the values and interests of the epistemic agent. Furthermore, Heather Douglas (2000) and others (Steel 2010; Wilholt 2009, 2013; Winsberg 2012; Biddle and Kukla 2017) have argued that such interest-dependent judgments occur throughout the research process. How data are classified and coded, which sorts of screening tests are used, which methods are employed in smoothing and correcting data, and indefinitely many other judgments involve this type of inductive risk balancing. For example, whether researchers classify slides of rat tumors as benign or malignant turns out to depend, to a very large degree, on the goal of the study and their stake in the outcome; pathologists in industry-funded studies overwhelmingly set the interpret ive bar much higher for seeing a malignancy in the same slides than do pathologists in government-funded studies (Douglas 2000). In such situations, there is no value-free notion of a correct choice of distribution of inductive risks, since any choice will involve trade-offs. Likewise, there is no value-free notion of “the correct epistemic standard.” We cannot section off the purely epistemic from other social and personal values and interests in scientific inference. The pragmatic encroachment theorists focus on individuals and when their beliefs count as knowledge; they want to know how values infect traditional epi stemic concepts like ‘justification.’ The inductive risk literature focuses not on the belief states of researchers, but on the act of accepting a hypothesis, which has
3 See Rudner (1953) and Hempel (1965) for classic formulations of the puzzle.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
44 Quill R. Kukla psychological components but also behavioral, social, and institutional components. The inductive risk theorists ask practical coordination questions, such as: “How does the infection of scientific inference with values affect our collective ability to trust and use scientific claims made by others?”; “How can we regulate science so that the role of values is not distorting or problematic, since we cannot eliminate that role?” But the general focus in both literatures is on the rejection of ‘pure’ standards of rationality and the recognition that good reasoning is inherently infused by and relative to socially variant interests. As Boaz Miller puts it: The bottom line [in both literatures] is that when research outcomes finally reach the context of application, they are already saturated with social value judgments, and reflect the various trade-offs between values that were made in the process of inquiry leading to them. Moreover, the indirect influence of social values is not a necessary evil that must be tolerated. Rather, it is necessary for achieving scientific objectivity because only it offers a non-arbitrary, principled, and relevant way. (Miller 2014, 261)
2.3 The Standpoint Dependence of Evidence and Justification “Standpoint theory” (or “standpoint epistemology”) began as a sub-discussion within feminist epistemology in the 1970s and ’80s (Smith 1974; Hartsock 1983, Harding 1986, 1992) and has grown to become a major approach within social epistemology. It has kept its feminist roots but expanded into a more general set of questions and approaches concerning the difference that social identity and position make to our epistemic agency—to what evidence is available to us, to how we can legitimately use that evidence in inference, and to what we can know. Standpoint theory has no single definition and its contours are hotly contested. But, roughly following a classic article by Wylie (2003), I define it as characterized by two theses: (1) Social location systematically influences our experiences, and shapes what we can know; knowledge is always (or often) achieved from a standpoint. (2) Marginalized standpoints provide epistemic advantages in at least some contexts. These two theses together lead to a third cornerstone of standpoint theory: (3) ‘Aperspectival’ knowledge is a mythical ideal, and good-quality reflection on our epistemic practices and justifications requires critical attention to who is doing the epistemic work, and from what social location.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Situated Knowledge, Purity, and Moral Panic 45 This critical attention to the identity of the knower, rather than just to the methods and standards they use to know, is what Sandra Harding calls “strong objectivity.”4 Strong objectivity, as a measure of epistemic quality, assesses the structured standpoints from which various people engage in epistemic practices, including the strengths and the limitations of those standpoints. According to this measure, an epistemic product such as a belief or an accepted hypothesis needs to be evaluated in light of its justification or warrant in the context of the standpoint from which it was generated, including evaluation of the quality and appropriateness of the standpoint to the epistemic task at hand. Strong objectivity also requires attention to the organizational features of knowledge production, which will include or exclude the voices of those with standpoints that are especially helpful. The standpoint epistemologist worries not just about how well positioned individuals are, on their own, to know various things, but also about how to organize institutions so as to enhance the production of good-quality, objective knowledge by bringing in the right people in the right ways, in the right relations to one another. This means that our epistemic practices are doubly ‘sullied’ by ineliminably social factors: No individual is a ‘pure’ aperspectival knower, and furthermore, we cannot do critical epistemology except by looking at social formations and organizations larger than individuals (Longino 1990; Solomon 2001; Biddle 2007). Why should we believe theses (1) and (2)? My goal here is not to launch a new or comprehensive defense of them. Rather, I would like to suggest that they are in fact commonsensically, mundanely true. There is nothing surprising or spooky about them. This doesn’t mean that there is no room for philosophers to poke holes in them. Instead, my claim is that they are commonsensical enough that the philosophical tendency to see them as radical claims in need of vigorous attack or defense is a product of ideology and moral panic, rather than of reasonable concern. Let’s consider some examples. Remember that the point of the examples is to be mundane rather than philosophically quirky or dramatic. (1) If I want to overhaul a space—a public building, perhaps—to make it as accessible as possible, trying to think abstractly about possible barriers is going to be dramatically more difficult than bringing in people with disabilities. Living in a body that is often misfit with and obstructed by the space around it gives one concrete, objective, thick-bandwidth knowledge. Asking for input from or collaborating with disabled people is clearly an epistemically efficient choice in this scenario. Trying to figure out how to make a space accessible without their input is likely to be clumsy and partial. People with mobility impairments don’t just
4 Harding’s strong objectivity is structurally analogous to Kant’s critical epistemology, which requires us to analytically examine the structure of the knower in order to judge the security of the knowledge. The difference, of course, is that Kant explicitly argues that there are necessary and universal features of knowers that are not socially variant (the categories and the forms of intuition).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
46 Quill R. Kukla have access to more facts about accessibility (though this may be true as well)— they have access to different, more flexible, more visceral, more immediate evidence than others can have. They can directly perceive barriers, openings, spaces, and so forth in different ways. Hansel Baumann, an architect specializing in Deaf architecture who helps design spaces at Gallaudet University, explains how Deaf students require broader hallways than hearing students, so that they can comfortably walk side by side and see one another as they walk. If one person has to drop behind another while walking, all conversation must cease. This makes complete sense once you know it. However, Gallaudet was originally designed to look like a mini-Ivy League campus, with old fashioned buildings with narrow halls. That this is a bad design is completely obvious from within a Deaf standpoint and easy to miss from outside of one (Baumann 2018).5 (2) Women are more likely to see sexism than men; people of color are more likely to see racism than white people; trans folks are more likely to see transphobia than cis folks. Why? Well, first of all, bigoted and discriminatory acts are more likely to happen around them, precisely because they are there! But also, since in most (but not all) contexts, bigotry is socially unacceptable, bigots are likely to hide their discriminatory views and behaviors from their ingroup. It is not accidental, nor a matter of ‘hypersensitivity’ or ‘paranoia’, that members of stigmatized groups have more and more detailed evidence concerning the bigotry that’s out there than do members of dominant groups, precisely because they are its target.6 (3) Living in the midst of a dense, diverse urban area inculcates a distinctive set of epistemic skills for moving through complex, diverse, sensorily rich spaces. City dwellers’ embodied patterns and habits have to be flexible, responsive, and adaptable, as we make our way through crowds, form queues, dart between people and across busy streets, efficiently hail cabs and negotiate subway stations, and interact at least in passing with a wide variety of people who move, speak, and occupy space very differently from one another. We can understand and distinguish between more accents. We know how to touch others appropriately in busy areas and how to avoid being touched when needed. None of these skills and perceptual capacities can simply be imparted by way of propositional assertions. The epistemic advantages that city-dwellers have are quite concrete. One can easily watch how fluently and efficiently an urban resident moves through their city and among its people, with minimal stress or hesitation, compared to a flustered tourist who will move slowly and haltingly and exhibit anxiety. (4) Standpoint theorists have traditionally emphasized the epistemic advantages that come from having a culturally and politically marginalized social 5 Interestingly, caring intimately for a person with mobility challenges can also give one a different kind of lived understanding and perception of barriers and spaces (Kukla 2007). 6 See, for instance, McKinnon 2015, who makes this point with respect to access to transphobia.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Situated Knowledge, Purity, and Moral Panic 47 position. But sometimes a marginalized standpoint comes with epistemic disadvantages. Consider the epistemic challenges that economically disadvantaged students face if they go to ‘good’ colleges.7 Even controlling for SAT scores and other traditional markers of academic potential, poor students often struggle to succeed at college in distinctive ways. They face direct epistemic impediments to navigating their classes and absorbing knowledge as well as their peers. For instance, if they have not grown up with the same habituated practices around money, lifestyle, and social relations as their peers, they are less likely to fit in well and learn from the kind of fluid socializing that characterizes college at its best. If they do not come from a family or neighborhood where most people have gone to college, they will know less about how to pick classes that suit their interests, how to talk to professors, what office hours are for, and so forth, and hence they will likely get less epistemic benefit out of being at college. They also face indirect epistemic impediments. For instance, they may be less able to take intellectual risks if their ability to afford college depends on their maintaining a near-perfect record of good grades, and they may be less able to buy optional books and materials. Implicit and explicit bias on the part of their professors may kick in if they ‘talk like a poor kid’, regardless of the sophistication of their comments.8 For all of these reasons, an economically disadvantaged student may well get less epi stemic benefit from college than will a wealthier classmate who is exactly as smart and academically well prepared as they are. Before going on I want to make two points about these examples. First, again, all of them may seem like cases of epistemic advantage or disadvantage that are too commonsensical or mundane to be of philosophical interest. They are just examples of how being a certain kind of person with certain experiences goes along with knowing some things, and not knowing others. Surely this is obvious, and not philosophically radical or challenging! But this is precisely my point. Remember that my goal is less to win arguments against sophisticated and elab orate defenders of aperspectivalism and epistemic purity, and instead to make the situated alternative seem everyday and unthreatening and not worth arguing against. Second, and crucially, each is a case where the epistemic advantage or disadvantage cannot be reduced to knowing (or not knowing) extra facts of the sort that can just be imparted via propositional assertions. There is no easy epi stemic substitute, in any of these cases, for actually living with a certain kind of materially and socially situated body. At the same time, they are not examples of mere physical skill either: In all four cases, embodied experience enables or
7 See, for instance, Madden 2014. 8 Or professors’ bias may actually kick in because she has sophisticated things to say. See, for instance, Martínez 2016. This piece focuses on the author’s Latina ethnicity but presenting as working class can pose similar risks to one’s epistemic credibility.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
48 Quill R. Kukla impedes epistemic access—it shapes what sort of perceptual evidence is available to the subject and what sort of inferences they are positioned to draw. I defined standpoint theory by two claims: (1) Social location systematically influences our experiences, and shapes what we can know; knowledge is always (or often) achieved from a standpoint. (2) Marginalized standpoints provide epistemic advantages in at least some contexts. I have argued that the first is not just true, but mundanely so. Turning to the second: Why would marginalization provide a distinctive epistemic advantage (especially given the fact that it can sometimes produce epistemic disadvantages as well)? At least three plausible reasons show up in the literature. First, Patricia Hill-Collins (1986), among others, has emphasized the epistemic advantages of having an ‘insider/outsider’ status. The materially incarnated social world is generally built for the convenience of the privileged, almost by definition, and the dominant world view is generally their world view. This means they are rarely pushed to take an outside perspective on their own stance or practices and how they are contingent, partial, or excluding. A recent egregious case of this was Michael Tooley, the septuagenarian straight, white man who chaired the University of Colorado philosophy department in 2014, when it was the site of multiple sexual harassment scandals. Tooley (2014) writes: I have been in half a dozen philosophy departments over the course of my career, and it does not seem to me that female members of those departments were treated differently in any way than male members. I did not, for example, see any differences between, on the one hand, the way in which male philo sophers interacted with female philosophers and, on the other, the way in which they interacted with each other. Nor did I see any prejudice against women faculty when it came to decisions to hire, to tenure, or to promote, or against female students when it came to admission to graduate school.
He floats the possibility that “maybe [he is] just an insensitive observer” but dismisses this worry as implausible or irrelevant. This sort of unjustified universal ization of one’s own perspective and failure to consider one’s own likely epistemic limitations is a distinctive weakness of those at the center. Those on the margins don’t have the luxury of remaining in their ‘own world’ and unjustly universalizing and totalizing their perspective, since their world is not independent, but constituted and subordinated by the dominant culture. Thus, the marginalized must in effect become ‘bilingual.’ They can see more clearly the workings and limitations of the dominant norms and culture from the outside, while also of
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Situated Knowledge, Purity, and Moral Panic 49 necessity becoming skilled at operating within it. They often have what Du Bois (1903) called ‘double consciousness,’ or the “sense of always looking at one’s self through the eyes of others.” This double vision places the self in social context. It is often socially painful, but it can be epistemically productive. Second and relatedly, those at the center often get no serious pushback against their views, since they have the power to present them as absolute. We all know older white men who have been spouting the same line about politics, metaphysics, or whatever it is for decades, because their voices have enough social authority that they can just talk obliviously over dissenters, whether literally or metaphorically. This credibility excess denies them the epistemic opportunity for their views to evolve when forced to grapple with other perspectives. Thus, privil ege can sustain ignorance. More generally, the ability to move through the social world and engage in social interactions with minimal risk can sustain ignorance. For example, those with normative social identities have less need to be able to read other people’s identities—others’ ethnicity or sexual orientation, perhaps— because their interactions with those different from them don’t put them at risk in the same way that they do for the marginalized. This can lead to the privileged being surprisingly unable to pick up on certain kinds of social communication and signaling. Third, Patricia Hill-Collins explored the epistemic advantages that come from being able to move relatively invisibly through privileged space. In Hill-Collins’s fictional example, a housekeeper solves a murder that baffles others by having this kind of access to facts not meant for ‘people’ to know and see, as she moves effect ively invisibly among the privileged people who employ her. Alison Wylie (2003) picks up on this example and makes this kind of standpoint advantage central. People will selectively reveal and conceal things from their peers, while the woman in the back cleaning the floor, emptying the trash, or refilling the coffee may get to see and hear things that ‘people’ were never supposed to hear. She sees the empty bottle of anti-depressant drugs in the VP’s trashcan, or the sudden look of exhaustion on a junior partner’s face once his superior leaves the room. A variety of techniques have been proposed to dodge or guard against the incursion of values and perspectives into the ‘core’ of knowledge. Often the strategy is to divide up epistemic practices, and to insist that only one part is ‘infected’ with values or standpoints, while the other, pure part is the real epistemic part. For example, in fending off the threat of pragmatic encroachment, Brian Weatherson writes, “I agree, evidentialism [that is, the position that only evidence determines justification] is false. And I agree that there are counterexamples to evidentialism from subjects who are in different practical situations. What I don’t agree with is that we learn much about the role of pragmatic factors in epistemology properly defined from these counterexamples to evidentialism.” He concedes that if two people have the same degrees of belief, it may count as a
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
50 Quill R. Kukla belief for one but not for the other. But this, he claims, is a philosophy of mind thesis, and not an epistemic matter: pragmatic context, in his view, affects whether you count as believing, and not whether you have a justification that makes this belief into knowledge (Weatherson 2005, 418). In a different divide-and-protect move made in the same spirit, Richard Jeffrey responded to inductive risk arguments by insisting that inferring from data and accepting or rejecting hypotheses is not part of science; scientists are just supposed to collect data and produce statistical information, which policy makers and others embroiled in value-and interest-laden projects can then use to make inferences. By his own description, “This account bears no resemblance to our ordinary conception of science” (Jeffrey 1956, 246). In the style of Kant, Jeffrey and Weatherson thus both shrink and strengthen the boundaries around what counts as the proper epistemic domain, in order to keep it pure. In all of the examples I offer above, there is still room for philosophers to try to save a role for pure, aperspectival epistemic practice. I have given no arguments that foreclose that possibility. But at a certain point, we need to ask why we are philosophically invested in working so hard to find some sort of pure, aperspectival epistemic cubicle or sanctuary. What do we gain by so carefully narrowing the domain of the properly epistemic (or the properly scientific) so as to excise dependency on situation, values, perspective, and interests? Why is there such philosophical resistance to these many and varied ways that epistemic practices and standards appear to be conditioned by situation?
3. Purity and Moral Panic I propose that the various divide-and-protect strategies are driven in no small part by ideological anxiety and moral panic. This is not intended as a total explan ation, nor as an invalidation or refutation of the many sophisticated attempts to parse out exactly how and to what extent epistemic standards can be situation independent. But I think it is a key force driving the dialectic of contemporary philosophical conversations. I will discuss four common categories of responses to the various arguments for situated knowledge, each of which is, in my view, primarily driven by moral panic. Some of these are responses that come directly from epistemology, and others circulate in the culture at large and shape the implicit background. Some are anxieties expressed by insiders who themselves defend some version of a situated knowledge thesis. What they have in common is a general anxiety around the idea of giving up on there being aperspectival, position-independent ways of knowing, and concomitantly on insisting upon the ineliminable relevance of the identity and social position of the knower to the assessment of her epistemic practices.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Situated Knowledge, Purity, and Moral Panic 51 (1) If we assign epistemic advantages to marginalized knowers, and insist on critical attention to who produces knowledge, we will devalue the great canon of Western literature, philosophy, and science. We must block these arguments in order to save the canon. (2) If knowledge is situated, it will be relativism and identity politics all the way down! We won’t have any unsituated, aperspectival stance from which to assess which standpoints are better than others, and so-called knowledge will turn into a free-for-all in which the loudest voices win. (3) Standpoint theories and their ilk essentialize and reify social identities, and hence they cannot accommodate human diversity and intersectionality. (4) We just need to work harder and more rigorously to carve out the bound aries around a pure space of epistemic practice unmarked by situation, because that’s where the true core of epistemology lies.
3.1 The Argument from Devaluing the Canon This first concern, familiar from countless op-eds and cultural commentators, does not require a serious philosophical response, and I take it that its roots in moral panic—and white male fragility—are clear. If attention to the epistemic work done by marginalized knowers ends up unseating the primacy of privileged knowers, this doesn’t constitute any kind of epistemic argument against it. Acknowledging and critically interrogating the fact that our traditional canonical texts were produced from particular situated perspectives only undermines their legitimacy if we assume that only aperspectival knowledge is valuable. But that’s exactly what’s being called into question. Furthermore, even if we end up needing to acknowledge that the standard ‘great thinkers’ had epistemic limitations caused by their social situation and identity, this doesn’t make them valueless; thoughts and texts don’t have to be perfect to be valuable. And if we decide that some of these texts, once put in context, really do have much less value than we thought, this still isn’t an argument against situated knowledge—it’s at worst a blow to our sentimental attachments. If the purity of the canon is challenged, and if other voices are valued in part because they are other voices, the fear is that this will somehow undercut the value or the primacy of a traditional set of texts—and with it, of course, a specific, privileged perspective marked by colonialism, race, and gender. By acknowledging that this is one perspective among many, we remove its automatic epistemic privilege and its claim to purity, and this is threatening to many. Recently, Reed undergraduates protested the Eurocentrism and male- dominance of their
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
52 Quill R. Kukla required first-year Humanities class. Of course, they were not objecting to the inclusion of white men on the syllabus, but rather calling for a more inclusive set of perspectives. The protests were covered in Inside Higher Education (Flaherty 2017), and virtually all of the comments were critical. A couple of typical ones read: My sense of things is that anything to do with ‘Western Culture’ is considered white and patriarchal by definition . . . If Western Civ. is inherently evil, it’s easy to throw out the whole package (free speech, due process, representative constitutional government) as a relic of old dead racist men. (Comment by CET, 11 September 2017) This is checkbox diversity: The students just want to have a certain number of authors who fall into the construct of “people of color.” I doubt that they are well-acquainted with the works they deride (and they might not be acquainted with any texts at all). (Comment by John Ahearn, 10 September 2017)
In these comments, we see the assumption that the valuation of alternative perspectives and their importance directly implies a devaluation of more traditional perspectives. But, of course, there is no reason why the first should imply the second. This is a widespread cultural anxiety rather than a philosophical thesis, yet it is widely reflected within the discipline, in suspicious and dismissive attitudes towards attempts to diversify traditional philosophical syllabi, for instance, as well as in undefended claims that of course students ‘must’ read Plato, Kant, etc., before moving on to non-Western works or works in feminist philosophy, or the like. My argument is not that reading the canonical texts is valueless, by any means. But the common assumption that no defense is needed for their special priority over less traditional texts and perspectives reflects ideology, as assumptions that seem to need no defense typically do. The underlying logic here is that valuing alternative perspectives poses a threat to philosophical knowledge and standards, rather than just an expansion and refinement of them.
3.2 The Argument from Fear of Identity Politics The essential argument from fear of identity politics proceeds as follows: (1) If there is no such thing as a pure, unmarked stance unshaped by situ ation, then there is no epistemically secure position from which to assess the epistemic quality of any other stance. (2) But if this is right, then no stance has any claim on being better than any other.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Situated Knowledge, Purity, and Moral Panic 53 (3) And if so, then epistemic practices and disagreements are no more than power plays, in which different groups engage in political jockeying over who gets to say what counts as true. (4) Finally, if epistemic disagreement becomes this kind of brute power war, women, trans folks, and people of color will win out over white cis men and define reality in ways that benefit them. Not every argument expressing epistemic alarm at the threat of identity politics has all these elements, but this basic line of thinking and its parts show up in various places. For instance, in a much-discussed blog piece, “The Problem of Me Studies,” Joseph Heath argues that people who speak from and about their own position of oppression or marginalization block anyone else from talking because everyone is scared to hurt their feelings (Heath 2015). It seems to me pretty evident that there is a lot of willingness to hurt marginalized and vulnerable people’s feelings and shut them down, but more to the point, Heath’s argument presupposes that philosophy has not already been ‘me studies,’ albeit with a distinctively privileged ‘me’ who takes himself to be the neutral case. More recently, Brian Leiter’s blog has featured a series of posts spreading the moral panic that trans folks are using brute force to block out anyone else’s voice and win arguments.9 Meanwhile, Susan Haack (1996) calls feminist epistemology that accepts the basic precepts of standpoint theory “preposterous” and equates it with the generalized rejection of objectivity and truth by way of an inference from (1) to (2) to (3), claiming that for standpoint theorists who think there is no science unmarked by the social position of the knowers, “knowledge is nothing but the product of negotiation among members of the scientific community” (79). As Elizabeth Anderson puts it, Haack is ‘alarmed’ by the possibility that knowledge is situated, because: behind her alarm lies a particular model of the interaction of evidential and political considerations in shaping inquiry. The model supposes that these considerations necessarily compete for control of inquiry . . . To the extent that moral values and social influences shape theory choice, they displace attention to evidence and valid reasoning and hence interfere with the discovery of truth. (Anderson 1995, 33)
Ian Hacking, interestingly, bites the bullet on a version of the argument above. He embraces both the antecedent and the conditional of (1), denying the existence of any epistemically secure position from which to assess the epistemic quality of any other stance: “We cannot reason as to whether alternative systems of 9 See, for instance, https://leiterreports.typepad.com/blog/2019/06/philosophical- discussion- of- trans-identity-a-guide-for-the-perplexed.html.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
54 Quill R. Kukla reasoning are better or worse than ours, because the propositions to which we reason get their sense only from the method of reasoning being employed. The propositions have no existence independent of ways of reasoning toward them” (Hacking 1992, 65). This leads him to step (2) and then step (3), and he accepts a version of the idea that disagreements from structurally different perspectives are fundamentally power battles. From Hacking’s neo-Foucaultian perspective, this is not an unwelcome conclusion. However, it is certainly not what proponents of strong objectivity, ecological knowledge, or pragmatic encroachment defend, and it plays into the alarmist narrative that situated knowledge accounts fundamentally challenge the security and objectivity of knowledge by reducing arguments to personalities and social positions. Step (1) presumes that the only way to think critically about anything is from an unmarked, pure position, but this is exactly what the standpoint theorist denies. The idea that we have to leap out of all standpoints in order to compare any standpoints begs the question in favor of a weirdly disembodied, amaterial view of how we reason, which is just the view the standpoint theorist is pushing back against. Indeed, in most fields other than philosophy, the idea that epistemic practices have to be pure and unlimited by any perspective is quite foreign. Scientific papers, especially in the social sciences, routinely include a section on ‘limitations,’ in which the authors explicitly reflect on how their own methods, access to evidence, and position as researchers limit the generalizability and transparency of the results. This isn’t taken as undermining all possibility of objectivity or of critical assessment, but as just an epistemically healthy acknow ledgment of the finitude of their resources and perspective. We can and routinely do use imperfect, finite, perspectivally bound tools to critique our own and others’ perspectives. The finitude of the tools is part of the more general finitude of the world. Sally Haslanger points out that not only is this point routine rather than alarming in the sciences, but it is a major reason why scientists require results to be replicable. Apparent knowledge that can’t be generated from multiple perspectives is not secure, no matter how solid the methodology seems to be from within (Haslanger 2019). Furthermore, step (1) does not imply step (2). Even if all epistemic starting points are imperfect, unstable, and limited (as surely they are), this does not mean that those with different standpoints cannot use their imperfect tools and standards to critically interrogate, engage with, and correct and improve one another. Some critics, and even some supporters of situated epistemology, especially standpoint theory, implicitly or explicitly assume that if standpoints are different from one another, then they must be totally incommensurable and immutable, as if people with different standpoints live in hopelessly discrete, monadic worlds. But there is no reason why people who have somewhat different access to evidence, whose interests and values lead them to make somewhat different inferences, who are positioned to ask somewhat different questions using somewhat
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Situated Knowledge, Purity, and Moral Panic 55 different frames, cannot talk to one another (partially and imperfectly) and critique one another. We can return to our analogy with Hegel’s critique of Kant here: Just because our epistemic tools are finite and ‘impure’ doesn’t mean we can’t use them to engage in critique and comparison. To refuse to use them because of their impur ity is to indulge in a kind of paralyzing epistemic risk aversion. There is no reason to wait around for a purity guarantee before leaping in and using the resources we have. We critique various sets of epistemic practices as best we can from our imperfect stances, and progress and hopefully broaden and enrich those stances as we go. Our epistemic tools and capacities are finite and limited, like all our tools and capacities. Why is this surprising or scary? It seems to be a special kind of hubris characteristic of philosophers in particular to think that we have to have perfect shared understanding and identical shared epistemic standards before we can go about the business of making epistemic progress with one another. Step (3) is also unjustified fear mongering. Let’s provisionally acknowledge that we need to attend to social identities and power relations in assessing epistemic standards and practices. That in no way implies that this is then the only thing we need to attend to. We lose none of our grip on epistemic virtues such as coherence, evidence responsiveness, or compelling conceptual analysis, and none of our grip on epistemic vices such as framing biases, overgeneralization, post hoc explanation, and so forth, just by adding in critical attention to the strengths, limitations, and distinctive features of an epistemic agent’s identity. Saying that identity plays a role in epistemic evaluation doesn’t imply that it completely determines this evaluation. Thus, standpoint epistemologies do not realistically risk reducing epistemology to identity politics or mere power wars. And even if we make it all the way through step (3), step (4) seems to involve a kind of panic-driven unjustified leap, given that all the empirical evidence points in the direction of white males still having the lion’s share of power when it comes to whose gets to speak, whose voice gets heard, and who gets taken seriously. There have been some heartening incursions into syllabi, departments, and science labs, in more progressive spaces, but it’s not as though marginalized knowers generally win at these power games.
3.3 The Argument from Intersectionality The argument here is that standpoint epistemology in particular assumes that there is such a thing as a or the women’s standpoint (or perhaps a black women’s standpoint, or whatever it may be). But, the worry goes, this seems to erase important differences between group members and to impute a stability to these marginalized identities that is both implausible and politically problematic. In other words, standpoint theory cannot accommodate intersectionality. This
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
56 Quill R. Kukla worry has overwhelmingly come from inside feminist and other anti-oppressive theory, but it still strikes me as being a fear-driven response that requires burdening standpoint epistemology with assumptions and metaphysical underpinnings it does not need. The risk here is that if we assume that a group shares a standpoint, this purportedly common standpoint will end up once again treating the most privileged voices from within a group as the neutral or ‘representative’ voices. ‘Women’s standpoint’ will end up being dominated by the voices of economically privileged, able-bodied white women, and so forth.10 And indeed, there is, of course, a long and painful tradition of straight, white women taking it upon themselves to speak for women in general; of black men taking it upon themselves to speak for black people in general; and of the most audible voices with the most social capital in the queer community being educated, white, gay men.11 Somehow, Caitlin Jenner managed to socially position herself (at least briefly) as ‘the voice’ of trans women, while poor trans women of color, disabled trans women, and so forth, who face horrific safety threats and discrimination, have little to no cultural voice, despite the rise of socially accepted trans superstars. So, these concerns are real and pressing. But there is no reason to infer from the fact that groups are internally epistemically diverse the conclusion that standpoint theory is misguided. There is nothing inherent in standpoint theory that requires us to believe that marginalized groups have distinctive standpoints. We can believe that our epi stemic resources and practices are shaped ineliminably by our gender, race, body size and functioning, sexuality, or whatever it may be, without believing that it does so in the same way for everyone in our group, or that it does so by way of our common experiences. The latter simply doesn’t follow. It can be clear that the way I am gendered has deeply affected my perceptual skills, epistemic resources and the like, without undergirding this with any claims about my sharing ‘a gender’ with a group of people who must have been epistemically affected in the same way. Nor is there any reason to measure how much of my standpoint is shared with other women—or with other white women, or other white, queer women, or other white, queer, culturally Jewish, economically secure women, or whatever the case may be. We just don’t need to ask about such commonalities in order to get on with the business of critically interrogating my situated position as a knower. Only ideological anxiety would drive one to insist on finding the bound aries around my proper epistemic group before we can get on with exploring the potentials and limitations of my standpoint.
10 See Hekman 199; Lugones and Spelman 1983. Wylie (2003) discusses this objection and responses to it. 11 This last point about gay male cultural capital was made vivid in fascinating and powerful ways in the documentary How to Survive a Plague (2012).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Situated Knowledge, Purity, and Moral Panic 57 It is clear to me, for example, that my experience in the academic discipline of philosophy is shaped by my gender—by the fact that others read me as a woman, for instance—and that I was trained up on seeing, arguing, questioning, interacting, and caring about issues in ways that were shaped from the beginning in gendered ways. I am also gendered differently than many other philosophers who are read as women. To name just a few points of divergence that reflect our different positions: I am not especially ‘femme’; I am assertive; my sense of humor and choice of hobbies tend to be socially marked as ‘masculine’; I have trouble ‘being ladylike’, which has both advantages and disadvantages when it comes to interacting with students and colleagues; and so forth. It would be absurd to claim that my male colleagues’ reactions to my assertive way of talking were not gendered, on the ground that other women are not so assertive and hence don’t experience those same reactions. If we were to try to find the shared kernel that makes up ‘women’s standpoint’ in philosophy, we would be left with nothing, or next to nothing. We would miss the rich and varied and ongoing ways in which almost every interaction and experience I have in the field is marked by gender. There is no reason why gender has to play the same role in any identifiable group of people’s lives in order for it to play a role, including a rich and epistemically significant one. It is true that early standpoint theorists emphasized group standpoints and that their language tended to be essentializing in the way that critics interested in intersectionality, such as Lugones and Spelman (1983), rightly problematized. Nancy Hartsock (1983), for instance, tried to ground standpoint epistemology in women’s shared relationship to domestic and childrearing work by analogy with Marx’s grounding of standpoint epistemology in wage workers’ shared relationship to the means of production. However, in the decades during which standpoint theory has matured, it has become generally agreed that this sort of essentializing move is both implausible and theoretically unnecessary (Intemann 2010; Kukla 2006; Wylie 2003). Hence the essentialism charge, while often politically important, seems to be both antiquated and philosophically beside the point when it comes to critiquing standpoint theory.
3.4 The Argument that We Should Just Try Harder This argument schema is not a critique of situated epistemologies so much as a methodological response to them. Its proponents take it as a given that needs no argument that conceding the situatedness of epistemic practices and justifications would be a troublesome last resort. For instance, Jon Kvanig writes nostalgically, “Whether one adopts a positive or a negative characterization of pragmatic encroachment, there are substantive grounds for rejecting the idea and endorsing the long-standing tradition of intellectualism in the theory of knowledge” (2011,
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
58 Quill R. Kukla 85). Meanwhile, Clayton Littlejohn, reviewing Fantl and McGrath’s Knowledge in an Uncertain World, which is one of the main book-length defenses of pragmatic encroachment, is open about his emotional reaction to arguments for situated epistemology: It was the best of books. It was the worst of books. Best because it was filled with interesting and original arguments about the nature and possibility of know ledge. Worst because the arguments were good, much better than I had hoped given that they were arguments against an attractive constellation of views. (Littlejohn 2010)
Within the pragmatic encroachment literature, even the defenders of encroachment consider it the ‘tainted’ option (Stanley 2005, 2016) and the unsituated alternative the ‘pure’ option. Many defenders of ‘tainted’ epistemology acknow ledge apologetically that maintaining purity would have been the preferable option. The very term ‘encroachment’ already invokes the metaphors of impurity, pollution, and a compromise of boundaries that ought to have been secure. Even Fantl and McGrath, the twin patriarchs of pragmatic encroachment, refer to their own view as ‘impurism’ (2009). This type of acceptance of situatedness also gets routinely called the “anti-intellectualist” position (Kvanig 2011; Neta 2007). This terminology presupposes that only pure, unsituated epistemic practices are ‘intellectual’, whereas it is exactly the metaphysics and nature of intellectual thought that is at stake in these debates. In “Is Epistemology Tainted?” Stanley acknow ledges, partly in response to my arguments, that this normatively loaded language, along with the rhetorical privileging of unsituated epistemology more generally, is a manifestation of ideology (2016). Since these are matters of ter minological choice rather than argument, all this language strikes me as a lovely example of the rhetoric of moral panic. Critics of the inductive risk literature proceed by acknowledging that values and interests ‘infect’ all sorts of parts of the scientific process, but insist that if we just cleave off these parts carefully enough, with a fine enough scalpel, and declare them to be outside of the ‘core’ of science, the real stuff, as it were, we will be fine. For instance, Hugh Lacey writes: Note that this thesis is consistent with moral and social values playing many roles within science, e.g., in motivating research interests, in setting the direction and determining the legitimation of applied research and applications, even in making judgments about the adequacy of available evidence and assessing the testimony of scientific experts—and in many other ways. Proponent of science as value free have never denied this. They hold only that at its core—where theories are accepted and research directions of basic research are set—moral and social
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Situated Knowledge, Purity, and Moral Panic 59 values (unlike cognitive values) have no proper role, and that no viable value-outlook should be especially favored by the advances of scientific understanding. Too often, science is value free has been dismissed by pointing to the legitimate play of values at the periphery rather than at the core of scientific practices. (Lacey 2005, 2, my emphasis)
This language of core and periphery sets up the central governing metaphor for Lacey. But he gives no argument for why the purportedly value-free domain is what counts as core rather than as peripheral to science.12 Regardless of whether Lacey manages to identify a part of the scientific process that is value free and aperspectival, which I remain agnostic about here, my interest is in his declar ation that this part is the ‘core,’ whereas the value-infested parts are mere peri pherals. The argument seems to be that the ‘core’ of science is value free and aperspectival because being value free and aperspectival is one of the most important things about science: “impartiality is a key value of all research practices conducted under any strategy” (Lacey 2005, 12). But key just seems to me to be another word for core here, and I don’t see any noncircular argument for accepting this priority claim. Why are so many philosophers working so hard to save an unsituated, aperspectival ‘core’ of epistemology, and why does it seem so clear, even to those who are pessimists about this project, that this would constitute the good stuff, the real deal? Or conversely, why does it seem obvious that arguments against such a pure space are arguments for a kind of tragic loss? In their massive, carefully documented history of objectivity, Lorraine Daston and Peter Galison show that aperspectivality is only one, relatively recent, interpretation of objectivity. Over time, we have understood objectivity in terms of fidelity to reality, trained judgment, reproducibility, and various other notions. “Knowledge that bears no trace of the knower” turns out to be a rather “young epistemic virtue” according to Daston and Galison (2007, 17). Perhaps it is time that we recognize that this is not a given and unquestionable epistemic virtue, but a historically constructed vision of what good epistemic practice looks like—one that was deeply undergirded by and constitutive of the erasure of the many kinds of epistemic agents who have been traditionally excluded from spaces for experts and authoritative knowledge producers.
12 Even Heather Douglas, who almost single-handedly reinvigorated the discussion of inductive risk and the ineliminability of values in scientific judgment, feels the need to insist that values can play an ‘indirect’ role, but may never play a ‘direct’ role Douglas 2000). Critics have struggled to make out this distinction and what makes one role more ‘direct’ than the other in her writing. But giving them an ‘indirect’ role does seem to rhetorically downgrade their ability to penetrate right to the core.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
60 Quill R. Kukla
4. Don’t Panic! Towards a Naturalized, Non-Ideal Epistemology Epistemic activities such as knowing, inquiring, and justifying—like absolutely everything else we biological creatures do—are ultimately material activities performed in the natural world, limited by our finitude, imperfection, and vulner ability to error and malfunction. It seems to me that the need to situate knowers in their contexts in order to critically evaluate their epistemic practices is no more mysterious or spooky than our need to situate any material process or phenom enon in its context in order to evaluate it. Nothing in this world is pure. We wouldn’t consider trying to understand an epidemic independent of the properties of the region across which it is spreading, or climate change independent of the material and social conditions in which it will occur, or the Algerian Revolution independent of the historical, political, and cultural relationship between Algeria and France, and between Europe and Africa. That we need to understand epistemic phenomena as situated material and social phenomena should be no more surprising. What would a naturalized, non-ideal epistemology look like? That is, what would it look like to do epistemology beginning from the assumption that we are all finite, natural beings constituted by our situation, incapable of purifying our stance, and operating with tools and methods and resources that are riddled with limitations and risks? What does it look like once we allow epistemology to admit that it is the normative study of a messy material set of social phenomena? A naturalized epistemology, I propose, would begin with the presumption that good (albeit imperfect) epistemic practices are possible, rather than trying to transcendentally ground their possibility from an idealized, perspectiveless location. It would study epistemic adequacy by analyzing particular empirical, socially situated practices, and teasing out their potentials and limitations, rather than by developing abstract standards against which particular epistemic practices are then measured. We can build up a picture of objectivity, not by beginning from an aperspectival ideal, but by slowly and piecemeal uncovering which practices and perspectives offer robust knowledge. This would be continuous with what Daston and Galison describe as: a view of objectivity as constituted from the bottom up, rather than from the top down . . . One becomes objective by performing objective acts. Instead of a preexisting ideal being applied to the workaday world, it is the other way around: the ideal and ethos are gradually built up and bodied out by thousands of concrete actions. (2007, 2)
Furthermore, a naturalized epistemology would acknowledge that most of the time, knowledge is a social enterprise. Most of our epistemic practices—both our everyday informal practices, and the more formal practices that make up
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Situated Knowledge, Purity, and Moral Panic 61 science—are collaborative and coordinative. Much of our epistemic labor is devoted to figuring out how to share information, negotiate disagreement, and generate hypotheses and knowledge together. We build up knowledge through laborious social interactions that require crossing language and disciplinary barriers and cultural gaps, and we do our best despite breakdowns in communication, financial interests, and ideological investments that skew our research, and so forth. Many of our epistemic norms are coordinative. We have implicit and explicit norms for when someone or something can be trusted, what different kinds of knowledge products can be used for, how to cite them, and so forth. We have standards, both loose and formal, for how different kinds of justifications can be translated into social practices (for instance, randomized controlled trials have a different place in evidence-based standards of medical care than do observational studies). None of these norms or practices will show up at all in a version of epistemology that considers only isolated, context-free, unsituated individual knowers. A non-ideal, naturalized epistemology would begin inside this messy reality, rather than extracting out an idealized individual and asking how he would know, were he a perfect epistemic agent and not finitely embodied or socially embedded. To the extent that there has been a push to transform epistemology into a dis cipline with a serious empirical component, it has looked starkly different from what I just described. Experimental philosophers who study epistemology do not study epistemic practices ‘in the wild,’ as I am advocating, nor do they take epi stemic practices to be social practices. On the contrary, most experimental phil osophy is aimed at designing and running experiments that isolate agents from other epistemic collaborators and from anything resembling the normal contexts of epistemic practice. Epistemological studies in experimental philosophy typic ally proceed by asking people whether artificial, described scenarios count as knowledge, rather than examining how people know (see, for instance, Weinberg et al. 2001; Turri 2013; and Swain et al. 2008, among others). This experimental model builds in several presumptions that I am challenging. These include: (1) Being in the midst of a situation is irrelevant to determining how we can know things within it; imagining a situation abstractly from the outside is just as effective. (2) We can decide whether something counts as knowledge by attending to an individual’s cognitive state singly, without attention to their interactions with other people or to the social organization of knowledge, and impersonally, without attention to who they are. (3) Epistemic practices are pure intellectual rather than materially embodied activities, so that our standards for thinking about them are roughly interchangeable with our standards for engaging in them.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
62 Quill R. Kukla Despite the emphasis on empirical experiment in this literature, it seems to me that all these assumptions are deeply antinaturalistic. They reinscribe aperspectivalist, anti-materialist assumptions, and they are disconnected from the messy social and natural reality of how we actually go about trying to know things.13 A naturalized, non-ideal epistemology of the sort I am describing will have more in common with the history and anthropology of science than with experimental philosophy as currently practiced, but unlike those disciplines, it will not shy away from normative claims. The purists see arguments that purport to demonstrate that some epistemic concept is ineliminably situated as threats to the security and rational orderliness of our epistemic practices. The central practical upshot of this chapter is my suggestion that we give up trying to carve out or defend a pure aperspectival core and get used to the idea that epistemic practices are socially situated, collaborative, and coordinative, performed by finite natural beings. This means that attention to the social organization of knowledge, to the distortions in epistemic practices caused by power relations, to the standpoints of the epistemic agents involved, and to the values, interests, and power relations shaping knowledge practices will be at the forefront of epistemology. Attending to how epistemic practices are performed by finite, diverse, socially embedded knowers is not a distraction or a contamination—it is simply what good critical analysis of a complex, natural, social phenomenon looks like.
References Anderson, Elizabeth (1995). “Knowledge, Human Interests, and Objectivity in Feminist Epistemology.” Philosophical Topics 23(2): 27–58. Baumann, Hansel (2018). “Deafspace: Lessons on Architecture, Empathy, and Cultural Agency.” Presentation at Georgetown University, Washington, DC.
13 Much experimental philosophy (including Weinberg et al. 2001 and Turri 2013) studies demographic differences in intuitions about knowledge, which may make it seem as if it is acknowledging situatedness. But first, once we acknowledge that in the natural world, epistemic practices are complex social patterns involving a great deal of coordination, it becomes odd, to say the least, to try to study them empirically on a person-by-person basis. This is so even if we note the demographics of each person. It’s a bit like trying to do epidemiology by studying individuals piecemeal; the social patterns and interactions are the phenomena under study, so we can’t just study individuals and then aggregate them back into groups. Second, noting demographic differences between individuals’ epistemic intuitions does not help us with the critical or normative interrogation of what different standpoints have to offer. By asking each participant to do the same small, contained intellectual task, we shed no light on the actual embodied differences between how different sorts of people access evidence, make inferences from evidence, develop justifications, and settle on knowledge. See Kukla 2016 for a detailed discussion of the limitations of traditional experimental epistemology and the differences between it and naturalized epistemology.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Situated Knowledge, Purity, and Moral Panic 63 Biddle, Justin (2007). “Lessons from the Vioxx Debacle: What the Privatization of Science can Teach Us about Social Epistemology.” Social Epistemology 21(1): 21–39. Biddle, Justin B. and Rebecca Kukla (2017). “The Geography of Epistemic Risk,” in Elliot and Richards (eds), Exploring Inductive Risk: Case Studies of Values in Science. Oxford: Oxford University Press, 215–38. Code, Lorraine (2006). Ecological Thinking. Oxford: Oxford University Press. Cohen, Stanley (1973). Folk Devils and Moral Panics: Creation of Mods and Rockers. Boulder, CO: Paladin. Daston, Lorraine and Peter Galison (2007). Objectivity. Princeton, NJ: Zone Books. DeRose, Keith (1992). “Contextualism and Knowledge Attributions.” Philosophy and Phenomenological Research 52(4): 913–29. Douglas, Heather E. (2000). “Inductive Risk and Values in Science.” Philosophy of Science 67(4): 559–79. Du Bois, W. E. B. (1903). The Souls of Black Folk. New York: Dover Publications. Fantl, Jeremy and Matthew McGrath (2009). Knowledge in an Uncertain World. Oxford: Oxford University Press. Flaherty, Colleen (2017). “Occupation of Hum 110.” Inside Higher Education (November 9, 2017), accessible with comments at www.insidehighered.com/ news/2017/09/11/reed-college-course-lectures-canceled-after-student-protestersinterrupt-class. Galison, P. (2003). “The Collective Author,” in Mario Biagioli and Peter Galison (eds), Scientific Authorship: Credit and Intellectual Property in Science. London: Routledge, 325–53. Goodchild, Michael (2006). “Geographic Information Systems”, in Stuart Aitken and Gill Valentine (eds), Approaches to Human Geography. New York: Sage, 251–62. Haack, Susan (1996). “Science as Social?—Yes and No,” in Feminism, Science, and the Philosophy of Science. New York: Springer, 79–93. Hacking, Ian (1992). “The Self-Vindication of the Laboratory Sciences,” in A. Pickering (ed.), Science as Practice and Culture. Chicago, IL: University of Chicago Press, 29–64. Harding, Sandra G. (1986). The Science Question in Feminism. Ithaca, NY: Cornell University Press. Harding, Sandra (1992). “Rethinking Standpoint Epistemology: What Is ‘Strong Objectivity?’. ” The Centennial Review 36(3): 437–70. Hartsock, Nancy (1983). “The Feminist Standpoint: Developing the Ground for a Specifically Feminist Historical Materialism,” in Sandra Harding and Merrill Hintikka (eds.), Discovering Reality. New York: Springer, 283–310. Haslanger, Sally (2019). “Standpoint Epistemology and the Authority of Critical Perspectives.” Lecture at Frei Universität Berlin, June 2019. Haugeland, John (1997). “Mind Embodied and Embedded,” in Having Thought: Essays in the Metaphysics of Mind. Cambridge, MA: Harvard University Press, 207–40.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
64 Quill R. Kukla Hawthorne, John and Jason Stanley (2008). “Knowledge and Action.” The Journal of Philosophy 105(10): 571–90. Heath, Joseph (2015). “The Problem of Me Studies.” In Due Course: A Canadian Public Affairs Blog (May 9, 2019), http://induecourse.ca. Hekman, Susan (1997). “Truth and Method: Feminist Standpoint Theory Revisited.” Signs 22(2): 341–65. Hegel, G. W. F. (1977). Phenomenology of Spirit, trans. A. V. Miller. Oxford: Oxford University Press. Hempel, Carl G. (1965). “Science and Human Values,” in Aspects of Scientific Explanation and other Essays in the Philosophy of Science. Cambridge: The Free Press, 81–96. Hill-Collins, Patricia (1986). “Learning from the Outsider Within: The Sociological Significance of Black Feminist Thought.” Social Problems 33(6): s14–s32. Huebner, Bryce, Rebecca Kukla, and Eric Winsberg (2017). “Making an Author in Radically Collaborative Research,” in Thomas Boyer-Kassem, Conor Mayo-Wilson, and Michael Weisberg (eds), Scientific Collaboration and Collective Knowledge. Oxford: Oxford University Press, 95–116. Ichikawa, Jonathan Jenkins (2017). Contextualizing Knowledge: Epistemology and Semantics. Oxford: Oxford University Press. Intemann, Kristen (2010). “25 Years of Feminist Empiricism and Standpoint Theory: Where Are We Now?” Hypatia 25(4): 778–96. Jeffrey, Richard C. (1956). “Valuation and Acceptance of Scientific Hypotheses.” Philosophy of Science 23(3): 237–46. Kant, Immanuel (1998). Critique of Pure Reason, trans. and ed. Paul Guyer and Allen W. Wood. Cambridge: Cambridge University Press. Kukla, Rebecca (2006). “Objectivity and Perspective in Empirical Knowledge.” Episteme: A Journal of Social Epistemology 3(1): 80–95. Kukla, Rebecca (2007). “Holding the Body of Another.” Symposium: Canadian Journal of Continental Philosophy 11(2): 397–408. Kukla, Rebecca (2012). “ ‘Author TBD’: Radical Collaboration in Contemporary Biomedical Research.” Philosophy of Science 79(5): 845–58. Kukla, Rebecca (2016). “Delimiting the Proper Scope of Epistemology.” Philosophical Perspectives 29: 202–16. Kvanig, Jonathan L. (2011). “Against Pragmatic Encroachment.” Logos & Episteme 2(1): 77–85. Lacey, Hugh (2005). Is Science Value Free? Values and Scientific Understanding. London: Psychology Press. Littlejohn, Clayton (2010). “Review of Fantl and McGrath, Knowledge in an Uncertain World.” Notre Dame Philosophical Reviews (May 1, 2010). Longino, Helen E. (1990). Science as Social Knowledge: Values and Objectivity in Scientific Inquiry. Princeton, NJ: Princeton University Press.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Situated Knowledge, Purity, and Moral Panic 65 Lugones, Maria C. and Elizabeth V. Spelman (1983). “Have We Got a Theory for You! Feminist Theory, Cultural Imperialism and the Demand for ‘the Woman’s Voice’. ” Women’s Studies International Forum 6(6): 573–81. Lynch, Michael P. (2016). The Internet of Us: Knowing More and Understanding Less in the Age of Big Data. London: W. W. Norton & Company. Madden, Vicky (2014). “Why Poor Students Struggle.” New York Times (September 21, 2014). Tiffany Martínez (2016). “Academia, Love Me Back.” Blog post (October 27, 2016), https://vivatiffany.wordpress.com/2016/10/27/academia-love-me-back. McKinnon, Rachel (2015). “Trans*formative Experiences.” Res Philosophica 92(2): 419–40. Miller, Boaz (2014). “Science, Values, and Pragmatic Encroachment on Knowledge.” European Journal for Philosophy of Science 4(2): 253–70. Miller, Boaz and Isaac Record (2013). “Justified Belief in a Digital Age: On the Epistemic Implications of Secret Internet Technologies.” Episteme 10: 117–34. Neta, Ram (2007). “Anti-Intellectualism and the Knowledge–Action Principle.” Philosophy and Phenomenological Research 75(1): 180–7. Nguyen, C. Thi (2018). “Echo Chambers and Epistemic Bubbles.” Episteme 17: 141–61. Rudner, Richard (1953). “The Scientist qua Scientist Makes Value Judgments.” Philosophy of Science 20: 1–6. Smith, Dorothy E. (1974). “Women’s Perspective as a Radical Critique of Sociology.” Sociological Inquiry 44(1): 7–13. Solomon, Miriam (2001). Social Empiricism. London: MIT Press. Stanley, Jason (2005). Knowledge and Practical Interests. Oxford: Oxford University Press. Stanley, Jason (2016). “Is Epistemology Tainted?” Disputato 42(8): 1–35. Steel, Daniel (2010). “Epistemic Values and the Argument from Inductive Risk.” Philosophy of Science 77(1): 14–34. Swain, Stacey, Joshua Alexander, and Jonathan M. Weinberg (2008). “The Instability of Philosophical Intuitions: Running Hot and Cold on Truetemp.” Philosophy and Phenomenological Research 76(1): 138–55. Tollefsen, Deborah (2007). “Group Testimony.” Social Epistemology 21(3): 299–311. Tooley, Michael (2014). “Why No Criticism of the Site Visit Report?” Blog post, http:// spot.colorado.edu/~tooley/Why_No_Criticism_of_the_Site_Visit_Report.html. Turri, John (2013). A Conspicuous Art: Putting Gettier to the Test. Ann Arbor, MI: University of Michigan Library. Weatherson, Brian (2005). “Can We Do without Pragmatic Encroachment?” Philosophical Perspectives 19: 417–43. Weinberg, Jonathan M., Shaun Nichols, and Stephen Stich (2001). “Normativity and Epistemic Intuitions.” Philosophical Topics 29(1/2): 429–60.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
66 Quill R. Kukla Wilholt, Torsten (2009). “Bias and Values in Scientific Research.” Studies in History and Philosophy of Science Part A 40(1): 92–101. Wilholt, Torsten (2013). “Epistemic Trust in Science.” British Journal for the Philosophy of Science 64(2): 233–53. Winsberg, Eric (2012). “Values and Uncertainties in the Predictions of Global Climate Models.” Kennedy Institute of Ethics Journal 22(2): 111–37. Wray, K. Brad (2006). “Scientific Authorship in the Age of Collaborative Research.” Studies in History and Philosophy of Science Part A 37(3): 505–14. Wray, K. Brad (2007). “Who Has Scientific Knowledge?” Social Epistemology 21(3): 337–47. Wylie, Alison (2003). “Why Standpoint Matters,” in Science and Other Cultures: Issues in Philosophies of Science and Technology. London: Routledge, 26–48. Yong, Ed. (2017). “The Absurdity of the Nobel Prizes in Science.” The Atlantic (October 3, 2017).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 19/03/21, SPi
4
Epistemology and the Ethics of Animal Experimentation Mylan Engel Jr.*
Each year 50–100 million animals are used as experimental subjects in scientific research.1 The costs—in terms of pain, injury, distress, and death—imposed on these animal subjects are enormous. Honest proponents of animal-based research acknowledge these costs but insist that such research is justified because its benefits to humans outweigh the costs imposed on its victims. My aim is to assess this purported defense of animal experimentation through the lens of epistemology. I focus on the use of animals in biomedical research, particularly pharmacological research, since if it is wrong to use animals in biomedical experiments aimed at the development of vitally important, potentially life-saving drugs, then it is a fortiori wrong to use animals to test unnecessary, trivial products such as a new floor wax or a new shampoo.2 In section 1, I identify some of the routine costs associated with animal experimentation. In section 2, I identify three commonsense moral principles that establish the prima facie wrongness of animal experimentation, i.e. they establish that animal experimentation is wrong unless there is an overriding moral reason that justifies inflicting pain, injury, distress, and premature death on the animal subjects involved. In section 3, I spell out “the benefits argument” in more detail and draw attention to the fact that the purported benefits of animal experimentation are principally epistemic in nature. In section 4, I discuss reliability and the crucial role it plays in both internalist and externalist epistemologies. These four sections together will show that harmful, non-therapeutic animal experimentation is morally permissible only if it is not known to be an unreliable method of acquiring beliefs about the origin, pathology, and proper treatment of human disease. In section 5, I demonstrate that animal
* Earlier versions of this chapter were presented at the Midsouth Philosophy Conference, the Russell Philosophy Conference, and the Rocky Mountain Ethics Congress. I would like to thank those in attendance for their questions and suggestions. Special thanks to Ramona Ilea, Bruce Russell, and an anonymous referee for their very helpful comments on earlier drafts. 1 Nuffield Council on Bioethics (2005, 7). 2 These new products are unnecessary because we already have all of the floor waxes and shampoos that we could ever need. In addition, the benefits are trivial, especially in comparison with the great harms inflicted on animals, and thus could not be justified by a cost–benefit analysis. Mylan Engel Jr., Epistemology and the Ethics of Animal Experimentation In: Applied Epistemology. Edited by: Jennifer Lackey, Oxford University Press (2021). © Mylan Engel Jr. DOI: 10.1093/oso/9780198833659.003.0004
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
68 Mylan Engel Jr. experimentation is in fact a highly unreliable, failed technology, and as such, it does not, and cannot, provide the epistemic benefits needed to justify its con tinued use. Consequently, section 6 concludes that animal experimentation is an epistemically and morally unjustified practice that ought to be abolished.
1. The Costs to Animals The overwhelming majority of biomedical experiments performed on non-human animals are non-therapeutic experiments. A non-therapeutic experiment is an experiment undertaken not to benefit the animal subject, but rather to advance scientific or medical knowledge. The overwhelming majority of animal experiments are also harmful (i.e. they cause, or are likely to cause, direct harm in the form of bodily injury, pain, infection, and death) to the animal subjects involved in the experiment. Indeed, the point of some experiments is to kill a certain percentage of the animal test subjects. For example, the point of the LD50 [lethal dose 50 percent] test is to determine the amount of a given substance required to kill half of the animal subjects in the test. To determine that amount, the animals are force-fed the substance until half of them die. Before that point is reached, most of the animals are very ill and in severe distress.3 Tom Regan highlights just some of the kinds of research conducted on animals: “Experimental procedures include drowning, suffocating, starving, and burning; blinding animals and destroying their hearing; damaging their brains, severing their limbs, crushing their organs, inducing heart attacks, ulcers, paralysis and seizures.”4 Other experimental procedures include irradiating animals, using them in compression and decompression studies, and subjecting them to electric shock.5 Much of this research, by the researchers’ own admission, causes the animal subjects extreme pain, injury, and distress.6 Many of these animals die as a result of the experiments. Those that survive are typically killed at the end of the experiment, having outlived their scientific usefulness. Consider one such experiment as described in the medical journal Surgery:
Beagle • Forty conditioned beagles (purchased from Marshall Research Animals, North Rose, NY), 6 to 7 months in age, received a 33 percent total body surface 3rd degree flame burn.
3 Singer (2009, 53–4). 6 Garrett (2012, 3).
4 Regan (2012, 108).
5 Ryder (1985, 79–82).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Epistemology and THE Ethics of Animal Experimentation 69 • The purpose: To determine the potential benefit or toxicity of Cornybacterium parvum vaccine, an immunomodulator, when given after severe burn injury. • Burning Procedure: In preparation of burning, the beagles were clipped, but not shaved, from the neck to the base of the tail. The animals were placed under general anesthesia and flame burned one side at a time, by placing kerosene-soaked gauze on their sides and igniting the gauze. The flame was extinguished after 60 seconds with a wet towel. Skin surface temperatures reached 180º C to 200º C [356º F to 392º F] before extinguishing. Immediately following the burn, intravenous fluid treatment using Ringer’s lactate solution was administered to the beagles for an 18-hour period. During this 18-hour period, “light anesthesia” was used. Thereafter, no a nesthesia was administered, since its continued use might have interfered with the test results. • Only animals which appeared to be alert and healthy on post-burn day 2 were entered into the study. [So, the actual number of beagles burned remains unclear.] The beagles were divided into two groups, a control group and an experimental group. On days 2 and 9 of the study, the 21 beagles in the experimental group were administered C. parvum vaccine (COPARVAX, lot 8B2705, CTM 2042, Burroughs Wellcome Co.) in a dose of 10 mg/m2 in 250 ml of saline solution. The 19 beagles in the control group were administered 250 ml of pure saline solution containing no vaccine. All animals were followed for survival for 90 days. • Results: Both test and control animals progressively lost weight. Animals that lost 30 percent of their pre-burn weight invariably died. Only 16 percent of the animals receiving the saline solution survived, whereas 48 percent of the animals receiving C. parvum survived. Total white counts and neutrophil function were the only values which were significantly better in beagles receiving C. parvum. (Stinnett et al. 1981, 237–42) Beagle was conducted by J. Dwight Stinnett and his colleagues at Shriners Burns Institute and Surgical Immunobiology Laboratory at the University of Cincinnati Medical Center. Support for the study was provided by the Shriners Burns Institute, by U.S. Public Health Service Grant No. AI-12936, and by the Burroughs Wellcome Company, maker of the C. parvum vaccine COPARVAX. The research protocols followed in Beagle as described above are consistent with the guidelines for “humane” animal care as outlined by the United States Department of Agriculture (USDA) and the National Institutes of Health (NIH).7 Experiments like Beagle are carried out at research facilities and universities every day. These experiments exact an extremely heavy toll—both in terms of suffering and in terms of premature death—on the animal subjects involved. 7 This fact alone strongly suggests that the Guidelines for Humane Animal Care have much to do with public perception, but little to do with humane care.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
70 Mylan Engel Jr.
2. Common Sense and the Prima Facie Wrongness of Animal Experimentation The prima facie case against animal experimentation rests on three widely accepted, commonsense moral principles that even the researchers themselves embrace. These principles are so central to our conception of morality that any moral theory that conflicted with them would be rejected on reflective equilibrium grounds. Since any adequate moral theory must cohere with these prin ciples, we can appeal to them directly when evaluating the moral status of animal experimentation. The principles are these: (P1) It is wrong to harm sentient animals for no good reason. (P2) It is wrong to cause sentient animals to suffer for no good reason. (P3) It is wrong to kill sentient animals for no good reason. These principles are not in dispute.8 Even the most steadfast defenders of animal experimentation accept them. For example, in The Animal Rights Debate, staunch animal experimentation advocate Carl Cohen explicitly endorses (P2) and (P3): “If animals feel pain (and certainly mammals do, though we cannot be sure about insects and worms), we humans surely ought cause no pain to them that cannot be justified. Nor ought we kill them without reason.”9 Elsewhere in the same volume, Cohen reiterates his commitment to (P2) and (P3): “Our obligations to animals arise not from their rights, I believe, but from the fact that they can feel pain and from the fact that we, as moral agents, have a general obligation to avoid imposing needless pain or death.”10 Peter Carruthers also directly endorses (P2):
8 As I have already noted, these principles are central to our understanding of morality. Together they specify an important part of the underived conceptual role of the concept of moral wrongness. By way of illustration, consider the following much-discussed example from Gilbert Harman: “If you round the corner and see a group of young hoodlums pour gasoline on a cat and ignite it, you do not need to conclude that what they are doing is wrong; you do not need to figure anything out; you can see that it is wrong” (Harman 1977, 4). Harman offered the example to show that some moral judgments are direct, as opposed to inferential. What is relevant about Harman’s example for present purposes is this: No one seriously doubts that burning a cat to death for no good reason is wrong. Treating a cat in such a way harms the cat, causes her to suffer, and kills her for no good reason, and we all judge such conduct to be wrong. For a more recent, non-fictional example, consider the public outrage that erupted when it was revealed that professional football player Michael Vick was guilty of sponsoring dog-fighting rings in which pit-bulls were forced to fight each other to the death. As with Harman’s cat, we are outraged that someone would cause these dogs such harm, suffering, and death for no good reason. We regard people who would engage in such conduct as morally deficient and/or depraved. These examples illustrate that principles (P1)–(P3) are partially constitutive of the very concept of moral wrongness and they confirm that no one seriously doubts (P1)–(P3). 9 Cohen (2001, 46), emphasis mine. To see Cohen’s commitment to (P2) here, we need only recognize that justification proceeds in terms of reasons. We are justified in causing an animal pain if, and only if, we have a good reason for doing so. If there is no good reason to cause an animal pain, then causing that animal pain cannot be justified. 10 Ibid., 226, emphasis mine.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Epistemology and THE Ethics of Animal Experimentation 71 . . . it will be useful to have a rough idea at the out-set of what our common-sense morality tells us about the status and appropriate treatment of animals . . . Most people hold that it is wrong to cause animals unnecessary suffering. Opinions will differ as to what counts as necessary . . . But all will agree that gratuitous suffering—suffering caused for no good reason—is wrong. (Carruthers 1992, 8)
Mary Midgley also expressly endorses (P2): “it is wrong to cause suffering for an entirely trivial issue.”11 These sentiments are not new. In 1813, Le Gallois endorsed (P2): “I own that it would be barbarous to make animals suffer in vain, if the object of the experiment could be obtained without it.”12 These prominent proponents of animal-based research explicitly acknowledge that it is wrong to harm, inflict suffering on, or kill animals for no good reason. If a practice causes harm, suffering, or death to animals for no good reason, then that practice is wrong and ought to be abolished. Since animal experimentation harms the animal subjects involved, causes them to suffer, and kills them prematurely, animal experimentation is wrong unless there is a good reason for conducting it.
2.1 Reasons vs Good Reasons Since the moral permissibility of performing harmful, non-therapeutic, lethal experiments on animals turns on whether or not there is a good reason to perform such experiments, it is important to be clear at the outset about what counts as a good reason. Not just any reason is a good reason to harm, inflict pain on, or kill an animal. To be a good reason, the reason must be morally weighty enough to justify the behavior in question; it must be morally weighty enough to override the most significant interests of the animal being harmed. An example will help illustrate the point:
Kitten Suppose I happen to enjoy the smell of live kittens being burned to death. That is a reason to pour gasoline on a kitten, light the gasoline, and burn the kitten to death, but it is not a good reason. My incredibly trivial interest in experiencing a particular smell sensation does not outweigh the kitten’s most significant interests in avoiding such horrific suffering and premature death. Consequently, it would be wrong of me to burn a kitten alive for that reason.
11 Midgley (1986, 219).
12 Le Gallois (1813, 19–21).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
72 Mylan Engel Jr. Similarly, if a researcher is going to inflict third-degree burns on forty conditioned beagles and deprive them of all pain medication 18 hours after the burning procedure, that researcher needs a good reason for doing so. Proponents of animal research acknowledge the heavy toll—both in terms of suffering and in terms of premature death—that animal experiments like Beagle exact on the animal subjects involved, but they insist that there is a good reason for conducting such experiments. Let us turn now to examine that reason.
3. The Benefits Argument There is only one serious attempt at justifying animal experimentation, namely, the benefits argument.13 This argument takes the form of a utilitarian cost–benefit analysis. According to the argument, animal experiments are justified because the benefits that humans receive from the experiments outweigh the costs imposed on the animal subjects. The benefits we allegedly receive from animal-based biomedical research are first and foremost epistemic. Experimenting on animal models is supposed to provide us knowledge of (or at least evidence or justified belief concerning) the origin and proper treatment of human disease. Indeed, animal-based biomedical research is claimed to be indispensable for acquiring such knowledge and evidence. Carl Cohen makes the point as follows:
13 To be sure, there are less-serious attempts to defend animal experimentation, but these attempts are either implausible or irrelevant. For example, one might claim that animals are incapable of feeling pain, but no reasonable person would accept such a view (and the researchers themselves don’t accept such a view), because the evidence of pain perception in animals is overwhelming. Here is some of that evidence: We witness pain behavior, not just reflex actions to noxious stimuli (protective pain), but subsequent pain-induced behavioral modification caused by bodily damage (restorative pain); we observe significant anatomical and neurophysiological similarity between humans and many animals (including all mammals and most vertebrates); endogenous serotonergic and opioid pain-control mechanisms are present in all vertebrates; efferent and afferent nerves (including A-delta fibers and C fibers) run throughout their bodies; analgesics and anesthetics stop animals from exhibiting pain behavior; and there is compelling (though unethically obtained) experimental evidence that the capacity to feel pain enhances survival value in animals, based on the self-destructive tendencies displayed by animals that have been surgically deafferented. Based on this cumulative observational, analogical, neurophysiological, and experimental evidence, no reasonable person can deny that animals can feel pain. One might also try to justify animal experimentation, as Cohen does, by claiming that animals don’t have rights, but whether or not animals possess rights is irrelevant to the moral status of animal experimentation. We all agree that it is wrong to cause unnecessary suffering. Since animals can suffer and suffering is intrinsically bad, we ought to cause them no suffering that cannot be justified. As noted in section 2, Cohen himself expressly endorses this principle. Since animals can suffer, and animal experimentation causes them to suffer, we need a reason morally weighty enough to override these animals’ interests in avoiding such suffering and premature death, if animal experimentation is to be justified. The benefits argument is the only serious attempt at providing such a reason. If it fails, animal experimentation is morally unjustified.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Epistemology and THE Ethics of Animal Experimentation 73 Before administering any new drug to any human being, we must have substantial evidence that the experimental compound is not poisonous, and the only way we can gather good evidence for that belief is through the results of prolonged trials using live animals . . . Long before investigators can even begin to measure the effectiveness of a new drug in humans, they must have good reason to believe that it is safe. To this end animal subjects must be used. (Cohen 2001, 76, emphasis mine)
In the early twentieth-century, French physiologist Claude Bernard defended the use of the “animal model” as follows: [Hospitals] are the first field of observation which a physician enters; but the true sanctuary of medical science is the laboratory; only there will he seek explanations of life in normal and pathological states by means of experimental analysis. . . . There, by experiments on animals, he will account for what he has observed in his patients, whether about the action of drugs or about the origin of morbid lesions in organs and tissues. There, in a word, he will achieve true medical science. (Bernard 1926, 35)
As Bernard sees it, the only way to come to know the pathology and proper treatment of human disease is to experiment on non-human animals. As such, the ultimate justification of animal-based biomedical research rests on the epistemic benefits that such research allegedly provides. To be sure, the benefits argument is sometimes couched in terms of the prac tical benefits—for example improved human health, disease prevention through vaccines, increased longevity, and symptom amelioration—that supposedly result from animal research.14 However, if there are such practical benefits, these benefits are ultimately derived from the epistemic benefits—the newfound knowledge 14 It is worth noting that animal experimentation advocates tend to dramatically exaggerate the practical benefits derived from animal- based biomedical research. For example, the American Medical Association’s White Paper defending animal research alleges that had animal-based research been restrained, “many millions of Americans alive and healthy today would never have been born or would have suffered a premature death. Their parents or grandparents would have died from diph theria, scarlet fever, tuberculosis, diabetes, appendicitis, and countless other diseases and disorders” (Hendee 1988, 17). Medical historians, John B. McKinlay and Sonya M. McKinlay, beg to differ: In general, medical measures (both chemotherapeutic and prophylactic) appear to have contributed little to the overall decline in mortality in the United State since about 1900— having in many instances been introduced several decades after a marked decline had already set in and having no detectable influence in most instances. More specifically, with reference to those five conditions (influenza, pneumonia, diphtheria, whopping cough and poliomyelitis) for which the decline in mortality appears substantial after the point of intervention—and on the unlikely assumption that all this decline is attributable to the intervention—it is estimated that at most 3.5 percent of the total decline in mortality since 1900 could be ascribed to medical measures introduced for the diseases considered here. (McKinlay and McKinlay 1977, 425)
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
74 Mylan Engel Jr. of (or evidence or justified belief concerning) the pathology and proper treatment of human disease that the animal experiments supposedly provide. To accurately assess whether or not biomedical animal experimentation provides us with the epistemic benefits needed to justify its continued practice, we first need to understand the role that reliability plays in justification and knowledge.
4. Reliability, Justification, and Knowledge Reliability plays an important role in both internalist and externalist theories of knowledge. Indeed, both internalists and externalists agree that the reliability of the method that gives rise to a belief is relevant to the justification and knowledge of that belief, though they disagree as to how it is relevant. Externalists with respect to epistemic justification typically follow Alvin Goldman’s (1979) lead and maintain that actual reliability is what matters for justification. On their view, a belief-producing method must be reliable in order to confer justification on the beliefs it produces. More precisely, S’s belief that p is justified if and only if S’s belief that p is produced by a reliable method (or process). Accordingly, if the method giving rise to S’s belief is actually unreliable, then the resulting belief is unjustified, regardless of whether S realizes it or not. There is no single internalist take as to how reliability impacts justification and knowledge, but, as we shall see, internalists do agree on one critically important point. We can see this point of agreement by contrasting several representative internalist views on reliability and justification. Matthias Steup (2004) defends what he calls internalist reliabilism—the view that a method M is a source of justification for S only if S has evidence that M is reliable. For Steup, the needed evidence of reliability takes the form of track-record evidence. If S recalls that method M has regularly given rise to true beliefs, then S has memorial track-record evidence of M’s reliability.15 If, however, S realizes that M has regularly given rise to false beliefs, then S has memorial track-record evidence of M’s As Hugh LaFollette and Niall Shanks observe, “most of the decrease in mortality is traceable not to medical intervention but to preventive measures, especially improvements in diet and sanitation” (1993, 116). LaFollette and Shanks go on to note: These conclusions should not be surprising. We only need to examine U.S. health statistics. The life expectancy increased 43% from 1900 to 1950—before the advent of many of these medical treatments and vaccines. Since 1950 life span has increased only 7.4%. Moreover, since the rate of mortality from motor vehicle accidents has decreased by more than 20% since 1950, that accounts for most of the increase in life expectancy [since 1950]. (1993, 128, n. 14) One still might think that a 3.5 percent total decline in morality due to medical interventions is significant. It remains to be seen whether we have those medical interventions due to animal-based biomedical research or in spite of such research. That issue will be discussed in section 5. 15 On Steup’s view, S needn’t form the belief that M is reliable in order for M to be a source of justification, but S must possess memorial track-record evidence of M’s past successes for M to be a source of justification for S.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Epistemology and THE Ethics of Animal Experimentation 75 unreliability. Importantly, on Steup’s view, if S has good reason (in the form of memorial track-record evidence) to believe that a given method M is unreliable, then M is not a source of epistemic justification for S. In short, Steup maintains that evidence against a method M’s reliability prevents M from being a source of justification (even if M is in fact reliable).16 John Pollock (1986) denies the need for track-record evidence of reliability. He holds that S needn’t know, believe, or have evidence that a defeasible reason R for p is a reliable indicator of p in order for R to confer justification on p, but “[d]iscovering that the present circumstances are of a type in which a defeasible reason is unreliable constitutes a defeater for the use of that reason.”17 As he puts it elsewhere: “Lack of reliability is a defeater . . . [I]f we acquire a reason for thinking that our reasons are unreliable, that defeats our reasoning and prevents our reasons from conferring justification.”18 Keith Lehrer (1990) contends that in order for S to be justified in believing that p on the basis of method M, S must accept that M is a reliable and trustworthy method of forming beliefs.19 Elsewhere, I have defended the following weaker internalist constraint on justification: “S is justified in believing that p only if S does not believe that p has been formed by an unreliable process.”20 Despite the disparity of these competing internalist views on the relationship between reliability and justification, they all entail the following internalist constraint on justification and knowledge: (IC) If S knows that belief-forming method M is unreliable on the basis of track-record evidence of M’s unreliability, then M is not a source of justification (or knowledge) for S. We can see these entailments as follows: If one knows that method M is unreliable on the basis of track-record evidence of M’s unreliability, then one has track-record evidence of M’s unreliability, which, on Steup’s view, prevents M from being a
16 Steup (2004, 407). As he puts it: Evidence against its reliability prevents sense experience from being a source of justification even if it is in fact reliable, as shown by the Reverse Evil Demon World. Evidence for its reliability turns sense experience into a source of justification even if it is in fact unreliable. That is what the Standard Evil Demon World shows. So, what seems to matter is not reliability itself, but evidence regarding reliability. (2004, 407) 17 Pollock (1986, 189–90), emphasis mine. 18 Pollock (1984, 114), emphasis mine. Although Pollock focuses on the unreliability of reasons, not methods, nothing in the present chapter will turn on that distinction. Suppose S’s reason R for thinking that drug X will be safe in humans is the fact that X tested safe in mouse models. If it is discovered that mouse models are unreliable predictors of drug safety in humans, then that knowledge defeats R and prevents S from being justified in believing that X is safe in humans (on the basis of R). So, if we discover that the method of testing drugs on mouse models is an unreliable method, we will ipso facto discover that R is an unreliable reason. 19 Lehrer (1990, 168). 20 Engel (1986, 66).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
76 Mylan Engel Jr. source of justification. On Pollock’s view, if S knows that M is an unreliable method, then that knowledge defeats any justification that M would have conferred on any M-produced belief (i.e. knowledge of M’s unreliability defeats M as a source of justification for Pollock). Lehrer contends that knowledge requires acceptance. So, if S knows that M is unreliable, then S accepts that M is unreliable. If S accepts that M is unreliable, then S does not accept that M is a reliable and trustworthy method of forming beliefs. Consequently, if S knows that M is unreliable, then M is not a source of justification for S, on Lehrer’s view. The minimalist internalistic constraint that I have defended also implies IC. If S knows that M is unreliable, then S believes that M is unreliable. Since, on my view, M is a source of justification for S only if S does not believe that M is unreliable, S’s knowledge of M’s unreliability prevents M from being a source of justification for S. Despite the nuanced differences of their views concerning the relationship between reliability and justification, internalists are committed to the view that known unreliability of a belief-forming method is sufficient to prevent that method from being a source of justification. Here we reach a rare point in epistemology where both internalists and externalists agree. Both internalists and externalists accept an epistemic principle I call known unreliability: (KU) If we know that a method of acquiring belief is unreliable on the basis of track-record evidence of its unreliability, then beliefs produced by that method are not justified (absent additional independent support).21 Examples of methods known to be unreliable include wishful thinking, hasty generalization, and mere guesswork. Since we know that these methods are highly unreliable, these methods are not sources of justified belief. Consequently, beliefs produced by any of these three methods are not justified (absent additional independent support).
4.1 Upshot of section 4 If we know that a method M of acquiring belief is unreliable on the basis of extensive track-record evidence of its unreliability, then M is not a source of justification (or knowledge). So, if there is extensive track-record evidence that animal-based biomedical research is an unreliable method of forming beliefs about the pathology and proper treatment of human disease, then, contra the benefits argument, 21 We have just seen that internalists are committed to the epistemic principle (KU). Externalists are also committed (KU), because knowledge is factive. If we know that a method is unreliable, then that method is unreliable, and actually unreliable methods are not a source of externalistic justification.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Epistemology and THE Ethics of Animal Experimentation 77 animal-based biomedical research is not a source of justification (much less knowledge) of the pathology and proper treatment of human disease. If animal-based biomedical research fails to provide the epistemic benefits it is purported to provide, then such research is neither epistemically nor morally justified and should be abolished. The burden of section 5 is to assess the reliability of animal-based biomedical research.
5. The Epistemological Case against Using Animals in Biomedical Research As noted in section 1, virtually all biomedical experiments performed on animals are non-therapeutic and cause significant harm and distress to those animals.22 Many of these harmful experiments cause the animal subjects severe pain, and virtually all of these experiments are ultimately lethal, since the animals are routinely destroyed at the end of the experiment.23 Is there a good reason to subject animals to these experiments? That is, is there a reason to perform harmful, non-therapeutic biomedical experiments on animal subjects that is morally weighty enough to override these animals’ interests in avoiding such harm, suffering, and premature death? As noted in section 3, the benefits argument is the only serious reason offered in defense of animal-based biomedical research. So, if the benefits argument fails, then animal-based biomedical research is not justified and ought to be abolished. The question we must answer, then, is this: (Q1) Does animal-based biomedical research provide sufficient epistemic benefits regarding the origin, pathology, and proper treatment of human disease to justify the continued use of animals in harmful, painful, non-therapeutic, and ultimately lethal biomedical experiments? 22 Mere laboratory confinement itself is so stressful for the animals as to be properly regarded as a psychological harm (see Balcombe 2004, 6–8). The experiments themselves are even worse, as the animals experimented on are virtually always intentionally harmed in some physical way. Some animals are intentionally infected with pathogens. Other animals have diseases thought to model human diseases intentionally induced, including artificially induced coronary artery disease, artificially induced strokes, and artificially induced cancers. Still other animals are irradiated or burned or maimed in other ways, such as intentionally induced spinal cord injuries and intentional limb amputations. Conducting these experiments requires, by its very nature, intentionally harming the animal subjects involved. 23 I am not claiming that virtually all of the substances that animals are forced to ingest in these biomedical experiments prove lethal to the animal subjects involved. What I am claiming is that the entire experimental procedure proves lethal because the animals are routinely killed at the end of the experiments to check for signs of effectiveness or toxicity in organs, or simply because the animals have outlived their experimental usefulness. The few animals that are not killed after the initial experiment are typically “recycled,” i.e. used in additional animal experiments, until they outlive their usefulness, at which point they are killed.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
78 Mylan Engel Jr. Since methods known to be unreliable provide no epistemic justification for the beliefs they produce, the answer to (Q1) turns on the answers to the following questions: (Q2) Is animal-based biomedical research a reliable method of forming beliefs about the origin, pathology, and treatment of human disease? (Q3) Do animal models reliably predict how drugs will react in humans? If the answers to (Q2) and (Q3) are negative, then the answer to (Q1) must also be negative. To answer (Q2) and (Q3), we need to determine whether or not animal-based research has a track record of success at modelling the origin, pathology, and proper treatment of human disease. That is the burden of section 5.1.
5.1 The Unreliability of Animal Models Biomedical animal experiments can mislead in four major ways along two distinct vectors—the safety vector and the efficacy vector: (i) They can mistakenly predict that a drug is safe for humans; (ii) they can mistakenly predict that a drug is unsafe for humans; (iii) they can mistakenly predict that a drug is effective in humans; and (iv) they can mistakenly predict that a drug is ineffective in humans.24 Of course, these mistakes can be compounded, since a drug that tests safe and effective in animal models may prove to be both unsafe and ineffective in humans. In what follows, I will explore how frequently animal models mislead in these various ways in order to assess just how reliable/unreliable animal-based biomedical research is. Consider the first way that animal experiments can mislead along the safety vector. Animal experiments can mistakenly predict that a drug will be safe in humans (since it was found to be safe in animal models in preclinical testing), when that drug is actually unsafe (i.e. lethal, teratogenic, or toxic) in humans. Such misleading results are called “false negatives,” because the drugs test “negative” for harmful effects in animal models but are subsequently discovered to have seriously harmful effects in humans. To see some of the deleterious effects that false negatives can cause, consider the oft-cited thalidomide tragedy. Thalidomide causes birth defects in humans. Indeed, thalidomide’s teratogenic effect in
24 Surgical techniques tested on animal models can mislead along these same four vectors. For example, in dogs, when veins are grafted onto occluded arteries so as to bypass the occlusion, the venous graft dilates and aneurysm develops. Venous grafts are now routinely used in human bypass operations with no similar effects, because removing a section of a vein is much safer than removing a section of an artery. However, the dog experiments led many surgeons to perform the more dangerous arterial grafts in humans and delayed the use of the safer venous grafts in humans (Greek and Greek 2000, 174–5).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Epistemology and THE Ethics of Animal Experimentation 79 humans was recognized early on, on the basis of clinical observation—women who had taken thalidomide during their pregnancies gave birth to babies without limbs. But, because thalidomide’s teratogenecity could not be readily replicated in other animals, thalidomide continued to be prescribed to pregnant women. Eventually, after testing thalidomide for teratogenecity in countless animal species, scientists were able to demonstrate a teratogenic effect in one breed of rabbit, but only at doses between 25 and 300 times the dose given to humans.25 After still more testing, thalidomide was found to have a teratogenic effect in monkeys, but at ten times the normal human dose.26 Tragically, there was a lag-time of five years between the original reliable human clinical data that clearly demonstrated thalidomide’s teratogenic effect in humans and scientists’ ability to produce a similar teratogenic effect in some other species. During those five years, thalidomide continued to be prescribed to pregnant women because it hadn’t yet been found to be teratogenic in other species. The result: Over 10,000 babies were born without limbs because researchers and regulators relied on the unreliable animal data and ignored the more reliable human clinical data.27 Had thalidomide been pulled from the market on the basis of the human clinical data, these birth defects would not have occurred, but it wasn’t pulled, pending “confirmation” in other species. The thalidomide tragedy provides a powerful illustration of the untold human costs that can be caused by reliance on misleading animal experiments. Thalidomide is a clear example of a drug that falsely tested negative for harmful effects, but it is hardly unique in that regard. When Eli Lilly introduced Opren as a potential cure for arthritis, the drug had passed all of the usual animal tests, including prolonged tests on rhesus monkeys—tests in which the monkeys received up to seven times the maximum tolerated human dose for a year. Despite being touted as a new “wonder drug,” Opren was pulled from the market after it caused over 3,500 severe adverse reactions and 61 deaths.28 Fenclozic acid, another anti-arthritis drug, had no ill effects on more than 13 species, but imme25 Greek and Greek (2000, 45). 26 Ibid. 27 Greek and Greek (2000, 45) give the following poignant description of the thalidomide tragedy: A German pediatrician named Widikund Lenz was the first to suggest a link between thalidomide and teratogenesis . . . Mothers who had taken thalidomide gave birth to babies with often shocking deformities. Most lacked developed limbs. The first recorded case of phocomelia secondary to thalidomide occurred on Christmas Day, 1956, but in 1957 the drug was released anyway. A clinician from Australia, W.G. McBride, confirmed thalidomide’s dangers. Alarmed he, Lenz, and others wrote letters to the distinguished medical publication The Lancet, reporting phocomelia in infants of mothers taking thalidomide. As the incidences of deformity increased, scientists frantically attempted to reproduce teratogenesis from thalidomide in animals of all varieties. They gave thalidomide to scores of animals looking for proof in animals of what they already knew occurred in humans—that thalidomide could cross the placenta and drastically damage unborn offspring—and they could find none. Since animal testing had not indicated a problem with thalidomide, its use persisted. Hence, animal testing delayed the recall of this highly teratogenetic drug. 28 Sharpe (1989, 95); LaFollette and Shanks (1993, 118); Singer (2009, 57); Greek and Greek (2000, 68).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
80 Mylan Engel Jr. diately caused liver toxicity in humans.29 Here is just a partial list of false negatives and their harmful consequences for humans:
▪ Amrinone—a medication for heart failure, was tested in mouse, rat, h amster, guinea pig, dog, and rhesus monkey models without ill effect. In humans, it caused thrombocytopenia (a lack of blood cells needed for clotting) in 20 percent of the patients taking it, some of whom died. None of the animal models predicted the occurrence of thrombocytopenia.30 ▪ Benzene—an industrial chemical, was found to cause leukemia in humans, but not in any laboratory animals. Animal models failed to predict this important cause of leukemia.31 ▪ Butazolidin (phenylbutazone)—a non- steroidal anti- inflammatory drug (NSAID) commonly used to reduce pain and inflammation in horses, when used in humans, produced more than 10,000 fatal cases of aplastic anemia.32 ▪ Chloramphenicol—an antibiotic judged safe after animal experiments, caused aplastic anemia in humans, which often proved fatal.33 ▪ Clioquinol—an antidiarrheal drug that tested safe in rat, cat, dog, and rabbit models, caused blindness and paralysis in humans.34 ▪ Diethylstilbesterol (DES)—a synthetic estrogen thought to prevent miscarriages on the basis of animal models, in humans caused spontaneous abortion, premature birth, and neonatal death. DES was subsequently discovered to have a teratogenic effect as well. It increased the risk of vaginal and cer vical cancer in the daughters of the patients whose fetuses weren’t spontan eously aborted or killed by DES.35 ▪ Fialuridine—an anti-hepatitis drug that worked well in woodchucks, caused liver damage in seven of fifteen people, five of whom eventually died. The other two required liver transplants.36 ▪ Flenac (Fenclofenac)—an NSAID that passed animal toxicity tests in ten animal species (mice, rats, guinea pigs, ferrets, rabbits, cats, dogs, pigs, horses, and monkeys), produced severe liver toxicity in humans.37 ▪ Flosint—an arthritis medication that was tested in rat, monkey, and dog models, all of whom tolerated the medication well, caused eight human deaths.38 ▪ Grepafloxacin (Raxar)—an antibiotic that was pulled from the market by Glaxo Wellcome after it was linked with severe cardiovascular events in humans and seven human deaths, was tested in mouse, rat, rabbit, and dog models before being approved for human use.39
29 LaFollette and Shanks (1993, 118). 30 Greek and Greek (2000, 67). 31 LaFollette and Shanks (1993, 118). 32 Pippen and Stoick (2005, 1). 33 Sharpe (1989, 95). 34 Greek and Greek (2000, 67). 35 Ibid., 61. 36 Ibid., 63. 37 Pippen and Stoick (2005, 1). 38 Ibid., 67. 39 Kollmeyer (1999); Moore (1999); Stahlmann and Schwabe (1997).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Epistemology and THE Ethics of Animal Experimentation 81
▪ Isoprenaline—an aerosol asthma inhaler, killed 3,500 young asthmatics in Britain, but cats were able to tolerate doses 175 times greater than those found dangerous to asthmatics.40 ▪ Novantrone—a cancer treatment, produced heart failure in humans, an effect not predicted by animal models, despite extensive testing on dogs.41 ▪ Practolol—a heart treatment, caused blindness and twenty-three deaths in humans, despite extremely thorough testing in animal models. After the drug was withdrawn, despite repeated attempts, scientists were unable to reproduce these results in another animal species.42 ▪ Rezulin (Troglitazone)—a drug approved by the FDA in 1997 to treat type 2 diabetes after it was shown to lower blood sugar in rats without producing adverse effects, caused severe fatal liver failure in humans and was withdrawn in 2000 after 391 deaths were attributed to the drug.43 ▪ Zimeldine—the first selective serotonin reuptake inhibitor (SSRI), caused the paralyzing illness Guillain-Barré syndrome, not predicted by animal tests, and was pulled from the market. This, in turn, delayed the marketing of Prozac, which proved safe in humans.44 ▪ Zipeprol—a centrally acting cough suppressant, produced seizures and comas in some of the humans who took it. Animal tests had given no warning of the severe neurological problems caused by the drug, despite the use of higher doses in the animal subjects.45 Drs. Ray Greek and Jean Swingle Greek (2000) have identified at least forty other examples of false negatives. Several of these “animal-safe” drugs caused heart attacks in humans, others caused kidney failure and/or strokes in humans, still others caused liver failure in humans, and yet others caused seizures in humans.46 One final example—Vioxx—poignantly illustrates the magnitude of harm to humans that can result from reliance on unreliable animal models. Nine of eleven studies on mice and rats showed the COX-2 inhibitor Vioxx to be safe for animal hearts and blood vessels. Moreover, six additional animal studies involving four different species showed Vioxx to be protective against heart attacks and vascular disease.47 This animal-based data proved tragically misleading. In humans, Vioxx caused an estimated 320,000 heart attacks, strokes, and cases of heart failure worldwide—140,000 of them fatal—before being pulled from the market in 2004.48 How have animal models faired in the war on cancer? The U.S. National Cancer Institute’s 25-year study that tested 40,000 plant species on mice to determine 40 Sharpe (1989, 95–6). 41 Greek and Greek (2000, 64). 42 Singer (2009, 57); Greek and Greek (2000, 67–8). 43 Pippen and Stoick (2005, 1). 44 Greek and Greek (2000, 62). 45 Singer (2009, 57). 46 Greek and Greek (2000, 61–8). 47 Physicians’ Committee for Responsible Medicine (PCRM) (2005, 13). 48 Anderegg et al. (2006, 11).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
82 Mylan Engel Jr. their anti-tumor activity was an utter failure. Robert Sharpe describes the results of this study as follows: As a result of the programme several materials proved sufficiently safe and effect ive on the basis of animals tests to be considered for clinical trials. Unfortunately all of these were either ineffective in treating human cancer or too toxic for general use (Farnsworth and Pezzuto 1984). Thus in 25 years of this extensive programme not a single antitumour agent safe and effective enough for use in patients has yet emerged, despite promising results in animal experiments. Indeed one former cancer researcher has argued that clues to practically all the chemotherapeutic agents which are of value in the treatment of human cancer were found in a clinical context rather than in animal studies. (Sharpe 1989, 91–2)
In this extensive, decades-long study, none of the plant-derived materials found safe and effective in treating mouse cancers proved safe and effective in treating human cancers.49 That’s a 100 percent failure rate. These unimpressive results are hardly surprising since, as Sidney Gendin explains, “Most mouse cancers are sarcomas (cancers arising in the bone, connective tissue, or muscle) and most human cancers are carcinomas (cancers arising in membranes).50 As a result, mice, the predominant species used in cancer research, make very poor models for understanding and treating human cancers. Even the industry’s own Lab Animal Magazine (2001) admits: “Mice are actually poor models of the majority of human cancers.”51 This knowledge is not new. The highly esteemed medical journal Lancet presciently warned, as far back as 1972, that, since no animal tumor is closely related to any cancerous tumor in humans, an agent which is active in the laboratory may well prove useless clinically.52 Cancer research isn’t the only area where animal-based research has proven ineffective and unreliable. Since the early 1980s, animal models have been used extensively in AIDS research, despite the fact that no other known species develops human AIDS. Much of this research has been conducted on chimpanzees 49 When one moves beyond plant species to all chemical compounds tested on animals in the war on cancer, the results are equally dismal. Greek and Greek describe the results: Between 1970 and 1985, fervent researchers subjected animals to over half a million compounds for anticancer effects. 500,000. Based on these animal experiments, only eighty compounds progressed to clinical trials. A mere twenty-four proved to have any anticancer activity in humans. Of the twenty-four, twelve went on to have a substantial role in chemotherapy . . . all twelve of these were analogues or chemical variations of previously-known chemotherapeutic agents. (Greek and Greek 2000, 128–9) 50 Gendin (1989, 203). Richard Ryder makes a similar point, noting that animals used in cancer research “perish miserably each year despite the fact that human cancer tumours cannot satisfactorily be produced in other species” (1985, 84–5). 51 Lab Animal Magazine 30 (June 2001): 13, cited in Anderegg et al. (2006, 3–4). 52 Lancet (1972, 827–8).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Epistemology and THE Ethics of Animal Experimentation 83 because chimpanzees are the only species to maintain the virus in their bodies after being injected with HIV.53 Since 1987, the National Institute of Allergy and Infectious Disease has funded over 100 clinical trials to test potential HIV vaccines. Every one of the 50-plus preventive vaccines and 30-plus therapeutic vaccines that were found safe and immunogenic against HIV/AIDS in primate studies has failed in human clinical trials.54 Yet another 100 percent failure rate. That is as unreliable as one can get.55 Animal experimentation advocates insist that we need to test drugs on animals to ensure that they are safe for human use. Every drug currently approved for human use by the FDA has undergone a battery of tests on a wide number of animal models. Despite all of that animal testing, adverse drug reactions are the fourth leading cause of death in the United States (behind only heart disease, cancer, and stroke).56 Every year, two million Americans are hospitalized and 100,000 die from the harmful effects of prescription drugs, at a cost of $136 billion per year.57 Worse still, the Food and Drug Administration (FDA) estimates that phys icians only report 1 percent of adverse drug reactions.58 So, the number of people who die each year from adverse drug reactions is probably much higher than the 100,000 deaths reported. Clearly, tests on animal models aren’t protecting us from adverse drug reactions. Thousands of people are dying each year because drugs react differently in people than they do in animal models.59 Drugs that test safe and effective in animal models can also mislead along the efficacy vector. Drugs found effective in animal models often prove ineffective in humans. For example, twenty-two drugs have been shown to be therapeutic for spinal cord injury in animal models, but not a single one of these drugs is
53 Sharpe (1989, 92). Despite retaining the virus in their bodies, chimpanzees do not go on to develop AIDS. 54 National Institute of Allergy and Infectious Diseases (2005). See also Anderegg et al. (2006, 4–5). 55 Recently, the macaque has become the non-human-primate species of choice for research on HIV/AIDS. The shift to the macaque was made because of the failure of chimpanzee-based research. The human immunodeficiency virus HIV-1 does not survive in macaques. So, researchers have created a new strain of virus, simian-tropic (st)HIV-1, by replacing the HIV-1 vif gene with the vif gene from simian immunodeficiency virus SIVMAC239. The new virus (st)HIV-1 survives in macaques, but it does not result in AIDS-like symptoms, unless the macaques are depleted of their CD8+ T cells. So, the macaque research is studying a different virus, and the new virus does not cause AIDS-like symptoms in macaques unless their immunes systems are artificially depleted of CD8+ T cells via anti- CD8-antibody infusion (Hatziioannou et al. 2009)—not exactly a promising model for the original HIV-1 virus as it functions in humans. 56 Regan (2012, 109). 57 Greek and Greek (2000, 16); Regan (2012, 109). 58 Regan (2012, 109). 59 One might think that, with this many people dying from adverse drug reactions each year, we need to test drugs on even more species so as to rule out unsafe drugs before they reach human trials, but such a procedure would dramatically limit our access to safe and effective drugs. Virtually all medications in use today “can be found to cause a serious side effect in some animal” (Greek and Greek 2000, 71). So, testing drugs on even more species would only exacerbate the problem of false positives. It would cause us to reject even more safe and effective drugs than we currently do. LaFollette and Shanks (1993, 118) make a similar point.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
84 Mylan Engel Jr. effective in humans.60 Call such results “false efficacy predictors.” False efficacy predictors don’t result in direct harm to humans. The drugs simply fail to work in humans. There are indirect harms, however. The billions of research dollars spent annually developing drugs that are safe and effective in animal models, but routinely found to be unsafe or ineffective in humans are wastefully being pumped into false leads and dead ends. These dollars could be far better spent both on effective prevention campaigns and on developing synthetic human models through in vitro reconstitution of disease-related human cell types or tissues that would more safely and reliably predict human response.61 If false negatives and false efficacy predictors were relatively rare, say 1 out of 100 or even 1 out of 10, animal experimentation might still be a reliable method of modeling drug safety and effectiveness in humans,62 but false negatives and false efficacy predictors are not rare. They are the norm. In prepared remarks presented on January 12, 2006, Andrew C. von Eschenbach, Acting Commissioner of the FDA, had this to say: “Consider just one stark statistic: Today, nine out of 10 compounds developed in the lab fail in human studies. They fail, in large part because they behave differently in people than they did in animal or laboratory tests.”63 Actually, the current failure rate of drugs that make their way to Phase I human clinical trials on the basis of preclinical animal testing is 92 percent.64 This clinical failure rate is split roughly equally between drugs that are too toxic and drugs that don’t work in humans. Accordingly, of all the drugs that make their way to human clinical trials on the basis of preclinical animal testing, approximately 46 percent prove to be false negatives and roughly 46 percent fall into the category of false efficacy predictors, barely better than flipping a coin along either vector. With its 92 percent clinical failure rate, preclinical animal testing is known to be an extraordinarily unreliable method of establishing the safety and effectiveness of pharmaceutical compounds in humans. Given its known extreme unreli ability, it follows from principle (KU) that preclinical animal testing is not a source of justification regarding the safety and effectiveness of drugs in humans. Contra Cohen, the fact that a drug has tested safe and effective in animal models provides no evidence as to how that drug will react in humans.65 60 Greek and Greek (2004, 18). 61 Just how much money are we talking about? Lester Crawford, Acting Commissioner of the Food and Drug Administration (FDA) in 2004, estimates that it takes a staggering $1.7 billion to produce one FDA-approved drug (Crawford 2004, 2). 62 I say ‘might still be a reliable method’ because whether animal models are reliable would still depend on how often they falsely predict that safe drugs are harmful and on how often they falsely predict that effective drugs are ineffective. 63 von Eschenbach (2006, 1). 64 In prepared remarks, Lester Crawford, Acting Commissioner of the FDA in 2004, reported that only 8 percent of the drugs that test safe and effective in animals, prove safe and effective in humans (Crawford 2004, 2). Also see Pippen (2007). 65 With roughly half of these failures being false negatives, relying on animal models to establish the safety of drugs is not only unreliable, it is also a dangerous way of testing drugs, since it puts the human subjects in Phase I, Phase II, and Phase III clinical trials at serious risk of grave harm and/or death.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Epistemology and THE Ethics of Animal Experimentation 85 Preclinical animal testing can mislead along the safety vector in the other direction as well. Pharmacological agents that are safe in humans can test unsafe (i.e. fatal, teratogenic, or toxic) in animal models. Such results are called “false positives,” since the drugs test “positive” for harmful effects in animals, despite being safe in humans. False positives mistakenly predict that a drug is unsafe for humans, when actually that drug is perfectly safe in humans. As for the efficacy vector, drugs that are completely ineffective in treating disease X in animal models can be highly effective in treating disease X in humans. I call these misleading results “false inefficacy predictors.” By my lights, animal-based false positives and false inefficacy predictors pose a far greater risk to human health and well-being than false negatives and false efficacy predictors, because, unlike the latter, false positives and false inefficacy predictors can, and do, result in drugs that are safe and effective in humans—lifesaving panaceas for human diseases—being pulled from development before ever making their way to human clinical trials. As long as we test potential drugs using unreliable animal models and table the development of drugs that might have proven safe and effective in humans on the basis of their deleterious effects in these animal models, we will inevitably continue to forego certain cures for human diseases. Your parent, partner, or child, now dying of cancer, might very well have been saved by a currently existing drug, which is both safe and effective in treating cancer in humans, that was shelved before making it to human clinical trials because it made rats and mice sick. We discover false positives in one of three ways: (i) Drugs that have been rejected in the United States by the FDA on the basis of adverse reactions in animal models often end up being approved for human use in other countries. When those drugs prove safe and effective for humans in those other countries on the basis of reliable human clinical data, the FDA sometimes reverses its earlier decision and grants the drug approval; (ii) drugs that have been approved for human use and found to be safe and effective in humans on the basis of post-marketing human clinical surveillance are sometimes subsequently tested on animal species that they weren’t previously tested on and found to be unsafe in one or more of those other species; and (iii) adverse drug reactions in animal models are sometimes simply ignored when they indicate an unwanted conclusion.66 In these cases, the drugs advance to human clinical trials despite their adverse reactions in animal models. Like their false-negative counterparts, false positives are not rare. Indeed, you probably have many such drugs in your medicine cabinet right now. Here are just a few examples of what are now known to be false positives: ▪ Acetaminophen (Tylenol)—causes renal failure and death in cats in low doses.67 ▪ Acetylsalicylic acid (Aspirin)—is teratogenic in rats, mice, dogs, rabbits, and monkeys, and poisons cats (but has no effects whatsoever in horses).68
66 Millstone (1989, 84). 67 Greek and Greek (2000, 71). 68 LaFollette and Shanks (1993, 118); Greek and Greek (2000, 71); Pippen and Stoick (2005, 2).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
86 Mylan Engel Jr.
▪ Cortisone—a widely prescribed corticosteroid, is teratogenic in mice.69 ▪ Depo-provera—a widely prescribe birth control pill, was rejected by the US FDA in 1973 because it causes cancer in dogs and baboons. Elsewhere around the world, women used the drug and found it safe and effective. Not until 1993 did the FDA approve the drug for use in the United States, despite extensive human clinical data of its safety and effectiveness. Reliance on misleading animal data denied US women access to this safe and effective birth control pill for twenty years.70 ▪ Digitalis—a heart medication that was discovered without the use of animal experiments, is successfully used to treat congestive heart failure and irregular heartbeat in humans, but its widespread use was delayed because animal models incorrectly predicted a dangerous rise in blood pressure.71 ▪ Furosemide (Lasix)—an important diuretic that for decades has been safely and effectively used in humans to reduce fluid retention during heart failure and other diseases, causes extensive liver damage in mice.72 ▪ Fluoride—a mineral routinely added to tap water and toothpaste to strengthen teeth and prevent cavities, causes cancer in rats.73 ▪ Ibuprofen (Advil, Motrin)—causes liver failure in dogs at low doses.74 ▪ Insulin—a lifesaving medication that is essential for regulating blood sugar in human diabetics, produces deformities in some laboratory animals.75 ▪ Morphine—is an opioid that sedates humans but stimulates cats. ▪ Prednisone—another routinely prescribed corticosteroid, causes cancer in some rodents. ▪ Streptomycin—a commonly prescribed antibiotic, is teratogenic in rats.76
One other example—penicillin—is worth noting because it drives home the potential risk to human health that false positives pose. The isolation of penicillin is widely regarded as one of the greatest medical discoveries of the twentieth century. With it, the age of antibiotics was born. Collectively, antibiotics have saved more lives than all other medical advances combined. Penicillin alone is estimated to have saved between 80 million and 200 million lives. Alexander Fleming, who observed penicillin kill bacteria in petri dishes, tested it on rabbits without ill effect. It was serendipitous indeed that he tested penicillin on rabbits rather than rodents. Penicillin is teratogenic in rats, and it kills guinea pigs and Syrian hamsters. Had Fleming tested penicillin on these rodent species, it probably would have never been approved for human use, and the age of antibiotics
69 Greek and Greek (2000, 72). 70 Ibid., 71–2. 71 Sharpe (1989, 97); Greek and Greek (2000, 72). 72 Pippen and Stoick (2005, 2). 73 Greek and Greek (2000, 74). 74 Ibid., 71. 75 LaFollette and Shanks (1993, 118). 76 Greek and Greek (2000, 70–6).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Epistemology and THE Ethics of Animal Experimentation 87 might never have come into being.77 Luck, and luck alone, prevented penicillin from being a permanently shelved false positive. If Fleming had unwittingly tested penicillin on rats and guinea pigs, somewhere between 80 million and 200 million people would have died from infections that could have been treated with penicillin. No doubt, we haven’t been so lucky with other life-saving pharmaceutical compounds that are now collecting dust in a laboratory somewhere, because they tested unsafe or ineffective in rodent species. There is no easy way to determine just how many safe and effective drugs for human diseases have been shelved by pharmaceutical companies because they either proved ineffective in other species or had adverse reactions in other species, partly because that information remains proprietary and partly because drugs that test highly toxic in animal models rarely make their way to human clinical trials, but there are reasons to think that the number of such drugs is probably quite large. For example, when the National Cancer Institute tested on mice 12 anti-cancer drugs currently being successfully used in humans, they found that 30 out of 48 times the drugs were ineffective in mice—a false inefficacy predictor rate of 63 percent.78 Animal models also prove terribly unreliable when it comes to testing substances for carcinogenicity, which makes them poor models for understanding the origin of many human cancers. Of 20 compounds known not to cause cancer in humans, 19 did cause cancer in animals—a false positive rate of 95 percent.79 Couple this information with the fact that 98 percent of the compounds tested in preclinical animal trials are killed by pharmaceutical companies before making their way to human clinical trials, largely on the basis of animal data80 and one can start to see the magnitude of the risk to human health posed by reliance on unreliable animal models. For every 600 drugs that enter preclinical testing on animals, only 12 ever advance to human clinical trials, and only 1 of these 12 receives FDA approval. That means for every 1 FDA-approved drug, 588 drugs are pulled from development by pharmaceutical companies without ever being tested on humans. In 2016, the FDA approved 22 new drugs. So, over 12,900 drugs were pulled from the market without ever being tested on humans. If the false positive rate were only 10 percent, 1,290 of these shelved drugs would have been safe for use in humans! Of the 98 percent of chemical compounds that are discarded by pharmaceutical companies, due to their toxicity or ineffectiveness in animal models, we will never know how many of them would have proven safe and effective in humans (unless we test them on humans), but there are good reasons to think that many of them would have been safe and effective in humans. 77 Fleming himself made a similar observation: “How fortunate we didn’t have these animal tests in the 1940s, for penicillin would probably never have been granted a license, and possibly the whole field of antibiotics might never have been realized” (cited in Greek and Greek 2000, 73). 78 Greek and Greek (2004, 17). 79 Ibid., 18. 80 Pippen (2007).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
88 Mylan Engel Jr. Consider one such reason: Occasionally, drugs that test toxic in animals make their way to human clinical trials anyway (cases where the animal data is simply being ignored). In August 2001, Mark Levin, CEO of Millennium Pharmaceuticals, presented data at the Drug Discovery Technology Conference in Boston that suggests just how prevalent false positives are. In the study that Levin presented, 28 potential new drugs were tested in rats for hepatotoxicity (i.e. liver toxicity). Seventeen of these drugs tested non-hepatotoxic in rats, and 11 tested hepatotoxic in rats. Twenty-two of these drugs advanced to human clinical trials anyway—14 of the 17 that tested safe and 8 of the 11 that tested positive for hepatotoxicity advanced. Of the eight that had tested hepatotoxic in rats, six were found safe (i.e. non-hepatotoxic) in humans—a false positive rate of 75 percent!81 This study provides compelling evidence that rat models are highly unreliable predictors of hepatotoxicity in humans. Since known unreliable methods are not a source of justification, the fact that a drug is hepatotoxic in rat models provides no evidence that that drug will be hepatotoxic in humans. While no one would suggest that 75 percent of the 98 percent of discarded drugs would have proven safe and effective in humans (since the tests reported by Levin focused solely on one form of toxicity), the large number of false positives of which we are already aware make it extremely likely that many of the discarded drugs would have proven safe and effective in humans, had we only tested them on humans and not relied on misleading data derived from highly unreliable animal models. We will never have access to those drugs because of our reliance on unreliable animal models. All those people whose illnesses would have been cured or ameliorated by those drugs unfortunately must continue to suffer from their illnesses because of our reliance on animal-based biomedical research.82 In these cases, far from making us better, the animal research is keeping us sick.
5.2 Why Animals Are Such Poor Models of Human Disease For decades, researchers have extensively used murine (i.e. mouse, rat, and related rodent) models to test drug candidates for subsequent human clinical trials, and very few of these trials have shown any success.83 Murine models are chosen not because of their phylogenetic similarity to humans, but because they are cheap, reproduce quickly, are easy to handle, have short lifespans, are standardized, and 81 Greek and Greek (2004, 17–18). Of the 14 drugs that tested safe (i.e. non-hepatotoxic) in rats that were subsequently tested in humans, 6 proved hepatotoxic in humans—a false negative rate of 43 percent. 82 This human suffering should also be factored into any cost–benefit analysis of animal research. If we are looking at all the relevant harms and benefits, then the considerable amount human suffering that continues because of misleading false positives should be counted among the substantial costs of animal-based biomedical research—something that is rarely mentioned in the literature. 83 Seok et al. (2013, 3507).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Epistemology and THE Ethics of Animal Experimentation 89 are resistant to intercurrent infections and because there is no strong lobby seeking to protect their interests.84 Erik Millstone (1989) has identified four conditions that must be met if animal models are to inform us of the likely effects of compounds on human health: (1) There must be “sufficient similarity between the anatomy, physiology and metabolism of experimental animals and those of humans to provide a basis for reliable extrapolation.” (2) There must be “a satisfactory match between lifestyles of the laboratory animals and the diverse circumstances of the human population whose health the testers wish to model and protect.” (3) There must be “well-defined systematic and valid rules for extrapolating from the results of animal tests to conclusions about humans.” (4) “Those rules of extrapolation need to be reliably and diligently followed by official regulatory institutions.”85 Millstone further observes that if conditions (1) and (2) are not met, then condition (3) cannot be met. Rats clearly fail to satisfy condition (1). As Robert Sharpe points out, comparisons of rats with humans “reveal major differences in skin characteristics, respiratory parameters, the location of gut flora, β-glucuronidase activity, plasma protein binding, biliary excretion, metabolism, allergic hypersensitivity and teratogenicity.”86 Greek and Greek explain how these physiological differences affect drug absorption and metabolism: Rats have no gall bladder. They excrete bile very effectively. Many drugs are excreted via bile so this affects the half-life of the drug. Drugs bind to rat plasma much less efficiently. Rats always breathe through the nose. Because some chemicals are absorbed through the nose, some are filtered. . . . Their gut flora are in a different location. Their skin has different absorptive properties than that of humans. Any one of these discrepancies will alter drug metabolism. (Greek and Greek 2000, 59)
Despite these gross physiological and biological differences, researchers and regulatory officials continue to assume that mice, rats, and other rodents accurately model human disease.87 Junhee Seok, a biomedical informatics specialist, and Dr. H. Shaw Warren, a specialist in infectious diseases at Harvard Medical School, put that assumption to the test. In their 2013 study (published in PNAS), they systematically evaluated, on a molecular basis, how well murine clinical models 84 Sharpe (1989, 97); Millstone (1989, 84). 85 Millstone (1989, 73). 86 Sharp (1989, 97). 87 Seok et al. (2013, 3507).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
90 Mylan Engel Jr. mimic human inflammatory diseases in patients. They used genome-wide expression analysis to compare human and murine-model genomic responses to inflammatory diseases and found that acute inflammatory stresses from different etiologies (i.e. burns, trauma, and endotoxemia) result in highly similar genomic responses in humans (R2 = .91), whereas the corresponding genomic responses (to burns, trauma, and endotoxemia) in mouse models correlate poorly with the human responses (R2 = 0.00–0.13). In particular, they found that individual gene activation in the human conditions was not predicted by the ortholog in the corresponding mouse model in either direction or magnitude.88 Quite the contrary: “Among genes changed significantly in humans, the murine orthologs are close to random in matching their human counterparts.”89 Since murine models react so differently to inflammatory stress, at the molecular level, they fail to satisfy Millstone’s first necessary condition for informative research.90 Murine models are not sufficiently similar to humans physiologically or cellularly to provide a basis for reliable extrapolation. This fact should come as no surprise once one compares human sensitivity to endotoxins with mouse resilience to endotoxins. The lethal dose of endotoxin for most strains of mice (5–25 mg/kg) is 1,000,000-fold greater than that of humans—as little at 30 ng/kg has been reported to cause shock in humans.91 These differences explain why every one of the nearly 150 clinical trials on murine models testing candidate agents intended to block the inflammatory response in critically ill humans has failed.92 With their 100 percent failure rate, murine models have proven to be maximally unreliable models of human inflammatory response and disease. Based on the findings of their systematic comparative analysis, Seok, Warren, et al. conclude with the following recommendation: the development of synthetic human models by in vitro reconstitution of disease-related cell types or tissues might similarly improve current disease models. . . . our study supports higher priority to focus on the more complex human conditions rather than relying on mouse models to study human inflammatory diseases. (Soek et al. 2013, 3507)
88 Ibid., 3510. 89 Ibid., 3507. 90 Millstone’s second necessary condition for informative research also is not met. Millstone explains why: while the human population is diverse in terms of genetics, health, diets, exposures, and environments, laboratory animal populations are often genetically uniform and free of environmental pathogens. They have access to unlimited quantities of food despite being confined in small cages which preclude much exercise. When the safety or efficacy of compounds are tested, this is almost always accomplished by adding one test substance at a time to animal diets, while patients receive several drugs at any one time . . . The possible effects of these chemical cocktails are not replicated in animal studies. (1989, 77) 91 Seok et al. (2013, 3511). 92 Ibid., 3507.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Epistemology and THE Ethics of Animal Experimentation 91
5.3 The Answers We are now in a position to answer the questions posed at the outset of section 5. (Q3) Do animal models reliably predict how drugs will react in humans? (A3) No. Quite the contrary, animal models are extremely unreliable at predicting how drugs will behave in humans. They routinely mislead along both the safety and efficacy vectors. They are extremely unreliable at predicting whether or not candidate drugs will be safe in humans. They are also extremely unreliable at predicting whether or not candidate drugs will be effective in humans. (Q2) Is animal-based biomedical research a reliable method of forming beliefs about the origin, pathology, and treatment of human disease? (A2) No. Animal- based biomedical research has a proven track record of unreliability when it comes to determining the origin, pathology, and treatment of human disease.93 (Q1) Does animal- based biomedical research provide sufficient epistemic benefits regarding the origin, pathology, and proper treatment of human disease to justify the continued use of animals in harmful, painful, nontherapeutic, and ultimately lethal biomedical experiments? 93 If biomedical animal experimentation is as unreliable at determining the origin, pathology, and proper treatment of human disease as I have argued, why do researchers continue to conduct such experiments? There are a number of reasons that such research continues, and none them have anything to do with the epistemic value of that research. First, animal experimentation is big business. It grosses over $100 billion per year (with some estimates as high as $1 trillion per year). With so much money at stake, the corporations that profit from the research have powerful lobbies that influence policy makers and regulators. This explains why the FDA mandates that all new pharmacological agents be tested in animal models in preclinical trials prior to human clinical trials, despite the fact that several FDA commissioners have acknowledged unreliability of such data. Second, animal experimentation is an entrenched technology that is handed down from research scientists to graduate students. Once a graduate student has perfected a scientific technique, say, learning how to implant electrodes in cat brains, they want to put that new skill to use. Third, like other university professors, professors in the sciences must publish or perish, and animal experiments make for easy publications. All one need do is take an experiment that has already been published and change one variable, say, the type of animal used or the amount of the drug administered, and publish the new results. Finally, and perhaps most importantly, corporations are never eager to give up possible legal protections. This helps to explain why, even though the FDA does not require that household products, personal care products, and cosmetics be tested on animals, many companies continue to test their products on animals. As Rod Preece and Lorna Chamberlain explain, “Companies that prefer using in vitro techniques because of their greater scientific reliability, reduced costs, and convenience, often repeat the tests on intact animals (despite the unreliability and invalidity of the latter tests) solely as a means of further protecting themselves against product-liability lawsuits” (1995, 70–1). Even in the case of biomedical research, Dr. James G. Gallagher, Director of Medical Research, Lederle Laboratories, acknowledged this as early as 1964: “Animal studies are done for legal reasons and not for scientific reasons. The predictive value of such studies for man is often meaningless—which means our research may be meaningless” (JAMA 1964, 35). Once product manufacturers, pharmaceutical companies, and the FDA start being held legally liable for the harms caused to humans by their unscientific reliance on demonstrably unreliable animal models, as they should be, their use of such models will likely begin to drop.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
92 Mylan Engel Jr. (A1) No. Since methods that are known to be unreliable are not sources of justification, evidence, or knowledge, and since animal-based research is known to be an extremely unreliable method for determining the origin, pathology, and proper treatment of human disease, animal-based research fails to provide the epistemic benefits needed to justify its continued use.94
5.4 Medical Research without Animal Models The only way to know how a drug will react in humans is to test that drug on humans. Researchers and regulators are keenly aware of this fact. That, of course, is why every drug must pass Phase I, Phase II, and Phase III human clinical trials before being approved for human use by the FDA. As noted above, 98 percent of the pharmacological compounds tested in animal models are pulled by pharmaceutical companies before making their way to human clinical trials. Instead of giving us knowledge of how these substances will react in humans, the animal-model paradigm actually prevents us from gaining such knowledge. Instead of providing us with epistemic benefits essential to curing human disease, by preventing us
94 Throughout this chapter, I have compiled extensive track-record evidence that demonstrates that animal-based biomedical research is an unreliable method both for determining the origin, path ology, and proper treatment of human disease and for determining how pharmacological compounds will behave in humans; and I have argued that since we know that animal-based biomedical research is an unreliable method based on this track-record evidence, such research is not source of justification or evidence about the origin and proper treatment of human disease. A referee expressed the following concern: Should animal-based biomedical research be viewed as a single method? Proponents of animal research might object that all manner of different methods are used across animal research and that it is thus incorrect to think of “animal research” as a single method. They might, in turn, argue that the ‘known unreliability’ principle applies to some, but not all forms of biomedical animal experimentation. This concern, of course, is the generality problem as applied to the epistemology of animal experimentation. Should we individuate methods coarsely and treat biomedical animal-based research as one monolithic method, or should we individuate methods more finely and, for example, treat rat-based research, mouse-based research, and chimpanzee-based research as different methods? The evidence compiled in this chapter shows that animal-based biomedical research, whether viewed monolithically as one overarching method or viewed severally as numerous distinct species– specific methods, is unreliable both in determining the origin, pathology, and proper treatment of human disease and at predicting how potential drugs will react in humans. The data provided show that rats, mice, hamsters, guinea pigs, rabbits, woodchucks, cats, dogs, horses, rhesus monkeys, macaques, baboons, and chimpanzees are unreliable models for human disease and unreliable pre dictors of drug safety and effectiveness in humans. As noted earlier in this section, in his prepared remarks, Lester Crawford, Acting Commissioner of the FDA in 2004, reported that only 8 percent of the drugs that test safe and effective in animals, prove safe and effective in humans (Crawford 2004, 2) Any method or collection of methods that mistakenly predicts that drugs will be safe and effective in humans 92 percent of the time is an extremely unreliable (and dangerous) method. And that 92 percent failure rate only takes into account the false negatives and false efficacy predictors. Crawford’s numbers don’t factor in the mispredictions generated by false positives and false inefficacy predictors, which render animal models even more unreliable than Crawford suggests. The bottom line: Animal-based biomedical research is an extremely unreliable failed technology, no matter how finely you slice it.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Epistemology and THE Ethics of Animal Experimentation 93 from testing the vast majority of pharmacological compounds in humans, the animal-model paradigm ensures that we remain ignorant as to how these drugs behave in humans. Should we then test drugs that have proven unsafe or ineffective in animal models directly on humans? We should, if we want to know how those drugs will react in humans. But can we test such drugs on humans without putting them at grave risk of harm or death? The answer is “Yes we can.” We can begin with in vitro tests on human tissues and cell layers and on human stem cells. Testing a candidate drug directly on human tissue in vitro provides information about how that drug will be absorbed and metabolized by human tissue. In vitro techniques were proven to be superior to animal studies at predicting human response over a decade ago and are now being used to identify disease mechanisms, drug targets, drug efficacy, and drug toxicity in virtually all types of human tissue.95 Candidate drugs can also be tested in silico using computer and mathematical models based on existing human clinical knowledge. These in silico tests provide important absorption, distribution, metabolism, excretion, and toxicity (ADMET) data that allow us to predict how a new drug candidate is likely to react in humans. In silico technology provides human ADMET predictions whose accuracy rivals that of in vitro methods.96 At some point, candidate drugs must be tested on intact organisms. Regardless, whether we test drugs on intact animal models (and get unreliable data) or not, eventually we must test drugs directly on humans, which is exactly what we do in Phase I, Phase II, and Phase III human clinical trials. Since the animal data cannot be reliably extrapolated to humans, we should stop collecting such data. Instead, we should bypass animal tests entirely and test those drugs that have demonstrated promise in in vitro and in silico tests directly on humans. Such tests can be performed safely in Phase 0 trials using microdosing technology. As Dr. John Pippin explains, human microdosing technology uses: radiolabeled trace doses (1-100 mcg) of candidate drugs to evaluate absorption, distribution, metabolism, and excretion in humans. These doses are less than 1 percent of that required to produce a pharmacological effect, thus there is virtually no risk for adverse effects. The radiation exposure is less than that obtained in a four-hour airplane flight. Positron emission tomography is used to acquire real-time data regarding drug disposition, and accelerator mass spectrometry is used to analyze parent drug and metabolite concentrations in blood, urine, and feces at specific intervals after dosing. (Pippin 2005, 16)
Microdosing technology provides accurate information about how a candidate drug will be absorbed, distributed, metabolized, and excreted in humans—something 95 Pippin (2005, 17).
96 Ibid, 16.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
94 Mylan Engel Jr. that no animal test can do. And, yes, it is safe. Microdosing technology “was endorsed by the European Agency for the Evaluation of Medicinal Products in January 2003 . . . , and has already been used to identify drug candidates for human Phase I trials.”97 We don’t need to rely on outdated, unreliable, malpredictive animal models to make significant advances in the treatment of human disease.
6. Conclusion We can learn a great deal about the ethics of animal experimentation by looking at the epistemology of animal experimentation. Nonhuman animals are demonstrably unreliable models of human disease. Since we know that animal-based research is an unreliable method for determining the origin and proper treatment of human disease, such research is not a source of evidence or justification (much less knowledge) about the origin and proper treatment of human disease. Given its unreliability, animal experimentation does not and cannot provide the epi stemic benefits needed to outweigh the harms inflicted on the animal subjects involved. A careful look at the epistemology of animal experimentation shows that such research is neither epistemically nor morally justified and should be abolished.
References Anderegg, Christopher, Kathy Archibald, Jarrod Bailey et al. (2006). “A Critical Look at Animal Experimentation.” www.mrmcmed.org/critical_look.pdf. Balcombe, Jonathan (2004). “Lab Stress 24/7.” Good Medicine 13(4): 6–8. Carruthers, Peter (1992). The Animals Issue. Cambridge: Cambridge University Press. Cohen, Carl, with T. Regan (2001). The Animal Rights Debate. Lanham, MD: Rowman and Littlefield. Crawford, Lester (2004). “Speech before Global Pharmaceutical Strategies Seminar.” Presented May 25, 2004. Engel Jr., Mylan (1986). “Coherentism Reliabilized.” Acta Analytica 2: 49–77. Garrett, Jeremy (2012). “The Ethics of Animal Research: An Overview of the Debate,” in Jeremy Garrett (ed.), The Ethics of Animal Research: Exploring the Controversy. Cambridge, MA: MIT Press, 1–15. Gendin, Sidney (1989). “The Use of Animals in Science,” in Tom Regan and Peter Singer (eds), Animal Rights and Human Obligations, 2nd edn. Englewood Cliffs, NJ: Prentice Hall, 197–208. 97 Ibid, 16.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Epistemology and THE Ethics of Animal Experimentation 95 Goldman, Alvin (1979). “What Is Justified Belief?,” in George S. Pappas (ed.), Justification and Knowledge. Dordrecht: D. Reidel, 1–23. Greek, C. Ray and Jean Swingle Greek (2000). Sacred Cows and Golden Geese: The Human Cost of Experimenting on Animals. New York, NY: Continuum. Greek, Jean Swingle and Ray C. Greek (2004). What Will We Do If We Don’t Experiment on Animals? Victoria, Canada: Trafford. Harman, Gilbert (1977). The Nature of Morality: An Introduction to Ethics. New York, NY: Oxford University Press. Hatziioannou, Theodora, Zandrea Ambrose, Nancy P. Y. Chung et al. (2009). “A Macaque Model of HIV-1 Infection.” Proceedings of the National Academy of Sciences [PNAS] 106(11): 4425–9. Hendee, William R. (1988). Use of Animals in Biomedical Research: The Challenge and Response: AMA white paper. Chicago, IL: American Medical Association. JAMA (1964). “Drug Industry Uncertain But Hopeful.” Journal of the American Medical Association 187(11): 35. DOI: http://doi.org/10.1001/jama.1964.03060240109061. Kollmeyer, Barbara (1999). “Glaxo Pulls Rexar [sic] because of Safety Concerns.” MarketWatch (October 27, 1999). LaFollette, Hugh and Niall Shanks (1993). “Animal Models in Biomedical Research: Some Epistemological Worries.” Public Affairs Quarterly 7(2): 113–30. Lancet (1972). “Testing Anti-Cancer Drugs.” The Lancet 299(7755): 827–8. DOI: https://doi.org/10.1016/S0140-6736(72)90808-2. Le Gallois, M. (1813). Experiments on the Principles of Life, trans. N. C. Nancrede and J. G. Nancrede. Philadelphia, PA: Thomas, 19–21. Lehrer, Keith (1990). Theory of Knowledge. Boulder, CO: Westview Press. McKinlay, John B. and Sonya M. McKinlay (1977). “The Questionable Contribution of Medical Measures to the Decline of Mortality in the United States in the Twentieth Century.” Health and Society (Summer 1977): 405–428. Midgley, Mary (1986). “The Case for Restricting Research Using Animals,” in Tom Regan and Peter Singer (eds), Animal Rights and Human Obligations, 2nd edn. Englewood Cliffs, NJ: Prentice Hall, 216–22. Millstone, Erik (1989). “Methods and Practices of Animal Experimentation,” in Gill Langley (ed.), Animal Experimentation: The Consensus Changes. New York, NY: Chapman and Hall, 72–87. Moore, Stephen D. (1999). “Glaxo Removes Raxar Antibiotic from Markets Due to Side Effects.” The Wall Street Journal (October 29, 1999), www.wsj.com/articles/ SB941049254106057001. National Institute of Allergy and Infectious Diseases (2005). “Clinical Research on HIV Vaccines.” (May 2005). Nuffield Council on Bioethics (2005). “The Ethics of Research Involving Animals.” http:// nuffieldbioethics.org/wp-content/uploads/Animals-Chapter-1-Introduction.pdf. PCRM (2005). “Animal Research on Trial.” Good Medicine 14(Autumn): 13–14.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
96 Mylan Engel Jr. Pippen, John J. (2005). “The Need for Revision of Pre-Market Testing: The Failure of Animal Tests of COX-2 Inhibitors.” Presented at the FDA Open Public Hearing (February 17, 2005). Pippen, John J. (2007). “Drug Development and Approval in the United States: Stages, Descriptions, Timetables, and Attrition Rates,” in manuscript. Pippin, John J. and Kristie Stoick (2005). “Dangerous Medicine: Examples of AnimalBased ‘Safety’ Tests Gone Wrong.” www.pcrm.org/sites/default/files/pdfs/research/ testing/exp/dangerous_medicine.pdf. Pollock, John (1984). “Reliability and Justified Belief.” Canadian Journal of Philosophy 14: 103–14. Pollock, John (1986). Contemporary Theories of Knowledge. Totowa, NJ: Rowman and Littlefield. Preece, Rod and Lorna Chamberlain (1995). Animal Welfare and Human Values. Waterloo, Ontario, Canada: Wilfrid Laurier University Press. Regan, Tom (2012). “Empty Cages: Animal Rights and Vivisection,” in Jeremy Garrett (ed.), The Ethics of Animal Research: Exploring the Controversy. Cambridge, MA: MIT Press, 107–24. Ryder, Richard D. (1985). “Speciesism in the Laboratory,” in Peter Singer (ed.), In Defense of Animals. New York, NY: Basil Blackwell, 77–88. Seok, Junhee, H. Shaw Warren, Alex G. Cuenca et al. (2013). “Genomic Responses in Mouse Models Poorly Mimic Human Inflammatory Diseases.” Proceedings of the National Academy of Sciences [PNAS] 110(9): 3507–3512. DOI: https://doi. org/10.1073/pnas.1222878110. Sharpe, Robert (1989). “Animal Experimentation—A Failed Technology,” in Gill Langley (ed.), Animal Experimentation: The Consensus Changes. New York, NY: Chapman and Hall, 88–117. Singer, Peter (2009). Animal Liberation, updated edn. New York: HarperCollins. Stahlmann, Ralf and Rudolf Schwabe (1997). “Safety Profile of Grepafloxacin Compared with Other Fluoroquinolones.” Journal of Antimicrobial Chemotherapy 40(Suppl. A): 83–92. Steup, Matthias (2004). “Internalist Reliabilism.” Philosophical Issues, 14: Epistemology: 403–25. Stinnett, J. et al. (1981). “Improved Survival in Severely Burned Animals Using Intravenous Cornybacterium parvum Vaccine Post Injury.” Surgery 89(2): 237–42. PMID: 7455909. von Eschenbach, Andrew C. (2006). “FDA Teleconference: Steps to Advance the Earliest Phases of Clinical Research in the Development of Innovative Medical Treatments.” Presented January 12, 2006.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
PART 3
EPI ST E MIC A N D D OX AST IC WRONG S
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 19/03/21, SPi
5
A Tale of Two Doctrines Moral Encroachment and Doxastic Wronging Rima Basu* In this chapter, I argue that morality might bear on belief in at least two conceptually distinct ways. The first is that morality might bear on belief by bearing on questions of justification. The claim that it does is the doctrine of moral encroachment. The second is that morality might bear on belief given the central role belief plays in mediating, and thereby constituting, our relationships with one another. The claim that it does is the doctrine of doxastic wronging. Though conceptually distinct, the two doctrines overlap in important ways. This chapter provides clarification on the relationship between the two, providing reasons throughout that we should accept both. The chapter proceeds as follows. First, in section 1, I present and defend the doctrine of moral encroachment, demonstrating how it stems from commitments about the role of morality in justification. Second, in section 2, I present and defend the doctrine of doxastic wronging, demonstrating how it stems from commitments about the role of morality in forming our interpersonal relationships. In section 3, I end by discussing the relationship between the two, presenting views that occupy all positions in logical space: those that accept both moral encroachment and doxastic wronging, those that accept just one of these doctrines, and those that accept neither. Ultimately, I suggest, the combination is to be preferred.
1. Moral Encroachment Moral encroachment is the view that moral considerations bear on the justification of belief. Critics of moral encroachment claim that moral considerations do not bear in any way on justification. Rather, justification involves only purely epistemic * For extensive discussion, this chapter owes a huge debt to Gabbrielle Johnson. Many thanks also to Jennifer Lackey, Brian Kim, and Maegan Fairchild for their written feedback. Also thanks to Renee Bolinger, Amy Flowerree, Georgi Gardiner, Stephanie Leary, Liz Jackson, and Cat Saint-Croix for discussions about the chapter.
Rima Basu, A Tale of Two Doctrines: Moral Encroachment and Doxastic Wronging In: Applied Epistemology. Edited by: Jennifer Lackey, Oxford University Press (2021). © Rima Basu. DOI: 10.1093/oso/9780198833659.003.0005
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
100 Rima Basu (i.e. non- moral) considerations.1 To answer the critic, arguments for moral encroachment must establish that moral considerations are necessary for the justification of at least some beliefs. So, let’s see how that is to be done. Belief aims at truth. However, often we get things wrong. Nonetheless, there are ways our beliefs can be better or worse, even when the truth of the matter eludes us, by being justified or unjustified. A traditional view of justification is evidentialism. According to evidentialism, what you should epistemically believe is a function of the evidence.2 Although evidence cannot guarantee truth, it’s the kind of thing that is truth conducive (i.e. it raises the likelihood of truth). For example, whether a drug is an effective treatment is not a matter settled by your particular interests or practical matters—for instance, stock prices or a desire for a quick cure. What would settle the matter is purely epistemic considerations. According to evidentialists, evidence, for example, of the effectiveness of the drug, is one such purely epistemic consideration. As a result, we arrive at what has often been called purism regarding justification and belief: that what you should epi stemically believe is solely a function of the evidence.3 The purist decree to abide by only one’s evidence makes belief formation impossible.4 This is because the evidence alone underdetermines what one should believe. This has become widely known as the underdetermination problem. To see this, consider the following case: Given the appearance of some distinctive dark, winged shapes, moving across my visual field, what should I believe? That visual evidence, joined with other factors, may license me to believe propositions such as: (1) There are things moving through the air in front of me. (2) There are birds flying in front of me. (3) There are jackdaws flying in front of me. (4) At least three jackdaws exist. Which proposition I do believe will depend on, among other things: how my perceptual abilities have developed (e.g. have I learned to discriminate different 1 The two aims of this chapter make adopting terminology difficult. One aim is to elucidate how moral concepts and epistemic concepts come apart in theorizing about the demands of morality on belief. Seemingly at odds with this is the other aim of the chapter, which is to argue that the set of considerations picked out by moral and epistemic concepts might be one and the same. To keep ideas distinct, I do my best to flag where uses of concepts like ‘justification’ and ‘epistemic’ are to the exclusion of the moral or practical by insertion of “pure” or its cognates. 2 See Feldman and Conee (1985) for the canonical statement of this view. 3 This presentation of purism is simplified for the ease of exposition. Purism isn’t a single thesis. The simplification adopted here is standard in many discussions of moral encroachment (see, e.g. Bolinger 2020). In what follows, I’ll consider more sophisticated versions of purism in turn. 4 Compare this to Antony’s (2001, 2006, 2016) criticisms of Dragnet Objectivity. See also, Johnson (forthcoming) and Stanley (2016). Thanks to Gabbrielle Johnson for drawing my attention to these conceptual connections.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Moral Encroachment And Doxastic Wronging 101 kinds of bird on the wing?); the background information I happen to have (e.g. do I know what a jackdaw is?); and my particular interests at that moment (e.g. what do I want to know or do now?). (Nelson 2010, 87)
More is needed to determine whether you should believe (1), (2), (3), or (4). The evidence alone doesn’t settle that question. Similarly, as Sarah Paul and Jennifer Morton (2018a, 2018b) have argued, from the view of purely epistemic consider ations, there is not a uniquely best evidential policy to have. There are multiple evidential policies that are rationally permissible for a given thinker to have from the point of view of purely evidential considerations. Thus, one might think the more that is needed include moral and practical considerations. This is what Berislav Marušić and Stephen White (2018, 112) gesture at when they say, “if there is more than one epistemically legitimate route to belief, there is space for morality to do some work.” However, we can reject this naïve version of purism without adopting moral encroachment, i.e. the claim that moral considerations must bear on the justification of at least some beliefs. This is because we can avoid the underdetermination problem without giving up on the spirit of purism by bringing in additional non-evidential, but purely epistemic (e.g. truth-conducive) resources to determine what we may permissibly believe.5 This brings us to a more plausible version of purism, let’s call it purism*. According to purism*: what you should believe is a function of only purely epi stemic considerations that may very well extend beyond the merely evidential but are crucially still neither moral nor practical.6 For example, consider the relevant alternatives approach in epistemology.7 Such a view has played an important role in explaining the grip that skepticism can have on us. Our evidence alone cannot settle skeptical worries because our evidence is consistent with the possibility that we are being radically deceived by a simulated reality (notice that this is a version of the underdetermination worry). In answering the skeptic, the relevant alternative theorist’s interest is in spelling out the epistemic grounds on which we can determine what relevant alternatives we may permissibly ignore to ensure that skeptical hypotheses are among them. This view has recently been used by Georgi Gardiner (2018, forthcoming) to provide an alternative explanation of the cases that purport to establish moral
5 Thanks to Gabbrielle Johnson for raising this concern. 6 This progression from purism to purism* has important structural similarities to progressions in thought about the value-free ideal in philosophy of science. The value-free ideal is often mistakenly regarded as the claim that values have no place in scientific inquiry. In actuality, the ideal allows that some virtues enter into scientific inference, so long as those virtues are epistemic, i.e. truth conducive. As Douglas (2016, 611), suggests, we should really regard it as the “epistemic-virtues-only-in-scientific- inference ideal.” See Johnson (forthcoming) for discussion. And thanks to Gabbrielle Johnson for drawing my attention to these similarities. 7 See, e.g. Dretske (1970); Stine (1976); Lewis (1996) among others.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
102 Rima Basu encroachment. Insofar as relevant alternatives are taken to be determined by purely epistemic considerations, this view is a version of purism*. For reasons that follow, I think that even this even more sophisticated defense of the view fails to exclude moral considerations. The case common to almost all accounts of moral encroachment centers on the Cosmos Club, where John Hope Franklin is mistaken for a staff member.8 Of rele vance to the case is that John Hope Franklin was the club’s first black member and at the time, still only one of a handful of black members, whereas almost all of the club’s staff members were black. Notice that given the demographics of the club, that someone is black makes it very likely that they are a staff member. Returning to the naïve version of purism, such demographic evidence alone would be sufficient to justify forming the belief that John Hope Franklin is a staff member. Proponents of moral encroachment, on the other hand, have denied that we should draw this conclusion on the basis of demographic evidence alone.9 Rather, they claim that we must also take into consideration various moral and ethical features of this scenario. For example, we might take into consideration the moral implications of mistaking one of the few black club members for a staff member, a paradigmatically racist move. Gardiner claims that this move from evidence alone being insufficient to the necessity of moral and ethical considerations is too quick. Again, a denial of purism does not entail moral encroachment because there are non-evidential epistemic considerations that can be brought to bear. For Gardiner, those non-evidential epistemic considerations include considerations of the relevant alternatives. By drawing on the relevant alternatives framework outlined above, she defends a version of purism* against moral encroachment. Her claim is that the feature in the case that explains the epistemic failure of the woman who mistakes John Hope Franklin for a staff member is her failure to consider a relevant alternative, namely, that John Hope Franklin is a club member. Truth, after all, renders an alternative relevant, as does whether the error possibility is a common source of error. Relevant alternatives, thus, are truth conducive because they tend us towards a more accurate understanding without appeal to moral considerations, and thus they are purely epistemic. In order to defend moral encroachment against purism*, it is necessary to establish either that these epistemic considerations alone are (still) insufficient for settling what to believe, or that these epistemic considerations themselves are not
8 The original case is from Gendler (2011), but it features as the motivating case for various accounts of moral encroachment. 9 Moral encroachment theorists differ on precisely what is objectionable in this case, i.e. whether it’s the belief formation (Moss 2018a, 2018b) or whether it’s the belief itself (Basu 2019a, 2019c). See Bolinger 2020 for an overview of the varieties of moral encroachment. What’s important for the view as presented here is that morality bears on whether or not one’s belief that John Hope Franklin is a staff member is a justified belief. Thanks to Brian Kim for pushing to be clear about this point.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Moral Encroachment And Doxastic Wronging 103 purely epistemic. I intend to defend moral encroachment in each of these ways. To do so, it helps to turn to a parallel debate in philosophy of science concerning the value-free ideal.10 Just as we have been debating the role of ethical and moral considerations in settling the question of what to believe, philosophers of science have long debated the role of ethical and moral considerations in settling the question of which scientific hypothesis to accept. Philosophers of science similarly recognize that the problem of underdetermination necessitates the use of some extra-evidential considerations in settling these questions. However, their extra-evidential consid erations come in the form of scientific virtues, for example virtues such simplicity, consistency, breadth, etc.11 A version of purism* that emerges in this discussion claims that whatever virtues scientists adopt, they ought to be only virtues that are purely epistemic. This has become known as the value-free ideal. Feminist philosophers of science have introduced two argumentative strategies against the value-free ideal. These include arguments from demarcation and arguments from inductive risk.12 Borrowing and adapting for our purposes, I present versions of each to leverage moral encroachment over purism*. First, a version of demarcation. Consider again Gardiner’s relevant alternative framework and the claim that all that is needed to avoid epistemic error in cases like the Cosmos Club is to properly settle which relevant alternatives ought to be under consideration. A common way of determining whether an alternative is in fact relevant is whether it is a common source of error. The problem with this account, however, is that it neglects how common sources of error are themselves a reflection of moral features of our environment. For example, it is not, in the context of the Cosmos Club, a common source of error to mistake a black person as a staff member, because the case takes place in a time when racial discrimin ation at the club renders a black person more likely to be a staff member. The alternative, where a black person in the Cosmos Club is a club member, is remote enough to merit dismissing it as irrelevant. The defender of the relevant alternatives will likely respond that the possibility that John Hope Franklin is not a staff member achieves the status of being among the relevant alternatives simply by dint of being true.13 A problem with
10 Much of framing here is inspired by discussions with Gabbrielle Johnson, who pointed me to the similarities between the two traditions and echoes analoguous points in Johnson (forthcoming). 11 See Kuhn (1977); Douglas (2016); and Johnson (forthcoming). 12 Prominent instances of the first are presented in Longino (1995, 1996), whereas prominent instances of the second are presented in Douglas (2000, 2003). Again, see Douglas (2016) and Johnson (forthcoming) for further discussion. 13 But notice how little this response does to help us with our initial problem of skepticism that relevant alternatives were supposed to answer. If the fact that the error possibility obtained renders it relevant, then as the skeptic likes to warn us, we can’t rule out that the error possibility that we’re all being radically deceived obtains. Thus, by the skeptic’s light, it’s always a relevant alternative that you might be a brain in a vat.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
104 Rima Basu this approach is its inefficacy in an equally problematic case in which it is true that John Hope Franklin is a staff member. Surely, we would still want the alternative that he is a club member to be among the relevant alternatives under consid eration. This is due to our moral considerations of the case, namely that it would be racist to not consider that alternative. Thus, we see morality playing a role in determining what the relevant alternatives are, as well as being reflected in the common sources of error we’re meant to avoid.14 This demonstrates the demarcation argument because it demonstrates the impossibility of teasing apart these extra-evidential considerations from the moral values that they result from and reflect.15 Second, an argumentative strategy using a version of inductive risk. This argument against purism* again adopts the strategy of claiming that extra-evidential epistemic considerations alone are insufficient for settling the question of whether to believe p. The general spirit of this argument claims that we are limited agents for whom uncertainty is inevitable. The risk of being wrong is ineliminable, and given this risk, morality must enter into our epistemic deliberations regarding whether the risk is worth it. This argument has a historic precursor in Richard Rudner. In his seminal paper, “The Scientist qua Scientist Makes Value Judgements”, he notes that “our decision regarding the evidence and respecting how strong is ‘strong enough’, is going to be a function of the importance, in the typically ethical sense, of making a mistake in accepting or rejecting the hypothesis.”16 Rudner gives the example of requiring a relativity high degree of confirmation or confidence with regard to the safety of a drug containing a lethal ingredient versus the not-as-high level of confirmation or confidence required for whether a machine stamping belt buckles is defective. The relevant difference between the two cases is, of course, the grave moral consequences of getting wrong the drug dosage and the relatively light moral consequences of getting wrong the defectiveness of the machine. Inductive risk arguments have likewise been iterated in the literature on pragmatic encroachment.17 Proponents of pragmatic encroachment, like their moral counterparts, 14 See Johnson (forthcoming) for this interpretation of Longino’s (1995, 1996) demarcation argument, namely that epistemic considerations are unwitting value proxies. 15 There could be additional ways—beyond common sources of error and truth—to establish the relevant alternative that John Hope Franklin is a club member, e.g. Lewis’s rule of salience. What I suspect, though I don’t have room to argue for here, is that any such rule that attempts to make this case while excluding the moral considerations will fail to make good on what I take to be an independent claim about the case: that the best explanation for why John Hope Franklin is a club member is a relevant alternative, even in cases where he’s not, is that it’d be racist to ignore this alternative. Thanks to Brian Kim for raising this objection. 16 Rudner (1953, 2), emphasis in original. 17 See, e.g. Fantl and McGrath 2002, 2009; Stanley 2005; Schroeder 2012; Ross and Schroeder 2014, among others. To the best of my knowledge, Kukla (2015) is the first to point out the similarities between discussions of inductive risk in philosophy of science and pragmatic encroachment. See also Stanley (2016). Thanks again to Gabbrielle Johnson for drawing my attention to the connections between these traditions.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Moral Encroachment And Doxastic Wronging 105 aim to establish that epistemic considerations alone are insufficient to establish knowledge, justification, epistemic rationality, etc. Pragmatic encroachers argue, unsurprisingly, that what is missing is the need to attend to various practical considerations before such epistemic matters can be settled. Moral encroachment extends the argument to include moral considerations.18 Notice, however, that Rudner’s original case is one lobbying for moral considerations, not mere practical ones.19 So, the original argument from inductive risk is an argument for moral encroachment. In addition to this sort of argument, there are two further moral risks to consider. The aforementioned inductive risk argument alone is enough to undermine purism* and instead establish moral encroachment as a general thesis. I end this section by noting the diversity of moral considerations that might enter into the evaluation of inductive risk and thereby lead to different varieties of moral encroachment.20 An area of exploration that I find particularly intriguing are possible forms of inductive risk arguments in the realm of belief that have been obfuscated by the discipline’s focus on the scientific context and its pragmatic analogues. Particularly, heretofore theories of inductive risk have focused naturally on the risk of drawing false conclusions, of getting things wrong. An unexplored kind of inductive risk that I conjecture we ought also to consider is that which comes from getting things right. I’m not confident that such cases where risk stems from accuracy are possible, but here I’d like to briefly explore the possibility by discussing violations of priv acy that occur when we infelicitously form true beliefs about others.21 What is the right to privacy? As Judith Jarvis Thomson (1975, 295) notes, “[p]erhaps the most striking thing about the right to privacy is that nobody seems to have any very clear idea what it is.” As she goes on to note, despite the fact that nobody seems to have a clear idea of what the right to privacy is, we nonetheless have strong intuitions about cases in which we’d say a right to privacy has been violated. Consider a standard case of privacy violation: someone breaking into your house and not only stealing your TV, but sticking around to rummage through your personal belongings, read through your love letters, scroll through
18 It is not obvious why this construal of moral encroachment is at odds with sophisticated invariantist frameworks that require a higher grade of knowledge for actions with high stakes. See, e.g. Reed (2010). So long as high-stakes scenarios can be determined in part by moral considerations, the invariant framework seems congenial to the approach advocated for here. Unfortunately, a comprehensive comparison of the two approaches is beyond the scope of this discussion, but thanks to Jennifer Lackey for pressing me to think about the relationship between the two. 19 I’m forced to omit discussion of a variety of purist responses to this line of thought. For a prom inent response to Rudner, see Levi (1960). 20 For instance, I and others have elsewhere made the argument that perpetuating injustice is one such consideration. See Bolinger (2020) for a survey of the varieties of moral encroachment. 21 See also Hunter (2018).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
106 Rima Basu your search history, etc. As Thomson (1975, 299) notes, “the burglar’s merely looking around in that way might make the episode feel worse than it otherwise would have done.” There is something about another person attending to these private things about us that makes the rights violation indisputable. That the burglar’s coming to know these facts about you constitutes a violation of privacy makes clear that we run at least some risk in coming to believe facts about others. Just as we keep our love letters hidden away, we keep parts of ourselves hidden away from others.22 In other words, we would prefer that others do not form true beliefs with regard to these private affairs. In fact, I believe that we run this risk merely in forming true beliefs about others, irrespective of whether we ourselves took the prying steps necessary for the unsanctioned release of that information. To see this, consider the case of doxing.23 Doxing, as David Douglas (2016, 199) notes, “is the intentional public release onto the Internet of personal information about an individual by a third party, often with the intent to humiliate, threaten, intimidate, or punish the identified individual.” Doxing is often done for a variety of motives, ranging from the desire to expose wrongdoing to cyber stalking and harassment. Key to doxing is that it makes private information about the person public. As Douglas notes, when a victim is doxed, merely entering the victim’s name into a search engine may end up revealing the victim’s private details. Part of the harm of doxing is the kind of epistemic terror that is inflicted on the victim of the doxing, thus risk of the privacy violation is irredu cibly intertwined with moral consequence. This consequence follows regardless of whether the searcher was the doxer. It does not obviously follow, however, if the information gleaned is false. Thus, weirdly, it seems the non-doxer runs an inductive risk in getting things right, not wrong. This, I maintain, is a risk that one needs to take into consideration when deliberating about whether to believe p.24 To recap, we’ve responded to the main obstacle to moral encroachment by demonstrating challenges that neither purism nor purism* can overcome unless they grant that moral considerations must bear on the justification of at least some beliefs. First, we saw the impossibility of teasing apart extra-evidential 22 Similarly, on Marmor’s (2015) explanation of the right to privacy, he notes that we have an interest in shaping and/or controlling how we appear to others. Thanks to Renée Bolinger for discussion on this point. Furthermore, there is a connection here to Dembroff and Saint-Croix’s (2019) work on agential identity and the interests we have that other people respect our identity, which I will return to in the discussion of holding in section 2. 23 Thanks to Amy Flowerree for suggesting this case. Examples such as these begin to broach cases of doxastic wronging. For instance, in discussion Cat Saint-Croix suggested that the right to be forgotten would be an example of a right to privacy that concerns what other people think of you. Another example might be the right to be known, as discussed by Jennifer Lackey (in progress). 24 As Brian Kim has pointed out to me, there are many ways we can mess up here and not all of them have to do with belief formation. As I’ve argued, you can go wrong by forming the true belief, but you can also go wrong by making information available, thereby making it possible for others to form justified true beliefs. One can also go wrong by making certain inquiries or questions salient to oneself and others. In Basu (in progress), I suggest that this last wrong is one that philosophers are particularly susceptible to when theorizing about others.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Moral Encroachment And Doxastic Wronging 107 considerations from the moral values that they result from and reflect. Second, we discussed a variety of types of moral consideration that get into evaluations of inductive risk: that forming some belief might eventuate in bodily harm to others, that we might perpetuate patterns of injustice, that we might violate a person’s privacy, etc. On the face of it, I see no reason to regard these as rivals. I don’t believe that there’s only one kind of moral consideration that should go into the evaluation of inductive risk.25 Ultimately, the question of what moral consider ations matter will be settled by first-order moral theories. For example, if one held a first-order moral theory that hypothesized that beliefs themselves could be wrong in various ways, this would be one sort of consideration that likely gets in. This is precisely the theory of doxastic wronging, which I turn to next.
2. Doxastic Wronging In section 1, we explored the extent to which morality bears on belief in virtue of belief ’s constitutive aim of truth. Aiming at truth led us to justification, and in justification, we saw the need for morality. In this section of the chapter, I want to set aside issues of justification to highlight another conceptually distinct avenue by which morality bears on belief. Although it’s right that beliefs aim at truth, they do so only in virtue of having intentional content that represents the world as being a certain way. This content provides a perspectival mode of presentation that mediates our relations to the external environment. This mediation is central to Frank Ramsey’s (1929 (1990), 146) idea that beliefs subserve the function of navigation: they are the “map by which we steer.” Not only do beliefs about the arrangement of space around us allow us to navigate the world more generally, so too do beliefs about people help us to navigate our social world. It is in virtue of belief ’s committing us to this content—content that represents, in the case of beliefs about another person, perspectival claims about that individual’s status in the world—that I conjecture solidifies belief ’s moral standing. By mediating our interpersonal relations to others, our beliefs about others bear moral weight. Doxastic wronging is the thesis that beliefs, in virtue of this standing, can sometimes themselves be the source of moral wrongdoing. Consider Grace, from the hit Netflix show, Grace and Frankie. During an interview, Grace reveals some beliefs she holds about her daughters, Brianna and 25 Some discussions of moral encroachment countenance only one kind of moral risk. Sarah Moss’s (2018a, 2018b) account of moral encroachment is motivated by the risks of beliefs, and she doesn’t think that moral encroachment can be properly motivated from the costs of the belief itself. I believe this erroneously gives the impression that moral encroachment comes in just one flavor. I myself have intimated as much by discussing doxastic wronging to the exclusion of other morally relevant consid erations in Basu (2019a). I regret not being more careful. For careful discussion, see Bolinger (2020).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
108 Rima Basu Mallory. She believes Brianna has run the family company into the ground; she also neither believes that Mallory is the smart daughter nor that she has made good use of her degree (that Grace paid for). Later, as her daughters pack up their desks, Grace is confused by why her daughters are upset with her but is willing to apologize for having said what she did. As Mallory naturally points out, “It’s not that you said all those terrible things, it’s that you actually believed them.” Here Mallory makes clear an overwhelmingly intuitive point: the source of the wronging is not in what was said, but in the belief itself. This exchange demonstrates three hallmarks of doxastic wronging: (1) doxastic wrongs are directed; (2) doxastic wrongs are committed by beliefs rather than the consequences of acting on a belief; and (3) doxastic wrongs are wrongs in virtue of the content of what is believed.26 Grace’s belief demonstrates all three of these features. First, her belief is directed: that is, she doesn’t merely do wrong, she does wrong to her daughters. Second, the wrong she commits is one of belief, not of word or deed. Third, and perhaps most fundamentally, the wrong she commits is a wrong in virtue of the content of what she believes. Elaborating on these in reverse order, I again take it that a belief ’s having the representational content that it does is essential to its moral standing (hallmark 3). The content represents her daughters as standing in a particular relation to the properties she attributes to them. For example, her belief about Brianna relates Brianna to attributes like being a bad CEO. Naturally, her daughters have a legitimate complaint about the picture that content paints of them. There are many features that factor into whether complaints on beliefs are legitimate that I will go on to elaborate, but for now, I take Brianna and Mallory’s case to be an uncontroversial one. However, beliefs are constituted not only by representational content, but also in placing the bearers of those beliefs in a particular relation to that content, namely one of committing to its being true (hallmark 2). Propositional attitudes involve both propositions (representational contents) and attitudes (relations to the contents). Beliefs are committal mental states; they involve committing the subject to the truth of the representational content. This is why beliefs can wrong: the representational content, together with that commitment, constitute a wrong in the belief itself, and not in any particular consequences that come from the belief. For example, if Grace merely said the relevant propositions, but wasn’t committed to the content’s being true (perhaps she thought lying would save the company), presumably her daughters would rightly feel differently. They might be upset that she lied, and suspicious of why she chose those particular lies, but these would be different complaints. Moreover, if she held a different attitude towards
26 For an introduction to these hallmarks, see Basu and Schroeder (2019, 181). There, Schroeder and I enumerate the hallmarks. However, the analysis to follow is novel and illustrates the extent to which my views on the subject have evolved. Thanks to Gabbrielle Johnson for many insights central to this analysis.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Moral Encroachment And Doxastic Wronging 109 the relevant proposition, for example, if she doubted or denied or feared the way the content represented her daughters, it seems that there would be no complaint at all.27 Thus, it’s in the belief itself—that her mother was committed to the particular content that she was—that her daughters were wronged.28 And finally, it is with respect to belief ’s being directed at others, and thereby, how beliefs relate us to others, that we are at greatest risk of wronging (hallmark 1). Let me turn to that now. As Rae Langton (1992, 486) notes, We don’t simply observe people as we might observe planets, we don’t simply treat them as things to be sought out when they can be of use to us, and avoid when they are a nuisance. We are, as Strawson says, involved.
In short, when it comes to people, there’s a different way of going about things. Further, this different way of going about things concerns not only how we treat them through our actions or our words, but also how we consider them in thought. As Marušić and White (2018, 100) have similarly argued, when it comes to persons, the core Kantian idea that underpins the idea of the categorical imperative is the following: “that our way of relating to people is categorically different from our way of relating to objects.” Persons, as ends-in-themselves, are not to be regarded or related to in the same way one relates to objects. Once you accept this general intuition, you might begin to wonder how we can capture this different way in which we ought to relate to others. As I’ve argued previously (see Basu 2019b), this being involved that Langton attributes to Strawson (1962) is the recognition that others’ attitudes and intentions towards us are important in a way that’s distinctive to the kind of things that we are, and that our treatment of others should reflect that importance. We are, each of us, in virtue of being social beings, vulnerable, and we depend upon others for our self-esteem and self-respect. Respect and esteem, however, are not mere matters of how we’re treated in word or deed, but also a matter of how we’re treated in thought. The implication of this (quite minimal) Kantian and Strawsonian picture is that people should figure in both our theoretical and practical reasoning in a way that is different from objects.29 We care how we feature in the thoughts of
27 The relationship between attitude, content, and doxastic wronging is complex. I discuss these points briefly at the end of the chapter, but ultimately leave the project of working out the complexities for future work (see n. 32). 28 I have elsewhere provided extensive arguments against the view that the wrong lies in the risk of acting on the belief (see Basu 2019a, 2019c). 29 I say “quite minimal” here because you don’t need to be a fully fledged Kantian to accept this picture; all you must accept is that people are different from mere objects and whatever it is that makes them different requires treating people differently from mere things. Even a consequentialist would accept this characterization of persons and our obligations to them, but perhaps they would extend it wider to anything capable of happiness or pain and say that those things should be treated differently than things that do not experience happiness or pain.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
110 Rima Basu other people and we want to be regarded in their thoughts in the right way; that is, doxastic wrongs are failures to regard people in the right way. Another way of capturing belief ’s role in mediating our relations to others that I quite like is found in Hilde Lindemann’s (2016) discussion of holding. Lindemann argues that given the kind of things that we are, there is a general moral obligation of holding others in personhood that extends to the attitudes we hold of one another. I take it that we can understand some doxastic wrongs as stemming from holding failures. The centerpiece of Lindemann’s theory of holding is that our identities are pieces of narrative construction that are constituted in part by others’ beliefs about us (as well as our beliefs about ourselves), and thus, we fundamentally depend on others. Holding, when done well, “supports an individual in the creation and maintenance of a personal identity that allows her to flourish personally and in her interactions with others.”30 Holding, when done badly, can be destructive. Doxastic wronging, then, captures the harmful or defective narratives that result in our beliefs concerning others. Thus, given the kind of thing that we are, there is a moral obligation to hold others well. The point I wish to emphasize here is that we have both a moral and a doxastic responsibility of holding one another. It matters how we hold others in our thought. The beliefs we have, after all, are constitutive of our relationships. This is made obvious by the role our beliefs play in contributing to the narratives of others. And as I’ve begun to argue in some new work, this is especially so in cases of parent–child relationships, like that of Grace and her daughters.31 When Grace thinks that Mallory isn’t the smart one or that Mallory has never done anything useful with her degree, Grace is neither creating nor maintaining a personal identity that would allow Mallory to flourish. We are especially vulnerable to our parents in this way. And, as I noted earlier, we are, each of us, in virtue of being social beings, vulnerable in our dependence on others for our self-esteem and self-respect. In a manner similar to the Kantian and Strawsonian picture I have been outlining, Lindemann suggests that “[w]e can think of requirements and prohibitions of this sort as falling under the general heading of an impersonally authoritative obligation to treat persons in a manner consonant with their value.”32 Consider Lindemann’s example of W. Elliott’s experience as a medical student. Elliot receives advice from a doctor who refers to one of his patients as “a plant”, and 30 Lindemann (2016, x). 31 As I discuss in Basu (in progress), the first hallmark of doxastic wronging provides resources for a view on which different relationships to others might result in different moral demands of the states themselves. There is something about relationships, especially our close relationships, that makes them a particularly rich space for all kinds of attitudinal wrongs more generally. However, it is unclear whether it is the closeness of the relationship that affects the degree to which we’re wronged, or whether it is the relationships themselves that change the kinds of wrongs that can be inflicted. I cannot say anything to resolve this thorny issue at this time, but I leave it to explore in future work. 32 Lindemann (2016, 23–4).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Moral Encroachment And Doxastic Wronging 111 who claims Elliot’s job is to make sure “the plant” is watered. As Lindemann notes, this is no way to treat a person. The patient is likely incapable of recognizing that the doctor is referring to her in that dehumanizing way. But, crucially, what wrongs her is to be thought of in this way, “because it pushes her outside the human community.”33 In sum, if our identities are narratively constituted through our interactions (belief, word, actions, all of it) with others, it follows that there is a moral obligation to hold others well.34 To reiterate, I take it that we can understand some doxastic wrongs as stemming from holding failures. Holding failures bear the three hallmarks of doxastic wrongs. First, holding failures are directed. Second, the failure is committed by the belief itself, not the consequences of the belief in action. Third, the failure is wrong in virtue of what is believed; that is, the patient has a legitimate complaint about the picture that content—the content that represents her as non- human, as a plant—paints of her. Ultimately, what these examples demonstrate is that we owe people more care in thought.
3. The Relationship between Moral Encroachment and Doxastic Wronging At this point, I’ve explained two conceptually distinct ways in which morality might bear on belief. The first is that morality might bear on belief by bearing on questions of justification. The claim that it in fact does is moral encroachment. The second is that morality might bear on belief given the central role belief plays in mediating, and thereby constituting, our relationships with one another. The claim that it in fact does is doxastic wronging. Though conceptually distinct, the two overlap in important ways. I take it that it is within this overlap that confusions about the relationship between the two doctrines have emerged. Overlap between the two doctrines occurs because moral encroachment opens the door for first-order moral theories to play a role in justification. Because doxastic wronging is a first-order moral theory about the nature of wrongdoing via belief, it naturally is a candidate for the kind of consideration that bears on justification. In other words, if you hold moral encroachment and you hold doxastic wronging, then you believe that doxastic wrongs bear on the evaluation of justification for belief.35 Alternative to the combination of the two views, one could 33 ibid. 34 There are likely limitations on holding; for instance, there are likely some narrative identities that we shouldn’t engage in this kind of holding with respect to. For more on this point, see Dembroff and Saint-Croix (2019), in particular 589–92. One might wonder to what extent it is necessary to take on the heavyweight metaphysical claims about identity here. Although I’m sympathetic to the identity- constituting narrative picture, I expect it is enough merely that these narratives exist and we are invested in them. Thanks to Maegan Fairchild for raising this question. 35 This is what I argue in Basu (2019a).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
112 Rima Basu hold either moral encroachment or doxastic wronging, but not both. For example, you might think moral encroachment is right for reasons provided by Rudner, but since doxastic wronging is not among your first-order moral theories, the sorts of moral considerations that enter into evaluations of justification will be limited to traditional first-order moral considerations, for example, allegiances to consequentialism (as Rudner himself adopts). Likewise, you might think doxastic wronging is right for reasons provided in section 2, but deny moral encroachment, and thereby think that questions of justification will be limited to trad itional epistemic considerations, for example, traditional allegiances to purism (or purism*). And, of course, you might reject both theses, in which case you will think questions of both justification and interpersonal relations can be settled without reference to how morality bears on belief at all. To get clear on these distinctions, it will help to discuss an oft-cited criticism of the combination of the two views. This is the redundancy objection. This objection has been put in several different ways.36 Here’s one way to put the objection: Marušić and White (2018, 99) argue that these considerations regarding adjustments of one’s evidential threshold as suggested by moral encroachment don’t seem sufficient to fully address what is morally significant about the case motivating moral encroachment. Thus, we’re left with the following question: “[w]hat normative work is morality left to do if it is conceived of as merely a derivative of the epistemic permissions and prohibitions?” By keeping doxastic wronging and moral encroachment separate, we can make clear that when you meet the higher evidential threshold in morally risky cases you may be epistemically, but not morally, off the hook. For example, if we consider Rudner’s scientist again, the moral risks of killing lots of people set the bar for confirmation or confidence extremely high. And we can imagine the case playing out in one of two ways. First, the drug doesn’t kill anyone and obviously no moral wrong is done. Alternatively, the drug kills a lot of people. What do we want to say in this case? It seems odd to say that no moral wrong is done; after all, people have died. Perhaps what we want to say is that the scientists did all that they could epistemically, thus they are epistemically off the hook, but not morally. Note that the same applies in cases where we combine moral encroachment and doxastic wronging. The complicating feature of these cases, however, is that it is in the formation of the belief itself that we run a risk of wronging. No matter how complicating this feature is, it seems to me that doesn’t change what we want to say. Moral encroachment demands only that we do our best epistemically by adjusting our evidential thresholds according to moral considerations. Moral 36 See, e.g. Begby (2018) and Gardiner (2018), who notice this point when they argue that if a belief is genuinely wrong, then the moral questions shouldn’t disappear after the question of justification is settled. This is right. Once we make clear that doxastic wronging and moral encroachment are conceptually distinct theses, it’s obvious that doxastic wrongs don’t reduce to failures to take into account moral considerations when evaluating justification.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Moral Encroachment And Doxastic Wronging 113 encroachment does not exonerate any and all wrongdoing that comes to fruition when we believe, even if we’ve properly adjusted our evidential thresholds. Once we’re clear about the difference between the two kinds of moral constraints on belief, and where they intervene on our doxastic lives, then the redundancy worry dissolves.37 Failing to keep these distinct can result in the kind of problem that Osborne (forthcoming, 5) notes, when he suggests that some versions of moral encroachment result in “our ultimate doxastic obligation [being] to our evidence.” It’s true that our ultimate epistemic- doxastic obligation is to our evidence, properly adjusted in light of moral considerations, but our ultimate moral-doxastic obligations exist independently. By keeping these two doctrines separate, we can also demonstrate how there are prominent views that already accept one doctrine, while rejecting the other. Doxastic partiality views seem to me to push us to accepting doxastic wronging while rejecting moral encroachment. For example, Sarah Stroud (2006) and Simon Keller (2004) argue that we have special responsibilities to our friends that require a kind of doxastic partiality; that is, we should not be neutral with regard to how we respond to evidence about our friends. But note that this non-neutrality doesn’t entail moral encroachment. For Stroud and Keller, these aren’t questions of justification, since they occur after justification has been settled (in the trad itional, purist sort of way). Similarly, Berislav Marušić (2015) argues that when it comes to beliefs about what we will do in the future, the stance we take to such beliefs is importantly different from the stance we take towards other beliefs. When it comes to what you will do in the future, you should believe against the evidence because whether you succeed is up to you. In this way, his view is similar to Stroud and Keller. Further, those who are partial to us, such as our friends, lovers, our spouse, etc. would be wronging us if they didn’t also believe against the evidence. In virtue of being close to you, they should exhibit a similar kind of doxastic partiality. In fact, Marušić takes it a step further by noting cases where it would be wrong to even be attending to these questions of evidence, evidential weight, justification, etc. To illustrate one such case, imagine that you are standing at the altar and are about to make your wedding vows. These vows include the promise to spend the rest of your life with your spouse-to-be. However, let’s suppose that the best evidence there is suggests that 50 per cent of marriages end in divorce.38 That puts your odds of spending the rest of your life with your spouse-to-be at chance. You’ve no reason to think you two are any different from any other couple that has stood where you now stand. You have no previous track record of marriage from which to draw on either. You are in every way just plain ordinary. So, now 37 Thanks to Maegan Fairchild for helping me get clear on this point. 38 This statistic is outdated, but let’s assume it for the purpose of this example.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
114 Rima Basu what should you do and what should you believe? Can you make a sincere promise to your spouse-to-be that you’ll spend the rest of your life with them? As Marušić notes, in such a case it’d be somewhat perverse to say, “I’ll be with you the rest of my life, but there is a significant chance I won’t,” even though saying that does reflect the weight of your evidence. Here, what morality demands is not only that you believe, but also that you disregard the evidence entirely. To say what you need to say in this case—that you will spend the rest of your life with them and only them—requires believing against the evidence. Because doxastic wronging can’t be reduced to moral encroachment, nor vice versa, one risk of accepting doxastic wronging (whether or not you accept moral encroachment) is that we are potentially led to cases of conflict between the moral and the epistemic. This is natural because, as I’ve outlined, there are two sources of demands on belief. One is justification, the other (in accepting doxastic wronging) is morality. Crucially, because these are distinct, the demands can pull in different directions. How worried about this should we be? According to Jennifer Saul, at least somewhat. As she notes, accounts that pit morality against epistemic rationality fit “exceptionally well with the right-wing narratives of politically correct thought-police attempting to prevent people from facing up to difficult truths; and of the over-emotional left, which really needs to be corrected by the sound common sense of the right. Anything that props up these narratives runs the risk of working against the cause of social justice.”39 That is, it seems a bad consequence of a view if, say, opposition to racism (what is morally required) leads one into epistemic irrationality.40 I think that the worry can be mitigated by adopting moral encroachment. What Saul’s quote demands is coordination between the moral and the epistemic, which is what moral encroachment provides. After all, having adequate justification is justification properly responsive to moral stakes. That being so epistemically responsive could nonetheless result in a moral wrong occurring is a possibility. But I don’t find this worrisome. We constantly have to choose between many conflicting demands. As Keller (2018, 34) notes, “the demands of human life are varied and conflicting, and the standards that apply to belief formation are varied and conflicting too.”41 In closing, I want to make some exploratory remarks about other ways the two doctrines can be pulled apart. That we can have cases of moral encroachment that don’t include doxastic wronging at all is evident. That we can have cases of doxastic wronging that don’t include moral encroachment is less evident.
39 Saul (2018, 238–9). For further defense of the no-conflict view that I’m now walking back, see Basu (2020). 40 This is the central puzzle in Basu (2019c). 41 See also Burge’s (2003, 509–10) discussion of the independence of various functional norms on belief formation, e.g. to aim at truth, to guide behavior, and to contribute to survival.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Moral Encroachment And Doxastic Wronging 115 Is there any token belief that triggers the demands of doxastic wronging, but not moral encroachment? It seems to me that if we accept both doctrines, then any token belief that runs the risk of doxastic wronging (notably, a belief about others) will trigger the demands of moral encroachment. I take it that doxastic wronging is a broader thesis than moral encroachment because there are doxastic states that aren’t within the purview of theories of justification. Thus, it stands to reason that we can pull the two apart in a particular case by considering these sorts of doxastic states. Recall the three hallmarks of doxastic wronging: (1) doxastic wrongs are directed; (2) doxastic wrongs are committed by beliefs, rather than the consequences of acting on a belief; and (3) doxastic wrongs are wrongs in virtue of the content of what is believed. Notice that all three of these hallmarks leave open that other mental states might fall under doxastic wronging. Belief is a committal mental state, but so are other mental states, for example, perception and, loosely, desire. This has prompted me to begin to explore whether the concept of doxastic wrongs extends to these other kinds of states. Initially, it seems to me plausible that it does. Focusing first on desire, a mother’s desire that her gay daughter marry a man seems to me to exhibit the same features that resulted in wronging in the belief case. Her desire has a particular problematic representational content (hallmark 3) and puts the mother in a committal relation to that content, namely to its becoming true (hallmark 2). It also seems to meet hallmark 1 by being directed at her daughter, and thereby contributing to a destructive narrative. Further, we again see reason to keep doxastic wronging and moral encroachment distinct. Moral encroachment wouldn’t be relevant in these further cases of doxastic wronging because other mental states such as perceptions and desires don’t function to be justified by the evidence at all. That is, morality can’t bear on the justification of these states because justification itself arguably doesn’t apply to these states. In summary, moral encroachment and doxastic wronging are not only conceptually distinct views, but also views that plausibly apply independently to token psychological state formation.42 Hopefully, what I’ve done is clarified the relationship between two doctrines in order to demonstrate the two distinct ways that morality might come to bear on belief. The first, concerning justification, leads us to moral encroachment. The second, concerning belief ’s role in our interpersonal relationships, leads us to doxastic wronging. With this, I maintain what I’ve said in previous work: when it comes to what we should believe, morality’s got bite. What I’ve failed to elucidate up to this point is the extent to which that bite is multifaceted and nuanced. To provide a complete catalogue of the intersections of morality and belief would 42 That is, at least not in the ordinary way assumed in discussions of justification. Though see Siegel (2015) and Jenkin (2020) for views that explore the applicability of epistemic justification to these other mental states.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
116 Rima Basu extend well beyond the current corpus. But what I hope to have made the case for here is just how thoroughly morality permeates our lives, including aspects of which we initially believed to be beyond its purview. It might be daunting to consider just how many demands morality makes on us, but a brief glimpse at the world should suggest that maybe we’re undercounting the wrongs of which we’re capable.
References Antony, Louise (2001). “Quine as Feminist: The Radical Import of Naturalized Epistemology,” in Antony, L. and Witt, Charlotte. E. (eds), A Mind of One’s Own: Feminist Essays on Reason and Objectivity. Nashville, TN: Westview Press, 110–53. Antony, L. (2006). “The Socialization of Epistemology,” in R. E. Goodin, and C. Tilly (eds), The Oxford Handbook of Contextual Political Analysis. Oxford: Oxford University Press, 58–77. Antony, L. (2016). “Bias: Friend or Foe?,” in M. Brownstein and J. Saul (eds), Implicit Bias and Philosophy, Volume 1: Metaphysics and Epistemology. Oxford: Oxford University Press, 157–90. Basu, Rima (2019a). “Radical Moral Encroachment: The Moral Stakes of Racist Beliefs.” Philosophical Issues 29: 9–23. Basu, Rima (2019b). “What We Epistemically Owe to Each Other.” Philosophical Studies 176(4): 915–31. Basu, Rima (2019c). “The Wrongs of Racist Beliefs.” Philosophical Studies 176(9): 2497–515. Basu, Rima (2020). “The Spectre of Normative Conflict: Does Fairness Require Inaccuracy?,” in Erin Beeghly and Alex Madva (eds), An Introduction to Implicit Bias: Knowledge and the Social Mind. Abingdon: Routledge, 191–210. Basu, Rima (in progress). “The Ethics of Doing Philosophy.” Basu, Rima (in progress). “The Parent Trap: How Doxastic Wronging Start at Home.” Basu, Rima and Schroeder, M. (2019). “Doxastic Wronging,” in Brian Kim and Matthew McGrath (eds), Pragmatic Encroachment in Epistemology. Abingdon: Routledge, 181–205. Begby, E. (2018). “Doxastic Morality: A Moderately Skeptical Perspective.” Philosophical Topics 46(1): 155–72. Bolinger, R. J. (2020). “The Rational Impermissibility of Accepting (Some) Racial Generalizations.” Synthese 197: 2415–31. Bolinger, R. J. (2020). “Varieties of Moral Encroachment.” Philosophical Perspectives 34(1): 5–26. Burge, T. (2003). “Perceptual Entitlement.” Philosophy and Phenomenological Research 67(3): 503–48, http://onlinelibrary.wiley.com/doi/10.1111/j.1933-1592.2003. tb00307.x/full.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Moral Encroachment And Doxastic Wronging 117 Dembroff, R. and Saint-Croix, C. (2019). “ ‘Yep, I’m Gay’: Understanding Agential Identity.” Ergo: An Open Access Journal of Philosophy 6: 571–99. Douglas, D. M. (2016). “Doxing: A Conceptual Analysis.” Ethics and Information Technology 18: 199–210. Douglas, H. (2000). “Inductive Risk and Values in Science.” Philosophy of Science 67(4): 559–79. Douglas, H. (2003). “The Moral Responsibilities of Scientists (Tensions between Autonomy and Responsibility).” American Philosophical Quarterly 40(1): 59–68. Douglas, H. (2016). “Values in Science,” in P. Humphreys (ed.), Oxford Handbook in the Philosophy of Science. pages. Oxford: Oxford University Press, 609–30. Dretske, F. (1970). “Epistemic Operators.” Journal of Philosophy 67(24): 1007–23. Fantl, J. and McGrath, M. (2002). “Evidence, Pragmatics, and Justification.” Philosophical Review 111(1): 67–94. Fantl, J. and McGrath, M. (2009). Knowledge in an Uncertain World. Oxford: Oxford University Press. Feldman, R. and Conee, E. (1985). “Evidentialism.” Philosophical Studies 48: 15–34. Gardiner, G. (2018). “Evidentialism and Moral Encroachment,” in K. McCain (ed.), Believing in Accordance with the Evidence: New Essays on Evidentialism. New York: Springer, 169–95. Gardiner, G. (forthcoming). “Relevance and Risk: How the Relevant Alternatives Framework Models the Epistemology of Risk.” Synthese. Gendler, T. (2011). “On the Epistemic Costs of Implicit Bias.” Philosophical Studies 156(1): 33–63. Hunter, D. (2018). “Directives for Knowledge and Belief,” in Conor McHugh, Jonathan Way, and Daniel Whiting (eds), Normativity: Epistemic and Practical. Oxford: Oxford University Press, 68–89. Jenkin, Z. (2020). “The Epistemic Role of Core Cognition.” Philosophical Review 129(2): 251–98. Johnson, G. (forthcoming). “Are Algorithms Value-Free? Feminist Theoretical Virtues in Machine Learning.” Journal of Moral Philosophy. Keller, S. (2004). “Friendship and Belief.” Philosophical Papers 33(3): 329–51. Keller, S. (2018). “Belief for Someone Else’s Sake.” Philosophical Topics 46(1): 19–35. Kuhn, T. (1977). “Objectivity, Value Judgement, and Theory Choice,” in The Essential Tension. Chicago, IL: University of Chicago Press, ch. 13, 330–39. Kukla, R. (2015). “Delimiting the Proper Scope of Epistemology.” Philosophical Perspectives 29: 202–16. Lackey, J. (in progress). “Epistemic Reparations and the Right to Be Known.” Langton, R. (1992). “Duty and Desolation.” Philosophy 67(262): 481–505. Levi, I. (1960). “Must the Scientist Make Value Judgments?” The Journal of Philosophy 57(11): 345–57.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
118 Rima Basu Lewis, D. (1996). “Elusive Knowledge.” Australasian Journal of Philosophy 74: 549–67. Lindemann, H. (2016). Holding and Letting Go: The Social Practice of Personal Identities. Oxford: Oxford University Press. Longino, H. E. (1995). “Gender, Politics, and the Theoretical Virtues.” Synthese 104(3): 383–97. Longino, H. E. (1996). “Cognitive and Noncognitive Values in Science: Rethinking the Dichotomy,” in L. Hankinson Nelson and J. Nelson (eds), Feminism, Science, and the Philosophy of Science.. Dordrecht: Kluwer, 39–58. Marmor, A. (2015). “What is the Right to Privacy?” USC Law Legal Studies Paper No. 14–13. Marušić, B. (2015). Evidence and Agency. Oxford: Oxford University Press. Marušić, B. and White, S. (2018). “How Can Beliefs Be Wrong?—A Strawsonian Epistemology.” Philosophical Topics 46(1): 97–114. Moss, S. (2018a). “Moral Encroachment.” Proceedings of the Aristotelian Society 118(2): 177–205. Moss, Sarah (2018b). Probabilistic Knowledge. Oxford: Oxford University Press. Nelson, M. (2010). “We Have No Positive Epistemic Duties.” Mind 119(473): 83–102. Osborne, R. C. (forthcoming). “What Do We Epistemically Owe to Each Other? A Reply to Basu.” Philosophical Studies. Paul, S. and Morton, J. (2018a). “Believing in Others.” Philosophical Topics 46(1): 75–95. Ramsey, F. (1929). “General Propositions and Causality”, in D. H. Mellor (ed.), F. P. Ramsey: Philosophical Papers. Cambridge: Cambridge University Press, 145–163. Reed, B. (2010). “A Defense of Stable Invariantism.” Nous 44(2): 224–44. Ross, J. and Schroeder, M. (2014). “Belief, Credence, and Pragmatic Encroachment.” Philosophy and Phenomenological Research 2: 259–88. Rudner, R. (1953). “The Scientist qua Scientist Makes Value Judgments.” Philosophy of Science 20: 1–6. Saul, J. (2018). “(How) Should We Tell Implicit Bias Stories?” Disputatio 10: 217–44. Schroeder, M. (2012). “Stakes, Withholding, and Pragmatic Encroachment.” Philosophical Studies 160: 265–85. Siegel, S. (2015). The Rationality of Perception. Oxford: Oxford University Press. Stanley, J. (2005). Knowledge and Practical Interests. Oxford: Oxford University Press. Stanley, J. (2016). “Is Epistemology Tainted?” Disputatio 8(42): 1–35. Stine, G. (1976). “Skepticism, Relevant Alternatives, and Deductive Closure.” Philosophical Studies 29: 249–61. Strawson, P. (1962). “Freedom and Resentment,” in Freedom and Resentment and Other Essays, Vol. 2008. London: Routledge. Stroud, S. (2006). “Epistemic Partiality in Friendship.” Ethics 116(3): 498–524. Thomson, J. J. (1975). “The Right to Privacy.” Philosophy & Public Affairs 4(4): 295–314.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 19/03/21, SPi
6
Predatory Grooming and Epistemic Infringement* Lauren Leydon-Hardy**
Predatory grooming is, broadly speaking, a preparatory process through which target individuals are primed, coached, or generally readied in some sense, for conduct that is exploitative in nature. Perhaps the most recognizable examples of grooming relationships are familiar from high-profile stories of abuse. In 2012, Malcolm Gladwell wrote of former Penn State assistant football coach Jerry Sandusky that he had “built a sophisticated, multimillion-dollar, fully integrated grooming operation, outsourcing to child-care professionals the task of locating vulnerable children—all the while playing the role of lovable goofball.”1 Predatory grooming is generally employed in circumstances where downstream exploitations constitutively involve participation or assent on behalf of targeted individuals. When the Penn State story broke, many wondered how this could be possible. How could so many boys, so many families, and so many administrators and officials, be caught up in a web of abuse that had remained hidden for so long? The answer, ineluctably, was that these victims had been groomed, exposed to sustained patterns of behavior aimed at rendering them acquiescent to—or even complicit in—conduct which, outside of the context of a grooming relationship, might otherwise have been readily recognized as harmful or exploitative. It is this epistemic aspect of the grooming relationship that I am interested in here. To understand grooming, we need to grasp not just the ethical dimension of the relationship, but also its epistemic element. At its heart, grooming is made possible by epistemic manipulation.
* This chapter includes potentially disturbing discussions of sexual violence and child abuse. ** I am exceedingly grateful to Gretchen Ellefson, Ian Elliott, Sandy Goldberg, Alex Guerrero, Jennifer Lackey, Trevor Nyman, Kathryn Pogin, Baron Reed, Stephen Turner, Nathan Weston, Stephen White, to audience members at the 2017 meeting of the Society for Exact Philosophy, and my co-participants at the 2018 Prindle Institute for Ethics: Applied Epistemology Retreat, all for invaluable comments and contributions to this project. 1 Malcolm Gladwell (2012). “In Plain View: How Child Molesters Get Away with It.” The New Yorker, September 24, 2012, www.newyorker.com/magazine/2012/09/24/in-plain-view. See also the case of Larry Nassar, former US Olympic Gymnasts physician, who recently plead guilty to numerous counts of sexual exploitation and assault, https://slate.com/sports/2017/11/larry-nassar-preyed-on- young-gymnasts-whod-been-taught-their-bodies-were-not-their-own.html.
Lauren Leydon-Hardy, Predatory Grooming and Epistemic Infringement In: Applied Epistemology. Edited by: Jennifer Lackey, Oxford University Press (2021). © Lauren Leydon-Hardy. DOI: 10.1093/oso/9780198833659.003.0006
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
120 Lauren LeydOn-Hardy As a basic strategy, grooming is motivated by a host of disparate ends; patterns of grooming behavior may emerge in diverse interpersonal settings. This variability is reflected in the ways that research on predatory grooming has developed. Some of this theoretical disunity owes simply to the scope of the analyses on offer. So, for example, much of the work on predatory grooming deals with child sexual exploitation at a high level of generality.2 More narrowly, though, a good deal of research focuses particularly on child exploitation online.3 Still other theorists emphasize particular interpersonal dynamics that may be vulnerable to grooming for the purposes of sexual coercion and exploitation, such as teacher/student4 or coach/athlete5 relationships. Notably, however, this latter body of research is not restricted to child sexual abuse; research here also models predatory sexual grooming in higher education and competitive athletics,6 as well as in professional mentoring and tutoring relationships.7 Further research seeks to understand grooming in peer-to-peer contexts.8 Yet further research on grooming addresses abuse that is not sexual in nature. For example, grooming patterns have been identified and studied in relation to the elderly and medically dependent.9 And in 2011, Drs Amy Hagopian and Kathy Barker argued in the American Journal of Public Health that recruiters for the US armed forces, deployed to high schools under the mandate of the No Child Left Behind Act, exercised grooming behavior in sustained efforts to recruit and enlist high-school aged children.10 Clearly, then, predatory grooming is not specific to any one form of exploitation, nor is it unique to any particular kind of relationship. In a recent interview, UK Ministry of Justice forensic psychologist Ian Elliott—whose work we will return to in section 1 below—discussed predatory grooming across the contexts of sexual predation and extremist recruitment online. Elliott explained: Say for example you have someone who has a burgeoning interest in a certain religion or a certain social movement, this could lead them to places that could make them slightly more vulnerable in the sense that their curiosity in a topic could be exploited. Again, the goal of that individual is perfectly legitimate, but it could lead them to a crossroads with people who would be interested in exploiting them. (www.operation250.org/ian-elliott) Part of Elliott’s point is that relevant vulnerabilities here needn’t be understood in terms of hardships, and the kinds of people vulnerable to predatory grooming needn’t be regarded as especially frail, or otherwise remarkable.
2 Cf. Craven et al. (2006); McAlinden (2006). 3 Cf. Kloess et al. (2017). 4 Knoll (2010). 5 Brackenridge (1997); Brackenridge and Fasting (2005). 6 Volkwein-Caplan et al. (2002). 7 Haring-Hidore and Paludi (1987). 8 Ashurst and McAlinden (2015). 9 Brandl et al. (2006). 10 Hagopian and Barker (2011). More recently, Representative Alexandria Ocasio- Cortez introduced a draft amendment in the House that would prohibit the use of video game, e-sports, and streaming platforms such as Twitch.com in recruitment efforts. www.vice.com/en/article/889vbv/ aoc-introduces-measure-to-stop-the-military-from-recruiting-on-twitch.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Predatory Grooming and Epistemic Infringement 121 What is prototypical of predatory grooming is not just that it works to coerce or otherwise engineer the cooperation of its victims in the course of their own exploitation. Ensconced in grooming relationships victims are party to their own injury, but on its own this is not unique to the phenomenon. Arguably quid pro quo arrangements (and even outright threats of violence) can have the effect of coercing individuals into acquiescing to, or participating in, abusive misconduct. In cases of predatory grooming, however, there is an essential dimension of cognitive or psychological interference, an integral element of unknowingness bound up in relationship itself. This can be contrasted with other forms of manipulation, in which the target is more fully in the know with respect to the nature of their circumstances. Quid pro quo arrangements are transactional; they involve a mutual agreement, even if that agreement is coercive. Or the victim of a long-con might be manipulated into participating in a bank heist. Though the behavior is coerced through manipulation (albeit manipulation that such a victim fails to track, ex hypothesi), she knows that she is robbing a bank. Here, the manipulation is located in her distorted motivation, rather than her distorted sense of reality. It is the difference between thinking, “I robbed the bank because he loves me” and “I am not being molested,” even as one is being molested. Victims of predatory grooming are almost entirely, and in many cases remain for years after the fact, completely in the dark as to the nature of the abuse that they suffer. When victims of abuse within grooming relationships do come to understand what has happened to them, their testimonies do not reflect a realization that they have been deceived (as with the dupe in a long-con) or coerced (as in cases of quid pro quo). The sense of epistemic obtrusion is more robust; it is a sense of having been changed, estranged from oneself, and thereby rendered complicit in (in some cases) years of abuse. Grooming is interpersonal, protracted, marked by asymmetries of power, serves to cultivate undue influence, and, most essentially, aims at masking abuse even by the lights of the abused.11 This is the sense in which grooming crucially involves the cultivation of an unknowingness in its victims. Grooming is not just a very complicated kind of lie; it is a sustained campaign of manipulation and coercion that is marked by a distinctively epistemic type of abuse. Groomers must hide in plain sight, even from their victims. Predatory grooming is thus properly understood as, among other things, an extreme form of epistemic misconduct: it is a preparatory process consisting of patterns of goal-directed behavior employed by would-be predators that can have the effect of untethering victims from their epistemic resources and from their ability to marshal those resources appropriately and responsively. 11 Might peer-to-peer cases of grooming be an exception to the point about asymmetries of power? I think this depends on your conception of power. It is plausible, for example, that in instances of grooming among peers there are distributions of social power (i.e. popularity) that make it possible.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
122 Lauren LeydOn-Hardy Predatory grooming has been studied primarily by forensic and social psychologists. There is tremendously valuable work in this literature; however, it inevitably fails to elaborate on the normative dimensions of the phenomenon. Forensic psychologists are asking one kind of question: how are groomers operationalizing their grooming programs? But we can ask another, equally illuminating question: why does grooming work? What is it like to be groomed, from the inside? We will see that grooming is made possible, in large part, by the ways in which predators are able to exploit and subvert the epistemic norms and expectations attendant to their relationships with their groomees. Grooming is about shaping the way a victim can, and will, reason about her relationship to her abuser. Thus, a more complete account of predatory grooming requires an analysis of its epistemic dimensions. This chapter examines a forensic psychological model of grooming behavior with an eye towards how we might amp up the explanatory power of the model by analyzing it in terms of the relevant norms. Understanding the epistemology of predatory grooming can help us to explain not just how it works, but how and why it is so harmful. In the latter sections of the chapter, I survey several conceptual resources from applied epistemology and show why I do not think these are quite apt to describe what we see in cases of grooming. Finally, I sketch a picture of the conceptual resource that I suggest can do the job. I call this epistemic infringement: a category of epistemic misconduct that works through the systematic contravention of interpersonal social and epistemic norms in a manner that subverts the epistemic agency of targeted individuals.12
1. A Generalized Social-Psychological Model of Predatory Grooming Grooming relationships center around patterns of goal- oriented behaviors, directed at some type of exploitation. The goals nested within these patterns of conduct can be varied, multiple, and hierarchical (in the service of one another). They also morph from context to context, and in response to interpersonal developments between perpetrators and their targets. Grooming relationships are dynamic; they develop over time and across milestones and decision points. Yet, understanding grooming in terms of goal-oriented behavior does not entail (though, certainly does not preclude) that perpetrators understand their behavior explicitly in terms of grooming, or even exploitation. For example, in cases of peer-to-peer grooming, it is, I think, particularly psychologically plausible that while we may identify goal-oriented patterns of behavior that culminate in 12 It is beyond the scope of this chapter to offer an account of epistemic agency. For my own view of what epistemic agency amounts to, see Leydon-Hardy (ms.).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Predatory Grooming and Epistemic Infringement 123 exploitative conduct, perpetrators may nevertheless fail to conceive of their own behavior in exploitative terms.13 Thus, at least in some cases, a lack of reflective self-knowledge should not be seen as a barrier to understanding the relevant patterns of misconduct as properly described as predatory grooming. The model I am working with comes out of work being done by Ian Elliott, the aforementioned forensic psychologist working with the UK Ministry of Justice.14 Elliott’s model is ideal for two reasons. First, Elliott’s work emphasizes generality. While much of the literature on grooming has specifically been geared towards modeling grooming patterns in child sexual exploitation, and particularly online child sexual predation, Elliott’s model aims at countenancing a plurality of grooming contexts.15 Second, Elliott’s model clearly explains the shifting and context-sensitive nature of the goal-oriented behavioral processes that are unique to grooming.16
Rapport
Incentive Sensitivity
Contact Disinhibition
Goal achievement
Phase 2: Disclosure
Security
Phase 1: Potentiality
13 Cf. Ashurst and McAlinden (2015). See also Stephen Darwall’s (2006) discussion of self- rationalizing behavior among “tyrants and batterers” in The Second Person Standpoint, 51 (discussed in Abramson 2014). 14 Elliott (2017). 15 Elliott’s model is generalized to countenance a plurality of sexual grooming scenarios. In private correspondence, we have discussed that there is no principled reason to suggest that this model (or something very close to it) is not further generalizable to additional exploitative motivations. For further discussion, see McAlinden (2006); Olson et al. (2007). 16 Elliott (2017: 7): “A self-regulation model of illicit grooming. Black solid lines/arrows indicate progress through the model; gray dashed lines/arrows indicate feedback loops.”
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
124 Lauren LeydOn-Hardy Grooming begins with what Elliott identifies as the potentiality phase.17 Here, a perpetrator cycles through four behavioral mechanisms with the aim of rendering a target amenable to exploitation. The effectiveness of these operations within the potentiality phase are tested in occasional cycles through the disclosure phase, where groomers assess victim receptivity to goal-related information and/or goal achievement. We will return to this concept of ‘testing’ in a moment, but first let me say a bit about each of the component mechanisms operant within the potentiality phase. The potentiality phase consists in four behavioral mechanisms: (1) rapport building; (2) incentivization; (3) disinhibition; and (4) security management. In rapport building, perpetrators work to establish and manage the quality of the relationship between themselves and their target. Rapport building aims at establishing positive aspects of the interpersonal dynamic, emphasizing mutual attentiveness and shared personality traits, cultivating shared experiences, and so on. Incentivization is the process by which perpetrators create and regulate incentives for targets. Incentives are highly context sensitive but will span a broad range of motivators and attractors, including gifts, money, compliments, praise, favors, opportunities, and more. Throughout the disinhibiting process, perpetrators work to reduce a target individual’s ability and/or willingness to resist. Disinhibition can involve psychological and/or emotional manipulation that might have a limiting effect on a target’s ability to react assertively or convincingly. Disinhibition might also be more flat-footed as with, say, the introduction of drugs or alcohol, substances which act as natural disinhibitors. In security management, perpetrators maximize control over information introduction, as well as third-party disclosures concerning the circumstances of the relationship. So, for example, in dependent-care grooming, this could mean delimiting conversations and quality time with a target individual’s family members; in online child sexual predation, this often means questioning the target about where their computer is located in the home, what their parents do for work, how often they are monitored online, and so on. Importantly, security management may also involve managing a groomer’s own relationship with other members of a target individual’s community or family. This is because positive relationships within the broader community can help to camouflage red flags. And, of course, notice that camouflaging red flags for third parties also has a sort of feedback effect on the epistemic position of the target individual relative to the underlying nature of the relationship: without friends and family voicing concern, one is less able to see or to seriously entertain the 17 Elliott in fact claims that grooming begins with the initial contact between the perpetrator (in Elliott’s language, the “protagonist”) and the target. In the spirit of theorizing about the grooming process with an eye to the most generalizable model, I want to avoid committing myself to this. Conceivably, there might be cases where a perpetrator and a target know each other prior to beginning the grooming process (e.g. in peer-to-peer cases a perpetrator may have matriculated with a target prior to initiating grooming behavior). This choice point should not undermine any further reliance on Elliott’s model for my purposes. However, it bears mentioning that choice points very much in this wheelhouse are emphasized throughout the literature on grooming, as theorists emphasize the challenges they face in articulating the process with clarity and specificity.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Predatory Grooming and Epistemic Infringement 125 possibility that something is awry.18 Each element of the model has subordinate goals nested within their behavioral processes. Within the rapport-building process, a single, nested goal might be to establish a set of shared interests. Establishing shared interests creates the context necessary for establishing a set of shared experiences. Shared experiences advance the overall rapport-building process, creating a context for introducing incentives. Finally, there is the disclosure phase. In disclosure, perpetrators aim to desensitize victims specifically to goal-relevant information and behavior. We will see all of this in much greater detail in section 2 as we look more closely at both hypothetical and real-world examples of grooming relationships. For now, though, I want to highlight that the path through the mechanisms of potentiality and disclosure is importantly non-linear; each individual process works interconnected to the others, in a cycle referred to as test–operate–test.19 In grooming relationships, test–operate–test describes how groomers cycle through the potentiality and disclosure phases, gathering information about the status and receptivity of the groomee and choosing succeeding mechanisms in light of the effectiveness of the previous. It is a self-regulatory mechanism, referring to the way in which groomers “take the temperature” of their victims. All four mechanisms of potentiality, as well as the disclosure phase, involve test–operate– test; a groomer might initiate an ‘operation’ of the disclosure phase with the introduction of goal-related information and ‘test’ for the groomee’s receptivity— does she recoil at the remark, or does she take it in stride? Depending on the groomee’s response to the test, the groomer might cycle back through potentiality, ‘operating’ one of the four constituent behavioral mechanisms, before ‘testing’ again to learn how the previous operation has affected the groomee’s receptivity. The general move is to “test out” some behavior from a chosen mechanism, assess the response, and try some more (or not), based on that assessment. Depending on the response, the groomer can try out a little more desensitization (the overarching goal of the disclosure phase), or cycle back to one of the four mechanisms of potentiality, where those mechanisms serve as a kind of
18 Grooming is interpersonally dyadic in the sense that only the groomee will ever be cycled through the disclosure phase of the model. Nevertheless, groomers may engage in rapport building, incentivization, and disinhibition with third parties. In the documentary Leaving Neverland, for example, a great deal of attention is paid to the incentive structures that Michael Jackson set up for the families of his victims. There, we see security management leveraging off incentivization, where many incentives are importantly directed outwardly (at the families) in order to maximize the effectiveness of security management, protecting the grooming relationship and allowing for phase two disclosures and goal achievement. Should we understand the groomer’s third-party relationships as grooming, in their own right? These relationships are certainly highly manipulative. However, insofar as security management relationships are managed in the service of disclosure and goal achievement, I am hesitant to describe these relationships as grooming, generally. 19 Elliott (2017, 7). In The Human “Interaction Engine” (2006) Stephen C. Levinson attributes the concept of test-operate-test to “the birth of cognitive science” when “Miller et al. (1960) suggested that the Test-Operate-Test-Exit (TOTE) unit should replace stimulus-response as the basic theoretical unit of human behavior: we test to see if the intended goal was achieved, if not, operate on it and try again. In cooperative interaction the only way to test is to see what the other person made of our actions.” (51)
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
126 Lauren LeydOn-Hardy inoculation against the desensitizing work of disclosure, while desensitization is most closely aligned with goal achievement.20
2. Epistemicizing the Model With Elliott’s model in view, let’s look at some cases. First, we can see the test–operate–test mechanism at work very clearly in this excerpt from Malcolm Gladwell’s essay on the Sandusky case: Sandusky started with wrestling, to make physical touch seem normal. In the shower, the boy initially turned on a showerhead a few feet from Sandusky. Sandusky told him to use the shower next to him. This was a test. The boy complied. Then came the bear hug. The boy’s back was touching Sandusky’s chest and his feet touched Sandusky’s thigh. Sandusky wanted to see how the boy would react. Was this too much too soon? The boy felt “weird” and “uncomfortable.” Sandusky retreated. The following week, Sandusky showed up at the boy’s home, circling back to test the waters once again. How did the boy feel? Had he told his mother? Was he a promising lead, or too risky? (Gladwell 2012, www.newyorker.com/magazine/2012/09/24/in-plain-view)
Here we see the test–operate–test mechanism cycles through iterations of disclosure/disinhibition (testing physical boundaries) and security management (checking third parties for information disclosure). As we have seen, t est–operate–test is the mechanism by which predators update their information and regulate their behavior accordingly. Test–operate–test is an effective information-gathering tool, but it can also be powerfully disorienting for the target. We will begin with a hypothetical case in the very early stages of a potential grooming relationship. Later, we will look at post-grooming/goal achievement testimony as well, but I think it is illuminating to begin this discussion with early-stage grooming relationships. It is easy to wonder (outside-looking-in) how a groomee could become entrenched in a relationship that the public typically only learns of once it is quite far along, by then readily identifiable as problematic. As with any abusive relationship, though, grooming relationships evolve over time, in a way that inside-looking-out, it can be difficult to spot the warning signs. Consider a case like Professor/Student co-author:
20 That is, if the goal is sexual abuse, the desensitizing behaviors will be sexual in nature; if the goal is financial exploitation, the desensitizing work will align with that goal. Thanks to Elliott for suggesting the metaphor of ‘inoculation’ in our discussion of the relationship between disclosure and potentiality.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Predatory Grooming and Epistemic Infringement 127 Student is new to Professor’s program and hopes to one day work in Professor’s area of research. Professor proposes co-authoring a paper with Student and suggests that Student join Professor for dinner to discuss the paper they will write, as well as Student’s other academic work. At dinner Professor proposes ordering a bottle of wine, which Student initially declines to share. “My mistake, I hadn’t realized my young co-author was quite so young!” Professor teases. Embarrassed, Student acquiesces.
Professional attention and professional opportunities are paradigmatically incentivizing, and alcohol is a natural disinhibitor. When Student expresses reluctance in the face of a perceived infraction of Professor/Student norms, a groomer has a choice. They might, for example, interpret resistance to the introduction of a disinhibitor as a cue to cycle back to rapport building, or alternatively to double down and introduce a different type of disinhibitor: teasing; that is, a disinhibitor that works to diminish reluctance by associating it with embarrassment. We noted earlier that the disinhibition behavioral mechanism serves to diminish a groomee’s ability to resist or to express discomfort. Disinhibition may be achieved through the introduction or drugs or alcohol, or by inculcating to the groomee a sense that reluctance will not receive uptake, that it is “no use,” or that it is interpreted (or interpretable) as insincere.21 Disinhibition may also be achieved through the introduction of associated costs or risks, as when inhibited behavior becomes associated with feelings of guilt, embarrassment, or paranoia. It should be noted that teasing, of course, does not necessarily serve as a disinhibitor in any criticizable sense, even in relationships marked by asymmetries of power. We can easily imagine circumstances in which a perfectly above-board professor engages in teasing behavior with a student for whom they indentify shyness as a challenge, or something in that ballpark; that is, we can imagine cases where teasing is a pedagogical tool, and we can imagine the worst-case scenario of Professor/Student co-author where what under different circumstances might have been a pedagogical tool is actually used in a quite different way. Indeed, this is all very much to the point: that the way in which all of these tools, strategies, and behaviors manifest are generally governed by the norms that structure our thinking about them, in the moment. We use interpersonal norms like we use a compass. I think I know true north, but if my compass is depolarized, then I’m in real trouble. Something like this is true of the norms here. In the good case, where Professor is trying to help a shy student emerge from a shell that is holding her back, Student can safely interpret being teased according to the healthful norms of mentorship and professionalization. She might think something along the lines of, this is a gentle way of encouraging me to be more outgoing. But notice that in the 21 Elliott (2017): “[D]isinhibition . . . seeks to reduce the target’s ability to respond genuinely, accurately, cogently, and convincingly.”
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
128 Lauren LeydOn-Hardy bad case, this same reasoning is available to Student, a useful fact for an unscrupulous Professor. Disinhibiting behavior diminishes a target individual’s ability to resist; one way of accomplishing this is to imply that a target individual’s reluctance is somehow misguided. The implication that Student is too immature, or inexperienced, might act as an effective reprimand against the backdrop of the invitation to co-author. This may work not only to make Student hesitant to express resistance, but, perhaps more troublingly, to dampen Student’s desire to resist, at all.22 Looking back, Student may reflect on their change of heart, “I wanted to drink with Professor (because I wanted to co- author with Professor).” Professor/ Student co-author may then illustrate multiple mechanisms of the grooming process in tandem. We see Professor inviting Student to appreciate their special rapport, predicated on a potent incentive, while further suggesting that Student’s resistance to the introduction of a disinhibitor is inappropriate given the circumstances. Invoking their co-authorship in Professor’s response might reasonably send a particular signal to Student, that the norms governing their relationship have shifted from strictly professor/student to those of co-authors, a kind of peerhood. And the norms governing that relationship are different; looser in some ways, perhaps more demanding in others. From an epistemic vantage point, this marks the introduction of uncertainty concerning what the operative norms governing the relationship are, or should be. The exact provenance of this uncertainty is slippery; it may issue from Professor’s actions (his behavior suggests that the norms are thus and so, so perhaps they are thus and so), from testimony or other sources of evidence (the things Professor says, how Professor says them), or from the changing attitudes of Student (perhaps I do, or ought to, want the norms to be thus and so). It might be objected that the example in Professor/Student co-author is implausibly benign. I think that if the case looks unambiguously harmless, this is because it is pitched at a level that is operationalizable in the earliest days of a grooming relationship. The thought here, and suggested earlier, is perhaps helpfully borne out by the familiar (if gross) fable of the boiling frog: abusive relationships never pick up at the point where the water is already boiling. Consider a separate, non-hypothetical case. In interviews with James Safechuck and his mother, Stephanie Safechuck, they describe how Michael Jackson was both exceedingly generous (incentivizing)—once palming James several hundred dollars after a movie night—and strikingly childlike (rapport building, security management). When Jackson invited James to his set trailer, Stephanie explains, “[James] said to me ‘He’s like a nine-year-old little boy.’ So, that made me feel comfortable.”23 Later, describing Jackson’s first visit to the Safechuck’s home, Stephanie recounts, “He came to our house and went through Jimmy’s closets, 22 For an excellent discussion of deformed desires, see Superson (2005). 23 Reed (2019, Pt 1).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Predatory Grooming and Epistemic Infringement 129 went through everything, acted like a little boy, giggling.” Here is James: “Michael would come to our house a lot . . . the more I would visit the more you would get sort of alone time with him. I brought him some toys, and he gave me some gifts. It’s more like hanging out with a friend that’s more your age. So it feels natural.” Knowing what we know now, this is gut-wrenching. But the picture from the inside is murkier. A wildly famous person reports that he is tremendously lonely to the parents of a talented young fan and forms a relationship with their family. He is goofy and childlike in his interactions with their son—why shouldn’t he be? He is also generous, yes, but he is Michael Jackson—he can be. And he gives the impression that he is grateful for their deepening familial intimacy. Stephanie Safechuck recalls, “I came to feel like he was one of my sons by how he behaved . . . He was a son I started to care for. He would spend the night. I’d wash his clothes.” From a certain vantage point, again, nothing predatory has happened here: a beloved popstar befriends the family of young dancer, who then receive all of the privileges of wealth and fame by association, and who in turn creates a warm, familial environment for this successful, though isolated, man. Again, none of this, on its own, is obviously predatory. And yet it is impossible to believe that these early days of incentivization, rapport building, and security management are anything short of pivotal in James Safechuck’s life. The teasing in Professor/Student co-author captures just one piece of the broader picture. Perpetrators cycle between behavioral mechanisms responsively, each move informed by the previous. The bite of disinhibiting behavior can be quickly eclipsed by switching to incentivization or rapport building, depending on what a perpetrator believes will be most effective in light of what they know about target receptivity, the particulars of their circumstances, the relevant power asymmetries, and so on. The test–operate–test mechanism models how groomers regulate the interpersonal dynamics of the relationship, as well as what target individuals can expect or anticipate, given the norms as they are established (and distorted) by the groomer. In Professor/Student co-author, we saw Professor enforcing potentially warped norms of collegiality connected directly to incentives that Professor introduces to Student: co-authorship, mentorship, fine dining, etc. By leveraging disinhibition in the form of a chiding remark about Student’s reticence to drink against a powerful stack of incentives, and in light of professional norms that Student assumes (norms of professionalization, mentorship, and so forth) Professor is in a position to shape Student’s expectations about what is normal in this context. I want to flag here two further points about the test–operate–test mechanism. The first concerns the test–operate–test mechanism as wielded by the groomer, and the second, as it may be experienced by the groomee. Part of why the test–operate–test mechanism is so effective is that it allows the groomer to leverage these behavioral mechanisms off one another. So, for example, incentivization coupled with, or followed by, disinhibiting behavior can work to powerfully
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
130 Lauren LeydOn-Hardy reinforce the effectiveness of both of those processes going forward. To see how, consider a second hypothetical, Professor/Student conference: Professor offers to pay Student’s way to a conference abroad but advises her to keep the funding a secret. The conference represents a significant professional opportunity; however, the financial assistance together with the injunction to keep it a secret gives Student pause. Student’s hesitancy provokes an expression of wounded remorse. Professor tells Student that he finds it deeply regrettable that she would ever suspect his behaviour as untoward. He rescinds the offer to pay Student’s way to the conference.
One thing to note here is that if Student had accepted Professor’s invitation, Professor would have had a profoundly successful cycle through both incentivization and rapport building, having established a shared secret as well as a shared experience, in addition to providing Student with the financial and professional resources to pursue a career-building opportunity. Nevertheless, this is no loss for Professor. What we see instead is disinhibition leveraging itself off incentivizing behavior in a way that can serve to shore up both mechanisms going forward. Professional and financial incentives are rescinded in response to an expression of doubt, implying that Student has wounded Professor by expressing reluctance or suspicion. One imagines Student feeling guilt, fear of further reprisal or lost opportunity, as well as disappointment at a professional opportunity foregone. But one also imagines Student experiencing creeping self-doubt: was the offer really problematic? Professors do (or ought to), after all, have an interest in the flourishing of their students, and it had been an exciting professional opportunity. Given the relevant professional norms, Student might reason that perhaps they had been wrong to be suspicious. Here, as in Professor/Student co-author, the relevant norms are doing a sort of double work: the very norms transgressed by the dubious invitation get appealed to in Student’s undermining her sense that something was wrong.24 All of this will work against Student’s epistemic health in 24 Another example of this kind of inversion of norms is at work in the victim testimony of Yazidi girls captured by ISIS. Rescued Yazidi girls report that while in captivity their captors told them that because they had been defiled, their community would never have them back. Whatever else we might say about purity norms of this sort, here we see the norms of the Yazidi community, which before their capture these girls regarded as healthful, turned against them in order to subvert their ability to reason about their circumstances, and to coerce them into participating in their exploitation. Indeed, when rescued, some Yazidi girls have reportedly declined to leave, initially insisting that they are members of the caliphate and married to their rapists. These behaviors and attendant traumas are described in explicitly epistemic language by both reporters and family members; the girls came to believe what their captors had told them and trying to unseat these beliefs is described as requiring an overabundance of counterevidence (see especially Episode 9 Pt II of The New York Times 2018 podcast Caliphate, www.nytimes.com/interactive/2018/podcasts/caliphate-isis-rukmini-callimachi.html). On December 18th, 2020 the Times issued an Editors’ Note indicating that parts of “Caliphate” failed to meet their evidentiary standards, writing, “In the absence of firmer evidence, “Caliphate” should have been substantially revised to exclude the material related to Mr. Chaudhry.” The Times declined to retract “Caliphate” in its entirety because Mr. Chaudhry’s narrative is only one part of the investigation. Given that the testimony cited here of the Yazidi girls’ and their family members is recorded
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Predatory Grooming and Epistemic Infringement 131 at least two ways. First, it will make it harder for Student to experience or recognize future doubt or suspicion. When Professor shifts from disinhibition (the rescinding of the offer) to rapport building (emphasizing their shared interests) or incentivization (new mentorship opportunities), the relief and lingering guilt will act as a powerful disincentive to regarding Professor’s future behavior with suspicion. Her ability to adjudicate future evidence is thus compromised. Second, this interaction has raised the stakes for Student expressing discomfort going forward. If something later seems or feels off, Student might think, “I’m not making that mistake again!” The interaction thus erodes Student’s perceptual and expressive competencies. And note that the latter issue of self-censorship is not a purely practical matter. Plausibly, if Student represses expressions of discomfort in future interactions, this will further warp the dynamic between Professor and Student in ways that will lead to certain objections not being discussed, certain assumptions being taken as true, and certain hypothesis not being considered—all of which will have a detrimental long-term effect on Student’s epistemic health. The norms that a target individual takes to govern a relationship can be exploited with the effect of constraining their expectations for, and interpretations of, various behaviors and interactions within the relationship. In Professor/ Student co-author, what makes Professor’s chiding remark effective is the suggestion that Student does not understand what would, or would not, be appropriate in this context. The idea here is to give the impression that while both parties agree on the appropriate norms (the norms of a professor/student relationship), Student is nevertheless mistaken about what those norms dictate, or how the norms apply. Thus, even as Professor transgresses healthful norms of mentorship, those same norms form a premise in Student’s reasoning that all is well. Part of what is happening here is the cultivation of a relationship over and above, or apart from the relationship that the target individual is aware of: a licit, surface-level relationship (Professor/Student) and an illicit, subterranean relationship. Groomers trade on the norms as they are understood relative to the licit relationship, while simultaneously cultivating the illicit relationship. Hiding in plain sight, they are camouflaged by the groomee’s misapprehension of which norms are doing what, given the strictures of the relationship as she takes it to be. We see this in Professor/Student conference, where Student explains away a compromising situation by appeal to the appropriate norms of a professor/student relationship. It is a contravention of those norms to put a student in such a compromising position, but under the right circumstances the very norms being contravened might also form the basis of an argument to the contrary. “Of course Professor invited me to that conference, it’s a great opportunity!” The desirability of the opportunity, the shame of the reprisal, and the risk of losing a mentor, together make the rationalization on the basis of the healthful norms compelling. first-hand, and because their stories should not be ignored in view of their proximity to other journalistic failings, I continue to include this footnote.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
132 Lauren LeydOn-Hardy An analysis of the test–operate–test mechanism in these normative terms illustrates how epistemically impactful it can be. Particularly in relationships marked by asymmetries of power (as grooming relationships so often are), a sudden leap from an expression of kindness to an expression of derision or resentment can be profoundly disorienting. And yet those sudden leaps can also become normalized, in part through the groomee explaining them away by appeal to the otherwise healthful norms of the licit surface-level relationship. In its own right, the dizzying interplay of these mechanisms—of skipping back and forth between incentivization, which can make the groomee feel special, and disinhibition, which can make the groomee feel distressed, or fearful—is central to the disarming and deluding effects of a grooming relationship. On the one hand, disinhibition can signal that there are costs to rocking the boat, and in this way, narrow the range of responses that a groomee takes to be appropriate. Disinhibitors coupled, or associated with incentivizing behavior, can also undermine a target individual’s confidence in their ability to discern the true nature of an interaction or experience. When Professor reacts to Student’s hesitancy, Professor signals that there are costs to not going along: lost opportunities, a crack in the veneer of Professor’s burgeoning mentorship. Yet a further element of this interaction is internal to Student’s own mind: wondering whether they were wrong to be suspicious, beginning to doubt their ability to adjudicate their evidence, and feeling remorseful for having misjudged. I’ve been trying to make the case that this phenomenon is importantly epistemic, a claim that I think is supported strongly by victim testimony and impact statements. We see this, for example, in much of the court testimony after the conviction of Olympic Gymnastics physician, Larry Nassar. Many of Nassar’s victims spoke about their inability to understand or to believe what was happening or had happened to them.25 Kyle Stephens testified that Nassar, who had been a friend of the family, began abusing her at a young age. As a child, Stephens tried to tell her parents what was happening, but Nassar claimed that she must be confused, and moreover, that should anything like that ever in fact happen to her, she should tell an adult right away. Her parents agreed. At Nassar’s trial, Stephens testified that, “It took the media coverage of the sexual abuse scandals in the Catholic Church and a friend confiding the details of her sexual abuse for me to realize that something was wrong.”26 Nassar’s response to a young Kyle Stephens highlights the roles of security management and rapport building in a grooming relationship, as well as the link between those mechanisms of grooming and the epistemically distorting consequences of being groomed. In Stephens’s testimony, 25 The Nassar trial provided an avalanche of relevant testimony from victims and professional associates, alike. For another fascinating narrative about Nassar’s grooming techniques see also Lauren Hopkins’s (2018) description of her relationship with him as a journalist writing about the world of gymnastics: www. huffingtonpost.com/entry/opinion-hopkins-gymnastics-nassar_us_5a673cfce4b0e5630073862f. 26 See Barajas (2018), www.pbs.org/newshour/nation/sexual-abuse-survivors-confront-former-usagymnastics-doctor-little-girls-dont-stay-little-girls-forever.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Predatory Grooming and Epistemic Infringement 133 we see a concrete and distinctly epistemic harm in the superabundance of evidence necessary for her to adjudicate the facts of her own abuse later in life.27 We also see the norms of the licit, surface-level relationship working in the service of cultivating the illicit, subterranean relationship. Even as Nassar raises the stakes for Stephens speaking up about his abuse by suggesting that she is at best mistaken, at worst a liar (disinhibition), he is further cultivating a relationship of trust with her parents (security management). Met with their daughter’s serious allegation, Nassar responds by counselling his accuser with care. Thus, even as he transgresses the norms of an adult/child relationship, his transgression is masked (both to Stephens and her parents) by invoking those norms directly: we are the adults, tell us if something like this ever happens to you. What we see, then, is that this subversion of the norms can have the effect of turning a victim’s epistemic resources against her: you’re supposed to trust your parents’ friends, and this makes their testimony and authority all the more evidentially weighty. In some cases, you might trust them even more than you trust yourself, your experiences, or your memories. The norms that abusers exploit in cultivating these relationships become the norms that victims draw on to explain away the abuse itself. A further example of this can be seen clearly in Rachel Denhollander’s statement, where she describes the way that community and professional norms of trust and expertise shaped her reasoning. She testified at Nassar’s sentencing hearing: I was barely 15 when Larry began to abuse me and as I lay on the table each time and try to reconcile what was happening with the man Larry was held out to be, there were three things I was very sure of. First, it was clear to me this was something Larry did regularly. Second, because this was something Larry did regularly, it was impossible that at least some women and girls had not described what was going on to officials at MSU and USAG. I was confident of this. And third, I was confident that because people at MSU and USAG had to be aware of what Larry was doing and had not stopped him, there could surely be no question about the legitimacy of his treatment. This must be medical treatment. The problem must be me. (Denhollander, 2018, www.cnn.com/2018/01/24/us/rachael-denhollander- full-statement/index.html)
Note how normative language features here in Denhollander’s deliberative reflection. We might think the expected reasoning would be: “This is wrong. If it is
27 Of course, this also looks like a straightforward case of gaslighting. I will flag here that I think gaslighting could describe the one-off testimony/denial here, and that gaslighting in the service of the broader grooming relationship is compatible with everything I’ve said so far. In Leydon-Hardy (ms.) I argue that gaslighting itself may constitute epistemic infringement. I will return to this thought in section 4.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
134 Lauren LeydOn-Hardy wrong, then he shouldn’t be doing it. So, he shouldn’t be doing it.” But, instead, we get this kind of reversed reasoning in her deliberating: “If this is wrong, then he shouldn’t be doing it. But he is doing it. So, it must not be wrong.” And notice that in the sentences that follow this deliberation, “medical treatment” and “problem” are clearly terms that are loaded with normative significance, such that it is clear Denhollander is invoking the relevant norms to reason her way out of the feeling that something is wrong. Again, we see something similar in Eliza Dushku’s testimony about the abuse she suffered from stunt coordinator Joel Kramer on the set of the film True Lies. Dushku writes, “I remember, so clearly 25 years later, how Joel Kramer made me feel special, how he methodically built my and my parents’ trust, for months grooming me.” Dushku draws a straight line from her abuser’s relationship with her parents to her own trust in him. We can also see the coercive and disorienting distortion of reality in her abuser’s portrayal of their relationship: “When he was ‘finished’, he suggested, ‘I think we should be careful . . . ,’ [about telling anyone] he meant. I was 12, he was 36.”28 Note how the abuse becomes a part of the rapport building: the violation is now a shared secret (a hallmark of rapport building) and this particular instance of rapport building then gets leveraged into a potent disinhibitor. If the fact of what he has done to her is a shared secret, then by implication, the responsibility for what he has done is also shared. It was her secret, too. Now, abuser and victim are “in it together”; their connection is deepened even as the stakes have gone up for speaking out, objecting, or asking for help. This dynamic of victim as her predator’s keeper can manifest in multiple ways across a variety of grooming relationships. I think both Denhollander and Dushku’s stories are examples of this.29 In Denhollander’s case, we saw her reverse reasoning herself, drawing on the virtues of her abuser (his stature, his medical expertise) in order to reinterpret the abuse. She’s doing an important part of the work for him. In Dushku’s case, there is an explicit suggestion of complicity. Both of these are strategies by which predators use a victim’s inner life against her; they are hiding behind their victims. In Dushku’s case, by making the secret not just Kramer’s but Dushku’s too, Kramer had established Dushku as co-conspirator, as a part of the thing that had (for Dushku) felt very wrong. Imagine in Professor/ 28 See Dushku (2018), www.facebook.com/OfficialElizaDushku/posts/1769957739689557. Indeed, there are many clear indications of the mechanisms of grooming at work in Dushku’s written testimony. For example, she writes that Kramer nicknamed her “jailbait” (disinhibition) and “made her feel special” (rapport building) before he invited her to a special party for crew members, where she could go swimming and have her very first sushi dinner (incentivization). We also see here the ways in which incentivization itself can be normatively loaded, inasmuch as the predator is holding out the promise of a new kind of relationship, which will be defined in part by its governing norms. This suggests that some kinds of incentivization cannot be understood without seeing them as normative goods. 29 I also think something like this is going on in Professor/Student Co-author when we imagine Student thinking, retrospectively, that she wanted to drink with Professor. The exculpatory work that these choices do in grooming relationships is an interesting and important feature of the epistemology of grooming.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Predatory Grooming and Epistemic Infringement 135 Student conference that Student had accepted Professor’s invitation and gone along with keeping the secret. Most of us would agree that Student would be innocent of any wrongdoing. Still, it is not hard to imagine how from the inside, Student might feel guilty about something—perhaps of accepting some unfair advantage, or of agreeing to keep the secret—and why Student might fear their peers learning of either the opportunity, the secret, or both. Indeed, these very fears would serve Professor well by further isolating Student, another tactic of security management. By now, we have seen several epistemically unhealthful elements at work in grooming relationships, both real and hypothetical: • misleading testimony/counter-testimony; • distorting norms: ∘ what they are, should be, how they apply, and whether one is tracking them accurately; • constraining evidence: ∘ inculcating false beliefs to third parties, rendering them defective sources of information/consultation; ∘ inculcating false beliefs to target individuals and blocking access to other sources of testimony through shame, guilt, requests or promises of secrecy; • eroding self-trust: ∘ denying through testimony or otherwise that one’s judgments are veridical/can be trusted (including expressions of dismay and disappointment, as well as outright denial); • impeding reasoning: ∘ undermining a target individual’s ability to draw certain inferences, to reason according to the available evidence, to acquire the necessary evidence to reason healthfully, to possess self-trust, or to confidently hold or express certain beliefs as a result of the epistemic fog arising from all of the above. An essential and distinctively epistemic thread runs through each of these cases. Kyle Stephens suffers a loss of knowledge—she knows as a child what has been done to her, but loses that knowledge until later in life, reacquiring her belief only with a superabundance of evidence. Rachel Denhollander’s self-trust, her ability to properly adjudicate her evidence, is undermined by the contravention of norms which she relies upon in interpreting her experience. When Professor suggests co-authoring, Professor is cultivating trust and influence with Student, elevating Professor’s perceived epistemic authority and encouraging practical and intellectual deference. When Professor suggests wine with dinner, Professor suggests that the norms of their relationship are compatible with this. And when Professor teases student, Professor further suggests that Student is unfamiliar with the appropriate norms, calling Student’s judgment into question, and
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
136 Lauren LeydOn-Hardy undermining Student’s self-trust. When Kramer suggests that he and Dushku share the secret of what he has done, Kramer eliminates from Dushku’s epistemic resources the testimonies and evidentiary perspectives of other adults who might have corrected this narrative, among other things. All of these interventions ramify through victims’ epistemic lives: they form more false beliefs, they develop unreasonable evidentiary standards, they trust authorities who are deceiving them. The element of unknowingness bound up in predatory grooming—of being victimized without being able to understand as much—is crucial to understanding the phenomenon. It is crucial to understanding why being subject to these patterns of behavior can be so epistemically destabilizing, and it is crucial to understanding the full breadth of the harm that comes to victims of grooming. So, the epistemology of grooming is crucial to understanding why it works, and how harmful it is. In section 3, I will consider three candidate epistemological concepts and show why none of them are apt to describe the distinctive epistemic harms of predatory grooming. In section 4, I will sketch a novel category of epistemic misconduct that I think gets it right.
3. Grooming as Epistemic Misconduct So far, I have argued that grooming works in part by interrupting a target individual’s ability to marshal her epistemic resources, adjudicate her evidence, and to respond authentically and assuredly to her environment and to what she knows. Grooming works by eroding an individual’s self-trust and judgment. This happens through the cultivation of a dual sense of deference to authority and a distortion of the relevant norms. For these reasons, I argue that grooming ought to be theorized as a form of distinctively epistemic misconduct. I will argue in this section that the current conceptual resources of applied epistemology are ill fitted to the aim of capturing that cultivated sense of unknowing that is at the heart of predatory grooming. I will give a few quick arguments for why some plausible candidates do not seem quite applicable (testimonial injustice,30 hermeneutical injustice,31 and epistemic violence32), before sketching an outline of the kind of resource I think gets at what is distinctive of this type of epistemic misconduct. Miranda Fricker describes testimonial injustice as the attribution of an undue credibility deficit to a speaker on the basis of identity-prejudicial attitudes on the hearer’s part. For example, when a victim of sexual violence is not believed due to stigmas about women and girls lying about regretted sex, she is also a victim of testimonial injustice. She receives a credibility deficit (a disbelieving response)
30 Fricker (2007).
31 Ibid.
32 Dotson (2011).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Predatory Grooming and Epistemic Infringement 137 motivated by an identity-prejudicial attitude (women and girls lie about sex). Predatory grooming involves asymmetries of power, and these plausibly track credibility distributions. It is unclear, however, that credibility distributions in grooming relationships are ultimately grounded in identity-prejudicial attitudes. Moreover, the identity-prejudicial grounding of instances of testimonial injustice is central to Fricker’s characterization of the phenomena. Philosophers have debated the scope of the relevant credibility distributions in these cases,33 however, it is nevertheless my sense that the received view of testimonial injustice holds that the phenomenon is importantly connected to identity-prejudicial attitudes. That just isn’t so in cases of grooming, where the impetus for misconduct is broadly understood to be opportunism rather than hostility, or perhaps even a desire to harm targeted individuals.34 It is psychologically plausible, for example, that some groomers understand their own motivations in terms of something like true love or devotion. More generally, though, grooming involves a great deal more than a misattribution of credibility, epistemically speaking, irrespective of any underlying explanation or motivation. Misleading testimony and counter-testimony, distorting norms, constraining evidence, eroding self-trust, and impeding reasoning—predatory grooming involves a host of complicated and dynamic epistemic interventions that are unrelated to the distribution of credibility. What about hermeneutical injustice? Here is Ishani Maitra’s characterization of this concept from her discussion of Fricker’s work: [Hermeneutical injustice] renders an agent unable to articulate aspects of her social experience even to herself, much less communicate them to others. Roughly speaking, someone suffers the latter kind of injustice, a hermeneutical injustice, when, as a result of (structural) identity prejudice, the hermeneutical resources available are too impoverished for her to articulate (to herself and to others) certain aspects of her social experience. (Maitra 2010, 207, emphasis added)
Right away we can see that, as with the characterization of testimonial injustice, hermeneutical injustice is fundamentally grounded in identity prejudice which, again, does not seem particularly relevant to grooming as a phenomenon. But, 33 Jennifer Lackey has argued that this picture of the relevant credibility distributions will miss paradigm examples of testimonial injustice, including cases that arise out of credibility excesses as well as cases involving the appropriate attribution of speaker credibility, where the asymmetry arises out of inflated self-attribution on the part of the hearer (cf. “Credibility and the Distribution of Epistemic Goods” in Kevin McCain (ed.), Believing in Accordance With the Evidence (2018) and “False Confessions and Testimonial Injustice,” (2020)). 34 A groomer might target a victim that they perceive as vulnerable to testimonial injustice (this may be useful for security management, among other things). However, as with other categories of epistemic misconduct, grooming might be facilitated by, but is not equivalent to, testimonial injustice.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
138 Lauren LeydOn-Hardy perhaps a deeper, more interesting point of divergence has to do with the particular kind of unknowing at work in grooming as compared to the unknowing of hermeneutical injustice. Hermeneutical injustices involve discursive lacunae, which emerge from a fundamentally unjust distribution of conceptual and discursive resources, rendering one incapable of articulating his/her own experiences either to others or, perhaps, to themselves. In grooming, the operant sense of unknowing is one that, as we have seen, is cultivated on behalf of the groomer. In cases of hermeneutical injustice, the experiences are lived absent the conceptual resources to describe what is, or has, happened. But concepts like ‘sexual assault’ or ‘inappropriate touching’ were not out of reach for Larry Nassar’s victims; indeed, those concepts are operant in Racheal Denhollander’s testimony of having reverse reasoned herself out of applying that very description to her own experience. What this shows is that the kind of unknowing at work in grooming is fundamentally dissimilar to the unknowing we’ve seen in cases of hermeneutical injustice. In grooming, the unknowing cannot be explained by any gap in the hermeneutical or conceptual resources (grounded in identity p rejudices or otherwise).35 Rather, it stems from a cultivated misapprehension of the applicability of those conceptual resources to one’s own circumstances. Finally, Kristie Dotson’s 2011 paper “Tracking Epistemic Violence, Tracking Practices of Silencing” furnishes us with the concept of epistemic violence. Epistemic violence is a tool for demarcating the silencing of marginalized groups. It is a concept grounded in considerations of testimony and credibility. Specifically, Epistemic violence in testimony is a refusal, intentional or unintentional, of an audience to communicatively reciprocate a linguistic exchange owing to pernicious ignorance. Pernicious ignorance should be understood to refer to any reliable ignorance that, in a given context, harms another person (or set of persons). Reliable ignorance is ignorance that is consistent or follows from a predictable epistemic gap in cognitive resources. (Dotson 2011, 238)
Dotson is concerned to account for the ways in which a knower might be harmed, owing to the speakers’ dependence on hearer reciprocity in linguistic exchanges. People need to be heard, and to be heard, speakers need uptake from their audiences. To elicit uptake, speakers need to receive their due credibility 35 Might the unavailability of the concept of grooming and the mechanisms of grooming contribute to the unknowing, in a way that the unknowing endemic to grooming is explained by that hermeneutical gap? Certainly, one imagines that anti-grooming efforts would focus, at least in part, on helping individuals and communities to recognize the signs of grooming, common grooming tactics, and so forth. This lends some support to the thought that perhaps some conceptual lacunae are, minimally, beneficial for predatory grooming. Nevertheless, and as with testimonial injustice, I think it is implausible that the full breadth of epistemic interventions involved in predatory grooming can be fully explained just in terms of hermeneutical injustice.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Predatory Grooming and Epistemic Infringement 139 assessment, which is to be treated in a certain manner by their audience. Sometimes, this may not happen because of identity prejudices (which Dotson calls “testimonial quieting,” and which we have seen a parallel discussion of in Fricker’s account of testimonial injustice); other times this may not happen because a speaker has reason to know, antecedently, that she will not receive uptake. This latter phenomenon Dotson calls testimonial smothering. In either case, Dotson argues that speakers or would-be speakers find themselves victims of epistemic violence. The question for our purposes is whether grooming is plausibly explained by a failure to communicatively reciprocate owing to pernicious ignorance. I don’t think this is the right analysis for grooming. It is true that there is (or at any rate, there may be) an element of an intentional refusal to communicatively reciprocate in linguistic exchanges between groomer and groomee, at least sometimes, as in the case where Nassar refused to acknowledge the truth when Kyle Stephens first spoke out. It is harder still, though, to explain that behavior as owing to pernicious ignorance. Groomers may not have reflective access to the grooming campaign that they are engaged in, but certainly they can and often do; indeed, that Nassar knew what he was doing seems essential to explaining his response to Stephens and her parents. Furthermore, in cases where there is a lack of reflective knowledge on behalf of the groomer—as might be the case in Professor/Student co-author—Dotson specifies that reliable ignorance “follows from a predictable epistemic gap in cognitive resources.” But the idea that in this case Professor is animated by a “predictable gap in cognitive resources” doesn’t seem right. It is precisely because Student has the reasonable expectation that Professor knows the appropriate norms that this kind of interaction works as a grooming scenario. So, even if Professor is to some extent in the dark about what he’s up to, this mental lapse is not owed to a predictable gap in Professor’s cognitive resources; to the extent that Professor is successful in grooming Student, it is precisely because Student is unaware of any such gap. In general, then, it seems that pernicious ignorance is an ill-fitting mechanism to posit at the explanatory heart of the epistemic misconduct undergirding predatory grooming. It matters, too, for our purposes, that the conceptual epistemic resources Dotson furnishes us with are characterized from the perspective of the perpetrator. This is a bedrock asymmetry between our projects. In section 1 of this chapter, I suggested that there might be a two-way street here between forensic psychology and epistemology: forensic psychology can provide philosophy with a conceptually rich model of the mechanisms of grooming. This gives us a way of thinking about what groomers are up to, what their patterns of behavior might look like. What I am claiming is that epistemologists have much to offer by way of enriching that model, to think about not just the phenomenon of grooming as a locus of epistemic misconduct, but further about the harms of predatory grooming. If, as social epistemologists, we primarily theorize grooming as a phenomenon from the perspective of the groomer, we lose out on an
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
140 Lauren LeydOn-Hardy important part of what we can bring to the conversation around predatory grooming: an account of how the relevant norms operate within the grooming relationship, in a way that makes grooming behavior particularly effective and, as such, harmful in a distinctly epistemic way. I’ll say more about this in section 4.
4. Epistemic Infringement I have been arguing that predatory grooming is importantly characterized by epistemic misconduct. I also think that the kind of epistemic misconduct that we see in grooming relationships can be seen in other social phenomena. I have just argued that several familiar epistemological concepts are ill suited to describe what’s going on in the cases I am interested in here. In this final section, I will first sketch a general epistemological analysis of the type of misconduct that we have seen in grooming cases. I will then elaborate on that concept by pointing to it at work in some other contexts—specifically, gaslighting and propaganda. Here’s a schema for the category of epistemic misconduct that I have in mind: Epistemic Infringement: To epistemically infringe on S is to systematically contravene the interpersonal social and epistemic norms that S takes to constrain her relationship to the infringer, in a manner that may encroach upon or undermine S’s epistemic agency.
On this understanding of epistemic infringement, it involves more than simply telling an untruth or assigning a localized credibility deficit. Those might contravene interpersonal social and epistemic norms, and even do so in a way that incurs significant and negative epistemic consequences for S, but they do not (at least in the typical situation) affect S’s epistemic agency itself. Affecting a person’s epistemic agency typically requires complex and temporally extended projects of deceit, misdirection, manipulation, and coercion. In cases of predatory grooming, for example, we see the slow erosion of a sense of self that is fundamentally seated in an individual’s epistemic life. To subvert an individual’s confidence in her own judgment, to make it such that she can no longer trust her own mind, one needs to do more than merely raise the specter of doubt—a part of her mental character needs to be damaged. In the foregoing discussion of grooming, we saw how this can be done by thoroughly subverting the attendant norms according to which individuals manage their epistemic resources, relative to a particular relationship. This is the sense in which I want to say that the core concept of epistemic infringement is the thought that this kind of abuse strikes at the epistemic agency of its victims. The notion of a distinct kind of epistemic agency is not uncontroversial. You might think (some do) that it just follows from the fact that doxastic voluntarism
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Predatory Grooming and Epistemic Infringement 141 is false that there is no tenable, or substantive account of epistemic agency. Considerations of this sort sometimes animate the thought that when it comes to epistemic agency, either there is no such thing, or whatever it is, it’s probably not very interesting; either your analysis of epistemic agency is thick in a way that makes it implausible, or thin in a way that makes it unilluminating. This is because avoiding doxastic voluntarism will involve stipulating a fairly weak connection between what one can choose and what one believes. Suitably weakened, it looks like the most substantive account of that relationship will amount to, say, the choice to set one’s gaze over here, rather than there (and so to exert some agency in one’s perceptual beliefs) or to read this rather than that (and so to choose which things about which to form beliefs). People do go in for this weaker kind of account—this is roughly what Pamela Hieronymi characterizes as “indirect” or “managerial” control, the weaker of two conceptions of epistemic agency that she defends.36 But many have suggested that if this is all that epistemic agency amounts to, the account, though perhaps uncontroversial, is not particularly substantive.37 So, a general theme in this literature seems to be that either one posits a controversially robust analysis of epistemic agency, in which case it’s interesting but deeply contested, or one posits a rather thin account, in which case it’s plausible, though not altogether interesting. I am less impressed than others by the move to dismiss a metaphysically thin conception of epistemic agency as philosophically light, but it’s beyond the scope of this chapter to give a full-throated defense of that picture. For present purposes, a flatfooted characterization of the notion of epistemic agency, one that just understands it in terms of the ability to marshal one’s epistemic resources with some healthful degree of autonomy, can help to explain the distinctly epistemic harms at work in predatory grooming, as well as in extreme cases of gaslighting, and in socio-structural examples of epistemic infringement such as propaganda. And even if you are the kind of theorist who goes in for something a little thicker than that, given that most accounts of epistemic agency incorporate something like reasons-responsiveness for their analysis, that conception of epistemic agency will be directly relevant to my account of epistemic infringement, given that infringement functions largely by damaging precisely that capacity to have one’s beliefs appropriately reasons-responsive.38 My argument here is just that your 36 Hieronymi (2009b). 37 Cf. Engel (2009), where he argues that ‘indirect voluntarism’ is inadequate to the task of countenancing a genuinely epistemic kind of agency, McHugh (2014), where he argues that cases of indirect control will not count as truly paradigm examples of regulating our doxastic states, or Setiya’s (2013) response to Heironymi’s (2011) where he argues that “If there is an interesting thesis of epistemic agency, it must go beyond the fact of belief revision” (183). 38 For a (non-exhaustive) survey of reasons-responsive accounts of epistemic agency, see Reed (2001, 2013, 2016a, 2016b); Hieronymi (2008, 2009a, 2009b); Marušic (2012); McHugh (2013); Peels (2015); and McCormick (2015). Engel (2009) and Setiya (2013) both defend the view that we can have doxastic responsibility (or be reasons responsive) without epistemic agency.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
142 Lauren LeydOn-Hardy preferred theory of epistemic agency shouldn’t be a hurdle for taking the foregoing analysis of epistemic infringement on board. Epistemic infringement is a kind of epistemic misconduct which works by contravening interpersonal and epistemic norms in a way that erodes or undermines an individual’s epistemic agency (i.e. her ability to manage her epistemic resources, adjudicate her evidence, appropriately interpret her circumstances, etc.). With that said, I want to briefly take a look at a couple of other examples that I think are properly described as epistemic infringement. I’ll start with gaslighting. Kate Abramson’s essential paper “Turning Up the Lights on Gaslighting” offers this introduction to the concept: [T]he phenomenon that’s come to be picked out with that term is a form of emotional manipulation in which the gaslighter tries (consciously or not) to induce in someone the sense that her reactions, perceptions, memories and/or beliefs are not just mistaken, but utterly without grounds—paradigmatically, so unfounded as to qualify as crazy. Gaslighting is, even at this level, quite unlike merely dismissing someone, for dismissal simply fails to take another seriously as an interlocutor, whereas gaslighting is aimed at getting another not to take herself seriously as an interlocutor. (Abramson 2014, 2, emphasis added)
There is an important difference that emerges here between gaslighting and other acts of silencing. In testimonial smothering, for example, speakers withhold their testimony not because they doubt their judgment, but because they have reason to believe that their testimony will not receive appropriate uptake. That is, because they doubt their audience’s judgment. In canonical examples of testimonial injustice, the harm happens at the locus of a credibility deficit; the prejudicial credibility assignment is itself—if not entirely, then at least significantly—the harm. For victims of testimonial injustice, then, we needn’t assume anything about the downstream effects of that kind of behavior on one’s overall epistemic well- being. And in neither case is the success of the particular silencing mechanism dependent on something so penetrating as gaslighting, which works by cultivating an internalized sense of mistrust. Gaslighting has this in common with predatory grooming. This suggests how the concept of epistemic infringement can work to supplement Abramson’s account of gaslighting. I’ve argued that this kind of phenomenon is crucially weaponized via the perversion of the norms governing certain kinds of relationships. We can see this very dynamic at work in Abramson’s central examples of gaslighting. For example, in her discussion of an engaged couple, Pat and Collier, Abramson argues that Collier’s manipulative dismissal of Pat’s professional aspirations rises to the level of gaslighting. In this case, it is not that Collier is simply telling Pat that he doesn’t think she has what it takes to succeed, but rather, Collier is worried about Pat; worried that her aspirations will lead to her disappointment. The implication is
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Predatory Grooming and Epistemic Infringement 143 that if Pat thinks she’s got what it takes, then she ought to think again, because this person who loves her is suggesting otherwise. Abramson writes, “It’s recognizable as gaslighting because Collier is so clearly trying to radically undermine Pat, and doing so through a combination of dismissing (e.g. ‘oh what’s the good of that, I mean after all’) and manipulating her (esp via her trust and love).”39 Abramson doesn’t invoke the language of norms, but I think it helps to make clear not just how these phenomena are deployed, but why they get a grip on us. Outside-looking-in, we can see what Collier is up to, but inside-looking-out it is also easy to see why Pat might be so marooned by this kind of maneuver. After all, Collier is her fiancée. When our closest people express concern, it is precisely because they are close to us that we are all the more inclined to take their concern seriously, to interpret their concern as well intentioned and well informed; that is, we invoke the norms of closeness in order to interpret their behavior. And so, the very norms of love and intimacy that Collier transgresses when he puts his ego ahead of Pat’s ambition become the norms to which she can appeal to assuage her creeping sense that something is wrong. We saw this same pattern in our discussion of predatory grooming, where the norms of teacher/student, medical professional/athlete, and so on—the very norms transgressed in the abusive accounts discussed above—were also crucially at work in obfuscating that abuse both to third parties and, critically, to the victims. So finally, and just briefly, I’d like to suggest that we can see this same pattern at work in (at least some) canonical instances of propaganda. For instance, in How Propaganda Works, Jason Stanley begins the chapter “Propaganda Defined” (emphasis my own) with a discussion of the Report on the Governability of Democracies to the Trilateral Commission (1975). That report argued for cultivating: epistemic authority, in the form of “experts,” as a means to instill practical authority over the masses. [. . .] to employ the vocabulary of scientific expertise in a political way, in effect using epistemic ideals as forms of coercion. Such mixtures of epistemic and practical authority, where epistemic authority is used to gain practical authority over the domain of democratically autonomous decision, tend to undermine the epistemic ideals. It leads to distrust of those who self-present as “scientific experts,” even when they want to warn us about the importance of vaccines or climate change. (Stanley 2015, 40, emphasis added)
Stanley is setting up this chapter with what he regards as an essential example of propaganda. And while this example is framed in terms of epistemic ‘ideals’ in a democratic society, it seems to me that we could swap in ‘healthful norms’ salva
39 Abramson (2014, 7).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
144 Lauren LeydOn-Hardy veritate. What I want to say here is just that this propagandistic strategy of appealing to the relevant epistemic norms (in this case, of science and expertise) in order to use those very norms for coercive, and epistemically corroding ends, is squarely in line with my characterization of epistemic infringement, scaled to the socio-structural order.40
5. Concluding Remarks The unifying theme of these social phenomena—grooming, gaslighting, and propaganda—is epistemic infringement. In each case, we see a pattern whereby the social and epistemological norms that ought to govern a particular relationship are surreptitiously transgressed with the effect of undermining epistemic agency. In these cases, the very norms being transgressed work to dampen the suspicion that something is amiss. One upshot of this is that, in cases of epistemic infringement, even from the perspective of the abused, it is often difficult to identify abuse as such. In general, these phenomena work by making it hard for their victims to acquire the knowledge they need to flourish within, and in light of, those relationships. My objectives in this chapter were essentially twofold: first, to bring the concept of predatory grooming to the table for philosophers; and second, to draw a line between what I think is the right epistemological analysis of predatory grooming and some other social phenomena of interest to philosophers. I hope this discussion proves useful in helping applied epistemologists make meaningful distinctions among these matters of interest, but also, that an analysis of predatory grooming through the normative lens of epistemology might find broader application for scholars thinking about, not just the mechanisms, but the harms of these kinds of behavior. The idea that they work precisely because they appeal to (and distort) what we take ourselves to know about the normative world around us strikes me as a vital component for understanding these phenomena, generally.
References Abramson, Kate. (2014). “Turning Up the Lights on Gaslighting.” Philosophical Perspectives 28(1): 1–30. Ashurst, Libby and Anne-Marie McAlinden (2015). “Young People, Peer-to-Peer Grooming and Sexual Offending: Understanding and Responding to Harmful Sexual Behaviour within a Social Media Society.” Probation Journal 62(4): 374–88. 40 For an account of collective epistemic agency see Leydon-Hardy (ms.)
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Predatory Grooming and Epistemic Infringement 145 Barajas, Joshua. (2018). “Sexual Abuse Survivors Confront Former USA Gymnastics Doctor: ‘Little Girls Don’t Stay Little Girls Forever’.” PBS, Public Broadcasting Service (January 16, 2018), www.pbs.org/newshour/nation/sexual-abuse-survivorsconfront-former-usa-gymnastics-doctor-little-girls-dont-stay-little-girls-forever. Brackenridge, Celia. (1997). “ ‘He Owned Me Basically . . . ’ Women’s Experience of Sexual Abuse in Sport.” International Review for the Sociology of Sport 32(2): 115–30. Brackenridge, Celia, and Kari Fasting. (2005). “The Grooming Process in Sport: Narratives of Sexual Harassment and Abuse.” Auto/Biography 13.1: 33–52. Brandl, Bonnie, Candace J. Heisler, and Lori A. Stiegel. (2006). “The Parallels between Undue Influence, Domestic Violence, Stalking, and Sexual Assault.” Journal of Elder Abuse & Neglect 17(3): 37–52. Callimachi, Rukmini, host. (2018). “Caliphate Chapter Nine: Prisoners, Part 2.” New York Times, 14 June 2018. caliphate.simplecast.com/episodes/chapter-nine-prisonerspart-two Craven, Samantha, Sarah Brown, and Elizabeth Gilchrist. (2006). “Sexual Grooming of Children: Review of Literature and Theoretical Considerations.” Journal of Sexual Aggression 12(3): 287–99. Darwall, Stephen. (2006). The Second-Person Standpoint: Morality, Respect, and Accountability. Cambridge, MA: Harvard University Press. Denhollander, Rachel. (2018). “Read Rachael Denhollander’s Full Victim Impact Statement about Larry Nassar.” CNN, Cable News Network (January 30, 2018), www.cnn.com/2018/01/24/us/rachael-denhollander-full-statement/index.html. Dotson, Kristie. (2011). “Tracking Epistemic Violence, Tracking Practices of Silencing.” Hypatia 26(2): 236–57. Dushku, Eliza. (2018). “Eliza Dushku Statement.” Facebook (January 13, 2018), www. facebook.com/OfficialElizaDushku/posts/1769957739689557. Elliott, Ian. (2017). “A Self-Regulation Model of Sexual Grooming.” Trauma, Violence, & Abuse 18(1): 83–97. Engel, Pascal. (2009). “Epistemic Responsibility without Epistemic Agency.” Philosophical Explorations 12(2): 205–19. Fricker, Miranda. (2007). Epistemic Injustice: Power and the Ethics of Knowing. Oxford: Oxford University Press. Gladwell, Malcolm. (2012). “In Plain View: How Child Molesters Get Away with It.” The New Yorker, 84–90. Hagopian, Amy and Kathy Barker (2011). “Should We End Military Recruiting in High Schools as a Matter of Child Protection and Public Health?” American Journal of Public Health 101(1): 19–23. Haring Hidore, Marilyn and Michelle A. Paludi (1987). “Sexuality and Sex in Mentoring and Tutoring: Implications for Women’s Opportunities and Achievement.” Peabody Journal of Education 64(4): 164–72.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
146 Lauren LeydOn-Hardy Hieronymi, Pamela (2008). “Responsibility for Believing.” Synthese 161(3): 357–73. Hieronymi, Pamela (2009a). “Believing at Will.” Canadian Journal of Philosophy 39(1): 149–87. Hieronymi, Pamela (2009b). “Two Kinds of Agency.” Mental Actions: 138–62. Hieronymi, Pamela (2011). “Believing at Will,” in D. Hunter (ed.), Belief and Agency. Calgary: University of Calgary Press, 149–87. Hopkins, Lauren. “Opinion|Larry Nassar Was a Master Manipulator, But He Didn’t Act Alone.” The Huffington Post (January 23, 2018), www.huffingtonpost.com/ entry/opinion-hopkins-gymnastics-nassar_us_5a673cfce4b0e5630073862f. Kloess, Juliane, Catherine E. Hamilton-Giachritsis, Anthony R. Beech. (2017). “Offense Processes of Online Sexual Grooming and Abuse of Children via Internet Communication Platforms.” Sexual Abuse: 1079063217720927. Knoll, James. (2010). “Teacher Sexual Misconduct: Grooming Patterns and Female Offenders.” Journal of Child Sexual Abuse 19(4): 371–86. Lackey, Jennifer (2018). “Credibility and the Distribution of Epistemic Goods,” in Kevin McCain (ed.), Believing in Accordance with the Evidence: New Essays on Evidentialism. New York: Springer Publishing. Lackey, Jennifer (2020). “False Confessions and Testimonial Injustice.” The Journal of Criminal Law and Criminology (1973-) 110(1): 43–68. Levinson, Stephen C. (2006). “On the Human ‘Interactional Engine’, in Nicholas Enfield and Stephen Levinson (eds), Roots of Human Sociality: Culture, Cognition and Interaction. New York: Berg. Maitra, Ishani. (2010). “The Nature of Epistemic Injustice.” Philosophical Books 51(4): 195–211. Marušić, Berislav. (2012). “Belief and Difficult Action.” Philosopher’s Imprint 12(18): 1–30. McAlinden, Anne-Marie. (2006). “‘Setting ’Em Up: Personal, Familial and Institutional Grooming in the Sexual Abuse of Children.” Social & Legal Studies 15(3): 339–62. McCormick, Miriam Schleifer. (2015). Believing Against the Evidence: Agency and the Ethics of Belief. London: Routledge. McHugh, Conor. (2013). “Epistemic Responsibility and Doxastic Agency.” Philosophical Issues 23(1): 132–57. Olson, Loreen N., Joy L. Daggs, Barbara L. Ellevold, and Teddy K. K. Rogers. (2007). “Entrapping the Innocent: Towards a Theory of Child Sexual Predators’ Luring Communication.” Communication Theory 17(3): 231–51. Peels, Rik. (2015). “Believing at Will is Possible.” Australasian Journal of Philosophy 93(3): 524–41. Reed, Baron. (2001). “Epistemic Agency and the Intellectual Virtues.” The Southern Journal of Philosophy 39(4): 507–26.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Predatory Grooming and Epistemic Infringement 147 Reed, Baron. (2013). “Fallibilism, Epistemic Possibility, and Epistemic Agency.” Philosophical Issues 23(1): 40–69. Reed, Baron. (2016a). “Having to Do with Knowledge.” Episteme 13(4): 549–54. Reed, Baron. (2016b). “Who Knows?” in Vargas, Miguel Ángel Fernández (ed.), Performance Epistemology: Foundations and Applications. Oxford: Oxford University Press. Reed, Dan. (2019). Leaving Neverland. Video file, www.hbo.com/documentaries/ leaving-neverland. Setiya, Kieran. (2013). “Epistemic Agency: Some Doubts.” Philosophical Issues 23(1): 179–98. Stanley, Jason. (2015). How Propaganda Works. Princeton, NJ: Princeton University Press. Superson, Anita. (2005). “Deformed Desires and Informed Desire Tests.” Hypatia 20(4): 109–26. Volkwein-Caplan, Karin, Frauke Schnell, Shannon Devlin, et al. (2002). “Sexual Harassment of Women in Athletics vs Academia.” Journal of Sexual Aggression 8(2): 69–82.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
PART 4
EPIST E MOLO GY A N D I NJU ST IC E
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 19/03/21, SPi
7
Epistemic Degradation and Testimonial Injustice Geoff Pynn*
It is obvious that you shouldn’t discount or reject someone’s testimony on the basis of a prejudice about them. Indeed, it is obvious that to do so is to wrong the speaker. But what is the nature of this wrong? This question, it turns out, is not that easy to answer. In this chapter, I will propose that we should think of the wrong of what Miranda Fricker calls testimonial injustice as a kind of degradation. The chapter offers a sketch, not an argument. Other accounts, some of which I will discuss here, are also illuminating. But the degradation account, I think, helps to bring out better than those other accounts the way in which testimonial injustice is a kind of insult to the dignity of the speaker. It may also provide some insight into our obligations to speakers. If the account is on the right track, we do something wrong and potentially harmful whenever we discount the testimony of a knowledgeable speaker. When we do so on the basis of a negative identity preju dice, we compound this wrong by degrading the speaker in a particularly dam aging way. Testimonial injustice, as characterized by Miranda Fricker, involves a speaker suffering what she calls a “credibility deficit” as the result of the operation of a negative identity prejudice on the hearer. A speaker who suffers a credibility def icit owing to some factor is perceived or assessed as having lower credibility than they otherwise would were that factor not present. The central cases of testimo nial injustice Fricker discusses involve speakers who are treated as less credible in virtue of the influence of prejudicial stereotypes concerning their race and gender on their audiences. In these cases, the credibility deficits they suffer result in their knowledgeable assertions being rejected by their audiences. The cases are familiar to anyone with a knowledge of the literature that has sprung up in response to
* Thanks to Chris Copan and an anonymous reader for Oxford University Press for very helpful comments. Thanks also to Jennifer Lackey and Lauren Leydon-Hardy for inviting me to present this work at the 2017 Central American Psychological Association meeting, and to my co-panelists, Ishani Maitra and Veronica Ivy, for their encouragement and helpful questions.
Geoff Pynn, Epistemic Degradation and Testimonial Injustice In: Applied Epistemology. Edited by: Jennifer Lackey, Oxford University Press (2021). © Geoff Pynn. DOI: 10.1093/oso/9780198833659.003.0007
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
152 Geoff Pynn Fricker’s book, but they are worth rehearsing, since they will play an important role in the discussion that follows.1 The first case involves Tom Robinson, one of the main characters in Harper Lee’s To Kill A Mockingbird. The story takes place in 1930s Alabama. Tom, a lowstatus black man, is falsely accused of raping a higher-status white woman. He gives accurate exculpatory testimony, but the racist jury, partly in virtue of their white supremacist prejudices, rejects it. Tom is wrongly convicted and killed while trying to escape. The second case involves Marge Sherwood, a character in the Anthony Minghella adaptation of Patricia Highsmith’s The Talented Mr Ripley. Sherwood suspects—accurately and on the basis of good evidence, as viewers know—that her fiancé, Dickie Greenleaf, was murdered by their friend, Tom Ripley. But Dickie’s father Herbert rejects her testimony to this effect, suggesting that she is speaking out of “female intuition” rather than conveying knowledge of the “facts”. His credibility judgment is shaped by misogynistic ideology. Tom escapes detection, and Marge suffers significant psychological distress, ultimately falling into a tragically ironic hysteria. As Fricker puts it, “We see poor Marge in some measure actually becoming what she has been constructed as: a hysterical female, apparently expressing herself in semi-contradictions, sticking to intu itions in the face of others’ reasons, unable to keep a grip on her emotions” (Fricker 2007, 88). These are central cases of testimonial injustice for Fricker because the negative identity prejudices involved in their production are systematic: they “track the subject through different dimensions of social activity—economic, educational, professional, sexual, legal, political, religious, and so on” (Fricker 2007, 27). Such examples deserve the most attention, both because of the grievous harms they involve and because of the part they play in sustaining broader social structures of exclusion and domination. Nonetheless, one-off instances of testimonial injust ice are possible, and need not involve such spectacular harms as those suffered by Tom and Marge. Fricker points out that there are other, much more localized instances of testimonial injustice that involve prejudices unrelated to such broad and relatively fundamental social identity categories as race and gender. For example, a panel of referees on a science journal might be led by a dogmatic prejudice against a certain research method to reject a submission that employs that method (Fricker 2007, 28–9). The spectacular cases are central instances of a kind that also occurs in much less spectacular form. Though it is easy enough to see that the credibility deficits suffered by Tom and Marge are cases of injustice, it can be difficult to characterize with precision what 1 In recent years, a number of writers have suggested that credibility deficits are not essential to testimonial injustice; see, e.g. Davis (2016). I think there is a way to extend the analysis I develop here to such cases, but I’m not especially confident about that. For the purposes of this chapter, I’ll assume that testimonial injustice always involves a credibility deficit. Like all other accounts of the phenom ena, mine will no doubt be only partial.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
Epistemic Degradation and Testimonial Injustice 153 makes them cases of injustice. One difficulty is just that there are many different worthy targets of moral and epistemic criticism in these cases. There is a noxious prejudice, a vile stereotype, and an oppressive ideology, each of which is defective along several different dimensions. There are multiple moral and epistemic fail ings on the part of the listeners. There are the particular grievous harms caused to each speaker by the rejections of their assertions. There are unjust social hierarch ies whose structures are reinforced through the exclusion of each speaker. Writers looking for wrongs to describe in these cases are faced with an embarrassment of riches. It may be that there is no distinctive injustice, harm, or wrong shared by all instances of testimonial injustice; rather, the category may best be construed as a family of more specific wrongs, each of which (somehow) involves the (some how) inappropriate operation of a (somehow) negative identity prejudice in the assessment of a speaker’s credibility. A pluralist account of the wrong of testimo nial injustice is tempting. Nonetheless, there does seem to be something normatively in common between Tom’s case, Marge’s case, and analogous cases all of us can identify from contemporary life and our own experiences. Though it may not be present in all cases of testimonial injustice, it is worth trying to articulate the normative defect apparently common in the central cases. Fricker says that the “primary harm” of testimonial injustice is that the “subject is wronged in her capacity as a giver of knowledge” (Fricker 2007, 44).2 This capacity is, or at least is conceived as, an essential dimension of human rationality, and in turn, of human value. Fricker describes the harm as an “insult”. She also suggests that it leads to a kind of degradation: The fact that the primary injustice involves insult to someone in respect of a capacity essential to human value lends even its least harmful instances a sym bolic power that adds a layer of harm of its own: the epistemic wrong bears a social meaning to the effect that the subject is less than fully human. When someone suffers a testimonial injustice, they are degraded qua knower, and they are symbolically degraded qua human. In all cases of testimonial injustice, what the person suffers from is not simply the epistemic wrong itself, but also the meaning of being treated like that. Such a dehumanizing meaning, especially if
2 Fricker tends to move fairly loosely between “injustice,” “wrong,” and “harm”. There are important differences between these concepts. One can commit an injustice without wronging anyone in par ticular (perhaps cheating on your taxes is like this); one can wrong someone without harming them (imagine betraying a friend’s confidence in a way that never causes them harm); and one can harm someone in a way that is neither unjust nor wrongful (it is generally assumed that state-sanctioned punishments can be like this). I think that this looseness has contributed to some of the disagreements in the literature on epistemic injustice. But being pedantic on this point in the text would be tedious, and I’m not sure anything would be gained. So, I’ll follow Fricker in using “harm,” “wrong,” and “injustice” more or less interchangeably.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
154 Geoff Pynn it is expressed before others, may make for a profound humiliation, even in cir cumstances where the injustice is in other respects fairly minor. (Fricker 2007, 44)
Fricker never returns to the theme of degradation in her book. Later on, she develops an account of the “primary wrong of testimonial injustice” that leans instead on Edward Craig’s genealogical account of the concept of knowledge, together with the concept of objectification. Drawing on the distinction Craig draws between informants and mere sources of information (Craig 1990, sec. V) she proposes that in cases of testimonial injustice, speakers are treated as the lat ter, rather than the former. This treatment, she argues, constitutes a kind of epi stemic objectification analogous to sexual objectification (Fricker 2007, 133ff.). The wrong of testimonial injustice is then interpreted through this Kantianfeminist lens. Fricker’s account is rich, subtle, and insightful. But as Gaile Pohlhaus points out, there is something odd about employing the concept of objectification to characterize the wrong done here. “Objective” sources of information—for example, the rings of a felled tree—present us with evidence we cannot reasonably reject. If we are aware that there are 14 rings on the tree, we cannot reasonably deny that the tree is 14 years old—at least, not without appeal to some further “objective” evidence. If we do, we will be rightly criticized as irrational, ignorant, or both. Clearly, this is not analogous to the situation with the testimony of a socially marginalized speaker. We are often free to reject their testimony without sanction; indeed, we may instead face informal sanction for treating their testi mony as an objective source of information. As Pohlhaus puts it, “It is precisely because Sherwood and Robinson are perceived as subjects that makes it possible for hearers in both cases to avert the kind of claim their testimony ought rightly to make upon hearers and to do so without the hearers risking the charge of irrationality” (Pohlhaus 2014, 104). Pohlhaus proposes instead that we characterize the primary wrong of testimo nial injustice in terms of what she calls the subject’s “derivatization”. A derivatized subject is treated as a subject (hence, not an object), but one whose capacities “exist solely in support of and never in tension with my own” (Pohlhaus 2014, 107). Unlike a full-blown subject, a derivatized subject is understood to be incap able of “contributing to epistemic practices uniquely; that is, from her own dis tinct lived experience in the world.” The primary harm of testimonial injustice, Pohlhaus suggests, is being treated as a derivatized subject. The derivatized sub ject’s epistemic capacities contribute to our common knowledge, but only insofar as they conform to the expectations within the contours of our own subjectivity. The derivatized subject’s capacities are thoroughly subordinated to our own, and hence incapable of prompting us to revise or reflect upon the deliverances of our own subject perspective.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
Epistemic Degradation and Testimonial Injustice 155 Like Fricker’s account, that of Pohlhaus is rich and insightful. But there are aspects of testimonial injustice that aren’t captured by the proposals of either Fricker or Pohlhaus. Consider a case presented by Ishani Maitra. Zara is reading the news online. She comes upon a piece whose author identifies himself as “a committed tea-partier”; she immediately deletes it, “cavalierly dismiss[ing]” the item, judging its author to have low credibility precisely because of her negative associations with tea-partiers. Maitra thinks that “Zara does no injustice to the writer by dismissing him. [. . .] [S]he simply doesn’t owe it to him to avoid a cred ibility deficit” (Maitra 2011, 199). And yet, she thinks, this is a case where the lis tener’s credibility judgment is influenced by a negative identity prejudice. Maitra lists several other cases where she thinks we lack an obligation to avoid making an identity-prejudicial credibility assessment: billboards, radio programs and podcasts, conversations overheard among strangers and acquaintances, and pub lic announcements. Yet in all of these cases, it seems that the speakers are either “objectified” in Fricker’s sense—treated as mere sources of information, rather than informants—or “derivatized” in Pohlhaus’s—treated as if their epistemic contributions were solely in support of, and not in tension with, any of my own capacities.3 So it is hard to see how, on either view, we could avoid counting these as wrongs of the same kind as testimonial injustice. Another question for both characterizations arises from cases where a speaker suffers an identity-prejudicial credibility deficit, but is not, in fact, speaking knowledgeably. Suppose that Tom Robinson had been lying on the stand. The racist jurors would have disbelieved him for the same racist reasons that they did in the original case. They would still be contemptible for assessing his credibility through the lens of a white supremacist ideology. But, while I am not quite sure what to say about this case, and while the case would clearly involve the same kind of rejection that flows from and reinforces a broader system of injustice, it does not seem to involve the same sort of personal wrong as is done to Tom in the original case. Or suppose that Marge was, in fact, in the grip of a fixed delusional system. Replace the good evidence she actually had of Tom’s guilt with figments generated by a feverish imagination. While Greenleaf would still be wrong to assess her testimony on the basis of a misogynistic stereotype depicting women as “hysterical,” it isn’t clear that he would have wronged her in the same way he did 3 Fricker discusses cases where the “hearer is right to judge her interlocutor as epistemically untrustworthy on the matter in hand, so that there is no injustice” (Fricker 2007, 135). One might think this is what is going on in Zara’s case. Of such cases, Fricker says there is some sense in which such a judgment “denies [the speaker’s] epistemic subjectivity” and hence counts as objectification. So, she concludes, some forms of epistemic objectification are ethically unproblematic. The key for Fricker is that in the ethically unproblematic cases, the credibility assessment is not “an instance of any general tendency on the hearer’s part to unfairly downgrade the speaker’s epistemic subjectivity.” But it seems to me that Zara doesn’t wrong the writer, even if her downgrade reflects an unfair ten dency. Except in special circumstances, we simply don’t wrong writers we don’t read, even if our basis for not reading them is unfair.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
156 Geoff Pynn in the original case. One ought not to discount testimony on the basis of prejudi cial stereotypes, but it doesn’t seem that the speakers in these hypothetical cases are entitled to the same sort of resentment as their original counterparts. In both cases, the listeners are open to criticism for their reliance on a negative identity prejudice, and they show disrespect to the speakers by dismissing their testimony on this basis. But neither one is a central case of testimonial injustice. However, in both cases, the speakers are objectified and derivatized in exactly the same way as they are in the central cases. If these differ from the central cases in some norma tively essential way, then treating a testifier as an epistemic object or derivatized subject cannot be the primary wrong of testimonial injustice. I think Fricker’s initial characterization of the wrong of testimonial injustice in terms of degradation can provide us with a more satisfying account. Let’s suppose that the normative unity in the central cases is that they involve epistemic degrad ation. Objectification and derivatization may result from or lead to degradation; they may also be ways of expressing or accomplishing degradation. But they are distinct forms of wrongful treatment. Degradation crucially involves a distinctive representational component. In pointing to the “symbolic power” and “social meaning” of the harm of testimonial injustice, Fricker is alluding to this dimen sion of the wrong. The victim of degradation is violated in a way that diminishes their status. The degraded person is represented as having a low social value rela tive to some salient threshold, typically one that is taken to be met by the degrader and their audience. We can see this diminishment at work in both of the central cases; moreover, the diminishment is accomplished via an assessment of the sub ject’s epistemic capacities. In the next section, I explore the concept of degrad ation more generally, before painting a fuller picture of epistemic degradation. * * * The notion of degradation is flexible and multifaceted. I intend to employ it in the sense in which it was introduced into the sociology and criminology literature by Harold Garfinkel (Garfinkel 1956). Garfinkel is interested in what he calls “status degradation ceremonies,” ritual expressions of moral indignation that function to strengthen group solidarity and commitment to shared norms and values by severely ostracizing a particular subject taken to have violated the group’s norms. A degradation ceremony occurs when “the public identity of an actor is transformed into something looked on as lower in the local scheme of social types” (Garfinkel 1956, 420). In the ceremony, a publicly acknowledged “denouncer” or “degrader,” held to speak on behalf of the “supra-personal values of the tribe,” enacts a ritual that transforms a “perpetrator,” who has transgressed those values, into a lower social type before a crowd of “witnesses”. The end result is that the perpetrator is demoted to a position alien to that occupied by the denouncer and witnesses: “He must be placed ‘outside,’ he must be made ‘strange’ ” (Garfinkel 1956, 423).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
Epistemic Degradation and Testimonial Injustice 157 Garfinkel was concerned with phenomena such as public executions, criminal trials, and military disciplinary proceedings, as well as analogous events in more tribal and kinship-based cultures. But while his focus was on formal ceremonies, many have extended his conceptual apparatus to less formal contexts: for example, a teacher’s scolding a student in front of the class (Bergner 1987), congressional hearings denouncing corporate executives (Cavender et al. 2010), discrediting by one’s peers in scientific research (Thérèse and Martin 2010), and spontaneous, violent demonstrations in reaction to news of Benedict Arnold’s attempted trea son (Ducharme and Fine 1995). My interest is not in the nature of a (successful or unsuccessful) degradation ceremony per se, but rather the concept of degradation implicit in Garfinkel’s discussion and those inspired by it. As I see it, there are three important ingredients of degradation in the relevant sense. First, degrad ation can only occur against a background hierarchy of identity types, where those who occupy a higher position in the hierarchy possess entitlements that those in a lower position don’t. The successfully degraded subject comes to occupy a lower position in the hierarchy than the degrader, and thereby lacks entitlements possessed by those in higher positions. Second, degradation is intrinsically public. A subject is successfully degraded only when their diminished social position is somehow made known to the broader community. Degradation in this sense cannot be a purely inward affair; shame is the private correlate of degradation. And third, degradation is performative. It involves treating the sub ject in a way that accords with and expresses their diminished status. To merely pronounce sentence upon a convicted criminal is not yet to degrade them. Their degradation is accomplished only once they have actually had their liberty revoked, which is expressed publicly through their being shackled, dressed in prison garb, and placed inside a cage. The natural home for the concept of degradation is a social environment struc tured around a hierarchy of ranks. In such an environment, one’s rank largely determines the treatment to which one is entitled. Perhaps more importantly, in such an environment, one’s basic social identity is constituted by one’s social pos ition. Degradation in such a context destroys the subject’s identity by stripping them of their rank. (Garfinkel says that in a successful degradation ceremony, a subject’s “total identity” is transformed.) Consider the example of Titus Oates, an English clergyman whose false testimony about the existence of a so-called “Popish Plot” against the life of King Charles II in 1678 resulted in, among other things, the wrongful execution of a number of Catholics. After the succession of James II in 1685, Oates was tried and convicted of perjury. In addition to a large fine, life imprisonment, and defrocking (a literal “de-grading” in a narrow sense, and one that caused much controversy among contemporary commentators, as it was not clear that it was within the court’s power to defrock), he was sentenced to be whipped through the streets of London and placed in the pillory three times a year for the rest of his life. In issuing its sentence, the court stated that the
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
158 Geoff Pynn purpose of Oates’s punishment was “that he should forever be Degraded,” ensur ing that he was never able to regain the high-status social identity that had made his 1678 testimony credible. James Q. Whitman, in his study of the divergence between American and European approaches to criminal punishment, discusses Oates’s case as an example of “status abuse,” where a high-status person receives a punishment ordinarily reserved for low-status persons, with the aim of enhan cing the severity of the punishment by increasing its degree of degradation (Whitman 2003, 28). That Oates was to be repeatedly and publicly subjected to treatment ordinarily reserved for low-status offenders was clearly part of the court’s intent.4 In contexts that are, officially at least, more egalitarian, a more salient kind of degradation occurs not when one’s place in the social hierarchy is diminished, but rather when one’s treatment strips one of the status of humanity itself. Human status is typically thought to confer various entitlements to treatment. Echoing Kant, the distinctive value that confers these entitlements is often referred to as human dignity. One might analyze human dignity in terms of the Judeo-Christian conception of humans as created imago Dei. Or it might be traced to some essen tial capacities thought to be possessed by humans as such; for example, the cap acity for reasoning, moral reflection, or self-consciousness. Both paths lead, as Jeremy Waldron puts it, to “an insistence that this ontological dignity is the birth right of every human” (Waldron 2008, 37). The entitlements violated when one is degraded along this dimension are not conferred by social rank; rather, they are one’s human rights. A subject whose degradation strips them of their “ontological dignity” is dehumanized. The torture and humiliation endured by the Iraqi pris oners at Abu Ghraib at the hands of American soldiers provides a good example of dehumanizing degradation. Put on leashes, forced to bark like dogs, and made to pose for photographs as if they were trained animals, these subjects were treated in ways that could be appropriate only if they were inhuman (Fay and Jones 2004). Both rank-based and dehumanizing degradation provide clear illustrations of its public and performative nature. The pain the victim of the pillory or public whipping endures, though severe, is mild compared to their public humiliation. Moreover, the subject in the pillory is not merely declared to be diminished in status; they are diminished in status by being pilloried. The prisoners at Abu Ghraib were (initially) degraded before a smaller audience consisting only of
4 The prohibition against “cruel and unusual punishment” in the eighth amendment to the US Constitution was copied from the English Bill of Rights adopted in 1689, in a political atmosphere that was charged by discussion of Oates’s treatment. Indeed, our main source of information for contem porary understandings of the relevant notion of “cruelty” comes from Oates’s petition appealing his punishment, and the legal discussion that it prompted. See Granucci (1969) for a detailed and fascin ating history. Whitman speculates that the “status drama” involved in Oates’s punishment was an essential part of its perceived “cruelty” (Whitman 2003, 157).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
Epistemic Degradation and Testimonial Injustice 159 their captors and each other. But in photographing their handiwork, their captors clearly had a broader audience in mind. And the performative nature of this par ticularly grotesque example of degradation needs no explanation. The public, performative nature of degradation is crucial to understanding its destructive power. Anyone who witnesses a public performance of degradation is invited to treat its subject as if they were low status, which—even without a formal cere mony of the sort that interested Garfinkel—causes the person’s social status to be lowered. The damage done to the degraded person’s self-understanding also depends upon these public and performative aspects. The subject’s awareness of their diminished public representation will tend to diminish their own concep tion of their status. They will come to identify with the degraded image of them selves they know is available to the onlooker’s gaze. So much for the structure of degradation. Arguably, degradation always harms its victim in a profound way, by threatening the destruction of their social iden tity. And wrongful degradation is clearly a very grave harm. When it results in the successful destruction of a person’s social identity, it has wide-ranging negative consequences for every aspect of their social and material existence, as well as to their self-conception. Milder forms of degradation may damage a person’s iden tity and self-conception in profound ways, without completely destroying it. In the rest of this chapter, I will develop the idea that identity-prejudicial credibility deficits wrong by degrading. If this is correct, it provides a unified account of the harm of testimonial injustice. * * * Degradation occurs against the background of a hierarchy of social types and the status-linked entitlements they involve. The wrong of testimonial injustice is done, first and foremost, to someone insofar as they are a speaker. Is there a way to construe speakers as subjects whose “rank” entitles them to various treatments, which will enable us to see the harm of testimonial injustice as a kind of degrad ation? In this section, I sketch a picture on which we can. The picture draws inspiration from a discussion that has exploded among epistemologists over the past few decades concerning the so-called “norm of assertion”. The claim that assertion is norm-governed has intuitive purchase because we can easily imagine cases in which we would resent or criticize a speaker for their assertion. Where there is legitimate criticism or resentment, there is a genuine wrong. Our feeling that our resentment or criticism in these cases would be legitimate lends credence to the idea that assertion is governed by a distinctive norm. Timothy Williamson kicked off the contemporary debate with the provocative suggestion that the rule “assert only what you know” was constitutive of the speech act of assertion (Williamson 2000). This suggestion has spawned a huge literature of counter examples and competitors.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
160 Geoff Pynn The debate about Williamson’s proposal can be seen as an exploration of the basic responsibilities of speakers. In what follows, I make two assumptions. The first is that assertion is epistemically normative, in the way presupposed in much of the literature responding to Williamson’s proposal. Assertions can be proper or improper, independently of any norms or rules that contingently apply to any particular assertion (e.g. that one be polite, discreet, helpful, and so on); more over, an assertion’s propriety depends essentially upon the epistemic position of the speaker vis-à-vis the content they have asserted (so, for instance, if the only requirement on proper assertion is that it be true (e.g. as in Weiner 2005), the assumption of epistemic normativity is false). The second is that the status rele vant to the epistemic normativity of assertions is knowledge, i.e. that Williamson’s idea is more- or- less on the right track. While I’m attracted to a generally knowledge-first approach to epistemic matters, the second assumption is made mostly for the sake of convenience. If the normativity of assertion is a matter of justified belief, reasons, certainty, or something else instead, that would require rephrasing the current discussion in different epistemic terms, but no deep revi sions, at least as far as I can tell. Suppose that a speaker knows what she asserts; she has discharged her asser toric obligations with respect to her assertion. An assertion, once made, must be dealt with by its audience. When an audience accepts an assertion, they somehow take its content on board, on the basis of its having been asserted by that speaker. Employing a picture of conversation familiar from Robert Stalnaker’s work, we can say that an assertion is accepted when its content is added to the shared pre suppositions of conversational participants or added to the common ground (Stalnaker 1978, 2002). The point of asserting something, for Stalnaker, is to expand the common ground by adding the assertion’s content to it. There are various ways an assertion may fail to be accepted, and hence fail to achieve its purpose. It may be outright denied; it be challenged; it may be ignored. These are different kinds of failures, but in each case, the speaker’s assertoric purpose is frustrated. They are prevented from changing the conversational “score” in the way their assertion was intended to do. In each case, the speaker’s assertion is rejected. Does a speaker who has discharged her assertoric duties thereby earn some entitlement to treatment? This is plausible. Just as some assertions are improper, so are some non-acceptances. Consider the following exchange: Helga can’t remember when the department meeting will be held, so she asks her colleagues if anyone knows. Sarah knows that it is scheduled for Friday, and confidently asserts this. “Are you sure?” Helga asks, “I don’t want to come in to campus on Friday unless I have to.” Yes, Sarah replies, she’s quite sure; she just read the email announcement an hour ago and put the meeting on her calendar.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
Epistemic Degradation and Testimonial Injustice 161 But though she has no evidence that Sarah is wrong, she replies, “Well, thanks, but I’m going to check with someone else first.”
Borrowing Williamson’s language, we can say that Sarah is entitled to feel some resentment at her treatment. And where there is legitimate resentment, there is a genuine wrong. One proposal for explaining the wrong in this case would be a norm governing acceptance directly parallel to that governing assertion. Mirroring the knowledge rule of assertion, we might suppose that there is a knowledge rule of acceptance, that enjoins an audience to accept any assertion that is known by the speaker. Or, equivalently, that the audience must reject an assertion only if it is unknown. This rule would explain why Sarah was entitled to feel resentment: she asserted something that she knew, but her assertion was rejected anyway. Obviously, there are other possible explanations for her feeling of resentment. There might be a different rule of acceptance. Or her resentment might be explained without positing a rule of acceptance. For the sake of the pic ture I’m creating, I will assume that something like the knowledge rule of accept ance is both on the right track and the best explanation for Sarah’s resentment, though a full defense of this idea would obviously require extensive further work. Nonetheless, to the extent that the rule can be put to work to explain the primary harm of testimonial injustice, some abductive evidence for its plausibility will have been supplied. One objection to the knowledge rule of acceptance is that it leaves no room for an audience to reject a known assertion that they have good reason to believe is unknown. Perhaps knowledgeable speakers possess a kind of authority that gives them a right against their assertions being rejected without good reason. A respon sible speaker does seem entitled to have their assertions treated with respect, which means having them weighed carefully by her audience. But if, in respon sibly considering the assertion, the audience finds good reasons to reject it, surely they are permitted to do so (perhaps they are obliged to do so). A knowledgeable speaker who finds themselves in a conversation with interlocutors whose evi dence prevents them from adding their shared knowledge to the common ground is unfortunate, but not wronged by their interlocutors. They cannot rightly resent their audience for their having the evidence they do and acting in accordance with it. They did the best they could, given their evidence. Note that a parallel complaint is often lodged against the knowledge rule of assertion. Speakers may be in an epistemic position that is internally indistin guishable from knowledge, and yet fail to know. According to the knowledge rule, such speakers do wrong by asserting. But this can seem implausible: such s peakers are doing the best they can, in light of their evidence. How can they be asserting improperly nonetheless? Borrowing one of Williamson’s examples, if I clearly see what looks like snow falling outside my apartment window, I am authorized to
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
162 Geoff Pynn assert that it is snowing, even if—unbeknownst to me—what I am seeing is actu ally fake snow being dropped from the roof for the sake of a scene being filmed by a crew on the street below. In response to this kind of objection, partisans of the knowledge rule typically distinguish between a blameworthy violation of the rule and a blameless one, or between (as DeRose 2002 puts it) primary and secondary propriety. To defend the knowledge rule against this kind of objection requires some sort of distinction between ways of going wrong. If this distinction can be made to work to defuse the objection on the side of speaker obligations, it should also be capable of defusing the concern on the side of speaker entitlements. Suppose I do know about the film crew and the fake snow, but my audience is convinced that the snow is real, and on this basis refuses to accept my assertion that it isn’t snowing. They may have good reason for thinking that I don’t know what I’m talking about, and hence good reason for their rejection. They reason ably take me to be wrong and reject on that basis. Hence, we may say that their violation of the acceptance rule was blameless or exhibited secondary, but not primary propriety. The discussion about the knowledge rule of assertion fur nishes the defender of the knowledge rule of acceptance with plenty of materials for excusing audiences who reject wrongly, but for good reasons. Even so, this kind of objection reflects a deep disagreement about the nature of the rules that structure conversation. One way of thinking about those rules is in terms of procedural propriety. For a speaker and audience to act as they should in conversation is nothing more than for them act with epistemic due diligence. By contrast, one might think of those rules in terms of optimal outcomes. According to the knowledge rule of acceptance, an audience is obliged to accept what a knowledgeable speaker asserts; put in somewhat different terms, a knowledgeable speaker doesn’t merely have a right against improper procedure; rather, they are entitled to a particular conversational outcome. The same distinction is familiar from other normative domains. Suppose that Victor is convicted of a crime and punished, despite being innocent. Is he wronged? One can imagine two different answers. A proceduralist might say that so long as his conviction followed all the procedural requirements of justice, Victor wasn’t wronged. Of course, those pro cedures are designed to reliably produce convictions only in cases of guilt. But even the best procedures can produce inaccurate outcomes from time to time. Nonetheless, according to the proceduralist, occasional punishments of the inno cent, however unfortunate, are not on that account unjust, so long as the innocent person was convicted according to a just procedure. From the perspective of just outcomes, however, Victor is clearly the victim of a grievous wrong, even if the procedure that led to his conviction was fair in principle and perfectly executed in this case. He may not have a right to complain about the procedures that were used to convict him, but he is still entitled to complain about the outcome of those procedures. An innocent person has a right, on the outcome-based view, against being punished for a crime they did not commit.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
Epistemic Degradation and Testimonial Injustice 163 I take it that in the case of criminal trials, there is substantial intuitive weight behind the outcome-based approach to justice. That an innocent person con victed of a crime would not thereby have had their rights violated seems like an awfully cold position to take. This is not, of course, an argument in favor of the outcomes-based approach to conversational propriety! But the comparison raises the question of why we might think that knowledgeable speakers are entitled only to a certain sort of respectful treatment, rather than to the accomplishment of the purposes their proper assertions were meant to accomplish. Why not think that a speaker whose known assertion is rejected is thereby wronged? Surely, such wrongs are often quite mild; perhaps the knowledge rule of acceptance, like the knowledge rule of assertion, is violated with great regularity. Indeed, it might be a common occurrence that, all things considered, the right thing for an audience to do is to violate its conversational obligations towards a knowledgeable speaker. In any event, this sketch is only the beginning of a defense of the knowledge rule. Moreover, the picture for which it is intended in the present context would be almost as well served by a toned-down procedural version of the rule, according to which known assertions may not be rejected without good reason for thinking they are unknown. This modification would necessitate some cumbersome revi sions to the discussion in the next section, but none of them would be impossible in principle. The use to which I will put the rule is compatible with the weaker, proceduralist form. But as I favor the stronger form, and because the discussion will be less cumbersome without the weakening, I’ll proceed with the outcomefocused rule in mind. * * * Throughout the testimonial injustice literature, it is commonly presupposed that the victims of testimonial injustice deserve to have their testimony taken more seriously, and given more credence, than their audiences do. It isn’t just that the operation of a negative identity prejudice results in the speaker receiving a lower degree of credibility than they otherwise would have done. The prejudice results in their receiving a less credibility than they should have. But we find little beyond the intuition that this is true to explain why it should be so. Fricker writes that, “Epistemological nuance aside, the hearer’s obligation is obvious: she must match the level of credibility she attributes to her interlocutor to the evidence that he is offering the truth” (Fricker 2007, 19). This maxim implies that a speaker whose credibility assessment doesn’t match “the evidence that he is offering the truth” is wronged. But this evidentialist maxim is not a plausible candidate for explaining why the speakers in the central cases should have received more credibility than they do. For one thing, it is quite unclear what evidence the audiences in the cases have, or what level of credibility they ought to have assigned the speakers, were they adhering to the maxim. What credibility
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
164 Geoff Pynn assignment to Marge’s suspicions would match Greenleaf ’s evidence? What is Greenleaf ’s evidence? Frankly, I have no idea how to answer either question—but despite my ignorance about these matters, I am confident that he assigns her less credibility than he should. Moreover, it isn’t difficult to imagine cases of testimo nial injustice where the listener does attribute to her interlocutor the degree of credibility that matches her evidence. A young white supremacist, raised in a virulently racist environment, may have scant evidence of the reliability of black speakers; moreover, his overall evidence may even support the idea that black speakers are especially unreliable: he has no evidence that his racist community is wrong, as he has simply never encountered any relevant prejudice-disconfirming information. Such a listener may simply lack evidence that a black speaker is offering the truth, and so assess their credibility in such a way as to line up per fectly with his evidence that the speaker is offering the truth. Nonetheless, he commits a testimonial injustice when he rejects a black speaker’s assertion on the basis of his racist stereotype. His prejudice leads him to accord the speaker less credibility than he ought to be accorded. Maitra points out that in the central cases, the speaker and hearer stand in “certain special interpersonal relationships.” In Tom Robinson’s case, the jurors have special obligations towards Tom that they fail to meet. Marge and Herbert have a “personal, quasi-familial relationship” that may involve “an obligation to take . . . testimony seriously” (Maitra 2011, 200). Perhaps particular entitlements and obligations flow from these relationships, and it is these that set the requisite level of credibility the violated speakers ought to receive. Without dismissing this suggestion, it is worth noting that it’s unclear how to make it work, even in these central cases. Juries don’t have a default obligation to assign a particular degree of credibility to witnesses. Of course, they may have an obligation not to allow their prejudices influence their credibility assessments, but it’s hard to see this as a spe cial obligation deriving from their status as jurors. Nor is it obvious that the rela tionship between a person and their child’s fiancé carries with it any particular obligations to assess credibility in any particular way. Families are complicated! The knowledge rule of acceptance explains quite straightforwardly what level of credibility the speakers in the central cases ought to receive. Tom and Marge make knowledgeable assertions, so the rule obliges their audiences to accept their assertions. Their audiences instead reject their assertions. With this rule in hand, there is no need to cast about for particular entitlements flowing from the special relationships between these speakers and their audiences, or to gesture towards vague evidentialist principles. To reject a known assertion is, given the rule, thereby to accord the speaker less credibility than the level to which they are entitled. Moreover, this entitlement is linked to a system of epistemic “ranks”: knowers and non-knowers. Someone who knows thereby has a superior epi stemic status to someone who doesn’t. So, to reject a known assertion is to violate the speaker’s entitlements in a way that exhibits the skeletal structure of
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
Epistemic Degradation and Testimonial Injustice 165 degradation: it is to treat a person with a high status as if they had a lower status. The status of the knower comes with an entitlement to acceptance and the audi ence who rejects a known assertion violates this status-linked entitlement, treat ing them in a way that would be appropriate only if they occupied the lower status of non-knower. A violation of a status-linked entitlement is not sufficient to degrade. The viola tion must also be public. But any explicit conversational rejection of a known assertion is public. A conversational sequence of assertion and subsequent rejec tion is open to all participants: which contents are and are not part of the com mon ground is a function of speaker presuppositions, which are intrinsically public attitudes. Moreover, the implicit rationale for the rejection compounds the public and performative nature of the violation. Given the knowledge rule of acceptance, to reject a speaker’s assertion is in part to represent the speaker as a non-knower, for the same sort of reason that to assert represents the speaker as a knower. In general, when there is a shared presumption that performing a certain action is permissible only if a particular condition is met, someone who performs the action thereby represents the condition as having been met. As Robert Brandom puts it, “Our social practices, as Wittgenstein emphasized, treat per formances as appropriate and in accord with those practices until and unless some specific question is raised about them” (Brandom 1983, 642). To reject a speaker’s assertion is, in effect, to represent them as lacking the status of someone who is entitled to shape the conversation’s common ground. Such a rejection thus has a structure apt for degrading. While the case is of course very mild, I think we can get a glimpse of the way that Helga’s rejection of Sarah’s assertion involves an element of degradation. In rejecting her assertion, Helga violates Sarah’s entitle ment to acceptance in a way that represents her as a non-knower. Provided the rejection is allowed to stand, participants will treat Helga as having rejected prop erly, and given the knowledge rule, that means they will treat Sarah as someone who doesn’t know what she asserts. Insofar as an improper rejection involves a public representation of a knowing speaker as a non-knower, it can degrade. Again, the form of degradation suffered by Sarah by means of Helga’s rejection is very mild. But the characteristic damage of degradation can be seen even here. Upon her rejection, she is faced with an unpleasant choice: she can either allow the diminishing representation to stand, or she can dig in and fight back. Either course risks undermining her public iden tity as a knowledgable speaker. To be passive is to accept her degradation; to fight back may produce embarrassment and possibly other secondary harms to her social identity, such as a reputation for uncooperativeness or a tendency to over react—and, of course, the fight may be lost, in which case the end result will be no better than had she allowed the degrading representation to stand in the first place. She may even begin to doubt herself: “I thought I knew, but I’m being treated as if I didn’t . . . is that right? Do I know?” Helga’s rejection violates Sarah’s
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
166 Geoff Pynn conversational entitlement in a way that produces a diminished representation, resulting in some combination of frustration, humiliation, disharmony, and decreased self-regard. Even if Helga’s rejection is an instance of epistemic degradation, that doesn’t mean we should treat it as an instance of testimonial injustice. If it is, surely it’s a marginal case. It lacks one crucial ingredient of Fricker’s central cases, which is their being rooted in systematic negative identity prejudices. That etiology gives the central cases the potential to degrade their victims quite significantly, or to reinforce a pattern of significant degradation already in place. In Tom’s case, the violation of the speaker’s entitlement is rationalized by or prompted by a vicious stereotype about black men. To the extent that the stereotype’s role in the rejec tion is publicly accessible, the rejection communicates a powerful negative mes sage about the speaker. By treating Tom as a non-knower in virtue of being a black man, the jury directly links his diminished epistemic status to a low-status social identity. Temporarily elevated on the witness stand, he is again brought low through the jury’s rejection of his testimony. Though the jurors do not give voice to their prejudice, Atticus Finch is able to see their rationale quite well: they are under the sway of a stereotype according to which “all Negroes lie, all Negroes are basically immoral beings, all Negro men are not to be trusted around [white] women.” In publicly violating his entitlement to acceptance by reliance on this negative stereotype, the jurors convey that the stereotype is true, both in general and of him in particular. The violation is profoundly diminishing, representing him as a lying predator in virtue of his race. This degrading structure is also apparent in Marge’s case. Her knowledgeable assertions are rejected by (this time, explicit) reference to a disparaging stereo type about women’s rationality. Evoking a stereotype of women as overly emo tional, irrational creatures to rationalize his rejection of Marge’s testimony, Greenleaf violates her entitlement to acceptance in a way that severely diminishes her conversational standing and hence her epistemic status. In this case, the inter ior aspect of degradation’s harm is vividly present as well. After repeatedly suffer ing the diminishing violation of her entitlement to acceptance, she is “reduced to a self-contradictory mantra, repeating emphatically to the incredulous Greenleaf, ‘I don’t know, I don’t know, I just know it’ ” (Fricker 2007, 88). The degradation account proposes treating such public violations of these speakers’ conversational entitlements to acceptance as the fundamental harm of testimonial injustice. Epistemic degradation consists in a public violation of a speaker’s epistemic status-linked entitlements. Given the knowledge rule, a know ing speaker whose assertion is rejected on the basis of an identity-prejudicial credibility deficit suffers a violation of her entitlement to acceptance. Any viola tion of the knowledge rule will tend to represent its victim as a non-knower, which is itself moderately degrading. But in the central cases, where the violation is understood to be rooted in a systematic negative identity prejudice, the
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
Epistemic Degradation and Testimonial Injustice 167 rejection will also represent the victim as a non-knower who is debased in all the ways encoded by the stereotype. In the cases just canvassed, the diminished rep resentation depicts the speaker as untrustworthy in virtue of his blackness or irrational in virtue of her femininity. Their epistemic degradation is magnified through the lens of a prejudicial stereotype that excludes them from the social and epistemic rank shared by the other conversational participants. Because degradation is intrinsically public, it can’t be accomplished through one individual’s thought alone. This helps to explain why cases like Maitra’s Zara don’t constitute testimonial injustices, despite their involving identity-prejudicial credibility deficits. Acceptance as characterized here occurs within the context of a conversation structured by shared presuppositions. Though the relationship between a writer and their potential audience is loosely analogous to a Stalnakerian context, it lacks the open communicative structure of a conversa tion. The reader’s cognitive reactions do nothing to modify the conversational “score”; there is no dynamic common ground to be influenced by the audience’s shifting presuppositions. Hence the norms governing assertion and acceptance in a conversational context are not in place in the “context” comprised of a writer and her audience. You can’t violate their conversational entitlement to acceptance unless you are engaged in a joint conversational exercise. And you don’t publicly diminish them simply by representing them to yourself in a low-status way. If epistemic degradation is the primary harm of testimonial injustice, then simply neglecting to read an article, even on the basis of a negative identity prejudice, doesn’t rise to the level of testimonial injustice. However epistemically irrespon sible it may be, merely ignoring their work isn’t sufficient to degrade a writer. In addition, the violation at the heart of the account is rejection, which is not the same as inattention. To the extent that an analogy can be drawn between conver sation and the interaction between a reader and writer, what Zara does is more akin to simply not engaging in conversation than to rejecting any assertions the tea-partier makes. Of course, a social practice of ignoring a speaker or writer, where that practice is rooted or rationalized by an identity prejudice, could easily constitute an epi stemic injustice. Indeed, this sort of phenomenon lies at the heart of Charles Mills’s notion of white ignorance, and what Kristie Dotson calls a “harmful prac tice of silencing.”5 Social practices are public; among their effects are the various meanings they have in the societies in which they occur. If the ignored speaker is knowledgeable, an extended degrading violation of broadly the same sort as occurs in Fricker’s central cases will occur. As the prejudicial practice of ignoring and its rationalization become more widely known, the diminished depiction of the writer or speaker conveyed by the relevant negative stereotype will be made
5 Mills (2007); Dotson (2011).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
168 Geoff Pynn available. So, while a one-off private identity-prejudicial credibility deficit won’t, on this account, constitute a testimonial injustice, because it does not degrade, such an instance may be part of a pervasive social practice of ignoring that does constitute a testimonial injustice, given the degrading effects such a practice would be liable to have. Finally, the degradation account gives us a way to distinguish the central cases from the weird ones where the speaker is not, in fact, exercising her capacity as a giver of knowledge. A lying speaker doesn’t assert what he knows, and so enjoys no entitlement to acceptance—at least, not in virtue of the knowledge rule of acceptance. An identity-prejudicial credibility deficit perpetrated against a nonknower thus doesn’t violate the speaker’s entitlements, and so isn’t an instance of the relevant kind of wrongful degradation, lacking (as it does) the primary harm at testimonial injustice’s core. Nonetheless, such a rejection may have some of the same harmful effects as wrongful degradation. Even if Tom were lying on the stand, the diminished representation made available by the jury’s prejudicial rejection would still be operative, and hence the public diminishment character istic of degradation would still be present. Thus, the incident could still be under stood as a member of the extended family of testimonial injustice, even though it doesn’t involve a violation of the speaker’s conversational entitlements, or an injury to them in their capacity as a knower. Hence the degradation account of testimonial injustice, insofar as it makes use of the rule of acceptance to charac terize the violation at the heart of the wrong, enables us to understand the harm of subjecting lying or delusional speakers to an identity-prejudicial credibility deficit, without requiring us to treat such cases as instances where the speaker is wronged in the same way as they are in a case of testimonial injustice. * * * I have attempted to make sense of Fricker’s fruitful suggestion that testimonial injustice involves a distinctive kind of degradation. In the central cases, where the speaker’s knowledgeable testimony is rejected on the basis of a systematic nega tive identity prejudice, the rejection constitutes a kind of severe degradation by rationalizing a violation of the speaker’s status-linked conversational entitlements in terms of a diminishing representation of them as a non-knower. Insofar as this account illuminates what is harmful in all cases of testimonial injustice, it pro vides some indirect evidence for the idea that knowledgeable speakers are entitled to shape conversational contexts in the way they intend. But even if you are unpersuaded that there is something like a knowledge rule of acceptance, the degradation account of testimonial injustice may still be useful. It sheds light on the structure of the harm at the heart of this form of injustice. Testimonial injust ice can be seen as a violation that diminishes its victim in the ways characteristic
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
Epistemic Degradation and Testimonial Injustice 169 of degrading treatment more generally. To elaborate this understanding, we can seek an account of the entitlements linked with various epistemic statuses and an understanding of how prejudicial credibility judgments produce degrading repre sentations of speakers whose epistemic status entitles them to be heard.
References Bergner, Raymond (1987). “Undoing Degradation.” Psychotherapy 24(1): 25–30. Brandom, Robert (1983). “Asserting.” Noûs 17(4): 637–50. Cavender, Gray, Kishonna Gray, and Kenneth Miller (2011). “Enron’s Perp Walk: Status Degradation Ceremonies as Narrative.” Crime, Media, Culture: An International Journal 6(3): 251–66. Craig, Edward (1990). Knowledge and the State of Nature. Oxford: Oxford University Press. Davis, Emmalon (2016). “Typecasts, Tokens, and Spokespersons: A Case for Credibility Excess as Testimonial Injustice.” Hypatia 31(3): 485–501. DeRose, Keith (2002). “Assertion, Knowledge, and Context.” The Philosophical Review 111(2): 167–203. Dotson, Kristie (2011). “Tracking Epistemic Violence, Tracking Practices of Silencing.” Hypatia 26(2): 236–57. Ducharme, Lori and Gary Fine (1995). “The Construction of Nonpersonhood and Demonization: Commemorating the Traitorous Reputation of Benedict Arnold.” Social Forces 73(4): 1309–31. Fay, George and Anthony Jones (2004). “Investigation of Intelligence Activities at Abu Ghraib.” Defense Technical Information Center (Report), U.S. Department of Defense. Fricker, Miranda (2007). Epistemic Injustice: Power and the Ethics of Knowing. Oxford: Oxford University Press. Garfinkel, Harold (1956). “Conditions of Successful Degradation Ceremonies.” American Journal of Sociology 61(5): 420–4. Granucci, Anthony (1969). “Nor Cruel and Unusual Punishments Inflicted: The Original Meaning.” California Law Review 57(4): 839–65. Maitra, Ishani (2011). “The Nature of Epistemic Injustice.” Philosophical Books 51(4): 195–211. Mills, Charles (2007). “White Ignorance,” in Shannon Sullivan and Nancy Tuana (eds), Race and Epistemologies of Ignorance. New York: SUNY Press, 11–38. Pohlhaus, Gaile (2014). “Discerning the Primary Epistemic Harm in Cases of Testimonial Injustice.” Social Epistemology 28(2): 99–114. Stalnaker, Robert (1978). “Assertion.” Syntax and Semantics 9: 315–22.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
170 Geoff Pynn Stalnaker, Robert (2002). “Common Ground.” Linguistics and Philosophy 25: 701–21. Thérèse, Sandrine and Brian Martin (2010). “Shame, Scientist! Degradation Rituals in Science.” Prometheus 28(2): 97–110. Waldron, Jeremy (2008). “Cruel, Inhuman, and Degrading Treatment: The Words Themselves.” NYU School of Law, Public Law Research Paper No. 08–36. Whitman, James Q. (2003). Harsh Justice: Criminal Punishment and the Widening Divide between America and Europe. New York: Oxford University Press. Weiner, Matthew (2005). “Must We Know What We Say?” The Philosophical Review 114: 227–51. Williamson, Timothy (2000). Knowledge and Its Limits. Oxford: Oxford University Press.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 19/03/21, SPi
8
My Body as a Witness Bodily Testimony and Epistemic Injustice José Medina and Tempest Henning Is testimony without words possible? For example, can a racially oppressed subject give testimony of her experiences of oppression without verbalizing them? This issue is particularly pressing when there are communicative obstacles for the oppressed subject to verbalize her experiences, as happens under conditions of hermeneutical injustice (Fricker 2007); that is, when the expressive and interpretative climate is so constrained or impoverished that there are no readily available resources to communicate adequately the experiences in question. This issue is also particularly pressing when there is no opportunity available for the oppressed subject to be treated as a credible speaker, as happens under conditions of testimonial injustice (Fricker 2007); that is, when unfair testimonial dynamics prevent the subject from speaking (“preemptive silencing”—Fricker 2007), or simply inhibit or intimidate the subject (“testimonial smothering”—Dotson 2011), or systematically discredit her verbal contributions. Hermeneutical and testimonial injustices strongly undermine and heavily constrain the communicative and epistemic agency of oppressed subjects, but they do not completely annihilate such agency: even without words available and without opportunities to speak or to be heard properly, the subject can still express herself through her body. It is for this reason that, for socially and politically engaged epistemology that wants to address issues of oppression, it is of the utmost importance to include bodily testimony, for such testimony is a key part of resisting oppression under conditions of epistemic injustice when the communicative (and especially verbal) agency of testifiers becomes unfairly constrained. Under such conditions, bodily testimony creates possibilities for sharing experiences, speaking up against oppression, and forming communities of resistance that verbal testimony cannot offer. Through bodily communication, epistemically marginalized subjects can signal to each other that there is something wrong they are experiencing; and, in this way, they can share experiences of oppression and start building a community of resistance. We are interested in elucidating how members of such a community engage in epistemic resistance;1 that is, how they 1 Epistemic resistance should be understood as covering any activity that fights against the different kinds of epistemic marginalization or epistemic oppression involved in epistemic injustice. For a full discussion of this notion, see Medina (2013). José Medina and Tempest Henning, My Body as a Witness: Bodily Testimony and Epistemic Injustice In: Applied Epistemology. Edited by: Jennifer Lackey, Oxford University Press (2021). © José Medina and Tempest Henning. DOI: 10.1093/oso/9780198833659.003.0008
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
172 José Medina and Tempest Henning fight back against their epistemic marginalization and struggle to exercise their communicative and epistemic agency. As we will explain below, a key component in building such a community of resistance is the communicative and epistemic solidarity that can be garnered through resistant bodily testimony. We will return explicitly to the community-building aspect of bodily testimony in section 4, but the role of bodily testimony in epistemic resistance (and in epistemic activism2) will be discussed throughout this chapter. In section 1, we will make conceptual space for bodily testimony in the literature on testimony, arguing for a way of broadening the notion of testimony by paying more attention to communicative possibilities afforded by our embodied positionality in communicative activities. In sections 2–3, we will develop our account of how to elucidate the testimonial force and content of non-verbal communicative acts such as gestures and facial expressions in relation to three elements: the communicative context, the embodied positionality of the communicator, and the communicative uptake that the audience gives, or fails to give, to the expressive behavior of the body. In section 4, we will explain how non-verbal communication figures in group testimony and the role that bodily testimony can play in building communities of resistance. Section 5 will offer a brief overview of our analysis and argumentation in this chapter, highlighting our main conclusions and suggestions for further research on bodily testimony in social and political epistemology.
1. Making Room for Bodily Testimony In the philosophical literature on testimony, one can find implicit (and sometimes explicit) definitions of testimony as concerning solely illocutionary speech acts. This narrow conceptual framing of testimonial acts as referring only to speaking and hearing is not usually defended but presumed. Few who theorize about the concept and practice of testimony include the body and the ways in which the body either contributes to or constitutes testimonial acts. Similarly, insufficient attention is paid to the ways in which epistemic injustice can also concern the bodily expressive behavior of marginalized individuals. Some resources for the conceptualization of bodily testimony can be found in feminist theory and critical race theory. Feminist epistemologists3 have argued that the body is a site of knowledge and knowledge generation. One of their most salient contributions to epistemology is the insistence that knowers are embodied and always located somewhere. The schema “S knows that P” may suggest that every S will have the same access to P; 2 See Medina (2018) and Medina and Whitt (2021). 3 See Bar On (1993), Code (1993), Grosz (1993), and Harding (1993, 2004).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Bodily Testimony and Epistemic Injustice 173 but the social positioning of each S can give differential access to P and can also produce partial ignorance of P (Code 1993, 39). In addition to an individual’s embodied positionality aiding or handicapping them in terms of knowledge acquisition, such factors also affect their bodily abilities to communicate said knowledge. The epistemology of testimony needs to pay attention to the ways in which different subjects can express their knowledge that P based upon their embodied positionality as much as it must take into account the ways in which various subjects can know P in ways that are specific to their embodied social positionality. For example, if an individual is raised within a culture that utilizes particular musical time signatures over others, then musical counting can be an embodied act where the counting is felt rather than thought. Particular bodies, dependent upon their social positioning, are more apt to display specific knowledge of time signatures via bodily movements, whereas other bodies who may be situated within a different social position may either lack the embodied ability to know the rhythms and the beats or lack the physical ability to demonstrate their knowledge of certain time signatures. Bodily mannerisms and gestures can be crucial for the transmission of knowledge, both in terms of the performative expression of this knowledge by embodied communicators and in terms of its reception by hearers with a consonant embodied sensibility. While the philosophical literature on testimony often restricts testimonial acts to illocutionary speech acts, there are definitions of testimony that are broad enough so as to include communicative acts of all sorts, verbal or not. Take, for instance, Jennifer Lackey’s definition of testimony, which has no stipulation that testimony must be verbalized: “S testifies that p by making an act of communication a if and only if (in part) in virtue of a’s communicable content, (1) S reasonably intends to convey the information that p or (2) a is reasonably taken as conveying the information that p” (2008, 30). According to this definition, testimony is a communicative act with informational content that is either intended by the communicator or reasonably ascribed by the audience. The gestures or mannerisms enacted by embodied communicators typically contain intended information and/or are reasonably interpreted as expressing communicative content by the hearers; and, therefore, under Lackey’s definition, they should qualify as testimony. Although Lackey tends to focus “on what speakers say,” her theory of testimony makes room for other communicative acts: “I advance a new theory that instead focuses on the linguistic or communicative items in testimonial exchanges—such as statements and other acts of communication” (2008, 2). Note that this makes conceptual room for bodily communication, which includes non-linguistic, communicative acts such as gestures and facial expressions. Given the in-principle inclusion of non-linguistic communicative acts within Lackey’s definition of testimony, it is within the purview of her theory to think about knowledge acquisition from testimonial exchanges as learning communicative contents from the words or from the body (e.g. bodily gestures or facial
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
174 José Medina and Tempest Henning expressions) of communicators. Expanding the focus of the study of testimony to include bodily communication can bring to the fore interesting complications for the analysis of testimonial dynamics. For example, issues of access to body language, and not only to verbal language, need to be taken into account; and this can call attention to important epistemic differences among recipients of testimony, depending on their capacity to access and to interpret the bodily aspects of communication. Consider, for instance, possible differences between the addressee of a testimonial act and the eavesdropper. Lackey suggests that there is no reason to think that the addressee of a testimonial act is epistemically better off than the eavesdropper: both the addressee and the eavesdropper have access to the content of the testimonial act and, therefore, are in identical epistemic predicaments with respect to the beliefs they form on the basis of the testimony (see Lackey 2008, 234). But although this seems to be true for verbal testimony, it is not necessarily true for bodily testimony (or for the bodily aspects of verbalized testimony): differently positioned eavesdroppers can have quite different access to the bodily aspects of the testimony, ranging from having limited access (e.g. not being able to see the entire face of the communicator from their angle and therefore missing some of the facial expressions such as a wink), to having no access at all (e.g. in the case of an eavesdropper who cannot see the bodily gestures and facial expressions of a communicator who happens to be seated behind her). Not only the bodily positionality of the recipient of testimony can make an important difference for the reception of testimony, but also the recipient’s attunement to bodily language and her capacity to interpret body language can make a crucial difference. Note that speakers of the same verbal language—for example, speakers of English—can be described as belonging to two different communitive groups, an in-group and an out-group, depending on whether or not they are able to interpret a particular set of bodily gestures. This means that someone c ommunicating in English with a diverse audience of English speakers may be able to communicate particular contents with her body language only to a particular subset of members of that audience, those who are attuned to and are able to interpret those particular gestures. Verbal languages are, of course, racially inflicted, and so is body language. In this chapter, we will focus on the bodily testimony of marginalized racial subjects, and we will pay attention to how the communicative contents of their bodily gestures are received and interpreted by differently positioned audience members and, in particular, by out-group and in-group audiences. The philosophical literature on testimony has not entirely neglected the body. Hume asserts that examination of the body should be used to determine the reliability of a testimonial exchange. For example, he points out that “suspicious” bodily language (e.g. edginess, restless facial expressions, etc.) can be grounds for calling the testimonial act into question. Exuberant gestures or gestures that communicate hesitation, for Hume, also warrants skepticism concerning the
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Bodily Testimony and Epistemic Injustice 175 testimony (Hume 1993, 75). But these arguments concerning the body within the testimonial literature situate the body in an auxiliary position. However, we contend that the body can be communicatively significant not only in modulating testimonial acts, but also in constituting testimonial acts so that the body of the communicator can function as a vehicle or expressive medium of testimony and become a source of knowledge acquisition in its own right.
2. Giving Bodily Testimony The communicable content and the performative force of a gesture cannot be identified unless we connect the gesture with three elements in the communicative act: the context of communication, the embodied positionality of the communicator, and the receptivity of the addressee(s) or audience.4 We will devote this section to the analysis of the bodily testimony in relation to the first two elements, the communicative context and the embodied positionality of the communicator; and we will devote section 3 to issues of testimonial receptivity and uptake. Think, for example, of the gesture of raising your hands above your head in front of a police officer; and consider the following two scenarios. If the subject raises her hands in a crime scene after a police officer shouts “Don’t move!”, in this communicative context the performative force of the gesture is one of surrender or submission, and the communicative content can be captured in the statement, “I am unarmed, mean no harm, and comply.” However, this gesture can carry a different performative force and express a different content in a different communicative context such as that of a Black Lives Matter demonstration. After the killing of Michael Brown in Ferguson, MO, in 2014, many people of color used the gesture of raising their hands above their head in silence5 to give testimony of racial profiling, of being stopped by the police, despite posing no threat or having done nothing wrong, of being scrutinized unfairly and often violently. In this communicative context, the hands-up gesture has testimonial performative force, giving expression to experiences of racial profiling and mistreatment by the police; and it is used by people of color to express the following content: “I too 4 In a contextualist view of communication such as ours, this is true both of verbal and non-verbal communicative acts. Although speech acts have the appearance of being more autonomous because the linguistic portion of utterances can be detached and analyzed independently of the particular context of communication, the full analysis of speech acts also requires paying attention to the communicative context, the embodied positionality of the speaker, and the receptivity of the audience. The importance of these three elements is only accentuated in the case of non-verbal communication because it is hard to see how the analysis of the force and content of the non-verbal communicative act can even begin without considering these elements. 5 The iconic hands-up gesture was also used in demonstrations while chanting the slogan “Hands Up, Don’t Shoot!” The slogan makes verbally explicit what the gesture typically means, echoing what Michael Brown was trying to express to his killer. But note that the gesture can express more than that; in particular, it can have a robust testimonial content, as we argue in this chapter.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
176 José Medina and Tempest Henning have experienced being perceived as a threat while posing none, I too have to walk with my hands up.” Our account of bodily testimony does not only address cases in which the expressive behavior of the body can be easily translated into words (as in the hands-up gesture), but also cases in which gestures and facial expressions carry testimonial force and content that cannot be fully captured in verbal utterances. For example, the iconic side-eye of Rep. Maxine Waters carries with it testimonial force and content in regard to the 45th President of the United States, but unlike the hands-up gesture, there is no single verbal utterance that can fully capture the content of Rep. Waters’ facial expression. That is not to say that one could not approximate the force and content of Rep. Waters’ bodily testimony in verbal descriptions. Her bodily way of giving testimony about her reactions to the President’s lies and personal attacks with her facial expressions could be described through verbal approximations such as “fed up,” “over it,” or “unamused.” Verbal approximations can paraphrase but not fully capture the content of the testimonial act in this case. A testifier who uses her body in this way to express a communicable content and a performative force can be marginalized when her bodily gestures and facial expressions are not taken seriously and given credence—an authority or credibility deficit that can track identity prejudice (if her bodily testimony is dismissed as an overreaction of a black women, for instance) and amount to a case of testimonial injustice. But if the testifier’s bodily expression is not even understood or recognized as containing any content, then the epistemic marginalization operates at a different level—at the level of intelligibility—and might qualify as a case of hermeneutical injustice (Fricker 2007). Fricker states that “testimonial injustice occurs when prejudice causes a hearer to give a deflated level of credibility to a speaker’s word.” (2007, 1) We want to include here cases in which the audience assigns unwarranted deflated credibility to a testifier’s bodily expressions (and not only to her words). For example, if an individual were to dismiss our knowledge of what it is like to be a person of color in philosophy because they harbor prejudices against people of color in the domain of academia and philosophy, this would be a case of testimonial injustice. This can happen in cases in which the philosopher of color verbalizes her experiences of marginalization in the profession and is unfairly distrusted, but it could also happen in cases in which the philosopher of color uses bodily gestures or facial expression to express such experiences and she is given no credibility. On the other hand, in hermeneutical injustice, the epistemic marginalization occurs at the level of intelligibility. In hermeneutical injustices, a gap (or inadequacy) in collective interpretive resources situates particular individuals at a hindrance when they are attempting to communicate their experiences. In the example of Rep. Waters’ facial expressions of disapproval and protestation, political pundits have dismissed her as a professional and credible body from whom knowledge
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Bodily Testimony and Epistemic Injustice 177 can be obtained, and their inability to read her body language adequately may be due to their lack of access to the resources needed to properly interpret her gestural testimony or just to outright willful hermeneutical ignorance (that is, unwillingness to master and use the hermeneutical resources that are readily available—see Pohlhaus 2014). In these cases of dismissal of bodily communication, whether because of deflated credibility or because of deflated intelligibility, a p erson of color’s ability to be counted as a knower by a dominantly situated individual is being undermined via their social positionality as understood through the lens of prejudice against bodies of color. Let’s look more closely at the role that body language can play in communicative dynamics around hermeneutical injustice. As her primary illustration of the phenomenon of hermeneutical injustice, Miranda Fricker uses the example given by Brownmiller about the communicative frustration women experienced when being sexually harassed at the workplace before the expression “sexual harassment” was available in the collective vernacular. Using this example, Fricker argues that both perpetrator and victim of sexual harassment were epistemically disadvantaged because “neither has a proper understanding” of what was happening (2007, 151). Fricker goes on to say that such a disadvantage “prevents her [the victim] from understanding a significant patch of her own experience: that is, a patch of experience which it is strongly in her interests to understand, for without that understanding she is left deeply troubled, confused, and isolated, not to mention vulnerable to continued harassment” (Ibid.). But it would be wrong to conclude that the victim in this predicament cannot express at all the wrong she is experiencing in being harassed and in being confined to communicative isolation about her harassment. The expressive powers of the body can be easily disregarded here. Our expressive capacity to protest exceeds the words available to us. It would be a mistake to think that a victim’s body after an incident (or repeated incidents) of sexual harassment remains inexpressive, that the body cannot communicate or protest the harassment via gestures, facial expressions, bodily posture, and other corporeal schemas. Even without the right words, a victim can nonetheless protest what she is going through without naming it. She can testify to at least some aspects of her experience of being violated through body language; and she can also use her body to express her predicament of not being understood or able to articulate properly the wrong she is experiencing. Through bodily expressive behavior, she can convey to others that she is, as Fricker puts it, “deeply troubled, confused, and isolated” and remains vulnerable to mistreatment. We contend that, even in hermeneutically impoverished contexts, the bodies of subjects who have been hermeneutically disadvantaged and lack adequate words can convey an understanding of what is occurring, and their bodies can protest and warn others. We argue for the possibility of testimonial knowledge that bodies can express in
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
178 José Medina and Tempest Henning order to foreclose assertions that hermeneutical lacunas utterly prevent an individual from expressing experiences for which there are no words available yet and from alerting others about incidents or phenomena they cannot verbalize. There can be a ‘this’ or a ‘that’ which has no name that is singled out and pointed at to others through bodily communication. For communicative success, a facial expression or gesture (such as Rep. Waters’ side-eye) requires the audience to have the relevant knowledge of bodily communication in order for the gesture or bodily expression to be registered as such and to be properly interpreted. If one has received no exposure to gestures of that kind in embodied communicative dynamics and lacks the proper knowledge or familiarity with that communicative use of the body, then the gesture can be missed or misread; it can be regarded as lacking any testimonial or communicative value at all. In some cases, failure to properly understand bodily testimony can result from the unavailability of the relevant interpretative tools or hermeneutical resources for processing the bodily expressions. In these cases, the hermeneutical ignorance that prevents understanding can be said to be a matter of epistemic bad luck.6 However, in other cases, the hermeneutical ignorance is willfully created and maintained because of lack of attention to or interest in how members of some communities communicate through body language. This cultivated disregard or epistemic neglect results in what Gaile Pohlhaus Jr (2014) has termed willful hermeneutical ignorance, which consists in dominantly situated knowers actively ignoring the hermeneutical resources that marginally situated knowers have created and use (2014, 716). This type of willful hermeneutical ignorance with respect to body language is precisely what interests us for the cases of epistemic injustice we are examining in this chapter: cases in which there is a failure of epistemic responsibility in the audience to whom bodily communication is addressed for not doing everything they can for understanding and believing the body as they should. In section 3, we will examine what it means to pay attention to the body as a responsible communicator and to give adequate testimonial uptake to the expressive behavior of the body.
6 This is what is suggested by Fricker’s standard cases of hermeneutical ignorance. However, in her more recent work, Fricker talks about the “shared pool” of hermeneutical resources in a more qualified way and makes room for dissonant meanings and interpretations that are not widely shared. In this sense, Fricker talks about “localised hermeneutical practices:” “fully functioning yet insufficiently widely shared hermeneutical practices;” “localised or in-group hermeneutical practices that are nonetheless not shared across further social space” (2016, 166 and 167). By pluralizing her view of hermeneutical resources, Fricker’s current view becomes polyphonic and converges with Medina’s, as she herself points out: Medina is right to emphasise that the intersectional ignorances created by the possession and non-possession of this or that cluster of interpretive concepts growing out of this or that area of social experience tell a ‘polyphonic’ or multi-voiced story of power and resistance, societal conceptual impoverishment and localised interpretive sophistication and creativity. (2016, 167)
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Bodily Testimony and Epistemic Injustice 179
3. Giving Testimonial Uptake to the Body Like speech acts, non-verbal communicative acts depend on the uptake of the addressee or audience for their success. A communicative act such as a gesture or a facial expression will not even be registered as such, let alone properly understood or answered, if the addressee or audience does not have the appropriate receptivity to process it and give it uptake. The receptivity of bodily language requires embodied hermeneutical and testimonial sensibilities; that is, being attuned to the performative force and communicative content of bodily communicative acts such as gestures and facial expressions. Receiving proper uptake does not mean that the audience agrees with the content the communicator intends to express, or that audience members become actively recruited to sustain the performative force of the gesture or facial expressions and to facilitate the illocutionary and perlocutionary acts in question. It only means that there is a communicative response or reaction consistent with the content and force of the communicative act; that is, that there is enough communicative support on the audience’s side to suggest that the content of the gesture or facial expression in question has been adequately processed and that its force has been understood as being directed at the achievement of particular communicative acts. Giving proper uptake requires a particular kind of hermeneutical and testimonial sensibility that makes communitive cooperation possible. Think, for instance, of cases of testimonial smothering, which involves a form of self-silencing that occurs when the communicator perceives her audience as unwilling or unable to provide appropriate uptake (Dotson 2011).7 If facial expressions such as rolling one’s eyes are used as a response to testimonial smothering, this is likely to be registered and processed as an expression of discomfort and protest only if audience members are equipped to detect testimonial smothering (i.e. equipped to understand the conditions under which self-silencing happens and the epistemic harm that results from it). Take, for example, the following case described by Cassandra Byers Harvin (1996) and analyzed by Dotson (2011) as a case of testimonial smothering. In “Conversations I Can’t Have” (1996), Harvin describes an encounter in a public library with a white woman who asked her what she was working on, and when Harvin answered that she was researching raising black sons in the United States, the white woman promptly replied, “How is that any different from raising white sons?” Harvin explains that the question, as well as the tone, gave her the distinct impression that her interlocutor thought that she was “making something out of nothing” (1996, 16). This clearly had a silencing effect on Harvin, who tells us that she politely pretended that she was running out of time and exited the conversation. 7 As Dotson puts it, “testimonial smothering [. . .] is the truncating of one’s own testimony in order to ensure that the testimony contains only content for which one’s audience demonstrates testimonial competence” (2011, 244).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
180 José Medina and Tempest Henning As Dotson explains, we can understand Harvin’s interlocutor as performing a micro-invalidation.8 Harvin could have protested the m icro-invalidation with facial expressions that gave testimony of her discomfort and her taking issue with the misplaced distrust—for example, by rolling her eyes or opening her eyes widely in defiance. Such bodily testimony could create the kind of communicative resistance and, if supported by others, the kind of communicative solidarity that can nullify the micro-invalidation, or at least mitigate the harm that results from it in terms of feeling communicatively inhibited. Harvin’s non-verbal reactions could have contributed to neutralizing the micro-invalidation, but of course, this would be communicatively effective only if there were subjects present who could understand Harvin’s interlocutor as performing a micro- invalidation. When alternative hermeneutical and testimonial sensibilities that depart from the narrow epistemic sensibility (or closed-mindedness) that triggers the testimonial smothering are present, such maneuvers of communicative resistance become possible. And note that communicative resistance to testimonial smothering and other forms of silencing present in micro-aggressions becomes available not only for the victims of such epistemic aggressions, but also for others present in the communicative exchange, for other participants as well as bystanders can also engage in what Medina has called epistemic micro-resistance (Medina 2017) to neutralize micro-aggressions. Offering the kind of bodily testimony that registers and reacts to the testimonial smothering is something that Harvin herself could have done, but also something that other participants in the exchange (or even bystanders or eavesdroppers) could have done to call out the intimidation, express solidarity, and pave the way to different communicative dynamics. But, again, this is only possible if the proper testimonial sensibility attuned to body language is available among participants in the communicative exchange. As this example can also illustrate, a communicator’s capacity to use the body to convey particular messages and achieve particular speech acts and an interlocutor’s ability to give proper uptake both depend on the embodied positionality of communicators. The racialized embodiment of the participants in the exchange that Harvin describes matters deeply. Just as it matters that Harvin’s interlocutor was white in order to put in context and in perspective the kind of challenge she was raising, the racialized embodiment of the communicator also matters for those who try to counter the micro-invalidation and engage in micro-resistance, whether verbally or non-verbally. This is not to say that white people could not have engaged in micro-resistance and expressed with their body language communicative support for Harvin. But their micro-resistance becomes specific (in its content and force) through their racialized embodiment, which is to say that a 8 As Dotson explains it, the micro-invalidation performed is “characterized by communications that exclude, negate, or nullify the psychological thoughts, feelings, or experiential reality of a person of color” (Dotson 2011, 247).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Bodily Testimony and Epistemic Injustice 181 white person rolling her eyes at Harvin’s interlocutor would be expressing the support of a white perspective that distances itself from the white perspective of the micro-invalidation. By contrast, a Latina rolling her eyes at Harvin’s interlocutor would be expressing communicative support for Harvin from a non-white, but also non-black perspective, which is also different from the specificity of the non-verbal resistance that black bodily language can exert. The proper hermeneutical and testimonial sensibility communicators need to have in order to give proper uptake to racialized, non-verbal testimonial expressions requires that they pay attention to the embodied positionality of participants in communication and to the racial dynamics their embodiment triggers. Different contexts contain different communicative dynamics and they impose different communicative burdens on participants. The questioning of Harvin’s interlocutor would have conveyed something quite different and would have had very different effects if it had not taken place in an informal setting in which the stakes are low. Think, for example, of the same critical questioning in a context in which people are expected to raise challenges and are entitled to demand justifications, perhaps because everything in those communicative contexts (such as many philosophical and scientific contexts) can be treated as a hypothesis or a claim in need of justification. In those contexts, the micro-invalidation of Harvin’s interlocutor would have to be resisted with different kinds of communicative maneuvers, those that include, for example, shifting the epistemic burden to the interlocutor so that she explains why she operates with the assumption that the education of children of all races in the United States is the same. In these more formal and regulated communicative contexts, bodily language can also be used, and in fact it is very often (perhaps unavoidably) used, even if unnoticed. And when in those contexts there are racial dynamics that become inhibiting, marginalizing, or silencing, bodily language can also be used to call them out and to resist them. The proper hermeneutical and testimonial sensibility communicators need to have in order to give proper uptake to racialized non-verbal testimonial expressions requires that they pay attention to the embodied positionality of communicators and to the racialized communitive cues and communitive dynamics as they surface and operate in particular contexts of communication. It is often the case that when women and people of color give bodily testimony about their experiences of oppression or marginalization, they receive adequate uptake only by similarly situated women and people of color who have had similar experiences, and not by male audiences or white audiences, whose members may not even register the bodily testimony as having specific communicative content or force at all, or may misconstrue it in their semantic or epistemic appraisal of what it conveys. Let us consider, from this angle, the case study that Karen Jones offers in her influential essay “The Politics of Credibility” (2002). The case study concerns the story of Fauziya Kassindja, who fled her native Togo in 1994 to escape a forced polygamous marriage and female genital mutilation. Judge
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
182 José Medina and Tempest Henning Donald Ferlise of the Immigration and Naturalization Service didn’t believe Kassindja’s story and he denied her application for asylum in 1995, stating: “I have taken into account the lack of rationality, the lack of consistency and the lack of inherent persuasiveness in her testimony and have determined that this alien is not credible” (quoted in Jones 2002, 158). As Jones notes in her analysis of this case, a German woman as well as a Nigerian man did believe Kassindja’s story and helped her in her journey before she appeared in front of Judge Ferlise. In her analysis, Jones offers good reasons to think that Judge Ferlise was epistemically irresponsible in his appraisal, which was guided by an arbitrary and antecedently established distrust: “Ferlise was known not to be sympathetic to female asylum seekers” (Jones 2002, 159). As Jones points out, in asylum petition cases, a judge is charged with appraising “the truth of stories” told by petitioners, not the truth of and consistency among individual sentences; and an initial low assessment of the petitioner’s trustworthiness affects how the story is heard and appraised: “Distrust puts in place a suspicious cognitive set that colors how we will interpret the words of another” (Ibid.). And, of course, distrust colors not only the reception of the testifier’s words, but also the reception of her body language and bodily testimonial expressions. Jones’ analysis focuses on the words of Kassindja’s testimony and on the content they express, but this analysis could be fruitfully expanded by paying attention also to the body language, as well as the bodily aspects of Kassindja’s speech. Note that the book co-authored by Kassindja on which the case study is based is entitled Do They Hear You When You Cry? (Kassindja and Bashir 1999). As the title of the book suggests, one’s embodied emotional expressivity can become a communicative and epistemic handicap for having one’s story properly heard and believed. Expanding Jones’ analysis of Kassindja’s testimony to include bodily communication can shed further light on why Judge Ferlise deemed her irrational, inconsistent, and unpersuasive, while others (specifically women and African men) who heard her testimony did not. Judge Ferlise seemed to lack an adequate testimonial sensibility to hear, interpret, and appraise both Kassindja’s verbal testimony and her bodily testimonial expressions. More importantly (and here is where Ferlise’s epistemically irresponsible behavior as a judge lies), he made no effort to acquire such (verbal and bodily) testimonial sensibility or to rely on those who have it. As an Immigration and Naturalization Services (INS) judge in charge of hearing asylum petition cases, Judge Ferlise had an epistemic obligation to familiarize himself with the emotional expressivity and body language of asylum seekers from different parts of the world before hearing their stories, or to rely on experts who can properly interpret such bodily testimonial expressions, just as a judge must resort to translators or experts of the relevant verbal language. Judge Ferlise’s epistemic irresponsibility is demonstrated by his inattentiveness to the specific bodily communication to be expected from female victims of traumatic experiences, but also by his inattentiveness to the cultural
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Bodily Testimony and Epistemic Injustice 183 specificity of Kassindja’s embodied testimonial expressions. Different hermeneutical and testimonial sensibilities with respect to bodily expressivity are typically available to different cultural and racialized groups. For this reason, when an out-group member is the recipient of testimony that contains bodily testimonial expressions, including crying and bodily expressions of distress, such recipient, in order to issue epistemically responsible assessments, must take into account the cultural specificity of the body language in question and must make sure that she has the appropriate bodily testimonial sensibility to interpret such body language. This form of epistemic responsibility in the appraisal of bodily testimony (and of the bodily aspects of verbal testimony) is heightened in cases where the stakes are high and/or the recipient of bodily testimony has been invested with authority to issue an official appraisal of the testimony, as was the case with Judge Ferlise. Lack of proper uptake of bodily testimony is worrisome because it can be communicatively and epistemically limiting. Moreover, it can amount to a significant communicative and epistemic disadvantage when there is a widespread inability to give uptake in the communicative environment in question. In such a communicatively impoverished environment, a testifier’s ability to give bodily testimony and to be epistemically appraised fairly can become severely constrained. But even under those adverse circumstances, there can be opportunities for the use of bodily communication in resistant testimony among non-mainstream subjects, a resistant bodily testimony that can go unnoticed by mainstream subjects who lack the relevant, bodily testimonial sensibility. In some cases, bodily communication can be a safer and more selective way of giving risky testimony than verbal communication, for communicators can choose to share experiences and testimonial information with those who have a particular kind of sensibility with respect to body language, while their communicative acts remain hidden from those who don’t. In this sense, in resistant bodily communication we can find something analogous to the resistant “double-talk” that queer theorists have identified in queer speech or discourse that contains a “subtext” that is only processed by those who understand non-dominant sexualities and to which a heterosexist sensibility is insensitive (see Sedgwick 2008). The expressivity of the body can offer a powerful way of giving testimony to circumvent verbal limitations when people cannot talk openly and safely about certain things. For example, in cases of racial testimonial smothering such as Harvin’s being communicatively inhibited at the library by her interlocutor’s negative attitudes, the speaker being smothered can communicate with her eyes and facial expressions to those with the right testimonial sensibility (i.e. sensitive to racial testimonial smothering) that she has encountered a conversational obstacle and has to exit the conversation. Communicators attuned to communicative resistances such as that of Harvin’s interlocutor can use bodily language to signal to each other that they are in an adverse communicative climate, warning each other about “conversations they can’t have” in that context, to use Harvin’s phrasing; and they often can do so
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
184 José Medina and Tempest Henning without the perpetrators of micro-aggressions of this sort noticing that their actions are being identified as creating a testimonial dysfunction. Converging with the subversive communicative possibilities that the body offers as highlighted by queer theory (Sedgwick 2008), in section 4 we will explore how the racialized body can be used in resistant testimony and in forging communities of resistance against communicative and epistemic injustices. In order to illustrate how non-verbal communication can be used in testimonial dynamics at the group level, it is instructive to see how demonstration techniques have been used in the Black Lives Matter movement to give bodily group testimony about racial profiling and mistreatment by the police.
4. Collective Bodily Testimony There is nothing more singular and individualized than a human body, and yet our bodies are interconnected, find themselves in similar predicaments, and undergo similar experiences. Our account of bodily testimony connects with, and contributes to, the recent discussions of collective testimony and of group-level testimonial injustices (Lackey 2014; Brady and Fricker 2016). In our discussion of bodily group testimony, we will remain neutral as to whether group testimony should be reducible to the individual testimonies that compose it (i.e. as the proper summation of such testimonies), or whether it should be understood in a non-reductionist and non-aggregative way so that the group itself is understood to be the bearer of the testimony and to have testimonial agency over and above the particular testimonial powers and activities of its individual members. Interestingly, bodily group testimony poses important challenges to both summative and non-summative accounts of group testimony. On the one hand, summative approaches would have to explain how it is that a bodily gesture becomes iconic within a social movement and acquires a communitive life of its own independently of the particular communicators enacting the gesture at any given time. But, on the other hand, non-summative approaches would have to explain how collective bodily testimony can have communicative and epistemic force over and above the expressive and testimonial powers of the particular bodies giving testimony. Bodily group testimony seems to be a case in which individual and collective aspects of testimony become inextricably tied in complex ways. Think, for example, of bodily gestures of group testimony used in Black Lives Matter demonstrations protesting against police violence and giving collective testimony of how unarmed black people are targeted by police: for example, the emphatic raising of hands above one’s head. The hands-up gesture echoes Michael Brown’s own bodily gesture before being shot dead in Ferguson. It is a gesture that demonstrators use to express solidarity with Michael Brown and other victims of police violence, calling special attention to those who lost their
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Bodily Testimony and Epistemic Injustice 185 life to police brutality and are no longer around to give testimony, but can lean on demonstrators who lend their bodies as a vehicle of communication. The handsup gesture is used to express shared experiences of racial profiling and mistreatment by the police, and its content can be captured by the following statement: “I too have experienced being perceived as a threat while posing none, I too have to walk with my hands up.” In rising their hands above their heads emphatically as a gesture of compliance that gets no proper uptake by the police, Black Lives Matter demonstrators are bearing witness to a trend, a pattern, and not just to their own (or anyone else’s) personal case of being targeted by the police, despite their emphatic statement of being unarmed and posing no threat. When a group bears witness to collective patterns or trends, bodily collective testimony seems to have an irreducible supra-individual quality that seems lost in accounts that would simply reduce it to individual instances that can be treated and fully understood in isolation from one another. In this sense, bodily collective testimony poses a challenge for reductionist accounts. It is not clear how summative approaches can show that from the summation of individual bodily testimonies the collective dimension of the testimony will emerge. On the other hand, it is also not clear how non-summative approaches can explain that, in bodily group testimony, individual gestures echo one another and their specificity cannot simply be subsumed in a generalization that loses sight of the embodiment of individual communicators. So, bodily testimony poses challenges for all adequate accounts of group testimony. At the very least, such accounts need to do justice to: (i) the shared, interpersonal communicative dynamics linking these bodily expressions, so that each bodily testimony is no longer understood in purely individualistic ways, but as a part of a cluster or network of expressive bodies that work together in communicative cooperation; while at the same time making sure that (ii), the specificity of the individual embodiment of communicators is not lost in a generalization that makes all instances of group testimony equivalent and interchangeable. In this sense, bodily group testimony calls at least for refinements of both summative and non-summative accounts so that they incorporate ways in which the generic and the specific, as well as the collective and the individual, aspects of embodied expressivity are preserved in those accounts.9 Bodily group testimony does not seem to be detachable from the particular bodies of the individuals offering testimonial expressions. White demonstrators can (and do) participate in raising their hands above their heads as a gesture of surrender, but they cannot convey the same message with their hands-up gesture as non-white demonstrators can. When enacted by white demonstrators, the 9 A good candidate for doing justice to these different crucial aspects of bodily group testimony is Lackey’s (2014) view, for which group testimony is non-summative and nonetheless reducible to the testimonial powers of individual communicators, such as representatives or spokespersons.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
186 José Medina and Tempest Henning hands-up gesture can be used to denounced that some citizens have to go out of their way to show that they pose no threat because they have been criminalized and they face arbitrary police surveillance and excessive police violence in their lives. What the hands-up gesture enacted by white demonstrators cannot add to the collective testimony is the racially biased dimension of being targeted by the police. Of course, white demonstrators can bear witness to their own experiences of encountering arbitrary police surveillance and excessive police violence, but not because of their race.10 More often than not, though, the hands-up gesture of white demonstrators simply expresses solidarity with those who have been unfairly targeted by the police while posing no threat, without making any claim about themselves: “I am here standing with my hands up in solidarity with those who are violently targeted while surrendering.” It is only in the case of non-white demonstrators that the hands-up gesture of Black Lives Matter protests can both express a personal experience of being exposed to police racial profiling and at the same time echo the experiences of others who have been racially profiled by the police, being thus a hybrid case of personal and group testimony. It is only demonstrators of color who can turn the hands-up gesture into a bodily group testimonial act of racialized police misconduct. Therefore, in this case of bodily group testimony, the expressive embodiment of people of color is irreplaceable; the communicable contents and performative force of the bodily group testimony would be different without those particular racialized bodies enacting the gesture. This suggests that at least in some cases of bodily group testimony it matters deeply which particular embodied subjects give expression to the bodily testimony. A demonstration that tries to denounce racial profiling of people of color through the enactment of certain gesture must contain demonstrators of color who enact the bodily testimonial acts, and such enactments by people of color must be given center stage. In cases of bodily group testimony like this one, there are constraints on the representativeness of individual members of a group and on their capacity to act as a spokesperson for the group. The idea that group testimony can be detachable from the members of the group and is capable of being expressed by individuals outside the group becomes particularly problematic in cases of bodily group testimony in which the embodiment of the testifier does matter. When bodily group testimony is about embodied experiences that members of the group share, such as being targeted by the police because of one’s racialized body, out-group members can verbally describe such testimony, but they cannot themselves, through their bodies, make contributions to the bodily group testimony as in-group members can. For, when out-group individuals try to do so, their bodily testimony travels outside the group: white people can, of course, use the emphatic hands-up gesture to give 10 The hands-up gesture of white demonstrators can mean, “I too have experienced being perceived as a threat while posing none, I too have to walk with my hands up,” but not “because of my race.”
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Bodily Testimony and Epistemic Injustice 187 testimony about being abused by the police, but not because of their racialized embodiment (the racial component that attaches to group affiliation vanishes). There are interesting limitations on how spokespersons can participate in collective testimony that are suggested by bodily group testimony. As our example shows, white demonstrators cannot be spokespersons for the bodily group testimony of Black Lives Matter because their embodiment does not match that of those targeted by the police because of their race, and embodiment here matters deeply for what is conveyed in the bodily group testimony. More generally, out-members cannot be spokespersons for the group in bodily group testimony when and because their distinctive embodiment is not apt for communicating the content and force of the testimonial acts in question.11 The study of bodily group testimony may also call for further refinements and specifications in other areas of the epistemology of testimony, for example, in the assessment of the reliability of testimony. Communicators often assess not only the reliability of the specific testimonial contents conveyed, but also that of the person or group giving testimony. Assessments of the reliability of testimonial contributions, as well as assessments of the reliability of the individuals and groups who are the bearers of testimony, should pay attention to the body as an indicator of reliability. In face-to-face communication, the bodily aspects of the testimony matter deeply for reliability assessments. And when it comes to bodily group testimony, the perceived reliability of the group (or lack thereof) prompted by bodily cues can facilitate or impede fair communicative dynamics. This is a reason why an analysis of bodily group testimony is crucial for understanding epistemic injustices and how to resist them. Because it is not easily detachable from the bodies used as vehicles of communication, the bodily testimony of a group seems to be more resistant to being ventriloquized and coopted by out-group members than the verbal testimony of in-group members. Collective bodily testimony offers a powerful source of in-group solidarity and a kind of testimonial visibility that gives center-stage to in-group members and makes them irreplaceable in testimonial dynamics, thus making it more difficult for testimonial marginalization to occur. More research is needed in social epistemology to explore how bodily testimony can contribute to building communities of resistance that can fight effectively against epistemic 11 In this sense, bodily collective testimony can pose a challenge even for the most cleverly designed approaches to group testimony such as Lackey’s (2014). Although on Lackey’s view, a group’s testimony is not to be understood as formed through aggregation, the group’s testimony can be reduced to that of one or more individuals, who are the source of that knowledge in question and act as spokesperson. In Lackey’s view, when one acquires knowledge from the testimony of a group, the group’s testimony is reducible to that of individuals who act as spokespersons who may or may not be members of the group in question. Despite the general plausibility of this view, there can be constraints on who can communicate on behalf of the group in cases of bodily group testimony when the embodiment of in-group members is necessary for the bodily testimonial expressions to express the content and force they are intended to have.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
188 José Medina and Tempest Henning injustices. As Charles Mills (2017) has pointed out,12 collective testimony puts issues of epistemic injustice in a new light; and we have further suggested that collective bodily testimony introduces further complications in epistemic group dynamics and brings to the fore thus far unexplored possibilities both for e pistemic oppression and for epistemic resistance.
5. Conclusion Many have argued that non-white bodies are negatively charged and credibility deficits attach to them. Our reflections on bodily testimony in this chapter have gone beyond this point, trying to establish that, under conditions of racial oppression, all racialized bodies (non-white as well as white), far from being neutral, are epistemically valued in different ways and receive very different kinds of communicative uptake, if they receive any at all. The analysis and evaluation of these different kinds of uptake give support to two theses. In the first place, we have argued that the differential valuing of racialized bodies and their bodily testimonial expressions results in epistemic injustices concerning the precariousness of the intelligibility and credibility of non-white bodily testimony. Second, we have argued that bodily communication can be used in resistant testimony and that bodily group testimony is particularly well suited for creating in-group communicative solidarity and for giving center-stage to in-group members in testimonial dynamics. We suggest that bodily communication and body language can provide (i) powerful ways of giving testimony to circumvent verbal limitations when people cannot talk openly and safely about certain things; and (ii) powerful critical tools for resisting epistemic oppression and for creating communicative solidarity against epistemic marginalization and epistemic injustice.
References Bar On, Bat-Ami (1993). “Marginality and Epistemic Privilege,” in Linda Alcoff and Elizabeth Potter (eds), Feminist Epistemologies. New York: Routledge, 83–100. Brady, Michael and Miranda Fricker (eds) (2016). The Epistemic Life of Groups: Essays in the Epistemology of Collectives. Oxford and New York: Oxford University Press. Code, Lorraine (1993). “Taking Subjectivity into Account,” in Linda Alcoff and Elizabeth Potter (eds). Feminist Epistemologies. New York: Routledge, 15–48.
12 Mills has argued that the marginalization of the collective testimony of oppressed groups shows how testimonial injustices and hermeneutical injustices become intertwined: “in at least some instances of group domination it will not be so easy to conceptually separate testimonial and hermeneutical injustice, since a case can sometimes be made that the hermeneutical problems come from the collective ‘testimony’ of the [oppressed group]” (2017, 251).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Bodily Testimony and Epistemic Injustice 189 Dotson, Kristie (2011). “Tracking Epistemic Violence, Tracking Practices of Silencing.” Hypatia 26(2): 236–57. Fricker, Miranda (2007). Epistemic Injustice: Power and the Ethics of Knowing. Oxford and New York: Oxford University Press. Fricker, Miranda (2016). “Epistemic Injustice and the Preservation of Ignorance,” in Rik Peels and Martijn Blaauw (eds), The Epistemic Dimensions of Ignorance. Cambridge: Cambridge University Press, 160–77. Grosz, Elizabeth (1993). “Bodies and Knowledges: Feminism and the Crisis of Reason,” in Linda Alcoff and Elizabeth Potter (eds), Feminist Epistemologies. New York: Routledge, 187–216. Harding, Sandra (2004). “Introduction: Standpoint Theory as a Sit of Political, Philosophic, and Scientific Debate,” in Sandra Harding (ed), The Feminist Standpoint Theory Reader: Intellectual and Political Controversies. New York: Routledge, 1–16. Harding, Sandra (1993). “Rethinking Standpoint Epistemology: ‘What Is Strong Objectivity?”, ’ in Linda Alcoff and Elizabeth Potter (eds), Feminist Epistemologies. New York: Routledge, 49–82. Harvin, Cassandra Byers (1996). “Conversations I Can’t Have.” The Progressive Woman’s Quarterly 5(2): 15–16. Hume, David (1993). An Enquiry concerning Human Understanding, ed. Eric Steinberg. Indianapolis, IN: Hackett Publishing Company. Jones, Karen (2002). “The Politics of Credibility,” in Louise Antony and Charlotte Witt (eds), A Mind of One’s Own: Feminist Essays on Reason and Objectivity. New York: Westview Press, 154–76. Kassindja, Fauziya and Layli Miller Bashir (1999). Do They Hear You When You Cry? New York: Delta. Lackey, Jennifer (2014). Essays in Collective Epistemology. Oxford and New York: Oxford University Press. Lackey, Jennifer (2008). Learning from Words: Testimony as a Source of Knowledge. Oxford: Oxford University Press. Medina, José (2017). “Epistemic Injustice and Epistemologies of Ignorance,” in Paul C. Taylor, Linda Martín Alcoff, and Luvell Anderson (eds), The Routledge Companion to Philosophy of Race. New York: Routledge, 247–60. Medina, José (2013). The Epistemology of Resistance: Gender and Racial Oppression, Epistemic Injustice, and Resistant Imaginations. New York: Oxford University Press. Medina, José (2018). “Resisting Racist Propaganda: Distorted Visual Communication and Epistemic Activism.” The Southern Journal of Philosophy 56: 50–75. Medina, José and Matt Whitt (eds) (2021). “Epistemic Activism and the Politics of Credibility: Testimonial Injustice Inside/Outside a North Carolina Jail,” in Heidi Grasswick and Nancy McHugh (eds), Making the Case: Feminist and Critical Race Philosophers Engaging Case Studies. Albany, NY: SUNY Press, 293–324. Mills, Charles (2017). “Ideology,” in Ian Kidd, José Medina, and Gaile Pohlhaus Jr (eds), The Routledge Handbook of Epistemic Injustice. London: Routledge, 239–64.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
190 José Medina and Tempest Henning Pohlhaus Jr, Gaile (2014). “Discerning the Primary Epistemic Harm in Cases of Testimonial Injustice.” Social Epistemology: A Journal of Knowledge, Culture and Policy 28(2): 99–114. Sedgwick, Eve Kosofsky (2008). The Epistemology of the Closet, 2nd edn. Los Angeles, CA: University of California Press.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
PART 5
E PIST E MOLO G Y, R AC E , A N D T HE AC A DE M Y
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 19/03/21, SPi
9
The ‘White’ Problem American Sociology and Epistemic Injustice Charles W. Mills
“How does it feel to be a problem?” (Du Bois, in Sundquist 1996a [1903], 101). This question—from the opening paragraph of the first and most famous chapter (“Of Our Spiritual Strivings”) of W. E. B. Du Bois’s most celebrated book, his 1903 The Souls of Black Folk—has classically epitomized the white American attitude towards black Americans from the postbellum period onwards. No longer enslaved, they were now somehow, through the promise of the Thirteenth, Fourteenth, and Fifteenth Amendments, to be incorporated as full members into the body politic, a body politic historically conceiving of itself as “a white man’s country.” Yet (possibly non-coincidentally) more than a century-and-a-half after the Civil War, the polity remains a racially divided one. The “Negro problem”—formally so designated in the subtitle of Gunnar Myrdal’s (2017 [1944]) massive 1300-page-plus, Carnegie Corporation-commissioned 1944 study, An American Dilemma: The Negro Problem and Modern Democracy, which was seen at the time and for many years afterwards as the definitive work on the subject—continues under a different designation to afflict the nation today. In this chapter, I will argue that this foundational (mis-)framing of the issue has had negative epistemic consequences not merely for popular consciousness, but, as just cited, the academy also. In the formal sociological treatment of race in the United States, certainly historically, and to a certain extent even now, we have exemplified for us an epistemic injustice on a scale so large that in a sense it shatters the boundaries conventionally assigned to the concept. Indeed, the mis-cognition involved is not at all confined to a single subject, but historically manifest in many other disciplines of the American academy also: for example, anthropology, criminology, history, political theory, and even international relations. As black American (and some progressive white American) theorists of the socio-political order have always insisted: to the extent that there is a Negro problem, it has to be contextualized within the larger structural matrix of the white problem (Ebony 1965). But the failure to recognize white oppression as the environing and shaping causal background has necessarily ramified detrimentally throughout the formal study of American society, a “white racial frame” (Feagin 2020) that has misoriented inquiry from the start. So though—for reasons of space—I Charles W. Mills, The ‘White’ Problem: American Sociology and Epistemic Injustice In: Applied Epistemology. Edited by: Jennifer Lackey, Oxford University Press (2021). © Charles W. Mills. DOI: 10.1093/oso/9780198833659.003.0009
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
194 Charles W. Mills will be focusing on sociology here, it needs to be appreciated that comparable indictments could be made of the historical development of many other subjects of the American university as well.
1. Ideology and Epistemic Injustice Miranda Fricker’s (2007) Epistemic Injustice: Power and the Ethics of Knowing has become one of the most influential interventions of the past quarter-century in any subfield of philosophy. If “social epistemology” as it was born and developed in the 1990s within mainstream epistemology largely ignored the non-mainstream insights on the subject coming from feminist and (earlier) Marxist perspectives on social cognition, Fricker impressively succeeded in bringing them into the debate. Indeed, so successful has her intervention been that the term (“epistemic injustice”) has now passed into routine disciplinary usage, often without any reference being made to Fricker herself. Moreover, unlike most philosophical innovations—limited to a strictly in-house impact, of no interest to the outsider—the concept lends itself to application across a wide range of other subjects. After all, knowing and not-knowing, getting it right and getting it wrong, are of necessity integral to all human enterprises, and hence of pressing concern to everybody. Thus, the recently published Routledge Handbook of Epistemic Injustice (Kidd et al. 2017) has chapters in the “Case Studies” section on such areas as law, digital environments, science, education, medicine and health care, mental illness, anthropology, and religion. However, with two or three exceptions (see the chapters on anthropology [Tsosie 2017] and on Eurocentrism in philosophy [Alcoff 2017]), the implications of a possible foundational epistemic injustice in the disciplines themselves have not, so far as I can glean from the book’s survey of contemporary literature in the field, been the subject of that much inquiry. Yet if one thinks about it, the possibility of large-scale testimonial and hermeneutical injustice in this area follows straightforwardly from Fricker’s own methodological starting point. In ideal theory, by definition, systemic injustice, epistemic or political, is non-existent. Whether in an idealized cognitive sphere (virtuous epistemic agents giving each other reciprocal epistemic respect and adhering to defensible protocols for the fair and unprejudiced examination of alternative factual claims and conceptual schemes) or an idealized socio-political sphere (virtuous moral agents treating other members of the society and the polity ethically, within a framework of just institutions), reality and the normative largely coincide. Cognitively, the intrinsic difficulties and complexities attending all investigations of the natural and social world will, of course, be an obstacle to the discovery of facts and the development of concepts appropriate for understanding the workings of the section of reality in question. But in a socio-political order that by stipulation is non-exploitative, no group interests linked to the perpetuation or overcoming of
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
American Sociology and Epistemic Injustice 195 structural social injustice will be at stake in the affirmation of, or challenge to, dominant belief-systems and hermeneutical frameworks. To be sure, as in the scientific endeavor in general, adherents of competing theories and paradigms will have vested group interests of their own—prestige, career success, Ivy League ambitions, disciplinary recognition and honors, research funding, intellectual “sunk costs,” etc.—in their theory or paradigm proving to be the correct one. See, classically, Thomas Kuhn’s (2012) The Structure of Scientific Revolutions. And these factors, often not tracking objectively justifi able protocols, will doubtless frequently interfere with getting things right. But because under ideal conditions they will not be linked to larger social structures and institutions of illicit group privilege and disadvantage, their shaping role will tend to be limited in their power and confined to the academy. Moreover, even under non-ideal social circumstances, there will be cases where the discipline in question is investigating matters not empirical in the first place (pure mathematics) or, though dealing with the physical world, exploring subjects so remote from the possibility of human manipulation and application that confirmation/disconfirmation of the rival theoretical candidates in question would have few, if any, practical implications. See, for a good example, the current debate between physicists over “Superstring theory” vs “Dark matter” as cosmic explanations (CNN 2018). So the crucial subset of cases will be those where (i) significantly non-ideal socio-political circumstances obtain (if “ideal” is taken literally to mean “perfect,” then even minor deviations from perfection will imply non-ideality, so it needs to be specified that the divergence is major); and (ii) the discipline’s purview is the study of some section of the social or natural or (putatively) supernatural world in which the truth/falsity of different candidates in the range of possible findings/ theories would, in conjunction with the appropriate assumptions, have significant implications for the justice or injustice of the existing socio-political order. (The stipulated “in conjunction . . . ” clause is a recognition of the fact that suitable theory-saving Duhem-Quine adjustments in auxiliary assumptions can always block presumptively disconfirming implications.) My concern here, then, is with systemically oppressive societies—“ill-ordered,” in the coinage used by Carole Pateman and me (2007, 5), as against John Rawls’s (1999, 4–10) ideal “well-ordered” societies. Whether in terms of gender or class or racial or colonial domination (or multiple combinations thereof), rival understandings of the social order and the social world are in these cases generally going to be closely linked with rationalizations of, or challenges to, its injustice. As such, factors outside of objective inquiry will play a critical role in determining which understandings usually prevail, since the group interests involved are not at all limited to the conflicting ambitions of the rival research teams of the investigative community but will extend to the constituent groups of the broader social system. If the society is characterized by male domination, for example (as Western societies of the past few millennia have generally been—bracketing the
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
196 Charles W. Mills contested question of the possible gender equity of hunter-gatherer societies), then how could the issue of female capability not be a “political” one? So, as indicated earlier, it will not just be the social, but sometimes the natural and (putative) supernatural, worlds that will also be the subjects of high-stakes, politically imbricated dispute. Think of biological determinism, for example, whether in its early pre-modern or more developed Enlightenment forms—the “scientifically” proven natural biological inferiority of women, the subordinate classes, and non-white races and colonized populations (Gould 1996; Saini 2018). Or, for ages less secular than our contemporary West, the role of religion, for example the Christian Church’s intellectual hegemony over and ratification of the hierarchies of the medieval order, extending for some believers well into modernity (in the famous words of the 1848 hymn “All Things Bright and Beautiful”: “The rich man in his castle/The poor man at his gate/God made them high or lowly/And ordered their estate”), not to mention Christianity’s abysmal record—so well known it presumably needs no documentation—in proclaiming women’s divinely ordained subordination. In cases like these, the depth and extensiveness of epistemic obstacles is obviously far greater than under ideal conditions, being rooted not just in the assumptions and conceptual frameworks in competition but in the interests of differently positioned groups in perpetuating or reforming the structures of the social order. Fricker’s (2007, ch. 7) own famous original example of hermeneutical injustice was the difficulty women would have in understanding and appropriately condemning what was happening to them as sexual harassment before the formal discursive introduction of the concept. But depending on how deviant from ideality, how “ill-ordered,” the society in question is (sexist, class-dominated, racist, colonial . . . ), such lacunae or gaps will not be isolated singletons, but part of much more extensive problematic ideational systems. In other words (and this was my [Mills 2017] own contribution to the Epistemic Injustice handbook), the concept of ideology originally classically associated with the left tradition needs to be resurrected and brought into the epistemic injustice debates. Albeit with a different vocabulary, usually “Continental” rather than “analytic” (though in its brief heyday, “analytical Marxism” did produce a few works on the subject—e.g. Torrance 1995), and with different political commitments, socialist rather than liberal, left theory has long been exploring the relation between social domination and social cognition. Analyzing Ideologie and doing Ideologiekritik was, in different ways and with different presuppositions, a central concern of Marx and Engels, Frankfurt School Critical Theory, Louis Althusser, and other strains of Marxism. Across these variants, the common overlapping theme was the holistic view of ideology as a negative cognitive phenomenon permeating the social order, not at all a matter of a few problematic hermeneutical gaps or misleading concepts, but an encompassing worldview, a way of (mis)understanding
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
American Sociology and Epistemic Injustice 197 the social world (and—as emphasized—sometimes sections of the natural and supernatural worlds also) and one’s place in it. Inevitably, then, according to this analysis, ideology will affect not merely ordinary “common sense” but the bodies of abstract technical thought that are tasked with providing us with a more rigorous understanding of that world, and that eventually become formalized as disciplines within the academy. So, it was not remotely the case that Ideologiekritik was to be limited to everyday consciousness, folk psychology, and folk wisdom. Rather, the claim was—going back to Marx’s engagement with what he called the “bourgeois economists” of his day— that ideology was to be found at the highest intellectual levels, insofar as, in a modern class society, even what purport to be the most sophisticated and refined theorizations are negatively influenced by class interests and the “appearances” of capitalism, thus leading to the rationalization of the class order. See, for example, such venerable (but still worth reading) left-wing collections of the 1970s as Ideology in Social Science: Readings in Critical Social Theory (Blackburn 1972) and The Political Economy of Science: Ideology of/in the Natural Sciences (Rose and Rose 1976). Marxism has, of course, fallen into abeyance since then (though it is now showing signs of revival), and as Fricker (2017) points out in a recent retrospect ive look at the etiology of her “epistemic injustice” concept, it was always flawed by its one-dimensional class reductionism: “As regards Marxism, for my purposes the monolithic social ontology of class . . . remained at that time riskily insensitive to other dimensions of difference.” Hence the need for a multidimensional view of group oppression, as manifest in her own work and in the work of others impressed by the usefulness of her development of a concept designed to “mak[e] sense of the lived experience of injustice in how a person’s beliefs, reasons, and social interpretations [are] received by others, even conscientious well-meaning others” (56). But where does “ideology,” one of the “invaluable bold abstractions” she acknowledges as a contribution from Marxism (55), fit in this rethinking? By contrast with Foucault, the other major figure whose “long shadow” she mentions, the relation assumed between “truth” and “power” was far less ambiguous in Marxist theory (56), given its Enlightenment pretensions—however naïve they may seem to us now—of developing a “scientific” understanding of society and history. Moreover, insofar as Fricker’s retrospective essay stresses the need for “continued strictness with regard to the remit of (what [she is] now labelling) ‘discriminatory epistemic injustice’,” a “bounded and specific” category, not relaxed “to embrace the generality of unfair interpersonal manipulations or . . . systemic riggings of the epistemic economy” (53), “ideology” would seem to have a natural role in our discussions. While admittedly in the activist socialist press, the term was, and is, sometimes employed in such a way as to imply deliberate and self-conscious conspiratorial obfuscation and deceit—the ruling class’s attempt to
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
198 Charles W. Mills fool as many of the people as much of the time as they can—such usage was not the standard one in academic left circles. Here, the presumption is that a structural dynamic rather than bad faith is crucial. So, such a framing is completely in keeping with Fricker’s insistence for both testimonial and hermeneutical injustice (in a section of her essay titled “Theorising the unintended”) that “the absence of deliberate, conscious manipulation is definitive” and “the hearer who cannot understand because she lacks sufficiently shared concepts with the speaker . . . . is not deliberately manipulating, concealing, or blanking anything.” Rather, “the cause of the injustice is structural” (54). Epistemic injustice, at least in Fricker’s conception, is not the same as conscious intellectual fraud. What I want to look at, then, is the workings of racism as an ideology, drawing on my earlier Routledge Handbook essay on the subject (Mills 2017), and how it might challenge Fricker’s categories when it is so influential as to affect the d isciplines themselves, which are, of course, precisely the intellectually prestigious subset of what she calls the society’s “shared hermeneutical resources.” With the revival of feminist theory in the 1960s, for example, the task that many second-wave feminists who had managed to get into the academy set themselves was the examination of how the purportedly objective subjects of political theory, history, anthropology, social science, medicine, and so forth had been molded by androcentrism, by taking males as the paradigm subject. And similarly, of course, for revisionist work motivated by opposition to racism and Eurocentrism. In such cases, the commonality and neutrality of the hermeneutical resources may be put into question, insofar as what count as acceptable “resources” are profoundly shaped by conflicting adversarial conceptions of how we should understand the genders and the races in relation to the social order. So, it may not be a matter of a few discrete gaps or problematic categories rendering incomplete or mildly tainting hermeneutical resources otherwise acceptable, requiring only a sharply targeted epistemic surgical insertion or excision for a successful oper ation. Instead, a more complex intervention is likely to be required, as with a cancer that has metastasized, spreading to other organs. The hermeneutical deficiency does not stand alone, but is articulated through other connections with many other linked concepts, background beliefs, characteristic framings, and micro/ meso/macro-explanations that serve both to embed it in a global view and to protect it from disconfirmation and discarding. So, Ideologiekritik may involve challenging and rethinking an entire conceptual system, and instituting a new reconstructed cognitive regimen to correct its pervasive epistemic injustice.
2. Racism: Psychological and Structural Let us now look at racism. I suggest that the two important senses relevant for our discussion will be the mental/psychological (racism as sentiment and/or belief)
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
American Sociology and Epistemic Injustice 199 and the institutional/societal (racism as structural domination/illicit advantage) and their reciprocal interlinkage through the concept of ideology. Start with the first. Over the past two decades plus, a range of competing pos itions has been advanced in critical philosophy of race on what the best conceptual analysis of (psychological) racism would be. Somewhat simplifying the debate, a basic contrast has emerged between volitional/affective accounts (racism as ill will, whether racial hatred or racial indifference) and doxastic/cognitivist accounts (racism as a set of beliefs about natural racial hierarchy). Tommie Shelby (2002, 2003) and I (2003) have argued (separately) that a doxastic/cognitivist account is superior to a volitional/affective account (Garcia 1996), and, of course, such an analysis is obviously what is required for entering a debate about epi stemic injustice, given the subject matter of that injustice. I have suggested elsewhere that racism can be analyzed as the claims that (i) the human race is naturally divided into different races; (ii) some races are superior to others; (iii) the metrics of this superiority/inferiority include some, or all, of the cognitive, the characterological, the aesthetic, the spiritual, and the physical, with the first two probably being the most important and most consistently invoked; and (iv) the degree of superiority/inferiority in the area in question is sufficient to warrant differential treatment of the races. Depending on the background theoretical framework and causal picture being presupposed, one could distinguish supernatural, biological, and cultural varieties of racism, though hybrid syntheses of two or more are by no means ruled out. Shelby (2003) has further developed the case that racism should best be seen as an ideology, thereby (in the left tradition) linking beliefs to “material” group interests and social structures. Since I drew on his work for my earlier Routledge Handbook chapter (Mills 2017), and since I doubt if I can say it any better than I did there, I will now repeat my gloss of his position: Shelby works with the concept of belief-sets as forms of social consciousness, an epistemically neutral concept that allows for both ideological and non-ideological versions. (So being a form of consciousness is a necessary, but not sufficient, condition for being ideological.) Four features characterize forms of consciousness: (i) the beliefs are widely shared (as against limited to a few); (ii) they form a seemingly coherent descriptive and/or normative system (in networks of apparent explanatory or justificatory power); (iii) they shape group outlooks (e.g., as part of “common sense”); and (iv) they significantly impact social action, interaction, and institutions, constituting part of our “life-world.” So forms of consciousness as such are not intrinsically problematic, and in fact are obviously necessary for us to navigate the world. Ideological forms of consciousness, however—the epistemically negative subset—are marked by (i) epistemic deficiencies of various kinds, for example false factual or normative claims, or tendentious framings and theoretical misrepresentations even when
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
200 Charles W. Mills true claims are involved; (ii) problematic origins, that is belief-formation and norm-uptake by the epistemic agent for bad reasons; and (iii) functionality for the establishment (when not yet existing), or reinforcement (when already existing), of systems of social oppression, via the epistemically deficient aspect of these forms of consciousness. (Epistemically respectable sets of beliefs can also serve the interests of privileged groups, so the role of this feature needs to be specified to avoid committing a version of the genetic fallacy.) Ideologies thus serve to justify, rationalize, legitimize, and/or obfuscate wrongful social domination. And what makes this account a materialist one is the crucial explanatory role played in the genesis and reproduction of these ideolo gies by group power and group interests within socio-economic structures and conditions. Vested group interests provide a causal explanation for the endurance of ideologies even when the evidence clearly refutes them. (103–4)
As can be appreciated, then, ideologies in this sense—at least when linked to the social groups in power over others—will generate widespread patterns of epi stemic injustice, both testimonial and hermeneutical. It will usually be the case that the derogation of the subordinated as incompetent knowers will be (a non-contingent) part of the ideology, while its defining conceptual frameworks, whether at higher or lower levels of abstraction, will take for granted the “standpoint” of the hegemonic group. It is not a matter (to cite Fricker’s criteria) of individual bad faith on the part of the privileged when they find it difficult to take the testimony of the subordinated seriously, or to find implausible (or even incomprehensible) hermeneutics assuming their equality, but the social inculcation of a world view pervasively disseminated throughout the social order and functional for its reproduction. So, these patterns of aprioristic belief and corresponding aprioristic skepticism, of conceptual frameworks foundationally structured by assumptions of the legit imacy of the ruling group’s domination of the social order, will affect both commonsense ideation and scientific disciplines. In the case of racism, our focus, it will not at all be “cabined,” “siloed,” but pervasively influential across a whole range of fields. After all, we are dealing here with the basic category of both the humanities and the social sciences (and some natural sciences): the human. So, claims about the natural divisions between humans—their innate characteristics, their (supposedly) ineluctable drives, their potentials and limitations—will inev itably ramify throughout any subject that investigates them, necessarily shaping the trajectory of the particular discipline in question. If the proper study of mankind is man, then understanding the intrinsic differences between these “men” is obviously of the most critical importance. The appropriate way to think of racism, then, is not as individual “prejudice,” but as a widely disseminated theory in its own right, a theory of racially differentiated human capacity and incapacity, either itself normative or (depending on how one wants to draw the boundaries) with
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
American Sociology and Epistemic Injustice 201 normative implications, that simultaneously functions as a kind of meta-theory constraining the line of development of other theories. Finally, racism as theory/political ideology in this overarching ideational sense is linked to racism in the institutional/societal sense. For some nations, racist thought and practice may be limited to particular tainted institutions in an overall healthy body politic. But in the case of the United States (a self-conceived modern Western liberal democracy that was a racial slave society—if one begins from colonial America—for nearly 250 years, and then, and still now, a racially segregated society, whether de jure or de facto, for the next 150 plus years), the racial nature of the polity should not be in the least controversial. So, it is not merely that racism as ideology has been central to the functioning of the nation’s key institutions, but also that the society as a whole can be categorized as historically a racist one, established on the subordination of blacks and Native Americans. Beliefs about non-white inferiority along with racialized frameworks of understanding have been central to white-supremacist institutions and a white-supremacist polity, epistemic injustice rationalizing or obfuscating social injustice.
3. Racism and American Sociology Against this background, then, let us look at the development of American sociology in the late nineteenth–early twentieth centuries as a case history in epistemic injustice on the disciplinary scale. In a sympathetic critique of, and friendly amendment to, Fricker, Elizabeth Anderson (2012) has argued for a supplementary account of “Epistemic Justice as a Virtue of Social Institutions,” since individual strivings for testimonial and hermeneutical justice are likely to be inadequate where “massive structural injustice” is involved. So just as well-meaning gifts to charity will not remedy structural poverty, a problem calling instead for the “redesign [of] economic institutions,” so in some cases we may need “to reconfigure epistemic institutions so as to prevent epistemic injustice from arising,” since “epistemic [injustice may] be embodied at the level of global systems of inquiry” (171). Obviously, the academic disciplines themselves are paradigm examples of such “epistemic institutions.” Likewise, in the opening chapter of the Routledge Handbook, Gaile Pohlhaus Jr (2017), one of the three editors, mentions “academic disciplines” as possible sources for “distorting understanding and stymieing inquiry” (13), where “not only individual vice” is at play but also the vices of “epistemic institutions” (17). What I am trying to do, then, is to illustrate this actuality with the real-life example of early American sociology (though some contemporary critics might reply that this temporal restriction is far too charitable). The study of society has in a sense always existed, but for the standard Western periodization it is traditionally in the nineteenth century that modern sociology
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
202 Charles W. Mills as a “science” of society is deemed to be born, whether in the early positivism of Auguste Comte, later figures like Max Weber and Emile Durkheim, or, for those with more radical sympathies, Karl Marx’s “materialist” conception of post-hunter-gatherer society as riven by class struggle, both synchronically and diachronically. By contrast with the Old World, though, the American “New World” was far more racially heterogeneous, so that the imperative of theorizing “race” was far more pressing. Manifest destiny had been fulfilled by the late nineteenth century, with the triumphant arrival of the Euro-settler population on the Pacific coast, and the final defeat of indigenous resistance. Not originally part of US society proper, native peoples would not formally be counted as citizens until the 1924 Indian Citizenship Act. But in the case of blacks, it was different. Here, one had a large population of millions of people, most of whom were ex-slaves, and unquestionably in some sense Americans, given at least nominally equal juridical status, as mentioned at the start, by what were known as the “Reconstruction” Amendments. Moreover, “race” was not at all a novel concept in the intellectual atmosphere, but from the colonial period onward, had been a central category of self- and other- American understanding both in everyday consciousness and elite thought (Jordan 2012; Smedley and Smedley 2018). So, the crucial questions were: How would a “science of society” be developed on American soil in these circumstances that would explain the workings of race? How would “race” itself be conceptualized—in a way that challenged, or in a way that ratified, pre-existing understandings? How would the differences between whites and blacks be framed? Whose theories would be allowed to contribute to the “common hermeneutical resources”? Which populations would be consulted as competent knowers and which would not? To what extent and in what ways would insights from the presumed neighbors of this new discipline be drawn upon? And above all, what overarching conceptualization of the United States as a nation would shape the entire research paradigm? Fortunately, we do not have to speculate about the answers to any of these questions, since we are looking back from the perspective of a century plus later. The story has already been told in two prizewinning books by American sociologists, Stephen Steinberg’s (2001 [1995]) Turning Back: The Retreat from Racial Justice in American Thought and Policy and, twenty years later, Aldon Morris’s 2015 The Scholar Denied: W. E. B. Du Bois and the Birth of Modern Sociology. Both books won the American Sociological Association’s (ASA) Oliver Cromwell Cox Award in their respective years for best book in anti-racist scholarship, so they are obviously particularly appropriate resources. I will also draw on Steinberg’s 2007 Race Relations: A Critique. (Interest disclosed: in writing this book, Steinberg himself drew on my “epistemology of ignorance” concept from my 1997 book The Racial Contract, so I hope this circle of mutual influence and citation will strike readers as virtuous rather than vicious.)
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
American Sociology and Epistemic Injustice 203 Let me now go through some of the possible obstacles to the achievement of epistemic justice in the development of this discipline to demonstrate how paradigmatically and systematically the history manifests them. Consider first the category of shared hermeneutical resources in the society at large. Even in its founding, a discipline does not begin ex nihilo, but from pre- existing elements deemed to be relevant (beliefs, concepts, frameworks, norms), however transformed, revised, and sophisticated they are going to be by the intended disciplinary construction. Yet, as emphasized earlier, the intellectual atmosphere of the United States of the late nineteenth century was deeply pervaded with racist assumptions, ideological “forms of consciousness” at both the scholarly and popular levels that had originally developed in the colonial period to justify and rationalize indigenous expropriation and African slavery. Moreover, anti-black racism would become particularly salient and consequential precisely because of the challenge of dealing with the newly emancipated millions of ex-slaves, in contrast to the diminishing significance (and smaller numbers) of indigenous peoples as components of the national population. Earlier in the century, polygenetic versions of “scientific racism” (multiple origins for humanity) had been very popular, above all in the American School of Ethnology, with support from the Swiss–American biologist and geologist Louis Agassiz. As Stephen Jay Gould (1996) has documented in his classic and prizewinning exposé, The Mismeasure of Man, the most important text produced by the school, Josiah Nott and George Gliddon’s 1854 Types of Mankind (2010), was not at all a fringe product, but an authoritative text regarded as the leading work of the time on the differences between humanity and subhumanity (a copy can be found online). Drawing on Samuel Morton’s craniometry, and including illustrations comparing blacks to chimpanzees and gorillas, the book gave scientific imprimatur to prejudices linking blacks with apes already long established in the white Western imagination. The ordinary white man and woman in the street, even if not themselves able to rehearse the technical proofs of African development from a separate branch of (sub)humanity, would at least be able to point with confidence at the learned Ivy League scholars who could. The post- 1859 (Origin of Species) Darwinian revolution would ultimately discredit polygenesis, which today survives only in peripheral racist circles. But scientific refutation of one variant of scientific racism by no means pre-empted the development of others. Social Darwinism, according to which some races— though of common human origin—were less evolved than others, closer to our ape ancestors (indeed hardly evolved from them), would become the new (monogenetic) racist scientific orthodoxy (Hawkins 1997). And once again, blacks were singled out as distinctively simian, their phenotype declaring their consanguinity (Hund et al. 2015). Indeed, some museum dioramas of the time actually located blacks as creatures in between apes and the fully human (whites). Reinforced by
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
204 Charles W. Mills the depictions of popular culture—the Jim Crow musical-comedy minstrel show, widespread cartoons of monkey- like blacks, caricatures in advertising—this iconography would become ubiquitous. And, to repeat, it had scientific backing from the canonized authorities. Were one to consult the 1910 Encyclopedia Britannica, for example, one would find an entry that explained the proven “mental inferiority” of “negroes” in comparison to whites as a consequence of “the premature closing of the cranial sutures,” that hindered further brain development (cited in Steinberg 2001, 30). Unsurprisingly, then, social Darwinist assumptions, as in Herbert Spencer’s work, would shape early (white) American sociology. In Morris’s (2015, 22, 115) diagnosis: Social Darwinism, embraced by the white founders of American sociology, claimed that society could be explained by natural laws . . . [A] hierarchy of races existed with superior races at the top, less superior ones in an intermediate position, and inferior ones locked at the bottom . . . Higher races inherited both superior genes and a socially transmitted superior culture . . . For social Darwinists, it was not possible to alter the racial hierarchy because it had been formed by unchanging natural laws under which a powerful natural selection process ensured the survival of the fittest.
Morris goes on to argue that Robert Park, who is standardly credited with being the father of American sociology of race through his “Chicago school” (i.e. the University of Chicago), was—contra the conventional wisdom—heavily influenced by social Darwinism in his famous “race relations cycle.” The four-stage cycle (competition, conflict, accommodation, and assimilation) was supposed to constitute a general pattern for understanding “race and ethnic relations in the modern world and . . . for humankind in the future” (116), “a predestined process of evolution through which racial groups must travel as they march to their ultimate destiny of becoming assimilated” (117). As Steinberg (2007, 49–57) points out, the very language used by Park euphemizes the realities of colonialism, conquest, racial slavery, and racial subordination. In Steinberg’s blunt summary: “We can say with only slight exaggeration that American sociology had its roots in an effort to provide erudite justification for racial hierarchy” (50). And Morris (2015, 117–18) adds that insofar as, for Park, “racial temperaments are key determinants differentiating racial groups,” shaping their culture and talents, he does indeed (contra the standard assessment) biologize race: For [Park], white Western Europeans, including American whites, were the superior races because they had developed sophisticated civilizations . . . Africa, by contrast, was populated by savages. Therefore, Africans and their descendants were different biologically and culturally from Europeans, with unique racial
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
American Sociology and Epistemic Injustice 205 temperaments and low levels of culture. Blacks occupied the bottom rung of the racial hierarchy because of crippling physical qualities and a lack of civilization. (115)
Moreover, a second point, such views were not at all confined to infant sociology. After all, if blacks were indeed the primary, or only, problem, then one would expect that their irredeemably problematic nature would manifest itself in other fields of study also. The neighboring disciplines—the putatively intellectually rigorous component of the society’s “collective hermeneutical resources”—were no more racially disengaged than sociology was. Far from being sympathetic friends of the defense, or at least detached bystanders, they were themselves allied with the prosecution. Anthropology, for example, was one of the first disciplines to be subjected to decolonial critique, unsurprising considering that it was a discipline forged directly out of the “colonial encounter” (Asad 1995 [1973]). Europeans had to make sense, within their mappings of the human, of the new creatures they were meeting in the colonial voyages, not merely for the goals of theoretical understanding but for the administration of empire. In other words, what would become “racial” taxonomies were not at all detached in-house academic exercises but guides to public policy. Indeed, in certain respects, anthropology and sociology overlapped, to the extent that understanding the “savage” in his native habitat would be essential to the likely success—if this daring venture was to be attempted— of integrating him into civilized, white society (Baker 1998). Criminology was likewise affected by the racialized understandings of the day, since the subject is not just about developing appropriate norms for punishment, but also about assessing the likelihood of certain populations being differentially disposed to be law breakers in the first place: “criminal races,” as black Americans were conceived of (Muhammad 2010). Hence the long history of the political use of the criminal justice system against African Americans, which has come to particular recent attention with Michelle Alexander’s book (2010), but which has always—if not to the same extent as today—prescribed special watchfulness for blacks. History might seem to the uninformed naïf to be just an account of crucial events. But understandings of these events, and what makes them crucial in the first place, will obviously be competing and contested. In the period, racialized theories of historical evolution were widespread, with some races seen as advanced and progressive (whites in general, but Anglo-Saxons in particular in the world of the “Anglosphere”), others as holding historical progress back (blacks in particular), and others as on their way out anyway (the category of “dying races,” as with Native Americans, though sometimes applied hopefully to blacks also in the postbellum period) (Horsman 1981). Political theory at best marginalized the fact of white political domination and at worst justified it (Smith 1997). Even international relations theory (IR) was affected, insofar as developing the best
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
206 Charles W. Mills strategy for dealing with the non-white races was seen as a global challenge for white elites across the planet (Vitalis 2015). The point is, then, that in societies characterized by deep structural oppression, hermeneutical obstacles will be far more extensive, multifaceted, and entrenched than a few missing concepts. The main axes of social subordination (here “race”) will act as powerful generators of fields of cognitive distortion. So “race” is in a sense crucial to understanding the problematic evolution of all these disciplines, but “race” with a very different signification than in mainstream conceptions. It is not merely that “race” has to be rethought, but that its linkages need to be reconceived. “Race” in the orthodox sense is present—indeed omnipresent—but in a guise refractory to permitting the development of an objective understanding of its referent’s actual place in the socio-political order. Its dominant (racist) conceptual representation is imbricated in a multidisciplinary body of theory, or a set of overlapping theories, predicated on natural racial superiority and inferiority. A progressive analysis of race trying to establish itself in this hostile intellectual arena thus has to struggle on multiple disciplinary fronts, in combative engagement with subjects not at all neutral, collegially joining in to assist the collective scholarly enterprise, but aggressively adversarial. Race as admittedly biological, but a biology of equality, or, more daringly still, race as not biological at all, but (in a contemporary vocabulary) a social construct, would be in fundamental conflict with contemporaneous framings. So, it is not just a matter of filling in a hermeneutical gap or targeting a particular misleading concept and finding a superior one to replace it, but of a theoretical revolution both in its conceptualization and the larger theoretical matrix in which that concept has been embedded, that reciprocally supports it. (As a useful analogy: think how revolutionarily enlightening at the time was second-wave feminism’s drawing of a conceptual demarcation between sex and gender.) What will therefore be required is an alternative theorization to compete with that of scholarly orthodoxy, necessarily entailing a corollary rethinking of constitutive frameworks, assumptions, norms, auxiliary micro- and meso-theories, and so forth—thus in crucial respects contesting the discipline itself, at least in its dominant incarnation. Third, as emphasized in section 2, the ideological is linked to the material. Social structures and institutions, particular configurations of group interests, constitute the material base of dominant-group ideologies, thereby serving to anchor them in oppressive socio-political systems whose resistance to reform is what in turn buttresses the ideologies’ resistance to disconfirmation. Slavery itself, of course, had been one such fundamental institution, its famous designation as “the peculiar institution” obfuscating its foundational role in the development of the United States and American capitalism in general (Baptist 2014). But the end of slavery did not at all mean the end of black racial subordination. Rather, after the betrayal of Reconstruction, a new system of racial exploitation was established in the South, complementing the less overt Northern variety. In Morris’s (2015, 6–7) assessment:
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
American Sociology and Epistemic Injustice 207 The majority of whites, both North and South . . . viewed [blacks] as an inferior race. Beyond ideology lay naked economic and political interests because southern white elites needed cheap labor akin to that provided by slaves if they were to remain a ruling aristocracy. During the late nineteenth and early twentieth centuries, white northern capitalists also required cheap labor to fuel the massive industrialization rapidly developing within the factory system embedded in American cities . . . The formal Jim Crow regime was hammered into place throughout the South during the last decade of the nineteenth century. Its purpose was to relegate blacks to the bottom of southern society, rendering them vulnerable to maximum exploitation . . . In the North, de facto racial segregation and widespread economic exploitation and political oppression differed only in degree from the southern Jim Crow regime.
Steinberg (2007, 53–7), citing work by Herman and Julia Schwendinger, and also R. W. Connell, would add that intra- and extra-continental imperialism needs to be recognized as another crucial “material” set of processes and institutional formations underwriting sociology’s social Darwinist affinities: [Today’s] mainstream sociologists fail to recognize the connection between these retrograde ideologies and the imperialistic projects of that period . . . Connell [observes] that the formative period of sociology was very much bound up with imperialist expansion. Contrary to what is generally assumed, the pioneers of sociology had a global perspective . . . What else was “manifest destiny,” if not a blueprint and ideological justification for imperial expansion from the Atlantic to the Pacific, including the annexation of a third of Mexico’s national territory? The end of the nineteenth century, when sociology was taking root as a discipline, was also a period of several overseas adventures, including the Spanish-American War, which led to the occupation of Cuba and the Philippines, followed by the annexation of Hawaii, Samoa, Guam, and Puerto Rico . . . Thus, Social Darwinism, with its hierarchy of higher and lower races, provided perfect justification [for these conquests]. (54–6)
As emphasized at the start, then: in unjust societies, rival understandings of the section of reality in question are going to be closely tied to rival prescriptions for the institutional creation and perpetuation, or reform and elimination, of unjust social structures. Think of pro-slavery views versus abolitionism, manifest destiny versus equitable treaty making with indigenous peoples, imperialism versus anti-imperialism, ongoing postbellum racial exploitation versus black incorporation on new, fair terms into the economy—in sum, racism versus anti-racism. So, the vested interests at stake will inevitably shape the contours of inquiry, generally resulting—if the power imbalance is unfavorable to progressive political reform— in massive ideological distortions, epistemic injustices on a macro-scale.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
208 Charles W. Mills Fourth, consider now another kind of “material” factor: restrictions on the class of respected cognizers. Anderson (2012, 165) points out that in cases of institutional epistemic injustice, “the larger systems by which we organize the training of inquirers and the circulation, uptake, and incorporation of individuals’ epistemic contributions to the construction of knowledge may need to be reformed.” If this may be true, even in the present more enlightened period, when racism has officially been repudiated, think how much more dramatically such reform would have been needed in the period we are considering! The derogation of blacks as reliable knowers is, of course, already baked into racism as a theory, primarily as cognitive defect (straightforward incapacity) but also as moral flaw, lack of virtue, slyness and dishonesty (a “vicious” unwillingness to tell the truth, even when known if personal or “racial” gain can thereby be achieved). So black testimony, insofar as it is ever solicited, will be aprioristically suspect. Correspondingly, the innate inferiority of “black minds” meant that their education could only be carried so far (Darby and Rury 2018). In the postbellum period, black education was, of course, generally segregated, and in the South in particular was radically underfunded, with the differential in resources going to white and black schools sometimes being as high as 10:1. Unsurprisingly, then, most African Americans were deprived of the training necessary to develop their cognitive skills (thereby circularly justifying—in white eyes—this allocation: “See? We told you how stupid they were”). Moreover, these barriers did not all diminish at the level of tertiary education. To begin with, of course, very few blacks would have been able to acquire the skills necessary for them to be viable candidates for such institutions in the first place. But in addition, patterns of discrimination at the university level meant that even fewer of this already tiny number would have been admitted to “white” institutions and a vanishingly small proportion of those would have had any chance of going on to graduate study. So, the training necessary to enable one to become a participant in the realm of academic discourse—and to develop, where called for, an intellectually sophisticated critique of that discourse—would have been out of reach for all but a handful. And even when the lucky and talented few received such a doctoral education, even from elite white institutions, the Jim Crow rules of the day meant that after graduation they would only have been able to get jobs at black institutions (historically black colleges and universities). If, in the academy in general, independent of race, prestige hierarchies always magnify and diminish, respectively, the voices of those teaching at elite and those consigned to lowly institutions, imagine how spectacularly this differential credibility gap is exacerbated by racial hierarchy. White scholars at white institutions versus Negro scholars (isn’t that oxymoronic?) at Negro institutions—whom are you going to believe? To say nothing of issues of higher teaching loads, lower pay, fewer resources, limited invitations to important conferences, and diminished access to the leading journals in the field. Separate and unequal in the academy inevitably mirrored separate and unequal in the society at large.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
American Sociology and Epistemic Injustice 209 It should be completely unsurprising, then, that well into the twentieth century, American sociology was an overwhelmingly “white” discipline. Steinberg (2007, 58) provides a dramatic picture: [H]ow could sociology not have been “white,” given that blacks were systematic ally excluded from the discipline, except for a token handful who were shunted off to teach in black colleges, cut off from the mainstream? Prior to World War I only fourteen blacks in the United States received a Ph.D. in sociology. Even as late as 1936 over 80 percent of black Ph.D.s were employed by Atlanta, Fisk, and Howard universities [HBCUs], and it was not until 1942 that a major university hired a full-time tenure-track African American professor. Only four blacks were among the 750 contributors to the American Journal of Sociology between 1916 and 1940. Thus, there was no significant black presence in the discipline to provide a reality check on the knowledge claims that were spewed forth by a white establishment.
Nor had this situation changed dramatically, even decades later: “In 1967 there were only 121 Negro doctorates in sociology in the entire nation, representing about one per cent of all sociologists” (14). Rather than “epistemic democracy” (Anderson 2012) in the field, it was more a matter of “epistemic whiteocracy.” No wonder, then, that as late as 1970, black radical sociologist Joyce Ladner (1998 [1970]) felt it necessary to publish an edited book calling for “the death of white sociology.” Fifth, consider as an exemplary illustration the fate of W. E. B. Du Bois (1868–1963). Long regarded as black America’s most impressive intellectual figure, with (depending on how one counts) twenty books and hundreds of articles to his credit, Du Bois had, by mainstream “white” standards, impeccable scholarly credentials. Not only was he Harvard’s first black Ph.D., but he also spent two years getting quantitative training at the University of Berlin and was in dialogue with Max Weber, correcting some of the German giant’s misconceptions on race (Morris 2015). But in a classic case of structurally determined (Fricker, Anderson) or over-determined, epistemic injustice, Du Bois’s prolific eight-decade career, beginning (for sociology) with the pathbreaking 1899 The Philadelphia Negro, could not overcome white sociological skepticism and exclusion, lacking in white eyes both testimonial credibility and hermeneutical acceptability. Moreover, as suggested at the start (and this is where the question is raised of the possible need for an expansion of Fricker’s notion to the broader disciplinary level), his unacceptable, indeed perhaps almost incomprehensible, challenge to regnant conceptions of “race” as biological racial superiority/inferiority brought into question the foundations not just of infant American sociology, but of all disciplines resting on the “obvious” truth of natural racial hierarchy. However, it must be admitted up front that, from a twenty-first-century “progressive” standpoint, Du Bois’s early writings on race are often pervaded with
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
210 Charles W. Mills what appears to be an uncomfortable degree of victim blaming. Du Bois himself often uses the phrase “the Negro problem” (Du Bois 1996b [1903], 1996c [1899]) and it is hard to avoid the conclusion that the anger he expresses in “Of Our Spiritual Strivings” is grounded more in outrage that he, Harvard Ph.D. and New England aristocrat, should be so upbraided, as against other manifestly less worthy Negroes with no such basis for protest. Nonetheless, he maintains from the start a principled opposition to racism, and even in an excerpt from the early 1899 Philadelphia Negro, “What is the Negro Problem?” (Du Bois 1996c [1899]), which finds him deploring “Negro crime” and its rooting in “Negro homes” that are “breeders of idleness and extravagance and complaint” (350), he does state at the outset that “the civilized world’s” denial of “Humanity” to “the Negroes of Africa” is the “widespread and deep-seated” “feeling” that “is, in America, the vastest of the Negro problems” (347). So, this broader contextualization rejects the position that black behavioral reform is going to be sufficient as a solution to the “problem,” and such structural emphases would become more prominent in his work as the new century wore on and he gravitated towards left theory. Unsurprisingly, then, the mature, alternative black radical paradigm of Du Bois and other black progressives of the period would fundamentally challenge the socio-political-historical assumptions of the white academy, and the related taken-for-granted understandings of the dominant white society of the time (or even now!). In one picture, “race” is location in a hierarchy of natural superiors and inferiors; in the other, “race” is location in a hierarchy of natural equals respectively privileged and disadvantaged by a structure of social oppression. In one picture, natural causality is centrally determinant; in the other, social causality is centrally determinant. In one picture, moral approbation (for the natural order of things) or at least an ostensibly wertfrei neutrality is what is appropriate; in the other, moral condemnation is what is called for. Finally, in one picture, no fundamental social change is necessary (or maybe even possible); in the other, prescriptions for radical, or even revolutionary, reform are clearly morally imperative. In sum, the black radical perspective constituted/constitutes a devastating indictment of the self-congratulatory white iconography of the United States as a new experiment in the history of mankind, a shining city on the hill that was going to be a moral beacon to the world. Is it any wonder that Du Bois, despite his manifold accomplishments, would remain unrecognized by the Jim-Crowed white academy, causing him to leave the United States in disgust for Ghana two years before his death and to take up Ghanaian citizenship? Only recently, with the work of Morris (2015) and others, has the longstanding black American demand to acknowledge Du Bois as the true father of American sociology begun to be taken seriously. Sixth, and finally, what were the implications for American sociology and its twentieth- century development of this marginalization of Du Bois and
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
American Sociology and Epistemic Injustice 211 canonization instead of Park and those following in his footsteps, if more culturally than biologically determinist? As both Morris and Steinberg emphasize, the result was a discipline that ignominiously failed in its self-appointed mission of being a “science of society,” presumptively able to uncover the deep structures and hidden dynamic of the social order. The dramatic social upheavals of the 1950s and 1960s—the black protests, demonstration, and later ghetto uprisings—were completely unanticipated in the pages of the leading sociological journals and the panels of the national ASA conferences of the preceding years. Steinberg (2001, 61–2) cites the crucial question posed in ASA president Everett Hughes’s 1963 presidential address: “Why did social scientists—and sociologists in particular— not foresee the explosion of collective action of Negro Americans toward immediate full integration into American society?” Hughes’s own answer was, unsurprisingly, inadequate, locked as he still was in the existing paradigm: “Our conception of social science is so empirical, so limited to little bundles of fact applied to little hypotheses, that we are incapable of entertaining a broad range of possibilities” (cited in Steinberg 2001, 62). But as Steinberg comments, this “narrow empiricism” was not methodologically neutral, but a manifestation of the political refusal to acknowledge the “big picture” of white racial domination (62). The structural epistemic injustice—testimonial, hermeneutical, and disciplin ary—done to the insights of the victims of this domination would blow back in the form of a disciplinary incapacity to comprehend within its framework of concepts and assumptions the reality of what was happening. Instead, as the decade wore on, and cities across the country went up in flames, it was the radical theory of Du Bois and other representatives of the alternative black sociological tradition that would furnish the necessary theoretical insight on the underlying forces at work. The 1960s “scholarship of confrontation,” in Steinberg’s (ch. 3) phrase, would gain an ironic vindication in the 1968 Kerner Commission Report, that—in effect dismissing decades of white sociological orthodoxy—flatly declared that “white racism is essentially responsible for the explosive mixture which has been accumulating in our cities since the end of World War II” (cited in Steinberg 2001, 78). A commission appointed by no less respectable an authority than President Lyndon Johnson and the US federal government was confessing that white American society was itself racist. The report’s recommendations were never acted upon, and the brief swing to the left and towards racial honesty would soon be reversed (hence Steinberg’s [2001] book title: Turning Back: The Retreat from Racial Justice in American Thought and Policy). Yet, if only for a moment, the racial truth had been told and the racial reality admitted. If American sociology today is somewhat more willing to accommodate the perspective of radicals, such as past ASA presidents Joe Feagin (Racist America [2014], The White Racial Frame [2020]) and Eduardo Bonilla-Silva (White Supremacy and Racism in the Post-Civil Rights Era [2001], Racism without Racists [2018]), and now current (2021) ASA president Aldon Morris (The Scholar Denied [2015]), with their unhesitating
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
212 Charles W. Mills employment of concepts of “institutional racism” and “white supremacy,” it is because it was eventually forced, despite itself, into a belated (partial) recognition of the real, underlying structural dynamic first identified by black radical theorists a century previously in their quest to expose and overturn both social injustice and epistemic injustice: the white problem.
References Alcoff, Linda Martín (2017). “Philosophy and Philosophical Practice: Eurocentrism as an Epistemology of Ignorance,” in Ian James Kidd, José Medina, and Gaile Pohlhaus (eds), Routledge Handbook of Epistemic Injustice. New York: Routledge, 397–408. Alexander, Michelle (2010). The New Jim Crow: Mass Incarceration in the Age of Colorblindness. New York: New Press. Anderson, Elizabeth (2012). “Epistemic Justice as a Virtue of Social Institutions.” Social Epistemology 26(2): 163–73. Asad, Talal (ed.) (1995 [1973]). Anthropology and the Colonial Encounter. New York: Humanity Books. Baker, Lee D. (1998). From Savage to Negro: Anthropology and the Construction of Race, 1896–1954. Berkeley, CA: University of California Press. Baptist, Edward E. (2014). The Half Has Never Been Told: Slavery and the Making of American Capitalism. New York: Basic Books. Blackburn, Robin (ed.) (1972). Ideology in Social Science: Readings in Critical Social Theory. London: Fontana. Bonilla-Silva, Eduardo (2018). Racism without Racists: Color-Blind Racism and the Persistence of Racial Inequality in America, 5th edn. Lanham, MD: Rowman & Littlefield. Bonilla-Silva, Eduardo (2001). White Supremacy and Racism in the Post-Civil Rights Era. Boulder, CO: Lynne Rienner. CNN (2018). “The Scientific Theories Battling to Explain the Universe” (August 17, 2018). Darby, Derrick and John L. Rury (2018). The Color of Mind: Why the Origins of the Achievement Gap Matter for Justice. Chicago, IL: University of Chicago Press. Du Bois, W. E. B. (2007 [1899]). The Philadelphia Negro: A Social Study. New York: Oxford University Press. Du Bois, W. E. B. (1996a [1903]). “The Souls of Black Folk,” in Eric J. Sundquist (ed.), The Oxford W. E. B. Du Bois Reader. New York: Oxford University Press, 97–240. Du Bois, W. E. B. (1996b [1903]). “The Training of Negroes for Social Power,” in Eric J. Sundquist (ed.), The Oxford W. E. B. Du Bois Reader. New York: Oxford University Press, 354–62. Du Bois, W. E. B. (1996c [1899]). “What is the Negro Problem?,” in Eric J. Sundquist (ed.), The Oxford W. E. B. Du Bois Reader. New York: Oxford University Press, 346–54.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
American Sociology and Epistemic Injustice 213 Ebony (August 1965). Special Issue: “The White Problem in America.” Feagin, Joe R. (2014). Racist America: Roots, Current Realities, and Future Reparations, 3rd edn. New York: Routledge. Feagin, Joe R. (2020). The White Racial Frame: Centuries of Racial Framing and Counter-Framing, 3rd edn. New York: Routledge. Fricker, Miranda (2007). Epistemic Injustice: Power and the Ethics of Knowing. New York: Oxford University Press. Fricker, Miranda (2017). “Evolving Concepts of Epistemic Injustice,” in Ian James Kidd, José Medina, and Gaile Pohlhaus (eds), Routledge Handbook of Epistemic Injustice. New York: Routledge, 53–60. Garcia, Jorge L. A. (1996). “The Heart of Racism.” Journal of Social Philosophy 27: 5–45. Gould, Stephen Jay (1996). The Mismeasure of Man, rev. and exp. edn. New York: W. W. Norton. Hawkins, Mike (1997). Social Darwinism in European and American Thought, 1860–1945: Nature as Model and Nature as Threat. New York: Cambridge University Press. Horsman, Reginald (1981). Race and Manifest Destiny: Origins of American Racial Anglo-Saxonism. Cambridge, MA: Harvard University Press. Hund, Wulf D., Charles W. Mills, and Silvia Sebastiani (eds) (2015). Simianization: Apes, Gender, Class, and Race, “Racism Analysis Series B Yearbook,” Vol. 6. Berlin: LIT Verlag. Jordan, Winthrop D. (2012). White over Black: American Attitudes toward the Negro, 1550–1812, 2nd edn. Chapel Hill, NC: Omohundro Institute of Early American History and Culture/University of North Carolina Press. Kidd, Ian James, José Medina, and Gaile Pohlhaus Jr (eds) (2017). The Routledge Handbook of Epistemic Injustice. New York: Routledge. Kuhn, Thomas S. (2012). The Structure of Scientific Revolutions, 50th anniversary edn. Chicago, IL: University of Chicago Press. Ladner, Joyce (ed.) (1998 [1970]). The Death of White Sociology: Essays on Race and Culture. Baltimore, MD: Black Classic Press. Mills, Charles W. (2003). “ ‘Heart’ Attack: A Critique of Jorge Garcia’s Volitional Conception of Racism.” Journal of Ethics 7(1): 29–62. Mills, Charles W. (2017). “Ideology,” in Ian James Kidd, José Medina, and Gaile Pohlhaus (eds), Routledge Handbook of Epistemic Injustice. New York: Routledge, 100–11. Mills, Charles W. (1997). The Racial Contract. Ithaca, NY: Cornell University Press. Morris, Aldon (2015). The Scholar Denied: W. E. B. Du Bois and the Birth of Modern Sociology. Oakland, CA: University of California Press. Muhammad, Khalil Gibran (2010). The Condemnation of Blackness: Race, Crime, and the Making of Modern Urban America. Cambridge, MA: Harvard University Press.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
214 Charles W. Mills Myrdal, Gunnar (2017 [1944]). An American Dilemma: The Negro Problem and Modern Democracy, 2 vols. New York: Routledge. Nott, Josiah Clark, and George Robins Gliddon. (2010). Types Of Mankind: Or Ethnological Researches, Based Upon The Ancient Monuments, Paintings, Sculptures, And Crania Of Races. Whitefish, MT: Kissinger Publishing. Pateman, Carole and Charles W. Mills (2007). Contract and Domination. Malden, MA: Polity Press. Pohlhaus Jr, Gaile (2017). “Varieties of Epistemic Injustice,” in Ian James Kidd, José Medina, and Gaile Pohlhaus (eds), Routledge Handbook of Epistemic Injustice. New York: Routledge, 13–26. Rawls, John. (1999). A Theory of Justice, rev. edn. Cambridge, MA: Harvard University Press. Rose, Hilary and Steven Rose (eds) (1976). The Political Economy of Science: Ideology of/in the Natural Sciences. London: Macmillan. Saini, Angela (2018). Inferior: How Science Got Women Wrong—and the New Research That’s Rewriting the Story. Boston, MA: Beacon. Shelby, Tommie (2003). “Ideology, Racism, and Critical Social Theory.” Philosophical Forum 34(2): 153–88. Shelby, Tommie (2002). “Is Racism in the ‘Heart’?” Journal of Social Philosophy 33(3): 411–20. Smedley, Audrey and Brian D. Smedley (2018). Race in North America: Origin and Evolution of a Worldview, 4th edn. New York: Routledge. Smith, Rogers M. (1997). Civic Ideals: Conflicting Visions of Citizenship in U.S. History. New Haven, CT: Yale University Press. Steinberg, Stephen (2007). Race Relations: A Critique. Stanford, CA: Stanford University Press. Steinberg, Stephen (2001 [1995]). Turning Back: The Retreat from Racial Justice in American Thought and Policy. Boston, MA: Beacon. Sundquist, Eric J. (ed.) (1996). The Oxford W. E. B. Du Bois Reader. New York: Oxford University Press. Torrance, John (1995). Karl Marx’s Theory of Ideas. New York: Cambridge University Press. Tsosie, Rebecca (2017). “Indigenous Peoples, Anthropology, and the Legacy of Epistemic Injustice,” in Ian James Kidd, José Medina, and Gaile Pohlhaus (eds), Routledge Handbook of Epistemic Injustice. New York: Routledge, 356–69. Vitalis, Robert (2015). White World Order, Black Power Politics: The Birth of American International Relations. Ithaca, NY: Cornell University Press.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 19/03/21, SPi
10
A Tale of Two Injustices Epistemic Injustice in Philosophy Emmalon Davis
1. Introduction This chapter has two aims. I articulate and defend a conceptual tool for thinking about epistemic injustice. I utilize this tool to develop a partial explanation for the persistent lack of diverse practitioners in academic philosophy.1 Academic phil osophy has been amply observed to lack diversity in (at least) two ways: (i) Academic philosophy lacks diversity with respect to the social identities of its practitioners. Women, gender minorities, and people of color are among those notably underrepresented and marginalized within the field.2 (ii) Academic philosophy lacks diversity with respect to the kinds of discourses in which its practitioners are engaged. Contributors engaged in race and gender discourses—including, for example, philosophy of race and feminist philosophy—are among those notably underrepresented and marginalized in the field.3 1 This chapter concerns academic philosophy as it is practiced in the United States and the so- called English-speaking world. While I focus especially on race and gender disparities in philosophy, I intend for my analysis to extend more broadly, where applicable, to other underrepresented groups (e.g. ability, class, sexual orientation, etc.). 2 Concerning women in philosophy, see Alcoff (2003); Haslanger (2008); Calhoun (2009); Antony (2012); Hutchison and Jenkins (2013); and Paxton et al. (2012). Concerning black women and women of color, see Yancy (2008); Belle [Gines] (2011); Russell (2019); Narayan (2003); Wilson (2017); and Lee (2017). Concerning trans and non-binary persons, see Bettcher (2018) and Dembroff (forthcom ing). Concerning blacks, see Yancy (1998) and Botts et al. (2014). Concerning indigenous persons, see Waters (2003) and Whyte (2017). Concerning Asian and Asian Americans, see Lee (2014) and Kim (2002). Concerning Latinx persons, see Gracia (2008); Yancy (2012); and Madva (2016). While this chapter aims to articulate commonalities that bear explanatory relevance for the underrepresentation and marginalization of the groups outlined above, my analysis may not resonate uniformly with every member within and across these groups. 3 Concerning feminist philosophy, see Walker (2005); Haslanger (2008); Rooney (2011); and Superson (2011). Concerning philosophy of race, see Outlaw (1996); Mills (1997); Haslanger (2008); and Yancy (2012). I also intend for my analysis to encompass a wider range of discourses, some of which are and some of which are not situated under the more general categories of race, feminist, or gender philosophy: e.g. philosophy of disability (Tremain 2018); trans philosophy (Bettcher 2018), black feminist philosophy (James 2014); queer theory (Salamon 2009); Africana philosophy (Outlaw 1996), indigenous philosophy (Whyte 2017), Indian political philosophy (Krishnamurthy 2016),
Emmalon Davis, A Tale of Two Injustices: Epistemic Injustice in Philosophy In: Applied Epistemology. Edited by: Jennifer Lackey, Oxford University Press (2021). © Emmalon Davis. DOI: 10.1093/oso/9780198833659.003.0010
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
216 Emmalon Davis As I show, these disparities are epistemically connected. Practitioners outlined in (i) and (ii) are both targets of epistemic injustice in academic philosophy; while the former are targeted in virtue of their social identities, the latter are targeted in virtue of the content of their contributions. To illuminate this connection, I dis tinguish between two varieties of testimonial injustice: identity- based and content-based testimonial injustice. Utilizing these twin concepts, I offer an epi stemic explanation that plausibly contributes to our understanding of the demo graphic disparities observed.4 Specifically, I argue that identity-based and content-based testimonial injustice are both prevalent in academic philosophy and that this prevalence introduces substantial barriers to participation for those targeted. Both identity-based and content-based testimonial injustice give rise to similarly injurious effects. For example, targets of either form of testimonial injustice may find it more difficult to shape the direction of inquiry, to secure credit for their contributions, and to find platforms through which to disseminate their ideas. They may be flatly ignored, dismissed, or interrupted. Their credibility may be openly questioned, and they may be pre- emptively denied opportunities to participate in an exchange. In a discipline like philosophy, the practice of which predominately transpires through written and spoken discourse, such epistemic and communicative hurdles are prohibitive. Thus, the dual and compounding effects of identity-based and content-based testimonial injustice in academic philosophy plausibly contribute to the lack of diversity concerning both the social identities of practitioners and the discourses in which practitioners are engaged. Diverse practitioners are disproportionately represented at every level of aca demic philosophy—from student majors to doctoral students to faculty. Empirical research has suggested that the most significant decline in representation occurs prior to the Ph.D. stage (Paxton et al. 2012; Beebee and Saul 2011; Botts et al. 2014). As this chapter demonstrates, epistemic injustice is harmful at all ranks of the discipline, but I argue that there is good reason to think its effects are par ticularly pernicious at the earliest stages of philosophical exposure. On the account I am developing, students who experience, or anticipate experiencing, testimonial injustice in the formal and informal spaces in which they are exposed to academic philosophy—for example, lecture halls, seminar rooms, departmen tal talks, hallways and common spaces, reading rooms, student clubs, etc.—are decolonial philosophy (Maldonado-Torres 2011), among others. While this chapter explores unifying factors contributing to a shared displacement within the field, a more comprehensive analysis would also examine commitments of individual discourses and unique forms of exclusion attendant to each. 4 See Thompson (2017) for a survey of proposed hypotheses and existing empirical research. As Thompson notes, the empirical literature has focused almost exclusively on philosophy’s gender gap, without sufficient attention to the underrepresentation of other identities or their intersections. Consequently, the range of hypotheses offered do not generally acknowledge the differential barriers experienced by non-white (especially black and indigenous), first-generation, disabled, trans, non- binary, etc. persons in pursuing higher education or academic careers in philosophy.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
EPISTEMIC INJUSTICE IN PHILOSOPHY 217 uniquely disincentivized from pursuing philosophy further. An unequal epistemic playing field provides ample reason for even the most promising practitioners to seek alternative educational and employment opportunities. The chapter proceeds as follows. In section 2, I distinguish between identity-based and content-based testimonial injustice and consider cases in which these injustices discourage participation in philosophy. In section 3, I articulate three ways in which a philosophical discourse may become social identity-coded, such that identity prejudice (or other unwarranted assessments) influence epistemic appraisals of the discourse and its contributors. In section 4, I distinguish my proposal from three extant accounts: the “different voices,” “schema clash,” and “culture of justification” accounts. In the remaining sections, I address potential objections to my analysis. In section 5, I respond to the objection that my analysis is unable to account for the “uniqueness” (at least amongst the humanities) of philosophy’s diversity problem. In section 6, I address the concern that my analysis employs a distinction without a difference; I argue that although identity-based and content-based testimonial injustice are inextricably linked, the phenomena should not be collapsed. In section 7, I respond to the challenge that content-based testimonial injustice is inappropriately characterized as an epistemic injustice, in that it lacks the central epistemic harm associated with traditional (e.g. identity-based) forms. I argue that the distinction between identity-based and content-based testimonial injustice illuminates two modes of epistemic subjectivity, both of which can be harmfully thwarted by epistemic injustice.
2. Identity-Based and Content-Based Testimonial Injustice As it has been conventionally—but not uncontestably5—defined, testimonial injustice picks out “the injustice that a speaker suffers in receiving deflated cred ibil ity from the hearer owing to identity prejudice on the hearer’s part” (Fricker 2007, 4). Identity prejudice, on this view, is prejudice that targets persons in virtue of their social group identities (e.g. their gender, race, ability, etc.). Testimonial injustice can lead to the dismissal or rejection of the target’s contri bution, or, operating pre-emptively, it can prevent the target from being appropri ately consulted or otherwise epistemically acknowledged. To illustrate, consider a doctor, Paul, who refuses to accept the testimony of a patient, Sammi, concerning her chronic pain, where this refusal is at least partially attributable to ableist and sexist stereotypes about the physical manifestations of chronic illness and women’s credibility concerning their own bodies. “You look fine to me,” Paul tells her, adding that most of his patients “are in wheelchairs.” Insofar as Sammi’s
5 See Medina (2011); Dotson (2012a); Anderson (2012); Pohlhaus Jr (2014).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
218 Emmalon Davis doctor assigns her an identity prejudicial credibility deficit, she experiences a familiar form of testimonial injustice. Let us call this identity-based testimonial injustice. Identity-based testimonial injustice can be contrasted with another form of testimonial injustice. A contributor may be prejudicially (or otherwise unjustly) assessed not only owing to social identity, but also owing to the kind of informa tion one attempts to convey. Indeed, the content of a contributor’s contribution can itself become social identity-coded in ways that provoke unwarranted epi stemic assessments of the contributor and contribution. To illustrate, consider several doctors at a medical conference who have gathered for drinks to discuss their favorite sessions. One doctor, Preston, begins to report to the group what he has learned during a session on fibromyalgia (a condition of recent interest to him, since several of his patients may be affected). Just as Preston begins to share, a fellow doctor, James, chimes in: “Preston, I’m gonna have to stop you right there. Everyone knows those patients are just looking for attention. There’s a rea son it only affects women!” Preston tries to respond, but James sharply interrupts Preston to tell the group about what he has learned in a session on new develop ments in robot-assisted joint replacement. As James’s commentary makes quite clear, his dismissal of Preston’s testimony owes itself to a gender association between fibromyalgia and women coupled with a prejudicial assessment of women (with chronic pain) as “attention-seekers.” James pre-emptively dismisses Preston’s contribution—essentially preventing Preston from sharing with the group what he has learned about fibromyalgia—in virtue of prejudice occasioned by identity-coded content in Preston’s contribution. Moreover, James calls into question Preston’s own credibility by suggesting that his contribution has no merit. Let us call this content-based testimonial injustice.6 Like its counterpart, content-based testimonial injustice can lead to the dismissal or rejection of the target’s contributions. It can also operate pre-emptively, when an audience pre vents a target from contributing when certain content is anticipated. In both cases, the underlying prejudices concern women and people with chronic pain. While Sammi experiences an identity-based testimonial injustice, in virtue of her social identity as a woman with chronic pain, Preston experiences a content-based testimonial injustice, in virtue of his attempt to engage others in a discourse associated with women and persons with chronic pain. The epistemic harm to Sammi is direct; her epistemic interests as a knower and contributor have been severely hampered by her doctor, affecting not only her standing as an
6 Relatedly, see Dembroff and Whitcomb (forthcoming) who argue for the existence of “content- focused epistemic injustice,” a similar phenomenon, which they distinguish from testimonial injust ice. Though our accounts developed independently, I take our shared recognition of the importance of content in epistemic injustice to be mutually supportive.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
EPISTEMIC INJUSTICE IN PHILOSOPHY 219 epi stem ic agent, but also, more broadly, her health and well- being. In Preston’s case, James’s dismissal reflects an attitude of epistemic denigration towards women and people with chronic pain and derails a discursive exchange in which a group of doctors might have learned something from their peer about fibromyalgia. In this way, James’s dismissal of Preston harms the epistemic inter ests of people with chronic illnesses and of women, more generally, in their capacities to be known, valued, and understood. Insofar as content-based testi monial injustice results in the dismissal and rejection of contributors whose con tributions are associated with or concern the interests of persons with unjustly devalued social identities, it takes on a broader epistemic (and social) significance. For our purposes, then, I contrast the following two forms of testimonial injustice: Identity-based: prejudice or other unjust assessments regarding a contributor’s social identity (e.g. gender, race, ability, etc.) influence an audience’s evaluation of the contributor’s epistemic standing (e.g. credibility, competence, value, etc.), compromising the audience’s willingness to consider or fairly engage the con tributor and contribution. Content-based: prejudice or other unjust assessments regarding social identity-coded content (e.g. gender-coded, race-coded, ability-coded, etc.) of a contributor’s contribution influence an audience’s evaluation of the contributor’s epistemic standing (e.g. credibility, competence, value, etc.), compromising the audience’s willingness to consider or fairly engage the contributor and contribution. Importantly, identity-based and content-based testimonial injustice can occur independently. That is, persons with non-dominant social identities can experience epistemic harms via identity-based testimonial injustice, regardless of the content of their contributions. Similarly, persons whose contributions are non-dominant social identity-coded can experience epistemic harms via content-based testimonial injustice, regardless of their social identities. It is also possible, however, for a contributor to be vulnerable with respect to both forms of testimonial injustice simultaneously. In such cases, the overlapping and intersecting epistemic effects for such contributors may be difficult to disentangle. As I show, however, the distinction is nonetheless conceptually and practically useful. Let us now turn to philosophy. To see how identity-based and content-based testimonial injustice disadvantage diverse practitioners in philosophy, consider the following. Lina is an undergraduate woman of color who has taken a few phil osophy courses but has yet to add it as a major. Lina greatly enjoys the ideas explored in her classes and values the skills she is developing. One of her profes sors has even suggested she consider pursuing a career in philosophy. But several experiences have rendered her uncertain about whether philosophy is the right
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
220 Emmalon Davis choice for her. (N.B. Although these cases are presented as vignettes, they are not hypothetical.) Page Forty-Five Lina is the only woman and the only student of color in her small in-class dis cussion group. The group has been assigned to discuss what (white male) philoso pher P means when he says x. Lina suggests several times that the group turn to page forty-five in the assigned text, a page she believes contains crucial clues con cerning what philosopher P means by x. The first few times Lina proposes the suggestion she is ignored, until finally one student jumps in quickly to respond that one should not read too much into any single passage. Throughout the exchange, the professor either does not notice she is being unfairly dismissed or simply does not intervene on her behalf. Lina stops participating in the conversa tion. After the discussion goes in circles for nearly ten minutes, another student states that he has found something important on page forty-five. The rest of the group dutifully turns to page forty-five, while that student reads aloud the passage Lina had underlined and starred in her own copy. At least she had the right idea, she thinks to herself. Student Colloquia Lina has finished a philosophy final paper exploring the influence of women in the meditative tradition before Descartes. Lina’s professor was very pleased with it and encouraged her to present the paper at the philosophy student colloquia club in order to get feedback and to meet more of the majors. As she is marking her name on the sign-up sheet to present, a male peer—who happens to be the organ izer—says he is happy to see that Lina is interested in presenting, since it’s usually the same few people, and the club could really use some new participants. The organizer then asks Lina what she is going to present on. After Lina shares her topic, the organizer scrunches up his face and says, “Here’s a suggestion—don’t present on that. Nobody wants to hear about that! Do you have any other papers?” Animal Ethics Lina is taking a course on animal ethics. She is intrigued, but not quite convinced, by the arguments she has been studying. In order to think through one of the arguments further, Lina decides to discuss it with another philosophy student with whom she has had a prior class, since he seemed to know a lot about phil osophy. When Lina tries to start a conversation with him, he is dismissive of the topic and mocks Lina’s enthusiasm. Before walking away, he jokes about shooting animals for the fun of it. Lina never broaches the subject again with him. A few months later, Lina discovers that this student has become a committed vegan. Surprised, Lina asks a mutual friend when this student became vegan. The friend reports that the student heard about moral veganism from another male peer
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
EPISTEMIC INJUSTICE IN PHILOSOPHY 221 (who also took the animal ethics class), which subsequently sparked his interest. Lina wonders why she was not able to sustain this student in conversation on the topic a few months prior. As these cases demonstrate, Lina is consistently denied the opportunity to make valuable contributions to discussion, to receive acknowledgement for suggestions that are originally her own, to start a conversation with peers that might result in philosophical insight (even personal growth), and to secure uptake for and gener ate interest in a new idea, where each of the aforementioned is integral to the pursuit of philosophy. While Fricker’s account of testimonial injustice primarily concerns assertions, Christopher Hookway (2010) has persuasively argued that something like testimonial injustice affects a much wider range of epistemic con tributions. In contrast to Fricker’s “informational perspective,” Hookway draws our attention to what he calls the “participant perspective” (157), a perspective from which other aspects of an epistemic exchange—for example, “asking ques tions, floating ideas, [and] considering alternative possibilities” (155)—become salient. I follow Hookway in acknowledging the epistemic significance of these further modes of participation. Thus, though I continue (for the sake of termino logical parsimony) to refer to the phenomena under investigation as “testimonial injustice,” I am expressly concerned with barriers to epistemic participation in this much broader sense.7 In each case, Lina’s interlocutors preclude her participation, but the distinction between identity-based and content-based testimonial injustice can provide a more nuanced diagnosis of each case. In “Page Forty-Five,” Lina is the only woman and the only person of color in her discussion group. It is presumably because of her social identity as a woman of color that she has difficulty securing the epistemic standing she needs to ensure that her contribution—which contains no ostensible identity-coded content—is taken seriously by her interlocutors; for the same exact contribution, when offered by a white male student, is received by her peers with no difficulty at all. In “Student Colloquia,” Lina is preemptively dismissed by her male peer because of what she intends to contribute, namely, a paper concerning the philo sophical influence of women in the meditative tradition before Descartes. For the organizer doesn’t object to the fact that Lina is signing up to present. In fact, he seemed quite happy at first. His reproach only comes after he discovers her pro posed paper topic. What marks Lina out for dismissal in this case—i.e. “don’t present on that”—is the gendered content of her intended contribution. (Indeed, it’s not hard to imagine that even a white, male student intending to present on the same topic might encounter some pushback.) 7 For those who prefer to adopt a different term instead of testimonial injustice, perhaps “epistemic participation injustice” would suffice.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
222 Emmalon Davis In “Animal Ethics,” Lina has difficulty sustaining her peer in a casual conversation on the topic of animal welfare, while another male student successfully sparks interest in the same topic on a different occasion. This case may be best theorized as a hybrid case. Lina’s peer may be dismissive of her both because of her social identity and because he perceives animal ethics to be an “un-masculine” topic.8 That Lina’s peer at least loosely connects animal ethics and masculinity might be evidenced by his comments about shooting animals for fun—a reference to hunting, stereotypically characterized as a (white) “man’s sport.” If some combination of identity-based and content-based dismissal is at work in this case, this could explain why Lina’s peer demonstrates increased willingness to engage the topic when it is later broached by a fellow male student. If my analysis is plausible, then “Page Forty-Five” illustrates identity-based testimonial injustice, while “Student Colloquia” illustrates content-based testimonial injustice. “Animal Ethics” illustrates a hybrid case. When epistemic injustice occurs with frequency in a given domain, those tar geted have reason to participate less often in such spaces or to avoid those spaces altogether (so as to avoid the annoyance, anger, or embarrassment associated with being targeted). Striving to succeed in a domain in which one is systematic ally and unjustly subjected to externally imposed epistemic and communicative disadvantages is, minimally, frustrating; at worst, it compromises one’s ability to reach one’s full potential in that domain. Lina is interested in philosophy. She has taken several courses. She has even received encouragement from faculty. In many ways, this is a best-case scenario. Nevertheless, it would hardly seem shock ing if Lina’s experiences lead her to conclude that, despite philosophy’s attrac tions, people with her identity or people with her interests, or both, are better off in other disciplines instead.
3. Social Identity-Coded Discourses in Philosophy On the view I am developing, content-based testimonial injustice occurs when the content of a contributor’s contribution becomes identity-coded, such that associations with a particular social identity become affixed to the content in question. But how does this happen? Taking gender and race as our starting point, let me develop this idea. Carol Cohn defines gender as “a system of meanings,” comprised of “ways of thinking, images and words that first shape how we experience, understand, and represent ourselves” (1993, 228–9). Profoundly entrenched and oppressively
8 For discussion, see Ruby and Heine (2011) and Rothgerber (2013).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
EPISTEMIC INJUSTICE IN PHILOSOPHY 223 resilient, the gender system permeates many facets of human life—from the toys and clothes with which children are outfitted, to the bathrooms people are expected to use, to the amount of compensation allotted for equivalent performances, to the opportunities available for social, political, and economic advancement, and so on. Gender associations usher in normative expectations that inform social practices of valuation and devaluation in accordance with prevalent gender stereotypes and ideologies (Valian 1998). Often, the association of a devalued gender identity with particular activities, traits, objects, or spaces transfers attitudes of devaluation thereto. Race, alongside gender, permeates and structures human activities and the spaces in which those activities take place. For example, race influences which regions receive aid in natural disaster (and how quickly); which neighborhoods are allocated effective resources for self-protection, economic development, and education; and which persons are presumed innocent or guilty when shopping in stores or walking down the street. Race gives rise to normative expectations, which determine, among other things, which spaces, places, and activities are val ued and which persons within those spaces are treated with respect. Accordingly, Charles Mills argues that: [t]he norming of space is partially done in terms of the racing of space, the depiction of space as dominated by individuals . . . of a certain race. At the same time, the norming of the individual is partially achieved by spacing it, that is, representing it as imprinted with the characteristics of a certain kind of space. So this is a mutually supporting characterization that . . . becomes a circular indictment: “You are what you are in part because you originate from a certain kind of space, and that space has those properties in part because it is inhabited by creatures like yourself.” (1997, 40–1)
As Mills suggests, the “racing” and “norming” of persons is co-constructed through the racing and norming of spaces (and the properties and characteristics thereby associated with them). Just as physical spaces can become gendered and/or racialized, so too can dia lectical spaces. A discourse, taken broadly to include spoken, written, signed, etc. forms of communication, becomes gendered or racialized when, as a result of race or gender associations, perceptions of the discourse and its practitioners are consciously or unconsciously shaped by gender or race ideologies. When a par ticular identity is socially devalued, this devaluation likewise serves to underwrite the devaluation of discourses associated with that social identity. To see how gen der and race ideologies influence value perceptions of philosophical discourses and their practitioners, consider this memory recounted by feminist bioethicist Hilde Lindemann (2006, W15):
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
224 Emmalon Davis A few months ago, I attended a workshop on metaethics. One morning at breakfast, as a handful of guy-philosophers and I lingered over our coffee, we played the game of ranking the various specializations within philosophy according to the prestige they enjoy—which not coincidentally, also inversely tracks gender. Here’s what we came up with: Philosophy of mind, philosophy of language, and metaphysics: The alpha-dominant philosophy, done by Real Men Epistemology and Philosophy of Science: Done by manly enough men Metaethics: Done by men who aren’t entirely secure in their masculinity Ethics, Social and Political Philosophy: Done by girls Bioethics: Done by stupid girls.
The entire hierarchy is presumptively white, and explicitly gendered discourses were not even ranked. But one can imagine what greater inclusivity might have looked like. Feminist philosophy, according to this logic, might be said to be done by angry, stupid girls; philosophy of race done by angry, stupid girls of color and angry, men of color who—depending on the racial stereotypes at play—either aren’t entirely secure in their masculinity or are hyper-masculine (but not in the “right” ways). And so on and so forth. Some variation of this prestige hierarchy is likely familiar to many philosophers. The imbedded juxtaposition of identity descriptors alongside relative intelligence descriptors (e.g. “stupid girls”) expressed in this hierarchy reflects a long history of prejudicial beliefs concerning the amount of intellect traditionally associated with each discourse, where discourses towards the top of the hierarchy are thought to require more intelligence than those down below. Consider, for ex ample, the sentiments expressed by P. F. Strawson—philosopher of language and mind, whose most widely discussed contributions to philosophy were, perhaps ironically, in ethics—who was known to have frequently joked that “he would turn to moral philosophy only when his powers were waning” (Snowdon and Gomes 2019). Indeed, Strawson openly characterized the subfield of ethics as less “intellectually gripping”9 than the other areas of philosophy to which he directed most of his attention. Against a backdrop in which only “Real Men” are purported to be endowed with the “powers” required to engage in “gripping” philosophy, the perception that philosophical ethics is easy and that it is “done by girls” is no innocent pairing.10 9 Strawson regarded his two ethics papers to “effectively embody all I have thought or have to say in a philosophical area which, important as I recognize it to be, I have never found as intellectually gripping as those to which I have given more attention” (Snowdon and Gomes 2019). 10 While some may be tempted to view Strawson as an outlier, see Lloyd (1984) and Rooney (1991) for discussion concerning the association of women and irrationality in the history of philosophy. See also Leslie et al. (2015) for discussion concerning expectations of innate brilliance and gender distri butions across academic disciplines.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
EPISTEMIC INJUSTICE IN PHILOSOPHY 225 Likewise, Charles Mills observes the existence of two separate “worlds”: on the one hand, the world of mainstream (i.e., white) ethics and political phil osophy, preoccupied with discussions of justice and rights in the abstract, on the other hand, the world of Native American, African American, and Third and Fourth World political thought, historically focused on issues of conquest, imperialism, colonialism, white settlement, land rights, race and racism, slavery, jim crow, reparations, apartheid, cultural authenticity, national identity, indigen ismo, Afrocentrism, etc. These issues hardly appear in mainstream political phil osophy, but they have been central to the political struggles of the majority of the world’s population. (1997, 4)
As Mills’ characterization illuminates, systemic racist, settler- colonial, and western-centric biases are activated within smaller intellectual spaces, shaping the ways in which racialized content is positioned within the field (if it is granted a position at all). Referring to this practice as “exceptionalism,” Kristie Dotson argues that this philosophical practice “involves the unfounded [. . .] exclusion of large bodies of investigation based upon the privileging of one group (or set of groups) and their investigations over others” (2012b, 11). The hierarchy informs, facilitates, and justifies exclusionary practices that govern not only who (i.e. what social identities) can participate in philosophy, but also about what (i.e. the con tent) one must contribute. Given these observations, we can articulate at least three ways in which a discourse can become social identity-coded. A discourse can become social identity-coded in virtue of: (1) assumptions concerning who the contributors in the discourse are; (2) assumptions concerning whose attitudes or interests the discourse reflects or serves; (3) assumptions concerning the relative ability or intellectual “power” (e.g. brilliance, objectivity, rationality, gravitas, etc.) believed necessary to engage in the discourse. Any of the aforementioned can produce epistemic injustice, but we should distin guish (1)–(3) for several reasons. First, certain discourses may turn out to be identity-coded in virtue of some, but not all, of these options. For example, ethics, as characterized by Strawson, may be gender-coded in virtue of (1) and (3), but not necessarily in virtue of (2). Second, the kind of testimonial injustice prompted by (1), (2), and (3) may differ, where (1) and (3) may be more likely to yield an identity-based testimonial injustice and (2) more likely to produce a content-based testimonial injustice. This distinction is important, as practices aimed to
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
226 Emmalon Davis reduce the harms of identity-based testimonial injustice (e.g. anonymous grading or peer review, etc.) will likely fail to address the harms of content-based testimo nial injustice; for how does one effectively “anonymize” the content of a contribu tion? Finally, the exclusionary effects mediated by (1), (2), or (3) may vary in intensity for different contributors and discourses, and they may combine together to produce novel harms. A gendered and racialized hierarchy of discourses largely determines who fashions the philosophical terrain, which questions are characterized as philo soph ic ally interesting, and what conversations are deemed worth pursuing. Higher in the hierarchy are those discourses that are thought (via 1) to be engaged primarily by “Real Men”; (via 2) to serve the interests or reflect the attitudes of “Real Men” (or to be so highly abstract as to serve the interests of no “man” in particular); or (via 3) to require a high level of “power” (e.g. innate intellectual ability, objectivity, rationality, gravitas, etc.) associated with “Real Men.” Situated much lower in the hierarchy are non-dominantly gendered and racialized dis courses; that is, those discourses thought (via 1) to be engaged primarily by women, gender minorities, and/or non-whites; (via 2) to serve the interests or reflect the attitudes of women, gender minorities, and/or non-whites; or (via 3) to not require high levels of ability or “power” for meaningful participation (where this last assessment also involves classist and ableist associations). Thus, identity-based and content-based testimonial injustice constitute dual mechanisms through which the epistemic norms of a prejudicial hierarchy are structurally preserved in philosophical spaces. Illustrating the first mechanism at work, black woman philosopher Anita Allen recalls the disturbing way in which a colleague attempted to discourage her from philosophy, openly asserting that she did not possess “enough candle power (i.e., intellect)” but possessed “too much juice (i.e., sensuality)” for the profession (Allen and Yancy 2018). Likewise, Yolanda A. Wilson (2017) writes of a philosophy professor who claimed that phil osophy lacks black participants because they can’t pass a logic requirement and of a white male peer who prevented her from collecting her department mail because he didn’t think a woman of color could possibly be a graduate student in the department. Or Sally Haslanger (2008), who tells of her professor, who shared that he believed women were incapable of being first-rate philosophers, owing to their inability to have seminal ideas. In these cases, demeaning race and gender stereotypes are mobilized to launch discipline-specific attacks. Such attacks target practitioners in virtue of their social identities (e.g. qua woman, qua black person, qua black woman, etc.) to undermine their capacities as epistemic and philosoph ical contributors. Illustrating the second mechanism at work, consider the statements of black feminist philosopher V. Denise James who recalls the following (2014, 192–3):
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
EPISTEMIC INJUSTICE IN PHILOSOPHY 227 After [a] senior philosopher questioned the possibility of my existence as a black feminist philosopher, I started calling myself a pragmatist because Dewey’s work on democracy came closest to helping me answer the “what do you do?” ques tion without betraying my growing desire to do philosophy differently. Pragmatist was a name I could stand on. I could locate myself in the profes sion . . . Without a “guy,” my status as a philosopher in these settings was tenu ous . . . With Dewey as my “guy,” the looks of confusion, of misunderstanding, of suspicion lessened, although they did not disappear. Dewey gave me access to a professional philosophical community. Although I smuggled in black feminist thought from outside of the discipline, as long as pragmatist was the name I called myself, I could be understood.
As James reports, the association of a discourse (e.g. black feminist philosophy) with a devalued social identity (e.g. black women) serves to exclude and discredit the discourse and its primary investigators in academic philosophy; but that same discourse, once associated with a valued social identity (e.g. Dewey, the white “guy”), is permitted greater inclusion and accreditation within the field. Likewise, feminist ethicist Margaret Urban Walker states that practitioners engaged in feminist philosophy are “vulnerable to being discredited for not knowing the more professionally central discourses,” while practitioners engaged in dominant discourses are permitted to “be blithely ignorant . . . or to have mistaken or car toonish impressions” of feminist discourses (Walker 2005, 161). This profound failure of epistemic reciprocity divests marginalized discourses and their con tributors of value, creates non-reciprocal responsibilities for those contributors, and ensures that when such discourses do collide, it is disproportionately on dominant terms. As James and Walker illustrate, their epistemic and philosophical participation is threatened in virtue of social identity-coded content in their contributions.
4. Not Another Version of . . . On the hypothesis I am developing, epistemic injustice constrains a contributor’s full epistemic participation in a given domain, providing targets with a reason to participate less often in such spaces or to avoid such spaces altogether. Contributors with non-dominant social identities and contributors engaged in non-dominant social identity-coded discourses are both targets of epistemic injustice in academic philosophy; thus, both kinds of contributors have reason to participate less in academic philosophical spaces or to avoid them altogether. In this section, I distinguish my proposal from three alternative accounts, namely, the “different voices,” “schema clash,” and “culture of justification” accounts.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
228 Emmalon Davis Consider first the “different voices” hypothesis. While many versions exist, the proposal generally seeks to explain demographic asymmetries in philosophy by appealing to “intrinsic” differences—either natural or socialized—between differ ent identity groups (e.g. gender, culture, etc.). On this hypothesis, philosophy’s lack of diversity would be explained by the fact that diverse practitioners are selected (or self-selected) out of the field, owing to intrinsic, group-based unsuit ability or disinterest. To this end, theorists have hypothesized differing group-based propensities or preferences concerning philosophical intuitions (Weinberg et al. 2001; Machery et al. 2004; Buckwalter and Stich 2014), argumentation styles (Moulton 1983), or modes of reasoning (Gilligan 1993; Orr 1989). Challenging this view, Louise Antony argues that, “the offending characteristic in philosophy must track some stable difference . . . or the explanation fails” (2012, 229). The existence of stable, intrinsic differences between social groups concerning their philosophical intuitions, their reasoning modes, or argumentation styles has not been reliably demonstrated, and the hypothesis is empirically unsupported.11 Moreover, the “different voices” hypothesis assumes controversial and unsubstan tiated forms of essentialism, and so we might also reject it on these grounds. In contrast to this hypothesis, my account proposes a stable—though, import antly, not intrinsic—difference shared by underrepresented practitioners: namely, an increased susceptibility to epistemic injustice in a range of philosophical spaces. First, my proposal explains the underrepresentation of persons with non-dominant social identities in virtue of their increased susceptibility to identity-based testi monial injustice. But the amount of credibility or epistemic standing a person is afforded is a decidedly extrinsic feature of that person in that it depends, in large part, on an audience’s perceptions of the contributor and contribution. Thus, my proposal does not depend on essentialist claims about these contributors con cerning, for instance, their actual abilities, interests, intuitions, or philosophical preferences. Unlike the “different voices” hypothesis, my account is sensitive to the intrinsic heterogeneity within non-dominant social identity groups, while honing in on a key similarity: namely, their increased susceptibility to epistemic injustice. On my view, the lack of identity diversity in philosophy is partially explained by this “stable difference.” In addition, however, my proposal also offers an explanation for the lack of content diversity observed in the field. As I have argued, not only are contributors with non-dominant social identities increasingly susceptible to epistemic injust ice in philosophy (via identity-based testimonial injustice), but so too are con tributors engaged in discourses associated with non-dominant social identities (via content-based testimonial injustice). On my account, the dual and com pounding effects of each form of testimonial injustice unify and partially explain 11 For example, see Fine (2010); Lam (2010); Adleberg et al. (2015), Kim and Yuan (2015).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
EPISTEMIC INJUSTICE IN PHILOSOPHY 229 the pervasive lack of both identity and content diversity in the field. Thus, my account explains more, while making fewer assumptions. Second, my view should be distinguished from the hypothesis discussed by Cheshire Calhoun (2009) and Sally Haslanger (2008) according to which the lack of identity diversity in philosophy is attributable to the difficulty diverse practi tioners may have in “imagining” (Calhoun 2009, 218) themselves as philosophers, owing to divergence between their conceptions of themselves and their concep tions of philosophy as white and hypermasculine. Drawing on Virginia Valian’s (1998) notion of a schema, Haslanger refers to these divergences as “schema clashes” (2008, 212). Haslanger states: It is difficult for women to feel “at home” in a hypermasculine environment since it requires sublimating potentially important aspects of identity; because some of the specific elements of masculinity that are emphasized in philosophy are also associated with whiteness, the same is true for [racial] minorities. (217)
As my analysis shows, I agree with these theorists that philosophy is white and hypermasculine, that it is broadly perceived as such, and that some (or even many) diverse practitioners must sublimate important aspects of their identity to feel “at home” in the discipline. On the hypothesis I am proposing, however, diverse practitioners do not feel “at home” in philosophy because they must contend with epistemic and communicative injustices when attempting to participate in the field. Thus, my hypothesis avoids generalizations concerning the self-conceptions of diverse practitioners. The problem that I have elucidated, namely, a disproportionate susceptibility to epistemic injustice in the field, is explanatory even in the absence of any schema clash on the part of diverse practitioners themselves. For on my account, it is not the diverse practitioners, but rather their interlocutors, who possess the comparatively curbed imaginations. Moreover, my account suggests that at least some of the characteristics that Haslanger associates with hypermasculinity (and hence, philosophy)—for example, competitive, combative, non-nurturing, and judgmental (217)—may be the wrong ones on which to focus. Consider the findings of a study by Thompson et al. (2016), which directly examined several hypotheses proposed to explain the underrepresentation of women in philosophy. The study examined the ease with which women participate in introductory philosophy classrooms, and—consistent with my proposal—revealed that women scored “significantly lower than men” with respect to their comfort speaking in class (2016, 9). Yet the researchers found no evidence to support the hypothesis “that women find Intro classes to involve overly aggressive or confrontational discussions” (13). This finding suggests that traits like competiveness and non-nurturance may be insuf ficiently explanatory. The researchers did note, however, that no questions directly
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
230 Emmalon Davis tested the hypothesis “that men in Intro classes dominate class discussion and may be called on more by Instructors, which women notice and dislike” (13). Concerning this hypothesis, a survey of existing empirical research cited several studies supporting it. The survey found “that women are less comfortable than men speaking in class, including expressing their opinions and asking/answer ing questions, and that this difference partially accounted for the relationship between gender of respondent and their willingness to take more philosophy” (Thompson 2017, 7). Particularly illuminating are the findings of a qualitative study (Lockard et al. 2017) utilizing female-identified focus groups.12 In describing the epistemic role of social identity in their classroom experiences, several students noted that, with the exception of a “selected” few women, “men dominated the class conver sations” (12) and that there were more men than women in their classes. Concerning her professor, one student reported that “the girls at the beginning of the semester had to have a little more weight behind what they said” (13), while several other students mentioned that professors played “favorites.” Alternatively, other students noted the role of identity-coded content, stating that they “wished that gender was a topic of conversation even when the material was not explicitly about women or gender” (12). The researchers concluded that “students’ ‘class room experiences’ . . . might be a more pronounced factor in students’ overall assessments of philosophy” (18). Of course, this data is limited both in terms of sample size and concerning the range of social identities studied. If, as I have been arguing, diverse practitioners are experiencing (or anticipate experiencing) epistemic and communicative disadvantages in philosophical exchanges in virtue of both identity and content, then our investigations ought to continue to explore this hypothesis further. Finally, let me distinguish my proposal from Kristie Dotson’s (2012b) assess ment of the field’s inability to sustain diverse practitioners, in which she identifies philosophy’s symptomatic “culture of justification” as a major contributing factor. Typified by the centrality of the question “How is this paper philosophy?,” Dotson’s account raises concerns about philosophy’s valorization of “justification as a method,” which privileges practices of legitimation preoccupied with a singu lar set of “presumed commonly-held, univocally relevant justifying norms” (6). She argues that philosophy’s privileging of this method—and the ensuing cul ture—creates an inhospitable environment for diverse practitioners in the field, who must disproportionately shoulder the burden of justifying their practices, projects, canons, and methodologies. She advocates instead for a culture of praxis 12 Diverse practitioners—especially at the undergraduate level—may not readily identify class room discomfort as a consequence of injustice and discrimination or may interpret epistemic barriers as evidence of their own philosophical, intellectual, or communicative inadequacies. Consequently, qualitative data in which practitioners describe their classroom experiences in greater detail are espe cially useful.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
EPISTEMIC INJUSTICE IN PHILOSOPHY 231 which recognizes a variety of canons, multiple methodologies of disciplinary validation, and the value of studying live concerns. I share with Dotson the view that a culture of praxis within philosophy would contribute to the development of a more attractive working environment. But my account should be distinguished from Dotson’s for two reasons.13 First, content- based testimonial injustice identifies a more general—and in some instances, more superficial—phenomena than that with which Dotson is primarily concerned. Because a philosophical discourse can become identity-coded in the absence of methodological divergence or departure from accepted legitimizing norms, content- based testimonial injustice may identify a broader range of contributions with respect to which diverse practitioners may be epistemically targeted. Second, however, the proposal I am developing suggests that operating alongside—perhaps even prior to—the burden of legitimizing specific projects, practices, or methodologies in philosophical spaces, practitioners with non-dominant social identities are first/also saddled with the burden of simply trying to participate in any capacity at all. On my proposal, a contributor may ex peri ence barriers to philosophical participation solely in virtue of social identity—regardless of content, project, practice, or methodology. That is, con tributors with non-dominant social identities may experience an inhospitable epistemic environment in philosophy, via identity-based testimonial injustice, while at the same time violating no justificatory norms.14 Thus, while a shift to a culture of praxis would undoubtedly increase the range of habitable options within the profession (thereby improving the environment for contributors engaged in non-dominant social identity-coded discourses), it would not necessarily improve the experiences of contributors with non-dominant social identities—and especially novice contributors—who (attempt to) engage philosophy’s more conventionally credited discourses. The epistemic environment in those discourses would likely remain inhospitable for people with their social identities. To this end, the dual mechanisms of identity-based and content-based testimonial injustice cast a wider explanatory net. Ultimately, determining which factors contribute to the demographic dispar ities in philosophy is an empirical matter, and it is no doubt true that myriad 13 It should be noted that Dotson states that she is focused on the “environment of professional philosophy for diverse practitioners who have made the choice to pursue philosophy as a career path” (emphasis added), and not—as I am presently—on “why traditionally conceived diverse peoples . . . are not attracted to philosophy as a career path” (5). Thus, while the view she articulates contributes to our understanding of this latter question, she did not intend to provide an explanation. Nonetheless, I consider her account here to highlight the relevant differences between our views. 14 Unless, for example, simply possessing a non- dominant social identity in the philosophy classroom constitutes a departure from justificatory norms on Dotson’s view. If this is the case, then I would argue that we ought to distinguish between justificatory norms that exclude persons in virtue of their social identities and justificatory norms that exclude persons in virtue of characteristics of their contributions. This distinction would not preclude the possibility that some justificatory norms could exclude on the basis of both, but it would allow us to track each phenomena independently.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
232 Emmalon Davis factors converge to create what Antony (2012) has called “The Perfect Storm.” In differentiating my proposal from related accounts, I am not claiming that these proposals have no explanatory role, nor am I suggesting that my proposal is the only, or even the most important, consideration at play. While my proposal does not purport to offer a full explanation, it can, I think, inform our understanding of why we continue to make so little progress in our efforts to diversify the field and especially why we see a substantial decrease in the proportion who pursue philosophy after their first experiences with it. Thus, the account I am proposing here should be understood to illuminate a hypothesis which warrants further investigation. As Gina Schouten notes, “we have a long way to go before we really know which mechanisms are at work. In the meantime, philosophers can deploy our own tools of inquiry to improve our discipline” (2016, 276). As I have been arguing, the distinction between identity-based and content-based testimonial injustice is one such tool.
5. The ‘Philosophical Uniqueness’ Objection One might worry that insofar as diverse contributors are at an increased risk for testimonial injustice, this will be true in any academic field and not just in phil osophy. Thus, one might object that my proposal is not suited to explain philoso phy’s peculiar demographic disparities. Indeed, philosophy fares particularly poorly in comparison to other humanities disciplines (instead, more closely resembling non-humanities fields like computer science and engineering) and this indicates there is something unique about philosophy’s diversity problem. Because the mechanisms (e.g. epistemic injustice) deployed in my hypothesis are generally widespread, my hypothesis would not explain the exceptional disparities observed in philosophy. Does my account have anything to say about this? I admit that the effects of epistemic injustice—both identity-based and content-based—are undoubtedly widespread (though perhaps not evenly spread) throughout the academy. The fact that my proposal can help explain why diverse practitioners may likewise be underrepresented in other academic disciplines is a virtue of my account. Despite the broader application of my proposal, however, I want to suggest several ways in which the effects of pervasive epistemic injustice will result in unique disadvantages for philosophers, and especially for novice philosophers at early stages of exposure. First, it seems to me that the effects of epistemic injustice will be uniquely problematic in the discipline of philosophy, a discipline which Laurence BonJour has characterized—in a book written for and taught to students—as “essentially dialectical in character” (2010, viii, my emphasis added). According to BonJour, philosophical practice consists of “arguments and responses and further
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
EPISTEMIC INJUSTICE IN PHILOSOPHY 233 arguments and further responses back and forth among the different positions on a given issue” (viii). While BonJour thinks engagement in this dialectical practice is essential for all practitioners in the field, he suggests that such participation is especially important for students in the beginning stages of philosophical inquiry. Indeed, BonJour states, it “is important for a student who wants to understand this dialectical development to become, to some extent at least, a participant rather than a mere observer” (viii). While thoughtful practitioners may challenge BonJour’s conception, this con ception of philosophy is commonly held and frequently communicated to many students upon their introduction to the field. Insofar as targets of epistemic injustice are denied an equal opportunity to participate in open-ended dialectical exchanges, a practice widely characterized as essential to doing philosophy, their ability to participate in the field will be greatly diminished. Indeed, if we recall the sorts of difficulties faced by the targets of testimonial injustice (lack of recogni tion for one’s contributions, difficulty shaping the direction of conversation, limited opportunities to disseminate ideas, frequent interruptions, being ignored or dismissed), this is precisely the sort of interference that renders participation in a dialogic exchange incredibly difficult (not to mention utterly unenjoyable), if not completely impossible. In addition to other challenges associated with “doing philosophy”—for example, parsing out difficult texts, constructing and evaluating arguments, etc.—targets of testimonial injustice must overcome substantial epi stemic and communicative hurdles to engage in basic levels of inquiry. Thus, while testimonial injustice is no doubt present in other disciplines, its effects will likely introduce discipline-specific disincentives in the philosophy classroom. Insofar as novice philosophers are taught that the very practice of philosophy essentially transpires through participation in dialectical exchanges, testimonial injustice in the philosophy classroom poses a distinct disadvantage. This is par ticularly important considering the fact that for many people, their first and per haps only exposure to philosophy as a discipline is in a college classroom, where this is not the case for other disciplines like, for instance, English.15 Second, epistemic injustice likely occurs with greater frequency in fields like philosophy precisely because diverse practitioners are so seriously underrepresented 15 One might object that English is similarly dialectical but does not share philosophy’s demo graphics. A difference may partly consist in the space between the characterization of a discipline as primarily (even centrally) dialectical and as essentially dialectical, where the latter is far more restrict ive. Many students in English departments study creative writing, technical writing, grammar, English language education, etc., none of which are typically characterized as “essentially dialectical” in BonJour’s narrow sense. More importantly, however, it seems to me no accident that while English has long validated a variety of communicative media (e.g. plays, autobiography, short stories, poetry, oral history), philosophy has not. For a variety of reasons, including economic necessity, expressive poten tial, political utility, and the preservation of marginalized intellectual traditions, the philosophical contributions of diverse thinkers have historically emerged in an abundance of media beyond the singularly prized “philosophical treatise.” Diverse practitioners may find happier homes in academic fields where the inclusion of these contributions is regarded as a feature, not a bug.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
234 Emmalon Davis in those spaces. There is good reason to think that the effects of identity-based testimonial injustice are less pronounced in exchanges between individuals with shared social identities, or, in the case of content-based testimonial injustice, between individuals with shared openness towards the content of a discourse. This is plausible for two reasons. First, shared knowledge of the reality of what it is actually like to have a certain social identity (or shared knowledge of the value of a discourse) fosters epistemic familiarity, removing the epistemic function of stereotypes or false ideologies on which non-group members or non-participants often rely to (mis)assess their interlocutors. Second, when non- dominantly situated persons experience and reflect upon the false and invalidating nature of the prejudicial evaluative practices to which they have been subjected, they are often better equipped than their dominantly situated counterparts to identify other instances of such practices at work. If this is correct, then the likelihood that diverse practitioners will experience testimonial injustice in a given exchange decreases as their representation in that environment increases. Two observations follow. First, philosophy classrooms and spaces in which contributors with non-dominant social identities occupy a critical mass—as opposed to being grossly outnumbered—generally reduce (although do not entirely eliminate) the likelihood of identity-based testimonial injustice. Second, philosophy classrooms and spaces in which contributors are engaged together with shared openness to non-dominant identity-coded content—as opposed to environments in which this content is denigrated or wholly absent—generally reduce (although again, may not entirely eliminate) the likelihood of content-based testimonial injustice. Each of these environments provide different benefits, and thus they do not perform the same ameliorative role, but both kinds of envir onments offer more favorable spaces in which diverse students could consider a future in philosophy.16 This observation is thus bittersweet. It is sweet, in that small but consistent shifts in the demographics of the field can produce more
16 Indeed, the average philosophy classroom is typically inadequate in one or both capacities and thus provides a subpar environment in which to recruit students to the field. Programs like PIKSI (Philosophy in an Inclusive Key Summer Institute), Compass Workshops, and the Rutgers Summer Institute for Diversity in Philosophy have served to develop the first sort of environment, in which diverse students are engaged in a broad range of philosophical discourses in spaces in which they occupy a critical mass. Unlike these programs, many departments do not have direct control over the identities of the students in their classrooms, but departments do have control over the content offered at the introductory level (and beyond!). Thus, one way to shift demographics is to provide adequate support for and encouragement of content diversity in classes. This extends beyond diversifying the contents of a “generalized” introductory syllabi but to also diversifying the range of courses that could be offered at the introductory level: for example, Philosophy of Race, Philosophy of Gender, Social and Political Philosophy, Chinese Philosophy, Philosophy of Disability, etc. In this way, diverse students would be able to choose from a wider range of content through which to engage philosophically, and such courses would naturally provide environments in which students enrolled would be more likely to share with one another an openness to the course content. The aim of cultivating classrooms in which students demonstrate openness to diverse content and to diverse identities is crucial, as merely diversifying a syllabus does not by itself reduce the possibility of testimonial injustice in the classroom.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
EPISTEMIC INJUSTICE IN PHILOSOPHY 235 sizeable impacts down the road. It is bitter, however, in that attracting diverse students into philosophy under the auspices that the field will thereby be hospit able to them is morally suspect. Indeed, there is an undeniable moral tension in recruiting diverse practitioners into an inequitable working environment because the environment needs more diverse practitioners to become equitable. From this perspective, it appears as though philosophy needs to have its diversity problem fixed before it can truly fix its diversity problem.
6. A Distinction without a Difference? Utilizing a distinction between identity-based and content-based testimonial injustice, I have developed a two-pronged epistemic explanation for the lack of diverse practitioners in academic philosophy. I argued that (i) contributors with non-dominant social identities and (ii) contributors engaged in non-dominant social identity-coded discourses are both susceptible to epistemic injustice in the field. While the former are targeted in virtue of their social identities, the latter are targeted in virtue of the content of their contributions; the injustices experi enced by contributors at the intersections are dual and hybrid. One advantage of my proposal, then, is that it connects the underrepresenta tion and marginalization of (i) and (ii), while avoiding circular analysis. But one might wonder whether a non-circular analysis is useful. There is evidence to sug gest that philosophers engaged in gender discourses in the field are more likely to be women and gender minorities and that those engaged in race discourses are more likely to be people of color.17 Given this overlap, why not just explain the underrepresentation of one in virtue of the other, or vice versa? Indeed, one might wonder whether a non-circular analysis—and hence the distinction between identity-based and content-based testimonial injustice—is useful. That is, does the analysis employ a distinction without a difference? First, although (i) and (ii) overlap, they are not identical. For instance, many (though not all) philosophers of race are people of color, and not all people of color in philosophy have interests in philosophy of race (or other race-coded
17 Haslanger (2010) found that feminist philosophy was among the most frequent areas in which women published and was the least frequent area for men. Schwitzgebel and Jennings (2017) found that the proportion of women in ethics—broadly construed to include social and political philosophy, philosophy of race, gender, and sexuality, aesthetics, normative ethics, applied ethics, and philosophy of law—“substantially exceeded the proportion in other areas of philosophy among U.S.-ranked fac ulty [Philosophical Gourmet Report], recent job placements, APA program participants, authorship in elite journals, highly cited authors, and targets of extended journal article discussion” (29). Paxton (n.d.) reported that the philosophical subfield of philosophy of race, gender, and sexuality is 81 per cent women. Botts et al. (2014) found that the top areas of specialization for US black philosophers were Africana, race, social and political, ethics, and continental philosophy; for black women philo sophers, these areas were race, ethics, continental, social and political, and feminism.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
236 Emmalon Davis discourses). Likewise, while most practitioners engaged in gender discourses are women and gender minorities, many women and gender minorities in philoso phy are not engaged in these discourses. My proposal is compatible with existing overlap without presuming overlap or relying on overlap to generate an explan ation. Persons with non-dominant social identities possess a variety of philosoph ical strengths and interests. Our efforts to diversify the field must be attentive to overlap without blurring the distinction between identity and philosophical interest. If we treat (i) and (ii) interchangeably or collapse them, we risk foreclos ing a wide range of philosophical opportunities for persons with non-dominant social identities—especially for those in the earlier stages of philosophical exploration. Second, distinguishing between (i) and (ii)—and thus between identity-based and content-based testimonial injustice—is useful because it enables us to articu late why structural and institutional efforts aimed to ameliorate one form of injustice can fail to address (and can even exacerbate) problems of the other. Consider, first, how efforts to ameliorate identity-based testimonial injustice can proceed in the absence of attention to content. Imagine a philosophy graduate program that, during the five years in which a graduate student might be enrolled, facilitates approximately 30 invited philosophy talks (averaging, say, 6 talks a year). Suppose that during this time, 100 per cent of the speakers are white and 80 per cent of the total speakers invited are cisgender men. Moreover, all of the invited speakers—regardless of social identity—give talks in mainstream epis temology, metaphysics, philosophy of language, mind, and, perhaps, occasionally ethics or history. Suppose, further, that the department eventually becomes con cerned with the demographics of their department colloquia and diligently works to introduce greater diversity in the race and gender identities of their speakers. Yet, as it turns out, none of the newly invited speakers work in areas outside the mainstream. While the department has made efforts to alleviate the effects of identity-based testimonial injustice, they have done nothing to remedy the effects of content-based testimonial injustice. Concerning the kinds of dis courses in which invited speakers are engaged, the resulting colloquia looks exactly the same as it did before. These exclusions are detrimental to the epistemic and professional lives of those practitioners who are pre-emptively passed over, but these harms have a more expansive reach. Invited talks serve to expose graduate students to philoso phy’s breadth and to demonstrate the range of inquiries it is possible for one to undertake as a philosopher; they provide students with opportunities to learn from and interact with other professionals who share their interests and to develop conversational competence in areas of philosophy about which they are less knowledgeable. Patterned content exclusions deprive students of these oppor tunities, effectively communicating to them that such content does not belong in philosophy departments. (This message is likewise communicated through
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
EPISTEMIC INJUSTICE IN PHILOSOPHY 237 curricula, concerning the kinds of courses offered and required for students to advance professionally.) Students unfamiliar with diverse discourses will continue to remain ignorant of them due to lack of exposure, and students with interests (or potential interests) in these areas will be less likely to pursue them if they perceive that such interests have no place in the profession more broadly. Some students may leave the discipline to pursue these interests elsewhere. Because graduate students already provide a substantial portion of existing undergraduate instruction (especially at the introductory level) and constitute the pool from which the next generation of faculty and instructors are selected, the effects of content- based testimonial injustice reverberate throughout the profession. Transformative interventions at the level of colloquia and curricula require a collective expectation that familiarity with and respect for diverse discourses are an integral part of a philosophy education; such interventions demand the intentional inclusion of valued and adequately compensated practitioners who knowledgeably engage in these investigations in their teaching and research. In the absence of explicit attention to content, efforts that increase identity diversity may produce limited effects at the level of discourse. By the same token, interventions aimed at increasing content diversity in phil osophy can fail to address identity-based inequalities. Consider, for example, the recent publication of a journal symposium on Black Lives Matter, in which no contributions from black philosophers were included. In an open letter to the edi tors, Chris Lebron (2017)—a black philosopher who has written extensively on Black Lives Matter—observed that, in the five years leading up to the symposium’s publication, the journal: has not published a single article on the philosophy of race: voting, elections, immigration, global markets, and animals have gotten their time in the journal’s sun. But as black Americans, and the philosophers who study racial inequality— a political philosophical problem—have directly engaged one of our era’s most sinister moral and political quandaries, the journal has failed to represent race in its pages. Maybe more damning, so far as I can tell, not one black philosopher has seen her or his work appear in the pages of your respected journal, on race or any other topic.
As Lebron notes, the journal’s publication record reveals a systemic two-fold lack: the contributions of black philosophers (regardless of content) and contributions to race discourses (regardless of the identities of the contributors) are perpetually absent.18 Against this backdrop, the publication of a special symposium on Black Lives Matter in which no black philosophers are included is particularly glaring. 18 One might respond that if the problem lies in the fact that black contributors did not submit their work, then the fault does not lie with the journal. This response is inattentive to the mechanisms
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
238 Emmalon Davis By failing to include any black thinkers—not even flagging the troubling irony of this deficiency—the possibility of black participation in that platform was pre-emptively foreclosed. To be clear, the claim that black-identified perspectives were wrongfully erased should not be confused with the claim that no non-black voices ought to have been included. As Lebron states, “[i]t is important that there be a range of viewpoints on a matter that is democratically urgent—we are all involved in this problem.” But the wholesale absence of black perspectives disre gards the contributions of persons whose identities are most centrally relevant to the content considered. What such a case illustrates is that efforts towards greater content inclusivity in a given domain can further exacerbate identity-based testi monial injustice. While identity-based and content-based testimonial injustice are inextricably linked, their remedies have not necessarily advanced hand in hand. Finally, the distinction between identity-based and content-based testimonial injustice allows us to identify the potentialities and pitfalls associated with vari ous positions which diverse practitioners in the field may occupy. This under standing can facilitate the development of more accurate guidance and mentoring structures for less advanced diverse practitioners as they navigate their own place in the discipline. Furthermore, efforts to implement institutional, structural, and interpersonal changes in the field will be more successful with greater knowledge of the unique forms of vulnerability or advantage that attach to each position. Indeed, not all who are marginalized in the field are equally marginalized. Those with comparative social privilege may enact epistemic injustice within the mar gins, and institutional remedies often favor the most privileged. To this end, consider first the position occupied by contributors picked out by (i) above exclusively. Because their primary engagements are in dominantly cred ited discourses, they are not generally at risk for content-based testimonial injust ice. Such contributors are, however, more likely to find themselves to be “the only one” with their identity in a given domain. Because, as I have argued, severe underrepresentation increases the likelihood that one will be targeted, such con tributors may face exceptional epistemic barriers—in the form of identity-based testimonial injustice—in exchanges with dominant interlocutors. In many instances, such barriers may entirely preclude opportunities for participation. This may explain the continued, disproportionate underrepresentation of persons with non-dominant social identities in dominantly credited philosophical dis courses; occupation within those spaces may be lonely and discouraging. When successfully attained, however, acceptance within a dominantly credited dis course can usher in sizeable advancements in one’s epistemic (and professional) through which testimonial injustice—operating pre-emptively and structurally—could unjustly prod uce this effect.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
EPISTEMIC INJUSTICE IN PHILOSOPHY 239 standing. For such contributors, the deployment of language and conceptual frameworks associated with dominantly valued discourses renders one’s partici pation acceptable—even welcomed—insofar as one’s philosophical contributions do not generally disrupt underlying hierarchies of value and displacement. Nonetheless, the privileged positions attained by such individuals provide unique platforms from which to actively implement structural change. What is called for are seismic shifts in the practices of inclusion, visibility, and accreditation, and those occupying a position of enhanced professional legibility can utilize their status to help facilitate these goals. By contrast, contributors picked out exclusively by (ii) above may experience increased susceptibility to content-based testimonial injustices in virtue of their engagement in dominantly devalued discourses, but their comparatively privil eged social identities largely shield them from more systemic forms of epistemic disadvantage. Because a privileged social identity can sometimes (though not always) inoculate against the force of content-based testimonial injustice, these contributors may be better positioned to engage devalued discourses in certain domains. For such contributors, however, increased susceptibility to content-based testimonial injustice is perfectly compatible with the fact that their participation in those discourses may itself perpetuate further epistemic inequalities against others. For example, some contributors may lack an appropriate standing—they may coopt contributions developed by more marginalized contributors or they may produce false or distorting information within a discourse. (This, too, can happen when contributors with devalued social identities are engaged in discourses associated with devalued social identities that are not their own.) Thus, mere participation in a discourse does not guarantee that one’s contributions within that discourse will be justified, even if one assumes some epistemic risk— via content-based testimonial injustice—in participating. Institutional change requires networks of allied participants transforming the practices through which knowledge is collectively produced, but such networks will foster further inequality if the participation of the relatively privileged serves to crowd out or appropriate contributions of those whose interests are centrally at stake. Consider, finally, those practitioners occupying positions at the intersections of (i) and (ii). These practitioners are vulnerable to an assemblage of identity-based and content-based testimonial injustices. Perhaps, owing to frequent acts of dis missal or discrediting, they have been prevented from participating in dominantly credited discourses in which they have genuine interests. Perhaps they have been pigeon-holed into discourses concerning their identities, as these discourses are prejudicially deemed more “fitting” for persons like them (Davis 2016). Some may be pressured or coerced into engaging in such discourses, where their epi stemic labor is thereby exploited for the benefit of dominant interlocutors and institutions (Berenstain 2016). Others may engage non-dominant discourses to
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
240 Emmalon Davis enrich connections with communities or to preserve and develop intellectual traditions that are centrally important to them. For others, engagement in such discourses constitutes what Camisha Russell identifies as deliberate occupation, much like “the way that one might occupy a lunch counter, an administrative building, or Wall Street . . . in protest or as a sort of demand for recognition” (2019, 184). Because of the unique relationship between their identities and the discourses they engage, these practitioners negotiate a constellation of epistemic double binds as they navigate the profession. Does one become a discredited contributor in a dominantly credentialed discourse or a credentialed contributor in a dominantly discredited discourse? Such practitioners may discover that the only discourses in which their contributions are not actively rejected are those discourses which centrally concern their social identities. Alternatively, they may be discredited for participating in identity-coded discourses precisely because they are judged to lack objectivity with respect to the subject matter. They may be assessed as too partial in their investigations or find that their contributions are dismissed as self-serving or self-interested. A double bind tightens into a Gordian knot. Since marginalized discourses remain largely separate from dominant-occupied discourses, however, these discourses may provide practitioners with intra-discursive epistemic advantages. Practitioners with non-dominant social identities are more likely to constitute a critical mass in these spaces, and many practitioners may find that such spaces offer comparably better epistemic envir onments in virtue of this fact.19 With greater freedom from the dictates of domin ant voices, marginalized discourses have long provided increased opportunities for philosophical participation, collective contestation, professional affirmation, and enhanced disciplinary (and interdisciplinary) imagination. But this observa tion by no means justifies their institutionalized marginalization. As Margaret Urban Walker states, “[s]egregation, here as elsewhere, does not conduce to equal respect” (2005, 160). Persons with non-dominant social identities deserve equal access to philosophical institutions, and institutionalized marginalization does not constitute such access. This is the overarching institutional context into which diverse students arrive. The interrelation between identity-based and content-based testimonial injustice is mediated by a plurality of individual perceptions and biases, systemic ideolo gies, organizational practices, structural hierarchies, and institutional networks of entrenched power. Epistemic justice in philosophy thus requires increased atten tion to the multiple ways in which these influences work together to advantage or 19 To say that an epistemic environment is comparably “better” is not to deny the real effects of unjust exclusionary practices or intra-level prejudices and hierarchies. The environment need not be an epistemic utopia to appear better than the alternatives.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
EPISTEMIC INJUSTICE IN PHILOSOPHY 241 disadvantage differently situated contributors across a wide range of contexts. As Lebron (2017) puts it, “diversity really is an ethically important ideal.” As this chapter has argued, diversity is also an epistemically important ideal.
7. The Epistemic Harm of Content-Based Testimonial Injustice Part of the task of this chapter was to conceptually distinguish between two forms of testimonial injustice. One might worry, however, that I have not adequately shown that content-based testimonial injustice is an epistemic injustice, in that I have not articulated its unique epistemic harm. The precise nature of the epi stemic harm of testimonial injustice is a matter of ongoing debate. On one view, for instance, testimonial injustice constricts a targeted contributor’s exercise of one’s own epistemic subjectivity, in accordance with dominant interests (Pohlhaus Jr 2014). On another view, epistemic injustice objectifies targeted contributors and demonstrates a basic lack of respect for their capacities as knowers (Fricker 2007). In central cases, the relevant social identities (and attendant preju dices) are the sort that “track” targets across a variety of social contexts—render ing them susceptible to further injustices in various domains (e.g. legal, economic, etc.). A commonality across various accounts is that testimonial injustice harms contributors in their epistemic capacities and takes on a broader—that is, sys temic—epistemic significance. Given this understanding, how ought we charac terize the harm of content-based testimonial injustice and to whom does it befall? Two worries present themselves in answering this question. First, one might worry that because content-based testimonial injustice is not directly (or, in hybrid cases, not exclusively) prompted by the targeted contributor’s social iden tity, it is not obvious in what way the contributor is harmed. Second, one might worry that because content-based testimonial injustice arises in fairly localized contexts, the associated harm therefore lacks any systemic significance. I address each in turn. Let me first articulate the epistemic harm experienced directly by targeted con tributors. Consider two cases introduced by Fricker. The first case concerns a group of scientists who, upon submitting a manuscript to a journal, receive a credibility deficit from a panel of referees harboring a “dogmatic prejudice” against the authors’ research method. The second case involves a small group of philosophers of science at an international science conference—predominately attended by research scientists and historians of science—where “simply falling into the identity category ‘philosopher of science’ renders one’s word likely to be dismissed as . . . vain speculations” (28). Though Fricker does not distinguish between identity-based and content-based testimonial injustice, she characterizes both cases as instances of testimonial injustice. But the testimonial injustices involved are, she says, incidental, rather than systematic. That is, the relevant
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
242 Emmalon Davis prejudices—about a scientific research method or the professional identity ‘philosopher of science’—are restricted to the highly localized academic contexts in which they arise. Thus, these are not the sort of structural prejudices that “render the subject vulnerable to any other kinds of injustice” (27) in additional domains, such as visiting a doctor’s office, filing through airport security, or taking out a loan, etc. To say that incidental testimonial injustices generally lack the broader social significance of systematic cases, however, is not to say that such harms are neces sarily morally inconsequential for those targeted. Indeed, says Fricker, “localized prejudices and the injustices they produce may be utterly disastrous for the sub ject, especially if they are repeated frequently [. . .] the accumulation of incidental injustices may ruin their life” (29). Thus, while systemic and incidental testimo nial injustice are different in kind, incidental testimonial injustice may nonethe less be harmful (and especially so when persistent). With this distinction in mind, what are we to make of the cases I have con sidered throughout this chapter, in which contributors engaged in non-dominant social identity-coded discourses experience testimonial injustice in academic philosophy. One might think these cases—which I have identified as content-based testimonial injustices—are analogous to the academic science cases described above. Accordingly, one could argue that the testimonial injustices experienced by contributors engaged in gender or race discourses in philosophy are inciden tal; that is, they are to be explained by the fact that such contributors employ methods against which mainstream philosophical audiences harbor “dogmatic prejudices” or that the professional identities “feminist philosopher” and “philosopher of race” are subject to localized identity prejudices in the field. Supposing this were true (as it may well be in some instances), classifying these testimonial injustices as incidental would not be to say that the targeted contributors are not experiencing testimonial injustices. Rather, it is to say that they experience epistemic harms that only disadvantage them in a particular domain. But the localized nature of the harm does not dissolve the claims to justice that those t argeted have against those who unjustly target them within that domain. Indeed, even if the harm did not take on any larger social significance, it would nonetheless warrant some ameliorative response within the localized contexts in which it arises.20 20 Because there is some risk that this conceptual tool will be misappropriated, I emphasize here that testimonial injustice picks out epistemic exclusions that are unjust. Thus, testimonial injustice cannot be used to theorize the justified marginalization or rejection of epistemic contributions that, for example, have been shown to be untenable (e.g. phrenology) or which exclusively serve to foster and preserve injustice (e.g. white supremacy). Excluding such content is not appropriately theorized as an injustice, not even incidental. Consequently, testimonial injustice will not be an appropriate mechanism for explaining all instances of epistemic exclusion in a given domain, and not all exclu sions require an ameliorative response.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
EPISTEMIC INJUSTICE IN PHILOSOPHY 243 This analysis of the harm of content-based testimonial injustice—as it affects the targeted contributor directly—concerns one mode of epistemic subjectivity, namely, regarding epistemic subjects in their capacities as subjects who know (who inquire, understand, etc.). Epistemic subjectivity, in this mode, concerns persons in their capacities to testify, to share, to investigate, and to otherwise engage meaningfully in an epistemic exchange. Both identity-based and content-based testimonial injustice can harm targeted contributors in this first mode. Identity-based testimonial injustice thwarts a contributor’s epistemic subjectivity in virtue of one’s social identity, where this harm is appropriately characterized as systemic. Content-based testimonial injustice thwarts a contributor’s epistemic subjectivity in virtue of identity-coded content in one’s contribution, where this harm is appropriately characterized as incidental. Because it is incidental, the epistemic significance of content-based testimonial injustice for the targeted contributor will thus depend upon a range of contextual factors. For example, when targeted contributors occupy comparatively privileged social positions, the constraints on their epistemic subjectivity introduced by content-based testimo nial injustice may be highly localized or infrequent (and in some cases, the effects may be rendered morally insignificant). For less privileged contributors, however, the epistemic harm will be more substantial, especially when the injustices are persistent. For those contributors who are vulnerable to both identity-based and content-based testimonial injustice, incidental content-based testimonial injustices may intertwine with and magnify the wider range of systemic constraints on their epistemic subjectivity. But an analysis that locates the epistemic harm of content-based testimonial injustice exclusively at the level of targeted contributors in individual exchanges will obscure its broader significance. This is because the cases I have considered are not wholly analogous to Fricker’s cases. As I have argued, content-based testi monial injustice is mediated by a perceived link between the content of a con tributor’s contribution and social identities that are systemically devalued. Thus, in the cases with which I have been concerned, it is not—or at least not exclu sively—a disfavored methodology or a highly localized professional identity that is the object of prejudice. Rather, contributors engaged in non-dominant social identity-coded discourses are subject to epistemic injustice because of identity prejudice concerning people with non-dominant social identities. Unlike prejudice against a disfavored method or professional identity, this prejudice is neither locally restricted nor domain specific; rather, the underlying prejudice is sys temic—so systemic, in fact, that it not only serves to discredit people with non-dominant social identities themselves, but also serves to discredit those who are engaged in discourses variously associated with them (and, especially, those dis courses perceived to advance their interests). Thus, more can be said in defense of
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
244 Emmalon Davis the claim that the harm of content-based testimonial injustice is not merely incidental, but also systemic. To identify the systemic epistemic harm of content-based testimonial injustice, we must consider a second mode of epistemic subjectivity. Epistemic subjectivity, in this second mode, concerns persons in their capacities as subjects to be known; that is, as subjects about whom knowledge can be produced and disseminated. As subjects to be known, persons and their interests are taken seriously, their experi ences are rendered legible, and their collective perspectives are engaged and understood. Content-based testimonial injustice tracks content by, about, and associated with persons with non-dominant social identities. By fostering the idea that such content does not matter, content-based testimonial injustice pre cludes broader engagement with discourses through which marginalized persons can be understood and appropriately valued. It thereby serves as a mechanism through which dominant audiences remain willfully ignorant of those discourses (Pohlhaus Jr 2012), despite the fact that they exist and that there are contributors who work to cultivate and share them (Mason 2011). In this way, content-based testimonial injustice contributes to a process of systemic erasure and foreclosure through which marginalized persons are rendered “unknowable,” a concept Kristie Dotson (following Fannie Barrier Williams) theorizes as “a trifold struc ture of disappearing” (2017, 426) constituted by disregard, disbelief, and dis avowal. As subjects to be known, persons and their experiences are regarded both as knowable and worth knowing. Thus, content-based testimonial injustice sys tematically thwarts the epistemic interests of marginalized persons in this sec ond mode.21
21 One might wonder how content-based testimonial injustice relates to hermeneutical injustice. Do the two overlap in some capacity or does one contribute to the other? Hermeneutical injustice picks out “the injustice of having some significant area of one’s social experience obscured from col lective understanding owing to a structural identity prejudice in the collective hermeneutical resource” (Fricker 2007, 155), where contra testimonial injustice, “no agent perpetrates hermeneutical injustice—it is a purely structural notion” (159). Like hermeneutical injustice, content-based testimo nial injustice involves structural identity prejudice, which, in the latter case, operates mechanistically through social identity-coded content. But while content-based testimonial injustice can operate purely structurally (consider, for instance, academic journal practices), many of the cases I have con sidered are perpetrated by identifiable agents in testimonial or otherwise dialectical exchanges. Moreover, unlike hermeneutical injustice, content-based testimonial injustice need not involve any conceptual lacuna or obscured intelligibility on the part of the marginalized, and it rejects the notion of a “collective” hermeneutical resource as neutral with respect to issues of group domination and marginalization. While it is possible for content-based testimonial injustice to thwart the creation of conceptual resources (and hence, for it to contribute to this aspect of hermeneutical injustice), it more frequently targets the broader dissemination of already existing resources. In this sense, the problem of content-based testimonial injustice is primarily communicative, rather than conceptual. It should be noted, however, that Fricker (2016) has since distinguished between midway, minimal, and max imal cases of hermeneutical injustice, where midway cases do not involve “any confused experiences whatever, but only frustratingly failed attempts to communicate them to members of an out-group” (167). Content-based testimonial injustice may overlap more extensively with these “midway” cases, but if the central problem of midway cases is largely communicative, it less clear why they should be
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
EPISTEMIC INJUSTICE IN PHILOSOPHY 245 The cooperative interplay between these two modes of epistemic subjectivity—to know and to be known—illuminates the interdependent nature of identity-based testimonial injustice and content-based testimonial injustice. Describing what she calls “imperial harm,” Camisha Russell (2019) argues that marginalized persons can be harmed in their capacities to be known when dominant others regard them as mere objects of knowledge. Such harms occur when dominant interlocutors and institutions purport to have produced exclusive or exhaustive knowledge concerning the marginalized, or where dominant discourses super sede discourses produced by those occupying the very subject positions under consideration. As Russell’s analysis reveals, respect for marginalized persons in their capacities to be known therefore requires respect for them in their capacities to know (and especially, for instance, about themselves, their identities, their communities, their histories, their needs, and their experiences). A proper regard for marginalized persons in their capacities as subjects who know, then, necessi tates a proper regard for those discourses associated with their collective inquiry, for such discourses are both process and repository of their knowledge. Thus, identity-based testimonial injustice and content-based testimonial injustice are inextricably linked. If epistemic justice requires the elimination of one, it requires the elimination of both.
Acknowledgements This chapter began as a letter to the philosophy faculty of my undergraduate institution, who, at the time of my graduation, invited me to “report back” on my experiences as a black female student in the department. This chapter has thus been in production for nearly a decade. I would like to thank audiences at the 2013 Diversity in Philosophy Conference, 2015 ESWIP, 2017 MAP UCLA Conference, 2017 Vancouver Summer Philosophy Conference (VSPC), 2018 MIT Minds of Our Own Conference, 2018 Princeton Workshop in Social Philosophy, and 2020 MAP Glasgow Workshop, as well as audiences in the philosophy departments at the New School for Social Research, Cornell University, and Northwestern University. For written comments and discussion, I am grateful to Elizabeth Anderson, Grace Helton, Samia Hesni, Jennifer Lackey, Ishani Maitra, Noralyn Masselink, Charles Mills, Wade Munroe, Taylor Rogers, Katherine Ward, Dennis Whitcomb, Allen Wood, Robin Zheng, and an anonymous reader. Thank you, finally, to Christina Van Dyke, for being the first to show me that—despite nearly every indication to the contrary—there was a place in philosophy for both me and my interests. Your peda gogy has always been my inspiration.
characterized as hermeneutical injustices, and not, for example, as structural testimonial injustices (Anderson 2012).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
246 Emmalon Davis
References Adleberg, Toni, Morgan Thompson, and Eddy Nahmias (2015). “Do Men and Women Have Different Philosophical Intuitions? Further Data.” Philosophical Psychology (28)5: 615–41. Alcoff, Linda (ed.) (2003). Singing in the Fire: Stories of Women in Philosophy. Lanham, MA: Rowman & Littlefield. Allen, Anita and George Yancy (2018). “The Pain and Promise of Black Women in Philosophy,” New York Times (June 18, 2018), www.nytimes.com/2018/06/18/opinion/ black-women-in-philosophy.html. Anderson, Elizabeth (2012). “Epistemic Justice as a Virtue of Social Institutions.” Social Epistemology 26(2): 163–73. Antony, Louise (2012). “Different Voices or Perfect Storm: Why are There So Few Women in Philosophy?” Journal of Social Philosophy 43(3): 227–55. Beebee, Helen and Jennifer Saul (2011). “Women in Philosophy in the UK: A Report by the British Philosophical Association and the Society for Women in Philosophy UK.” Joint BPA/SWIP Committee for Women in Philosophy (September 8, 2011), www.swipuk.org/notices/2011-09-08. Belle, Kathryn S. [formerly Kathryn T. Gines] (2011). “Being a Black Woman Philosopher: Reflections on Founding the Collegium of Black Women Philosophers.” Hypatia 26(2): 429–37. Berenstain, Nora (2016). “Epistemic Exploitation.” Ergo, an Open Access Journal of Philosophy 3(22): 569–90. Bettcher, Talia (2018). “ ‘When Tables Speak’: On the Existence of Trans Philosophy.” Daily Nous (May 30, 2018), http://dailynous.com/2018/05/30/tables-speak-existencetrans-philosophy-guest-talia-mae-bettcher. BonJour, Laurence (2010). Epistemology: Classic Problems and Contemporary Responses. Lanham, MA: Roman & Littlefield. Botts, Tina F., Liam K. Bright, Myisha Cherry, Gunter Mallarangeng, and Quayshawn Spencer (2014). “What Is the State of Blacks in Philosophy?” Critical Philosophy of Race 2(2): 224–42. Buckwalter, Wesley and Stephen Stich (2014). “Gender and Philosophical Intuition,” in Joshua Knobe and Shaun Nichols (eds), Experimental Philosophy: Volume Two. Oxford: Oxford University Press, 307–46. Calhoun, Cheshire (2009). “The Undergraduate Pipeline Problem.” Hypatia 24(2): 216–23. Cohn, Carol (1993). “Wars, Wimps, and Women: Talking Gender and Thinking War,” in Miriam G. Cooke and Angela Woollacott (eds), Gendering War Talk. Princeton, NJ: Princeton University Press, 227–46. Davis, Emmalon (2016). “Typecasts, Tokens, and Spokespersons: A Case for Credibility Excess as Testimonial Injustice.” Hypatia 31(3): 485–501.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
EPISTEMIC INJUSTICE IN PHILOSOPHY 247 Dembroff, Robin (2020). “Cisgender Commonsense and Philosophy’s Transgender Trouble.” Transgender Studies Quarterly 7(3): 399–406. Dembroff, Robin and Dennis Whitcomb (forthcoming). “Content Focused Epistemic Injustice,” in Tamar Szabo Gendler and John Hawthorne (eds), Oxford Studies in Epistemology. Oxford: Oxford University Press. Dotson, Kristie (2012a). “A Cautionary Tale: On Limiting Epistemic Oppression.” Frontiers: A Journal of Women Studies 33(1): 24–47. Dotson, Kristie (2012b). “How Is This Paper Philosophy?” Comparative Philosophy 3(1):3–29. Dotson, Kristie (2017). “Theorizing Jane Crow, Theorizing Unknowability.” Social Epistemology 31(5): 417–30. Fine, Cordelia (2010). Delusions of Gender: How Our Minds, Society, and Neurosexism Create Difference. New York: W. W. Norton & Company. Fricker, Miranda (2007). Epistemic Injustice: Power and the Ethics of Knowing. Oxford: Oxford University Press. Fricker, Miranda (2016). “Epistemic Injustice and the Preservation of Ignorance,” in Rik Peels and Martijn Blaauw (eds), The Epistemic Dimensions of Ignorance. Cambridge: Cambridge University Press, 160–77. Gilligan, Carol (1993). In a Different Voice: Psychological Theory and Women’s Development. Cambridge, MA: Harvard University Press. Gracia, Jorge J. E. (2008). Latinos in America: Philosophy and Social Identity. Malden, MA: Blackwell. Haslanger, Sally (2008). “Changing the Ideology and Culture of Philosophy: Not by Reason (Alone).” Hypatia 23(2): 210–23. Haslanger, Sally (2010). “Preliminary Report of the Survey on Publishing in Philosophy.” The APA Newsletter on Feminism and Philosophy 10(1): 10–15. Hookway, Christopher (2010). “Some Varieties of Epistemic Injustice: Reflections on Fricker.” Episteme 7(2): 151–63. Hutchison, Katrina and Fiona Jenkins (eds) (2013). Women in Philosophy: What Needs to Change? New York: Oxford University Press. James, V. Denise (2014). “Musing: A Black Feminist Philosopher: Is That Possible?” Hypatia 29(1): 189–95. Kim, David Haekwon (2002). “Asian American Philosophers: Absence, Politics, and Identity.” American Philosophical Association Newsletter 1(2): 25–8. Kim, Minsun and Yuan Yuan (2015). “No Cross-Cultural Differences in the Gettier Car Case Intuition: A Replication Study of Weinberg et al. 2001.” Episteme 12(3): 355–61. Krishnamurthy, Meena (2016). “Decolonizing Analytic Political Philosophy.” Philosopher Blog (June 3, 2016), https://politicalphilosopher.net/2016/06/03/ meenakrishnamurthy.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
248 Emmalon Davis Lam, Barry (2010). “Are Cantonese-Speakers Really Descriptivists? Revisiting CrossCultural Semantics.” Cognition 115(2): 320–9. Lebron, Chris (2017). “An Open Letter to the Editors of the Journal of Political Philosophy; or How Black Scholarship Matters, Too.” Philosopher (May 24, 2017), https://politicalphilosopher.net/2017/05/24/an-open-letter-to-the-editors-of-thejournal-of-political-philosophy-or-how-black-scholarship-matters-too. Lee, Carole (2014). “Asian Americans, Positive Stereotyping, and Philosophy.” APA Newsletter on Asian and Asian-American Philosophers and Philosophies 14(1): 2–7. Lee, Emily S. (2017). “Interview,” in George Yancy (ed.), On Race: 34 Conversations in a Time of Crisis. New York: Oxford University Press, 231–6. Leslie, Sarah-Jane, Andrei Cimpian, Meredith Meyer, and Edward Freeland (2015). “Expectations of Brilliance Underlie Gender Distributions across Academic Disciplines.” Science 347(6219): 262–5. Lindemann, Hilde (2006). “Bioethics’ Gender.” The American Journal of Bioethics 6(2): W15–W19. Lloyd, Genevieve (1984). The Man of Reason: “Male” and “Female” in Western Philosophy. Minneapolis, MN: University of Minnesota Press. Lockard, Claire A., Helen Meskhidze, Sean Wilson, Nim Batchelor, Stephen BlochSchulman, and Ann J. Cahill (2017). “Using Focus Groups to Explore the Underrepresentation of Female-Identified Undergraduate Students in Philosophy.” Feminist Philosophical Quarterly 3(4): 1–29. Machery, Edouard, Ron Mallon, Shaun Nichols, and Stephen P. Stich (2004). “Semantics, Cross-Cultural Style.” Cognition 92(3): B1–B12. Madva, Alex (2016). “Implicit Bias and Latina/os in Philosophy.” APA Newsletter on Hispanic/Latino Issues in Philosophy, 16(1): 8–15. Maldonado-Torres, Nelson (2011). “Thinking Through the Decolonial Turn: PostContinental Interventions in Theory, Philosophy, and Critique—An Introduction.” Transmodernity: Journal of Peripheral Cultural Production of the Luso-Hispanic World 1(2): 1–16. Mason, Rebecca (2011). “Two Kinds of Unknowing.” Hypatia 26(2): 294–307. Medina, José (2011). “The Relevance of Credibility Excess in a Proportional View of Epistemic Injustice: Differential Epistemic Authority and the Social Imaginary.” Social Epistemology 25(1): 15–35. Mills, Charles W. (1997). The Racial Contract. Ithaca, NY: Cornell University Press. Moulton, Janice (1983). “A Paradigm of Philosophy: The Adversary Method,” in Sandra Harding and Merill Hintikka (eds), Discovering Reality. Springer, Dordrecht, 149–64. Narayan, Uma (2003). “What’s a Brown Girl Like You Doing in the Ivory Tower? Or, How I Became a Feminist Philosopher,” in Linda Alcoff (ed.), Singing in the Fire: Stories of Women in Philosophy.. Lanham, MA: Rowman & Littlefield, 81–93.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
EPISTEMIC INJUSTICE IN PHILOSOPHY 249 Orr, Deborah (1989). “Just the Facts Ma’am: Informal Logic, Gender and Pedagogy.” Informal Logic 11(1): 1–10. Outlaw Jr, Lucius (1996). On Race and Philosophy. New York: Routledge. Paxton, Molly (n.d.). “Potential Sources of Gender Disparities in Philosophy: An Examination of Philosophical Subfields and Their Relation to Gender,” American Philosophical Association, https://cdn.ymaws.com/www.apaonline.org/resource/ group/bf785b0d-eb59-41f8-9436-1c9c26f50f8e/Paxton,_Gender_Disparities_i.pdf. Paxton, Molly, Carrie Figdor, and Valerie Tiberius (2012). “Quantifying the Gender Gap: An Empirical Study of the Underrepresentation of Women in Philosophy.” Hypatia 27(4): 949–57. Pohlhaus Jr, Gaile (2014). “Discerning the Primary Epistemic Harm in Cases of Testimonial Injustice.” Social Epistemology 28(2): 99–114. Pohlhaus Jr, Gaile (2012). Relational Knowing and Epistemic Injustice: Toward a Theory of Willful Hermeneutical Ignorance. Hypatia 27(4): 715–35. Rooney, Phyllis (1991). “Gendered Reason: Sex Metaphor and Conceptions of Reason.” Hypatia 6(2): 77–103. Rooney, Phyllis (2011). “The Marginalization of Feminist Epistemology and What That Reveals about Epistemology ‘Proper’,” in Heidi E. Grasswick (ed.), Feminist Epistemology and Philosophy of Science. Springer, Dordrecht, 3–24. Rothgerber, Hank (2013). “Real Men Don’t Eat (Vegetable) Quiche: Masculinity and the Justification of Meat Consumption.” Psychology of Men & Masculinity 14(4): 363–75. Ruby, Matthew B. and Steven J. Heine (2011). “Meat, Morals, and Masculinity.” Appetite 56(2): 447–50. Russell, Camisha (2019). “On Black Women, ‘In Defense of Transracialism,’ and Imperial Harm.” Hypatia 34(2): 176–94. Salamon, Gayle (2009). “Justification and Queer Method, or Leaving Philosophy.” Hypatia 24(1): 225–30. Schouten, Gina (2016). “Philosophy in Schools: Can Early Exposure Help Solve Philosophy’s Gender Problem?” Hypatia 31(2): 275–92. Schwitzgebel, Eric and Carolyn Dicey Jennings (2017). “Women in Philosophy: Quantitative Analyses of Specialization, Prevalence, Visibility, and Generational Change.” Public Affairs Quarterly 31(2): 83–106. Snowdon, Paul and Anil Gomes (2019). “Peter Frederick Strawson,” in Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy, https://plato.stanford.edu/archives/ spr2019/entries/strawson. Superson, Anita (2011). “Strategies for Making Feminist Philosophy Mainstream Philosophy.” Hypatia 26(2): 410–18. Thompson, Morgan (2017). “Explanations of the Gender Gap in Philosophy.” Philosophy Compass 12(3): 1–12.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/03/21, SPi
250 Emmalon Davis Thompson, Morgan, Toni Adleberg, Sam Sims, and Eddy Nahmias (2016). “Why Do Women Leave Philosophy? Surveying Students at the Introductory Level.” Philosopher’s Imprint 16(6): 1–36. Tremain, Shelley (2018). “Philosophy and the Apparatus of Disability,” in Adam Cureton and David Wasserman (eds), Oxford Handbook of Philosophy and Disability. Oxford: Oxford University Press, 82–99. Valian, Virginia (1998). Why So Slow? The Advancement of Women. Cambridge, MA: MIT Press. Walker, Margaret (2005). “Diotima’s Ghost: The Uncertain Place of Feminist Philosophy in Professional Philosophy.” Hypatia 20(3): 153–64. Waters, Anne (2003). “Structural Disadvantage and a Place at the Table: Creating a Space for Indigenous Philosophers to Be More Proactively Involved in Decision Making Forums Affecting the Emergence and Impact of Indigenous Philosophers of the Americas.” APA Newsletter on Native American and Indigenous Philosophy 3(1): 2–5. Weinberg, Jonathan, Shaun Nichols, and Stephen Stich (2001). “Normativity and Epistemic Intuitions.” Philosophical Topics 29(1 and 2): 429–60. Whyte, Kyle (2017). “Indigenous Research and Professional Philosophy in the U.S,” in Meena Krishnamurthy (ed.), Featured Philosopher Blog (February 3, 2017), https:// politicalphilosopher.net/2017/02/03/featured-philosopher-kyle-whyte. Wilson, Yolonda Y. (2017). “How Might We Address the Factors that Contribute to the Scarcity of Philosophers Who Are Women and/or of Color?” Hypatia 32(4): 853–61. Yancy, George (ed.) (1998). African-American Philosophers: 17 Conversations. New York: Routledge. Yancy, George (2008). “Introduction: Situated Black Women’s Voices in/on the Profession of Philosophy.” Hypatia 23(2): 155–9. Yancy, George (ed.) (2012). Reframing the Practice of Philosophy: Bodies of Color, Bodies of Knowledge. Albany, NY: SUNY Press.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
PART 6
EPIST E MOLO GY A ND F E M I NI ST PE R SPE C TI V E S
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 19/03/21, SPi
11
Rape Culture and Epistemology Bianca Crewe and Jonathan Jenkins Ichikawa*
1. Introduction This chapter concerns a question more pressing than we wish it were: how should institutions and individuals react to sexual assault allegations that haven’t been established in legal settings? For example, if student A accuses student B of sexual assault, but doesn’t make a report to law enforcement, what formal university responses might be appropriate? (Should B be barred from campus parties? Or expelled? What kinds of accommodations to A are appropriate?) In general, the question of what to do in a given circumstance has a lot to do with epistemology. The right action depends on the reasons one has available, and the availability of reasons is an epistemic matter. On one widespread approach, it’s a matter of knowledge.1 If you know that a student is a rapist, you are in a very different practical situation than you are if you don’t. One tempting idea about our central question is that institutions and individuals should be deferential to law enforcement, effectively outsourcing the epistemology to the criminal justice system. For example, this line of thought has it that a university should take care to be neutral as to the truth of any allegation, * Authors contributed equally: names are listed alphabetically. We are grateful to Nate Bemis, Nathan Cockram, Annaleigh Curtis, Jeremy Dawson, Logan Fletcher, Jennifer Freyd, Sandy Goldberg, Jade Hadley, Cassie Herbert, Marc Hewitt, Carrie Jenkins, Glynnis Kirchmeier, Jennifer Lackey, Lauren Leydon-Hardy, Clayton Littlejohn, Lucia Lorenzi, Graham Moore, Carla Nappi, Phyllis Pearson, Kathryn Pogin, Alan Richardson, Kyle da Silva, Joe Slater, Chris Stephens, Rodrigo Valencia, Jordan Wadden, Nancy Wu, and Michel-Antoine Xhignesse, for discussion of drafts of this chapter and for inspiration and discussion related to this material. We’re especially grateful to Alex Guererro, who provided helpful and detailed comments on an early full draft. Thanks also to Cam Gilbert, Kristin Conrad, and Mira Kuroyedov for help proof-reading. Work for this project was funded in part by a SSHRC Insight Grant on rape culture and epistemology. 1 Central views in this tradition include Hawthorne and Stanley (2008); Fantl and McGrath (2009); and Weatherson (2012). Ichikawa (2017, 139–42) defends the centrality of knowledge in action in terms of reasons. We work within the knowledge framework in this chapter, although much of what we want to say is translatable into competing frameworks relating between epistemology and action. For example, if one thinks that all of one’s justified beliefs are among one’s practical reasons, a central question will be whether, in the kinds of cases we will go on to discuss, one justifiably believes that a sexual assault occurred. The considerations regarding rape culture, skepticism, and status quo biases we will go on to discuss seem equally applicable in that framework; the considerations regarding con textualism might or might not be similarly applicable. (Cohen (1999) and Wedgwood (2008) defend the context-sensitivity of justification ascriptions.) So, while we are working within a particular, admittedly controversial, framework, we hope that the interest of our project will extend to theorists in competing traditions. Bianca Crewe and Jonathan Jenkins Ichikawa, Rape Culture and Epistemology In: Applied Epistemology. Edited by: Jennifer Lackey, Oxford University Press (2021). © Bianca Crewe and Jonathan Jenkins Ichikawa. DOI: 10.1093/oso/9780198833659.003.0011
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
254 Bianca Crewe and Jonathan Jenkins Ichikawa unless and until the law enforcement process delivers its verdict. If a student is convicted of sexual assault in a fair criminal trial, a university should treat it as settled that the student is guilty of sexual assault and react accordingly. But unless and until that happens, the university should do nothing.2 We think such temptation to deference is a powerful social meme, although it does not tend to be endorsed explicitly by university officials. One simple reason for this is that in many jurisdictions, the law directly prohibits this kind of deference. Since 1980 in the United States, for example, Title IX has required universities to have sexual harassment and assault policies beyond simple reference to law enforce ment, on the grounds that sexual violence is a kind of gender discrimination.3 In Canada, Ontario and British Columbia instituted mandatory sexual assault policies for universities in 2017. When students report sexual assaults by university members to their universities, the allegations are typically investigated by a university process separate from law enforcement. If university investigatory committees conclude that the allegations are true, they may take action on this basis—firing, expelling, or otherwise discip lining the person found to have committed assault, for example. One common complaint about this kind of action is that it is insufficiently deferential; it amounts to “taking the law into their own hands,” or, in cases in which legal pro ceedings are initiated but no conviction is reached, “putting their judgment over the law.” We think the temptation to deference is influential, even among univer sity administrators; certainly, it is influential among the wider population. (You can find expressions of this deferential attitude near the top of every list of com ments on news articles about university sexual assault policies.) We will argue that this sort of deference is mistaken, for at least two kinds of reasons. One is that, as anti-rape activists have often emphasized, law enforcement procedures themselves are deeply flawed when it comes to their practice of dealing with sexual assault. Another has to do with the connections between knowledge and action, and between knowledge and knowledge ascriptions. We also think that these disparate considerations are more closely interconnected than it may at first appear. We begin in section 2 with a discussion of some of the challenges to adequate law enforcement responses to sexual assault, including the operation of sometimes transparent social and political factors that constitute rape culture. In section 3, we begin to foreground more explicit epistemological questions, focusing on the interaction of rape culture with the epistemology of testimony. Sections 4–5 take up the relationship between epistemology, practical reasoning, and action, 2 There are stronger and weaker forms of deference one might adopt. One might defer to law enforcement only in those cases in which the matter is being adjudicated legally; but proponents of the deferential stance often hold that one should suspend judgment even if the question has never been legally adjudicated. 3 For some legal discussion, see Curtis (2017) and Johnson (2017).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Rape Culture and Epistemology 255 explaining how skeptical pressure motivates a conservatism that benefits the sta tus quo. We introduce, in section 6, contextualism about knowledge ascriptions as a solution to some of the challenges that have emerged and offer a further criti cism of many deferential stances; in section 7 we show how many of the social and political challenges discussed early in the chapter recur with respect to metasemantic considerations.
2. Rape Culture and Law Enforcement Epistemic deference to law enforcement makes sense only to the degree to which law enforcement’s epistemic procedures are reliable.4 So, the idea that institutions and individuals should so defer rests on particular assumptions about the contexts in which sexual assault and assault allegations occur, and about the formal and informal channels of justice in place to deal with them. Articulating these assumptions and clarifying certain relevant features of the social and political context will allow us to show how they are bound up with harmful epistemic idealizations. When institutions or individuals defer to the investigative and procedural standards of police and the law, this bespeaks a high degree of confidence in these standards and procedures. We think such confidence betrays an ignorance of the social and political context in which sexual assault occurs, in which victims testify about their experience (formally and informally), and in which individuals, institutions, the police, and the law respond to said testimony. For example, cultural attitudes surrounding sexual assault, women’s credibility,5 and patriarchal and misogynist presumptions of access and entitlement to women’s bodies are significant determinants in how juries rule in sexual assault cases. These issues also have a lot to do with whether or not sexual assault is ever disclosed or reported.6 4 There are also important non-epistemic grounds against deferring to law enforcement. For example, one might have political motivations for questioning the legitimacy of police and legal systems as authorities and instruments of justice. One could, for example, explain politically motivated non-deference in terms of both the lack of reliability exhibited in the epistemic procedures of law enforcement and as the historical and contemporary violence enacted by the state in the protection of certain interests. Though we lack the space to explore these reasons in depth, we see them as part of a continuum with the epistemic reasons we examine throughout. 5 Women are not the only demographic that experiences sexualized violence, but the cultural atti tudes that produce such things involve an ideology that is deeply gendered. Here, we will speak of “women” not only as the demographic category that experiences the highest rates of sexual assault, but also as an ideological category with built-in patriarchal baggage that renders it particularly susceptible to some of the political and epistemic problems discussed here. In cases where non-women experi ence and disclose their experiences of sexual assault, the widespread association of their experiences with typically feminine ones (and the even scarcer hermeneutical resources available to them) consti tutes part of the difficult social situation that they face, both in terms of disclosing and being believed. 6 Anderson (2003); Du Mont et al. (2003); George et al. (1992); Koss (1985); McAuslan (1998).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
256 Bianca Crewe and Jonathan Jenkins Ichikawa These social and political factors are sometimes theorized as elements of rape culture—a cultural environment where sexual assault and sexualized violence is a normative or expected type of interaction.7 This account demands that analysis of sexualized violence (and dominant understandings thereof) occur with an eye to the cultural and political factors that entrench and reproduce it, including spurious beliefs about race, gender, and sexuality. Rape culture is therefore visible not only in the statistics of reported sexual assault, but also in the cultural dis course that produces these statistics and other undisclosed experiences, and it conditions responses to allegations of sexual assault.8 Importantly, patriarchal and misogynist values work in tandem with many other oppressive hierarchical systems in producing the social conditions structuring rape culture. Colonialism, racism, homophobia, ableism, and classism produce populations which are often especially vulnerable to sexualized violence. For example, racist stereotypes and the structural oppression faced by women of color work in tandem to render this demographic at high risk for sexual assault, and likely to experience difficulties in reporting their experiences.9 These vulnerabil ities are compounded when, for example, the women in question are refugees, asylum seekers, or immigrants,10 if they do not have stable access to housing,11 or if they have disabilities.12 One reason we are less optimistic about the ability of the law to deal with infractions stemming from patriarchy and misogyny has to do with the political and ideological affiliations of legal systems themselves. We think deferential attitudes mistakenly assume that no such affiliations exist. This vision rests on an idealized (and often unarticulated) account of the relationship between law and the cultural context in which it is embedded—one which many critical legal theorists have sought to bring to light. They argue that, often, liberal democratic institutions not only permit, but also conceal and reify the subordination of women under the guise of formal neutrality. In Towards a Feminist Theory of the State, Catharine MacKinnon suggests that inasmuch as the state and its insti tutions are constituted in a cultural context, they will tend to reflect that culture. This means that, for example, a patriarchal culture is apt to yield and mutually reinforce a patriarchal state. According to MacKinnon, “women are oppressed socially, prior to law, without express state acts, often in intimate contexts,”13 and therefore states and institutions that are committed to negative freedom and judicial neutrality—and that see themselves as socially transcendent rather than embedded—will not be able to address injustices that arise from this configuration. 7 Brownmiller (1975); Hampton (1999). 8 Buchwald et al. (2005); Burt (1980). 9 Foley et al. (1995); Donovan and Williams (2002); George and Martínez (2016). 10 Hynes and Lopes (2000); Ward and Vann (2002); Wenzel et al. (2004). 11 Kushel et al. (2003); Tyler et al. (2010). 12 Khalifeh et al. (2015); Basile et al. (2015). 13 Mackinnon (1989, 165).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Rape Culture and Epistemology 257 Attempts are increasingly being made to update laws and procedures s urrounding sexual assault—the criminalization of sexual assault within marriage, for example,14 or the slow phasing out of the requirement of corroboration of sexual assault by witnesses.15 Although these developments reflect important cultural milestones, and doubtless improve the lives of many, there is still ample evidence16 and theoretical work on the ways in which the law and rape culture continue to cooperate. This situation arises from, and is compounded by, abstrac tions in law and in our conceptual analysis. Epistemologists, critical legal theorists, and other scholars are increasingly attentive to how power, ideology, and oppressive cultural conditions affect theor izing, and to how the idealizations they employ sometimes abstract away from highly relevant contextual factors. Charles Mills’s work is a central case in point. In his work on ignorance and idealization in philosophical theorizing, Mills emphasizes the gap between oppressive social conditions and theoretical idealiza tions, writing that “abstracting away from realities crucial to our comprehension of the actual workings of injustice in human interactions and social institutions”17 provides little practical directives for the achievement of social justice, and, moreover, often obscures the existence of the very conditions that exacerbate injustices—such as, for example, sexism and racism. Ultimately, Mills argues that this precludes achievement of the ideal-as-idealized model, leading to situations in which theoretical idealizations serve to bolster the social and political status quo, and are employed, at least in part, for this reason. MacKinnon’s work is focused on a similar gap: namely, the one assumed between the private and public sphere, and, more generally, between what is deemed to be amenable to politicization and legal treatment and what is not. MacKinnon and Mills focus on political philosophy, but their arguments apply equally to theorizing in epistemology, as we will show.
3. Knowledge and Testimony Law is not the only domain implicated in this kind of deferral of epistemic responsibility: folk-conceptions (and double-standards) regarding the relationship between knowledge and testimony are also significant. Our epistemic reactions to testimony are heavily influenced by many cultural and political factors. Often, we take testimony as, for all relevant purposes, conclusive for its contents. If you tell your friends where you went for dinner last night, they will probably take that as 14 In Canada, in 1983 with the update of Bill C-127, and in 1993 in all 50 of the United States. 15 In the 1980s and 1990s, in North America. See Tang (1998). 16 For instance, in statistics on untested evidence kits and underreporting in the United States—see Ritter (2011) and Rennison (2002). 17 Mills (2005, 170).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
258 Bianca Crewe and Jonathan Jenkins Ichikawa settling the question, even if there is something riding on it—if it would settle a bet, for instance. It is, of course, possible to raise skeptical challenges to such cases: you might be lying. In cases in which we are wholly reliant on someone’s testimony, the possibility that they are not telling the truth does constitute a skeptical scenario— it’s a way that things could be that your evidence doesn’t, in some strict sense, rule out.18 But we do not always treat the possibility of skeptical scenarios as undermining knowledge. In at least a wide variety of cases, our orientation toward testimony is a non-skeptical one: when someone tells us that P, we end up justified in believing that P. And ordinarily, if the speaker knows that P and tells us that P, we end up ourselves knowing that P. That testimonial knowledge is common is both common sense and philosophical orthodoxy.19 But we do not always take this kind of stance. Sometimes we interrogate testi mony; sometimes we explain it away or demand proof. As much recent literature in both social psychology20 and epistemic injustice21 has shown, social and cul tural factors influence the degree to which testimony is considered credible. The identity of both speaker and hearer, as well as the content being conveyed, influ ence the reception of testimony at the individual and broader cultural level. Testimony that conveys a personal experience of sexual assault—especially in cases in which no one other than those involved in the interaction have first-hand knowledge of what occurred—is often met with skepticism. In other words, a dif ferent epistemic standard is employed in these cases that is not present in the reception of other kinds of testimony. Furthermore, the determination of whether or not the testimony of the individual alleging assault counts as knowledge is frequently cast in terms of measuring the reliability of two people against each other, without regard for important contextual features. This is a mistake, because some of those contextual features might incline members of the community, juries, and the police to assume that one individual is much more reliable than the other, for reasons having to do with stereotypes and structural power. This is visible in what, in highly publicized controversies regarding sexual assault, is submitted as evidence that might decrease the reliability of the individual who is testifying that they were assaulted: their past sexual behavior, their choice of clothing, their friends, their hobbies, etc. Leigh Gilmore addresses this in her discussion of Anita Hill’s testimony in the Clarence Thomas confirmation hearings:
18 Lewis (1996, 553) lays out this way of thinking of skeptical scenarios. 19 See, e.g. Lackey (2008) or Adler (2015) for the philosophical canon. 20 Frohmann (1991); Mack (1993); Mulder et al. (1996); George et al. (2016); Snyder et al. (1977). 21 Alcoff (1999); Jones (2002); Fricker (2007).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Rape Culture and Epistemology 259 Common notions like “Nobody knows what really happened” and “It’s a case of he said/she said” deflect inquiry into what we might come to know and even what we already do know. Moreover, they participate in narrowing attention to the immediate time frame of crisis or scandal, restricting the frame temporally, and ensuring that no adequate context for understanding will emerge. For these reasons, Toni Morrison’s comment on the hearing provides pivotal guidance: “For insight into the complicated and complicating events that the confirmation of Clarence Thomas became, one needs perspective, not attitudes; contexts, not anecdotes; analyses, not postures. For any kind of lasting illumination, the focus must be on the history routinely ignored or played down or unknown.” (Gilmore 2017, 57 citing Morrison 1992, x)
When individuals testify that they were sexually assaulted, their testimony is often considered to be less reliable than most other testimony. It is often suggested that they are either lying—out of animus, or to seek attention, or because the ‘survivor’ identity is socially valuable—or deluded. Psychological discourse surrounding the harms of sexual assault often unwittingly contribute to rendering those who testify about their experience unreliable in the eyes of their audience. For example, in her book on the politics of sexualized violence, Alison Healicon notes that therapeutic articulations of the category of “rape victim” inhibit the effective communication of the experience, as individuals belonging to this category are understood to have a damaged grasp on reality.22 A “credibility conundrum” ensues:23 if, according to the available hermeneutics, the trauma of sexual assault damages a potential testifier specifically in terms of their capacities qua testifiers, then it will be impossible for their testimony to convey knowledge. These assumptions and idealizations constitute a part of rape culture and are invoked strategically to further entrench that culture. However, they do not necessarily arise out of the ill will of particular individuals or communities. Following Miranda Fricker’s (2007) Foucauldian account of power in Epistemic Injustice, we take it that the lack of available hermeneutical resources for the communication and uptake of sexual assault narratives is a collective and structural problem, and is not necessarily withheld by particular individuals. Part of our aim in this chapter is to render visible the biases and automatic socio-political alignments embedded in fundamental epistemological concepts and assumptions. In sum, rape culture has acute epistemic effects. It conditions the reception of testimony that alleges the occurrence of a sexual assault and restricts the her meneutical resources that marginalized individuals can draw on to communicate their experience. The results of this are visible in the difference between idealized notions of how culpability, evidence, and knowledge are determined in cases 22 Healicon (2016).
23 Jordan (2004).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
260 Bianca Crewe and Jonathan Jenkins Ichikawa where sexual assault is alleged and what actually tends to occur. Many social scientists have studied what happens when individuals disclose that they have been sexually assaulted, in a variety of informal or institutional settings. By and large, those who seek to communicate their experience of sexual assault tend to suffer serious social and emotional costs for it.24 It is also important to note that the belief that there are hermeneutical and legal resources available to address such injustices are held much more readily by privileged groups and individuals: in a 1992 study on the disclosure of sexual assault, Gail Elizabeth Wyatt found white women far more likely to disclose their assaults than black women, who “do not anticipate that they will be protected by traditional authorities and institutions.”25 Furthermore, because the reasons that some individuals or communities might be adverse to reporting to legal or penal authorities are often not broadly under stood in dominant culture, failure to report is often construed as indicative of the weakness of a case and the perceived non-credibility of the person in question, rather than institutionalized racism, classism, sexism, ableism, etc. The assumption that the testimony of people who claim to have been sexually assaulted is unreliable is poorly supported by the available evidence, as in fact false allegations of sexual assault are exceedingly rare.26 Of course, it is possible to find examples of false assault allegations. However, as in the more general examples discussed above, this does not undermine the general reliability of such reports. Indeed, it is essential to the epistemology of testimony generally that although sometimes testimony is misleading, it typically issues into know ledge. Reliability needn’t be perfect reliability. But, as we have seen, the fact that sexual assault narratives are often quite trustworthy does not mean that they are, in fact, trusted.
4. Knowledge and Action As alluded to in section 1, there are important connections between epistemology and practical decision-making. A prerequisite to knowing what to do is knowing how things are. One way to encode that relationship is via a straightforward relationship between knowledge and reasons for action, according to which one’s reasons constitute all and only that which one knows.27 On this framework, to know something is to have it conclusively settled—or at least, conclusively enough
24 Davis et al. (1991); Golding et al. (1989); Ullman (2000); Ahrens (2006); Dworkin et al. (2016); Wyatt (1992). 25 Wyatt (1992: 86). 26 Lisak et al. (2010). 27 The philosophical literature on reasons is vast and complicated by diverse terminological choices. Ichikawa (2017, 140–1) articulates a version of the approach we are using here, and situates it into broader discussions of ‘internal’ and ‘external’ reasons, ‘possessed’ and ‘unpossessed’ reasons, and ‘reasons’ and rationalizing action.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Rape Culture and Epistemology 261 for one’s practical purposes. If you know that P, you don’t have to worry about whether not-P. Conversely, if uncertainty about P renders the possibility of not-P a serious practical consideration, you do not know that P. The knowledge–reasons principle explains why skepticism and inaction are so closely connected. It explains, for example, why someone concerned about the possibility that an accused rapist is innocent will face pressure to deny knowledge of his guilt. “I know that he is guilty, but we shouldn’t punish him because of the chance that he is innocent” is a contradiction. The knowledge–reasons principle says that if something is known, it is available as a reason. If something is not known, then it implies that it’s not a reason.28 This points to a normative complication of our discussion above. We have argued that unwarranted doubt, a manifestation of a pervasive rape culture, tends to attach to testimony about sexual assault; this often leads to inaction in a way that we consider an important mistake. But note that if one has doubt—even unwarranted patriarchal doubt—one may eo ipso fail to know. And this, given the knowledge–reasons principle, implies that one lacks reason to act. Jennifer Nagel’s work on “epistemic anxiety” offers an example of how these ideas might interact. Nagel characterizes epistemic anxiety as a kind of feeling of uncertainty; when one is epistemically anxious, one doesn’t consider the question to be settled. It is possible to have epistemic anxiety even when one’s evidence is conclusive; having conclusive evidence and feeling as if one has conclusive evidence are different states. Epistemic anxiety, Nagel says, is inconsistent with knowledge because it is inconsistent with (outright) belief, which is itself necessary for knowledge.29 Epistemic anxiety is not automatically rational. If one is epistemically anxious, one may be accurately reflecting the weakness in one’s epistemic position, but one might be experiencing unwarranted doubt. Suppose someone doesn’t know that a sexual assault has occurred because of epistemic anxiety. The knowledge–reasons principle says the fact of the sexual assault is not among their reasons; it’s easy to argue they don’t have sufficient reason to act. Does this vindicate a deferential response? There are two strategies available to undercut this suggestion. One is to argue that epistemic rationality affirmatively requires belief in the cases in question. One ought not be epistemically anxious, because being so is epistemically prohibited
28 We recognize that this is a controversial commitment (though, as fn. 27 suggests, identifying just which views conflict with what is not straightforward, due to diverse terminological choices). The broader point is uncontroversial: there is a substantial connection between epistemology and appro priate action. The knowledge–reasons principle is one way of articulating it. As we said at the start, we expect that our central points are translatable into other frameworks. 29 See, e.g. Nagel (2010). Note that the ‘belief ’ condition in the classical ‘justified true belief ’ theory of knowledge stood for a kind of full commitment, similar to certainty or being sure. Gettier (1963). See Ichikawa (2017, 224–6) for discussion.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
262 Bianca Crewe and Jonathan Jenkins Ichikawa under the circumstances.30 On such a picture, unwarranted doubt cannot render someone’s inaction epistemically faultless. However, many epistemologists are skeptical of “positive” epistemic norms, and therefore do not take failure to believe something that one could have known to be an epistemic failure.31 Theorists of this stripe hold that epistemology is only in the business of identifying beliefs it is epistemically permissible to hold, but that there are no genuine positive epistemic norms. In our view, this tradition gives an unwarranted privileged place to skepticism. It holds that skepticism is always epistemically permissible. But even theorists who limit epistemic normativity to prohibitions of belief typically allow that other sources of normativity—moral and pragmatic norms, for example—can interact with epistemic norms to generate obligations to believe.32 They insist that skepticism is epistemically permissible but are open to the idea that positive moral or pragmatic norms interact with negative epistemic ones to require belief. Whether or not one accepts positive epistemic norms, then, there is room to say that in some cases of epistemic anxiety regarding sexual assault testimony, one ought not have that anxiety, and one ought to believe the allegations. Skepticism is a mistake of some kind. As we have seen, skepticism often works in tandem with oppressive social conditions to discredit the knowledge claims of marginalized individuals and groups. Our own sympathies are in favor of positive epistemic norms; we think that in these cases, failure to believe is as much an epistemic error as is believing something that isn’t warranted. Undue epistemic anxiety is therefore on our view an epistemic error, in addition to being (in this case) a moral error. But whether or not one agrees with us about positive epistemic norms, one can agree that one should know.
5. Pragmatic Encroachment and Epistemic Idealizations So far, we have pointed to some reasons for concern about the uptake of sexual assault allegations—their reception is processed in an oppressive political and cultural context, where motivated doubt is a significant possibility. One might accept this much, however, and deny that such doubt is mistaken. Maybe the natural tendency towards epistemic anxiety, in the face of sexual assault
30 See, e.g. Ichikawa and Jarvis (forthcoming) for an application of this line of thought in other kinds of cases. 31 See, e.g. Nelson (2010) and Littlejohn (2012, 47). 32 For instance, Nelson (2010, 92–3) and Littlejohn (2012, 48) both allow that non-epistemic goals interact with epistemic restrictions on permissible belief to allow for hybrid obligations to believe.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Rape Culture and Epistemology 263 allegations, is appropriate and justified, and deference to the legal system is the most appropriate cautious response. There is a tendency towards skepticism when the perceived importance of a question is high. One way to motivate the kind of skepticism and deference we are discussing, then, is to emphasize the high practical stakes involved in responding to sexual assault allegations. For example, a person might in good faith believe that the seriousness of sexual assault allegations indicate that they are better handled in a legal setting. We have argued that this faith is often misplaced: harmful idealizations obscure the relationship between law and culture, and these idealizations are apt to reproduce oppressive conditions under the guise of formal neutrality. Skeptical or deferential stances towards sexual assault allegations might also be motivated by the assumption that sexual assault is a “private” matter. However, the distinction between “public” and “private” is itself often mobilized in the service of maintaining an oppressive status quo. For example, in The Sexual Contract, Carole Pateman argues that the traditional liberal distinction between the public and private sphere often situated women in the latter. Furthermore, this tradition assumes the private sphere to be outside the scope of legal or political recourse, with the effect that violence or unequal power relations in this sphere are naturalized and obscured.33 In discussions of the high stakes surrounding sexual assault allegations, it is also common to emphasize the seriousness of the consequences for those who have been accused. This in effect emphasizes the risked harm of a false finding of responsibility. Perhaps, in the face of risks such as these, it takes a stronger evidential situation to count as knowing. This, broadly speaking, is the thought behind so-called ‘pragmatic encroachment’ in epistemology.34 Pragmatic encroachment theorists hold that knowledge doesn’t just depend on “truth-relevant” features like one’s evidential situation, but also on practical considerations like how high the stakes are. In particular, the higher the stakes, the more it takes to know.35 Applied to the present discussion, the pragmatic encroachment strategy might suggest that the gravity of the accusation, and its social and practical costs, are enough to raise the bar for knowledge. It’s harder to know about sexual assault via testimony than it is to know other things, because the stakes are higher for beliefs about sexual assault. This strategy could explain institutional reluctance to take action against individuals accused of sexual assault: the costs of being wrong are just so high.
33 Pateman (1988). 34 The term was introduced by Jon Kvanvig in a “Certain Doubts” blog post in 2004. 35 Central defenders of pragmatic encroachment include Stanley (2005); Fantl and McGrath (2009); and Weatherson (2012). See Weatherson (2017) for a sympathetic contemporary overview of the position.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
264 Bianca Crewe and Jonathan Jenkins Ichikawa We are pessimistic about justifying this kind of reticence to believe via ragmatic encroachment. The explanation sketched assumes certain perspectives p that should not be taken for granted. In particular, it emphasizes the risks and costs to the person accused of sexual assault and ignores those for the person who testifies that they were assaulted. In other words, it is important to note that what one considers to be a high-stakes situation (and the perspective from which stakes are analyzed) is itself conditioned by ethical and epistemic factors flowing from the political and ideological context. It is, for example, revealing of our cultural and political commitments that the framing of sexual assault controversies and trials often highlight the harms men and boys will experience as a result of being accused or found guilty of sexually assaulting someone. Kate Manne has labelled this phenomenon “himpathy.”36 Likewise, the preoccupation with the potential for false accusations (especially given their infrequent occurrence) demonstrates that the one group’s potential harms far outweigh the potential harms to another group in the cultural consciousness. Such considerations (and rape culture more generally) must be brought to bear when theorizing pragmatic encroachment in the context of responding to sexual assault allegations. Pragmatic encroachment is motivated by a kind of risk aversion. The implicit model is one that treats suspension of judgment as the default—to make a judgment is to undergo the risk of being wrong. Even if the pragmatic encroachment model explicitly recommends further investigation rather than the suspension of judgment, we take it that attention to the non-ideal and socially embedded con texts of “high-stakes” epistemic situations might in fact reveal that skepticism and the felt need for further investigation are very closely connected. Indeed, their effects are identical, especially in cases in which one’s own investigatory capacities are limited, and the very meaning of “investigation” is steeped in oppressive assumptions. In addition to motivating skeptical or deferential positions, the public/private divide also affects how harms and stakes are characterized. On the pragmatic encroachment mode, these characterizations are often dependent on longstanding cultural beliefs about sexual assault as a “private” matter implicating only the accuser and the accused. This stance limits the extent to which rape culture can be articulated as a harm at a collective or societal level, or indeed whether sexual assault can be understood in terms of pernicious cultural discourse, rather than the mere ill will or criminality of individual perpetrators. This stance also limits the extent to which communities and institutions can be held accountable for preventing sexual assault: doing nothing, whether under the guise of searching
36 Manne (2018, 196–201).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Rape Culture and Epistemology 265 for further evidence or deferring to law enforcement, often implicitly condones the behavior of those who perpetrate sexualized violence. The problems with the pragmatic encroachment model are also related to our discussion of negative and positive epistemic norms in section 4. Ultimately, the idea that suspending judgment is the way to “play it safe” ignores the potential risk of failing to acquire available knowledge; moreover, following the injunction to “seek out further evidence” can, via selective investigation, provide further opportunities for one’s inquiry to tilt in favor of confirming certain (patriarchal, etc.) starting points.37 It assumes, in effect, that the status quo—with the subject declining to form a belief on the matter—is normatively acceptable. But failure to form available knowledge can be costly and harmful. This is true in general, but especially obvious in the case of testimony of sexual assault. Moreover, without an adequate understanding of how knowledge attributions are affected by rape culture, the resultant harms will be made invisible and further entrenched. This is because epistemic idealizations of the kind that Mills discusses work in tandem with a strong bias towards the status quo, both in epistemology and in broader society. If pragmatic encroachment is understood in terms of some putatively “objective” standard of what counts as a high-stakes situation, it is likely that the perspective of socially dominant groups and individuals will be the yardsticks we employ, precisely because their perspectives are taken to be neutral. MacKinnon’s work on the character of law in a patriarchal society and Mills’s account of ideologically motivated idealizations are both intimately tied to this point. In effect, they caution that we must be wary of taking as neutral what is in fact ideologically loaded. Some of the intuitions that are cited to motivate pragmatic encroachment in cases like these are susceptible of alternate explanations along these lines. A general tendency to favor the status quo—reinforced by the cultural elements it favors—is part of the story.38 A self-serving bias is surely another. It is disruptive to believe a sexual assault allegation—especially if it is against someone with whom one has any kind of significant relationship. A friend of one of the authors has recounted a story of discussing her sexual assault with the woman married to her assailant. In declining to listen, the wife explicitly cited the cost that believing would bring. “If it’s as bad as you say,” she told her, “believing it would destroy my marriage.” Here, of course, she is citing the high cost of believing even if it is true, so the pragmatic encroachment mold doesn’t fit in the usual way. But given the high cost of belief in general in cases like these, it’s not difficult to imagine a kind
37 Compare the parallel point about scientific evidence in Levy (2019). 38 Hundleby (2016) suggests that we teach a “status quo bias” alongside the more canonical formal and informal fallacies.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
266 Bianca Crewe and Jonathan Jenkins Ichikawa of motivated invocation of the high cost of being wrong. It is easier to do nothing (viz, suspend judgment) and rationalize it via pragmatic encroachment.39 Transparency regarding how cultural and ideological commitments structure our inquiry is the first step to a version of philosophy that addresses oppression. Explaining the phenomena that we have been investigating in terms of pragmatic encroachment must address the value- laden assumptions structuring that explanation. A commitment to a particular perspective (one that is automatically privileged in a rape culture) goes into the appeal to pragmatic encroachment to explain why knowledge cannot be conveyed in the cases we’re considering here, and to fail to render this explicit as a commitment stemming from the social con text of knowledge is problematic for ethical and epistemic reasons. The project of rendering political commitments and the ideological context of knowledge explicit can also be seen as arising within feminist standpoint theory, especially with regard to the claim that the epistemic privilege of marginalized communities arises from their understanding of this context. For example, stand point theory takes the occupation of a marginalized social location to enhance understandings of the workings of power, as marginalized people must negotiate both their own environments, where power and oppression are necessarily confronted, and the environments of the socially dominant, who have the privilege of remaining ignorant about the many of the features of their social world.40 It is the juxtaposition of these environments that creates, for marginalized knowers, a sense that dominant accounts of the world are decidedly non- neutral and value-laden. When considering pragmatic encroachment, the workings of the legal system and testimony as applied to sexual assault, scholarship on epistemic privilege, motivated ignorance, and idealization all help to draw out why the foundations from which we reason often have invisible political and social implications. As Adrienne Rich notes, in a context that privileges and universalizes the experience of dominant groups (in this case, men), “objectivity has been little more than male subjectivity.”41 Furthermore, values and political commitments are often only visible as such when they are different from the norm. We will expand upon this status quo bias, and related ideologically motivated skepticism, in what follows. 39 Charles Mills develops a similar idea in his work on ignorance. In The Racial Contract (1997), he theorizes that the psychic and moral conditions that made possible slavery, colonization, and the con tinued structural privileging of whiteness involve the conceptual erasure of the personhood, property, and communities of racialized peoples. In the cases Mills considers in The Racial Contract, as in the example of believing disruptive testimony regarding sexual assault, self-transparency and a genuine commitment to understanding social realities would interfere with the positive self-conception of individuals and communities in privileged epistemic and social positions. As Mills puts it, these epis temologies of ignorance operate as a “pattern of localized and global cognitive dysfunctions,” though they are “psychologically and socially functional” (Mills 1997, 18). 40 Patricia Hill Collins theorizes this epistemic position in terms of the “insider-outsider” (1990, 1991); this line of thought also appears in Narayan (1988). 41 Rich (1994, 3).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Rape Culture and Epistemology 267
6. Contextualism and Knowledge without Courtroom Knowledge We turn now to another aspect of the problem with excessive deference to law enforcement, declining to act on the grounds that guilt hasn’t been proven in a legal setting. In a just and equitable criminal system, defendants are considered innocent in light of the law unless and until they are proven guilty. An inevitable consequence of any such system is that some genuinely guilty offenders will be acquitted, because, although they are guilty, they haven’t been proven to be so. This is widely thought justified on the grounds that it is better for guilty people to walk free than it is for innocent people to be wrongly incarcerated. We accept these tenets of liberal democracy. Because of prosecutors’ burden of proof, it obviously does not follow from the fact that someone has not been convicted of some criminal offense that they did not commit that crime. The phrase “innocent until proven guilty” doesn’t mean that anyone literally is innocent.42 So much, we take it, is pretty obvious. After all, some people are guilty of crimes that no one ever found out about, and so were never prosecuted. But this is even true of people who have been acquitted, perhaps because the evidence against them was insufficient. Courts sometimes use the misnomer of “finding the defendant not guilty”—they might say more precisely that they “did not find the defendant guilty.” Some defendants are guilty of the crimes of which they were acquitted. They are “innocent in the eyes of the law,” but they are guilty. According to the thought behind the deferential stance we critique, if there wasn’t evidence sufficient for a criminal conviction, this typically demonstrates that we can never know whether someone is guilty of an offense, and so cannot responsibly take any action in response. This is the thought behind the suggestion that, for example, if a sexual assault finding hasn’t been established in court, it would be inappropriate to act on that accusation, because this would be putting our own judgment over that of the law. For the reasons outlined in section 2, we do not think that, when it comes to reacting to sexual assault, it is always wrong to put one’s own judgment over that of the law. As enforced, the law is deeply flawed in many ways, including systematic ones. But we also think it is a mistake to suppose that in general, acting on one’s own sexual assault finding, when there is insufficient evidence for criminal remedies, even amounts to putting one’s own judgment over that of the law. Suppose a sexual assault allegation is found insufficiently well-supported to justify legal remedy. From this, it is very natural to infer that the evidence available 42 Article 11 of the Universal Declaration of Human Rights has it that “[e]veryone charged with a penal offence has the right to be presumed innocent until proved guilty according to law in a public trial.”
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
268 Bianca Crewe and Jonathan Jenkins Ichikawa was insufficient for knowledge that the assault occurred. If we assume that a uni versity fact-finding procedure has access to the same body of evidence as law enforcement, then university fact finders don’t have evidence sufficient for knowledge either. So, they too ought not to act on the allegation. One way to resist the argument is to deny the link between knowledge and actionability. Another, discussed in section 5, is to invoke pragmatic encroachment, suggesting that knowledge is possible only in the lower-stakes scenario, perhaps holding that the stakes of university findings, while serious, are not so weighty as those involved in legal decisions about whether someone ought to be convicted of a crime. We focus now on a different avenue of resistance: the contextualist one. Contextualism about knowledge ascriptions is the view that sentences containing “knows” are context sensitive. Like indexicals (“I”, “now”, “that”), gradable adjectives (“humid”, “skillful”, “feminist”), and other terms, “knows” is susceptible to diverse semantic implementations in different conversational contexts.43 For example, conversations that focus on skeptical ideas might tend to evoke higher standards for “knows”. One influential motivation for contextualism is its ability to explain patterns of intuition of this type. Perhaps in courtroom conversational contexts, high epistemic standards are in operation (perhaps because of the weighty practical matters at stake); consequently, “we cannot know what happened” might be true in the jurors’ conversation. But in other contexts, even given the same evidence, when different actions are contemplated, “we know that a sexual assault occurred” might be true. Contextualism is controversial, but in our view, it has much to commend it.44 A contextualist can tell a simple story about differential standards for action in criminal contexts and other contexts, such as university action: one can count as “knowing” that someone has committed a serious offense in a conversation about how the university ought to respond to it, without counting as “knowing” it in a conversation about whether the state ought to incarcerate the perpetrator. This framework straightforwardly invalidates the tempting argument that, since there wasn’t evidence sufficient for criminal conviction, there is no way to know what happened, so it would be unwarranted to act on the basis of a mere allegation. Here is an example of the kind of fallacious argument we have in mind. In 2016, after the University of British Columbia announced a plan for John Furlong to keynote a high-profile university event, activists protested the invitation, citing 43 Influential early discussions of contextualism include Stine (1976); DeRose (2009); Lewis (1996); and Cohen (1999). 44 DeRose (2009) and Ichikawa (2017) give extended sympathetic treatments of contextualism, including an emphasis on the connections between knowledge and action. Note also that even if con textualism is false in general, there is a strong case to be made that legal settings often invoke a dis tinctive sense of “knows”, according to which one “knows” only that which one is permitted to take on board for the purpose of the legal discussion. Thanks to Alex Guerrero here.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Rape Culture and Epistemology 269 dozens of allegations of child abuse, including sexual abuse, by former students from the Canadian First Nations schools where Furlong taught in the 1970s. UBC briefly rescinded the invitation, but quickly reversed that decision. UBC President Santa Ono apologized to Furlong, writing that the withdrawal was made “without proper consideration of its potential impact on Mr. Furlong or his family.” He reinstated the invitation, with this statement: Notwithstanding what led to the decision to cancel Mr. Furlong’s keynote address, I have made it my decision as president of the university to reverse course because it is simply the right thing to do. I decided this after better informing myself with the facts, including Mr. Furlong’s stellar reputation in the fields of business, leadership and sport, the diverse views of our many stake holders, and, as importantly, the judicial record. The British Columbia Civil and Supreme Courts have ruled in favour of Mr. Furlong in every matter that has come before them. The university had no basis to put its judgment above theirs. https://president.ubc.ca/homepage-feature/2017/01/09/ubc-reinstates-johnfurlong-as-keynote-speaker/
Two things are notable about this passage. First, it is a clear example of the ‘himpathy’ phenomenon we discussed in section 5, in this case reflecting settler-colonial priorities in addition to patriarchal ones. In this example, the interests of the accused are made salient, and those of the accusers are ignored. Second, it is an example of a fallacious argument from the lack of a legal finding of guilt to the conclusion that UBC lacks relevant knowledge to justify choosing a different speaker. The testimony of dozens of former students (some of which were given in sworn affidavits) did, in our view, put UBC officials in a position to know that Furlong was not a good choice of someone to venerate. This result, we think, is significant on its own—it refutes the deferential thought one hears every time one engages in public discourse about university sexual assault policies: “leave this criminal matter to law enforcement.” Even if we set aside the many problems with the way law enforcement is practiced, because criminal contexts tend to evoke high epistemic standards, their conclusions do not always bear directly on what counts as “known” in other contexts, like university ones. Since we know many things that haven’t been proven in court,45 the suggestion that we ought to act only on that which has been proven in court amounts to the suggestion that we ought to ignore many of the things we know. This is implausible in general; for example, when it comes to ignoring the known fact that one’s student is a rapist, it is negligently harmful. 45 One might say: since we count as “knowing” many things relative to the standards appropriate for the actions we are contemplating.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
270 Bianca Crewe and Jonathan Jenkins Ichikawa Indeed, the argument is sometimes taken even further, to deny the appropriateness of individual actions in response to sexual assault allegations. Here is an anecdote illustrating the thought.46 We recently observed a discussion about a call for papers for a philosophy workshop dedicated to the work of Professor Vine, an influential philosopher who had been accused of sexual misconduct by multiple former students and employees. Another philosopher, Alvin, saw the call for papers, and suggested that, in light of serious and credible allegations against Vine, academics should not submit papers for this workshop. But Wade, another academic, took exception to Alvin’s advice, arguing that the allegations against Vine, though serious, were unproven. Unless one had add itional evidence, Wade said, beyond the detailed testimony that was then public, it would be irresponsible and unfair to Vine to take such an action as declining to submit to this workshop. As our discussion above should make clear, we have several points of disagreement with Wade in this story. For one thing, a detailed public allegation by a credible witness, as there was in this case, is in fact excellent evidence of wrongdoing. For another, Wade’s emphasis on the possible harms to Vine leave out a very important part of the story—Vine’s accusers are harmed by discounting or setting aside their testimony; participating in the workshop despite their testimony would have contributed to the normalization of sexual assault in professional philosophy. Third, Wade’s argument problematically assumes that suspending judgment on the allegations is the safe, “neutral” choice, and that taking a stance by refusing to submit to the workshop is an act of radicalism that requires a strong justification. But there is no clear reason why submitting to the workshop is the choice of default; indeed, in an intuitive sense, “not submitting a paper to a workshop” isn’t an action at all—it’s what one would do by default, absent some good enough reason to submit it. Assuming otherwise is an example of a harmful idealization in the sense of section 5. Fourth, however, this case is also an illustration of the possible contextualist fallacy. In a context of a jury deliberation, perhaps “we cannot know whether these allegations are true” is true, because in that context, very strong epistemic standards are in effect; this simply does not imply that such sentences are true in more ordinary contexts, such as those operative in deciding whether to submit a paper to a workshop. To suppose otherwise is to assume that very high epistemic standards are always in effect, with the result that we “know” very little. Such skepticism is a recipe for inaction. So far, we have not said much about how it is that epistemic standards are generated, beyond a gesture at the idea that when the stakes are higher, there is a tendency to employ stronger standards for “knows”. In section 7, we foreground
46 Because the conversation recounted took place in a private forum, we use pseudonyms here.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Rape Culture and Epistemology 271 this question. We will suggest that there is an interesting connection between the ideas about rape culture mentioned in section 2 and the metasemantic machinery of contextualism discussed in section 6.
7. Epistemic Standards, Normativity, and Rape Culture Contextualism is a claim about the truth conditions of natural language sentences involving ‘knows’. It is motivated, at least in part, by descriptive facts about the way ‘knows’ sentences are used, and it suggests that those uses amount to true ones. In this way, contextualism is conservative, rather than revolutionary. It is, in typical instances, motivated by a methodological commitment to preserving the truth of ordinary knowledge ascriptions.47 Nevertheless, we think it would be a mistake to suppose that contextualism supports any kind of complacency about knowledge ascriptions. On the contrary, the normative issues we began with about rape culture and biases towards the status quo interact deeply with contextualism and knowledge ascriptions, pointing to an undertheorized avenue for political engagement. In this final section, we explore those connections. The result will be an example of the sort of complex relationship between feminist epistemology and contextualism gestured at in Brister (2017, 58). We begin with two important metasemantic questions. First: can factors outside of individual speakers’ knowledge and control influence the epistemic standards they employ? And second: are there substantive normative questions about what epistemic standards one ought to employ? If the answer to the first question is yes, then it will be relatively easy for speakers to make false ‘knows’ claims, due to an incorrect assessment of their own standards. If the answer to the second question is yes, then, even if someone is saying something that is true, it might be that they ought to be speaking in a context where that sentence wouldn’t be true. Each of these possibilities corresponds to a kind of normative objectivity about knowledge ascriptions, pointing to ways in which whether one is proceeding properly may extend beyond the subject access. An example of objectivity of the first sort has been defended in a series of papers by Christopher Gauker.48 Gauker’s project emphasizes the mind-independence of contextual parameters, holding that, for example, whether a skeptical possibility is relevant depends more on the features of the situation at issue than on the speaker’s attention or intentions. On such a view, speakers might falsely believe that a certain skeptical worry is contextually relevant—this would lead to false ‘knowledge’ denials. For example, in a knowledge ascription involving testimony 47 DeRose (2009) is particularly clear about this kind of motivation; see especially 56–9. 48 See, e.g. Gauker (2008) and Gauker and Mion (2017).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
272 Bianca Crewe and Jonathan Jenkins Ichikawa about sexual assault, one might hold that objective features of one’s situation imply a less skeptical epistemic standard, even if the speaker thinks a more stringent standard is operative. If so, individuals may face significant challenges in recognizing their own contextual situations; if one is unaware of the operation of these tacit contextual standards, one may be a very poor judge of knowledge ascriptions. Improved self-awareness, including awareness of one’s contextual situation, is crucial.49 This points to the structural nature of the problem: since rape culture isn’t under the immediate control of the individuals negatively affected by it or perpetrating it, the development of alternative hermeneutical resources for understanding and responding to sexual assault must involve a collective cultural effort, as well as the cultivation of certain epistemic virtues sensitive to particularities of context.50 Even if one allows that speakers have considerable fiat over the relevant features of their conversational contexts, there is still room for a more objective assessment. Suppose, for example, that the context sensitivity of ‘knows’ is largely a matter of something like David Lewis’s “Rule of Attention,” according to which whether an alternative is relevant depends on whether the speaker is thinking about it.51 As has often been observed, Lewis’s is a skeptic-friendly version of con textualism: if one is thinking about a skeptical scenario, and a subject’s evidence doesn’t rule that scenario out, then in one’s context, the subject can’t “know” that the scenario doesn’t obtain. For example, if a speaker is thinking about the possi bility that a witness is lying, then, in the speaker’s context, no one will count as “knowing” the fact testified unless their evidence is sufficient to rule out a lie. On such a picture, speakers will have considerable control over the relevant features of their conversational contexts; compared to Gauker’s framework, we’d expect mistaken knowledge ascriptions to be rarer on a Lewisian story. But this doesn’t imply that there aren’t substantive normative questions about knowledge ascriptions. Truth or falsity of utterances isn’t the only thing we might be interested in evaluating. We may assume that subjects have the ability to implement whatever contextual parameters they like; that doesn’t mean there aren’t norms governing these conversational decisions. For one example of the kind of view we are thinking of, consider Sarah Moss’s (2018) discussion of belief in harmful stereotypes. Moss is interested in cases where statistical evidence seems to support a probabilistic claim that it would be 49 As a rough analogy, consider Miranda Fricker’s discussion of hermeneutical marginalization. Some aspects of some people’s experiences are particularly difficult for people to comprehend because they lack the necessary hermeneutical resources. This is not a failure that can be diagnosed ‘from the inside’; there is no simple recipe for detecting and avoiding it. Instead, one must develop a disposition of openness to experiences that are difficult for one to understand, recognizing that one’s hermeneut ical resources are limited. 50 This is one explanation for some of the value of consciousness-raising : see Kelland (2016). 51 Lewis (1996, 559–60). The subject’s attention is one of several factors that influence relevance in Lewis’s framework.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Rape Culture and Epistemology 273 intuitively morally problematic to accept. To take one of Moss’s examples, suppose one knows on general statistical grounds that most of the women in a given office building are administrative assistants, and that one encounters a woman there, with no additional evidence relevant to her role. Moss wishes to deny that one can know on these grounds that she is probably an administrative assistant. To do this, she invokes a broadly Lewisian form of contextualism, along with what she calls the “Rule of Consideration.” This is a moral norm to the effect that “in many situations where you are forming beliefs about a person, you morally should keep in mind the possibility that they might be an exception to statistical generalizations.”52 The idea here is that one morally ought to be thinking about certain possibilities, such as the possibility that the woman one sees is a statistically unusual woman who is probably a manager, rather than an administrative assistant. And if one were thinking about such a possibility, it would count as relevant for the purpose of the knowledge ascription. If one flouts the Rule of Consideration, one might speak truly with an utterance ascribing “knows that she’s probably an administra tive assistant” to someone in this situation. But Moss’s view is that one has a moral obligation not to flout the Rule of Consideration. If one ignores the possibility that this woman is exceptional, one has spoken truly, but behaved badly. This isn’t the venue to adjudicate the plausibility of Moss’s particular suggestion; we mention it as a vivid example of the ways in which moral considerations and contextual parameters can interact. There can be questions about what kinds of epistemic standards are appropriate. Moreover, most contextualists hold that speaking as if a given standard is in effect has a tendency of putting that standard into effect.53 So knowledge denials in debatable cases—such as testimony—will have a tendency of raising the standards for knowledge. An assertion to the effect that “there’s no way to know what really happened” can, if left unchallenged, lead to a more skeptical context where that sentence is true. Given the connections between knowledge ascriptions and action, such skeptical standards can have dramatic sociopolitical effects. For reasons such like these, Esa Diaz-Leon has recently suggested that, given contextualism, a knowledge ascription carries two important roles in a conversation. In addition to semantically expressing its content, it also conveys the pragmatic suggestion that the standard employed is the proper one. Diaz-Leon writes, about a disagreement between a Moorean who ascribes “knowledge” and a skeptic who denies “knowledge”, that: Moore and the skeptic (might) express compatible propositions at the semantic level, but they still express a genuine disagreement. In particular, their disagreement can be explained at the pragmatic level. That is, when the skeptic
52 Moss (2018, section 10.4).
53 Lewis (1979) is the locus classicus.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
274 Bianca Crewe and Jonathan Jenkins Ichikawa utters [a ‘ knowledge’ denial], she is expressing the proposition that Moore does not satisfy certain (very high) standards of justification, whereas when Moore utters [a ‘knowledge’ ascription], he is expressing the proposition that he does satisfy certain lower standards of justification. These two propositions are compatible. But the skeptic is also conveying the information that “knows” should be used in a way so that one would count as knowing a proposition only when one satisfies certain very high standards, whereas Moore is also conveying the information that “knows” should not be used in that way. (Diaz-Leon 2017, 75)54
Diaz-Leon’s emphasis in her paper is to explain the sense of disagreement that persists in cases like these, where contextualism seems to allow that the speakers are asserting compatible truths. But her proposal here, with which we are sympathetic, offers an important framework for thinking about an additional political power of knowledge ascriptions, given contextualism. The flexi bil ity of contextualism corresponds to a flexibility regarding which testimony to take seriously. An extreme example will make the connection vivid. Imagine a community that systematically employs high standards when talking about what women do or do not “know” and whether one can “know” on the basis of women’s testimony, and low standards when talking about the parallel issues for men. In this community, people will react to women’s descriptions of their experiences with utterances like, “People sometimes misinterpret other people’s actions, so although she thinks he was being sexist, she doesn’t know it.” But when men describe their experiences, skeptical possibilities don’t tend to come to mind, so people say things like, “He knows what happened because he saw it himself.” And female testimony will be greeted with the skepticism that comes with high standards: “We can’t know unless we have some way to make sure she isn’t lying,” with no similar skepticism about male testimony. This is a thought experiment, though not, we think, a terribly far-fetched one. The point is to demonstrate that contextualism generates a dimension allowing the truth of utterances to come rather far apart from their normative aptness. Although there are many things deeply wrong with the misogynistic society posited, a tendency to make false knowledge ascriptions needn’t be one of them. The corollary is: one can object to assertions that reflect and contribute to rape culture without insisting that the contents of the assertions are false. One way in which assertions can contribute to rape culture is by managing epistemic standards in a way friendly to perpetrators of sexual assault. 54 Diaz-Leon is applying to “knows” a strategy she attributes to Plunkett and Sundell (2013), who make a similar suggestion about moral evaluative language.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Rape Culture and Epistemology 275 “He-said–she-said” discourse is a case in point. If a woman accuses a man of sexual assault, and he rebuts that there was sexual contact, but that it was consen sual, this is literally a case in which “he said” one thing and “she said” something inconsistent. Pointing this out tends to have the conversational effect of inducing speakers to employ high epistemic standards, such that “We can’t know what really happened” will come out true. But that doesn’t mean we should modify the context in this way. If this is one of the many examples where “her” testimony is much more credible than “his”, it may well be that lower standards are appropri ate, such that we can “know” that misconduct occurred. (This will be particularly plausible when informal action is being contemplated—i.e. decisions about whom to invite to a party, or whether to submit to a conference.) These actions indicate what is acceptable in the context of shared social life, and do not hinge on a par ticular definition of relevant communities (e.g. the state) or of the standards of that community (e.g. proof beyond a reasonable doubt). The ability to affect the conversational context in a particular way is a kind of social power. Not all parties to the disagreements Diaz-Leon describes are on equal footing. Like most social power, this ability is not distributed impartially; in a rape culture, those seeking to raise the standards governing sexual assault alle gations will have the field tilted in their favor. So, we agree with Evelyn Brister’s (2009) suggestion that it is helpful and informative to look to contextualism through a feminist lens, with an eye towards the deployment of social power. Brister emphasizes that, in a contextualist framework, “skepticism is a kind of power grab that puts other epistemic agents on the defensive and at a disadvan tage. While some might believe that knowledge is power, the skeptic recognizes denying knowledge is power.” She continues: Recognizing the power dynamics of skeptical arguments allows us to examine the interests that philosophical skepticism serves. To begin with, philosophical skepticism distrusts all knowledge claims and avoids falsehood at the cost of dismissing possible truths: this represents a deeply conservative attitude toward the production of knowledge. For this reason, skepticism poses a special danger to novel claims that may already face a higher burden of proof and where skeptical objections may then prove decisive. In this way philosophical skepticism can exacerbate existing conditions of epistemic unfairness. (Brister 2009, 682)
And: The skeptic would deny a collective responsibility for coordinated inquiry—and since all the institutions of scientific investigation, justice, and political account ability depend on decision making guided by knowledge claims, the skeptic’s change of context undercuts our ability to act on the basis of justified beliefs. (Brister 2009, 683)
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
276 Bianca Crewe and Jonathan Jenkins Ichikawa We agree with both thoughts here. To Brister’s observations, we add that real-world skepticism isn’t a monolithic raising of standards, the way introduc tory epistemology courses sometimes imagine that it is; it is a tool that is used selectively and strategically to promote entrenched interests. So, the flexibility of knowledge ascriptions, given contextualism, comes along with significant social power. The decision to employ some standards rather than others is a political one. It shouldn’t be surprising, then, to suggest that this power tends to be wielded in a way that tends to protect the interests of the status quo. In general, social power tends to be used to reinforce itself. The case of contextualism is particularly interesting in this respect, due to its largely invisible operation. Notoriously, contextualism posits context sensitivity in natural language, about which many ordinary speakers are unaware. This is part of the reason why con textualism is a controversial thesis. Some epistemologists have rejected context ualism on the grounds that it is implausible that language could exhibit this kind of context sensitivity without ordinary speakers’ recognizing it.55 We are sanguine about this “semantic ignorance” problem; we see no particular tension in the idea that natural language exhibits surprising features ordinary speakers may be con fused about. What we would like to emphasize is that this confusion contributes to speakers’ ability to wield social power in an unreflective way. If one is unaware of the choice point of selecting an epistemic standard in a given situation, one may employ various biases, under a misapprehension of objectivity. If, for example, one has a natural tendency to favor the testimony of the socially powerful over that of marginalized people, one might, in a way similar to the society in our thought experiment above, employ higher standards in some cases than others, without realizing it. We think there’s good reason to suspect this is reasonably common.56 In The Epistemology of Resistance, José Medina (2013) argues that structural inequality distorts epistemic relations between privileged people and oppressed people, resulting in a “numbness” among the privileged to injustices against the oppressed. As a consequence, the default epistemic positions of privileged communities and individuals are often oppressive ones. Applying this to our investigations here means that even without overt sexist biases or outright ill will, it is possible to make knowledge ascriptions in a way that maintains and supports 55 Schiffer (1996) was particularly influential along these lines. See also Hawthorne (2003, 107–9) and Stanley (2005, 115–20). Greenough and Kindermann (2017) give a contemporary overview of this kind of challenge and offer a response on behalf of the contextualist. 56 This might be an alternate way of implementing some of Miranda Fricker’s ideas about testimo nial injustice. Testimonial injustice is a “credibility deficit owing to identity prejudice in the hearer” (Fricker 2007, 28). A credibility deficit is naturally thought of as a lower-than-deserved confidence in one’s credibility. An alternate proposal, inspired by our remarks above, might characterize the inappropriate invocation of high epistemic standards as a kind of credibility deficit. Perhaps one can commit an epistemic injustice, not by having too low an opinion of someone’s credibility, but by set ting too high a bar for accepting their word.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Rape Culture and Epistemology 277 rape culture, and does violence to those who are attempting to be heard. (Indeed, “possible” here is an understatement. It’s what we should expect.) As we have argued, putatively “neutral” responses like suspending judgment or “doing nothing” can themselves be substantive contributions to, and reflections of, rape culture. This is true both because it normalizes bad behavior, contributing to the general idea that it is acceptable, and, given contextualism, because it amounts to accommodating harmful conversational contexts. Indeed, contextualism is valuable in part because of the way it deals with skepticism, but attention to the ways in which it can be used as a conservative force is crucial to applying contextualism in an anti-oppressive way. Transparency regarding when harmful conversational contexts are being activated is itself a significant achievement for social justice. Following feminist standpoint theorists, we think marginalized individuals are typically situated best to know when this is occurring. Our project here has been in many ways a continuation of the critical work of feminist philosophers who argue that the “view from nowhere” is in fact often a view from a very particular and situated location. However, acknowledging how tendencies towards conservatism, risk aversion, and alignment with the status quo can have oppressive effects is not merely a critical endeavor. Indeed, though rape culture is a structural phenomenon, there are things that we can do collectively and as individuals to resist its negative epistemic effects. The implicit biases that motivate the reception of sexual assault narratives are not unchangeable. Moreover, we do hold people morally accountable for addressing such biases, as well as their active, motivated epistemic ignorance regarding certain politically loaded concepts or identities.57 Consciousness raising is also invaluable as both a means of introducing new hermeneutical resources into the public eye, and a way for individuals to come to recognize their own contexts. Feminist theorizing of the second wave—the heyday of consciousness raising—tells us that “the personal is political.” Accounting for this at the level of epistemic theory has been part of our project here.
References Adler, Jonathan (2015). “Epistemological Problems of Testimony”, in Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy, summer edn. https://plato.stanford. edu/archives/sum2015/entries/testimony-episprob. Ahrens, Courtney E. (2006). “Being Silenced: The Impact of Negative Social Reactions on the Disclosure of Rape.” American Journal of Community Psychology 38(3–4): 263–74. 57 See, e.g. Mills (2005, 2007); Sullivan (2006); Holroyd (2012); and the first five chapters in Brownstein and Saul (2016).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
278 Bianca Crewe and Jonathan Jenkins Ichikawa Alcoff, Linda. (1999). “On Judging Epistemic Credibility: Is Social Identity Relevant?” Philosophic Exchange 29(1): 1. Anderson, Michelle (2003). “Prostitution and Trauma in U.S. Rape Law.” Journal of Trauma Practice 2 (3/4): 75–92. Basile, Kathleen C., Matthew J. Breiding, and Sharon G. Smith, (2016). “Disability and Risk of Recent Sexual Violence in the United States.” American Journal of Public Health 106(5): 928–33. Brister, Evelyn (2009). “Feminist Epistemology, Contextualism, and Philosophical Skepticism.” Metaphilosophy 40(5): 671–88. Brister, Evelyn (2017). “Feminism and Contextualism,” in Jonathan J. Ichikawa (ed.), Routledge Handbook of Epistemic Contextualism. London: Taylor & Francis, 57–67. Brownmiller, Susan (1975). Against Our Will: Men, Women, and Rape. New York: Penguin Books. Brownstein, Michael and Jennifer Saul (eds) (2016). Implicit Bias and Philosophy, Vol. 2: Moral Responsibility, Structural Injustice, and Ethics. Oxford: Oxford University Press. Buchwald, Emile, Pamela Fletcher, and Martha Roth (2005). Transforming a Rape Culture. Minneapolis, MN: Milkweed Editions. Burt, Martha R. (1980). “Cultural Myths and Supports for Rape.” Journal of Personality and Social Psychology 38: 217–30. Cohen, Stewart (1999). “Contextualism, Skepticism, and the Structure of Reasons.” Philosophical Perspectives 13: 57–89. Curtis, Annaleigh (2017). “Due Process Demands as Propaganda: The Rhetoric of Title IX Opposition.” Yale Journal of Law and the Humanities 29, https://ssrn.com/ abstract=2980500. Davis Robert C., Ellen Brickman, and Timothy Baker (1991). “Supportive and Unsupportive Responses of Others to Rape Victims: Effects of Concurrent Victim Adjustment.” American Journal of Community Psychology 19: 443–51. DeRose, Keith (2009). The Case for Contextualism. Oxford: Oxford University Press. Diaz-Leon, Esa (2017). “Epistemic Contextualism and Conceptual Ethics,” in Jonathan J. Ichikawa (ed.), Routledge Handbook of Epistemic Contextualism. London: Taylor & Francis, 71–80. Donovan, Roxanne and Michelle Williams (2002). “Living at the Intersection: The Effects of Racism and Sexism on Black Rape Survivors.” Women and Therapy 25(3/4): 95–105. Du Mont, Janice, Karen-Lee Miller, and Terri L. Myhr (2003). “The Role of ‘Real Rape’ and ‘Real Victim’ Stereotypes in the Police Reporting Practices of Sexually Assaulted Women.” Violence against Women 9(4): 466–86. Dworkin, Emily R., Samantha L. Pittenger, and Nicole E. Allen (2016). “Disclosing Sexual Assault within Social Networks: A Mixed-Method Investigation.” American Journal of Community Psychology 57(1–2): 216–228.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Rape Culture and Epistemology 279 Fantl, Jeremy and Matthew McGrath (2009). Knowledge in an Uncertain World. Oxford: Oxford University Press. Foley, Linda A., Christine Evancic, Karnik Karnik, Janet King, and Angela Parks (1995). “Date Rape: Effects of Race of Assailant and Victim and Gender of Subjects on Perceptions.” Journal of Black Psychology 21(1): 6–18. Fricker, Miranda (2007). Epistemic Injustice. Oxford: Oxford University Press. Frohmann, Lisa (1991). “Discrediting Victims’ Allegations of Sexual Assault: Prosecutorial Accounts of Case Rejections.” Social Problems 38 (2): 213–26. Gauker, Christopher (2008). “Zero Tolerance for Pragmatics.” Synthese 165(3): 359–71. Gauker, Christopher and Giovanni Mion (2017). “The Mind-Independence of Contexts for Knowledge Attributions,” in Jonathan J. Ichikawa (ed.), Routledge Handbook of Epistemic Contextualism. London: Taylor & Francis, 455–64. George, Linda K., Idee Winfeld, and Dan G. Blazer (1992). “Sociocultural Factors in Sexual Assault: Comparison of Two Representative Samples of Women.” Journal of Social Issues 48: 105–25. George, William H. and Lorraine J. Martínez, (2016). “Victim Blaming in Rape: Effects of Victim and Perpetrator Race, Type of Rape, and Participant Racism.” Psychology of Women Quarterly 26 (2): 110–19. Gettier, Edmund (1963). “Is Justified True Belief Knowledge?,” Analysis 23(6): 121–3. Gilmore, Leigh (2017). Tainted Witness: Why We Doubt What Women Say about Their Lives. New York: Columbia University Press. Greenough, Patrick and Dirk Kindermann (2017). “The Semantic Error Problem for Epistemic Contextualism,” in Jonathan J. Ichikawa (ed.), Routledge Handbook of Epistemic Contextualism. London: Taylor & Francis, 305–20. Golding, Jacqueline M., Judith M. Siegel, Susan B. Sorenson, M. Audrey Burnam, and Judith A. Stein (1989). “Social Support Sources Following Sexual Assault.” Journal of Community Psychology 17: 92–107. Hampton, Jean (1999). “Defining Wrong and Defining Rape,” in Keith BurgessJackson (ed.), A Most Detestable Crime: New Philosophical Essays on Rape. New York: Oxford University Press, 118–56. Hawthorne, John (2003). Knowledge and Lotteries. Oxford: Oxford University Press. Hawthorne, John and Jason Stanley (2008). “Knowledge and Action.” Journal of Philosophy 105(10): 571–90. Healicon, Alison (2016). The Politics of Sexual Violence: Rape, Identity, and Feminism. London: Palgrave Macmillan. Hill Collins, Patricia (1990). Black Feminist Thought: Knowledge, Consciousness, and the Politics of Empowerment. New York: Routledge. Hill Collins, Patricia (1991). “Learning from the Outsider Within,” in Mary Margaret Fonow and Judith A. Cook, (eds ), Beyond Methodology: Feminist Scholarship as Lived Research. Bloomington, IN: Indiana University Press, 35–9.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
280 Bianca Crewe and Jonathan Jenkins Ichikawa Holroyd, Jules (2012). “Responsibility for Implicit Bias.” Journal of Social Philosophy 43(3): 205–362. Hundleby, Catherine E. (2016). “The Status Quo Fallacy: Implicit Bias and Fallacies of Argumentation,” in Michael Brownstein and Jennifer Saul (eds ), Implicit Bias & Philosophy, Vol. I: Metaphysics and Epistemology. Oxford: Oxford University Press. Hynes, Michelle and Barbara Lopes Cardozo (2000). “Sexual Violence against Refugee Women.” Journal of Womens Health and Gender Based Medicine 9(8): 819–23. Ichikawa, Jonathan Jenkins (2017). Contextualising Knowledge. Oxford: Oxford University Press. Ichikawa, Jonathan Jenkins and Benjamin W. Jarvis (forthcoming). “Hybrid Virtue Epistemology and the A Priori,” in Dylan Dodd and Elia Zardini (eds), The A Priori: Its Significance, Grounds, and Extent. Oxford: Oxford University Press. Johnson, Antuan M. (2017). “Title IX Narratives, Intersectionality, and Male-Biased Conceptions of Racism,” 9 Georgetown Journal of Law & Modern Critical Race Perspectives 57–75. Jones, Karen (2002). “The Politics of Credibility,” in Louise M. Antony and Charlotte E. Witt, (eds ), A Mind of One’s Own: Feminist Essays on Reason and Objectivity, 2nd edn. Boulder, CO: Westview Press, 154–176. Jordan, Jan (2004). “Beyond Belief?: Police, Rape and Women’s Credibility,” Criminology & Criminal Justice 4 (1): 29–59. Kelland, L. (2016). “A Call to Arms: The Centrality of Feminist Consciousness-Raising Speak-Outs to the Recovery of Rape Survivors.” Hypatia (31)4. Khalifeh, H., Moran, P., Borschmann, R., et al. (2015). “Domestic and Sexual Violence against Patients with Severe Mental Illness.” Psychological Medicine 45(4): 875–86. Koss, Mary P. (1985). “The Hidden Rape Victim: Personality, Attitudinal, and Situational Characteristics.” Psychology of Women Quarterly 9: 193–212. Kushel, Margot B., Jennifer L. Evans, Sharon Perry. et al. (2003) “No Door to Lock: Victimization among Homeless and Marginally Housed Persons.” Archives of Internal Medicine (163)20: 2492–9. Lackey, Jennifer (2008). Learning from Words. Oxford: Oxford University Press. Levy, Neil (2019). “Due Deference to Denialism: Explaining Ordinary People’s Rejection of Established Scientific Findings.” Synthese 196, 313–327. Lewis, David (1996). “Elusive Knowledge.” Australasian Journal of Philosophy 74(4): 549–67. Lewis, David (1979). “Scorekeeping in a Language Game.” Journal of Philosophical Logic 8(1): 339–59. Lisak, David, Lori Gardinier, Sarah C. Nicksa, and Ashley M. Cote (2010). “False Allegations of Sexual Assualt: An Analysis of Ten Years of Reported Cases.” Violence against Women 16 (12): 1318–34. Littlejohn, Clayton (2012). Justification and the Truth-Connection. Cambridge: Cambridge University Press.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Rape Culture and Epistemology 281 Mack, Kathy (1993). “Continuing Barriers to Women’s Credibility: A Feminist Perspective on the Proof Process.” Criminal Law Forum 4 (2): 327. MacKinnon, Catharine (1989). Towards a Feminist Theory of the State. Cambridge, MA: Harvard University Press. Manne, Kate (2018). Down Girl: The Logic of Misogyny. New York: Oxford University Press. McAuslan, Pamela (1998). “After Sexual Assault: The Relationship between Women’s Disclosure, the Reactions of Others, and Health.” Dissertation manuscript, Wayne State University, Detroit, MI. Medina, José (2013). The Epistemology of Resistance: Gender and Racial Oppression, Epistemic Injustice, and Resistant Imaginations. Oxford: Oxford University Press. Mills, Charles (2005). “ ‘Ideal Theory’ as Ideology.” Hypatia. (20)3: 165–184. Mills, Charles (1997). The Racial Contract. New York: Cornell University Press. Mills, Charles (2007). “White Ignorance,” in Shannon Sullivan and Nancy Tuana (eds), Race and Epistemologies of Ignorance. Albany, NY: SUNY Press. Morrison, Toni (1992). “Introduction: Friday on the Potomac,” in T oni Morrison (ed.), Race-ing Justice, En- Gendering Power: Essays on Anita Hill, Clarence Thomas, and the Construction of Social Reality. New York: Pantheon, vii–xxx. Moss, Sarah (2018). Probabilistic Knowledge. Oxford: Oxford University Press. Mulder, Marianne R. and Frans Willem Winkel (1996). “Social Workers’ and Police Officers’ Perceptions of Victim Credibility: Perspective-Taking and the Impact of Extra-Evidential Factors.” Psychology, Crime and Law 2(4): 307–19. Nagel, Jennifer (2010). “Epistemic Anxiety and Adaptive Invariantism.” Philosophical Perspectives 24(1): 407–35. Narayan, Uma (1988). “Working Together across Difference: Some Considerations on Emotions and Political Practice.” Hypatia 3 (2): 31–48. Nelson, Mark T. (2010). “We Have No Positive Epistemic Duties.” Mind 119(473): 83–102. Pateman, Carole (1988). The Sexual Contract. Oxford: Polity Press. Plunkett, David and Tim Sundell (2013). “Disagreement and the Semantics of Normative and Evaluative Terms.” Philosophers’ Imprint 13(23): 1–37. Rennison, Callie Marie (2002). Rape and Sexual Assault: Reporting to Police and Medical Attention, 1992–2000, NCJ Publication No. 194530. Washington, DC: US Department of Justice, Bureau of Justice Statistics. Rich, Adrienne (1994). Blood, Bread, and Poetry: Selected Prose 1979–1985. New York: Norton. Ritter, Nancy (2011). The Road Ahead: Unanalyzed Evidence in Sexual Assault Cases, NCJ Publication No. 233279. Washington, DC: US Department of Justice Office of Justice Programs, National Institute of Justice (NIJ). Schiffer, Stephen (1996). “Contextualist Solutions to Scepticism.” Proceedings of the Aristotelian Society 96(1): 317–33.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
282 Bianca Crewe and Jonathan Jenkins Ichikawa Snyder, Mark, Elizabeth D. Tanke, and Ellen Berscheid (1977). “Social Perception and Interpersonal Behaviour: On the Self-Fulfilling Nature of Social Stereotypes.” Journal of Personality and Social Psychology 35: 656–66. Stanley, Jason (2005). Knowledge and Practical Interests. Oxford: Oxford University Press. Stine, Gail C. (1976). “Skepticism, Relevant Alternatives, and Deductive Closure.” Philosophical Studies 29: 249–61. Sullivan, Shannon (2006). Revealing Whiteness: The Unconscious Habits of Racial Privilege. Bloomington, IN: Indiana University Press. Tang, Kwong-leung (1998). “Rape Law Reform in Canada: The Success and Limits of Legislation.” International Journal of Offender Therapy and Comparative Criminology (42)3: 258–270. Tyler, Kimberly A. and Morgan R. Beal (2010). “The High-Risk Environment of Homeless Young Adults: Consequences for Physical and Sexual Victimization.” Violence and Victims 25(1): 101–15. Ullman, Sarah E. (2000). “Psychometric Characteristics of the Social Reactions Questionnaire: A Measure of Reactions to Sexual Assault Victims.” Psychology of Women Quarterly 24(3): 257–71. Ward, Jeanne and Vann, Beth (2002). “Gender-Based Violence in Refugee Settings.” Lancet 360(Suppl.): 13–14. Weatherson, Brian (2017). “Interest-Relative Invariantism,” in Jonathan J. Ichikawa (ed.), Routledge Handbook of Epistemic Contextualism. London: Taylor & Francis, 240–53. Weatherson, Brian (2012). “Knowledge, Bets, and Interests,” in Jessica Brown and Mikkel Gerken (eds ), Knowledge Ascriptions. Oxford: Oxford University Press, 75–103. Wedgwood, Ralph (2008). “Contextualism about Justified Belief.” Philosophers’ Imprint 8(9): 1–20. Wenzel, Suzanne L., Joan S. Tucker, Marc N. Elliott, Grant N. Marshall, and Stephanie L. Williamson (2004). “Physical Violence against Impoverished Women: A Longitudinal Analysis of Risk and Protective Factors.” Womens Health Issues 14(5): 144–54. Wyatt, Gail Elizabeth (1992). “The Sociocultural Context of African American and White American Women’s Rape.” Journal of Social Issues. 48 (1): 77–91.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 19/03/21, SPi
12
Feminist Pornography as Feminist Propaganda, and Ideological Catch-22s Aidan McGlynn*
1. Introduction The second-wave anti-pornography movement has come to be identified with figures like Susan Brownmilller, Andrea Dworkin, Catharine MacKinnon, and Gloria Steinem, and (fairly or not) with a stance that embraced some form of censorship of some pornographic materials.1 This tendency has been strikingly true in discussions of pornography within analytic philosophy too, with attempts to refine and defend MacKinnon’s claims that pornography subordinates and silences women, made with an eye towards issues about freedom of speech, dom inating the literature since Rae Langton and Jennifer Hornsby’s agenda-setting papers published in 1993. However, the Dworkin- and MacKinnon- helmed movement against pornography grew out of an earlier period of activism across the United States which had a different conception of the nature of the problem posed by pornography and of the right way to tackle that problem. In particular, these earlier second-wave feminists saw pornography as a kind of ‘anti-female propaganda’, but rather than taking this to be a distinctive problem with pornog raphy, these feminists took it to place pornography on a spectrum with other representations of women, such as advertising and album covers. Crucially, they were against censorship or other legalistic responses of any kind, and instead advocated for education and consumer activism (Bronstein 2011).2 This earlier strand in anti-pornography feminism has the potential to be an inspiration to contemporary philosophical work on pornography too; in particular, the * I am very grateful to Catarina Dutilh Novaes, Katharine Jenkins, and Ellie Mason, as well as an audience at the Edinburgh Epistemology Research Group, for discussion and feedback. 1 Dworkin and MacKinnon’s controversial ordinances made harms due to the production or distribution of pornography actionable as a violation of civil rights, rather than calling for state censorship of a more familiar sort, though the philosophical literature tends not to make much of the differences (e.g. Langton 1993). For useful discussion, see Mikkola 2019: chapter 4. 2 The idea that propaganda is ‘anti-women propaganda’ comes up throughout Carolyn Bronstein’s history and interpretation of the second-wave anti-pornography movement. For influential published sources of the idea that pornography is misogynist propaganda see, e.g. Brownmiller (1975) and Russell and Lederer (1977); and within philosophy, see Langton (1990, 2012). Aidan McGlynn, Feminist Pornography as Feminist Propaganda, and Ideological Catch-22s In: Applied Epistemology. Edited by: Jennifer Lackey, Oxford University Press (2021). © Aidan McGlynn. DOI: 10.1093/oso/9780198833659.003.0012
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
284 AIDAN M C GLYNN emphasis on propaganda and education suggests that pornography should be just as much the concern of epistemology as it has been of speech act theory.3 This is the approach to the philosophical issues raised by pornography adopted in this chapter. The claim that pornography causes, or is itself a form of, harm to women has been fiercely contested both within philosophy and more generally, with the height of the dispute being, of course, the so-called ‘Porn Wars’, a key aspect of the ‘Sex Wars’ of the late 1970s and early 1980s. There is not even a consensus on what pornography is, with some characterisations already building in controversial claims: consider, for example, Andrea Dworkin and Catharine MacKinnon’s definition of pornography as sexually explicit material that subordinates women (e.g. MacKinnon 1987, 176).4 Here, I want to discuss the idea that pornography is not inherently problematic or harmful, and so I will need a more neutral starting point. I will take pornography to be sexually explicit materials (texts, images, videos, games, and so on) which have the primary purpose of sexually arousing their audience. This is close to what we ordinarily take ourselves to be talking about when we use the term ‘pornography’ and it does not prejudge any issues about how pornography is to be evaluated. What is typically called mainstream pornography covers the mass-produced, commercially driven, largely heterosexual pornography which is the paradigm people tend to have in mind when they think of pornography. Characterised in these terms, mainstream pornography is heterogeneous in terms of how it is produced, the kind of audience it is made for, the kind of medium it employs, the kind of content it has and how that content is presented, and so on. Still, there are a number of tendencies which explain why many people find it and the industry that makes it deeply problematic: for example, its heteronormativity; the way it presents all human sexuality in a way that caters almost exclusively for the ‘male gaze’; the way it presents desire for certain kinds of bodies (typically: white, cisgendered, non-disabled, and slim) as normal and the rest as fetishes; the way it presents acts such as anal sex and ejaculation onto the body or face as close to universally desirable to women; the way it sometimes endorses and perpetuates the myths that women are and should be always sexually available to men and that they enjoy rape; the way it eroticizes degrading and violent acts towards women; and the ways that it can exploit and harm the performers, particularly but not exclusively the women, involved in its production.5 3 I do not mean to suggest that epistemology has entirely neglected pornography; consider, for instance, Langton’s work on pornography and maker’s knowledge (collected in Langton 2009), or Fricker’s epistemological account of one way in which pornography might silence and objectify women (2007: ch. 6 but see McGlynn 2019 for criticism). 4 See Langton (1993) for the classic philosophical defence of the subordination claim, and see Jenkins (2017) for a recent alternative defence. 5 For recent content analyses of popular mainstream pornography titles that provide support for some (though not all) of these worries, see Bridges et al. (2010) and Klaassen and Peter (2015).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
FEMINIST PORNOGRAPHY AS FEMINIST PROPAGANDA 285 However, the characterisation of pornography offered above also leaves room for feminist (or more broadly, egalitarian) pornography.6 At a first approximation (to be refined shortly), this is material that is sexually explicit and intended primarily to arouse, but which is ethically produced with the aim of depicting egalitarian and manifestly consensual sexual acts. My interest in this chapter is with the idea that feminist pornography might act as a kind of feminist propaganda. Pornography appears to be a particularly powerful spreader of sexual ideology.7 However, while mainstream pornography tends to peddle a harmful, sexist sexual ideology, some feminists, including some philosophers, have expressed the hope that feminist pornography could harness pornography’s persuasive force—its propagandic power to shape the attitudes, and perhaps the behavior, of its consumers—but change the message. The idea that feminist pornography might be a particularly—perhaps even a uniquely—effective vehicle for a more egalitarian, positive view of human sexual relations gives an epistemological spin to Annie Sprinkle’s oft-quoted remark that the answer to bad porn is not to make no porn, but rather to make better porn.8 In this chapter, I will critically examine the claim that feminist pornography can play this counter-propagandic role with respect to mainstream pornography, and I will draw the disappointing conclusion that it most likely cannot. Sections 2 and 3 set up that discussion, with section 2 clarifying what feminist pornography is, for the purposes of this chapter, and section 3 examining some ways in which promoting feminist pornography, so understood, can seem like a measured, positive, and promising response to the problems associated with mainstream porn ography. Central to these, I suggest, is the proposal critically explored in this chapter, namely that feminist pornography can function as feminist propaganda in a way that makes it a particularly, or uniquely, effective counter to the propagandic power of mainstream pornography. In section 4, I argue against this proposal, focusing in particular on the two different accounts of how pornography works as propaganda that have been recently proposed by philosophers: one by A. W. Eaton and Catarina Dutilh Novaes, and one that I have explored, building on Jason Stanley’s recent account of political propaganda. I argue that each of 6 The ‘mainstream’/’feminist’ contrast is useful, but it is admittedly misleading to the extent that it suggests that feminist pornography is never produced by the mainstream pornography industry. See Taormino et al. (2013, 16) and Mikkola (2019, 209–10). 7 In McGlynn (2016), I described the harmful ideology as ‘misogynist’, but I prefer the term ‘sexist’ now in light of Kate Manne’s useful distinction between sexism and misogyny; sexism is an ideology that attempts to justify women’s lower status within patriarchal society, while misogyny is the ‘law- enforcement’ branch of patriarchy, enforcing the norms that keep women in that subordinate social position (Manne 2018, especially ch. 3). Within this scheme, my interest is with sexist pornography. Is there a sense in which pornography can be misogynistic in Manne’s sense too? Almost certainly, but I won’t pursue this thought here. (One might immediately think of so-called ‘revenge porn’ in this connection, but while undoubtedly misogynistic, this is misnamed; Claire McGlynn has proposed ‘image-based sexual abuse’ as a more appropriate label (e.g. McGlynn et al. 2017). 8 Eaton (2017, 252, fn. 6) traces this quote to a newspaper article published in 1999.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
286 AIDAN M C GLYNN these leaves us in a kind of ideological catch-22 situation: the former sees feminist pornography as having the power to reshape our sexual desires and other affect ive attitudes, but does not explain how we are to persuade consumers of mainstream pornography to switch prior to that reshaping; while the second sees the propagandic force of pornography as lying in the way it exploits and spreads preexisting sexist ideology, leaving it mysterious how feminist pornography could effect a shift in sexual ideology, even in principle. One final clarification is in order. I offer no argument against the production or promotion of feminist pornography. Nor do I intend to suggest that there are no good arguments for its production or promotion; for example, I do not dispute the potential of feminist pornography to present or construct women as sexual subjects rather than sexual objects (e.g. as argued in Ziv 2015). My target is a specific kind of epistemic interest in feminist pornography, as a counter to the sexist propaganda of much mainstream pornography, and feminist pornography can have lots of other positive roles to play, even if it cannot play that one.9
2. What is Feminist Pornography? We might suspect that distinguishing a category of feminist pornography will be a somewhat problematic task, given the notorious obstacles faced by earlier attempts to distinguish pornography from erotica. However, we should not be overly pessimistic. Some of those obstacles concern the attempt to re-appropriate the word ‘erotica’; for example, Andrea Dworkin influentially argued that: . . . in the male lexicon, which is the vocabulary of power, erotica is simply highclass pornography: better produced, better conceived, better executed, better packaged, designed for a better class of consumer. As with the call girl and the streetwalker, one is turned out better but both are produced by the same system of sexual values and both perform the same sexual service. (1981, 10–11)
The worry is that the distinction is nothing but an attempt to ‘buy oneself out of trouble, e.g. by writing class markers into the definitions of problematic practices’ (Finlayson 2016, 138). However, feminist pornography is explicitly a subcategory of pornography—there is no attempt to sweep this fact under the carpet—and the idea is precisely that it is not ‘produced by the same system of sexual values’. That said, we need to acknowledge that some of the force of Dworkin’s worry remains. Any attempt to draw morally significant distinctions within the class of pornog raphy will be fallible, controversial, and will reflect the values of those doing the
9 For further relevant discussion, see Mikkola (2019, 217–28).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
FEMINIST PORNOGRAPHY AS FEMINIST PROPAGANDA 287 distinguishing. I do not think that this makes characterising morally and epi stemically significant subcategories of pornography impossible,10 but we do need to recognise our own fallibility and be alert to the possibility of self-serving biases. A second issue concerns whether it is the intended audience of a pornographic work, how it is made, or what kinds of things it depicts that makes it feminist in the relevant sense. For example, Amalia Ziv characterises feminist pornography as that which is ‘geared towards a female audience or a mixed one, rather than towards a predominately male audience’ (2015, 3). We also have so-called ‘fairtrade’ and ‘ethical’ pornography: pornography made in ways that aim to guarantee certain standards of health and safety for those involved in its making, and to minimise the chances that they have been victims of human trafficking (Sullivan and McKee 2015, 41–3). Finally, we have pornography that aims to present a more egalitarian picture of gender and sexuality, and so tries to avoid sexist, racist, classist, ableist, fattist, transphobic, and heteronormative tropes, tries to place an emphasis on explicit consent and on female sexual agency, and so on. These different characterisations need not be seen as rivals; for example, the editors of the influential volume The Feminist Porn Book (Taormino et al. 2013) offer a characterisation of feminist pornography that includes elements of all three. Feminist pornography ‘evolved out of and incorporates elements from the genres of “porn for women,” “couples porn,” and lesbian porn’, and it is made by people who ‘emphasize the importance of their labor practices in production and their treatment of performers/sex workers; in contrast to norms in the mainstream sectors of the adult entertainment industry, they strive to create a fair, safe, eth ical, consensual work environment’. But feminist pornography also ‘uses sexually explicit imagery to contest and complicate dominant representations of gender, sexuality, race, ethnicity, class, ability, age, body type, and other identity markers’ and it ‘creates alternative images and develops its own aesthetics and iconography to expand established sexual norms and discourses’ (Taormino et al. 2013, 9–10, 15). So feminist pornography overlaps with, but is not identical to, pornography that is made for women or couples, and it is both ethically produced and has content and imagery consistent with, or in the service of, feminist aims.11 In this chapter, I am not going to be especially interested in feminist pornog raphy in the first sense—pornography created for women. That is not intended as a dismissal of the importance or philosophical interest of pornography aimed at women. But when we are considering the potential for pornography to play a positive, educational role, our focus is primarily on pornography intended for a male audience; pornography aimed in the first instance at women is not really to
10 I am disagreeing here with Ziv (2015, 14). For related views, see van Brabandt (2017) and Zheng (2017, 193). The issue demands more space than I can give it here. 11 Taormino and her co-editors stress that they want the boundaries of what counts as feminist pornography to remain open and negotiable: see (2013, 18).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
288 AIDAN M C GLYNN the point. That leaves us with the interlinked second and third senses in which pornography can be feminist: in how it is produced and in its content (compare Eaton 2017).
3. Feminist Pornography as an Escape from the ‘Porn Wars’ It is not difficult to see the appeal of the proposal that the influence of mainstream pornography should be countered with feminist pornography. There are several related points to make. First, the proposal seems to allow us to recognise that mainstream pornography is, or at least might be, having a negative effect on the attitudes, and perhaps even the behaviour, of men without this recognition pushing us towards an endorsement of anything that looks like a form of censorship, and without being labelled as ‘anti-sex’ or ‘anti-pleasure’; in other words, it promises to occupy clear, blue water between the two polar positions in the ‘Porn Wars’ (compare Taormino et al. 2013, 14–15). Likewise, we might hope to clear a path through the tangled debate over what the empirical evidence tells us about whether pornography causes harm to women.12 This remains a fiercely debated issue, and as Eaton has argued at length (2007), much of the empirical work appealed to by both sides is problematic and the debate has largely proceeded without careful pinning down of what is meant by ‘pornography’, ‘cause’, or ‘harm’.13 Still, despite this ongoing uncertainty and disagreement, there is room for legitimate concern with mainstream pornog raphy. As MacKinnon has stressed (1994, 26), there are no good grounds for accepting that such pornography does not cause harm to women; our situation with respect to pornography is arguably not like our situation with respect to vaccines and autism. Moreover, there are other relevant sources of information in addition to the kinds of empirical studies that have been the central focus of the debate, including the testimony of rapists who have discussed the role played by pornography in their crimes and the testimony of the survivors of those crimes. And, as Eaton and Lorna Finlayson have both emphasised, anti-pornography feminists have also bolstered their case with appeal to reflection on what Eaton calls the ‘intimate connection between pornography’s representational content (i.e. what is represented in pornographic works) and the nature of the criminal acts in question’ (2007, 704–5: compare Finlayson 2016, 149–50). Where Eaton
12 For three recent overviews, see Sullivan and McKee (2015, ch. 3); Tarrant (2016, ch. 6); and Finlayson (2016, ch. 9). These authors all explicitly identify as feminists, yet reach different, sometimes completely opposing, conclusions to each other; this debate remains, as it always has been, one that takes place in part within the feminist movement itself. 13 Eaton’s (2007) paper aims to rectify this situation, identifying and distinguishing a range of harms that might be relevant, as well as (plausibly, to my mind) spelling out the kind of causal claim that should be at issue in the debate.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
FEMINIST PORNOGRAPHY AS FEMINIST PROPAGANDA 289 and Finlayson differ is in their views of the significance of this kind of analysis of what mainstream pornography is like. Eaton takes it to be suggestive, but no more than that. Finlayson wants to give it much more weight, acknowledging that it fails to provide a demonstration of a causal connection between pornography and harm to women, while arguing that this is not the point. However, even if we suppose that these kinds of considerations give us only a relatively weak, defeasible reason to think that mainstream pornography is harmful to women, that might well be enough to justify taking action of a measured sort. Supporting the production of pornography that does not raise these worries and which in fact promotes healthy attitudes towards women and sexuality seems relatively measured; censorship and other legalistic responses seem to call for much more by way of evidence of harm.14 Finally, the proposal that mainstream pornography should be countered with feminist pornography seems less vulnerable to worries about overgeneralisation than censorship. Feminist anti-pornography theorists have often distanced themselves from attempts to justify anti-pornography stances that appeal to obscenity. It is what pornography does that makes it problematic, not what it represents (e.g. MacKinnon 1994, 16; Langton 1993, 38). A film could feature the same sexually explicit or violent images, and so have the same content as a work of pornog raphy, but be used for (say) educational purposes. On the face of it, there is a troubling flipside to this point. There may be all kinds of media that is not as sexually explicit as contemporary pornography, but which communicates the same kinds of messages and potentially has the same kinds of effects; we might think here of some young adult literature, pop music and music videos, and Hollywood films, to name just a few.15 So, if we take there to be a compelling case for the censorship of pornography, based on ‘what it does’, it seems like it will be difficult to avoid commitment to censorship on a much larger scale, which seems very problematic in both theory and practice. There seems to be no analogue of this overgeneralisation worry with respect to feminist pornography. These three points all offer apparent advantages of the promotion of feminist pornography over the idea that the state or legal system can play a starring role in addressing the problems associated with mainstream pornography. However, even if these advantages prove to be genuine, so far all that is on offer is a comparative claim; promoting feminist pornography is better than responses like state censorship. The final argument we will introduce in this section, which will be our focus in the rest of the chapter, fills precisely the gap left here; if successful, it offers a reason to think that feminist pornography is a particularly—perhaps even a uniquely—good response to mainstream pornography. The argument 14 Nick Novelli made a version of this point in a reading group on Eaton (2007). 15 Katharine Jenkins raised this concern in her comments on an early version of McGlynn (2016): see McGlynn 2016, 341–2, fn. 17 for discussion.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
290 AIDAN M C GLYNN starts from a diagnosis of (at least one reason) why mainstream pornography should trouble us; much of it functions as propaganda spreading a sexist sexual ideology. Let me make some preliminarily remarks on how I am understanding the notions of propaganda and ideology here, so that this thesis comes into focus. I take propaganda to be media (moving or still images, texts, and so on) or speech that acts as a spreader of ideology, where the characteristic feature of ideological attitudes is that they are relatively insensitive to evidence.16 This is roughly what Jason Stanley calls the ‘classical sense of propaganda’ (2015, 48, italics in original), and he quotes a particularly clear formulation from Randal Marlin, according to which propaganda is the ‘organized attempt through communication to affect belief or action or inculcate attitudes in a large audience in ways that circumvent or suppress an individual’s adequately informed, rational, reflective judgment’ (Marlin 2002, 22). This characterisation does seem to capture something central and essential to the notion of propaganda, but it is thin and uninformative, and so leaves a lot of room for different, richer accounts that offer a more substantive story about what propaganda and ideology are, and how they relate to each other (Stanley 2015, 50–1). In section 4, I consider two such accounts, each of which provides an interpretation of the thesis that pornography can function as propaganda, and I will offer objections to the proposal that feminist propaganda can be effective feminist propaganda on each of these interpretations.
4. Propaganda and Ideological Catch-22s The first proposal I will consider has been defended by A. W. Eaton (2007, 2017) and Catarina Dutilh Novaes (2017, 2018), and a similar idea can be found in Maes (2017).17 The core idea is that pornography can act as propaganda in the sense that it shapes its consumers’ desires and other affective responses through something like a process of Aristotelian habituation.18 Aristotle’s own concern 16 I intend to remain as neutral as possible on the mental ontology of ideology; however, I am not sure how neutral I can be. See Gendler (2008, 565–6) and Stanley (2015, 188–90) for disagreement on whether there can be ideological beliefs. 17 Mikkola has also expressed sympathy for Eaton’s views (2019: 228–30), though she regards it as ‘highly tentative and speculative’ given the absence of evidence that feminist pornography is likely to be an effective counter to mainstream pornography. Maes does not make the connection to propa ganda in his paper. In fact, Eaton does not use the term ‘propaganda’ in her published work on these issues either, though a talk at the University of Edinburgh called ‘Propaganda, Pornography, Pictures, and Persuasion’ did explicitly make the connection. I am glossing over some differences between Eaton and Dutilh Novaes here in order to present a more unified target for the discussion; I will comment on the main differences below. 18 We can see this account as a much more plausible development of the idea that pornography causes men to adopt a harmful sexual ideology through a process of primitive conditioning (as MacKinnon and Dworkin at times seem to be suggesting: see e.g. MacKinnon 1987, 181 and 1994, 11 and Dworkin 1978, 289).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
FEMINIST PORNOGRAPHY AS FEMINIST PROPAGANDA 291 was with habituation towards virtuous responses to things, the idea being that ‘the disposition to feel properly about some object is inculcated in a subject by repeatedly getting the subject to have that feeling with the right intensity toward the object’ (Eaton 2017, 250). However, there is nothing in the model itself that prevents it from providing an account of habituation in the direction of vice. On this model, pornography may well cause its audience to adopt a problematic sexual ideology, but the kind of causation at stake is probabilistic rather than deterministic: As I have argued elsewhere (2007), a causal model of pornography’s harms need not be crudely deterministic, not need it portray pornography’s audiences as unthinkingly and automatically associating stimulus with response, nor should it focus on the extreme but rarer kinds of effects at the expense of the mundane and quotidian. The causal model that I think antiporn feminists ought to adopt (a) is probabilistic, (b) holds ceteris paribus, (c) and is cumulative; that is, it insists on the importance of repeated engagement—Aristotle’s “habituation,” after all, takes time. (Eaton 2017, 251)
The bad news, on this account, is that mainstream pornography is particularly well suited to shape the affective responses and desires of its consumers in line with misogynist ideology, due to the ‘vividness’ with which it presents its content, and in particular due to the way that it ‘eroticises’ the subordination of women (Eaton 2007, 678–80 and 2017, 251–2; Dutilh Novaes 2018). The good news, the suggestion runs, is that these same features make feminist pornography a powerful way of counteracting the effects of mainstream pornography. Eaton writes: . . . this conviction in the power of representations to shape our sentimental lives cuts both ways: pornographic representations that promote sexist taste can do— and, we believe, have done—damage, but by the same token, pornographic representations promoting gender equality can do good. (Eaton 2007, 252)
Commenting on this passage, Dutilh Novaes signals her agreement: In other words, what is intrinsic to pornography is not a particular vision of gender and sexuality being promoted, but rather its power to give rise to sexual arousal [. . .] What feminist pornographers seem to be saying is this: let us appropriate this powerful tool, pornography, in order to promote a different, more egalitarian vision of gender and sexuality. (Dutilh Novaes 2018)19 19 Dutilh Novaes goes on to argue that feminist pornography is propaganda given the characterisation of propaganda due to Sheryl Tuttle Ross (2002). However, it is clear from other places in Dutilh Novaes’s paper (such as the passage just quoted) that the crucial feature of pornography that makes it
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
292 AIDAN M C GLYNN There is much in the accounts of Eaton and Dutilh Novaes that I find plausible and congenial. Unfortunately, neither spells out the positive proposal in any detail, so while it is attractive in theory, I am deeply sceptical that it can deliver on its promise in practice. The awkward question is this: what does the promotion of feminist pornography involve, such that it may function as an effective counter to mainstream pornography? Having done so much to clarify notions like ‘harm’ and ‘cause’ in this debate, Eaton is uncharacteristically vague here. In her (2007) paper, she suggests ‘encouraging the production of egalitarian pornography’ (693), and more recently she writes that feminists are committed ‘to embracing forms of pornography that serve to shape our collective erotic tastes in the direction of gender equality’ (2017, 244) and that ‘we have good reason on feminist grounds . . . to champion this new form of pornography’ (2017, 256). However, she does not say anything specific about what such embracing and championing actually involves. Dutilh Novaes is more explicit about the limitations of her argument, and so while she raises the question of whether feminist pornography is effective propaganda and ‘whether it will have the uptake its creators hope to achieve’ (2018, 1424), she does not attempt to answer them in her paper.20 The problem is that there does not seem to be a promising answer to this kind of question, given that at present most of the pornography out there and most of the pornography consumed by men is mainstream rather than feminist. The trick is to turn that situation around by in some sense promoting feminist pornography, rather than restricting access to mainstream pornography, since otherwise we are just endorsing a kind of censorship. Moreover, this has to be done in such a way as to escape the apparent catch-22 nature of the situation; crudely put, it is hard to see why men would watch feminist pornography rather than mainstream porn ography prior to the process of habituation that will, if things go well, instill a better sexual ideology in them by shaping their tastes in the direction of the former rather than the latter. Let us consider what promoting feminist pornography might involve. Some countries have experimented with publically funding the production of feminist pornography; however, such experiments have tended to be short lived, due to public backlash (Sullivan and McKee 2015, 19–20). Moreover, by itself this policy does not provide an escape from the catch-22; we would be financially supporting the production of pornography that the target audience does not particularly want to watch. A suggestion that more directly takes on the challenge of the catch-22 is to somehow ensure that feminist pornography finds its intended apt to be effective propaganda is the one highlighted by Eaton, namely the vivid and eroticized way it presents its content. Another difference between her position and Eaton’s is that Dutilh Novaes is much less wedded to the idea that Aristotelian habituation plays a significant role (personal communication), and so some of the critical points to follow may have less force against her account or may need to be reframed. 20 See also the final paragraph of Dutilh Novaes (2017).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
FEMINIST PORNOGRAPHY AS FEMINIST PROPAGANDA 293 audience; that is, we need to mandate watching it. The details of how this might work raise questions that do not seem to have easy or remotely comfortable answers. At what age should this process of habituation begin? In what kind of setting? But a more basic point is that any version is liable to raise some serious moral qualms. Moreover, Eaton stresses in the passage quoted above that habitu ation takes time and repeated exposure to the representations in question. We need to be clear, then, that we are talking about repeatedly exposing people to pornography they probably would not choose to watch of their own accord, with all that brings in its wake, practically and morally speaking. What might the promotion of feminist pornography look like, when falling short of censoring its mainstream competition or requiring exposure to it even when this is against someone’s will? The problem here is not that it is hard to think of possible answers to this question, but rather that the options do not encourage optimism that feminist pornography can act as an effective counter to mainstream pornography. For example, perhaps sufficiently old teenagers could be encouraged towards feminist pornography as an alternative to mainstream pornography as part of sex education.21 Done appropriately, this does not strike me as a bad idea as far as it goes; but it doesn’t go very far. High-quality relationship and sex education (RSE) remains spotty even in places like the United Kingdom. Nor is pornography standardly included on the RSE curriculum, though there is increasing pressure for it to be. So an immediate uphill struggle is involved in even getting anything like this on the curriculum. But suppose, counterfactually, that feminist pornography was standardly advocated to teens of an appropriate age as a standard part of RSE. There is no particular reason to expect that there would be any substantial uptake as a result of this recommendation, especially amongst those who have already had their tastes shaped by mainstream pornography. We’re still confronting the same catch-22; short of forcing people to watch feminist pornography (perhaps repeatedly), it’s not clear how making fem inist pornography more available and raising awareness is going to have much effect when the target audience are precisely those whose sexual tastes have been shaped by the ideology of mainstream pornography. There are some gaps in this reasoning, which might perhaps be exploited. First, watching feminist pornography might be incentivised without being mandated, as a way of escaping from the catch-22; people might choose to watch something that does not accord with their current sexual tastes if they are offered some other reward to do so. Second, while many people will baulk at the idea of the government funding the production or dissemination of any kind of pornography, private individuals might financially support feminist pornography as a way of ‘promoting’ it. Third, even if feminist pornography is unlikely to naturally find an 21 This is one kind of ‘soft nudge’, whereby people are encouraged towards a better option without the worse options being made unavailable (Mikkola 2019: 102–3).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
294 AIDAN M C GLYNN audience with those whose sexual tastes have already been shaped by mainstream pornography, it might play a propagandic role in shaping the tastes of those who haven’t yet absorbed the ideology of mainstream pornography; those whose engagement with mainstream pornography is the result of merely following the path of least resistance in an attempt to educate themselves about sex, perhaps, rather than those who actively seek out material that may contribute to gender inequality.22 The first of these proposals is difficult to assess in such an abstract form, and it would be worth considering what kinds of incentives (if any) might be both effective and ethical; I won’t attempt that project here. The second suggestion seem unobjectionable in principle, though I again worry that it’s not likely to have much effect in practice, since relatively few people are going to spend their money this way. The third suggestion raises trickier issues. One worry with it is that the sexist ideology pushed by mainstream pornography is continuous with (if perhaps more extreme than) the sexist ideology that people are exposed to from a very early age in a patriarchal society such as ours, and the sexual tastes of boys have already been shaped in ways that fit with the images of women promoted by mainstream pornography. That suggests that even boys who haven’t already internalised the ideology of mainstream pornography may have been already primed to find such pornography more appealing than its feminist rival.23 There’s much more that could be said, but let us consider an alternative interpretation of the claim that pornography can function as propaganda, one that I have proposed and explored in an earlier paper (McGlynn 2016). As we have seen, Eaton and Dutilh Novaes stress the idea that pornography presents inequality in an especially vivid way and in a way that eroticises it. An alternative thought lies at the heart of Jason Stanley’s recent account of propaganda and its relationship to ideology. Propaganda, on Stanley’s account, involves a contribution to public discourse that is presented as an embodiment of certain ideals, and which either tends to support those ideals ‘by emotional or other nonrational means’ (2015, 53) or tends to undermine those ideals. In principle, either supporting or undermining propaganda might be unproblematic, depending on whether the ideals being supported or challenged are worth preserving or rejecting. The characteristic feature of ideology is, as we have been assuming throughout, its relative insensitivity to evidence, though Stanley argues that this is no obstacle to there being ideological beliefs.24 This central feature of ideological beliefs does not ren22 I am grateful to Ellie Mason for raising this possibility. Dutilh Novaes (personal communication) has suggested a different role that making feminist pornography more readily available and raising awareness might play, providing an alternative for men who are already aware of the problems with mainstream pornography, or who feel it is having an adverse effect on their relationships. I don’t wish to dispute the importance of this potential role in the slightest. 23 For a different kind of worry with Eaton’s proposal, particularly (but not exclusively) as it might apply to our racial sexual preferences, see Zheng (2017, 189–90). 24 As far as I can tell, the details of Stanley’s broader account of propaganda and its relationship to ideology do not turn on this commitment, but it is hard to be sure.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
FEMINIST PORNOGRAPHY AS FEMINIST PROPAGANDA 295 der them false or otherwise problematic; for example, someone’s belief that we should be open to listening to divergent views might be so intimately connected to that person’s self-conception that it is not something they could readily revise, but this seems to be unproblematic and perhaps even epistemically beneficial (Stanley 2015, 197). The kind of propaganda that Stanley is really concerned with, though, is undermining propaganda directed at the ideals taken to be valuable in a liberal democracy, such as autonomy and equality, and this kind of problematic propaganda goes hand in hand with ‘flawed’ ideologies: Undermining propaganda is a claim that is presented as embodying a political ideal, but that is in the service of the kind of goal that tends to undermine that ideal. What this means is that the success of undermining propaganda depends on two things. First, it depends on people having beliefs that are resistant to the available evidence, the evidence that reveals the tension between goal and ideal. Secondly, since undermining propaganda conceals a contradiction of sorts, the beliefs that are resistant to evidence must themselves be flawed in some way. (2015, 178)
The ‘contradiction of sorts’ is between the ideal and the goal, and Stanley’s point is that propaganda can only be effective in a democracy when this is masked, and that this masking can only done by beliefs which are not only ideological, but have a further epistemic flaw. Flawed ideologies, in Stanley’s sense, are ones which create ‘barriers to the acquisition of knowledge’ (2015, 200). The fact that a belief is ideological is not problematic per se, but when that belief is immune to revision and promotes and sustains ignorance, then it is flawed in Stanley’s sense. Flawed ideology masks the contradiction between ideal and goal. An example will help (Stanley 2015, 68–9). Calls for strict voter ID laws present themselves as embodiments of the ideal of equality: ‘one person, one vote’. However, voter fraud levels are miniscule, and the real reason politicians call for such policies is that they politically disenfranchise certain groups: in the United States, the effect is to disenfranchise large numbers of black and Hispanic voters who tend to vote Democrat. How could one fail to notice, or see the significance of, the ‘contradiction’ here between what the ideal of equality demands and the actual and intended effect of such laws? Well, suppose one had a belief to the effect that black and Hispanic Americans are not real Americans, and so do not deserve to play a role in determining the political course of the country.25 This is clearly a belief that will sustain certain kinds of ignorance, so it is flawed in the relevant sense. And if 25 Ta-Nehisi Coates describes the ideology that Donald Trump’s birtherism tapped into as ‘that modern recasting of the old American precept that black people are not fit to be citizens of the country they built’ (2017, 342).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
296 AIDAN M C GLYNN one holds this belief, then the outcome that many of these people are stripped of their vote will seem in line with, rather than in tension with, the ideal of equality; the contradiction is masked. Most pornography is not propaganda in quite Stanley’s sense, because pornog raphy is rarely a contribution to public political discourse (McGlynn 2016, 334). Perhaps some pornography does count as such a contribution; for example, Playboy regularly features essays on political issues, such as free speech, and so if we look at the publication as a whole it arguably counts as a pornographic publication which is intended to make a contribution to public political discourse. Even if this example is accepted, it is surely the exception rather than the rule. However, the basic dynamic of propaganda in Stanley’s sense arguably is present in much (though by no means all) mainstream pornography; such pornography depicts women in positions of subordination, but as choosing and enjoying this.26 It presents itself as an embodiment of the ideals of autonomy and equality, but depicts women in ways that are inconsistent with those ideals, and does so in a way that arguably tends to undermine those ideals. For example, Deep Throat was explicitly marketed as a depiction of female sexual liberation, but it instead depicted Linda Boreman being forced to engage in throat sex against her will, and that depiction is thought to have led to other women being coerced (see e.g. MacKinnon 1987; Langton 2005; and McGlynn 2016). On Stanley’s account, the effectiveness of undermining propaganda—the way in which it can have an effect on someone’s attitudes and behaviour without engaging their ‘rational will’—is due to the way that propaganda exploits pre-existing flawed ideology in order to mask the contradiction between ideals and goals, rather than due to the vividness of the way it presents its content (though this vividness may certainly enhance its effect). This is crucially important when we think of pornography along the lines suggested by Stanley’s account of propaganda; there does not seem to be even a prima facie case that feminist pornography might act as propaganda in a way that is an effective counter to mainstream pornography. The problem can again be thought of as a kind of ideological catch-22; the kind of feminist ideology that would need to be in place for feminist propaganda to be effective is precisely what we need feminist pornography to spread. In contrast, the kind of sexist ideology exploited and spread by mainstream pornography is, unfortunately, already very widespread. There is a basic asymmetry here, which makes it very hard to see how feminist pornography could even in theory act as effective femin ist propaganda, in anything like Stanley’s sense of propaganda. This pessimism may seem unwarranted. After all, as noted above, Stanley himself explicitly argues that there can be propaganda which aims to undermine a 26 Of course, some mainstream pornography is overtly sexist and misogynist, not even maintaining the pretence that the women depicted are not being humiliated, degraded, or harmed, and such material may count as propaganda in a different sense.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
FEMINIST PORNOGRAPHY AS FEMINIST PROPAGANDA 297 problematic ideal (2015, 63–8; 2016, 288–9); he calls such propaganda a form of civic rhetoric (2016). Have I just argued that this kind of civic rhetoric, as understood by Stanley, is impossible? If so, what are we to make of the numerous examples that he offers? The details of Stanley’s examples are not always entirely clear, unfortunately. In some of his examples, he does not explicitly identify the political ideal being undermined, or detail the role of ideology in enabling that challenge (e.g. 2016, 289). In some of his other examples, it is clearer which ideal is being undermined; for example, in a discussion of John Coltrane’s take on ‘My Favourite Things’, Stanley identifies the problematic ideal as ‘a white aesthetic ideal’ (2015, 64), and in a discussion of Tommie Shelby’s work, Stanley identifies the ideal as that of ‘obedience to authority and the pursuit of legitimate work, taken unrestrictedly’ (2015, 65). However, the role of pre-existing ideology in undermining these ideals remains somewhat unclear. One proposal might be that civic rhetoric in the form of undermining propaganda exploits, for example, racist or sexist ideology in order to expose the bankruptcy of the ideals with which such ideology is associated. However, Stanley’s discussion of Shelby’s (2007) paper suggests that this is not what Stanley has in mind; rather, the idea seems to be that the target ideal is undermined once we reject an associated flawed ideology: The paper is presented as an attempt to justify a white ideology that we are in conditions of justice, and so the ideals of obedience to authority and pursuit of legitimate work can be exceptionless. However, what the paper in fact does is undermine the exceptionless ideal, by rejecting the presupposition that we are in a just state [. . .] Shelby’s paper targets the exceptionless generalization that one should be obedient to authority and pursue legitimate work. It seems at first to rely on that generalization, and pursue the question of how it is to be applied. Shelby’s discussion however eventually reveals that it is false by rejecting the idealizing presuppositions, using the example of the unjust conditions in “the dark ghetto.” (2015, 66)
What this reveals is that there is a crucial difference between some of Stanley’s cases of undermining propaganda to which he does not draw enough attention. All of the cases Stanley offers meet his characterisation of undermining propa ganda, in that they involve contributions which are presented as the embodiment of some ideal which in fact they have the goal and effect of undermining. But there is a significant difference between cases in which this disparity between the ideal and the true goal must continue to be masked and cases in which this disparity is an initial ruse, which hooks in a certain audience before the true goal is revealed. Most of Stanley’s examples of problematic undermining propaganda are of the former sort, but his description of Shelby’s project looks more like the latter.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
298 AIDAN M C GLYNN The significance of this distinction is that the role of pre-existing flawed ideology seems to be quite different in these two ways in which undermining propaganda can function. When the contradiction between ideal and goal needs to remain masked, on Stanley’s account, it is flawed ideological beliefs that enable this kind of masking, as we have seen above. However, the passage on Shelby’s paper, quoted above, suggests that the second model for undermining propaganda involves the rejection of pre-existing ideological beliefs as part of the demonstration that the ideal in question is bankrupt. Having refined our understanding of undermining propaganda by drawing this distinction, let us return to the topic of feminist pornography. Above I argued that feminist pornography cannot function as undermining propaganda in Stanley’s sense due to the dependency of such propaganda on pre-existing ideol ogy. However, that worry does not touch the kind of account of undermining propaganda as civic rhetoric which Stanley in fact offers—at least for some of his examples—since the role of pre-existing ideology is quite different in that account. So, contrary to the initial concern that opened this discussion, nothing I have said rules out the possibility of this kind of civic rhetoric, or suggests that Stanley’s examples, properly understood, are failures. Indeed, this discussion suggests another way to interpret the thesis that femin ist pornography might function as undermining propaganda; perhaps feminist pornography might work as propaganda on something like the model suggested by Stanley’s treatment of Shelby’s project. In order to make this precise, we would need to identify the ideal which feminist pornography pretends to embody as a kind of ruse. But even before we try to settle any such details, the proposal looks unpromising. It’s unpromising, first of all, as a description of feminist pornog raphy as it now exists. Producers of feminist pornography are sometimes rather uneasy at using that label, since they can worry that the ‘feminist’ label will drive away their male audience (Tarrant 2016, 167–8), but such pornography is nonetheless not presented as an embodiment of some problematic ideal: as a sheep in wolf ’s clothing, so to speak. Nor does this seem like a promising strategy for getting feminist pornography’s message out. It has been estimated that there are over ten thousand new pornographic films released each year, and so putting out a feminist pornographic film disguised as one that embodies sexist ideals seems like a fairly sure-fire way to fail to find the kind of audience one wants.
5. Concluding Remarks My starting point in this chapter has been the idea that pornography is better approached as an epistemological issue, as an issue about what makes propaganda effective and what form an education that counteracts that effect needs to take, rather than primarily an issue about freedom of speech and censorship. It is in
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
FEMINIST PORNOGRAPHY AS FEMINIST PROPAGANDA 299 this sense that this chapter (as well as much of the work it has engaged with) is more continuous with the earlier incarnations of the feminist anti-pornography movement than with the writings and activism of figures like Dworkin and MacKinnon.27 My argument, however, has sounded a rather pessimistic note. Feminist pornography promises to allow us to harness the propagandic power of pornography and repurpose it as an educational tool for equality. But whether this is even so much as possible in theory depends on what it is about pornog raphy that makes it effective propaganda. As I have argued, the prospects seem much better if this effectiveness is down to the vivid and eroticised way that porn ography presents its content than if it is due to the ubiquity of pre-existing sexist ideology. And even if feminist pornography can in principle act as feminist propaganda, it remains very difficult to see how it might effectively discharge this role in practice. Mainstream pornography plays particular roles in the lives of its audience, and it is hard to see how we can replace it in those roles with feminist pornography without censorship or some other kind of practically—and perhaps morally—problematic intervention into people’s lives. That said, in closing I want to return to a point I stressed in section 1, namely that nothing I have argued here suggests that feminist pornography cannot be a force for good. All pornography should be ethical and fair trade in its production, and there is plausibility to the idea that pornography that depicts a positive, egalitarian vision of gender relations and sex can play a valuable role in people’s lives—if they choose to let it.
References Bridges, Ana J., Robert Wosnitzer, Erica Sharrer, Chyng Sun, and Rachel Liberman (2010). ‘Agression and Sexual Behavior in Best-Selling Pornography Videos: A Content Analysis Update’. Violence Against Women 16: 1065–85. Bronstein, Carolyn (2011). Battling Pornography: The American Feminist AntiPornography Movement, 1976–1986. Cambridge: Cambridge University Press. Brownmiller, Susan (1975). Against Our Will: Men, Women and Rape. West Drayton, Middlesex: Pelican Books. Coates, Ta-Nehisi (2017). We Were Eight Years in Power: An American Tragedy. London: Hamish Hamilton. Dutilh Novaes, Catarina (2017). ‘Pornography as Propaganda: Feminist Pornography and Egalitarian Values’. University of Groningen Philosophy, Politics, and Economics blog,(March23,2017),www.rug.nl/filosofie/ppe/blog/blog-23-03-2017-pornography-aspropaganda-feminist-pornography-and-egalitarian-values.
27 It is worth stressing, though, that it’s no part of my project to dismiss Dworkin and MacKinnon’s contributions to this topic; indeed, McGlynn (2016) draws heavily on MacKinnon’s work.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
300 AIDAN M C GLYNN Dutilh Novaes, Catarina (2018). ‘Pornography, Ideology, and Propaganda’. European Journal of Philosophy 26: 1417–26. Dworkin, Andrea (1978). ‘Pornography and Grief ’, reprinted in 1980 Laura Lederer (ed.), Take Back the Night: Women on Pornography. New York: William Morrow and Company, Inc., 286–91. Dworkin, Andrea (1981). Pornography: Men Possessing Women. London: The Women’s Press. Eaton. A. W. (2007). ‘A Sensible Antiporn Feminism’. Ethics 117: 674–713. Eaton, A. W. (2017). ‘Feminist Pornography’, in Mari Mikkola (ed.), Beyond Speech: Pornography and Analytic Feminist Philosophy. Oxford: Oxford University Press, 243–57. Finlayson, Lorna (2016). An Introduction to Feminism. Cambridge; Cambridge University Press. Fricker, Miranda (2007). Epistemic Injustice: Power and the Ethics of Knowing. Oxford: Oxford University Press. Gendler, Tamar Szabó (2008). ‘Alief in Action (and Reaction)’. Mind & Language 23: 552–85. Hornsby, Jennifer (1993). ‘Speech Acts and Pornography’. Women’s Philosophy Review 10: 38–45. Jenkins, Katharine (2017). ‘What Women Are For: Pornography and Social Ontology’, in Mari Mikkola (ed.), Beyond Speech: Pornography and Analytic Feminist Philosophy. Oxford: Oxford University Press, 91–111. Klaassen, Marleen and Jochen Peter. (2015). ‘Gender (In)equality in Internet Pornography: A Content Analysis of Popular Pornographic Internet Videos.’ Journal of Sex Research 52 (7): 721–35. Langton, Rae (1990). ‘Whose Right? Ronald Dworkin, Women, and Pornographers’, reprinted in Rae Langton (2009). Sexual Solipsism: Philosophical Essays on Pornography and Objectification. Oxford: Oxford University Press, 117–64. Langton, Rae (1993). ‘Speech Acts and Unspeakable Acts’, reprinted in Rae Langton (2009). Sexual Solipsism: Philosophical Essays on Pornography and Objectification. Oxford: Oxford University Press, 25–64. Langton, Rae (2005). ‘Autonomy-Denial in Objectification’, reprinted in Rae Langton (2009). Sexual Solipsism: Philosophical Essays on Pornography and Objectification. Oxford: Oxford University Press, 221–40. Langton, Rae (2009). Sexual Solipsism: Philosophical Essays on Pornography and Objectification. Oxford: Oxford University Press. Langton, Rae (2012). ‘Beyond Belief ’, in Ishani Maitra and Mary Kate McGowan (eds), Speech and Harm: Controversies Over Free Speech. Oxford: Oxford University Press, 72–93. MacKinnon, Catharine (1987). Feminism Unmodified: Discourses on Life and Law. Cambridge, MA and London: Harvard University Press.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
FEMINIST PORNOGRAPHY AS FEMINIST PROPAGANDA 301 MacKinnon, Catharine. (1994). Only Words. Cambridge, MA and London: Harvard University Press. Maes, Hans (2017). ‘Falling in Lust: Sexiness, Feminism, and Pornography’, in Mari Mikkola (ed.), Beyond Speech: Pornography and Analytic Feminist Philosophy. Oxford: Oxford University Press, 199–220. Manne, Kate (2018). Down Girl: The Logic of Misogyny. New York: Oxford University Press. Marlin, Randal (2002). Propaganda and the Ethics of Persuasion. Peterborough, Ontario: Broadview. McGlynn, Aidan (2016). ‘Propaganda and the Authority of Pornography’. Theoria 31: 329–43. McGlynn, Aidan (2019). ‘Testimonial Injustice, Pornography, and Silencing.’ Analytic Philosophy 60: 405–17. McGlynn, Claire, Erika Rackley, and Ruth Houghton (2017). ‘Beyond “Revenge Porn”: The Continuum of Image-Based Sexual Abuse’. Feminist Legal Studies 25: 25–46. Mikkola, Mari. (2019). Pornography: A Philosophical Introduction. Oxford: Oxford University Press. Ross, Sheryl Tuttle (2002). ‘Understanding Propaganda: The Epistemic Merit Model and Its Application to Art’. Journal of Aesthetic Education 36: 16–30. Russell, Diana E. H. and Laura Lederer (1977). ‘Questions We Get Asked Most Often’, reprinted in 1980 Laura Lederer (ed.), Take Back the Night: Women on Pornography. New York: William Morrow and Company, Inc., 23–9. Shelby, Tommie (2007). ‘Justice, Deviance, and the Dark Ghetto’. Philosophy and Public Affairs 35: 126–60. Stanley, Jason (2015). How Propaganda Works. Princeton, NJ: Princeton University Press. Stanley, Jason (2016). ‘Precis of How Propaganda Works’. Theoria 31: 287–95. Sullivan, Rebecca and Alan McKee (2015). Pornography: Structures, Agency and Performance. Cambridge: Polity Press. Taormino, Tristan, Celine Parreñas Shimizu, Constance Penley, and Mireille MillerYoung (eds) (2013). The Feminist Porn Book: The Politics of Producing Pleasure. New York: The Feminist Press. Tarrant, Shira (2016). The Pornography Industry: What Everyone Needs to Know. Oxford and New York: Oxford University Press. van Brabandt, Petra (2017). ‘In/Egalitarian Pornography: A Simplistic View of Pornography’, in Mari Mikkola (ed.), Beyond Speech: Pornography and Analytic Feminist Philosophy. Oxford: Oxford University Press, 221–41. Zheng, Robin (2017). ‘Race and Pornography: The Dilemma of the (Un)Desirable’, in Mari Mikkola (ed.), Beyond Speech: Pornography and Analytic Feminist Philosophy. Oxford: Oxford University Press, 177–96. Ziv, Amalia (2015). Explicit Utopias: Rewriting the Sexual in Women’s Pornography. Albany, NY: State University of New York Press.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
PART 7
E PIST E MOLO G Y A ND SE X UA L C ONSE NT
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 19/03/21, SPi
13
Epistemic Responsibility in Sexual Coercion and Self-Defense Law Hallie Liberto
Consider Al and Betty. Al meets Betty at a bar, summons her to the back alley behind the bar, and suggests that she have sex with him. She does so—not because she wants to have sex with him, but because he has a gun in his holster the entire time, and she fears that he might use it if angered by a sexual refusal. In fact, she believes the gun to be an implicit threat—one that Al wants her to understand by displaying the gun so prominently. Both in United States law and in philosophical literature there is disagreement about whether Betty has consented to have sex with Al, given her motivating reason. Was it reasonable for Betty to believe that she was being threatened with violence? Was it reasonable for her to believe that she could not escape unharmed without having sex with Al? Even if the court deemed this sex act non-consensual, that is not enough to determine whether Al has criminally raped Betty. This hinges on whether he believed, or believed rea sonably, that Betty was choosing to have sex with him under threat.1 If Betty’s belief was deemed reasonable then the actus reus would be satisfied: she suffered non-consensual sex with Al. However, rape convictions in the United States also require a mens rea condition be met—criminal intent. Consider next Yvette and Zachery. Yvette shouts some racial slurs at Zachary at a bar, visibly angering Zachery. They later meet in a back alley. Zachery approaches Yvette with an angry face. Yvette sees that Zachery has his hands in his pockets. Based on his race, his clothing, and his demeanor, Yvette believes that Zachery is the kind of man who carries a gun (though she does not see one) and who uses it when sufficiently mad—and he looks mad now. In fear for her life, Yvette takes out her own gun and shoots Zachery to death. In many US states, Yvette can appeal to her fear as an affirmative defense if she is charged with murder or unlawful killing; she need only convince the jury that a reasonable person might have felt afraid for her life. In some states, Yvette is actually immune from 1 Note that this was only true in United States Law as of the Sexual Offense act of 2003. Before 2003, defendants only needed to convince a jury that they had a sincere belief, however unreasonable (Wortley 2013).
Hallie Liberto, Epistemic Responsibility in Sexual Coercion and Self-Defense Law In: Applied Epistemology. Edited by: Jennifer Lackey, Oxford University Press (2021). © Hallie Liberto. DOI: 10.1093/oso/9780198833659.003.0013
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
306 Hallie Liberto prosecution if she can convince a jury in a pre-trial hearing that she held a reasonable belief that Zachery might kill or gravely injure her. Was Al responsible for double-checking with Betty, to make sure that she did not feel coerced into sex due to the presence of the gun? Was Yvette responsible for waiting until she had more information before shooting Al to death? Or were their initial beliefs sufficiently reasonable to justify their actions? Due to questions like these, the criminal laws pertaining to sexual coercion and killing in self-defense are concerned with epistemic responsibility. In what follows, I make a case for how we should assign and handle the adjudication of epistemic responsibility in these two areas of criminal law. I will explain why it is problematic to try to assess whether those accused of sexual coercion and unjustified killing had reasonable beliefs, or whether they acted like reason able people. I will present an alternative suggestion by Hubin and Healey with regard to cases of sexual coercion that I will label the “reasonable expectation from state” (REfS) standard. I will argue that adopting a REfS standard for adjudicating both self-defense and sexual coercion cases is better than the “reasonable person” standard currently used. However, contra Hubin and Healey, I argue that expectations from the state toward victims of these criminal cases—expectations that ascribe epistemic responsibility to the victims—are misdirected. This is true both in cases of sexual coercion and killing in self-defense. My project focuses on a very specific form of moral responsibility that is epi stemic in nature: the moral responsibility to gather information so as to be better justified in one’s beliefs before proceeding to act. I take this to be a form of epi stemic responsibility. This is the type of responsibility that I will be referring to when I say “epistemic responsibility” in this chapter. Before I begin my argument, I want to identify some questions related to sexual coercion and killing in self-defense that I will not handle in this short chapter. I am not going to attempt to answer any question about what is consent. Consent might be a purely internal event—occurring inside a person’s head—an internal choice made that grants permission to another person.2 Alternately, it might involve an external, communicative component—like a verbal “yes,” or a set of actions that manifest willingness.3 My arguments about epistemic responsibility will not hinge on the answer to this question—though how my conclusions could be applied to particular laws will vary depending on what communicative acts might be required for consent. In the United States, criminal trials dealing with both sexual coercion and killing in self-defense try to ascertain the weight of threats—just how bad it would be for a victim if a threat was carried out. For instance, we might ask in a sexual coercion case: was the threat made by a man weighty enough to limit the woman’s 2 For examples of the internal account of consent, see Alexander (1996). Also see Hurd (1996). 3 For an example of the internal plus communicative account of consent, see Dougherty (2015).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Coercion and Self-Defense Law 307 options such that her sexual compliance did not constitute consent? If he threatened to shoot her, then courts would say: the woman’s sexual compliance was not consent. If he threatened to sing a show tune, then courts would say: the woman’s sexual compliance did count as consent. In a case of self-defense, we might also ask about the weight of the threat posed by the attacker; did the attacker pose a significant enough threat to justify killing him in self-defense? For instance, if the attacker meant to dismember you, then you would be justified in killing him. If the attacker was merely trying to forcibly pierce your ears, then it would not be justified in killing him.4 Of course, there are many hard cases—threats that are not so serious as death and dismemberment, nor so trivial as singing and ear piercing. However, I am not going to explore the seriousness of threats in this chapter. Finally, in what follows, I will usually refer to person A as male, and person B as female, when I cannot refer to them simply as A and B. I do this in order to acknowledge that the vast majority of sexual assaults in the United States, and indeed the world, are male-on-female. Much of what makes threats in the sexual arena credible threats is the consistent evidence we gather from real-life experiences and current events that remind us that men who threaten women are willing to carry through with their threats.5 Let us return to the cases that motivate the questions of this chapter. Here they are, simplified: Perceived sexual coercion: Person A perceives person B as making a threat to hurt A if A does not have sex with B. A’s belief in B’s threat motivates A to have sex with B. Perceived killing in self-defense: Person Y perceives person Z as making a threat. Y’s belief motivates Y to shoot and kill Z. We can ask questions about each person’s beliefs. Each may be morally relevant. Let us start with persons A and Y, since they are the ones whose beliefs motivate behavior that jeopardizes another person, and who stand to be defendants in criminal cases dealing with killings and sexual assault. Person A might not know that person B believes that she is under threat and that A is motivated to have sex with A only because of the perceived threat. Before the Sexual Offenses Act of 2003, person A’s false belief would have been, by itself, 4 This is assuming a proportionality criterion for using force in self-defense. For a description of this account, and its various nuances, see Feinberg (2001). 5 Scott Anderson has recently made a powerful argument that it is our cultural narratives—derived from real events—that makes man’s threat against women credible. It is this credibility that makes sexual coercion possible. He thinks that women cannot rely on men assuming that they are under threat during sexual solicitation, making sexual coercion harder and more unlikely for women to perpetrate, as it would have to be explicit (Anderson 2016).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
308 Hallie Liberto exonerating. If he sincerely held the belief that B consented, then courts could not charge him with rape, no matter how unreasonable his belief.6 However, today person A could not be exonerated on the grounds of a completely unreasonable false belief that person B consented. Legal theorists and philosophers argued that the mens rea condition could be met by recklessness or negligence, not just intent—and that it is reckless to have sex with someone and to reject the possibility that she did not consent on insufficient information.7 In court, whether person A’s belief was reasonable is decided by the jury, provided that, “in determining whether a belief was reasonable, the jury must give regard to ‘all the circumstances’.8 When it comes to cases of sexual coercion (as opposed to sexual offense cases that hinge upon incapacitation, alcohol, or age), the relevant question for the jury is this: was it reasonable for the man to believe that the woman knew that she was able to safely refuse sex? Here are some considerations that we might hope that the jury takes into account when attempting to answer this question: Did person A have a weapon that he knew to be visible to person B? (For instance, was A carrying a gun?) Was person A aware of any barriers to person B’s safe ability to leave person A’s company? (For instance, were the couple in a remote location? Was person B relying on A to drive her home from the remote location?) Did person A have any occupational or legal authority over person B that A could have reasonably recognized as a weighty barrier to B safely refusing sex with A? (For instance, was person A the prison guard, military superior, or boss of B?) We might hope that the jury takes these considerations into account if we think that person A is epistemically responsible for his false belief regarding B’s consent in light of these circumstances. Let us now consider person Y. Person Y might falsely believe that person Z is trying to kill or gravely injure her and, because of this false belief, kill person Z. If her belief is reasonable, then she cannot be convicted of murder. In many states, she cannot even be tried for murder if her belief is deemed reasonable in a pre-trial
6 See fn. 7 for examples of court cases in which sincere but unreasonable beliefs were all that was required for rape trial acquittals. 7 The cases that prompted legal theorists and philosophers to discuss the reasonableness of belief were not ones of sexual coercion, but ones of forcible rape. In Regina vs Cogan (1975) and Regina vs Morgan (1976) (both cases handled in British courts), husbands arranged for their own wives to be raped by engaging other men to forcibly have sex with them. In both cases, the husbands had told the men committing the forcible sex that their wives consented. During the forced sex, the women resisted, screamed and, in one case, screamed to her children to call the police. In both cases, the men committing the forcible sex were exonerated on the grounds that they believed that the women had consented. For accounts of these cases, see Hubin and Haely (1999). See also McGregor (2005, ch. 2). 8 This provision has been criticized for failing to clarify whether internal features of the mind of the accused—such as a mental illness—count as part of ‘all the circumstances.’ (Wortly 2013, 187–8).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Coercion and Self-Defense Law 309 hearing.9 Whether in trial or in a pre-trial hearing, the reasonability of the belief is determined by the jury. Was person Y’s belief that person Z was attempting to kill or gravely injure her a reasonable belief? Here are some considerations that we might hope that the jury takes into account when attempting to answer this question. Was person Z brandishing a weapon in the direction of person Y? Was person Z physically attacking person Y with lethal or potentially lethal force? Did person Z tell person Y that he was going to kill or gravely injure person Y, and was this verbal threat credible from the perspective of person Y? If the answer to all of these questions is “No,” then we might think that person Y was not justified in killing Z, even if Y did believe herself to be in mortal danger. Both in cases of sexual offense and in cases of killing in self-defense, a court might consider whether particular beliefs held by the victims were reasonable. For instance, in certain sexual coercion cases, judges have ruled that women could not have reasonably believed themselves to be unable to safely leave the company of the man, or unable to safely refuse sex. For this reason, they might rule that the act of sex was consensual—even if the woman sincerely believed herself to be trapped or under threat of violence. Similarly, a court might consider whether a person killed in self-defense could have reasonably known that he was being perceived as a threat to person Y, and if he could have done anything to change this perception. If so, then a jury might deem person Z to have forfeited his right to life—to have, essentially, brought about his own death. I am going to present two famous cases in order to illustrate the sort of considerations presented to juries when they are asked to evaluate the reasonableness of a person’s beliefs. One case is of alleged sexual coercion and one deals with killing in self-defense.
State vs Rusk (1981) A woman, Pat, met Mr Rusk at a bar through a mutual friend. Pat agreed to give Rusk a ride home. Pat alleged that when she stopped to let him out, Rusk invited her inside and she twice refused. Rusk reached over and turned off the ignition and took the keys from her. He told her to come upstairs. His apartment door remained unlocked and there was a phone on the table that Pat claimed that she did not see. Rusk exited the room for a few minutes, came back, undressed Pat, and had sex with her. According to Pat’s testimony: She cried; he choked her
9 In these pre-trial immunity hearings, the burden of proof is on the defendant to show that her belief was reasonable—unlike in a criminal trial, wherein the prosecution must prove beyond a rea sonable doubt that the defendant did not hold a reasonable fear. For a thorough explanation of how this works in Florida, see Alvarez (2017).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
310 Hallie Liberto lightly, and she asked, “If I have sex with you, will you let me leave without killing me?” After the sex act, Pat asked for her keys back. Rusk walked her to her car and gave back her keys, and asked if he could see her again. She drove away. Rusk claims that Pat voluntarily came up to his apartment, denies that he took her keys, and denies that she asked, “If I have sex with you, will you let me leave without killing me?” He says that she only started crying after they had had sex. Though Rusk was initially convicted of second-degree rape (coerced sex), on appeal, the decision was overturned. The court ruled that even if Pat’s claims were all true, Rusk did not use force or make threats sufficient to justify a reasonable belief in Pat that he would hurt her. Further, even if he had taken her keys, the court thought it was unreasonable that Pat would not have thought that she could scream or run to a neighboring apartment for help.10
State vs Zimmerman (2013) Zimmerman, an armed man participating in an informal “community watch,” followed an unarmed, black teenager, Trayvon Martin through his housing community. Zimmerman was in his car initially, Martin on foot. Zimmerman reports that he found Martin suspicious because he was walking slowing in the rain and looked like he might be on drugs. Zimmerman called the police’s non-emergency hotline to report Martin. The police first asked some questions that suggest that Zimmerman should continue to watch Martin (i.e. “What is his race?), but ultim ately told him not to continue pursuing Martin. However, Zimmerman continued to pursue him and got out of his car. Zimmerman testified that, while he was outside his car, Martin jumped out from behind a bush and attacked Zimmerman. According to Zimmerman, Martin got on top of him, punched him in the nose, slammed his head against the pavement, and was beginning to reach down his leg for Zimemrman’s gun. At that point, Zimmerman screamed for help and shot Martin. Because Martin is dead, he could not provide testimony. However, his family members who listened to the recording from a neighbor’s 911 call reported that the voice screaming “Help” was Martin’s, not Zimmerman’s. Also, Martin’s friend, Rachel Jeantel, who was on the phone with Martin up until the confrontation between Martin and Zimmerman began, reported that Zimmerman approached Martin. Martin knew that he was being followed and was trying to get away, fearing that Zimmerman was a “pervert.” Jeantel encouraged Martin to get away, suggesting to Martin that the adult pursuing him might be a rapist.11
10 State of Maryland vs Edward Salvatore Rusk, No. 142. September Term, 1979. Court of Appeals, Maryland. 289 Md 230; 424. A.2d 720; 1981 Md LEXIS 165, 8. 11 Alvarez and Buckley (2013). See also Jonsson (2013).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Coercion and Self-Defense Law 311 Testimony from neighbors watching from their windows differed in who was believed to have been on top of the other in the physical fight leading to the gun shot. Some reported that it was Martin, some that it was Zimmerman. The forensic expert testified that Martin was probably on top when he was shot. However, expert testimony was also given suggesting that the wounds that Zimmerman received (a bloody nose and bumped head) were minor, and not consistent with lethal force.12 In his closing remarks, Zimmerman’s attorney asked the courtroom to be silent for four minutes. He then told the jury to consider that this was the amount of time that Martin had available to him to go home, and he chose not to do so. He ended by saying, “Trayvon Martin caused his own death.” The jury acquitted Zimmerman of all charges.13 * * * In both of these cases, the victims’ reasonability was called into question. Why is this relevant? In State vs Rusk, the jury was asked to determine whether Pat had a reasonable fear that she could not safely leave Rusk’s company. They determined that she did have a reasonable fear, and the state overturned their ruling—saying that she did not. Since her fear was unreasonable, it did not vitiate her sexual consent. In State vs Zimmerman, the defense lawyer asked the jury to take into account the opportunity that Martin had to run away home—an opportunity he did not take. Even though the jury was meant to deliberate upon the reasonability of Zimmerman’s fear for his own life, the lawyer prompted the jury to also consider whether Martin had the opportunity to flee. There are many interpretations of why this might have been relevant to Zimmerman’s innocence or guilt—though the lawyer himself did not stipulate his reasons. Here is the first line of reasoning that might be interpreted from the defense lawyer’s closing remarks: If Martin had known that Zimmerman had perceived him as a lethal threat, and Martin was not a lethal threat, he would have acted in a way to try to change Zimmerman’s perception. Over the course of four minutes, Martin reasonably could have known (or: it would have been unreasonable for Martin to believe anything but the fact) that Zimmerman perceived him as a lethal threat. Hence, the fact that he did not attempt to change Zimmerman’s perception is evidence that he was a lethal threat. This evidence renders Zimmerman’s own belief reasonable, and the killing justifiable. Here is the second line of reasoning that might be interpreted from the defense lawyer’s closing remarks: Over the course of four minutes, Martin reasonably could have known (or: it would have been unreasonable for Martin to believe 12 Alvarez and Buckley (2013).
13 Ibid.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
312 Hallie Liberto anything but the fact) that Zimmerman perceived him as a lethal threat. By failing to walk away, Martin chose to remain and menace Zimmerman. This choice constituted the forfeit of Martin’s right to life, and justified Zimmerman in killing him. Either of these lines of reasoning can explain the lawyer’s final pronouncement, “Trayvon Martin caused his own death.” There might be alternate interpretations of the lawyer’s closing remarks. However, I think that these are the two explanations that could possibly be relevant to the decision that the jury was asked to make about the innocence or guilt of Zimmerman. I am not asking the reader to believe that the lawyer had one or both of these lines of reasoning in mind during his remarks. I only ask them to recognize that these are explanatory interpretations of his remarks, in that they show how the lawyer’s exercise could be relevant to the jury’s task. Both lines of reasoning require the jury to consider whether Martin could have reasonably believed, or reasonably failed to hold the belief, that he was perceived as a lethal threat to Zimmerman, and whether he should have acted differently on this basis. When it comes to sexual offence cases, feminists have long since pointed out the problems that arise when juries are asked to gauge the reasonableness of fear, or of a belief in another person’s consent. Courts are prone to use “reasonable man” standards. For instance, in State vs Rusk, it might be convincing to a court that Pat could have run away, in consideration of what a reasonable man would have done in her place. However, a male judge or a jury full of men might fail to understand how a woman might be differently situated when stranded without her car keys in an unknown neighborhood, late at night. Feminist philosophers suggest that courts appeal to a “reasonable woman” standard in determining the actus reus of criminal cases in sexual assault and coercion. If a reasonable woman would submit to sex rather than face running through an unknown neighborhood at night, then her sexual compliance cannot constitute consent. Further, a reasonable woman, unlike a man, might not assume that she should outrun her assailant. Joan McGregor points out that resisting with force is a very masculine instinct. Women are raised to submit and endure because their resistance against a man will be futile and cause them more harm. In light of this difference, a lack of physical resistance is not a form of consent.14 Hubin and Healey agree with the feminists’ concerns, but do not think that substituting a “reasonable woman” standard will help in the prosecution of rape cases. Even if this alternative standard would help establish the actus reus of sexual offenses, it would not help establish the mens rea. Even if Pat was found by the appeal court to have been the victim of non-consensual sex, this would not mean that Rusk could be convicted of rape. As long as he held a reasonable belief that Pat did consent, then he must be acquitted. While a “reasonable woman” standard could be appropriately applied to Pat, a “reasonable man” standard must be 14 McGregor (2005, 41). See also Estrich (1987, 65).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Coercion and Self-Defense Law 313 applied to Rusk—and, courts deem reasonable men to believe that women are consenting to sex with them if they do not provide a show of physical resistance.15 Hubin and Healey suggest that instead of adopting a “reasonable person” standard of any type, we instead adopt a legal standard based on: What is it rea sonable for the state to expect of a person? Is it reasonable for the state to expect someone in Rusk’s position to halt his sexual advances and ask if Pat wants to go home untouched? Remember that Pat claims to have asked Rusk, “If I have sex with you, will you let me leave without killing me?” Is it reasonable for the state to expect Rusk to respond by saying, “I’m not going to kill you whether or not you have sex with me.”16 Is it reasonable for the state to expect Pat to ask more directly, “If I don’t have sex with you, will you kill me?” Is it reasonable for the state to expect Pat to attempt to make an emergency phone call, when the man she believes to be her assailant might return at any moment? Hubin and Healey think this proposal is virtuous because it does not require members of a court of law to differentiate between features of a person’s mind and features of their circumstances (e.g. could someone’s prior experience of being raped count as a feature of their circumstance)? They also think that it avoids making any particular generalizations about what a man versus a woman would reasonably do or reasonably believe. They point out that a “reasonable woman” standard assumes that one standard should apply to all women. However, in assessing State vs Rusk, they point out: If she [Pat] were a six-foot, 200 pound, black-belt karate instructor confronted by a five-foot, 140 pound, partially paralyzed assailant, then, ceteris paribus, it might be reasonable to treat her failure to scream, call for help, leave or employ physical force to resist the assault as in indication of consent. Alternatively, if she is a five-foot, 90 pound, woman who has been previously raped, confronted by a six-foot, 200 pound, assailant, then, ceteris paribus, it seems unreasonable for the law to demand of her that she engage in these things in order not to be judged to have given consent . . . On this view, we are not ultimately interested in what a reasonable agent (woman, man or person) would do; we are, at bottom, interested in what it is reasonable for us to require (expect) individuals to do. The gender, if any, of some hypothetical reasonable agent just drops out of the picture. (Hubin and Healey 1999, 136–7)
I think that the main virtue of Hubin and Healey’s proposal is that it does not leave open the following strategy for defense teams. With reasonable person standards, a defense team need only conjure up a single, possible explanation of 15 Hubin and Healey (1999, 136). 16 Note that State vs Rusk was heard before the Sexual Offense Act of 2003, and so the reasonability of Rusk’s belief was not called into question at the time.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
314 Hallie Liberto how a reasonable man might have come to think that a woman was consenting to sex. Think how quickly a student can come up with a possible counterexample to a generally true proposition, how easily any of us could tell a made-up story of how a genuinely innocent, adult man was duped into believing that a 15-year-old girl was really 18 (an excellent, fake passport—designed just to entrap unsuspecting men into sex with a minor). Since the burden of proof is on the shoulders of the prosecution, all that a defense needs to accomplish is to convince a jury that there is some way that a reasonable man could have held a belief that the woman was consenting to sex. Curley suggests that if a man is going to claim that he had a reasonable belief in a woman’s consent, “then we must assume that he considered the possibility that she was not consenting, and rejected it.”17 Carole Pateman points out that if a man really considered this possibility, that the woman was not consenting, he certainly couldn’t reasonably reject this possibility if she were verbally refusing sex or physically resisting sex in any way.18 A virtue of Hubin and Healey’s proposal is that we might be able to say, uniformly, that the state can reasonably expect that person A consider the possibility that person B is not consenting, and only reject it if he can find no evidence for that possibility being the case. If person B is saying, “No,” then he cannot reject the possibility, and should not have sex with B. If there is a weapon present, then he cannot reject the possibility. If she expresses to him a fear that he might kill her, then he cannot reject the possibility. If he has taken away her most obvious means of safe retreat (e.g. her car keys), then he cannot reject the possibility. If someone in this situation defends himself by saying, “I didn’t think of these things,” that cannot matter. It was reasonable for the state to expect him to consider these things. Hubin and Healey are mistaken to think that the state can reasonably expect that victims of sexual coercion can reasonably be expected to fight back or flee, depending on their size and strength. Even if Pat was a 200-pound black belt in karate, the state should not demand that she physically resist sex rather than comply with Rusk’s demands, and that her compliance amounts to consent. After all, even a very strong person can still expect to be somewhat injured by a fight. Further, Pat might be unwilling to do the physical harm to Rusk that might be required to incapacitate him, such that she could flee. Pat’s unwillingness to use her considerable strength in self-defense should not be interpreted as her consent to sex. That she voiced her refusal, that she communicated her fear of being killed—the state can demand no more than this of Pat (in fact, I think it cannot 17 Curley (1976, 348). 18 See Pateman (1980). She says, “The circumstances in which a man might so deliberate, and come to an honest, but mistaken, belief about consent, are unlikely to be straightforward or simple-and surely would preclude explicit, prolonged manifestations of refusal by the woman and threatened or actual physical violence by the man” (160).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Coercion and Self-Defense Law 315 demand this much, since oftentimes fear paralyzes rape victims, and they are unable to say anything). Rusk must consider and reject the possibility that Pat is consenting, and her silence is not grounds for rejecting the possibility. It is rea sonable for the state to demand that person A does not reject the possibility that B is not consenting on the basis of B’s silence. Part of the reason why I take person B to be without much responsibility— certainly not much epistemic responsibility—is because she does not choose to act in a way that has high stakes for another person. If she complies and has sex with person A, even though her belief that person A is coercing her is completely unreasonable, person A does not suffer for it. However, if the belief is completely unreasonable, then person B might have a duty to refrain from holding person A accountable for the non-consensual sex that person B endured—at least, short of new information that better justifies her belief or fear, learned subsequently. Now, Hubin and Healey’s proposal for handling reasonability in sexual offense law can also be helpfully applied to self-defense law. After all, asking judges and juries to consider what a reasonable person would do in Zimmerman’s situation is subject to bias—perhaps not gendered bias, as is the case with sexual coercion. Instead, the reasonable person standard is subject to racial and ethnic bias when it comes to self-defense cases. John K. Roman, from the Urban Institute, analyzed FBI data and found that the race of a shooter and the race of the person being shot both make a big difference in the prosecution of defendants who use self-defense strategies in their trials. Roman was comparing states that have “stand-your-ground” laws with states that do not have such laws (note that Zimmerman’s defense team did not formally make use of this particular defense law).19 He found that there is a remarkable discrepancy in the success of convincing a jury (or judge) that a killing is justified, depending on the races of the persons involved. In states with “stand-your-ground” laws, white people shooting black people is found to be justifiable, for reasons of self-defense, 17 per cent of the time (versus 9 per cent of the time in states without “stand-your-ground” laws), whereas, in all states, black people shooting white people is found to be justifiable, for reasons of self-defense, only 1 per cent of the time.20
19 “Stand-your-ground” laws allow people to kill an aggressor instead of requiring someone who comes under attack to retreat to a safe place. These laws are controversial extensions of the age-old “Castle laws” (laws that allow you to shoot someone intruding into your home). “Stand-your-ground” laws say that not only your home, but any public place in which you have a right to be is a place from which you have no duty to retreat. The controversy surrounding these laws became part of the cultural debate during Zimmerman’s trial, even though Zimmerman’s defense team did not choose to make a case using Florida’s “stand-your-ground” law. See Alvarez and Buckley (2013) and also Jonsson (2013) for discussion of these laws during the time of Zimmerman’s trial. 20 See Roman (2013).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
316 Hallie Liberto Juries are more likely to think that a killer is justified when his target is a black person—especially a black man or boy. This is because they are more likely to agree that his belief that the other person is going to kill or gravely injure him/her is reasonable. If we were to apply a REfS standard, instead of a reasonable person or reasonable belief standard to self-defense killings, this could help us eliminate the racial component of a jury’s deliberations. We should not ask juries to consider: Did Zimmerman have a reasonable belief that Martin was going to kill him? Could a reasonable man have believed such a thing? If we only considered Martin’s youth and relatively small size, compared to Zimmerman, it would be surprising that a jury thought this belief to be reasonable. After all, Martin was not in a physical brawl with Zimmerman that involved even remotely serious injuries. It is only in consideration of Martin’s race, and American stereotypes of black men and boys, that the jury’s decision is not surprising. Recall that the defense attorney suggested that Martin had four minutes to go home, and did not, before being shot, and that his refusal to do so meant that he caused his own death. Earlier, I suggested that there were only two ways that the lawyer’s suggestion could possibly bear upon the jury’s decision to deem Zimmerman’s belief reasonable: (i) We presume that Martin, if reasonable, would have known that he was perceived as a lethal threat, and if he wasn’t a threat, then he would have done whatever he could to change Zimmerman’s perception. So, if he does not act to change this perception, in the course of four minutes, than Zimmerman’s belief is rendered reasonable. Alternately, (ii) we might say that Martin—knowing he is perceived as a threat—forfeits his right to life by choosing to remain and menace Zimmerman. Note that these accounts are both flawed. It is not clear that Martin has a responsibility to change the perception of Zimmerman. Even as a teenager, Martin has a right to walk around the neighborhood, make purchases, and linger in the rain while talking on the phone. If he is engaging in no violent or illegal behavior, then it is not his responsibility to prove to others that he is not a threat. Zimmerman’s stalking behavior was the behavior that necessitated some sort of signal, from Zimmerman to Martin, that Zimmerman was not a threat to Martin. This is true, regardless of what happened in the subsequent, four confrontational minutes. If anyone behaved in a way, at the outset, that justified the other person’s initial, violent confrontation, it is Zimmerman. Even if Martin did attack Zimmerman (using mild force—based up the forensic testimony), this attack needs to be viewed in the context of Zimmerman following Martin and instilling in Martin a reasonable fear of Zimmerman. It could well be that Zimmerman forfeited his right against being punched in the nose. If the court thinks that “going home” is the way for Martin to signal that he is not a threat to Zimmerman, then Martin is being held to a standard that is different than Zimmerman. Zimmerman had four minutes to go home—in fact, he had many more minutes after being told by the police hotline to stop pursuing
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Coercion and Self-Defense Law 317 Martin. It is this very double standard that made the case noteworthy to academics and political commentators during Zimmerman’s trial. Though Zimmerman did not formally appeal to Florida’s “stand-your-ground” laws in his defense, those laws were informally used by the defense team and by Judge Nelson, overseeing the trial. The jury was instructed to keep in mind that Zimmerman had no responsibility to retreat to a safe spot instead of shooting Martin, that he could stay and “meet force with force,”21 whereas, the defense lawyer ended his closing statements with the suggestion that Martin should have gone home and that, by failing to do so, he caused his own death. For these reasons, both of the explanations of why Martin’s failure to go home justified Zimmerman killing him are steeped in double standards—perhaps racial, perhaps age-related. The jury agreed with the defense, however. Applying the REfS standard to cases of killing in self-defense leaves less room for juries and judges to implement racial double standards. What is it reasonable for the state to demand of someone in Zimmerman’s situation? This is a question that can be asked and answered without specifying the race of the people involved. Legal codes should be established that specify what type of information gathering, warning, or endurance of minor injuries can be expected or demanded by the state. For instance, maybe it is reasonable to demand that Zimmerman (or anyone in his situation) endure a bloodied nose or a bump to the head without shooting, especially if he was the one who initially pursued and confronted his assailant. Maybe it is reasonable for the state to expect and demand that Zimmerman (or anyone in his situation) tell his assailant that he has a weapon and will use it. Maybe it is reasonable for the state to demand that I never shoot someone who is not making physical contact with me, and is not brandishing a weapon at me. If legal policy was written in this way, the juries would not have to determine the reasonability of a defendant’s beliefs, a project that might involve racial biases. Juries would simply need to determine whether someone met the demands of the state before killing in self-defense. Can the state reasonably expect or demand anything from someone in Martin’s situation? Of course: The state can reasonably expect that Martin will not attack or threaten Zimmerman, unprovoked. Of course, there needs to be policy that helps us adjudicate what it means for Martin to be provoked by Zimmerman (e.g. is stalking a provocation that justifies a punch to the nose?). However, even if someone in Martin’s position falls short of meeting the state’s reasonable expect ation, it does not follow that he has signaled a lethal threat to Zimmerman, nor that he has forfeited his right to life. For this reason, it is not clear that the answer to the question, as it pertains to Martin, will have much bearing on the answer to the question as it pertains to Zimmerman.
21 See Jonsson (2013).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
318 Hallie Liberto I want to point out that the REfS standard helps resolve racism in sexual coercion cases as well as self-defense cases. In the United States, black men are perceived as more violent, and are more likely to be viewed as threatening. For this reason, American juries have a history of being more willing to convict a black man in a rape trial, including trials in which men are accused of sexual coercion.22 There is a longstanding cultural narrative about black men as rapists—one that motivated countless lynchings of black men in the hundred years after the American Civil War.23 When there are existing cultural narratives about a particular type of person posing a threat, it is easier to view actual or imagined threats as credible.24 If both black and white men accused of rape are subject to a REfS standard for determining whether their sexual partners are consenting, then juries will only be asked to determine whether the man met those standards. For instance, perhaps the state can reasonably expect that person A will not have sex or solicit sex with a new partner while armed, lest she believe that she is under threat. If so, then this can standard can be applied equally to white and black men. The jury need only consider whether person A was armed at the time of the sex act or in the time proceeding the sexual act when he made the sexual request. The REfS standard also does some work to prevent racism against black, female victims of rape from rearing its head in rape trials. Reasonable man and reason able belief standards in sexual offense cases have failed black women. There is a persisting American attitude towards black women that they are “un-rapable” because they are in a constant state of desiring and wanting sex.25 This attitude is especially dehumanizing, not because of the way that it sexualized black women (it is not clear that sexualizing a person is a way of dehumanizing them), but because it assumes that the woman’s rights against sexual interference are con tinuously waived. This assumption produces the result that the woman lives, de facto, without sexual rights. So, when a jury is asked whether a reasonable man would have believed that a black woman consented to sex, the jury is likely to say yes—no matter what means the black woman used to resist or refuse the sexual encounter. I should note that this phenomenon bears out whether the male defendant is black or white.26 It seems that American attitudes that black women are “un-rapable” because they are continuously waiving their sexual rights is even stronger than our presumption that black men are racists. The REfS standard would do some work to remove the opportunity for racist judgments in jury deliberations. Instead of being asked what a reasonable man would believe about the woman’s consent, the jury would be asked whether the man had met the rea sonable expectations of the state to be sure of the woman’s consent—expectations that would be the same whether the victim was black or white. 22 See Bar On (1999). See also Roberts (1997) and Davis (1981). 23 See Biss (2009 [2011]). 24 See footnote 5 above. Scott Anderson does not apply his theory to racial narratives that serve to make the threat perceived by white people by black people more credible. 25 Bar On (1999, 206). See also Collins (2000 and 2005, 223). 26 See Donovan and Williams (2002). See also: Foley et al. (1995) and George and Martinez (2002).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Coercion and Self-Defense Law 319 Note that in order for the REfS standard to have the advantages that I suggest, allowing for less opportunity for racism in jury deliberations, the expectations of the state need to be outlined ahead of time, and be made known to the public. This is true both in sexual coercion cases and in self-defense cases. Hubin and Healey think that it is important that a jury consider what the state can reason ably expect of individuals, given all of the particulars of the individuals in question (e.g. the 200-pound, black belt in karate who is the victim of sexual coercion). I have already explained that I think this is problematic. Further, prescribed expectations—ones that can be applied to all people regardless of race, and be made well known to the public—can still be nuanced. For instance, perhaps if sexual partners have a long history of non-coercive sex with each other, then the state should require less pre-sexual investigation into the other’s consent. What I mean is: Consider a man whose wife has refused sex from him many times, without coming to any harm. This man can reasonably believe that her sexual agreement counts as consent, without asking her ahead of time if she knows that she can safely say “No.” Consider our traffic laws. In general, a red light means stop. Everyone knows this. However, in the United States, if everything is calm, and you can see in all directions, and there is no special sign prohibiting it, you may turn right on red. This nuance to the law is well known—and causes little trouble. In this short chapter, I have argued that reasonable person or reasonable belief standards do a poor job of handling both sexual coercion cases and killing in self-defense cases—and for similar reasons. Our attitudes about what a reasonable man or woman would have believed in a particular situation are subject to racist and sexist constructions. American juries seem incapable of thinking that a rea sonable man would do much to ascertain whether his partner is truly consenting to sex. American juries seem prone to think that reasonable men and women, when confronted with a black man or boy, will believe themselves to be in mortal peril. Applying the REfS standard to both cases better avoids these discriminatory applications. I have argued that, contra Hubin and Healey, it does not make sense to apply this standard to victims of self-defense killings and sexual coercion as a way of deliberating about the accused.
References Alexander, Larry (1996). “The Moral Magic of Consent (II).” Legal Theory 2: 165–74. Alvarez, Lizette (2017). “Florida Posed to Strengthen ‘Stand Your Ground’ Defense.” New York Times (March 15, 2017). Alvarez, Lizette and Cara Buckley (2013). “Zimmerman Is Acquitted in Trayvon Martin Killing.” New York Times (July 13, 2013). Anderson, Scott (2016). “Conceptualizing Rape as Coerced Sex.” Ethics 127 (October 2016): 50–87.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
320 Hallie Liberto Bar On, Bat-Ami (1999). “‘The Scottsboro Case’ on Responsibility, Rape, Race, Gender, and Class,” in Keith Burgess-Jackson (ed.), A Most Detestable Crime. New York: Oxford University Press, 200–210. Biss, Eula (2009). Notes from No Man’s Land: American Essays, original edn, 2009. Minneapolis, MN: Graywolf Press. Collins, Patricia Hill. (2000). Black Feminist Thought: Knowledge, Consciousness, and the Politics of Empowerment, 2nd edn. New York: Routledge. Collins, Patricia Hill. (2005). Black Sexual Politics: African Americans, Gender, and the New Racism. New York: Routledge. Curley, Edwin M. (1976). “Excusing Rape.” Philosophy and Public Affairs 5: 325–60. Davis, Angela (1981). “Rape, Racism, and the Myth of the Black Rapist,” in Women, Race, and Class. New York: Vintage Books. Donovan, Roxanne and Michelle Williams (2002). “Living at the Intersection: The Effects of Racism and Sexism on Black Rape Survivors.” Women and Therapy 25(3/4): 95–105. Dougherty, Tom (2015). “Yes Means Yes: Consent as Communication.” Philosophy and Public Affairs 43(3): 224–53. Estrich, Susan (1987). Real Rape. Cambridge, MA: Harvard University Press. Feinberg, Joel (2001). “Abortion and the Conflict of Claims,’ in George Sher (ed.), Moral Philosophy: Selected Readings, 2nd edn. Belmont, CA: Wadsworth/Thomson Learning, 735–47. Foley, Linda A. et al. (1995). “Date Rape: Effects of race of Assailant and Victim and Gender of Subjects on Perceptions.” Journal of Black Psychology 21(1): 6–18. George, William Henry and Lorraine J. Martinez (2002). “Victim Blaming in Rape: Effects of Victim and Perpetrator Race, Type of Rape, and Participant Racism.” Psychology of Women Quarterly 26: 110–19. Hubin, Donald C. and Karen Haley (1999). “Rape and the Reasonable Man.” Law and Philosophy 18(2): 113–39. Hurd, Heidi (1996). “The Magic of Consent.” Legal Theory 2: 121–46. Jonsson, Patrik (2013). “Racial Bias and ‘Stand-Your-Ground’ Laws What the Data Show.” Christian Science Monitor (August 6, 2013). McGregor, Joan (2005). Is It Rape? On Acquaintance Rape and Taking Women’s Consent Seriously. Farnham, Surrey: Ashgate Publishing Ltd. Pateman, Carole (1980). “Women and Consent.” Political Theory 8(2): 149–68. Roberts, Dorothy. (1997). Killing the Black Body: Race, Reproduction and the Meaning of Liberty. New York: Vintage Books. Roman, John K. (2013). “Race, Justifiable Homicide, and Stand Your Ground Laws: Analysis of FBI Supplementary Homicide Report.” Urban Institute, https://www. urban.org/sites/default/files/publication/23856/412873-Race-Justifiable-Homicideand-Stand-Your-Ground-Laws.PDF.pdf. Wortley, Natalie (2013). “Reasonable Relief in Consent under the Sexual Offences Act 2003.” Journal of Criminal Law 3: 184–8.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 19/03/21, SPi
14
Sexual Consent and Epistemic Agency Jennifer Lackey*
It is widely accepted that knowledge is sufficient for epistemically permissible action; that is, if someone knows a given proposition, then it is said to be epistem ically permissible for this person to act on it. To the extent that this is denied, it is argued that either more, or less, than knowledge is required, such as certainty or justified belief. In this chapter, I show that being able to act on knowledge that someone has consented to sex provides an interesting challenge to this framework. In particu lar, I argue that someone may know that another consents to sex and yet it may still be epistemically impermissible to act on this knowledge. This is clearest when the knowledge of the consent in question is secondhand rather than firsthand. When this happens, the problem is not that more, or less, than knowledge is needed to warrant action, but, rather, that a particular kind of epistemic support is required, one that involves testimony from the consentee herself. This is due to the fact that the consentee is uniquely positioned with respect to the question of her own consent, both agentially and epistemically. I further argue, however, that it doesn’t follow from this that knowing firsthand that another consents to sex is sufficient for it to be epistemically permissible to act on this knowledge. For someone might also have background beliefs, either in general or about another in particular, that function as defeaters for such action. Thus, a single instance of testimony granting consent needs to be viewed in a broader evidential framework, one where this piece of evidence alone might not be enough to warrant action on this occasion. In this way, I defend what I call the total evidence view of the epistemology of sexual consent. This shows that, con trary to what is widely argued, affirmative consent policies are too weak,1 rather than too strong. * I’m very grateful to Tom Dougherty, Thomas Grundmann, Alex Guerrero, Jonathan Jenkins Ichikawa, Lauren Leydon-Hardy, Hallie Liberto, Trevor Nyman, Isabella Reed, Baron Reed, and participants in the 2018 Summer School in Philosophy at the University of Cologne for valuable comments on an earlier draft of this chapter. 1 While there is some variation in affirmative consent policies, there is also significant convergence. According to the University of Minnesota, which is fairly representative, “Sexual assault is: 1) actual or attempted sexual contact without affirmative consent; or 2) a threat to engage in contact that would be, if the threat were carried out, sexual contact without affirmative consent. . . . Affirmative consent is freely and affirmatively communicated words or actions given by an informed individual that a sober
Jennifer Lackey, Sexual Consent and Epistemic Agency In: Applied Epistemology. Edited by: Jennifer Lackey, Oxford University Press (2021). © Jennifer Lackey. DOI: 10.1093/oso/9780198833659.003.0014
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
322 Jennifer Lackey The upshot of these considerations is that determining whether sexual consent has been given, especially in light of how high the stakes can be, requires that agents be epistemically responsible, where this can go beyond what is required in standard cases of action.
1. The Knowledge Norm of Practical Rationality In the current philosophical literature, knowledge is often taken to be the norm governing practical reasoning. For instance, Timothy Williamson maintains that the “epistemic standard of appropriateness” for practical reasoning can be stated as follows: “One knows q iff q is an appropriate premise for one’s practical reason ing” (Williamson 2005, 231).2 Similarly, according to John Hawthorne and Jason Stanley, “Where one’s choice is p-dependent, it is appropriate to treat the proposition that p as a reason for acting iff you know that p” (Hawthorne and Stanley 2008, 578).3 I have elsewhere4 called the thesis found in these passages the knowledge norm of practical reasoning, or the KNPR, and formulated it as follows: KNPR: It is epistemically appropriate for one to use the proposition that p in practical reasoning if and only if one knows that p.
reasonable person under the circumstances would believe communicate a willingness to participate in the sexual contact.” https://policy.umn.edu/hr/sexharassassault. Other universities with affirmative consent policies include the University of California, Texas A&M, University of Virginia, Indiana University, Stanford University, Yale University, University of Illinois, Urbana- Champaign, and University of New Hampshire. California (https://leginfo.legislature.ca.gov/faces/billNavClient. xhtml?bill_id=201320140SB967) and New York (www.governor.ny.gov/news/governor-cuomo-orders- comprehensive-statewide-review-compliance-enough-enough-law-protect) are states with affirmative consent standards. For instance, according to California’s Senate Bill 967: “Affirmative consent” means affirmative, conscious, and voluntary agreement to engage in sexual activity. It is the responsibility of each person involved in the sexual activity to ensure that he or she has the affirmative consent of the other or others to engage in the sexual activity. Lack of protest or resistance does not mean consent, nor does silence mean consent. End Rape on Campus describes the affirmative consent standard, and its history, as follows: In 2014, California garnered widespread attention when Governor Brown signed the nation’s first affirmative consent standard for colleges to use in campus sexual assault cases. The law established that consent is a voluntary, affirmative, conscious, agreement to engage in sexual activity, that it can be revoked at any time, that a previous relationship does not constitute consent, and that coercion or threat of force can also not be used to establish consent. Affirmative consent can be given either ver bally or nonverbally. Additionally, the law clarified that a person who is incapacitated by drugs or alcohol, or is either not awake or fully awake, is also incapable of giving consent. (http://endrapeon campus.org/yes-means-yes) 2 This is the insensitive invariantist version of this view. The contextualist version is: “A first-person present-tense ascription of ‘know’ with respect to a proposition is true in a context iff that proposition is an appropriate premise for practical reasoning in that context” (Williamson 2005, 227). 3 Hawthorne and Stanley restrict their conditions to “p-dependent choices” since p may simply be irrelevant to a given action. I will ignore this complication in what follows, as the cases discussed in this paper all involve p-dependent choices. 4 See Lackey (2010).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Consent and Epistemic Agency 323 As stated, there are two dimensions to the KNPR; one is a necessity claim and the other is a sufficiency claim. More precisely: KNPR-N: It is epistemically appropriate for one to use the proposition that p in practical reasoning only if one knows that p. KNPR-S: It is epistemically appropriate for one to use the proposition that p in practical reasoning if one knows that p. While both versions of the KNPR are of interest, my focus here will be on only the sufficiency claim.5 Such a thesis, though widely accepted, is not always expressed explicitly in terms of practical reasoning. John Hawthorne, for example, claims that “. . . we operate with a conception of deliberation according to which, if the question whether p is practically relevant, it is acceptable to use the premise that p in one’s deliberations if one knows it . . .” (Hawthorne 2004, 30). Since it is clear that the kind of acceptability Hawthorne has in mind is epistemic in nature, and deliberation about a practically relevant question is another way of talking about practical reasoning, this passage amounts to an endorsement of the KNPR-S. Other characterizations of the sufficiency claim are not cast in terms of reasoning at all, but focus instead on broader aspects of practical rationality. According to Jeremy Fantl and Matthew McGrath, “S [knows] that p only if S is rational to act as if p” (Fantl and McGrath 2002, 78).6 This is logically equivalent to S’s knowing that p being sufficient for S’s rationally acting as if p; and while acting as if p may be slightly stronger than using the proposition that p in prac tical reasoning, this thesis is certainly in the same spirit as that found in the KNPR-S. Stronger still is the claim found in the following passage from Stanley (2005): “To say that an action is only based on a belief is to criticize that action for not living up to an expected norm; to say that an action is based on knowledge is to declare that the action has met the expected norm” (Stanley 2005, 10). Here, Stanley suggests that knowledge is sufficient not only for being properly epistem ically positioned to rely on p in practical reasoning, but also for being so positioned to act on p. For the sake of clarity, I have called the broader view connecting knowledge with both practical reasoning and action the knowledge norm of practical rationality7 and formulated the sufficiency version of it as follows: 5 This condition has also been put in terms of knowledge being sufficient for one to be in a “good enough epistemic position to . . . rely on p in practical reasoning” (Brown 2012, 123). I should note, however, that Brown herself does not endorse this sufficiency claim. 6 The explicit formulation of this quotation refers to justified belief, rather than to knowledge. But immediately following this passage, Fantl and McGrath write, “. . . it might seem that we are imposing an unduly severe restriction on justification and therefore on knowledge” (Fantl and McGrath 2002, 79). It is clear, then, that Fantl and McGrath intend for this condition to apply to both justification and knowledge. 7 To keep the knowledge norm of practical reasoning and the broader knowledge norm of practical rationality distinct, I shall refer to the former as the KNPR and the latter as the KNPR*.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
324 Jennifer Lackey KNPR-S*: It is epistemically appropriate for one to use the proposition that p in practical reasoning, to act as if p, and to act on p, if one knows that p. Since the arguments that follow apply to all versions of the connection between knowledge and practical rationality thus far discussed, the KNPR-S* will be the central focus of this chapter. It should be noted that the KNPR-S* holds that knowledge is sufficient for ren dering practical reasoning and action epistemically appropriate, but there may be other senses of propriety in which knowledge fails to be so sufficient. For instance, I may clearly know that my drunk colleague is making a fool of himself at a departmental party, but it may nonetheless be inappropriate for me to act as if this is the case by, for example, confronting him about his behavior. It may be imprudent because it would strain our friendship; or it may be impolite because it would be utterly embarrassing to him; or it may simply be pointless because he won’t remember my actions the next day anyway. Thus, my confronting my friend may be inappropriate in all of these ways, while nonetheless being epistemically proper and thereby in keeping with the KNPR-S*. Now, the KNPR-S* has a great deal of intuitive appeal. If I decide to leave for the airport an hour later than was expected, my knowing that the relevant flight was delayed seems sufficient to render such a conclusion epistemically permis sible. If my choice is questioned, appealing to my knowledge adequately meets the challenge, while offering anything less—such as my suspecting that the flight is delayed, or being justified in believing that it is—does not. Moreover, the KNPR-S* has significant theoretical power. For instance, it is often noted that while it is epistemically inappropriate to rely on the proposition that one’s lottery ticket will lose in one’s practical reasoning and relevant actions if one has merely probabilistic evidence for this conclusion, it is epistemically permissible to so rely on this proposition once one has learned the results of the lottery.8 The KNPR, combined with the thesis that one possesses knowledge that one’s lottery ticket will lose in the latter, but not the former, case, can easily explain this data. Similarly, it is frequently observed that when low standards are in play within a contextualist framework, such as in DeRose’s case of self-attributing knowledge that the bank is open when it is not especially important that he deposit his pay check, it seems epistemically appropriate to rely on the proposition in question in practical reasoning and action.9 In contrast, when high standards are in play, such as when DeRose has just written a very large and important check for which he needs to ensure that adequate funds will be available in his bank account, it seems inappropriate to so rely on this proposition.10 Once again, the KNPR-S*,
8 See, for instance, Hawthorne (2004). 10 See, again, DeRose (2002).
9 See DeRose (2002).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Consent and Epistemic Agency 325 combined with the view that one can self-attribute knowledge in the former, but not the latter, case, can account for this data with ease. Finally, the KNPR-S* is said to explain what makes knowledge distinctively important or valuable. According to Fantl and McGrath, “If you know that p, then p is warranted enough to justify you in ϕ-ing, for any ϕ” (Fantl and McGrath 2009, 66). Thus, for any instance of practical reasoning or action, ϕ, one’s know ing that p is sufficient for epistemically justifying one in ϕ-ing. Such a principle, Fantl and McGrath argue, “secures the distinctive importance of knowledge” (Fantl and McGrath 2009, 182).11 In a similar spirit, Hawthorne claims that “. . . the importance of the concept of knowledge consists, in large part, in such [a] connection . . . as [that between knowledge and action]; in turn, it seems likely that any view that severs such [a connection] will be highly disruptive to our intuitive sense of the epistemic landscape” (Hawthorne 2004, 31). And Stanley maintains that rejecting the connection between knowledge and action “devalues the role of knowledge in our ordinary conceptual scheme” (Stanley 2005, 10). Despite the intuitive plausibility and theoretical power of the view that know ledge epistemically suffices for practical rationality, I have argued elsewhere that such a thesis is false. In particular, I show that it is not epistemically permissible to use the proposition that p in practical reasoning, to act as if p, or to act on p in certain cases involving what I call isolated secondhand knowledge.12 My aim in this chapter is to explore related, though importantly different, sorts of considerations against the KNPR-S*, ones involving knowledge of sexual consent. While my con clusion will be that knowledge that someone consented to sex is epistemically insufficient for the corresponding practical reasoning and action, and thus that the KNPR* is—once again—shown to be false, my ultimate goal is to shine a light on the important connections between sexual consent and epistemic agency.13
2. Knowledge of Sexual Consent To begin, let’s consider the following scenario: Secondhand Knowledge: Sally reports to a handful of her male college student friends that she consents to having sex that evening with one of their 11 More precisely, they write: “We have not of course considered all possible knowledge-free accounts of knowledge-level justification. But our discussion gives us reason to think that, at least given fallibilism, there is no such account. If, indeed, there isn’t, it looks like KJ secures the distinctive importance of knowledge” (Fantl and McGrath 2009, 182). 12 See Lackey (2011, 2016a). 13 It is worth noting that the arguments in this chapter also bear on standard views of the moral permissibility of acting on consent. For instance, Alan Wertheimer writes, “We often treat consent as a premise of a moral argument. We say that if B consents to A’s doing X to B, then it is permissible for A to do X to B” (Wertheimer 2001, 375). Given what I argue below, doubt can also be cast on a claim of this sort.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
326 Jennifer Lackey classmates, Sam. The students know Sally well, she has never withdrawn consent in the past after giving it, and they have no reason to believe that this occasion is unusual in any relevant respect, so they reliably convey this information to Sam. Given that Sam has very good reasons to believe that his classmates are trustworthy, and has no evidence to the contrary, he comes to know that Sally consents to having sex with him that evening. There are a couple of points to note about Secondhand Knowledge. First, Sam’s knowledge that Sally consents to having sex with him that evening is entirely secondhand, which means that its epistemic support is provided by the testimony of others—in this case, his classmates.14 This is to be distinguished from cases in which someone offers consent through a power of attorney, spokes person, or other proxy agent.15 Suppose, for instance, that I give power of attor ney to my husband to make decisions about my medical procedures while I am unconscious. When my husband signs the consent form to amputate my left leg in order to save my life, the surgeon’s knowledge of my consent is not secondhand in this way. That is to say, my husband is not passing on my consent through his testimony; rather, it has been delegated to my husband to make decisions of this sort for me, and so there is a sense in which the doctor’s knowledge is firsthand. In what follows, I am interested only in cases of secondhand knowledge of con sent, and so I will set aside instances in which someone has the authority to act on behalf of another. Second, giving consent to sex should clearly be distinguished from merely expressing a desire to have sex. Your saying that you wish to sell your house is different from consenting to sell your house; your expressing your desire to marry me is not the same as consenting to marry me; and your reporting your desire to have sex with someone is not thereby to consent to having sex with, say, him or her.16 What precisely is the difference? While a detailed explanation lies outside the scope of this chapter, the beginning of an answer can be found in the follow ing passage from a recent paper by Hallie Liberto on sexual consent: Ordinarily we have obligations not to intimately touch, use, or harm each other’s bodies and property. These obligations are sometimes referred to as duties to not trespass. These duties constitute weighty moral reasons due to the welfare- related significance of individuals being able to control what happens to their bodies and property. These obligations can also be described as duties that 14 See Lackey (2011). 15 See Ludwig (2014) and Lackey (2018a, 2018b) for a discussion of what Ludwig calls “proxy agency.” This type of consent might also be understood as “imputed” or “delegated” consent. For a dis cussion of some of the issues surrounding this sort of consent, see Westen (2004) and Wrigley (2007). 16 Granted, there might be cases in which consent is given through the reporting of a desire, such as your saying, “I want to have sex with you right now.” But this does not mean they are one and the same.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Consent and Epistemic Agency 327 correspond to bodily rights and property rights. When we render consent to another individual, we waive that individual’s obligation to refrain from doing one of these things. In the language of moral rights, consent operates to waive our moral rights against being intimately touched, used, or harmed by the per son to whom we have rendered consent. A particular act of sexual consent can waive some sexual rights and not others. Jones might waive her right against Smith engaging in vaginal intercourse with her, but retain her right against Smith engaging in anal intercourse with her. (Liberto 2017, 128)
As Liberto notes, consent involves the waiving of a right or,17 at the very least, the releasing of another person from a duty that he or she would otherwise have towards the one consenting. The same is not true of the mere expression of a desire. You may, for instance, say that you would like to sell your house, but you might not be able to for financial reasons, and thus you are not waiving your right to your property simply by reporting your desire. Similarly, you might say that you want to have sex with someone, but not be willing to do so because you are already in a committed, monogamous relationship. Thus, the mere expression of your desire does not release the other person from the duty to refrain from touch ing you. This distinction is important to note, as it is often conflated in cases of contested sexual consent.18
17 For a nice development of this view, see Dougherty (2015). 18 It is also worth noting that while the scenario itself in Secondhand Knowledge is fictional, some of the key aspects at issue are relevant to a variety of actual cases. For instance, Derrick Rose and two of his friends were recently accused of gang-raping a woman who reports being too intoxicated to have consented to sex. In his defense, Rose claims that a text from the woman earlier that morning is at least in part responsible for giving consent to him and his friends to have sex with her. An article from the sports channel ESPN describes the matter as follows: There was no text message that specifically gave Derrick Rose permission to have sex with his ex-girlfriend, but for the NBA star the signs were clear, he testified in a lawsuit that accuses him and two friends of raping the woman. It started with a morning text message from the woman saying Rose was the reason she ‘wakes up horny,’ including a photo of herself in her nightshirt and continued with other texts about sex throughout the day. Rose said Friday he surmised that the ‘horny’ text had triggered consent from that point on, and that was only reinforced for him by their sexual history, sex acts she engaged in with him and his per sonal assistant that night at his house and an invite hours later to her apartment. ‘I was assuming that all of us going over there that she wanted to have sex with all of us,’ Rose testified in a matter- a- fact demeanor. (https://apnews.com/article/2bb51ca0c584439fbc6d365 0bbda5971) Here, Rose claims that the text he received that morning is not only at least partly constitutive of the woman consenting to have sex with him that night, but also of her consenting to have sex with his two friends. While it is clear to me that the woman does not even defeasibly consent to having sex with anyone by virtue of sending the text in question, it might still be asked whether Rose’s reporting to his friends that the woman did consent to sex could ever be epistemically sufficient for them to act on it. Moreover, related to the second point made in the text about the distinction between giving consent to sex and merely expressing a desire to have sex, it should be clear that in the case involving Derrick Rose, the strongest conclusion that can be drawn from the text message that the woman sent earlier that morning is that she expressed a desire, but certainly not that she consented to sex, let alone
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
328 Jennifer Lackey With these points in mind, I now want to make a case for concluding that despite knowing that Sally consents to sex in Secondhand Knowledge, Sam is not appropriately epistemically positioned to act on this knowledge by, for instance, initiating sexual activity with her. The first point on behalf of this con clusion is intuition. If Sam has no other evidence about whether Sally consents to sex with him beyond the testimony of his friends, including no direct informa tion from Sally through verbal and/or non-verbal communication, it seems clearly problematic to think that he has proper epistemic grounds for having sex with her. For instance, if Sam initiates physical contact with Sally and she con fronts him about why he would think such advances are welcome, she would almost certainly not be satisfied with him responding that he knows from his friends that she consents to having sex with him. She would, instead, reasonably expect Sam to obtain consent directly from her. It might be thought that the explanation for this intuition is due entirely to an epistemic asymmetry resulting from the possibility of the withdrawal of consent. In particular, since consent can be withdrawn at any point, and since only Sally can withdraw her consent to have sex with Sam, she has an evidential advantage over anyone else regarding whether there is knowledge of her consent at a given time. Given this, the reason why we have the intuition that it is not epistemically appro priate for Sam to act on his knowledge that Sally consents to sex in Secondhand Knowledge is because Sally might have changed her mind and withdrawn her consent without Sam having access to this fact. But then we don’t have a case that clearly poses a problem for the KNPR-S*, for what might be doing the intuitive work here is doubt that Sam can really know that Sally consents to sex at T2 entirely on the basis of his friends’ testimony at T1. By way of response, I will make two points. First, people can withdraw consent with respect to many different kinds of actions, and yet we do not thereby con clude that knowledge of such consent is problematic. I might consent to having my oral surgeon remove my wisdom teeth a week before it is scheduled, and I can certainly withdraw it up until the moment I lose consciousness, but we would not say that my oral surgeon does not know that I consent to the procedure the day before it occurs. The same is true in many other cases—if I consent to your stay ing in my house over the summer, or using my car, or cutting my hair, there is no obstacle to your knowing that I so consent merely because I may have changed my mind in the last few minutes. But then in the absence of an argument that sexual consent is importantly different in this respect, the mere fact that Sally might withdraw consent should not preclude Sam from knowing that she has given it. Second, to the extent that this worry lingers, we can certainly modify Secondhand Knowledge to lessen its force. We can, for instance, imagine with multiple people. As should be clear, failure to appreciate this distinction can result in a broad range of harmful violations.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Consent and Epistemic Agency 329 that Sam learns from his friends that Sally consents to sex with him minutes before he initiates physical contact with her, and that he knows that Sally has never withdrawn sexual consent in the past. Despite this, Sally is still reasonable to expect that the epistemic basis of Sam’s belief will include her own firsthand expression of consent. It would not seem at all out of place for her to be frustrated or upset by the suggestion that the friends’ testimony, no matter how reliable and well-grounded, is epistemically adequate for Sam to initiate sexual contact with her. This shows that the mere fact that Sally has the most up-to-date information about the status of her consent cannot explain the intuition that it is not epistem ically permissible for Sam to act on his knowledge. Nevertheless, there is clearly an epistemic asymmetry between the one giving consent and everyone else, and this is where we can begin to move beyond mere intuition in our explanation of the verdict in Secondhand Knowledge. At the very least, the one consenting is surely in an epistemically privileged position rela tive to others concerning whether she consents. Suppose, for instance, that Sam says that he knows that Sally consents to sex at a given time, and she denies that this is the case. There is no question that Sally has the final word—that her testimony trumps any amount of evidence that might be produced to the contrary.19 More generally, barring unusual circumstances,20 we ought to defer to the one doing the consenting regarding whether she consents to sex, even in cases of disagreement. In this way, we might say that a consentee is an authority with respect to whether she has consented to sex. Let us now turn to getting a better grasp on how to under stand this crucial notion and the role it plays in explaining why knowledge of sexual consent cannot be entirely secondhand if it is to warrant action.
3. Authority The question I will take up in this section is what the nature is of the authority we have regarding whether we consent to sex. An obvious contender is that the authority is epistemic—specifically, that a consentee is epistemically authoritative with respect to whether she consents to sex. One way to understand this is that whatever the justification- or warrant-conferring properties are of belief, consen tees have more, or better, access to them. For instance, if reliability is central, then consentees are more reliable than others with respect to whether they consent; if evidence is key, then consentees have more, or better, evidence than others 19 Notice, of course, that it is a different question whether Sally ever consented to sex with Sam, which others might be better placed to know if, for instance, Sally’s memory is unreliable because of a recent head injury. 20 “Unusual circumstances” include, for instance, advance directives and perhaps other pre- commitment strategies, such as “Take my keys away when I’ve been drinking, even if I resist turning them over.” Thanks to Alex Guerrero for raising this point.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
330 Jennifer Lackey regarding whether they consent; and so on. Notice that this view is not at odds with what was argued in section 2, as I showed there only that consentees having an evidential advantage about whether consent has been withdrawn cannot do the relevant explanatory work in Secondhand Knowledge. The proposal here leaves open the possibility that the evidential advantage in question might go beyond this limited information about the withdrawal of consent. To explore this further, it may be helpful to consider the view defended in a recent paper by Matthew Parrott,21 where he examines the nature of the authority that we seem to have with respect to psychological self-ascriptions—such as beliefs, desires, and intentions. Just as I argued is the case with consent, Parrott claims that we ought to defer to psychological self-ascriptions, even when there is disagree ment. For instance, if I report believing that my Fiat is white and my daughter dis agrees, you should defer to me about my belief, even if you side with my daughter about the actual color of the car. Interestingly, however, Parrott denies that this “deferential trust” can be explained epistemically, and he cites two reasons for this. First, he argues that in self-ascriptions of psychological attitudes, offering add itional evidence, such as behavioral, worsens the epistemic case for deferential trust. If, for instance, I report that I believe that someone is stalking me and I offer as evidence for this that I repeatedly catch myself looking out my window at work to make sure no one is there, Parrott claims that you should have less, rather than more, reason for deferring to my psychological self-ascription. He writes: . . . it looks like the more a person explicitly appeals to behavioral evidence, the less justification we have for deferentially trusting what she says. From an epistemic perspective, this is very difficult to understand. Generally, when a person cites more evidence for an assertion, it either becomes more credible or, at worst, there is no effect on the justification for trusting what she says. (Parrott 2015, 2219)
Given that more evidence seems to result in less justification for deferential trust, Parrott concludes that the authority in question cannot be epistemic in nature. Second, he argues that “the practice of deferring to psychological self-ascriptions looks importantly different from the kind we show to epistemic authorities” in that the relevant authority “can be easily shared” in the latter case (Parrott 2015, 2220). An expert mechanic can share his knowledge of the workings of a car, and the entitlement to deference can be similarly shared. The same is not the case, however, in cases of psychological self-ascriptions. Both of these considerations are true of consent, and yet I do not think that this forces us to conclude that the authority in question is not epistemic in nature. 21 Parrott (2015).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Consent and Epistemic Agency 331 With respect to the first consideration, we can certainly imagine a case in which someone citing behavioral evidence on behalf of her consenting to sex is taken to weaken your justification for deferring to her. Suppose, for instance, that Sally reports as evidence of her consent that she drank a glass of red wine, and that she only drinks red wine when she consents to sex, so she must be consenting to sex. In such a case, it might be reasonable to take the behavioral evidence that Sally is citing as reason for less deference to her report of consenting. But I don’t see at all how this shows that the authority in question isn’t epistemic (nor do I see the argument for this in the case of psychological self-ascriptions). Parrott claims that “From a strictly epistemic standpoint, it is very hard to see why there would be any case in which the justification for deferring to someone’s assertions would be eroded by her citing additional evidence for the truth of what she said” (Parrott 2015, 2220). This, however, doesn’t seem difficult at all. More evidence— even good evidence—can result in less epistemic justification quite straightfor wardly if the evidence conflicts with the existing evidence or if, when combined, it raises doubts about the overall reliability of the source in question. I might, for instance, report to you that one (and only one) employee at the Art Institute of Chicago is planning an inside theft of a painting, and then offer additional evidence that it is security guard Paul because this is what I was reliably told. Given that you do not have any independent information about the situation, your justification for deferring to me, let’s suppose, is high. But now suppose further that I later report to you that I have additional evidence that one (and only one) employee is planning an inside theft, and offer reliable testimony that it is Peter.22 This is additional evidence for the truth that an employee is planning a theft at the Art Institute, and we can certainly stipulate that all of the testimony is reliable, but because the additional evidence that it is Peter conflicts with the earlier testimony that it is Paul (because we are assuming that it is only one security guard who is so planning), your justification for deferring to me should quite reasonably decrease. In other words, from my offering individually reliable pieces of evidence it doesn’t follow that they are jointly reliable, and so additional evidence might erode your justification for deferring to my testimony because, when taken together, they raise doubts about my overall reliability as a source on the matter. Similar considerations apply in the case of consent. Sally might report that she consents to sex, and Sam might defer to her because he reasonably interprets her as speaking from her unique, first-person position as the agent giving the con sent. When she goes on to supplement this with behavioral evidence involving her drinking a glass of red wine, there is a plausible explanation for why Sam’s justification for deferring to her could decrease with the addition of this testi mony. In particular, in light of the new evidence, Sam might wonder whether she 22 I discuss a case like this in greater detail in my (2016b), which is a modified version of one found in Goldman (2014).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
332 Jennifer Lackey in fact occupies a unique epistemic position in the first place. Compare: I report to you that I have a headache and you defer to me about this as an authority about my own pain. I then tell you that I took an Advil, so I must have a headache. If all you receive is this additional evidence, you might reasonably doubt whether I was ever reporting to you that I have a headache from the privileged position regard ing my pain states you assumed me to have, and so your justification for deferring to me might decrease. My diagnosis is that something similar explains the cases that are of interest to Parrott about self-ascriptions of psychological states. Thus, this first point fails to motivate the conclusion that the authority at work in Secondhand Knowledge cannot be epistemic in nature. Parrott’s second point, if you recall, is that epistemic authority can be shared in standard cases, yet the authority connected with psychological self-ascriptions cannot. Similar considerations apply to sexual consent; that is, my epistemic authority with respect to whether I consent to sex is not shareable in the way that knowledge about mechanics is. But it is entirely unclear to me why this is a prob lem for an epistemic reading of the authority in question in either case. I surely have a unique epistemic position vis-à-vis my pain states. If my doctor and I dis agree over whether I am in pain, you should defer to me every time. Assuming that my authoritative position here is at least partly grounded in the presence or absence of pain phenomenology, this is clearly not shareable with others. Yet why would we be at all inclined to thereby conclude that my authority is not epistemic? My suspicion is that Parrott is assuming that evidence is, by its very nature, pub licly accessible and shareable. However, this assumption should be rejected, as phenomenal states are some of the most epistemically secure sources of evidence, and yet they are in no straightforward way shareable. Neither consideration offered by Parrott, then, should lead to the conclusion that the authority in sexual consent is not epistemic. Nevertheless, I do not want to defend the view that the sense in which one has epistemic authority with respect to sexual consent is simply a matter of the con sentee having greater access to relevant evidence, much like the expert historian about World War I has more, or better, evidence than I do about the cause of the war. Instead, I want to argue that the authority one has regarding sexual consent is inextricably linked with one’s agency, and that this, in turn, gives rise to a form of epistemic authority. Interestingly, Parrott contrasts epistemic authority with what he calls agentive authority. With respect to the authority we have concerning our psychological attitudes, he writes: I am uniquely a cognitive agent with respect to . . . [my psychological] attitudes. I am the only one who is in a position to do something, namely constitute their existence or character directly on the basis of reasons. Because I am the only
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Consent and Epistemic Agency 333 one, having this capacity endows me with a special kind of agentive authority for my attitudes. (Parrott 2015, 2224, original emphasis)
According to Parrott, then, we have agentive authority regarding our psycho logical attitudes because our agency constitutes their existence on the basis of reasons. Moreover, he is clear that this kind of agency is not epistemic in nature, using an example of an umpire at a baseball game to illustrate this: The umpire is an authority for whether the pitch that has just been thrown is a strike. This is because her judgments, and only her judgments, constitute strikes. Notice that the umpire has this authority even when she is not in a better pos ition than others to know whether the pitch is a strike. The umpire cannot be ignorant but loyal fans or computers are typically in just as good of, if not a much better, epistemic position for knowing whether a pitch is a strike. Their knowledge in no way undermines the umpire’s authority precisely because it is not grounded in her comparative epistemic standing. (Parrot 2015, 2224)
While I do not wish to take on board what Parrott says about agentive authority regarding psychological attitudes, I do want to endorse something in the neigh borhood with respect to sexual consent. To begin, note that I am uniquely an agent with respect to whether I consent to sex in that I am the only one who is in a position to do something that is constitutive of its existence and character. In other words, only I am capable of exercising my agency in such a way that waives my moral right against being intimately touched. Because of this, my having this capacity endows me with a special kind of agentive authority regarding my sexual consent. Indeed, sexual consent is plausibly regarded as unlike many other kinds of consent, even ones where the stakes are very high, in the extent to which it is linked with agency. I can, for instance, grant authority to others to perform many actions on my behalf—speak, make purchases, agree to medical procedures, respond in court, and so on. But I cannot grant authority to another to consent to sex on my behalf without violating my sexual autonomy in problematic ways.23 While agentive authority is crucial to understanding sexual consent, how ever, I disagree with Parrott that this concept is independent of epistemic author ity in the way he claims. I have already rejected his arguments that would show that the authority at work in Secondhand Knowledge cannot be epistemic in nature, but here I want to briefly examine his argument above. Consider the 23 If “granting authority” to consent strikes some as too strong, this point can be reframed in terms of delegation. For instance, I can (non-objectionably) delegate to others the decision of whether I perform many actions, but I cannot (non-objectionably) delegate to another the decision of whether I will have sex. I’m grateful to a comment from Alex Guerrero that led to the inclusion of this note.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
334 Jennifer Lackey example of the umpire and the authority she has with respect to whether a pitch is a strike: according to Parrott, while the umpire has agentive authority by virtue of her judgment constituting a given strike, a local fan might be in a better epistemic position than the umpire to know whether the pitch is a strike. Hence, he con cludes, agentive authority is not epistemic in nature. Parrott’s argument here, however, is mistaken. If the umpire’s judgment alone is constitutive of a strike, then a loyal fan would be in a better epistemic position only if she had better access to the umpire’s judgments than the umpire does herself. But barring highly unusual circumstances, such as mindreading, this is simply not the case. What the loyal fan may have better access to than the umpire are the facts of the case regarding whether, given the rules of baseball, a strike occurred. But if Parrott wants to say that these facts, at least in part, determine whether a strike has occurred—in other words, if he wants to grant that the umpire could be wrong— then the umpire’s judgment is not constitutive of a strike. Thus, there is a dilemma: either the umpire’s judgment is constitutive of a pitch’s being a strike or it isn’t. If it is, then agentive authority brings with it epistemic authority since the umpire has better access to her own judgments than anyone else (barring unusual circum stances). If it isn’t, then the umpire might not be in a better epistemic position than others regarding whether a strike has occurred, but then she also doesn’t have agentive authority over this matter. Either way, Parrott fails to show that agentive authority is not at least partly epistemic in nature. Similar considerations apply with respect to sexual consent. While I have argued that consentees have agentive authority with respect to whether they con sent to sex, my view is that this brings with it a kind of epistemic authority. I am not saying that agentive authority just is epistemic authority, nor that the former is reducible to the latter. Rather, I am saying that by virtue of the exercising of one’s agency being constitutive of whether one consents to sex, one is typically better epistemically positioned than anyone else to know whether one is doing so (again, barring highly unusual cases, such as mindreading). Otherwise put, one is almost always in a uniquely privileged epistemic position with respect to the question of sexual consent at that time. Thus, consentees have both agentive and epistemic authority regarding this question. It should be noted that this claim is supported by both of the dominant views of the metaphysics of consent. According to the “attitudinal view,” morally valid consent can be given merely by adopting a certain mental attitude. Heidi Hurd, for instance, argues that the relevant mental state is the consentee’s intending that the one receiving the consent engage in the consented-to conduct.24 Peter Westen holds that the mental state is the consentee’s acquiescing to the conduct in ques tion.25 And Larry Alexander maintains that the mental state is that of the 24 See Hurd (1996). 25 See Westen (2004).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Consent and Epistemic Agency 335 consentee’s “waiving her right that the conduct not occur” (Alexander 2014, 107).26 As Dougherty notes, the intention in question on the attitudinal view “would have to be formed by a competent agent who is free from coercion and serious deception. But when such a person forms the intention, the attitudinal view will consider this necessary and sufficient for morally valid consent” (Dougherty 2015, 229). In contrast, the “performative view” of sexual consent grants that an intention is necessary for morally valid consent, but “denies that a mere intention can be sufficient, countering that valid consent also requires communication” (Dougherty 2015, 229). The performative view allows that the requisite communication can be non-verbal, perhaps expressed through a nod or even silence if, for instance, it is made explicit in the context that silence will be taken as communicating assent. “Thus, the debate between the attitudinal view and the performative view is not about whether communication must be explicit or verbal; it is about whether communication is required at all” (Dougherty 2015, 230).27 Notice that both the attitudinal and the performative views agree on the neces sity of certain mental states for morally valid consent and, moreover, that such states cannot be coerced or brought about through deception.28 Given this, such views support my claims above about the agentive and epistemic authority that consentees have regarding the question of whether they consent to sex. Most of the mental states at issue—intending, acquiescing, and waiving—involve the exercising of the consentee’s agency, especially when it is stipulated that such states cannot be forced or the result of being deceived. Moreover, standard views of self-knowledge maintain that we have privileged access to these sorts of mental states,29 thereby granting us epistemic authority with respect to them. So, even if communication is also required for morally valid consent, the consentees them selves will always be at an epistemic advantage in the normal cases, since they will have privileged access to their own mental states. 26 In so doing, Alexander claims that one forgoes “one’s moral complaint against another’s act” (Alexander 1996, 166). I should note that acquiescing and waiving might be best understood as actions that are constituted by mental states, but I will follow others, such as Alexander (2014), and characterize them as mental states in what follows. 27 Since my aim in this chapter is to make progress on the epistemology of sexual consent, I will, to the extent that it is possible, remain neutral on what consent actually is. I should note, though, that communication need not be understood on the performative view as necessarily firsthand. I can certainly successfully communicate to my daughter that we are meeting for dinner at 7:00 p.m. by telling my husband this and knowing that he will convey it to her. This is not using my husband as a spokesperson but, rather, counting on him to pass on this information to my daughter in a standard testimonial exchange. Similarly, for those drawn to the performative view of sexual consent, we can build into Secondhand Knowledge that Sally communicates her consent to her friends k nowing that they will pass it on to Sam. Given what I say about the agentive and epistemic authority that con sentees have with respect to their consenting states, however, this does not undermine the central arguments I give against the epistemic permissibility of Sam acting on this knowledge. I’m grateful to a comment by Jonathan Jenkins Ichikawa that led to the inclusion of this note. 28 For an extended discussion about the relationship between deception and consent, see Dougherty (2013). 29 See, for instance, Gertler (2011) for a discussion of the standard views of self-knowledge.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
336 Jennifer Lackey This brings us back to Secondhand Knowledge. When Sam comes to know from friends that Sally consents to sex with him, he learns something about her rather than from her. In this way, Sam bypasses Sally’s agency in acquiring knowledge of her consent. This is problematic not only from a moral point of view, but also from an epistemic one. Given that Sam’s having a false or unjusti fied belief would result in a significant harm being done to Sally, the stakes are high, and so he needs to be in a very strong epistemic position in order to act on his belief that she consents.30 But notice: it doesn’t help for Sam to just pile on more of the same kind of epistemic support on behalf of his belief that Sally con sents to sex. For instance, it won’t do for him to hear this from five or six friends, rather than four of them. This is because the problem with Sam’s knowledge is not one of quantity—where he needs more and more evidence, or greater and greater reliability of the testifiers. Rather, what he needs is an altogether different kind of epistemic support for his belief; namely, he needs firsthand evidence from Sally herself that she consents to sex.31 This is due to Sally’s being uniquely positioned with respect to this question, both agentially and epistemically, and so Sam has to learn this directly from her. Anything else, no matter how much of it there is, falls short. This point can be made clearer by considering a modified version of Secondhand Knowledge: suppose that Sam has far less support for his belief that Sally consents to sex along the quantity spectrum such that he cannot be said to know this, yet part of the basis for his belief includes Sally’s firsthand testimony that she consents.32 When we compare this case—which we can call Secondhand Knowledge*—with the original one, it seems clear that Sam is in a better epistemic position to act on his belief here. In other words, even if Sam’s belief has weaker epistemic support overall, the fact that it includes Sally’s 30 There is no need to grant a thesis as controversial as pragmatic encroachment, according to which the standards for knowledge can vary, depending on the stakes (see, for instance, Fantl and McGrath 2002 and 2009) in order to maintain that stakes have epistemic significance. For instance, it may be irrational for me to not exercise greater epistemic caution around peanuts than you do when my child has a life-threatening allergy to them, and yours does not. This doesn’t necessarily mean that given the same evidence, you know, say, that there aren’t peanuts in this slice of cake, while I do not. Rather, it might be the case that we both know that there aren’t peanuts in the cake, but—given the incredibly high stakes—I need a grade of knowledge closer to certainty to act on this knowledge, while you do not. (See Reed 2010 and 2013 for a defense of this view.) Similarly, you and I might both know that consent has been given, and while this knowledge might be sufficient for you to borrow someone’s car, I might need a grade of knowledge closer to certainty when sexual consent is concerned. 31 In fn. 30, I discussed how high stakes can require a higher grade of knowledge, such as certainty, for epistemically permissible action. What I am suggesting here is that stakes might also require knowledge of a different kind, such as firsthand, for epistemically permissible action. 32 It might be asked how Sam could fail to know that Sally consents to sex, despite her testifying directly to him that she does. There are many ways this can happen, one of which is that Sam has rea son to doubt the reliability of Sally’s testimony. Perhaps, for instance, he doesn’t know her well and she’s had a couple of drinks and so he isn’t sure that her ability to consent hasn’t been impaired. I will discuss this in further detail later in the chapter.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Consent and Epistemic Agency 337 testimony makes him better positioned epistemically for action. This is evidenced by the fact that if he were pressed on why he initiated sexual contact with Sally, his response that she told him that she consents to sex would be universally regarded as better than that his friends told him that she does. This result says something interesting about the relationship between know ledge of sexual consent and action. The debate about what is sufficient for epi stemically permissible action takes place entirely along the dimension of quantity of epistemic support. For instance, those who reject the KNPR-S* argue that it is either weaker than knowledge, such as justified belief, or stronger, such as cer tainty. If I am right here, however, then the very framing of this issue is deeply misguided, at least with respect to the question of sexual consent.33 In particular, instead of focusing on whether more or less epistemic support is needed, we need to look at the kind of support that is present. This means that the agent herself is critical to the epistemology of sexual consent, since one needs to learn from her, not about her, in order to properly act.
4. The Total Evidence View of the Epistemology of Sexual Consent In section 3, I argued that knowledge of sexual consent is insufficient for action when the knowledge fails to include firsthand evidence from the consentee her self, thereby showing that the KNPR-S* is false. But perhaps even more import antly, I showed that the agency of the consentee plays a crucial role in the epistemology of sexual consent. What I want to do in this section is to extend the focus on agency, though from a slightly different direction. Given what was argued in section 3, it might be thought that having firsthand evidence of sexual consent from the consentee herself is sufficient for epistemically permissible action. In what follows, however, I will argue that being a responsible epistemic agent vis-à-vis sexual consent requires more than this; in particular, it requires basing one’s belief about sexual consent on one’s total evidence, including relevant background information, both general and specific. This has the surprising result that, contrary to what is widely argued, affirmative consent policies are too weak, rather than too strong. To begin, consider the following: suppose that we have been friends for years and I know that it is a firmly held belief of yours that your hair should not be cut. It is a core tenet of your religion that your hair is sacred, which you have embraced since childhood, and you frequently speak about how your religious beliefs are 33 As mentioned in section 1, I argue elsewhere that there are other cases that falsify the KNPR-S*, though the considerations there are quite unlike the ones regarding sexual consent here.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
338 Jennifer Lackey central to your identity as a person. Moreover, I have never heard you express any doubts about either your religion or this belief in particular. Now suppose that one night we’re out together, you hand me a pair of scissors, and you say, expli citly, that you consent to me cutting your hair. Is it epistemically permissible for me to act on this by cutting your hair? I want to say no, and the reason for this is that your testimony that you consent to your hair being cut is one piece of evidence that needs to be considered against the background of all of the other relevant evidence. I have years and years of accumulating massive amounts of information that should make me doubt the reliability of this single instance of testimony. When viewed in its entirety, there seems to be no good epistemic reason why this one piece of evidence ought to swamp an entire body of excellent evidence that conflicts with it. Indeed, this is reminiscent of cases where a single confession is taken as definitive of guilt of a murder, even in the face of overwhelming evidence to the contrary. Of course, details might be added to both of these cases that impact the verdict in different directions. If, for instance, we’ve been drinking multiple bottles of wine together, or you have been under massive amounts of stress lately, then I have even further evidence for doubting the reliability of your testimony that you consent to a haircut. Similarly, if the confession is made by a juvenile, or under conditions of fear or sleep deprivation, then there is greater reason to question its veracity. On the other hand, if you tell me after you hand me the scissors that you have been thinking about this for many months, and you go on to report to me compelling reasons for wanting to cut your hair, then my doubts might be adequately assuaged. The same holds in the second case: if the defendant persists in maintaining the confession long after he leaves the interrogation room, and is able to explain away the conflicting evidence, then it may be epistemically appro priate to believe in the defendant’s guilt. I want to say that similar considerations apply in the case of sexual consent. We bring to each of our experiences a host of background information, both in gen eral and about the particular situation at hand. For instance, we have beliefs about typical human desires, how people respond to alcohol, what is standardly taken to be violating, and so on. We know that a great many people do not like to be humiliated or abused, do not have sex in public places, do not have sex with mul tiple partners on one night, and so on. We also know that the stakes are extremely high for getting it wrong with regard to these matters, and so we had better err on the side of caution. It would be far more harmful for us to wrongly believe that someone likes being humiliated and abused, for example, than to wrongly think that she does not. Thus, when a one-off moment radically deviates from what our previous experiences tell us, responsible epistemic agency tells us to refrain from acting until we are able to acquire further information.34 34 For a similar view, see Guerrero (2021).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Consent and Epistemic Agency 339 This is what I call the total evidence view of the epistemology of sexual consent. On this view, in order to be epistemically proper, one’s beliefs about whether sex ual consent has been given need to be based on one’s total body of relevant evi dence. The total evidence includes the testimony offered at the moment consent is purportedly given, but also background information, both about sexual consent in general and about the consentee in particular. The recent case of Brock Turner, who is the Stanford swimmer found guilty of raping a woman while she was unconscious, helps illustrate this point. In a letter the victim read aloud to Turner, she writes: Even in [Turner’s] story, I only said a total of three words, yes yes yes, before he had me half naked on the ground. Future reference, if you are confused about whether a girl can consent, see if she can speak an entire sentence. You couldn’t even do that. Just one coherent string of words. Where was the confusion? This is common sense, human decency. According to him, the only reason we were on the ground was because I fell down. Note; if a girl falls down help her get back up. If she is too drunk to even walk and falls down, do not mount her, hump her, take off her underwear, and insert your hand inside her vagina . . . You said, If I wanted to get to know her, I should have asked for her number, rather than asking her to go back to my room. I’m not mad because you didn’t ask for my number. Even if you did know me, I would not want to be in this situation. My own boyfriend knows me, but if he asked to finger me behind a dumpster, I would slap him. No girl wants to be in this situation. Nobody. I don’t care if you know their phone number or not (www.buzzfeed.com/katiejmbaker/heres-the-powerful-letter-the-stanford- victim-read-to-her-ra?utm_term=.pvMDyRrzq#.jol9kMXGR).
Despite the fact that Turner claims the victim consented to sexual activity through the three “yeses” she purportedly offered, this letter makes clear that he has back ground evidence that should have challenged this. For instance, even if one says “Yes,” this may be insufficient evidence that consent has been given if one cannot “speak an entire sentence,” is on the ground because she has fallen down, is out side in a public place behind a dumpster, and so on. Otherwise put, one’s justifica tion for believing that consent has been given might be defeated by other evidence one possesses, or should possess. There are two central kinds of defeaters that are typically taken to be incompat ible with justification. First, there are what we might call psychological defeaters, which can be either rebutting or undercutting. A psychological defeater is a doubt or belief that is had by S and indicates that S’s belief that p is either false (i.e. rebut ting) or unreliably formed or sustained (i.e. undercutting). Defeaters in this sense
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
340 Jennifer Lackey function by virtue of being had by S, regardless of their truth-value or epistemic status.35 Second, there are what we might call normative defeaters, which can also be either rebutting or undercutting. A normative defeater is a doubt or belief that S ought to have and indicates that S’s belief that p is either false (i.e. rebutting) or unreliably formed or sustained (i.e. undercutting). Defeaters in this sense func tion by virtue of being doubts or beliefs that S should have (whether or not S does have them), given the presence of certain available evidence.36 Applying these considerations to the Turner case, if the victim said “Yes” to sex, but had been drinking and could not put together a coherent sentence, or had fallen down behind a dumpster, then Turner had a defeater for his justifica tion that consent had been given. There are at least two different ways of under standing this. First, there might be a doxastic or normative rebutting defeater if he either believed, or should have believed, that these are signs that the victim is too intoxicated to give morally valid consent. This, in turn, would indicate that his belief that consent had been given is false. Second, there might be a doxastic or normative undercutting defeater if Turner either believed, or should have believed, that his belief that consent had been given is unreliably formed or sustained. Given that intoxication often negatively impacts decision-making abilities, Turner has reason to believe that the basis of his belief—namely, the victim’s testimony— might not be an epistemically secure one. Of course, defeaters themselves are candidates for defeat, so he may acquire additional evidence that changes his epistemic status. If the victim began to talk in coherent sentences, for instance, or explained that she had in fact consumed very little alcohol that evening, then the evidence to the contrary might be defeated. But in the absence of such additional evidence, it is clear that Turner’s epistemic position with respect to whether the victim consented did not warrant action. It is worth focusing on the connection between intoxication and epistemic defeat in more general terms, as it has application far beyond the single case of Brock Turner. Now, one question that might be asked is whether intoxication in fact always undermines morally valid consent. Otherwise put, we might ask whether there could be what Alan Wertheimer calls morally valid “intoxicated consent.”37 This is not the point I wish to address here. Rather, what I am con cerned with is what happens epistemically when one has reason to believe that a 35 For various views of what I call psychological defeaters see, e.g. BonJour (1980, 1985); Nozick (1981); Pollock (1986); Goldman (1986); Plantinga (1993); Lackey (2008); Bergmann (1997); and Reed (2006). 36 For discussions involving what I call normative defeaters, approached in a number of different ways, see BonJour (1980, 1985); Goldman (1986); Fricker (1987, 1994), Chisholm (1989); Burge (1993, 1997); McDowell (1994); Audi (1997, 1998); Williams (1999); Lackey (2008); BonJour and Sosa (2003); Hawthorne (2004); Reed (2006); and Goldberg (2017). What all of these discussions have in common is simply the idea that evidence can defeat knowledge (justification) even when the subject does not form any corresponding doubts or beliefs from the evidence in question. 37 See Wertheimer (2001).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Consent and Epistemic Agency 341 purported consentee is intoxicated, regardless of whether morally valid consent in fact obtains in a particular case. And here, what I want to say is that evidence that one has, or should have, that someone is intoxicated generally provides an undercutting defeater for believing that morally valid consent to sex has been given. As men tioned above, we know that intoxication can, among many other things, cause confusion, impair judgment, lower inhibitions, and lead to memory loss. People often do things while they are intoxicated that they not only would never do while sober, but also have no recollection of doing once they’re sober. Moreover, the stakes are very high when sexual consent is at issue, as sexual autonomy, and viola tions of it, are typically extremely important to people. Because of this, the very presence of intoxication on the part of an apparent consentee should elicit doubt in others about the reliability of the basis of their beliefs that consent has been given. More precisely, the grounds for believing that consent has been offered is stand ardly testimony or non-verbal behavior, but intoxication should give rise to ques tioning whether such signs are reliable indicators of morally valid consent. Since doubt about one’s basis of belief can be sufficient for functioning as an undercut ting defeater, justified belief of intoxicated consent is often defeated. What is more, consentees are often not drinking alone: those receiving the consent are often intoxicated, too. If you’ve consumed a significant amount of alcohol, then you have evidence that your ability to properly read cues is impaired, and thus that your belief that someone has consented to sexual activity, especially under unusual circumstances, such as behind a dumpster, is unreliably formed. This would bring along a further level of undercutting defeat. Now, as mentioned above, it does not follow from this that one can never have an undefeated belief that one has consented to sex while one of the parties is intoxicated. There might be additional evidence that one has that defeats the undercutting defeaters in question. If, for instance, two people have been in a relationship for quite some time and have learned how to read one another’s behavior, especially while drinking, then the doubt elicited by the intoxication might be defeated by evidence that their respective judgments are reliable on this occasion. In addition to evidence at the general level, there might also be information that is unique to the situation that conflicts with consent. Just as in the haircutting example, if someone has ample evidence that one’s partner would not be consent ing to sex under the present circumstances, then positive cues conveyed at the time may be epistemically inadequate to warrant action. Let’s just imagine a vari ant of the case above: suppose that A knows that it is a firmly held belief of B’s that B should wait until marriage to have sex. It is a core tenet of B’s religion that sex should take place only in the context of a marriage, which B has embraced since childhood, and B frequently speaks about how her religious beliefs are central to her identity as a person. Moreover, A has never heard B express any doubts about either her religion or this belief in particular. Now, suppose that, one night, A and
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
342 Jennifer Lackey B are out together and B says, explicitly, that she consents to having sex with A. Is it epistemically permissible for A to act on this? As with the haircutting example, my answer is no, and it is even a firmer no, given that the stakes are often higher regarding unwanted sex than they are for unwanted haircuts. In this case, A has massive amounts of evidence against B consenting to sex outside of the context of marriage and only one piece of testi mony that conflicts with this. A’s total body of evidence, then, clearly tells against consent. As I said above, new information can be acquired that changes A’s epi stemic status, such as B telling A that she has been reassessing her view about sex before marriage for months, or that she recently abandoned her religion. But in the absence of this, acting on the belief that consent has been given, even when based on hearing that this is so firsthand from the consentee, is unwarranted.38 A question that might be asked here, however, is whether refusing to consent to sex is similarly one piece of evidence to be weighed against all of the other evi dence one has on the matter. If so, does it follow that despite someone explicitly saying no to sex, it might be epistemically proper for one to believe that she actu ally does consent, given all of the relevant background information? By way of response, notice that while my total evidence view is committed to saying that a refusal of sex is indeed one piece of evidence, it doesn’t follow from this that dif ferent pieces of evidence cannot be weighed more, or less, heavily. Testimony from both an expert and a novice regarding whether chemotherapy versus sur gery is the best treatment option for a particular kind of cancer, for instance, are quite different evidentially. While they are both pieces of evidence, the one from an oncologist utterly swamps that of, say, a concerned neighbor. In the same way, even though there is a sense in which a “no” to sex is one piece of a total body of evidence, it, too, utterly swamps all of the other evidence from an epistemic point of view. Given this, there are no concerns that, in ordinary cases, someone’s refusal of sex might be compatible with another’s total evidence epistemically jus tifying belief that sexual consent has been given. The broader point that should be emphasized here is that one needs to be a responsible epistemic agent in all matters, but perhaps even especially when sexual consent is at issue, and this might at times require more than affirmative consent. Just as I need to weigh new evidence with what I currently have in my possession when believing that it will rain today or that my student did not commit the crime for which he is convicted, so, too, the total evidence view says that we need to take into account all of the relevant evidence when believing that someone has con sented to sex. If, for instance, a woman has been drinking heavily and seems to be
38 This point might also be supported by a general moral caution principle of the sort defended in Guerrero (2007).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Consent and Epistemic Agency 343 consenting to sex in a bar with all of her friends watching, this one instance of apparent consent should be weighed with all of the background evidence avail able that most people would never have sex under such circumstances.39 Overall, then, responsible epistemic agency requires refraining from action in such a case. This raises an interesting question about the relationship between epistemic permissibility and moral permissibility. Suppose in the example above, B does, in fact, consent to having sex, but A’s overall evidence doesn’t support A’s believing this with justification. If B consented to sex, and thus waived her right to not be touched by A, doesn’t it follow that it is morally permissible for A to act accord ingly, despite the fact that this isn’t epistemically permissible?40 If so, it looks like there can be a significant gap between epistemic and moral evaluation when sex ual consent is concerned. By way of response, notice that the mere fact that consent is given doesn’t mean that it is not morally blameworthy to act on it for other reasons. Suppose, for instance, that unbeknownst to you, I sign a document giving you ownership of my Fiat, which I then drop in the mail. Before you receive the document or even know about its existence, you come over to my house and drive off with my Fiat. From your perspective, you are clearly stealing my car and are thus morally blameworthy for theft, despite the fact that I had already consented to your taking it. For instance, if I learn that you took my car in a state of ignorance of this sort, my assessment of your moral character would surely be negatively impacted: I wouldn’t want to be your friend, I wouldn’t trust you with my belongings, and so on. Thus, the wrong-making feature of your action is not my lack of consent, but your lack of regard for me and my belongings. I would say something similar in cases where sexual consent is given, but one is not epistemically positioned to know that it has been. Acting on this is morally permissible in the same way that driving off with my car in the above case is permissible—strictly speaking, con sent has been given, but moral assessment goes beyond the violation of rights. If A knows that B’s deepest values have always involved abstinence until marriage, for instance, then even if momentary consent is given, A might be morally blame worthy for not being sensitive to B and her well-being. Sure, A might point to the “Yes” that was given at the time to prove that B’s rights weren’t violated, but B might expect far more than this from a loving, respectful partner. Thus, the epi stemic and overall moral evaluation of the person and the relevant action involve
39 See, for instance, www.nytimes.com/2014/07/13/us/how-one-college-handled-a-sexual-assault- complaint.html. 40 Recall, for instance, Wertheimer’s claim: “We often treat consent as a premise of a moral argu ment. We say that if B consents to A’s doing X to B, then it is permissible for A to do X to B” (Wertheimer 2001, 375).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
344 Jennifer Lackey significant convergence, and thus such a case does not turn out to be one where epistemic and moral permissibility come apart. One aspect of the conclusion from this section that is particularly surprising is that when affirmative consent policies are criticized,41 they are almost universally regarded as too strong. They are said to turn ordinary human interactions into sexual offenses,42 to “ruin” good men,43 to deny due process to the accused, and so on.44 If I’m right, however, such policies not only often drop out of the most basic requirements of responsible epistemic agency, but they are also too weak, as affirmative consent is only one piece of evidence.45 Thus, on the total evidence view, “Yes” should not always be taken to be epistemically sufficient for action.
5. Conclusion We have seen that applying epistemological tools to questions about sexual con sent can shed light on a range of phenomena. In particular, the unique agential and epistemic position that consentees occupy with respect to sexual consent shows that knowledge, particularly when it is secondhand, is not sufficient for epistemically permissible action. We have also seen, however, that firsthand con sent is also not sufficient, as the testimony of consentees is one piece of evidence that needs to be weighed with all of one’s other relevant evidence about the mat ter. The total evidence view of the epistemology of sexual consent thus requires that we be responsible epistemic agents, and base our beliefs regarding whether such consent has been given on our total body of relevant evidence. This has important implications, not only for affirmative consent policies that are being hotly debated, especially on college campuses, but also for some of the most sig nificant moral questions facing us today.
References Alexander, Larry (1996). “The Moral Magic of Consent II.” Legal Theory 2: 165–74. Alexander, Larry (2014). “The Ontology of Consent.” Analytic Philosophy 55: 102–13.
41 See fn. 1. 42 Seewww.washingtonpost.com/posteverything/wp/2015/05/20/feminists-want-us-to-define-theseugly-sexual-encounters-as-rape-dont-let-them/?utm_term=.9297b8f3bf34. 43 See http://theweek.com/articles/443335/big-problem-californias-new-sexual-consent-law. 44 See http://reason.com/archives/2015/02/03/how-affirmative-consent-rules-threaten. 45 This, of course, assumes that policies regarding sexual consent should require responsible epi stemic agency. While this strikes me as quite plausible, a defense of this lies beyond the scope of the present chapter.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Consent and Epistemic Agency 345 Audi, Robert (1997). “The Place of Testimony in the Fabric of Knowledge and Justification.” American Philosophical Quarterly 34: 405–22. Audi, Robert (1998). Epistemology: A Contemporary Introduction to the Theory of Knowledge. London: Routledge. Baker, Katie J. M. (2016). “Here’s the Powerful Letter the Stanford Victim Read to Her Attacker.” BuzzFeed News (June 3, 2016), www.buzzfeed.com/katiejmbaker/heresthe-powerful-letter-the-stanford-victim-read-to-her-ra?utm_term=.pvMDyRrzq#. jol9kMXGR. Bergmann, Michael (1997). “Internalism, Externalism and the No-Defeater Condition.” Synthese 110: 399–417. BonJour, Laurence (1980). “Externalist Theories of Epistemic Justification.” Midwest Studies in Philosophy 5: 53–73. BonJour, Laurence (1985). The Structure of Empirical Knowledge. Cambridge, MA: Harvard University Press. BonJour, Laurence and Ernest Sosa (2003). Epistemic Justification: Internalism vs. Externalism, Foundations vs. Virtues. Oxford: Blackwell Publishing. Brown, Jessica (2012). “Assertion and Practical Reasoning: Common or Divergent Epistemic Standards?” Philosophy and Phenomenological Research 84: 123–57. Burge, Tyler (1993). “Content Preservation.” The Philosophical Review 102: 457–88. Burge, Tyler (1997). “Interlocution, Perception, and Memory.” Philosophical Studies 86: 21–47. Chisholm, Roderick M. (1989). Theory of Knowledge, 3rd edn. Englewood Cliffs, NJ: Prentice-Hall. DeRose, Keith (2002). “Assertion, Knowledge, and Context.” The Philosophical Review 111: 167–203. Dougherty, Tom (2013). “Sex, Lies, and Consent.” Ethics 123: 717–44. Dougherty, Tom (2015). “Yes Means Yes: Consent as Communication.” Philosophy and Public Affairs 43: 224–53. End Rape on Campus (n.d.). “ ‘Yes’ Means ‘Yes’ and Affirmative Consent.” http:// endrapeoncampus.org/yes-means-yes. Fantl, Jeremy and Matthew McGrath (2002). “Evidence, Pragmatics, and Justification.” The Philosophical Review 111: 67–94. Fantl, Jeremy and Matthew McGrath (2009). Knowledge in an Uncertain World. Oxford: Oxford University Press. Fricker, Elizabeth (1994). “Against Gullibility,” in Bimal Krishna Matilal and Arindam Chakrabarti (eds), Knowing from Words. Dordrecht: Kluwer Academic Publishers, 125–61. Fricker, Elizabeth (1987). “The Epistemology of Testimony.” Proceedings of the Aristotelian Society, supp. 61: 57–83. Gertler, Brie (2011). Self-Knowledge. New York: Routledge.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
346 Jennifer Lackey Goldberg, Sanford (2017). “Should Have Known.” Synthese 194: 2863–94. Goldman, Alvin. I. (1986). Epistemology and Cognition. Cambridge, MA: Harvard University Press. Goldman, Alvin. I. (2014). “Social Process Reliabilism: Solving Justification Problems in Collective Epistemology,” in Jennifer Lackey (ed.), Essays in Collective Epistemology. Oxford: Oxford University Press, 11–41. Guerrero, Alexander A. (2007). “Don’t Know, Don’t Kill: Moral Ignorance, Culpability, and Caution.” Philosophical Studies 136: 59–97. Guerrero, Alexander A. (2021). “The Epistemology of Consent,” in Jennifer Lackey (ed.), Applied Epistemology. Oxford: Oxford University Press. Hawthorne, John (2004). Knowledge and Lotteries. Oxford: Oxford University Press. Hawthorne, John and Jason Stanley (2008). “Knowledge and Action.” Journal of Philosophy 105: 571–90. Hurd, Heidi M. (1996). “The Moral Magic of Consent.” Legal Theory 2: 121–46. Lackey, Jennifer (2008). Learning from Words: Testimony as a Source of Knowledge. Oxford: Oxford University Press. Lackey, Jennifer (2010). “Acting on Knowledge.” Philosophical Perspectives 24: 361–82. Lackey, Jennifer (2011). “Assertion and Isolated Secondhand Knowledge,” in Jessica Brown and Herman Cappelen (eds), Assertion. Oxford: Oxford University Press, 251–75. Lackey, Jennifer (2016a). “Assertion and Expertise.” Philosophy and Phenomenological Research 92: 509–17. Lackey, Jennifer (2016b). “What Is Justified Group Belief?” The Philosophical Review 125: 341–96. Lackey, Jennifer (2018a). “Group Assertion.” Erkenntnis 83: 21–42. Lackey, Jennifer (2018b). “Group Lies,” in Eliot Michaelson and Andreas Stokke (eds), Lying. Oxford: Oxford University Press, 262–84. Liberto, Hallie (2017). “Intention and Sexual Consent.” Philosophical Explorations 20: 127–41. Ludwig Kirk (2014). “Proxy Agency in Collective Action.” Nôus 48: 75–105. McDowell, John (1994). “Knowledge by Hearsay,” in Bimal Krishna Matilal and Arindam Chakrabarti (eds), Knowing from Words. Dordrecht: Kluwer Academic Publishers, 195–224. Nozick, Robert (1981). Philosophical Explanations. Cambridge, MA: The Belknap Press. Parrott, Matthew (2015). “Expressing First-Person Authority.” Philosophical Studies 172: 2215–37. Plantinga, Alvin (1993). Warrant and Proper Function. Oxford: Oxford University Press. Pollock, John (1986). Contemporary Theories of Knowledge. Totowa, NJ: Rowman and Littlefield. Reed, Baron (2006). “Epistemic Circularity Squared? Skepticism about Common Sense.” Philosophy and Phenomenological Research 73: 186–97.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Sexual Consent and Epistemic Agency 347 Reed, Baron (2010). “A Defense of Stable Invariantism.” Noûs 44: 224–44. Reed, Baron (2013). “Fallibilism, Epistemic Possibility, and Epistemic Agency.” Philosophical Issues 23: 40–69. Stanley, Jason (2005). Knowledge and Practical Interests. Oxford: Oxford University Press. Wertheimer, Alan (2001). “Intoxicated Consent to Sexual Relations.” Law and Philosophy 20: 373–401. Westen, Peter (2004). The Logic of Consent: The Diversity and Deceptiveness of Consent as a Defense to Criminal Conduct. Burlington, VT: Ashgate Publishing. Williams, Michael (1999). Groundless Belief: An Essay on the Possibility of Epistemology, 2nd edn. Princeton, NJ: Princeton University Press. Williamson, Timothy (2005). “Contextualism, Subject-Sensitive Invariantism and Knowledge of Knowledge.” The Philosophical Quarterly 55: 213–35. Wrigley, Anthony. (2007). “Proxy Consent: Moral Authority Misconceived.” Journal of Medical Ethics 33: 527–31. Young, Cathy (2015). “Feminists Want us to Define Those Ugly Sexual Encounters as Rape. Don’t Let Them.” Washington Post (May 20, 2015), www.washingtonpost. com/posteverything/wp/2015/05/20/feminists-want-us-to-define-these-uglysexual-encounters-as-rape-dont-let-them/?utm_term=.9297b8f3bf34.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 19/03/21, SPi
15
The Epistemology of Consent Alexander A. Guerrero*
Moral, legal, and political philosophers have spent a great deal of time thinking about what consent is, and how consent can play the apparently transformative role that it appears to with respect to making some otherwise impermissible and objectionable conduct permissible and unobjectionable. These are hard and important metaphysical and moral questions. In this chapter, I draw attention to the fact that—depending on our account of what consent is—there may be hard and important epistemological questions about consent as well. I will suggest that debates about consent often mistake epistemological issues for metaphysical ones. People who defend so-called “affirmative consent” views sometimes are accused of, or even take themselves to be, offering a new (and controversial) view about the nature of consent. I think this is a mistake. I argue that the right way of understanding “affirmative consent” requirements is as a view about what is required, epistemically speaking, before one can justifiably believe that another person has consented. This view will be justified, if it is, because of background views about epistemic justification and the way in which epistemic justification interacts with moral norms governing action. Embracing an “affi rma tive consent” standard does not commit one to embracing either an “attitudinal/ mental” view (consent requires only the right mental attitude) or a “performative/ behavioral” view (consent requires the right mental attitude plus some communicative performance regarding that attitude) of the nature of consent. In this chapter, I offer a necessary component of any plausible view of what consent is, and then set out plausible norms governing the epistemology of consent, given this component. I then discuss the implications of this view in two real-world contexts in which consent figures prominently: consent to sex and consent to medical treatment. Importantly, the focus throughout is on moral questions, not legal ones. The aim of the chapter is to clarify what is required for various actions to be morally permissible. There is then the further question of * Thanks to Yuval Abrams, Hrafn Asgeirsson, Kristen Bell, Ryan Doerfler, Doug Husak, Jennifer Lackey, Hallie Liberto, Olivia Odoffin, Cosim Sayid, Daniel Viehoff, Gerardo Vildostegui, and audiences at the Workshop on the Public Discourse of Contested Sexual Consent at the University of Connecticut in September 2016 and the 2018 Rutgers- Columbia Undergraduate Philosophy Conference for helpful comments on earlier versions of this chapter. Thanks, too, to Tom Dougherty, who provided very helpful comments on this chapter in October 2017 as a referee for this volume. Alexander A. Guerrero, The Epistemology of Consent In: Applied Epistemology. Edited by: Jennifer Lackey, Oxford University Press (2021). © Alexander A. Guerrero. DOI: 10.1093/oso/9780198833659.003.0015
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Consent 349 how, or whether, the law (and comparable policies of, say, universities) should be crafted to align with morality. I discuss this question only briefly, although it is obviously an important part of the full “affirmative consent” discussion. Here is an overview of the argument: (1) Consent requires that a person have a specific mental state. More specific ally, an agent, A, consents to some state of affairs, SA, only if A has an attitude (implicit or explicit) of affi rmative endorsement toward SA. (2) Consent, when present, makes morally permissible some conduct that would otherwise be morally impermissible. (3) An agent, B, can non-culpably act as if another person, A, consents to some state of affairs, SA, only if B has a justifiable belief that A has an attitude of affirmative endorsement towards SA. I will suggest that we should accept (1). Indeed, I am inclined towards something stronger; namely: (Attitudinal View): Consent is a mental attitude towards a state of affairs. Specific mental attitudes are necessary and sufficient for consent. More specific ally, an agent, A, consents to some state of affairs, SA, if and only if A has an attitude of affirmative endorsement towards SA. The Attitudinal View says that mental attitudes are not only necessary for consent, but also that they are sufficient for consent. I will not give a defense of this view in this chapter, but I think that hesitation in embracing the Attitudinal View is typically the result of a failure to appreciate the truth and moral significance of (3). In particular, there are cases that appear to be similar to each other, but which differ in this respect: No Consent: An action of B’s that is wrong because A did not consent to B’s action, and B’s action is morally impermissible without A’s consent. Objectionable Moral Risk: An action of B’s that is wrong because B took an objectionable moral risk with respect to whether A consented or not, given B’s epistemic standing with respect to the question of whether A consented. In some cases of Objectionable Moral Risk, people might misdiagnose the moral problem as one of No Consent. The wrong in these cases might not be that the person did not consent. The wrong might be that the person did not have a justified belief that the person did consent, and, as (3) makes clear, it is morally objectionable to act as if one did have such a justified belief when one did not. Furthermore, acting in a case where one takes an objectionable moral risk regarding another’s consent is a way of disrespecting that person, and acting without sufficient regard
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
350 Alexander A. Guerrero for that person’s standing as a moral agent—it is just a different way of doing this than acting without that person’s consent. Additionally, in some cases, the relevant state of affairs to which the person does or does not have an attitude of affirmative endorsement can include facts about the state of affairs with respect to communication and explicit performance. For example, Joe might have an attitude of affirmative endorsement to this state of affairs: Talk: Moe asks if I want to have sex and then we talk about it and then (if the conversation goes well) we have sex while not having an attitude of affirmative endorsement towards this state of affairs: No Talk: Moe has sex with me without us ever talking about it In such a case, conversation or verbal performance will be relevant to whether any sex between Joe and Moe is consensual, but because of Joe’s specific attitudes, not because of the nature of consent. Section 1 of the chapter will explain and make the case for (1) and (3), and will discuss why some might have required something like expression or communication as part of the metaphysical account of what consent is. I will suggest that cases people have used to motivate the “performative” view of the metaphysics of consent really just motivate an appreciation of (1) and (3), and should leave us agnostic about whether the “attitudinal” or “performative” view of consent is correct. Section 2 of the chapter will focus attention on the epistemological details concerning justifiable belief about whether others consent to some state of affairs. The claims defended here are related to broader issues concerning the interaction of moral and epistemic norms, part of a debate about what I have called “moral epi stemic contextualism”1 and what others have discussed under the heading of “moral encroachment.”2 The basic set of questions concerns (i) the management and justification of our beliefs (typically the jurisdiction of epistemic norms); and (ii) the use of our beliefs as the basis of action, in high stakes moral contexts. Section 2 continues the overall argument of the chapter, focusing on the epistemology of consent: (4a) Whether (i) it is justifiable for B to believe that A consents to some state of affairs, SA, or (ii) whether B knows that A consents to SA, depends, in part, on the moral context, the moral stakes of the situation 1 Guerrero (2007, 69). Although I used the phrase “moral epistemic contextualism” in that paper, in the commonly accepted terminology that is now standard in the epistemological literature, the view I discuss would be better described as a form of “subject-sensitive invariantism,” where the subject sensitivity comes from the moral stakes in question. I discuss this at greater length in section 2. 2 Fritz (2017); Pace (2011); Basu (2018); Moss (2018); Gardiner (2018); Basu and Schroeder (forthcoming).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Consent 351 or (4b) Whether it is morally objectionable for B to act based on justified belief or knowledge that A consents to SA—and whether B is non-culpable for acting as if “A consents to SA” is true—depends, in part, on the moral context. These two premises both suggest that the moral stakes matter. But they make different suggestions as to how they matter. The first (4a) suggests that moral stakes affect epistemic considerations of justification and knowledge. The second (4b) suggests that moral stakes affect what we can do on the basis of propositions that we believe—even propositions that we justifiably believe or know. It could be that both of these suggestions are correct, that only one of them is, or that neither of them is correct. I will discuss these possibilities in section 2, arguing that we should accept at least one of (4a) or (4b). The argument then continues: (5) Some cases in which consent is important are, as a result, cases with high moral stakes, and so B must possess more/stronger evidence (the result of perhaps correspondingly more investigation) in order for B (i) to justifi ably believe that A consents or (ii) to non-culpably act as if A consents. (6) In particular, cases for which sexual or medical consent are important are cases in which there are high moral stakes. (7) In cases in which B is contemplating an action which is permissible only if A consents to states of affairs involving sexual activity or medical intervention, B must possess more/stronger evidence (perhaps requiring greater investigation) (i) to justifiably believe that A consents or (ii) to non-culpably act as if A consents. Exactly what “more” a person must do will be context dependent. But it is plausible that this is what makes “affirmative consent” the normatively appropriate standard with respect to sexual consent: an epistemic account about when we can justifiably believe that others have consented to engaging in sexual activity, or about when we can permissibly act on such a belief, not a metaphysical claim about when there is consent. We can remain agnostic regarding whether the “attitudinal” or “performative” account of the nature of consent is correct, while embracing “affirmative consent” ideas. I will make similar suggestions with respect to informed consent requirements in the medical context.
1. Consent and Beliefs about Consent What is consent? Let’s start by demarcating some central, uncontroversial properties of consent.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
352 Alexander A. Guerrero First, consent is always attached to an agent: consent is always someone’s consent, it always has a subject. I consent, or you consent, or she consents, or we consent. Second, consent always concerns some specific object: there is always something that someone is consenting to, even if that something has imprecise or underspecified boundaries. I consent to the surgery, you consent to be governed by this person for the next four years, he consents to having sex with that man later this evening, they consent to allowing the car dealership to withdraw $10,000 from their bank account. To remain suitably abstract to cover the broad range of things to which consent can be given, I suggest that we adopt the view that consent is always consent to a state of affairs with some (perhaps imprecisely specified) temporal duration. These states of affairs will often include the actions of other people, but they may only involve the granting of certain (perhaps unexercised) permissions to particular people. These are two uncontroversial descriptive claims about consent: consent always has a subject and consent always has an object. Here is a third uncontroversial claim about consent: when it exists, it is morally transformative. Consent alters how it is morally permissible to treat or to engage with the subject of the consent, the consenting person(s). The precise alteration involved depends on the state of affairs to which consent has been given. When there is consent, some states of affairs are morally permissible which would have previously been morally impermissible.3 We could countenance something like “superfluous” consent—something like consent, but in cases in which it is not needed, and in which it does nothing, morally speaking. I consent to the continued existence of tomatoes, you consent to my continuing to have a relationship with my parents, they consent to 1917 having been earlier in time than 1922, and so on. But this is a parasitic use of the idea of “consent.” The main reason this is so is that consent is plausibly connected to moral permissions or moral rights that we have regarding what we can permis sibly do and how we can be permissibly treated by other agents. A short story here goes like this: we have various moral rights, these rights come along with correlative duties that others have (typically negative duties: duties not to harm us, touch us, rule over us, impede us in going over here or over there, and so on), and consent releases some people from these duties vis-à-vis us.4 But I have no moral right regarding whether tomatoes continue to exist, whether you continue to have a relationship with your parents, and so on. So, this is not real consent; this is playing at consent or joking about consent.
3 Of course, consent alone may not be sufficient to make an act morally permissible. The fact that someone consents to sex does not make it permissible to have sex with that person if by doing so one will thereby kill 100 other people (fill in the details as you like). 4 Tom Dougherty has a nice discussion of this picture (2015).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Consent 353 Let us take these three claims as accepted by all: consent has a subject, consent has an object, and consent is morally transformative. The controversy comes not with noting that consent is morally transformative, but in giving an account of why or how it has this effect. There are two main views, the “attitudinal view” and the “performative view.”
1.1 The Attitudinal View On the “attitudinal” view, consent is a mental state that an individual might have, an attitude toward some real or imagined state of affairs. On this metaphorical picture, there is the terrain over which we as individuals get to move as we see fit, over which we get to govern. This is the province of our autonomy, the real and metaphorical space over which our beliefs, attitudes, commitments, and values ought to determine what happens. These are the things that we get to have control over, or that we normatively ought to have control over, at least, even if it doesn’t always work out this way. The suggestion on the attitudinal view is that an agent, A, consents to some state of affairs, SA, if and only if A has a particular attitude towards SA. Consent is morally important on this view because consent is essentially a part of an individual’s attitudinal map, a plot of where she’d like to go—based on her values—if it were (as it should be) up to her. The precise details of the required attitude or mental state are a matter of some debate. On one view, the attitude is an intention with some specific content.5 On another view, the attitude is one of acquiescing in the consented-to conduct.6 On yet another view, the mental state is an attitude of waiving one’s right that some particular conduct not occur, where this is a matter of being disposed not to object to another’s conduct based on that right.7 On another, the attitude might be a desire. On the view that I find most plausible, the attitude is one of what we might call “affirmative endorsement” towards some state of affairs, SA. This attitude might be a combination of beliefs, desires, and intentions, but it is perhaps most plaus ible just to think of it in terms of a robust pro-attitude towards SA. “Robust” and “affirmative” in the previous sentences are meant to indicate the depth and relative integration of the attitude in the agent’s overall set of beliefs, desires, and values. What it is meant to rule out are the “pro-attitudes” that can be generated towards states of affairs simply by dramatically and artificially limiting the option
5 Hurd (1996). 6 Westen (2004, 25–63). Or see Ferzan (2016). On Ferzan’s view, the person giving consent must only have the view that another person’s action is “okay with them.” 7 Alexander (1996, 2014).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
354 Alexander A. Guerrero set of states of affairs that the agent believes might be brought about. “Your money or your life” might lead one to have a highly contingent pro-attitude towards the state of affairs in which you give the robber all of your money. But this will just be, on this terminology, “mere” acceptance, or “mere” endorsement, or a “nonrobust” pro-attitude towards that state of affairs. Of course, spelling out the details here is notoriously difficult, as a quick assessment of the literature on coercion and exploitation makes evident. We will end up owing an account of when baselines are artificial (rather than “natural”), when a choice stems from an agent’s “deep” (rather than “shallow”) values, or when a choice counts as an authentic and robust expression of an agent’s autonomy, rather than something else. And it is made more difficult because, of course, when the agent hands over her money, rather than foregoing her life, this is a deep reflection of what she values, what really matters to her, and so on. Trying to specify the kinds of backgrounds against which an agent can count as consenting to some state of affairs is a difficult task. In real life, we acknowledge that some situations are so fraught with comprom ise and coercion that consent is impossible (or something close to that). Consider, for example, rules that mandate that all sex between correctional officers and people who are currently incarcerated counts as non-consensual. The important point is that we need to operate with a relatively thick understanding of the normative significance of agreement to some state of affairs obtaining, an understanding on which agreement is related to foundational ideas of autonomy and self-respect.8 On this view of “robust” or thick agreement, it can be reasonable to refuse to agree to a bad situation, even if that situation is not the worst situation, and even if that situation is better than one’s other options. And we cannot assume the robustness or reasonableness of agreement to some state of affairs simply because on the agent’s ranking of the options that she faces, that state of affairs is her highest-ranked option. To focus on relatively clear cases of consent, we can restrict our focus to those cases in which the balance of A’s affirmative endorsement of or pro-attitude towards some state of affairs, SA, is due to positive features of SA—features of SA that appear as good or valuable to A, by the lights of A’s evaluative framework— rather than due to the badness of features of A’s alternatives, SA1, SA2, etc. This is a placeholder of a view, but it will be sufficient for the purposes of this chapter.
1.2 The Performative View The performative view of consent maintains that, although having a particular mental attitude is necessary for an agent to consent, this is not sufficient for 8 For discussion, see Guerrero (2017).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Consent 355 consent.9 Instead, the performative view requires a particular mental attitude plus a communicative act of some kind (whether verbal or non-verbal) that serves to communicate the existence of this mental attitude.10 This kind of view focuses less on the moral importance of allowing individuals to follow their moral map through autonomy-central aspects of the world, and more on an explicitly interpersonal dance of non-permission, explicit waiver, and subsequent specific permission. On this view, although it is important that A has the appropriate mental attitude towards SA, that is, by itself, not enough to do anything morally transformative. Instead, the fact of A’s mental attitude towards SA must be communicated to B, in order for anything morally transformative to occur. In the standard case, we imagine that SA includes B doing something to or with A, something that would otherwise be morally impermissible. A, by both having the appropriate mental attitude and communicating this fact to B, waives a certain kind of moral right against B, eliminates a moral duty that B otherwise would have with respect to A, and thereby makes morally permissible action of B’s that involves A and which would otherwise be morally impermissible. All of this happens through some public communication—in the standard case, some communication (verbal or non-verbal) directly from A to B. There is more that could be said about both the attitudinal and the performative views of consent. For now, I want to draw attention to two other claims. First, both of these views are committed to (1) or something like it: (1) Consent requires that a person have a specific mental state. More specifically, an agent, A, consents to some state of affairs, SA, only if A has an attitude (implicit or explicit) of affirmative endorsement towards SA. This is worth stressing in part because some have discussed views on which consent is just about the requisite kind of performative or communicative act. On this kind of view, in relatively normal conversational contexts (involving competent speakers of the language, etc.), saying, “I consent to SA” would be both necessary and sufficient for consent, whether the person had any particular mental attitude or mental state or not. We can call these “bare performance” views of consent. These views run into a host of problems. We might, for example, think that saying “I consent to SA” in a particular context might be the result of fear,
9 At least, this is the view that has most recently been referred to as the “performative” view by Tom Dougherty, among others (Dougherty 2015). Alan Wertheimer discussed a view that he called the “performative view” on which no particular mental state is even necessary. On that view, which he defended in early work, an appropriate communication of consent is sufficient for consent (Wertheimer 2003, 144). 10 This view, described by some as the “hybrid” view, is presented and defended in, e.g. Owens (2011) and Miller and Wertheimer (2010).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
356 Alexander A. Guerrero intimidation, and a host of other considerations compatible with the person having a robust attitude of non-endorsement towards SA. On either the moral entitlement to chart one’s own course over this terrain or the moral waiver views of consent, it is hard to see why this utterance would be morally transformative. Even if we treat cases in which there is coercion or duress differently than normal contexts, we still might think a simple assertion of “I consent to SA” (made on a whim, or whatever) might be troublingly little basis for others to rest on, particularly if SA includes things like performing surgery on A, sending A to war, or having sex with A. It is plausible that the mental state behind the assertion is what really matters. Even if we don’t go this far, we might think that it definitely does matter, even if it is not all that matters. Some who might have been tempted by these “bare performance” views of consent might have been worried about what we might call “innocent recipients.” These people are the recipients of what looks for all the world like perfectly valid consent, but which is actually the product of something other than the usual causal chain that runs from a mental attitude of affirmative endorsement towards SA to a communicative act of “I consent to SA.” Instead, the “I consent to SA” they receive is the product of someone else’s hypnosis, coercion, deception, or some kind of spontan eous and unreflective impulse. (Obviously, the case is different if the recipients of what looks like consent are themselves responsible for getting the person to engage in the speech act.) These “innocent recipients” have, by hypothesis, done nothing wrong. They may have even been very careful to ask the person in question all the right questions. The “bare performance” views can say: the person said, “I consent to SA” and either there was no coercion, duress, etc., or the recipient of the consent was not aware of it nor responsible for it, so: morally transformative consent. But one needn’t adopt this implausible view of consent in order to say plausible things about innocent recipient cases. Recall, for example, claim (3): (3) An agent, B, can non-culpably act as if another person, A, consents to some state of affairs, SA, only if B has a justifiable belief that A has an attitude of affi rma tive endorsement towards SA. This claim does not imply, implausibly, that the person who says, “I consent to SA” while under hypnosis actually does consent to SA. But it does allow that what an innocent recipient does may be morally non-culpable, if her belief is justifiable. Note, however, that (3) is just a necessary condition on morally non-culpable action—there will be other conditions that would likely have to be satisfied as well. One of those conditions might be that A actually does consent to SA. If that is correct, then it may seem that the “innocent recipient” is back on the hook. Even if they satisfy (3), this other condition may not be satisfied, and so it may not be morally permissible for the innocent recipient to act as if A consents.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Consent 357 Again, however, there is something more plausible that we can say, if we endorse a related principle regarding the distinction between moral blameworthi ness and objective moral wrongdoing: Mistaken Belief: If B has a justifiable belief that A has an attitude of affirmative endorsement towards some state of affairs, SA, then B is not culpable for acting as if A consents to SA—even if it turns out that A does not consent to SA, and even if this means that what B does (given that A does not consent to SA) is objectively morally wrong. If we accept Mistaken Belief, then in some cases in which “innocent recipients” act without actual consent, they do something that is objectively morally wrong, but for which they are not blameworthy.11 In one familiar way of speaking about such things, we might say that they have an excuse for what they do.
1.3 Knowing Other Minds If we accept this as a response to the worry of “innocent recipients,” then we have no reason not to accept what seems independently plausible: consent requires that the consenting individual has a specific mental attitude towards some state of affairs. But this leaves all of us with a problem: the problem of other minds. There is a familiar kind of deep philosophical skepticism here: how can we even know that the people around us even have minds. Let us leave that worry aside, as moral concerns about consent only really get off the ground if we are engaging with creatures with minds. Still, given the mental state component of consent (whether it is all that is required, or just part of what is required), there will be the hard issue of when we have a justifiable belief that a person has the requisite mental attitude. There are several significant difficulties here. The first is just the familiar difficulty about knowing what is going on in the minds of other people. This is just a generalization of the problem Whitney Houston focuses on: 11 This is, arguably the correct response to a case such as the one that Wertheimer discusses in defending the “bare performance” view (which he calls the “performative” view) against the view on which a mental state and communication of that mental state are both necessary for morally transformative consent. Wertheimer discusses a case in which an inattentive patient thinks that she is only authorizing a biopsy when she signs a consent form for a lumpectomy, although her physician did explain the procedure and the form to the patient (2003, 148). Wertheimer suggested that the physician did enough to ascertain that the patient consented, and so the expression of consent makes it morally permissible to perform the lumpectomy. A more plausible view, I think, is that performing the lumpectomy in these circumstances (when the patient does not possess a mental state endorsing or authorizing this state of affairs) is objectively morally wrong, although not morally blameworthy, and perhaps “subjectively” morally permissible, if one wants to countenance this category.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
358 Alexander A. Guerrero How will I know (Don’t trust your feelings) How will I know How will I know (Love can be deceiving) How will I know How will I know if he really loves me?
There are various sources of evidence we might have to help us assess whether a particular person has a particular mental state, such as the mental state of being in love with Whitney Houston. One of these is the direct testimony of the person whose mental states are in question. We might (if not too shy) ask the person directly and see what they say. This is defeasible evidence, of course. The person might have reasons to misrepresent their mental states, or they may be self-deceived about their own mental states. A second source of evidence is the testimony of third parties regarding the person in question’s mental states. Maybe their best friend made an inquiry on your behalf, and they report back to you. Again, this is defeasible evidence, with even more ways of going astray; it inherits all of the possible problems with direct testimony, as well as possible concerns about third-party deception, misrepresentation, miscommunication, and so on. A third source of evidence, which is just a more general category that encompasses the first two, is all of the perceptual and observational evidence we might gather regarding the person’s behavior (actions, reactions, emotions) or the behavior of others that might be used to draw inferences about the person’s mental states. As with the other sources, there are familiar ways in which this can go wrong. Indeed, we are even capable of being misled as to whether some entity even has mental states. Assuming we get that right, we can still be in significant error regarding the nature and specific content of those mental states. As Houston herself suggests, there are familiar sources of error in this regard. One is biased or misleading interpretation of evidence, due to one’s psychology and background investment through processes described by such labels as “motivated reasoning,” “emotional reasoning,” “confirmation bias,” and so on. Subtly misread actions, facial expressions, and similar ostensible clues to the inner life of others are the bread and butter of many artistic creations: songs, films, Jane Austen novels. These are things that can be hard to get right. On the other hand, we often feel we know exactly what another person feels, or believes, or endorses: because they tell us, because we can see it all over their face, because of what they do. The empirical literature on what is called “mindreading” suggests some skepticism about our abilities in this arena is warranted. Shannon Spaulding, summarizing that literature, notes: Most people think of themselves as pretty good at understanding others’ beliefs, desires, emotions, and intentions. However, social psychologists have discovered
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Consent 359 that we are significantly worse at mindreading than we think we are . . . We consistently and substantially overrate our ability to accurately judge others’ mental states and interpret social interactions . . . the consensus from the empir ical literature is that mind misreading is very common. (Spaulding 2016, 423)
Spaulding discusses some of the implications of this, noting particularly some of the common causes of mind misreading: “(1) we are too cognitively taxed to engage in thorough information search, (2) we pay attention to superficial cues, (3) we are biased by self-interest, (4) we fail to understand our own mental states, (5) and we inappropriately deploy stereotypes.”12 Obviously, there will be cases involving belief about the consent of others in which these potential sources of mindreading error are present. Importantly, it is plausible that the more experience that we have interacting with a person, the easier it is to correctly infer what their mental states are, based on what they say and do, based on what we observe of them. This might help us avoid attending to superficial cues and using stereotypes in interacting with others. There are ways of helping to pay greater attention to avoid self-interest biases, and familiar ways of getting more and better evidence about what a person’s mental attitudes are towards various possible or actual states of affairs. In the case of the mental state relevant for consent, which I have described as an attitude of affirmative endorsement towards some state of affairs, there are some further issues worth discussing explicitly. First, the attitude of affirmative endorsement might vary in its details, depending on the state of affairs towards which it is directed. In particular, it might be connected to more or less emotional and affective response and to very different kinds of emotional responses, it might be more or less phenomenologically vivid, it might be more or less transparent and known to the individual whose attitude it is (even if we want to require that it always be at least somewhat introspectively available), and it might be more or less integrated into the person’s overall set of beliefs, values, and other psychological attitudes. As a result, the attitude of affirmative endorsement might be relatively passive or active in its effects. Sometimes we consent to states of affairs that do not require us to do very much at all, just to accept the state of affairs. On other occasions, consent to a state of affairs will be connected to all kinds of other intentions and actions. In these ways, consenting to serious surgery is different than consenting to having your picture appear in your employer’s monthly newsletter. Consenting to take part in a religious conversion ceremony is different than consenting to sex. Consenting to be governed by a candidate whom one views as an effective but uninspiring pragmatist is different than consenting to be governed by a candidate whom one
12 Spaulding (2016, 434).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
360 Alexander A. Guerrero idolizes and loves. Consenting to join the army is different than consenting to be the emergency contact listed on a friend’s medical form. This kind of diversity is not well captured by views that require that the mental state in question be an intention or a desire, or so it seems to me. Sometimes the attitude is connected to, or even caused by, intentions and desires. But it need not always be. Second, in addition to knowing whether a person has the mental attitude of affirmative endorsement towards some state of affairs, there is the difficulty of knowing precise (or even relatively precise) facts about the state of affairs towards which the person has the attitude, particularly given the aforementioned diversity with respect to those attitudes. The object of consent is clearly of crucial import ance, and on this account, it is a component (the object) of the relevant mental attitude. But it is also hard to know the boundaries of the state of affairs to which consent has been given, even when one knows that the person in question possesses an attitude of affirmative endorsement to some state of affairs. There are some practices that aim to fix the object of consent with precision. Consider, for example, the work that goes into drafting explicit contractual language, or that guides the creation of multipage forms in the medical context aimed at securing “informed” consent to a particular medical procedure. But these are the exceptional case. Many of our interactions do not involve this much clarity—either to the subject whose consent it is, or to those of us who may be interested in what, if anything, the person consents to. This is one reason, for example, that we might expect consent to sexual activity to be a complicated matter, and why it is normatively important to be attuned to the details of what is going on with the other person throughout the duration of a sexual encounter. A person might not have explicit views about the exact states of affairs to which they are consenting when beginning certain activities, even if they begin with some kind of consent to states of affairs in some general category. Third, an additional difficulty is that the specific mental attitude of the person consenting is subject to change over time, depending on the temporal duration of the state of affairs to which the person consents. Very different than promises or legal contracts, it is plausible that consent can be withdrawn or altered at any moment up until the consented-to state of affairs is concluded. If the state of affairs I have consented to has not yet begun or is still ongoing, I can withdraw my consent. Even if a person says that “I consent to SA” while possessing the requisite mental attitude regarding SA, they might still change their mind (no longer having the attitude of endorsement toward SA), without being able to say anything. In such a case, it is plausible that they no longer consent to SA. A person scheduled for surgery who is conscious but can no longer communicate, but who decides that she doesn’t want the procedure and no longer has an attitude of affirmative endorsement to the state of affairs in which the procedure is conducted, no longer consents to the surgery, although it may be true that she previously had consented to it. This is one reason to make as certain as possible that consent in a particular
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Consent 361 case is likely to be stable over time. Now, it is true that in some cases, if a person’s consent is made known to others at time T1, if it is altered or revoked or newly withheld at time T2, that may leave those others in the position of “innocent recipients” if they have not been apprised of this change from T1 to T2. But that does seem to be the situation people are in under those circumstances. And that follows on either the Attitudinal or the Performative views of consent, given that both require that the person currently have a specific mental attitude—not just that they have had the specific mental attitude at some point in time. A fourth difficulty is that, once we have noted that consent involves a mental attitude of affirmative endorsement towards some state of affairs, there is the question of whether it can include implicit attitudes of affirmative endorsement, as well as explicit ones. And then there would be the question of whether both implicit and explicit attitudes of affirmative endorsement both are of equal normative significance. There are different ways of understanding implicit attitudes. On one understanding of them, they differ from explicit attitudes in being both relatively automatic and introspectively inaccessible. This is the sense of “implicit” people use when speaking of, for example, implicit bias. A second sense of “implicit” is more familiar from discussion of implicit or tacit belief, where the difference concerns whether a representation with a certain content is actually “present” or “occurrent” in one’s mind in a certain way. So, for example, in the case of belief, one might believe some proposition p explicitly if a representation with that content is actually present in the mind, whereas one believes some proposition p only implicitly or tacitly if one believes p, but the mind does not currently possess a representation with that content. In the case of attitudes of affirmative endorsement, we might say that one can have an attitude of affi rma tive endorsement towards some state of affairs either explicitly (where the person is presently conscious of this attitude and is attending to it) or implicitly (where the person has this attitude towards the state of affairs, but is not presently conscious of this attitude, nor is she attending to it). Of course, in both cases, an account is still owed about what it is to believe or to possess an attitude. It is very plausible that we believe things beyond those things that we are presently attending to. I believe that there is not a monkey sitting on my shoulder as I type this. I believe that my car is where I parked it earlier today. I believe that the Library of Congress has a copy of John Rawls’s A Theory of Justice. And so on. I believed these things implicitly or non-occurrently, prior to writing those sentences. While coming to think about them, or while writing them, or just after writing them, I believe them explicitly and occurrently. These beliefs are not automatic, nor are they introspectively inaccessible. That is not the relevant sense of “implicit” here. Similarly, it seems that we have many implicit attitudes toward states of affairs. I hope that my cousins in Miami are doing well. I regret that I never had a chance to meet my paternal grandfather. I consent to having my photograph on the departmental website. And so on. Again, I think I had these
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
362 Alexander A. Guerrero attitudes implicitly or non-occurrently, prior to writing those sentences. After or while writing them, I attend to those attitudes, making them explicit. But I don’t think I thereby create new attitudes that I didn’t previously possess. Of course, in the context of consent, it may be controversial whether implicit attitudes can ever be sufficient to do the normative work we expect from consent. I think they can play this normative role. It is plausible, for example, that I currently have the implicit attitude of affirmative endorsement towards this state of affairs: that if I have suffered a devastating and debilitating accident, then I am treated by medical professionals to the best of their abilities, even if I am not conscious and cannot discuss their plans with them. I think we have lots of standing attitudes of affirmative endorsement of this sort, and they are (most of the time) held only implicitly, at least until they are brought to our attention and thereby made explicit. It is true that in the vast majority of cases of consent, there will be occasion to make an attitude of affirmative endorsement towards some state of affairs shift from being implicit to being explicit, as the state of affairs is presented to one for consideration and response. But I think that in many cases of what people call “imputed,” or “counterfactual,” or “hypothetical” consent, there is actual consent, although it is implicit—in the same way that I actually believe that there is not a monkey sitting on my shoulder, although (prior to writing this sentence) I believe that only implicitly. This is controversial, however, and some might not follow me in thinking this. We might bypass some of these difficulties if we switched from talk of attitudes to talk of traits, understood as “broad-track dispositions to behave and cognize” in particular ways.13 If we went down that route, we could talk not of our attitudes of affirmative endorsement, but instead about our dispositions to respond in various ways towards various states of affairs. Roughly: we respond positively and attract ively (in various ways) to those states of affairs that we endorse and negatively and aversively (in various ways) to those that we do not endorse or that we reject. Thinking of consent in this way would be revisionary, but mostly because we are used to thinking of consent only in relatively explicit, discrete, episodic modes—something close to saying “I do” or signing our name on a dotted line— rather than as a standing trait or disposition. But if we think about what seems to be the correct normative picture in the case of, for example, consent to sexual activity, it is not obvious to me that this other dispositional, fluid, responsive picture of consent is a non-starter. Indeed, it might seem plausible, depending on what the object of consent is in a particular case. Although we can promise without saying, “I promise to ____ ,” much of our promising does involve actual explicit statement to that effect. Consent is mostly not like that. We consent to all kinds of things, all day long, without ever once saying the word “consent.” And it
13 See Machery (2016).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Consent 363 is not obvious that some kind of relatively fluid, dispositional account is a non-starter when we think of consent. Consider, for example, the view discussed in section 1.1, on which the relevant mental state is an attitude of waiving one’s right that some particular conduct not occur, where this is a matter of being disposed not to object to another’s conduct based on that right.14 But I will leave this issue to the side, here, hoping only to have convinced the reader that implicit attitudes might well be part of the picture of the metaphysics of consent. A fifth and final difficulty is that if we accept that a mental attitude of affi rma tive endorsement is at least a component of consent, then we should be open to the possibility that consent itself comes in degrees, rather than only being a binary, all-or-nothing state. The clearest route to this conclusion would be to accept a picture on which attitudes of affirmative endorsement towards states of affairs come in degrees. In some cases, we might completely endorse a state of affairs, or completely reject it. But in other cases, our attitude towards the state of affairs might be more mixed. I might have an attitude of tentative but affirmative endorsement—I’m at 70 per cent, not 100 per cent. As a result, we should think that I “kind of ” consent to the state of affairs, or that I mostly consent, or we might say that the encounter was largely or mostly consensual. This is perhaps an unsettling realization, since it might be nice to think consent comes entirely in black or white. But I think that is implausible as a characterization of actual moral life. We arguably should, as a moral matter, strive to engage with others in ways that are entirely consensual, rather than mostly consensual or only somewhat consensual. But it seems a mistake to think that all interactions must be classified as fully or not-at-all consensual as a conceptual matter. And so it is not an embarrassment for an account of the metaphysics of consent that it has this feature. This, again, seems to push us away from seeing consent as any kind of analogue of promise. We might try to avoid this conclusion by identifying some distinctive kind of attitude that is the ‘consent’ pro-attitude, and which we either have or do not have towards all possible states of affairs. Or we could identify some threshold point such that ‘consent’ only exists if one has the attitude of affirmative endorsement that registers at or above this threshold. Such responses should be familiar from discussion of degrees of belief and attempts to avoid various problems that arise when we think of belief as something other than an on-or-off notion. I don’t think these responses are particularly plausible or needed, and I won’t at any rate pursue them further here. In this section, I have argued for a particular picture of the metaphysics of consent, one on which, crucially, a person consents to a state of affairs, SA, only if the person has a particular mental state: namely, an attitude of affirmative
14 Alexander (1996, 2014).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
364 Alexander A. Guerrero endorsement to SA. I suggested that both Attitudinal and Performative views of consent should endorse this necessary condition. And I suggested that with this component comes a host of distinctive difficulties. The attitude of endorsement itself can be different in its nature and connection to other mental states and attitudes, depending on the state of affairs towards which it is directed. The attitude of endorsement will have a particular object, a particular state of affairs, but the precise boundary of that object can be unsettled and/or difficult to discern. And the attitude of endorsement can change over time, be implicit as well as explicit, and plausibly comes in degrees. Noting these components of the metaphysics of consent is essential for getting clear on the subject of the next section: the difficulty and complexity that can be involved in the epistemology of consent—knowing whether a person consents to some state of affairs, the robustness of that consent, whether the person continues to consent to that state of affairs, and what precisely it is to which the person consents. I will suggest that none of this is easy and, given the moral significance of consent, it complicates the epistemological story we should tell.
2. High Moral Stakes and the Epistemology of Consent It is morally important for us to know whether other people consent to various states of affairs. This is particularly true in cases in which the states of affairs do, or might, involve us doing things to and with other people that would be impermissible without their consent. In section 1 of the chapter, I offered the following claim: (3) An agent, B, can non-culpably act as if another person, A, consents to some state of affairs, SA, only if B has a justifiable belief that A has an attitude of affi rma tive endorsement toward SA. I also suggested that we should accept this claim: Mistaken Belief: If B has a justifiable belief that A has an attitude of affirmative endorsement towards some state of affairs, SA, then B is not culpable for acting as if A consents to SA—even if it turns out that A does not consent to SA, and even if this means that what B does (given that A does not consent to SA) is objectively morally wrong. Both of these claims fix our attention on whether an agent has a justifiable belief that another person consents to some particular state of affairs. We might also want to know that another person consents to some particular state of affairs, to the extent that knowledge encompasses both justification and truth of
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Consent 365 the proposition believed, as well as some (controversial to state) connection between those two things. Thus, we might endorse an even stronger necessary condition than (3): (3—knowledge) An agent, B, can morally permissibly act as if another person, A, consents to some state of affairs, SA, only if B knows that A has an attitude of affirmative endorsement towards SA. This condition might be thought appropriate, at least if we are looking to describe objective permissibility, rather than mere subjective permissibility (if we want to countenance this latter category). So, we have moral reasons to be concerned with whether we justifiably believe or know that other people consent. And there are reasons, spelled out in section 1, to see this as somewhat difficult, due to the metaphysical facts about consent: in particular, due to the fact that consent includes as a significant component whether other individuals have particular mental states.
2.1 Moral Stakes, Moral Encroachment, and Justification Here is a natural question: does the moral significance of these beliefs about consent affect how much individual agents must do in order to be justified in holding these beliefs, or to count as knowing that a person consents to a specified state of affairs? In earlier work, I suggested the following: The more morally significant the actions that a belief in p (or absence of a belief in p) will support or license, the more stringent the epistemic demands that must be met before one can act as if one is justified in believing that p. Importantly, this “increase” in the epistemic demands is required by moral con siderations, not epistemic ones. We might also think that what is at stake, morally, in believing p actually alters when one is justified in believing p, but I am primarily concerned just with the question of when one is justified in acting as if one is justified in believing p. (Guerrero 2007, 69)
This suggests two relevant possibilities, which are reflected in these two different components of (4): (4a) Whether (i) it is justifiable for B to believe that A consents to some state of affairs, SA; or (ii) whether B knows that A consents to SA, depends, in part, on the moral context, the moral stakes of the situation
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
366 Alexander A. Guerrero or (4b) Whether it is morally objectionable for B to act based on justified belief or knowledge that A consents to SA—and whether B is non-culpable for acting as if “A consents to SA” is true—depends, in part, on the moral context. These two claims both suggest that the moral stakes matter. But they make different suggestions as to how they matter. The first, (4a), suggests that moral stakes affect epistemic considerations of justification and knowledge. The second, (4b), suggests that moral stakes affect what we can do on the basis of propositions that we believe—even propositions that we justifiably believe or know. It could be that both of these suggestions are correct, that only one of them is, or that neither of them is correct. I think that we should accept at least one of (4a) or (4b) and will try to make that case here. It is worth briefly connecting these views to other related views regarding ways in which non-epistemic considerations (considerations that do not have ‘some sort of intimate connection with truth’15)—whether pragmatic or moral—can affect epistemic justification and knowledge. Some have argued that there is a connection between knowledge and action, so that if one knows that p, then p is warranted enough to justify you in φ-ing, for any φ.16 But if one goes in for that kind of claim—tying epistemic considerations regarding knowledge to considerations of when one is justified in acting, or when one is justified in using p in some chain of reasoning—then one might argue that non-epistemic features seem to affect what one can permissibly do, either on pragmatic or moral grounds. The easiest way to see this is to compare pairs of cases in which individuals in each of the cases have the same evidence or justifying basis for a belief, but what they are planning on going on to do on the basis of that belief is different. Consider: Imagine two different people, each of whom is attempting to assess whether there are any people inside an abandoned house. The first person, Jack, is attempting to determine whether anyone is inside because he is trying to determine how many people live in the town. The second person, Jill, is attempting to determine whether anyone is inside because she is charged with demolishing the house. (Guerrero 2007, 68)
15 Ichikawa and Steup (2016). 16 Fantl and McGrath (2009, 66), building on earlier work (2002, 2007). And compare: “If a subject knows that p, then she is in a good enough epistemic position to rely on p in her practical reasoning” (Brown 2008); and “Where one’s choice is p-dependent, it is appropriate to treat the proposition that p as a reason for acting iff you know that p” (Hawthorne and Stanley 2008, 578).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Consent 367 Now imagine that Jack is just curious about this; nothing important or official turns on his count. And imagine that Jack does what is ever required for him to come to justifiably believe that there is no one in the house. Perhaps he asks some people who live nearby, knocks on the door, peeks his head inside, yells a few times, and then comes back and does this a week later. If he sees no sign of anyone after having made these investigations, it is plausible that he has done enough to justifiably believe that no one lives in the house. And, if it is true that no one lives in the house, then it seems plausible that he knows that no one lives in the house. But then consider Jill, who is investigating the question of whether anyone lives in the house because her construction company is charged with demolishing it. Imagine that she does all the same things as Jack in terms of investigation. She asks some people who live nearby, knocks on the door, peeks her head inside, yells a few times, and then comes back and does this a week later. If she sees no sign of anyone after having made these investigations, is it plausible that she has done enough to justifiably believe that no one lives in the house? And, if no one lives in the house, is it plausible that she knows this? Given what they have done to investigate and given their evidence, it is very natural to think both (i) that Jack justifiably believes that no one lives in the house and that Jack knows that no one lives in the house; and (ii) that Jill is not justified in believing that no one lives in the house and that, even if it is true that no one lives in the house, that she does not know that. One reason we might believe (ii) is that we accept the connection between knowledge and justifiable action articulated above, and we do not yet think that Jill would be justified in demolishing the building based on her belief that no one lives in the building, so we infer that she must not know that no one lives in the building.17 This line of reasoning can be used to motivate what has come to be called “subject-sensitivity,” the claim that “the truth-value of a knowledge attribution depends on features of the attribution’s subject’s environment that are not paradigmatically epistemic features.”18 Perhaps the most common form of subject-sensitivity that has been discussed is sensitivity to stakes, with most of the emphasis having been on the practical stakes to the agent, but with some now focusing on the moral stakes involved. James Fritz formulates the basic argument schema here as follows: (1) Knowledge-Action Link: If S knows that p, then S is warranted enough to act (believe, prefer) as if p. (2) Action-Environment Link: Feature F of S’s environment (where F is not a paradigmatically epistemic feature of an environment) makes a difference as to whether S is warranted enough to act (believe, prefer) as if p. 17 James Fritz has a nice presentation of argument schemas of this kind (2017, 4–6). 18 Ibid., 1.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
368 Alexander A. Guerrero (3) Therefore, Subject-Sensitivity: The truth-value of a knowledge attribution depends on features of the attribution’s subject’s environment that are not paradigmatically epistemic features. (Fritz 2017, 4–5) The suggestion would then be that features of Jill’s environment make a difference to whether she is warranted enough to act as if p, and if she cannot act as if p, then she must not know that p. Why should we accept that (2) is true, that it is the non-epistemic features, namely the stakes, that are doing the work? The answer: because of cases like the one involving Jack. This argument, if accepted, gets us to subject-sensitivity, and to versions of both pragmatic and moral encroachment (we can identify Jack and Jill cases, varying the stakes along both personal prudential or pragmatic dimensions, or moral dimensions). This, in turn, lends support to (4a), once we note that consent is of high moral significance: (4a) Whether (i) it is justifiable for B to believe that A consents to some state of affairs, SA; or (ii) whether B knows that A consents to SA, depends, in part, on the moral context, the moral stakes of the situation. This is fine as far as it goes, but it does require one to sign on to the KnowledgeAction Link to get to subject-sensitivity or stakes-sensitivity. And we might be tempted to reject that claim, or claims like it.19 We might also find that accepting pragmatic and/or moral encroachment is a high price to pay; one that might independently give us pause. One worry is that it is troubling to accept the idea that Jack and Jill can have the exact same evidential support for some proposition, such as the proposition that no one is in the house, and yet be in different places in terms of epistemic justification or know ledge with respect to that proposition.20 A second worry is that it can seem that this encroachment of the moral or pragmatic on the epistemic is introducing these concerns in the wrong place, putting weight on the epistemic side, rather than on the practical side concerning what a person is going to do with what she believes or even knows. Reflecting on Jill’s situation, we might not object to Jill holding the belief that the house is empty, and she could even use this belief for a whole bunch of purposes—just not all purposes. If we move away from some simple knowledge-action connection, we might accept something like (4b) in thinking about the Jack and Jill cases. Recall: (4b) Whether it is morally objectionable for B to act based on justified belief or knowledge that A consents to SA—and whether B is non-culpable for acting as if “A consents to SA” is true—depends, in part, on the moral context. 19 See, e.g. Jennifer Lackey, “Sexual Consent and Epistemic Agency,” Chapter 14, this volume. 20 For arguments in this direction, see Gardiner (2018).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Consent 369 The Knowledge-Action Link suggests that if we know that p, then we can always act as if p. But we might decide that we don’t want to go in for pragmatic and moral encroachment, for a host of reasons. We might then sever this connection, and say: given that they have exactly equivalent evidence, Jack and Jill are either both justified in believing that no one is in the house, or neither of them is. Either they both know, or they both don’t know. But leave those epistemic questions aside. The important difference these cases bring out is a practical one. They each believe p with a certain level of justification, based on a certain set of evidence. Beliefs at that level of justification, based on that kind of evidence, can warrant some actions, but not all actions. On this picture, there are normative limits on what the beliefs can be used to rationalize or justify or ground, based on the level of justification of those beliefs. Jack can non-culpably do what he is contemplating doing, based on a belief that is as justified as his is. But Jill cannot non-culpably do what she is contemplating doing, based on a belief that is as justified as hers is. On this kind of view, we might reject (4a), while endorsing (4b). There is more that might be said here, but it seems that in response to Jack and Jill kinds of cases, we are left with two possible views: View One—Moral Encroachment: moral stakes can affect epistemic justification and knowledge. Contexts differ, morally speaking, in that they differ in the moral significance of the actions under consideration which a belief in p (or absence of belief in p) will, or might, license or be used to justify in that context. As the moral stakes increase, the exact same level of evidentiary support can result in different consequences with respect to whether an individual justifiably believes that p or whether they know that p. In Jack’s context, he justifiably believes and knows that there is no one in the house. In Jill’s context, she does not justifiably believe or know that there is no one in the house. View Two—Scope of Use Tied to Epistemic Justification: moral stakes can affect what actions we can take based on certain beliefs, depending on the level or kind of justification we have for those beliefs. In particular, as the moral stakes increase, beliefs that would have been “actionable” for us might become “unactionable”21 with respect to certain actions—incapable of supporting, justifying, or being the rational ground for those actions—despite nothing changing about the justification with which we hold the beliefs in question. Jack’s belief that there is no one in the house can serve as justification for the action he is contemplating. Jill’s belief cannot serve as justification for the action she is contemplating. Jill’s belief could serve this ground if she came to hold it with even greater justification—by, for example, investigating much more extensively within the house.
21 This terminology comes from Guerrero (2007, 69).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
370 Alexander A. Guerrero If we endorse View One, then we should endorse (4a). If we endorse View Two, then we should endorse (4b). If we accept both View One and View Two, then we should accept both (4a) and (4b). And these two views are not incompatible. In the foregoing, I hope to have made it plausible that in responding to Jack and Jill cases, we should endorse at least one of these two views, and so we should accept at least one of (4a) or (4b). It is outside the scope of this chapter to argue decisively for one or both of these views but let me at least say that View Two should be familiar. As I noted in previous work: [w]e might think that legal rules of evidence are based on similar considerations— there might be cases in which an agent is justified in believing p as a result of hearsay, but we don’t think that a belief formed on these grounds should be used as grounds on which to punish someone. (Guerrero 2007, 70)
In general, we will require that beliefs be held with greater confidence and with greater evidential support if those beliefs are going to be the basis on which we take morally significant actions—assuming that not taking those actions is a clearly morally permissible option. The slogan here is something like: scope of use should be sensitive to strength of justification. As one’s justification for belief increases, one can use that belief for more purposes, and in particular, for more morally serious purposes, in terms of deliberation and action. There is a question of whether we should think knowledge is different than justified belief, or whether it represents something of a special case. Perhaps knowledge requires a level of justification such that knowledge is always fully actionable, always capable of serving as a rational justifying ground for actions, dependent on the known proposition. This is certainly what is suggested by those defending the Knowledge-Action link and similar claims. In that case, the level of justification for knowledge would have to be quite high, at a level of justification such that it permits full actionability, across all contexts, even morally serious ones. If that is so, then it would suggest that neither Jack nor Jill have sufficient justification to constitute knowledge in the abandoned house case, since it seems that neither of them is permitted to perform actions such as destroying the house. Another possibility is that although we are always permitted to act on what we know, and to take any actions at all that are dependent on the known proposition, this is true only in the sense of objective permissibility, rather than subjective permissibility. The thought here relies on the idea that individual agents may not always know that they know the things they in fact know.22 Let’s say there is some 22 We do not always know that we know the things we know, and any account of knowledge should be able to accommodate that fact. The reason I think this is that even though we will be aware of our evidence and corresponding justification for some proposition, with that evidence and corresponding justification rising to a level sufficient for knowledge, we still might not know whether the proposition supported by the evidence is true. Some accounts of knowledge will rule out this possibility, or come
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Consent 371 proposition p, such that it is the case that if p is true, then a number of “p-dependent” actions are permissible, and if p is false, then those p-dependent actions are impermissible. Trivially, then, if knowledge of a proposition requires that the proposition be true, then knowledge of p entails that all p-dependent actions are permissible. But that is permissible in an objective permissibility sense: permis sible given the objective facts of the world. There may still be an issue of blame worthiness or the related idea of moral culpability or subjective permissibility, as distinct from objective moral rightness or objective permissibility. Here’s why: the agent may not know that they know—they may not even believe that they know. And they may have inadequate justification for the belief that they know that p. So, I might be in a situation where I in fact know that p, but it would still be morally blameworthy for me to act as if I know that p, because I do not know that I know that p. I might not even believe that I know that p. Imagine the person who, pressing Jill, asks, “But do you know that no one is in the house?” Even if Jill has the evidence that Jack has, and even if in fact this evidence is sufficient to justify a belief that the house is empty, and even if it is true that the house is empty, she still might pause, and might well say “No.” This could be because the bar for knowledge is high, as contemplated earlier. But it could also be that there is a difference between knowing that p, and knowing that one knows that p, or even believing that one knows that p. (This may depend, of course, on the extent to which one opts for an internalist or externalist conception of knowledge—among other things.) In certain morally significant stakes cases in which we lack this kind of introspective awareness of the fact of our knowledge, we should not act (at least if we know that refraining from acting is morally permissible, and we know that we know this), even if we in fact do know the proposition in question. If we act in such a case, even if we in fact do know the proposition in question, we act in a way that is morally blameworthy, although we do nothing that is close to doing so, by requiring so much in the way of evidence and justification. Consider, for instance, the idealized, strong conception of knowledge discussed by Hintikka: Suppose we say that evidence for a proposition, P, is conclusive iff it is so strong that, once one discovers it, further inquiry cannot give one reason to stop believing P. The concept of knowledge used by many philosophers seems to be a strong one on which one knows P only if one’s evidence for P is conclusive in this sense. It is plausible that the KK principle holds for this strong concept of knowledge. (1970, 145–6) But this is, as Hintikka acknowledges, an unrealistically strong conception of knowledge. And even conclusive evidence in this sense is not enough to guarantee truth. Even leaving that aside, it is plausible that we know many things for which we do not have conclusive evidence in this sense. Our ordinary concept of knowledge is not like this strong one. Still, many have defended the so-called “KK principle” and luminosity about knowledge, where the KK principle says that, for any proposition p, if one knows that p, then one knows that one knows it. For a variety of attacks on the KK principle, see Alston (1980); Feldman (1981); and Williamson (2000). For recent defenses of that principle, see Stalnaker (2015); and Greco (2014). It is outside the scope of this chapter to engage this question fully. What I say here can be interpreted conditionally: if it is possible to know that p without also knowing that one knows that p, then an agent might be blameworthy for acting as if she knows that p, even when she does in fact know that p.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
372 Alexander A. Guerrero objectively impermissible. We destroy the house, and no one is harmed, because the house was empty. But we took a risk, from our subjective vantage point. And we were aware that we might be taking a risk, given that we did not believe that we knew the house to be empty. And it was not a risk that we were compelled by other morally significant factors to take. And all of that remains true, even if we in fact knew that the house was empty. In summary, then, we might accept View One: Moral Encroachment, and then we should endorse (4a). Or we might reject Moral Encroachment, but endorse View Two: Scope of Use Tied to Epistemic Justification, and then we should endorse (4b). If we accept both View One and View Two, then we should accept both (4a) and (4b). If we accept View One, then Jack knows, and Jill doesn’t. If we accept View Two, then perhaps Jack and Jill both know, or neither of them do, but either way, neither can take morally significant actions based on their beliefs, given their level of justification (whether sufficient to constitute knowledge or not). I hope to have made the case that we should accept either View One or View Two (or both), and so should accept either (4a) or (4b) or both. Where does that leave us regarding the epistemology of consent?
2.2 Moral Stakes and the Epistemology of Consent The foregoing discussion suggests that moral stakes matter, either by directly affecting whether a person holds a justified belief or knows some proposition, or by affecting whether a person can permissibly act as if they justifiably believe or know some proposition, whether they can act based on a justified belief or know ledge. Connecting these conclusions to the general argument, we get the next several premises: (5) Some cases in which consent is important are, as a result, cases with high moral stakes, and so B must possess more/stronger evidence (the result of perhaps correspondingly more investigation) in order for B (i) to justifi ably believe that A consents or (ii) to non-culpably act as if A consents. (6) In particular, cases for which sexual or medical consent are important are cases in which there are high moral stakes. (7) In cases in which B is contemplating an action which is permissible only if A consents to states of affairs involving sexual activity or medical intervention, B must possess more/stronger evidence (perhaps requiring greater investigation) (i) to justifiably believe that A consents or (ii) to non-culpably act as if A consents. Premise (5) should be uncontroversial, if we have taken on board either (4a) or (4b) or both. (Whether both disjuncts hold depends on whether (4a) and (4b) are
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Consent 373 both true.) And (6) is uncontroversial, even when left relatively vague as to what “importance” involves. Here is one way of understanding importance: sexual and medical consent are important in that engaging in sexual activity with or medical interventions on someone who does not consent to those states of affairs constitutes a serious moral wrong. There might be other cases in which not getting someone’s consent is also wrong, but not seriously so. Perhaps, for example, I should not borrow your coffee mug without your consent. Or perhaps I should not take your picture without your consent. These are less important, morally. If we accept the argument so far, then, we are left with (7), which requires more, epistemically, of those who are taking certain actions that require the consent of another person.23 Exactly what “more” a person must do, or what better/ stronger evidence they must have, will be hard to state in a general way. The suggestion I want to make here is that it is plausible that this is what makes “affi rma tive consent” a more normatively appropriate standard with respect to sexual consent: an epistemic account about when we can justifiably believe that others have consented to engaging in sexual activity, or about when we can permissibly act on such a belief, not a metaphysical claim about when there is consent. In this sense, proponents of affirmative consent may be advocating for something importantly new in the norms regarding sex and consent. But what they are advocating for is a higher bar for what counts as enough evidence to justifiably believe that others have consented to engage in sexual activity, or when it is permissible to act on such a belief, not a new metaphysical view about consent.24 Consider the difficulties identified in section 1 about coming to know the mental states of others. On the picture of the metaphysics of consent defended above, a person consents to a state of affairs, SA, only if the person has a particular mental state: namely, an attitude of affirmative endorsement to SA. The precise details of the attitude of endorsement itself can be different in its nature and connection to other mental states and attitudes, depending on the state of affairs towards which it is directed. The attitude of endorsement will have a particular object, a particular state of affairs, but the precise boundary of that object can be unsettled and/or difficult to discern. And the attitude of endorsement can change over time, be implicit as well as explicit, and plausibly comes in degrees. This makes it genuinely hard to know whether we have another person’s consent—knowing whether a person consents to some state of affairs, the robustness of that consent, whether the person continues to consent to that state of affairs over time, and what precisely it is to which the person consents.
23 Although I do not follow Tom Dougherty in settling on the performative, communication- requiring view of consent or seeing it as perhaps motivated by concerns of the sort mentioned here, Dougherty agrees with the broad idea that as the stakes matter more, one’s evidence has to be better (clearer, stronger, less ambiguous, etc.) (Dougherty 2015, 248). 24 Thanks to Olivia Odoffin for pressing this point.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
374 Alexander A. Guerrero It suggests practices that involve regular efforts made to ‘check in’ and communicate thoughtfully and carefully with others about what is happening, what they would like to happen, whether it falls within the scope of what they endorse, whether they still possess the relevant attitude of endorsement, and so on.25 These will vary depending on the context, of course. In the medical context, efforts have been made to get “informed consent” which focuses on making sure that the person actually has an attitude of endorsement towards the states of affairs that will, or might, obtain during and after the proposed medical intervention, or towards the states of affairs in which certain actions are taken, understanding the attend ant risks. With respect to “affirmative consent” standards in the domain of consent to sexual activity, these should not be understood as requiring an explicit question and an explicit affirmative verbal “Yes” to that question, although they have been caricatured that way in the popular media.26 Instead, “affirmative” here is captured by the idea offered in section 1, in which consent is a robust, affirmative pro-attitude towards some state of affairs. Some have offered the idea of “enthusiastic” consent as capturing the relevant idea. The issue then becomes an epistemological one: How can we know when someone has this attitude of robust endorsement towards some state of affairs, such as having sex? What kind of evidence do we need? Given the foregoing discussion, what constitutes adequately strong evidence? We can understand that the arguments in favor of affirmative consent are an attempt to shift how people think about what constitutes evidence of consent, if we accept the idea that consent is this kind of enthusiastic, robust attitude of endorsement, rather than either just an explicit verbal “Yes” or the absence of a verbal “No.” Consider, for example, what the popular writer Tara Culp-Ressler says about affirmative consent standards:
25 For an excellent recent discussion of issues relating to the pragmatics and ethics of such communication, see Kukla (2018). 26 See, for instance, the discussion in Amanda Hess’s clarificatory piece about California bill SB-967 Student Safety: Sexual Assault, which as of 2015 is now the law of California. She writes: The California legislature is weighing a bill that would require college students to secure “affi rmative consent” from their partners at every stage of sexual activity . . . You may have heard of this bill as the one that would require students to draft up a written sex contract before bed or constantly proclaim “yes, yes, yes!” at every slight readjustment, thereby prac tically redefining most sex as rape. The Fresno Bee editorial board interpreted the bill to mean that “ ‘yes’ only means ‘yes’ if it is said aloud.” The Daily Californian, the independent student newspaper of UC–Berkeley, also claimed that affirmative consent is necessarily verbal. RH Reality Check advanced the game to approvingly say that affirmative consent requires “a verbal or written yes.” If consensual sex entailed that level of consent, millions of couples would be unsuspectingly raping one another every night of the week. But the bill doesn’t actually require those things. It calls for “an affirmative, unambiguous, and conscious decision by each participant to engage in mutually agreed-upon sexual activity.” (Hess 2014) Hess helpfully points out that all kinds of non-verbal cues can, in principle, constitute sufficient evidence that a person affirmatively consents.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Consent 375 The current societal script on sex assumes that passivity and silence—essentially, the “lack of a no”— means it’s okay to proceed. That’s on top of the fact that male sexuality has been socially defined as aggressive, something that can result in men feeling entitled to sex, while women have been taught that sex is something that simply happens to them rather than something they’re an active participant in. It’s not hard to imagine how couples end up in ambiguous situations where one partner is not exactly comfortable with going forward, but also not exactly comfortable saying no. Under an affirmative consent standard, on the other hand, both partners are required to pay more attention to whether they’re feeling enthusiastic about the sexual experience they’re having. There aren’t any assumptions about where the sexual encounter is going or whether both people are already on the same page. At its very basic level, this is the opposite of killing the mood— it’s about making sure the person with whom you’re about to have sex is excited about having sex with you. (Culp-Ressler 2014)
It is worth pausing to note the epistemic notes being sounded here. Passivity and silence do not constitute evidence of consent, because consent is not about the absence of an explicit “No.” Instead, both partners should be attentive both to their own attitudes towards what is occurring, and to whether one’s partner has an attitude of affirmative endorsement towards what is happening. This might require explicit verbal communication, but it need not. Importantly, in discerning whether another person consents to some state of affairs, one is required to use one’s total evidence regarding that question. It is epistemically inappropriate to fixate, for example, on a single speech act while ignoring extensive and clear non-verbal cues, body language, relevant personal history, and so on.27 This is just as true in the medical context. It would be inappropriate for a doctor to see a signature on a consent form as sufficient evidence of consent if the patient is exhibiting clear anxiety, confusion, and lack of comprehension regarding the details of what was being consented to, or if the patient has adamantly refused to consent to this very procedure for months, and there is no explanation of the apparent change of heart, etc. In both kinds of situations, it is incumbent on the person seeking consent to probe further, to investigate what is going on with the other person, to attempt to resolve or explain apparent contradictions in the evidence the person has, and so on. And this is so because of the kind of action that is being contemplated. The bar is higher for the evidence needed to justifiably believe that a person consents, or to non-culpably act as if the person consents.
27 For a similar point, see Lackey, Chapter 14, this volume. See also Husak and Thomas (2001, 102), in which they defend a “totality of circumstances” approach to assessing whether a belief about consent (even a mistaken belief) is reasonable or not.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
376 Alexander A. Guerrero This is one place where I part ways with, for example, Douglas Husak and George Thomas, who argue that “empirical generalizations about the means by which consent to sex is typically given or withheld should be used to assess the reasonableness of a mistake about consent.”28 My thought here is that broad empirical generalizations, although fine for forming some kinds of beliefs, will not be enough in these cases. As Husak and Thomas note, these kinds of discussions are often mired with potential for miscommunication as both men and women attempt to navigate interpersonal cues, sexist and patriarchal background social pressures and expectations to behave and perform in certain ways, and much else. All of this affects communication. They suggest that because of this, some mistaken beliefs regarding consent will be reasonable because of the way in which men (in particular) might have been socialized to read and understand verbal and non-verbal behavior. Their view is that generalizations about how consent is viewed by men and women should inform judgments about reasonableness or justifiability of mistake. They don’t commit to specific claims about which generalizations will be used to show reasonableness, but they do say things of this sort: It seems likely that many men who use minor force to overcome resistance at the initial stage believe that consent to penetration will be forthcoming (even if the consent is to unwanted sex). It is difficult to imagine how minor force can accomplish the rape of a woman who resists . . . Perhaps some women become passive and signal their non-consent only weakly, if at all, once their initial resistance is ignored. In these cases, the man who used minor force at an earlier point could believe that, despite her lack of cooperation, the woman now wants to proceed—or is at least willing to proceed. Whether this belief is reasonable, of course, is a different question. (Husak and Thomas 2001, 100)
Although they don’t answer that question, they do suggest that the way to answer it will be to look at what is typically or generally true of men and women in these situations. That is troubling if many men have been conditioned in various sexist and patriarchal ways; it is also troubling if whatever is typical for the majority in terms of communication comes to govern, regardless of the specifics present in a particular case. My view is that it is imperative on people to gather significantly more in the way of evidence in the face of any kind of doubt, or even uncertainty, about the presence of consent. Many of the cases they suggest might end up as reasonable mistakes are ones in which it is implausible that the individuals involved actually investigated the question of consent enough to count as justified, or to act on their belief, given what I have argued above. And that remains so even if such
28 Husak and Thomas (2001, 88).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Consent 377 mistakes are currently statistically common, due to sexist practices, sexual miseducation, and other social norms. (It is worth stressing that morally blame worthy wrongdoing can be statistically common and has been—at least in some domains—for all of human history.) Focusing on broad empirical generalizations regarding how people express their consent in some domains, even if combined with one’s personal history with such expressions, will typically not be enough to yield reasonable or justifiable belief about consent—not without close attention to the evidence one is encountering in the specific case. Husak and Thomas are worried that a test is needed for when a mistaken belief is reasonable, and their test—whether a reasonable person (in some cases, really a reasonable man) in those circumstances might believe that consent was present, where that is to be assessed regarding what is statistically common or normal—is the only one on offer. A better test might be: Would a reasonable person (or, perhaps better, a reasonable woman) believe that consent was being unambiguously given under these factual circumstances? And that can be shown, perhaps, based on a totality of the evidence test, including evidence about what are normal ways of registering unambiguous consent. This might seem to change what is needed, metaphysically, before certain actions are morally permissible: shifting from consent to unambiguous consent. But to reiterate the central theme of the chapter, the better way of understanding this is as a needed epistemic corrective, given the warped social dynamics that exist in sexual communication and interpersonal interactions, particularly between men and women. If people investigate only for evidence of consent—or really, something that might better be described as evidence that is compatible with consent—many will come to have unjustified, mistaken beliefs about whether real, valid, normatively transformative consent is present. When they act on those beliefs, they cause harm, they act in ways for which they are morally blameworthy, and they engage in unexcused consent violations. The underlying aim remains making sure that consent is present; the route to achieve that aim is to err on the side of caution, in terms of how much evidence one gathers before acting. Don’t stop investigating upon coming to believe that one’s evidence is compatible with the person consenting, or even with it being likely that they consent. If people look instead for unambiguous consent, this makes it more likely that their beliefs about consent will be justified and correct.
2.3 Morality, Law, and Mens Rea On this picture, then, there are many things that one can do that are morally blameworthy. In this section, I will set out the different types of actions that can be morally blameworthy, and the way in which an individual’s beliefs or mental states are relevant. After discussing the moral case, I then briefly consider the way
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
378 Alexander A. Guerrero in which these moral distinctions might be—and perhaps ought to be— reflected in law. These first two kinds of morally blameworthy action are familiar, and it is uncontroversial that agents are typically morally blameworthy in both cases (assuming that there were no other compelling moral reasons that required X to φ, despite the lack of consent).
(1) (2) (3) (4)
Knowing consent violation: X performed an action, φ. X’s φ-ing is only morally permissible with W’s consent. W did not consent to X’s φ-ing. X knew that W did not consent to X’s φ-ing.
(1) (2) (3) (4)
Reckless indifference to consent violation: X performed an action, φ. X’s φ-ing is only morally permissible with W’s consent. W did not consent to X’s φ-ing. X was recklessly indifferent to whether W consented to X’s φ-ing.
In these cases, W did not consent, and X either knew this fact or was recklessly indifferent to this question. One way of understanding “reckless” indifference here is indifference that is itself morally culpable. There might be some cases in which indifference is not morally culpable; those are not the cases covered by (4). And the (2) condition explicitly states that, for actions of this sort, a necessary condition of moral permissibility (although by no means a sufficient condition!) is that the person in question consents to the action. What I have argued above, however, suggests that there are other ways in which one can act that would also be morally blameworthy. Consider:
( 1) (2) (3) (4)
Unexcused consent violation: X performed an action, φ. X’s φ-ing is only morally permissible with W’s consent. W did not consent to X’s φ-ing. (a) X mistakenly believed that W consented to X’s φ-ing, but (b) X’s belief about W’s consent was not justified.
Here, X did believe that W consented, but this belief was not justified, and so X’s action constitutes an unexcused consent violation.29 It is an “unexcused” consent 29 An alternative explanation is to focus on so-called “secondary” duties of investigation. In a recent paper, Tom Dougherty argues that affirmative consent standards might be understood by reference to a duty of due diligence, which includes a duty to “adequately investigate” whether consent is present (if needed). He doesn’t say much specifically about what this might require, but briefly
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Consent 379 violation because if X’s belief was justified—even given the heightened requirements defended earlier—then X might be excused for φ-ing. X might not be morally blameworthy. But there is no such excuse if the belief is unjustified. In these cases, it might be that this belief was grossly negligent—falling below some basic standard of justification for any kind of belief. But the suggestion in this chapter is that, because of the moral significance of the action being contemplated, the bar for what is required for justified belief about consent is higher, so that what might sometimes be enough to believe something non-negligently or reasonably would not be enough in these cases. It seems clear that what is done in cases like this will often be morally blame worthy and morally objectionable. This is clearest in those cases in which (i) X is morally or epistemically blameworthy for having an unjustified belief about W’s consent or in which (ii) X is aware that X’s belief about W’s consent is based on somewhat shaky evidence and in which not acting is clearly a permissible option for X. Consider a case in which X believes that W consents, but knows that he doesn’t know that W consents, and intentionally doesn’t take any of several easy steps to try to get more and better evidence about whether W consents. Imagine further that X’s evidence that W consents is very weak—clearly not enough to justify the belief. If X decides to go ahead and have sex with W anyway under these circumstances, and if W in fact does not consent, it seems that X is morally blameworthy for going ahead and acting anyway. More controversially, there are still more ways in which people can act in ways for which they are morally blameworthy.
( 1) (2) (3) (4)
Objectionable moral risk (reckless indifference): X performed an action, φ. X’s φ-ing is only morally permissible with W’s consent. W did consent to X’s φ-ing. X believed that W did not consent, or X was indifferent regarding W’s consent (not having considered the question). Objectionable moral risk (unjustified belief):
(1) X performed an action, φ. (2) X’s φ-ing is only morally permissible with W’s consent. (3) W did consent to X’s φ-ing. suggests that what is adequate might well vary with context and stakes. The wrong in some of these cases then might be, on his terminology, the failure to live up to this duty of due diligence, which would, on his view, explain the permissibility of holding people culpable in cases in which the advocates of affirmative consent policies think there is culpability. That seems to mischaracterize the wrong, suggesting that what was morally objectionable was failing to live up to this “secondary” investigative duty, rather than violating a primary duty not to act in this way given the lack of consent, or a primary duty not to act in this way given one’s poor epistemic position regarding the question of consent. But a full discussion of this route is outside the scope of this chapter (Dougherty 2018).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
380 Alexander A. Guerrero (4) (a) X correctly believed that W consented to X’s φ-ing, but (b) X’s belief about W’s consent was not justified. These cases are different than those above because they are cases in which W did in fact consent. The objectionable conduct here stems not from there being an instance, of, say, non-consensual sex, but because of X’s taking an objectionable moral risk with respect to whether W consented or not, given X’s epistemic standing with respect to the question of whether W consented. Many think that cases of this sort, where one takes an objectionable moral risk, but the risked outcome does not in fact materialize, are less morally objectionable—even if whether the bad outcome materialized or not is not under the control of the agent. I don’t want to wade into the controversy regarding the existence of moral luck in terms of outcomes, but I do want to stress that these cases may involve morally blame worthy conduct, even if it may not be as morally bad as the previous kinds of cases. One reason that this conduct is morally blameworthy is that it often involves taking a significant and unnecessary moral risk, given one’s epistemic situation.30 For example, it is plausible that if X knows that not φ-ing is morally permissible, and if φ-ing is only morally permissible if W consents to state of affairs SA, then X ought not to φ unless X knows, or at least justifiably believes, that W consents to SA.31 If X fails to even consider the question of consent, or considers it only inadequately, and goes ahead and acts anyway, this is a way of disrespecting that person, acting without sufficient regard for that person’s standing as a moral agent—it is just a different way of doing this than acting without that person’s consent. There is more that might be said to ground the claim that all five of these cat egories of action constitute what is often, or typically, morally blameworthy conduct—conduct that is blameworthy because of the pairing of a certain kind of action with a certain kind of epistemic situation regarding the consent of another. Some actions covered by these descriptions are already illegal. In some states, for an act to constitute rape still requires the use of force, not just the fact of non-consent, even if known. But in many states, there has been a shift to including as rape those actions in which one person has sex with another while knowing that the person does not consent. But many of these actions—including the unexcused consent violations and both kinds of objectionable moral risk actions—are not
30 I discuss this kind of case at length in Guerrero (2007). 31 One significant difference between sexual consent cases and medical consent cases is that, in the former cases, it is almost always completely morally fine to refrain from acting—there is nothing, or very little, at stake, morally speaking, in refraining from engaging in sexual activity on a particular occasion. That is not always true in the medical context. There may often be quite a bit at stake in the decision to refrain from offering some medical intervention. That changes the calculus, since it means that both acting and not acting are high stakes, morally speaking.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Consent 381 currently defined as illegal. It is too large a project to argue that these all should be criminalized, nor to argue for some specific level of punishment in response to them. I do want to suggest, however, that at least one argument against making them illegal is unsound: the argument that suggests they ought not to be illegal because the conduct included is not morally blameworthy, and moral blameworthiness is a necessary condition for permissible criminalization. Kimberly Ferzan, for example, is worried about affirmative consent policies on the grounds that they might expand criminal law to cover cases in which there is no moral blameworthiness. She writes: If . . .potential defendants are punished in order to cause social change or to protect women by creating prophylactic rules, then we are punishing individuals who are nonculpable as to what we really care about (nonconsensual sex) in order to accomplish our goal (better and more accurate communication about consent). We are punishing the morally innocent. We should pause before punishing the innocent for the collective good. (Ferzan 2016, 421)
Ferzan is right that introducing certain kinds of requirements of “affirmative consent” might criminalize some morally innocent behavior, if these requirements mandate verbal consent. As I have suggested, we needn’t require anything like verbal expressions of consent in every case. But if what I’ve argued so far is correct, then Ferzan is wrong to suggest that the only thing we do or should care about is culpability regarding “nonconsensual sex.” Although it would be true that we might punish some individuals who are non-culpable as to non-consensual sex if we made illegal the conduct described in the “objectionable moral risk” cases, we would not be punishing the “morally innocent.” Those individuals would be engaging in morally blameworthy conduct, much like the criminally reckless or intoxicated driver who doesn’t hit anyone, or (in the reckless indifference versions) like those who attempt but do not succeed in perpetrating crimes such as homicide and rape.32 We really do, and should, care about this kind of conduct, too. The foregoing has focused on moral claims, not legal ones. But the analysis might be extended to the law of rape and sexual assault as well—and, indeed, many who focus on affirmative consent doctrines think that they should be doctrines about the law and policy of rape and sexual assault. Current law regarding rape and sexual assault is very far from aligning with what many believe to be the moral facts concerning these matters. For a long time, rape was defined in the criminal law as requiring the use of force, so that the 32 In almost all jurisdictions, if attempted rape is criminalized, it is not a felony offense, and it usually is only even a possible offense if it involves sexual interaction with people who are not legally capable of consent: minors, currently incarcerated people, etc.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
382 Alexander A. Guerrero actus reus of rape included the use of force. Reform efforts in the 1980s resulted in many jurisdictions including an alternative actus reus of sex without consent, or “rape by non-consent,” but in most of these jurisdictions this alternative is limited to circumstances in which one of the parties, by legal definition, cannot consent due to their age, employment relationship, or state of intoxication.33 Even for those proportionally few cases in which the actus reus can be this kind of legally defined non-consensual sex, the mens rea of rape is still such that it limits rape to cases in which the defendant (i) intends or knows that he is engaging in nonconsensual sex; or (ii) is recklessly indifferent as to the victim’s lack of consent. If the defendant mistakenly believes that the victim consents, then the element is not satisfied, and acquittal is required—even if the mistaken belief is not reasonable or justified.34 Thus, under current law in Anglo-American jurisdictions, even unexcused consent violations (in the above terminology) are not criminalized. Nor, of course, are either of the two kinds of objectionable moral risk cases. It is outside the scope of this chapter to make the case for correcting this, but here let me just join Marcia Baron, Helen Power, and Kari Hong, among others, in suggesting that substantial reform in this area might well be appropriate. Power, in particular, argues for including what I’ve called “unexcused consent violations” under the law of rape, as instances of third-degree “negligent” rape—cases in which a person is procedurally or substantively negligent in forming a belief that the person with whom they are having sex consents to doing so.35 Neither Power nor Hong go as far as I might, in that both still require that the actus reus in question actually include that one engages in an act of non-consensual sex with another person. I hope to have both suggested that there is a substantial case for moral blame worthiness in these cases—such that might undergird legal criminalization and punishment—and suggested that we should perhaps see the bar for non-negligent, justified belief formation about consent as being higher than is commonly thought. Of course, much more would need to be said in full defense of further extending the criminal law, and in the details of that possible extension.
33 For discussion and review of this area of law, see Pineau (1989); Archard (1999); Baron (2001); Power (2003); Hong (2018). 34 For extended discussion of the law in the United States, see Hong (2018, 270–89). For similar discussion of the law in the United Kingdom, see Power (2003, 379–87). Some have suggested that having an unjustified or unreasonable belief regarding consent can itself constitute recklessness, or that it should constitute recklessness. Power, discussing Antony Duff ’s Intention, Agency, and Criminal Liability (1990), suggests that his view is that “the individual who unreasonably believes that the victim consents is reckless, in the sense that the belief itself demonstrates practical indifference” (Power 2003, 391). But this is not the law in either the United States or the United Kingdom. 35 Power (2003, 381). Hong defends a “malice” standard for the mens rea of rape, in which any of knowledge of non-consent, intent to have sex with someone who doesn’t consent, or reckless indifference to consent could satisfy the mental element for rape. She explicitly states that negligence in forming a belief about the other’s consent would not suffice on her standard (Hong 2018, 303).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Consent 383
3. Conclusion In this chapter, I have argued for a certain view about one necessary component of consent: the presence of a mental state of robust, affirmative endorsement toward a state of affairs. I have argued that, given this component, we all face hard epistemological questions about when others consent. Additionally, these questions are morally very significant—they arise in high-stakes moral contexts. These two features give rise to a number of important epistemological issues. I have argued that we should accept either moral encroachment or the view that the scope of use of our beliefs and even knowledge may be sensitive to our level of justification for those beliefs. In either case, it will mean that agents must do more, epistemically speaking, before they can justifiably believe that another person consents, or non-culpably act as if another person consents. We may, for example, need to look beyond the words that people say, to pay attention to non-verbal cues and personal history, and to inquire and investigate further when we encounter ambiguity or contradictory signals. On this picture, “affirmative consent” standards can be understood as responding to, and even articulating, these epistemological concerns, rather than as offering a new metaphysics of what constitutes consent. If we accept this picture, this has implications both for how we ought to view the morality of various actions, given the agent’s mental states, and for what we ought to think about the law governing violations of consent.
References Alexander, Larry (1996). “The Moral Magic of Consent (II).” Legal Theory 2: 165–74. Alexander, Larry (2014). “The Ontology of Consent.” Analytic Philosophy 55(1): 102–13. Alston, William (1980). “Level-Confusions in Epistemology.” Midwest Studies in Philosophy 5(1): 135–50. Archard, David (1999). “The Mens Rea of Rape: Reasonableness and Culpable Mistakes,” in Keith Burgess-Jackson (ed.), A Most Detestable Crime: New Philosophical Essays on Rape. Oxford: Oxford University Press, 213–29. Baron, Marcia (2001). “I Thought She Consented.” Philosophical Issues 11: 1–32. Basu, Rima (2018). “Can Beliefs Be Wrong?” Philosophical Topics 46(1): 1–17. Basu, Rima and Mark Schroeder (2019). “Doxastic Wronging,” in Brian Kim and Matthew McGrath (eds), Pragmatic Encroachment in Epistemology. Abingdon: Routledge, 181–205. Brown, Jessica (2008). “Knowledge and Practical Reason.” Philosophy Compass 3: 1135–1152.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
384 Alexander A. Guerrero Culp-Ressler, Tara (2014). “What ‘Affirmative Consent’ Really Means.” Think Progress (June 24, 2014), https://thinkprogress.org/what_affirmative_consent_actually_ means_ca665b32b388. Dougherty, Tom (2018). “Affirmative Consent and Due Diligence.” Philosophy and Public Affairs 46(1): 90–112. Dougherty, Tom (2015). “ ‘Yes Means Yes’: Consent as Communication.” Philosophy and Public Affairs 43(3): 224–53. Duff, Antony (1990). Intention, Agency and Criminal Liability. Oxford: Blackwell. Fantl, Jeremy and Matthew McGrath (2002). “Evidence, Pragmatics, and Justification.” Philosophical Review 111(1): 67–94. Fantl, Jeremy and Matthew McGrath (2009). Knowledge in an Uncertain World. Oxford: Oxford University Press. Fantl, Jeremy and Matthew McGrath (2007). “On Pragmatic Encroachment in Epistemology.” Philosophy and Phenomenological Research 75(3): 558–89. Feldman, Richard (1981). “Fallibilism and Knowing that One Knows.” Philosophical Review 90(2): 266–82. Ferzan, Kimberley (2016). “Consent, Culpability, and the Law of Rape.” Ohio State Journal of Criminal Law 13: 397–439. Fritz, James (2017). “Pragmatic Encroachment and Moral Encroachment.” Pacific Philosophy Quarterly 98(1): 643–61. Gardiner, Georgi (2018). “Evidentialism and Moral Encroachment,” in Kevin McCain (ed.), Believing in Accordance with the Evidence: New Essays in Evidentialism. New York: Springer, 169–98. Greco, David (2014). “Could KK Be Ok?” Journal of Philosophy 111(4): 169–97. Guerrero, Alexander A. (2007). “Don’t Know, Don’t Kill: Moral Ignorance, Culpability, and Caution.” Philosophical Studies 136: 59–97. Guerrero, Alexander A. (2017). “Political Functionalism and the Importance of Social Facts,” in Kevin Vallier and Michael Weber (eds), Political Utopias. Oxford: Oxford University Press, 127–50. Hawthorne, John and Jason Stanley (2008). “Knowledge and Action.” Journal of Philosophy 105(10): 571–90. Hess, Amanda (2014). “ ‘No Means No’ Isn’t Enough. We Need affirmative Consent Laws to Curb Sexual Assault.” Slate (June 16, 2014), www.slate.com/blogs/xx-factor/ 2014/06/16/affiratie_consent_california_weighs_a_bill_that_would_move_the_ sexual.html. Hintikka, Jaakko. (1970). “ ‘Knowing that One Knows’ Reviewed.” Synthese 21: 141–62. Hong, Kari (2018). “A New Mens Rea for Rape: More Convictions and Less Punishment.” American Criminal Law Review 55: 259–332. Hurd, Heidi (1996). “The Moral Magic of Consent.” Legal Theory 2: 121–46.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Consent 385 Husak, Douglas and George Thomas (2001). “Rapes without Rapists: Consent and Reasonable Mistake.” Philosophical Issues 11(1): 86–117. Ichikawa, Jonathan J. and Matthias Steup (2016). “The Analysis of Knowledge,” in Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy, https://plato.stanford. edu/archives/win2016/entries/knowledge_analysis. Kukla, Rebecca (2018). “That’s What She Said: The Language of Sexual Negotiation.” Ethics 129(1): 70–97. Machery, Edouard (2016). “De-Freuding Implicit Attitudes,” in Michael Brownstein and Jennifer Saul (eds), Implicit Bias and Philosophy, Vol. 1: Metaphysics and Epistemology. Oxford: Oxford University Press, 104–29. Miller, Franklin G. and Alan Wertheimer (2010). “Preface to a Theory of Consent Transactions: Beyond Valid Consent,” in Franklin G. Miller and Alan Wertheimer (eds), The Ethics of Consent. Oxford: Oxford University Press, 79–105. Moss, Sarah (2018). Probabilistic Knowledge. Oxford: Oxford University Press. Owens, David (2011). “The Possibility of Consent.” Ratio 24(4): 402–21. Pace, Michael (2011). “The Epistemic Value of Moral Considerations: Justification, Moral Encroachment, and James’ ‘Will to Believe’. ” Noûs 45(2): 239–68. Pineau, Louis (1989). “Date Rape: A Feminist Analysis.” Law and Philosophy 8: 217–43. Power, Helen (2003). “Towards a Redefinition of the Mens Rea of Rape.” Oxford Journal of Legal Studies 23: 379–404. Spaulding, Shannon (2016). “Mind Misreading.” Philosophical Issues 26(1): 422–40. Stalnaker, Robert (2015). “Luminosity and the KK Thesis,” in Sandford Goldberg (ed.), Externalism, Self-Knowledge, and Skepticism: New Essays. Cambridge: Cambridge University Press, 19–40. Wertheimer, Alan (2003). Consent to Sexual Relations. Cambridge: Cambridge University Press. Westen, Peter (2004). The Logic of Consent: The Diversity and Deceptiveness of Consent as a Defence to Criminal Conduct. Burlington, VT: Ashgate Publishing. Williamson, Timothy (2000). Knowledge and Its Limits. Oxford: Oxford University Press.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
PART 8
E PIST E MOLO G Y A ND T HE IN T ER NET
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 19/03/21, SPi
16
The Internet and Epistemic Agency Hanna Gunn and Michael Patrick Lynch
1. Introduction For most people, the internet is now the most dominant source of socially useful knowledge. Its widespread use has made knowledge more accessible, more widely distributed, and more commonly produced. But the internet is also widely seen—and not just by philosophers—as raising a number of distinct epistemological problems. Some of those problems concern the metaphysics of knowledge—the extent to which knowledge via the internet is understood as outsourced, or even extended, knowledge. Others concern the type of knowledge the internet can give us—whether, for example, the knowledge we gain by using our digital devices is a kind of testimonial knowledge. In this chapter, we will focus on a third issue: how our uses of the internet to gain information affect our epistemic agency—or our capacity to take responsibility for our own epistemically relevant mental states and our wider contributions to our epistemic environment. In the early days of digital technology, the internet was generally seen as increasing our epistemic agency for the simple reason that it made information (and presumably knowledge) more accessible and widespread. But in recent years, such optimism has been tempered due to the rise of fake news, massive amounts of misinformation online, and the average consumer’s seeming credulity with regard to what they read on social media. The chapter will outline these objections and critically examine them, arguing that both our epistemic agency and our ability to responsibly exercise such agency can be undermined by some uses of the internet, even as those same uses increase agency in other ways.
2. Epistemic Agency An “agent” is traditionally thought to be someone (or something) capable of taking responsibility for and having some degree over the authorship of their actions. But even more fundamentally, an agent is someone who can affect change, and an “epistemic agent”, on this reading, is someone capable of bringing about Hanna Gunn and Michael Patrick Lynch, The Internet and Epistemic Agency In: Applied Epistemology. Edited by: Jennifer Lackey, Oxford University Press (2021). © Hanna Gunn and Michael Patrick Lynch. DOI: 10.1093/oso/9780198833659.003.0016
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
390 Hanna Gunn and Michael Patrick Lynch specifically epistemic change.1 The ability to bring about epistemic changes is, we take it, fundamental to epistemic life. Epistemic agency, then, is about taking responsibility for the changes that we make within our epistemic lives. Where can we bring about epistemic changes? We believe it useful to distinguish between a “narrow-scope” and a “wide-scope” understanding of epistemic change; we take it that most work in the literature on epistemic agency has focused only on what we will define as narrow-scope epistemic changes. Narrow-scope epistemic changes are those confined to our internal epistemic states, while wide-scope epistemic changes are those that take place outside of ourselves and within either our interpersonal interactions or our epistemic communities more broadly. Intuitively, being responsible for narrow- and wide-scope changes will pose different obligations and require different behaviours of us as epistemic agents. We begin by discussing this distinction further, and then showing why we think it is a useful one to make when considering the impact of the internet on epistemic agency. We propose that narrow-scope epistemic changes are the kinds of changes brought on through processes of critical reflection on our epistemically relevant internal states (e.g. our beliefs, attitudes, abilities, and commitments). Through these processes of reflection, we may be able to foster within ourselves, for example, more justified and true beliefs. But we may also be able to change our epistemic attitudes—become more conscientious or intellectually humble, for example. When we exercise our epistemic agency narrowly, we take charge of various aspects of our reflective cognitive life.2 Epistemic agency is a concept discussed in the epistemic literature, and typically is often thought of in something like the narrow-scope way that we have just described. The responsibilities often discussed concern ways to minimise the deleterious consequences of our cognitive biases and heuristics and how to employ various epistemic principles for truth seeking. While we take it that epistemic agency does, importantly, involve these narrowscope or internally oriented activities, a significant amount of epistemic activity occurs between persons: collaborative research, participating in conferences, deliberating with others, teaching or learning from our teachers. We will call these wide-scope epistemic changes. Put more generally, wide-scope epistemic changes include contributing to a shared body of knowledge (by adding to or revising it), changing social epistemic norms or practices, proposing new 1 An alternative understanding of epistemic agency can be found in Lynch (2013), where epistemic agency is understood in terms of our commitment to, and our acting from, particular basic epistemic principles. 2 For instance, Kornblith’s account of justified belief as responsible belief (1983, 34). There is an interesting crossover between work on epistemic agency in the narrow-scope sense and research on doxastic voluntarism; see Alston (1988) and Steup (2017). For further discussions of the possibility of responsibility for the beliefs one forms, see McHugh (2013) and Olson (2015). For discussions of the very possibility of epistemic agency in a narrow-scope sense, given issues around doxastic voluntar ism, see Engel (2013) and Setiya (2013).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Internet and Epistemic Agency 391 epistemic methods, or changing someone else’s beliefs. All of these interpersonal epistemic actions or activities are also ways to affect epistemic changes. And so, if we are interested in our ability to bring about epistemic changes more generally, then it is worth considering how we can take responsibility for these wide-scope epistemic changes, i.e. take responsibility for exercising our epistemic agency in collaborative epistemic activities. For the purposes of evaluating the internet and its influence on our epistemic lives, we think it can be particularly valuable to consider wide-scope epistemic changes. Certainly, we can also productively think about the ways that our experience on the internet changes how we go about forming beliefs, and which of our biases the internet might help to alleviate or exacerbate, and so on. Miller and Record (2013) take up exactly these issues and discuss our responsibility for our beliefs formed by the use of internet-mediated information.3 Increasingly, though, we are learning about the ways that instant communication, social media, and ready access to massive databases like Wikipedia are changing how we interact with one another epistemically. For example, the viral nature of (mis)information through platforms like Twitter and the isolated epistemic communities that form on social media networks like Facebook seem to make it far easier for us to cause others to gain new beliefs. As the internet brings us together in new ways, it also alters the ways we participate in interpersonal epistemic activities. As a result, we propose to consider here a reading of epistemic agency that considers our ability to take responsibility for both the narrow-scope and the wide-scope epistemic changes we can bring about online. A very wide-scope understanding of epistemic agency is described by Reider (2016) in the following way: “epistemic agency might be broadly construed as anything that is responsible for the assessments, acquisition, dissemination, and retention of knowledge” (x).4 This definition remains neutral on what exactly epistemic agents are (e.g. individ uals, groups, digital computers), but shares the core theme of a responsibility for epistemic changes in a wide array of epistemic activities. We restrict our discussion of epistemic agents to individuals. In accordance with this, we can describe our goal here as putting a spotlight on some of the ways that we engage with others in acquiring, assessing, disseminating, and otherwise sharing knowledge on the internet, or pursue these activities for our own epistemic ends. We noted earlier that on familiar, narrow-scope understandings of epistemic agency, being responsible involves mediating one’s acquisition of belief or ameliorating bias, etc. When we widen the scope of epi stemic agency to include interpersonal epistemic actions, then we will also widen 3 They argue that when an epistemic agent does not understand the processes of filtering and personalisation used online, the beliefs they form from these sources are less justified. 4 Reider (2016) provides a useful introduction to the subject of social epistemology and epistemic agency. The collection is organised around two anchor articles representing different approaches in social epistemology: Goldberg (2016) and Fuller (2016).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
392 Hanna Gunn and Michael Patrick Lynch the scope of what it takes to be a responsible epistemic agent. We will describe some ways we think the internet changes interactive or social epistemic activ ities—the wide-scope exercise of epistemic agency. The concept of responsible epistemic agency, generally speaking, provides some normative conditions on making epistemic changes: for example, how one ought to update beliefs in light of new evidence. We propose that responsible epistemic agents are—at least—those who meet their epistemic obligations. And while a full account of responsible epistemic agency is obviously beyond the scope of this chapter (which concerns the impact the internet has on the development and exercise of our agential capacities), we can note that on the inclusive narrow- and wide-scope reading we favor, these obligations will extend beyond believing what is true and avoiding believing what is false. In particular, we believe an account of responsible epistemic agency that attempts to address the wide-scope changes we have suggested will also include at least two further conditions relating to how the agent engages in bringing about epistemic changes. The first is that responsible epistemic agents work not only to develop epi stemic virtues, merit, and capacities that help them to responsibly change their epistemic environment, but also to develop the capacities that enable them to recognise and respect these epistemic traits in others. The second is that respon sible epistemic agents treat other epistemic agents with a certain kind of respect that demonstrates a willingness to learn from them. We will motivate these two conditions over the following sections by describing ways that the internet— through features of the web and through particular popular platforms—affects our ability to assess others and to learn from them. We will further develop these two conditions by modelling them on the familiar model of appraisal and recognition respect in ethics but presenting what we take to be a plausible epistemic application of these concepts (Darwall 2006). We sketch this proposal now, but present a fuller picture in section 4.1. Secondary to our goal of evaluating how the internet might be affecting our ability to be responsible epistemic agents then, is to propose that we can productively think about epistemic obligations via the appraisal and recognition model of respect. Our first condition requires that responsible epistemic agents work to appraise the positive epistemic traits in others by, for example, recognising their epistemic merits and treating them with the respect they deserve for having, for instance, knowledge. Responsible epistemic agents, when thinking about wide-scope epi stemic change, also appreciate and take into account the merit and skills of other knowers and treat them accordingly. Our second condition can be understood as requiring that responsible epistemic agents treat one another with an epistemic recognition respect. Epistemic recognition respect requires, at least, avoiding responding to others with epistemic arrogance. We propose that epistemic arrogance involves dismissive behaviour towards other epistemic agents and is a way of excluding them from participation in epistemic life.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Internet and Epistemic Agency 393 Why are these conditions important for assessing how the internet affects our ability to responsibly bring about epistemic change? Consider, for e xample, the paucity of information one has to draw on when looking at online forum comments in order to decide the degree to which a stranger is deserving of appraisal respect. Internet platforms may make it more difficult to appropriately appraise, and thus appropriately respond to, others online. Consider also how long it takes one to discredit and ignore (or scroll past) someone’s contribution to a forum if they have a name or picture that identifies them, for example, as on the opposite side of the political spectrum as oneself. This swift discrediting is an action of denying epistemic recognition respect—people like that aren’t people I have to take seriously. Internet platforms often make it very easy to make such judgments about others by encouraging us to present such identifying features about ourselves, or alternatively, encourage us to sort ourselves into social groups that make it easier to identify in-group and out-group members. In thinking through the ways that our epistemic interactions and activities online help us to seek truth, to develop epistemic merit or skill, to recognise these skills in others, to listen to or learn from others, etc., we can assess some of the ways that the internet is impacting our epistemic lives. We should want to know, then, whether the internet as an epistemic environment helps or hinders our development of epistemic agential capacities and skills. Does it make us more or less likely to engage in research? Does it help us maintain good epistemic policies such as fact-checking or does it make us more likely to “side” with our “team”? Are we subject to greater or lesser quantities of deceptive information online? All of these are possible impingements on our ability to be responsible knowers and on our ability to responsibly bring about epistemic changes in both the narrowand wide-scope senses. In the following sections, we consider ways that the internet can support and undermine our ability to meet the epistemic obligations we have proposed for epistemic agency. We begin with a discussion of how the internet might be expanding our epistemic agency.
3. Has the Internet Expanded Epistemic Agency? While we are interested in this chapter in primarily spelling out threats that our uses of the internet pose to responsible epistemic agency, it is plausible that our life online has expanded our epistemic agency in certain ways as well—that is, that it has, in various ways, increased our ability to responsibly take control of our epistemic lives in both the narrow- and wide-scope senses. Perhaps the single most heralded upside of the changes in informational technology we have been experiencing for the past two decades is the “democratisation”
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
394 Hanna Gunn and Michael Patrick Lynch of knowledge that the internet is sometimes said to provide. One way to understand this claim that the internet has democratised knowledge is to see it as a proposal that the internet expands our basic epistemic agency. There are at least three ways in which this might be said to be true. First, and most obviously, the internet, like the printing press before it, has made bodies of knowledge more widely available. The possibility of massproduced books lowered the price at which knowledge could be bought and sold. Consequently, it brought such knowledge—and the possibility of literacy—to millions of people who lacked access to it previously. Web 2.0 has greatly expanded this process, while also changing both the sheer amount of different kinds of information available and the speed at which that information can be accessed. In this way, then, we are more able to gain knowledge (narrow-scope change) and share knowledge (wide-scope change) with one another. A good example concerns internet access itself. Try Googling, “How many people have access to the Internet?” and sources such as Wikipedia and the International Telecommunications Union in Geneva will tell you that while roughly 94 percent of Sweden has internet access, and 84 percent of Americans, only 2.1 percent of the population of Chad does. Yet the availability of these stat istics are themselves a good example of the sort of information that just a few years ago you would have to go to a large research university to find, or rely on journalists to report. While billions of people continue to have no access to it, millions have immediate access to the sort of information they wouldn’t have had just a decade ago. Epistemic access to knowledge is not only less concentrated, but the ability of one to fulfil one’s epistemic obligations in acquiring and disseminating knowledge is less concentrated. In short: “more information to more people” is one obvious way that the internet can make knowledge—or its acquisition—“more democratic.” Assuming that the information made available is reliably sourced and accurate, then the body of knowledge to which one has access has expanded. And one might plausibly take this to mean that we have expanded both our choices about what to believe and the degree and ease of meeting one’s epistemic obligations. Our ability to quickly verify or falsify simple claims of fact, for example, can make it easier to meet a norm to align one’s belief to the evidence. Thus, increased availability and access would seem to have expanded our epistemic agency in both the narrow- and wide-scope senses. Narrowly, increased access to reliable sources of information enhances our ability to meet our reflective epistemic obligations. Widely, having more access to more information allows agents to more reliably and quickly contribute to shared knowledge communities. Second, and building on this last point, another manner in which the internet arguably expands our epistemic agency by making it more “democratic” is by making knowledge production more inclusive. That is, it allows people not in positions of power to contribute to the production of knowledge. This happens in
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Internet and Epistemic Agency 395 several different ways. Open-source software like Mozilla’s Firefox web browser is one kind of example. When bugs or vulnerabilities turn up in Firefox software, a diverse and widespread community of volunteers work on fixes and plug-ins. Open-source software operates similarly to an online co-op. It is software by the people, for the people. These open-source features include more individuals in the development of tools that serve epistemic ends like disseminating, accessing, and acquiring knowledge.5 Another example of epistemic inclusivity is the growing number of open-access, research-sharing sites. Here people can both access and contribute to the growth of knowledge without necessarily being highly placed or grant-supported researchers. Third, the internet is also used in ways that expand epistemic participation (and inclusivity), and therefore enhance epistemic agency, through the development of challenge-specific prizes.6 Firms like InnoCentive, for example, serve as platforms from which to launch problem-centered, open-access competitions. Such competitions have been useful sources of innovation in scientific research for centuries. The British Crown spurred a huge leap forward in marine navigation in the seventeenth century, for example, by offering a prize for a device that could calculate a ship’s longitude—resulting in the invention of the marine chronometer. Competitions like this work partly because they provide an incentive for “fresh eyes” on the problem. Indeed, researchers Jeppesen and Lakhami, in their 2010 study of InnoCentive, suggested that there is an inverse relationship between a solver’s likelihood of solving a problem and his or her degree of expertise in the field in question. And the same study also found that women significantly outperformed men as problem solvers on the site—despite, and possibly because, they are often more on the edges of the “scientific establishment.” As Brabham writes, this means that, for example, “a biologist may fare better than a chemist would at solving a chemical engineering problem” (2013, 21). That is inclusivity of a very obvious sort. Individuals who have worthwhile contributions to make to these epistemic activities, as indicated by the success of these crowdsourcing initiatives, are able to participate when they would have been excluded prior to this use of the internet. This clearly increases the epistemic agency of those included, and it also allows for their epistemic merits to be recognised in ways that would previously have been ignored. So, there is no doubt that the internet has changed how we distribute, produce, and reveal knowledge, and as a result, expanded our epistemic agency. But using the language of “democratisation” to describe these changes obscures as much as
5 Of course, web browsers also serve many other non-epistemic ends; for instance, they also serve communicative ends through chat clients and email. 6 See Standing and Craig (2018) for a discussion of the ethics of a variety of online crowdsourcing activities.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
396 Hanna Gunn and Michael Patrick Lynch it describes. It ignores the fact that these changes don’t often increase our ability to take responsibility for our epistemic lives; they sometimes undermine it—that is, they can inhibit or prevent us from exercising our epistemic agency in a responsible way. We now move to a discussion of threats to responsible epistemic agency online. We start with a discussion of how various features of the internet seem to support epistemically arrogant behaviour, and how various forms of personalisation of online content can undermine responsible epistemic agency.
4. Threats to Epistemic Agency 4.1 Information Personalisation and Epistemic Arrogance While the internet does make information (and sometimes knowledge) more widely accessible and available, a number of recent studies suggest that this very fact can actually undermine our ability to be responsible epistemic agents, and perhaps undermine epistemic agency itself. Much of what we know (or think we know) now we “Google-know,” in the sense that we acquire information online from websites and social media (Gunn and Lynch 2018; Lynch 2016). The sheer availability of information, and its perceived reliability, has been argued to actually increase people’s confidence in their knowledge and make them think they know more than they do. Thus, some researchers suggest that our ability to access expert opinion so easily and immediately online can cause us to “lose track of our own reliance on it and [distort] how we view our own abilities” (Fisher et al. 2015, 675). This conclusion, also supported by recent work by Fernbach and Sloman (Fernbach et al. 2013; Sloman and Fernbach 2017), in turn indicates that we can overestimate how much we individually know because “externally accessible information is conflated with knowledge ‘in the head’ ” (Fisher et al. 2015, 682). Certain studies, for instance, suggest that searching for explanations on the internet increases the likelihood that we think we know more than we do, even about an unrelated topic—even controlling for factors such as content and time. One way this can impact epistemic agency considered in the wide-scope sense is that it is possible that the sheer accessibility of information itself can lead to increased epistemic arrogance. By epistemic arrogance, we mean an unwillingness to update one’s belief system in the face of evidence and experience supplied by others. As Dillon (2003), Tanesini (2016), and Lynch (2018) have argued, one chief harm of epistemic arrogance is that those who are arrogant in this way fail to give others epistemic recognition respect. Since showing such respect is an important feature of being an epistemically responsible agent when making wide-scope epistemic changes, epistemic arrogance can undermine responsible epistemic agency. In order to spell this out, it is important to flesh out these terms. As Darwall (2006) has influentially argued, there are at least two kinds of respect, and both
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Internet and Epistemic Agency 397 kinds are relevant here in different ways. Recognition respect differs from what Darwall calls “appraisal respect.” Appraisal respect is a kind of esteem—as when we respect someone’s performance as a scientist, athlete, or chef. Recognition respect, on the other hand, is not esteem, but a recognition of dignity or authority; “it concerns not how something is to be evaluated or appraised but how our relations to it are to be regulated or governed” (2013, 19). In the epistemic realm, appraisal respect is essentially esteem for another’s epistemic virtues. You respect someone in this way when you see her as either knowledgeable about some subject, for example, or as open-minded, or perceptive, or careful. Epistemic recognition respect, on the other hand, is “realized in our epistemic conduct in relation to her” (2013, 20). That is, “in giving her views weight or authority in deciding what to believe ourselves” (ibid.).7 We accord someone epistemic recognition respect when we recognize her as a fellow reasoner, judger, or knower who has the authority to both form their own opinions and to give reasons to defend them. When you ignore or discount someone’s views because you justifiably take them to have little or no expertise on the subject at hand, one is neither close- minded nor arrogant. You show a justified lack of appraisal respect towards the person and their views on the matter. To the extent that someone is deeply epi stemically arrogant about some aspect of his world view, he shows a lack of recognition respect towards others. He is inclined to discount others’ testimony not simply because he thinks they are less informed, but because he is inclined to see their contributions as fundamentally having no positive epistemic impact on his own. He is unwilling to see their views as relevant to his own opinions, and to that extent he does not see them as equal players in the epistemic game of giving and asking for reasons. As we noted in section 2, the internet can make it very easy to write off the contributions of others online as worthless and irrelevant. Our earl ier example suggested that we are often quick to dismiss others because of their picture, name, or other identifying features that might mark them out as having, for example, a political affiliation that conflicts with one’s own. We suggested that the judgment that is made, perhaps implicitly, is something like, “I don’t have to consider the opinions of people like that.” In this way, epistemic arrogance can contribute to testimonial injustice (Fricker 2007). If I discount the knowledgeable testimony of a member of a marginalised group simply because of their (perceived) identity, I am acting in an epistemically unjust way, and I may fail to appreciate some important fact as a consequence. But if, moreover, I am unwilling to countenance that I could learn something 7 It is worth noting here that we are diverging from Darwall’s own discussions of epistemic respect. In particular, we take it that epistemic respect will have a scope beyond truth and truth seeking—we take this to be warranted, given the deeply social nature of many of our epistemic practices. Thus, affording epistemic respect is not merely a matter of truth seeking, but also a matter of the appropriate social-epistemic conduct.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
398 Hanna Gunn and Michael Patrick Lynch from any member of that group about some aspect of my world view, then to that degree the injustice is a violation of recognition respect. Fundamentally, as Dillon has noted, that is because, as with arrogance generally, “it is a violation of respect for others because it both denies the intrinsic dignity of others and strikes at their self-respect” (2003, 195). To fail to recognize that another may have something to teach you—even if the lesson you draw isn’t perhaps quite the same as they intended—means that you not only fail to recognize their dignity as an epistemic agent, but also that you implicitly undermine their self-respect as fellow knowers. It seems uncontroversial that identity-prejudiced credibility deficits of the sort Fricker is concerned with happen routinely online. So, to the extent that the internet facilitates and encourages the quick dismissal of others on the basis of some superficial features of their world view or identity, then it may contribute to a culture or practice of denying epistemic recognition respect to others. In this way, it undermines responsible epistemic agency by making it easier to fail to meet some of our (wide-scope) epistemic obligations to other epistemic agents. The very nature of some online platforms and of the digital personalisation of online content contribute to this concern about epistemic arrogance. They do so because online personalisation can limit the range of persons we encounter online, thus making us less familiar with a wider range of epistemic perspectives and reasons. Echo chambers and filter-bubbles are two phenomena that have garnered significant recent attention, in part for these reasons. Echo chambers are online platforms where one is exposed only to information from like-minded individ uals (Bakshy et. al. 2015). A filter-bubble, by contrast, is a result of technology such as personalised search engines, media feeds, and social media circles, which make us informationally isolated in a “bubble” of information that reflects our pre-existing beliefs (Pariser 2011). The personalisation of social media platforms and search engines is one feature of the internet that is particularly worrying for reasons of responsible epistemic agency. With respect to narrow-scope concerns, echo chambers and filter-bubbles threaten responsible epistemic agency by making it likely that we will be duped into false beliefs or misinformation and make it more likely that we will fail to appropriately survey the range of relevant evidence on offer. When many users of these platforms are not aware of the risks posed by echo chambers and filter-bubbles, then they can unwittingly come away with the impression that they have done their epistemic due diligence.8 The ways that digital personalisation inhibits our ability to get to the truth because it narrows down our ability to access potentially relevant information is a critical topic for present social epistemologists. We take it that the ways in which the
8 We note again the work by Miller and Record (2013) on this topic, and their discussion of what is required for justified belief when using personalised internet services or platforms.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Internet and Epistemic Agency 399 internet might encourage epistemic arrogance is an equally pressing interpersonal social epistemic topic. An obvious corrective for epistemic arrogance resulting from the personalisation of online spaces is to intentionally seek out alternative spaces—be these communities, news sources, blogs, or Twitter feeds. However, even the well-intentioned epistemic agent who seeks to expand her world view is still likely to have her newsfeed customised because of who she interacts with, what she “likes”, and what she shares.9 Our ability to get around digital personalisation is limited, it is part-and-parcel of the online epistemic environment, and this doesn’t look to be something that will change anytime soon. This is one way in which the internet might always present a barrier for responsible epistemic agency. The ability to selectively create online communities can lead to the creation of spaces where epistemic recognition respect is frequently denied to particular persons, while simultaneously reinforcing (or greatly increasing) the in-group members’ credence in their own beliefs. This raises a combination of narrow- and wide-scope concerns for responsible epistemic agency. When we are very isolated within circles of like-minded people, what can result is an alienation from people whose world views differ from ours. If this is sufficiently strong, then we can fail to recognise them as taking part in the same broad epistemic joint activities as the people we do interact with frequently—trying to make sense of our world and our place in it, the political task of finding the good life for us all, etc.10 This perceived rubicon between our world view and theirs, our reasons and theirs, our experts and theirs, can make it easy to completely discount people as having anything worth contributing to our epistemic life and ends. Quality epistemic collabor ation and rational persuasion would seem to depend on our treating one another with epistemic recognition respect; if we are sufficiently alienated, then we lose this common ground. The familiar unpleasantness of online interactions has not gone unnoticed in academic research. A 2016 poll by the Pew Research Center and Elon University asked 1,537 technology experts, scholars, corporate practitioners, and government leaders whether they thought “public discourse online [would] become more or less shaped by bad actors, harassment, trolls, and an overall tone of griping, distrust, and disgust?” (Rainie et al. 2017). The results showed that 42 percent expected no major change, while 39 percent expected that online social inter action would be increasingly shaped by negative activities online. The same respondents felt empowered by online communication technology, but nonetheless expressed pessimism about the future quality of online interaction. 9 See the Wall Street Journal’s “Blue Feed, Red Feed” for an insight into just how different one’s Facebook news feed can appear, given one’s (apparent) political affiliation based on the personalisation algorithms used by Facebook to deliver content (Wall Street Journal 2015). 10 See Sunstein (2017) for an extended discussion of how social media personalisation affects our democratic ideals.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
400 Hanna Gunn and Michael Patrick Lynch One of the themes noted from the poll in response to this question was that, “the ever-expanding scale of Internet discourse and its accelerating complexity make it difficult to deal with problematic content and contributors” (our emphasis). This supports some of our concerns about the impact of the internet for both narrowscope and wide-scope epistemic changes; in particular, for our ability to responsibly form beliefs (“content”), to appraise others online, and to engage in respectful epistemic interactions online (“contributors”). On a positive note, Frost-Arnold (2016) proposes an account of “hopeful trust” as a tool for overcoming epistemic ignorance that results from privilege and prejudice. What is particularly salient for our present discussion is that the model she develops is based on case studies of online efforts to challenge ignorance. As she notes, while there are many people who participate in online discussions of prejudice to share their experiences with similar persons, and there are those who participate to antagonise others, there are many people who engage because of a sincere desire to help others “unlearn their privileged habits and prejudiced stereotypes” (2016, 514). To show someone hopeful trust is to be a speaker, S, who acts in such a way as to present to the prejudiced hearer, H, a vision of H that is unprejudiced, and where S trusts that H can become this unprejudiced version of themselves. S’s hopeful trust in H then serves as a motivation for H to overcome their prejudice. Frost-Arnold discusses successful cases of hopeful trust online to show its effectiveness as a strategy for addressing hostile epistemic agents on the internet. These case studies and the account of hopeful trust inject some optimism into this discussion, and also present a way that epistemic agents can actually work to engage with others in respectful ways that might positively impact the development of online epistemic communities. We have suggested in this section that the internet poses threats to responsible epistemic agency by the ways that it seems to support epistemically arrogant and exclusionary behaviours. In addition, we have highlighted some of the ways that the personalisation of online spaces can lead us unwittingly into echo chambers, filter-bubbles, and away from a productively diverse range of epistemic perspectives. In section 4.2, we consider how the sheer volume of information on the internet is making it difficult to responsibly create narrow- and wide-scope epi stemic change.
4.2 Fake News and Information Pollution Perhaps no phenomenon more dramatically captures the present atmosphere than the rise of “fake news,” in which pure fiction masquerading as truth (like posts that the Pope had endorsed Donald Trump) may have spread wide enough to influence political elections. Yet fake news reports are actually just one kind of information pollution—a particularly virulent kind, no doubt, but not the only
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Internet and Epistemic Agency 401 kind. Other kinds include propaganda, overtly false political advertising placed on social media sites, Twitter bombast—and the use of White House press briefings to assert falsehoods, no matter what the evidence may show. The rise of information pollution poses serious barriers to epistemic agents by making the epistemic “landscape” particularly hostile. The most obvious problem with any kind of information pollution is that it is a kind of lie for political gain. But framing the issue solely in terms of lying actually underplays and mischaracterises the grander deceptions being perpetuated inside the internet’s funhouse of mirrors. In section 2, where we looked at the ways that the internet can enhance epistemic agency, we described a variety of features that bring more people into epistemic activity and give them access to more and more information. However, we also described these with the caveat that the information we can access online is reliably sourced and accurate information, which simply is not true of much of the information found online. Lying is not quite the same thing as deception. To lie on standard accounts is to deliberately say what you believe to be false with the intention of deceiving your audience.11 I can deceive you without lying (silence at a key moment, for instance, can be deceptive). And I can lie to you without deceiving. That may be because you are skeptical and don’t believe me, but it may also be because what I say is inadvertently true. Either way, you are lied to but not deceived. That might suggest that deception occurs when someone is actually caused to believe what is false. “Deception,” as philosophers say, is a “success term.” But that’s only halfway there. Deception can happen even without false belief. Consider the common shell game. The con man presents three shells, one of which has a coin underneath. He moves the shells around and asks you to pick the shell with the coin. If done right, it looks easy, but isn’t. Using sleight of hand, he distracts you so that you can’t track the right shell and know where the penny is. But one can lack knowledge without having a false belief. One can be simply confused, and that is typically the case with such tricks. You don’t know what to think, and so simply guess. You can be deceived not only by believing what is false but by not believing what is true. The use of social media to spread political misinformation online is analogous in some ways to a giant shell game.12 Propagandists often don’t care whether 11 Not all accounts of lying accept that one needs to have an intention to deceive. For instance, Saul (2012) replaces the intention-to-deceive condition with the condition that one instead needs to say what one believes to be false, whilst taking oneself to be in a warranting context. On the assumption that online speech is also generally taken to be speech in a warranting context, as we typically assume is the case with offline speech, then there will be little difference here for our main points. This may not be a warranted assumption online, but we do not have the space here to explore the truthfulness norms of internet speech. 12 The “shell-game” caused by information pollution isn’t a special problem for epistemic agents, but rather a potential problem of any epistemic environment. We thank an anonymous reviewer for the following example that helps to make this clear. Consider how a digital news-aggregation program
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
402 Hanna Gunn and Michael Patrick Lynch everyone, or even most people, really believe the specific things they are selling (although it turns out that lots of people always do). They don’t have to get you to actually believe that the penny is under the wrong shell. They just have to get you confused enough so that you don’t know what is true. That’s still deception. And it is this kind of deception that for-profit conspiracy sites have been particularly adept at spreading on social media. No doubt, some percentage of people actually believed such postings, but a far greater number of people come away ever so slightly more doubtful of what is true. They didn’t know what to believe.13 Information pollution, then, is making it harder to exercise epistemic agency responsibly. It does so by permeating online sources of information with bad information—making it harder to get knowledge. It also makes it harder to pick out reliable and credible sources of information, undermining our abilities to appropriately appraise others as knowers. Given the techniques that are often used in this online shell game, the very biases and heuristics that we might work hard to avoid are exploited. Our epistemic merit and skills can be directly undermined by crafty propagandists online. This makes it harder not only to sort information responsibly, but also to continue to develop good epistemic skills and capacities. This is not unlike hypothetically learning to knit with someone throwing in additional needles and strands of yarn—while you might be able to wade through the mess, it’s much harder for you to keep an eye on what is actually important. There was a time when someone would say something outrageously false (“The Moon landing was faked”), and it would be ignored by most folks with the reason ing that “If that was true, I would have heard about it by now.” By that, they meant “heard about from credible, independent sources.” Filters (primarily, editors) worked to not only weed out the bad, but to make sure the truly extraordinary real news made it to the surface. The internet has made that reasoning less common. Fewer people reject crazy claims based on the grounds that they hadn’t heard about them before now, because chances are they already have heard about them, or something close to them, from the sites that tend to confirm their biases. To be more precise: outrageous claims may seem more credible when they are made within your epistemic bubble. That makes you more susceptible to taking fake news seriously, or at least not taking it un-seriously. It makes you easier to confuse.
would fare in an epistemic environment swamped with propaganda and lies. Without clear indications of what is reliable, the program may struggle just as much as a person to select credible information. The connection to epistemic agency that we want to draw attention to is just how acting epistemically responsibly is made more difficult for epistemic agents online, given widespread—often intentional—information-polluting practices. Moreover, as the example may help to show, relying on digital aggregation may not help us to filter out the quality information from the propaganda. 13 See Vosoughi et al. (2018) on how Twitter is a very effective platform for the spread of false information. They present the disturbing conclusion that misinformation spreads faster by humans than by automated bots on the network, and that the misinformation spread is more memorable than true information.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Internet and Epistemic Agency 403 In section 4.1, we discussed how the personalisation of the internet, either through search engines or social media platforms, can ensconce us in these epistemic bubbles that affirm what we already believe. We might see this as a world of information specially designed for us, a restricted “safe zone” of reasons and evidence designed to keep us epistemically unchallenged. Information pollution, by contrast, can lead to a situation where basically any belief can gain the appearance of being justified—all reasons abound online. Thus, information pollution can cause us to fail to distinguish between conspiracies and quality information or informants by swamping us with so much information that it is incredibly challenging to engage in epistemically responsible ways.14 We have discussed in this section a range of examples of how information pollution can make for a hostile online epistemic environment. What information pollution does, we might say, is make us lose control of this epistemic environment. It does so by swamping the “land” with bad or deceptive informants, by subtly feeding in sponsored content in place of editorial content, by nudging us to share fake news. This raises many risks for epistemic agents: it makes it more difficult for us to gain true beliefs, to exercise our due diligence in research, to contribute truthfully and informatively in dialogue, to develop and exercise intellectual virtues. In short, it makes it increasingly difficult to fulfil one’s epi stemic obligations. These are not new risks, but the internet and the new media it brings with it does appear to be changing how we get information, share it, and the ways we respond to one another. That being said, there is no real analog of the Tweet to be had—short, viral, instant messages that can reach millions of people in very little time. We can point, then, to some forms of online interaction that raise unique interpersonal epistemic challenges and in part because of their viral, information-polluting nature. In our final section, we discuss some of the complications that online anonymity poses to responsible epistemic agency. We take it that this particular feature of the internet both expands or supports epistemic agency, while making it harder to fulfil our epistemic obligations to ourselves and others.
4.3 Anonymity and Epistemic Agency Online personas and pseudonyms provide a way for individuals to speak their minds without having to reveal their identity. This can be empowering, 14 Aikin and Talisse (2016a, 2016b) argue for a distinction between “spitballing” and “swamping” as two tools used to control public political discussion. Spitballing involves making as many contributions to a conversation as quickly as one can in an effort to derail the conversation. Swamping involves making “overtly bizarre or inflammatory” claims, in effect to force other people to discuss them, instead of discussing other, perhaps more important or pressing, topics. Information pollution can make use of both of these tactics and they both serve to inhibit the responsible exercise of epistemic agency.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
404 Hanna Gunn and Michael Patrick Lynch particularly for people speaking out against oppression, abuse, or in other contexts where they may fear retaliation or ridicule. In this way, our epistemic agency can be enhanced on the internet by enabling us to act as testifiers or informants. In times when our ability to speak as ourselves is undermined or chilled, anonymity can be the best enabler.15 The internet brings with it simple and immediate ways to speak anonymously—morally, politically, and epistemically this is can be very powerful. On the other side of the coin, though, anonymity also makes our lives as epistemic agents more difficult, simply because when our informants are anonymous it is difficult to assess their credibility, their associations or conflicts of interest, or their track record. The internet is now a major source of testimonial knowledge for many people, and it brings with it two core epistemic concerns: we regularly don’t know who we are talking to and we typically don’t have access to information about their trustworthiness (Gunn and Lynch 2018). Anonymity undermines reliabilist methods for evaluating testimonial knowledge, and so poses threats to responsible epistemic agents. When we don’t know who our sources are, it’s fairly hard to meet normative conditions for forming beliefs through reliable sources. Anonymity also threatens our ability to afford epistemic trust appropriately; being overly cautious can cause us to ignore, distrust, or dismiss reliable sources. The ability to place epistemic trust well is crucial to our epistemic lives because of the sheer amount of knowledge we gain through the testimony of others. Elsewhere, we have argued that a reductionist theory of epistemic trust is appropriate online because of the difficulties in evaluating online testifiers (Gunn and Lynch 2018). Within the literature on testimony, a reductionist account is one that argues that the epistemic merit of testimony is reducible to the merit of the sources underpinning that testimony.16 Thus, we are entitled to trust some particular testimony depending on the merit of the source of that testimony, for example, inductive inference, perception, or memory. Our support for a reductionist account of the trustworthiness of testimony on the internet requires that epistemic agents work to actually determine the credibility of the sources we are looking at. That is, epistemic trust on the internet is earned, not afforded prima facie. Responsible epistemic agents, if the reductionist account is plausible for online testimony, will need to assess the credibility or justificatory support of
15 See Dotson’s (2011) presentation of testimonial smothering for an account of self-policed speech in the face of prejudiced audiences. 16 Reductionist theories divide fairly evenly into “global reductionism” and “local reductionism.” Proponents of global reductionism often take testimony itself to be a generally reliable source of information, and thus on the basis of the general reliability of testimony we are entitled (prima facie) to trust testimony (e.g. Van Cleve 2006). Proponents of local reductionism instead typically argue that we are entitled to trust particular speakers, on particular issues because of their reliability in the past (e.g. Fricker 1994).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Internet and Epistemic Agency 405 online testifiers. Anonymity on the internet, though, presents us with strangers that we cannot so evaluate. Roughly along these lines, Robert Fellmeth has argued in an interview with Ralph Nader that we do not have a right to online anonymity based on the First Amendment to the US Constitution (Fellmeth and Nader 2017). Fellmeth argues that online anonymity undermines the rights of listeners to choose who they listen to and undermines their ability to evaluate speakers for credibility.17 Key to his argument is that listeners have epistemic rights based on the First Amendment to develop their own informed point of view. This requires not only protections on speech, but also entitlements for audiences to know who they are listening to, and in addition, that audiences have a range of viewpoints available to them. Anonymity not only prevents us from being able to evaluate someone as credible, but it can make us easy targets for propaganda and potentially drastically reduce the pool of informants without our knowledge. Knowing the identity of who we interact with, then, is important not only for our personal beliefs but for the health (and diversity) of our collective pool of knowledge. We do not intend to evaluate Fellmeth’s interpretation of the Constitution, but his argument gets right at the problem for epistemic agency we are interested in here. Anonymity is widely regarded as a positive design feature of the internet, but it is potentially hazardous for responsible epistemic agency. If having justified beliefs based on the testimony of others online requires that one is able to assess the credibility of the testifier, then anonymity undermines this requirement for responsible epistemic agents. The reductionist proposal for affording epistemic trust to online testifiers recommends skepticism in the face of anonymous testimony online—unless there is some other way of assessing the credibility of the source independent of knowing the particulars about the individual(s) offering the testimony.18 The idea that listeners as epistemic agents might be entitled to some information about who they are listening to that enables them to assess their credibility is a striking proposal. The suggestion would plausibly amount to a positive epistemic obligation to make ourselves identifiable to our audiences so they can assess our trustworthiness or credibility. This might follow from wide-scope concerns about
17 ‘People talk about the right to speak and free speech on the utterance side. But that’s only one part of it. The other part of it is the right of the audience to weigh the credibility of the speaker. Who is that? What are their biases? What’s their expertise? The first amendment is not just defending the right of people to bleat, to make noise. It has a purpose in terms of ascertaining the truth, and developing the points of view, and educating people. And the identity of the speaker is critical to that function.’ (Fellmeth and Nader 2017) 18 For instance, anonymous posting on reliable platforms, anonymous testifiers with institutional affiliations, group testimony from credible organisations, but where the individual members of that group cannot be identified, or even platforms like Wikipedia, where we might take the reviewing procedures for content to be sufficient to establish credibility. We take it that assessing what exactly counts as testimony online is a substantive issue that deserves more careful treatment than we provide here.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
406 Hanna Gunn and Michael Patrick Lynch being a source of epistemic change for other persons. From the testifier’s perspective, we might say that responsibly trying to change someone else’s mind online requires that one’s testimony be presented in such a way that one’s audience can evaluate one’s credibility. If one only testifies anonymously, one might be failing to participate in epistemic activities responsibly because one doesn’t present one’s testimony with relevant information about the reliability of that testimony. From the perspective of an inquirer assessing an online testifier, by contrast, if responsibly receiving online testimony requires that one appropriately appraise the testifier, and such appraisal requires that one’s interlocutor disclose who they are, then we might think that our interlocutor ought to make this information available to us. For Practical reasons the pernicious consequences of widespread fake news and information pollution might weigh in favour of more paternalistic measures surrounding online testimony. For instance, if we do not have good reason to think that the majority of people are likely and/or capable of being appropriately skeptical on the internet, then we may place demands on those who create content or provide content-sharing platforms to introduce paternalistic quality measures—one such measure could be the elimination of anonymity. This would, of course, depend on which side “wins out” in an analysis of the benefits of anonymity over its elimination online, an issue we cannot settle here. In this section, we have expanded on the tension between online anonymity and responsible epistemic agency. We have shown that this has at least two aspects: first, that forming beliefs responsibly from the anonymous online testimony of others might require that we largely remain skeptical of anonymous online testimony and resist forming beliefs from it; second, that what it takes to responsibly contribute to epistemic activities online might require that we in fact do not testify anonymously because in doing so we might fail to respect an epi stemic entitlement of our audience to facts that enable them to assess the credibility of what we say.
5. Conclusion In this chapter, we have been concerned with the effect the internet has on our epistemic agency (our ability to make epistemic changes in our environment), whether those changes are narrow or wide. Narrow-scope changes concern our abilities as epistemic agents to bring about changes to our epistemic attitudes through, for example, acquiring justified beliefs or developing intellectual virtues. Wide-scope changes concern our abilities to contribute to and participate in collective epistemic activities, for example, contributing to a research team or developing new methods for inquiry.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Internet and Epistemic Agency 407 While the internet expands epistemic agency by providing more people with more access to epistemic goods, by making epistemic life more inclusive, and by opening up opportunities for epistemic participation, the internet also raises a range of specific challenges for responsible epistemic agents. In particular, the various ways that content and social interaction are personalised online make it difficult to meet one’s narrow-scope obligations to engage in quality research, and often make it easy for epistemic agents to become arrogant inquirers and fail to meet their wide-scope obligations. The rise of fake news and information pollution online make the epistemic environment a hostile one to well-intentioned epistemic agents and can encourage the sharing of false or deceptive information, or simply make us more susceptible to this information. Finally, online anonymity can potentially undermine both the responsible acquisition and dissemination of knowledge via testimony. The internet’s influence on our epistemic lives is vast, and philosophers and other researchers are just starting to come to terms with the extent and depth of that influence. This chapter is one part of that effort; much more work remains to be done before we fully understand the ways in which information technology impacts our ability to be responsible knowers.
References Aikin, Scott F, and R. B. Talisse (2016a). “Spitballing. 3 Quarks Daily.” (June 20, 2016), www.3quarksdaily.com/3quarksdaily/2016/06/spitballing.html. Aikin, S. F. and R. B. Talisse (2016b). “Swamping. 3 Quarks Daily.” (July 11, 2016). www.3quarksdaily.com/3quarksdaily/2016/07/swamping.html. Alston, William P. (1988). “The Deontological Conception of Epistemic Justification.” Philosophical Perspectives 2: 257–99. Bakshy, E., S. Messing, and L. A. Adamic (2015). “Exposure to Ideologically Diverse News and Opinion on Facebook.” Science 348(6239): 1130–2. Brabham, D. C. (2013). Crowdsourcing. Cambridge, MA: MIT Press. Van Cleve, J. (2006). “Reid on the Credit of Human Testimony,” in Jennifer Lackey and Ernest Sosa (eds), The Epistemology of Testimony. Oxford: Oxford University Press, 50–74. Darwall, Stephen L. (2006). The Second-Person Standpoint: Morality, Respect, and Accountability. Cambridge, MA: Harvard University Press. Darwall, Stephen L. (2013). Honor, History, and Relationship: Essays in SecondPersonal Ethics II. Oxford: Oxford University Press. Dillon, Robin S. (2003). “Kant on Arrogance and Self-Respect,” in Chester Calhoun (ed.), Setting the Moral Compass: Essays by Women Philosophers. Oxford: Oxford University Press, 191–216.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
408 Hanna Gunn and Michael Patrick Lynch Dotson, Kristie (2011). “Tracking Epistemic Violence, Tracking Practices of Silencing.” Hypatia 26(2): 236–57. Engel, P. (2013). “Is Epistemic Agency Possible?” Philosophical Issues 23(1): 158–178. Fellmeth, Robert and Raplph Nader (2017). “The Difference between Liberal and Progressive.” (December 9, 2017), https://ralphnaderradiohour.com/the-differencebetween-liberal-and-progressive. Fernbach, Philip M., Todd Rogers, Craig R. Fox, and Steven A. Sloman (2013). “Political Extremism is Supported by an Illusion of Understanding.”: Psychological Science 24(6): 939–46. Fisher, Matthew, Mariel K. Goddu, and Frank C. Keil (2015). “Searching for Explanations: How the Internet Inflates Estimates of Internal Knowledge.” Journal of Experimental Psychology General 144(3): 674–87. Fricker, Elizabeth (1994). “Against Gullibility,” in A. Chakrabarti and B. K. Matilal (eds), Knowing from Words. Amsterdam: Kluwer Academic Publishers, 125–161. Fricker, Miranda (2007). Epistemic Injustice: Power and the Ethics of Knowing. Oxford: Oxford University Press. Frost-Arnold, K. (2016). “Social Media, Trust, and the Epistemology of Prejudice”. Social Epistemology 30(5–6): 513–31. Fuller, Steve (2016). “A Sense of Epistemic Agency Fit for Social Epistemology,” in Patrick Reider (ed.), Social Epistemology and Epistemic Agency: Decentralizing Epistemic Agency. Lanham, MD: Rowman & Littlefield: International. Goldberg, Sanford C. (2016). “A Proposed Research Program for Social Epistemology,” in Patrick Reider (ed.), Social Epistemology and Epistemic Agency: Decentralizing Epistemic Agency. Lanham, MD: Rowman & Littlefield: International. Gunn, Hanna K. and Michael P. Lynch (2018). “Googling,” in David Coady (ed.), The Routledge Handbook of Applied Epistemology. London: Routledge. Jeppesen, Lars B. and Karim R. Lakhani (2010). “Marginality and Problem-Solving Effectiveness in Broadcast Search.” Organization Science 21(5): 1016–33. Kornblith, Hilary (1983). “Justified Belief and Epistemically Responsible Action.” Philosophical Review 92(1): 33–48. Lynch, Michael P. (2013). “Epistemic Commitments, Epistemic Agency and Practical Reasons.” Philosophical Issues 23(1): 343–62. Lynch, Michael P. (2018). Arrogance, Truth, and Public Discourse.” Episteme 15(3): 283–96. Lynch, Michael P. (2016). Internet of Us: Knowing More and Understanding Less in the Age of Big Data. New York: Liveright. McHugh, Conor (2013). “Epistemic Responsibility and Doxastic Agency.” Philosophical Issues 23(1): 132–57. Miller, Boaz and Isaac Record (2013). “Justified Belief in a Digital Age: On the Epistemic Implications of Secret Internet Technologies.” Episteme 10(2): 117–34.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Internet and Epistemic Agency 409 Olson, Dustin (2015). “A Case for Epistemic Agency.” Logos and Episteme 6(4): 449–74. Pariser, Eli (2011). The Filter Bubble: What the Internet Is Hiding from You. London: Penguin Press. Rainie, Lee, Janna Anderson, and Jonathan Albright (2017). “The Future of Free Speech, Trolls, Anonymity, and Fake News Online.” Pew Research Center: Internet and Technology, www.pewinternet.org/2017/03/29/the-future-of-free-speechtrollsanonymity-and-fake-news-online. Reider, Patrick J. (2016). “Introduction: What is Social Epistemology and Epistemic Agency?,” in Social Epistemology and Epistemic Agency: Decentralizing Epistemic Agency. Lanham, MD: Rowman & Littlefield: International. Saul, Jennifer M. (2012). Lying, Misleading, and What is Said: An Exploration in Philosophy of Language and in Ethics. Oxford: Oxford University Press. Setiya, Kieran (2013). “Epistemic Agency: Some Doubts.” Philosophical Issues 23(1): 179–98. Sloman, Steven and Philip Fernbach (2017). The Knowledge Illusion: Why We Never Think Alone. London: Penguin Random House. Standing, Susan and Standing Craig (2018). “The Ethical Use of Crowdsourcing.” Business Ethics: A European Review 27(1): 72–80. Steup, Matthias (2017). “Believing Intentionally.” Synthese 194(8): 2673–94. Sunstein, Cass R. (2017). #Republic: Divided Democracy in the Age of Social Media. Princeton, NJ: Princeton University Press. Tanesini, Alessandra (2016). “ ‘Calm Down, Dear’: Intellectual Arrogance, Silencing and Ignorance.” Aristotelian Society Supplementary Volume 90(1): 71–92. Vosoughi, S., and D. Roy, and A. Sinan (2018). “The Spread of True and False News Online.” Science 359: 1146–51. Wall Street Journal (2015). “Blue Feed, Red Feed.” (August 19, 2019), http://graphics. wsj.com/blue-feed-red-feed.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 19/03/21, SPi
17
How Twitter Gamifies Communication C. Thi Nguyen*
Twitter is now one of our primary venues for public discourse. But it is not a neutral or transparent medium. Twitter shapes how we interact, who we interact with, and—perhaps most importantly—it suggests specific goals for those interactions. Twitter doesn’t just provide a speaking platform, nor are its effects confined to algorithmic filtering. Twitter shapes our goals for discourse by making conversation something like a game. Twitter scores our conversation. And it does so not in terms of our own particular and rich purposes for communication, but in terms of its own pre-loaded, painfully thin metrics: Likes, Retweets, and Follower counts. And if we take up Twitter’s invitation and internalize those evaluations, we will be thinning out and simplifying our own goals for communication. Let’s take a step back. Twitter is at once pluralistic in its scope and monolithic in its technological form. Twitter is pluralistic because it offers relatively open access to powerful resources for public discourse. Anybody can form an account, and anybody with the right feel for the medium, it seems, can gather enormous numbers of followers. Twitter democratizes access to large-scale communication, which once had been held by a relatively small number of media companies.1 At the same time, Twitter is monolithic, because everybody who uses Twitter must communicate through the same interfaces and must be subject to the same algorithms.2 What is the impact of so much public discourse being shoveled through one platform? Other discussions of Twitter have focused on the enforced shortness of tweets, the influence of hidden algorithmic filtering, the promotion of group * I’d like to thank Mark Alfano, Matthew Carlson, Helen Daly, Jon Ellis, Max Hayward, Aaron James, Jennifer Lackey, Michael Lynch, Elijah Millgram, Alison Rieheld, Adriel Trott, and Matt Strohl for their help with this chapter. Key ideas for the chapter emerged from my work with Bekka Williams on moral outrage porn—including the notion of instrumentalization. 1 My understanding of Twitter here has been particularly informed by Zeynep Tufekci’s (2017) thoughtful analysis of the advantages and disadvantages of this pluralism. 2 This pattern of thought has been particularly influenced by Tufekci’s (2018) discussion of how the internet era has democratized communication, but at the same time subjected all online communication to a few very small portals. The internet is democratic because anybody can put up a web page, but it is monolithic because we all find web pages using Google Search—so Google Search’s algorithm becomes an all-powerful control on our collective attention. Obviously, in the background, is Marshall McLuhan’s (1964) discussion of the impact of medium over content.
C. Thi Nguyen, How Twitter Gamifies Communication In: Applied Epistemology. Edited by: Jennifer Lackey, Oxford University Press (2021). © C. Thi Nguyen. DOI: 10.1093/oso/9780198833659.003.0017
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
How Twitter Gamifies Communication 411 polarization, the lack of accountability mechanisms, and the collapse of conversational contexts (Sunstein 2009, 46–96; Marwick and boyd 2011; Miller and Record 2013; Frost-Arnold 2014; Rini 2017). I would like to focus on another basic feature of Twitter—one whose importance and impact, I think, has not been adequately appreciated. Twitter gamifies communication by offering immediate, vivid, and quantified evaluations of one’s conversational success. Twitter offers us points for discourse; it scores our communication. And these game-like features are responsible for much of Twitter’s psychological wallop. Twitter is addictive, in part, because it feels so good to watch those numbers go up and up. In fact, the design of Twitter and its scoring mechanisms have been significantly informed by design strategies fostered in the Las Vegas gambling industry—strategies which overtly seek to increase the addictiveness of their products.3 The clear scoring system brings with it another very game-like aspect: a clear and unambiguous ranking. We usually don’t emerge from a party with a ranked list of who the best conversationalists were. Twitter, on the other hand, offers both short-term rankings (Likes and Retweet numbers for each tweet) and long-term rankings (Follower counts). Most importantly, the rankings are entirely unambiguous. Unlike conversation in the wild, I can know exactly how well each tweet did, and I can instantly compare my overall popularity with that of any other user. This can provide all sorts of pleasures: the thrill of victory, when we see those numbers tick up; and the sense of long-term achievement, presented in precise and unquestionable quantitative form. Supporters of gamification say that it is a technology for increasing motivation. Gamification can supposedly imbue everyday activities with all the fun and excitement of a game.4 Here, then, is an optimistic view of Twitter: by gamifying public discourse, Twitter increases overall participation, and so helps us to reap the rewards of public discourse—such as a more fully politically engaged populace. I do not accept the optimistic view. Crucially, I don’t think that gamification merely increases our motivation to perform an activity, while preserving all the original goods of that activity. Gamification increases our motivation by changing the nature of the activity. Often, the goals of ordinary activity are rich and subtle. When we gamify these activities, we change those goals to make them artificially clear. Games are more satisfying than ordinary life precisely because game-goals are simpler, clearer, and easier to apply. In games proper, this simplification isn’t particularly problematic, because the goals are peculiarly artificial. Game 3 Natasha Dow Schull’s Addiction by Design (2012) offers a thorough look at the technological innovations of the gambling industry to optimize the addictiveness of their products. Since that book, Schull has been vocal about how those technologies have been adopted by gaming and social media companies (Madrigal 2013; National Public Radio 2014; Seymour 2019). 4 This point was put most influentially by Jane McGonigal (2011). For critical discussion, an excellent starting place is Walz et al. (2015).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
412 C. Thi nguyen activities, and their associated goals, are usually kept secluded from ordinary life. But there is no such protective separation when we gamify ordinary activities. To reap the motivational benefits of gamification, we must reshape the ends which govern our real-life activities. Pre-gamification, the aims of discourse are complex and many. Some of us want to transmit information or to persuade; some of us want friendship. Some of us want to join together in the pursuit of truth and understanding. Twitter gamifies discourse and, in so doing, offers us re-engineered goals for our communicative acts. Twitter invites us to shift our values along its pre-fabricated lines. We start to chase higher Likes and Retweets and Follower counts—and those are very different targets. Others of us may come to Twitter already interested in popularity and status. For those, the gamification of Twitter may not represent such a radical change in the basic content of their goals. But even for those already interested in popularity, Twitter can change the way in which they conceive of popularity—by making highly salient a handful of specific metrics for popularity. Like and Follower counts are not the only way to conceive of popularity, but they the measure that Twitter highlights. What’s more, the effect of Twitter’s gamification, across the community Twitter users, will tend towards levelling and flattening the diversity of values. Insofar as Twitter’s gamification motivates its users, then it will drag all of its users’ communicative values in the same direction—towards the same metric. Gamification homogenizes the value landscape. And this phenomenon will help explain some of the more socially toxic aspects of Twitter. The technology invites us to focus our cares on the narrow task of getting points and going viral. And that goal is in tension with our interest in having morally sensitive and openhearted communication. This gamification invites us, instead, to view communication through the lens of competition, victory, and success on Twitter’s very specific terms. Let me emphasize the fact that Twitter offers us an invitation to change our values. Twitter will not change our values for us. It is a system designed to offer us pleasure in return for simplifying our values—but we still have to take up that offer. But it does ease the way for us considerably, by offering a pre-prepared and seductively designed pathway. Of course, Twitter isn’t the only place where gamification influences communication, discourse, and collective understanding. We can see similar effects with Facebook’s Likes, YouTube’s click-through and watch-through counts,5 academic citation rates, and more. But here I wish to explore, in detail, how gamification impacts discourse and knowledge production in one particular instantiation, as an opening step towards understanding life in the time of quantification.
5 The YouTube examples were suggested by Mark Alfano.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
How Twitter Gamifies Communication 413
1. Games and Gamification Why gamify? If there is a Bible to the contemporary gamification movement, it is Jane McGonigal’s book, Reality is Broken: How Games Make Us Better and How They Can Change the World.6 McGonigal provides a clear—and very influential— argument for gamification. Ordinary life, she says, is quite painful. Everyday activities, like work, education, and chores, are dull and repetitive. But luckily, she says, we already have an extremely effective technology for eliminating drudgery: computer games. By importing key design features from modern gaming into ordinary activities, we can transform daily life into something far more enjoyable (McGonigal 2011). In so many modern computer games, we voluntarily engage in what looks, from the outside, like pure drudgery. Many games, particularly computer roleplaying games, involve what’s known as “grinding”—performing simple, repetitive activities to slowly build up various in-game points and currency. Grinding can involve killing easy enemies, over and over again, for experience points and gold—or laboriously gathering piles of ingredients in order to craft equipment. Why are people willing to engage in such drudgery in their spare time when they avoid such activities like the plague in real life? The answer seems to lie in the powerful feedback and reward mechanisms available in games, especially contemporary computer games. In such games, we are given immediate rewards for our achievements in the form of points, leveling up, achievement badges, and the like. Games quantize our successes, making our progress clear and vivid. McGonigal emphasizes, in particular, how games offer us a steady sense of progress and victory, through a constant stream of clear feedback, in the terms of the accumulation of points (52–63).7 So why can’t we borrow those feedback and reward mechanisms and slather them over real-life activities? We’ve seen, in recent years, many efforts to gamify the workplace and the school. Business entrepreneurs seem particularly interested in gamification’s ability to increase worker productivity by increasing worker motivation. Disney famously gamified its hospitality workforce, providing leaderboards and rankings for speedy performance. Notably, the system increased productivity, at the expense of also increasing the injury rate. And workers hated the system, calling it the “electronic whip”—and saying that they couldn’t help being motivated by it, even though they detested the intrusion (Gabrielle 2018). We’ve seen the introduction of gamification into fitness, with technologies like FitBit and Strava offering game-like structures of points, rankings, and leaderboards for exercise. We’ve 6 For a more technique-oriented design manual for gamification, see Chou (2015). 7 The mechanisms for this are complex. McGonigal provides a survey of the empirical literature; for a more pessimistic counterpoint; see Schull’s work (2012) on game addiction and its relationship to points. Schull’s account stresses the way in which the exact timing of the quantized reward in certain game designs triggers addictive surges of serotonin.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
414 C. Thi nguyen seen gamified education in schools, and in various apps. The popular language learning app DuoLingo gamifies language learning by offering its users points and virtual medals for achieving various daily goals, like learning new vocabulary words. McGonigal and her fellow gamification advocates are optimistic about the utopian potential of gamification. In McGonigal’s picture, gamification is an unalloyed good: it simply removes drudgery and adds pleasure. But her optimism depends on believing that gamification can achieve these psychological goods while adequately preserving the value of the activity. When we understand the source of gamification’s motivational power, we will see the problem with McGonigal’s optimism. Gamification involves a trade: it increases our motivation in an activity by narrowing and simplifying the target of that activity—which, in turn, changes the nature of the activity. And this may be fine when the activity has a naturally simple target, as is possibly the case with language learning. But the goals of discourse are many and subtle, and gamification threatens to destroy much of that diversity and subtlety. The usual view among gamification advocates is to treat games and gamification as providing the same sort of value. Insofar as games are good, the story goes, then gamification must also be good, since it makes life more like a game. But this view conceals the profound differences between games proper and the gamification of real-world activities. To understand that, we will need a clearer account of the nature and value of games. Let me summarize my account of games, which I have developed in excruciating detail elsewhere.8 Games, I have argued, are the art form that works in the medium of agency. The game designer doesn’t just create characters, stories, and environments. The game designer sculpts the temporary agency that the player will occupy during the game. They design, not only a world, but who the player will be in that world. I do not just mean that the game designer provides a fictional backstory for a character. They design the essential agential structure of the ingame actor. They designate what the in-game agent’s abilities and affordances will be—whether they will be a jumper, a shooter, a builder or an information gatherer. And, most importantly, the game designer sets the in-game agent’s motivations by setting the goals of the game. And the game player submerges themselves in this sculpted agency, temporarily. Game playing involves the temporary adoption of an alternate set of goals. Why do all this? For one thing, our goals in game-life are so much clearer than in ordinary life. In ordinary life, our goals are often obscure. We often don’t know exactly what we’re doing—or we find our reasons hard to articulate and difficult to apply. And we are beset with a confusing welter of values—both from within
8 The present account relies on material drawn from Nguyen (2017, 2018, 2020).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
How Twitter Gamifies Communication 415 our own value system and from the bruising value complexity of the social world. But games offer a relief from all that. While playing a game, we know exactly what we are trying to do—and afterwards, we know exactly how well we have done. Success in a game is clear and unmistakable. There are points. And game values usually fit neatly with one another. In ordinary life, our values are hard to balance. I care about spending time with my loved ones, raising my children right, writing good philosophy, enjoying myself in rock climbing, staying healthy, and eating delicious food. Not only are my values often in tension, but there is usually no way to precisely compare them. How do I compare achievements under one of these goals against sacrifices in another? What, exactly, is the cost–benefit analysis for choosing between working today or taking my children to the aquarium? But with games, there is usually a clear central currency of value. A game tells me to achieve victory points and then tells me exactly how many victory points things are worth.9 The goods of a game are readily commensurable, by design. In ordinary life, values are often inchoate, subtle, and difficult to apply. But in games, values are easy. Games offer us a momentary experience of value clarity. They are a balm for the existential pains of real life. In games, we know exactly what we are doing and why we are doing it. And when we are done, we know exactly how well we have done. Games offer us a momentary respite from the value confusion of the world. It is relatively easy for the game designer to create value clarity because the values in games are entirely artificial. The game designer can just tell us what to care about, and we simply care about it for a while. This is part of what it means to say that agency is the medium of games. The in-game agencies—their abilities, their motivations—are the plastic medium which the game artist manipulates to achieve their effects. But when we seek to gamify ordinary life, we are trying to impose value clarity on a pre-existing thicket of values. This is the worry with Twitter. Twitter can grant us the emotional security and existential relief of value clarity, but we must adopt Twitter’s narrowed targets in exchange.
2. How Twitter Changes Discourse McGonigal views gamification as providing nothing but a motivational boost. The analysis I have offered shows the problem with that view. We get those extra motivational elements—pleasure, fun, engagement—in exchange for 9 Some other games offer a few different currencies of success, but even then, those various currencies are usually compatible. In many computer role-playing games, for example, I am offered both experience points and gold, with no clear explanation of which I am to pursue. Though there is no direct exchange rate between the two currencies, they go hand-in-hand. Usually the path to more experience points is through more gold, and vice versa. So, for success I can aim to maximize both.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
416 C. Thi nguyen substantively changing the goals of the activity, and so changing the activity itself. The gamified design of Twitter influences discourse by inviting its users to change the goals of their participation in discourse—to simplify those goals in exchange for pleasure.10 Let me stipulate a bit of terminology. Let us call the designed technology which offers points and scores “design for gamification.” And let us use “gamification” to refer to those cases when a player interacts with design for gamification and actually adopts those points and scores as primary motivators during the activity— when the activity actually does become something like a game for them. Notice that you can need to actually adopt these clear goals, at least for the moment, to get the pleasures on offer. Consider some of our ordinary goals for communication. We may wish to collectively pursue truth and understanding, or to promote empathy for one another. But Twitter’s scoring mechanism invites us to replace those values with another, much simpler goal: that of maximizing one’s Likes, Retweets, and Follower counts. Twitter’s measures are a radically simplified—and quite impoverished— rendition of the wide plurality of values for communication we might hope to find across a community of conversers. For one thing, we have evidence aplenty that what makes something go viral is not its truth, or the degree to which it promotes understanding. Recent studies have shown that tweets loaded with strong moral emotions, like outrage, are far more likely to go viral, via an effect that researchers call “moral contagion” (Brady et al. 2017). But a gamification booster might resist this portrayal. They might suggest that Twitter’s scoring mechanism does an adequately good job of reflecting the true plurality of values. Perhaps individuals have their own values for which they come to Twitter. But those values guide when each individual user decides to Like, Retweet, or Follow. Thus Likes, Retweets, and Follower counts serve as useful measures of overall success against a plurality of values, since they function to aggregate individual approval. But being guided by an aggregate measure of the audience’s approval is a far cry from being guided by one’s own internal sense of value. First, pressing Like is a quick reaction. It typically records a user’s positive first-impression response to a tweet. So, if we evaluated our communicative attempts by their Like counts, we 10 My view here is much opposed to Ian Bogost’s famous argument that “gamification is bullshit.” Bogost’s argument here is that gamification is bullshit because the term ‘gamification’ was used so flexibly and variably by corporate profiteers that the term was essentially useless—that it was a pure buzzword, with no content (Bogost 2011). As my discussion shows, gamification is a specific phenomenon with clear techniques and identifiable consequences. The fact that salespeople have used the term poorly does not undermine the usefulness of the term itself. Interestingly, I do think that gamification is bullshit, but in a different sense. Harry Frankfurt’s discussion of bullshit can be read in the following way: bullshit is an activity that has been diverted from its usual goal (Frankfurt 2005). In that specific sense, I do think gamification is bullshit. (Bogost cites Frankfurt’s discussion of bullshit, but Bogost misses much of the specificity of Frankfurt’s analysis.)
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
How Twitter Gamifies Communication 417 would be effectively biased in favor of tweets that users immediately enjoy. We would be effectively biased against slow-burn content—against those ideas that lingered in the memory and revealed their depths slowly. It seems far more likely that a user will Like a tweet if it, say, expresses a view that they already agree with, than one that presents a challenging or subtle view that the user will have to wrestle with for a while. This is not because tweets somehow can’t be profound by their very nature. Rather, it is a feature of how Twitter’s interface captures the data to feed its metrics. A user might eventually come to appreciate a challenging tweet, but they are far less likely to go back and find that tweet, weeks later, to press Like. Slow appreciation is far less likely to be captured by the system and be counted towards that tweet’s score. And insofar as we have become gamified, then we will judge our own communicative success in terms of that recorded score. Second, Twitter scoring emphasizes the total number of Likes, rather than, say, the depth of engagement or lasting effect of a particular communication. This sort of problem plagues all sorts of large-scale value aggregations. Consider Matt Strohl’s criticism of the movie-review aggregator site Rotten Tomatoes. Rotten Tomatoes surveys the online reviews of a movie and reduces them each to a simple binary: was it a positive or negative review? And then Rotten Tomatoes produces an aggregate percentage of positive reviews. Notice, says Strohl, how this influences the results. A movie which strikes every critic as a little bit above average will score 100 per cent on Rotten Tomatoes and show up at the top of the heap. A movie which divides the critics—which some critics find utterly brilliant and other critics find baffling—will show up with a 50 per cent score, and appear, numerically at least, as a mediocrity. But great movies, says Strohl, rarely please everybody. Much of the most important art is difficult and utterly divisive. But the filtering and aggregating mechanism of Rotten Tomatoes ends up expressing a mathematical preference for more blandly agreeable material (Strohl 2017). Twitter’s aggregation method produces a similar effect. Sometimes, when I’m teaching, I say something to a whole class that I doubt will reach most students, but that I strongly suspect will resonate with one or two students. And often, that’s good enough for me. But Twitter scores each tweet with a simple binary measurement: either we Like a tweet, or we don’t; either we Retweet, or we don’t. This binary data collection screens off, at the input stage, any considerations of depth of impact or profundity of connection. Then Twitter automatically, and very visibly, aggregates the results of that binary input. Twitter’s scores make highly salient the number of users with positive reactions, while de-emphasizing the quality of any particular interaction. Insofar as we come to be motivated by Twitter’s scores, then the aim of our communication will be subject to a similar biasing effect as with Rotten Tomatoes. We will prefer those communications that appeal to the greatest number—even if that appeal is marginally positive—rather than those communications that might reach a smaller number more deeply.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
418 C. Thi nguyen Third, Twitter scoring aggregates user interests into a single monolithic statistic, which threatens to diminish the plurality of values for which we collectively communicate. Let’s assume, for the moment, that every Twitter user Likes those tweets in a way that accurately reflects their particular interests in communication. (In other words, assume that the last two problems don’t apply.) Even so, gamification will result in a homogenization of the values for which various actors communicate. Pre-gamification, each tweeting user will be motivated by their own particular values in communicating, giving us a diversity of communicators with different and distinctive motives. Such a diversity of interests is quite healthy, epistemically speaking. Cognitively diverse communities do better at figuring things out.11 Suppose, now, that the entire community succumbs to gamification and start chasing popularity by Twitter’s metric. Post-gamification, we have a body of communicators identically motivated to satisfy the same mixed populace. We’ve replaced a diversity of motivations with a motivational monolith. (Here’s one way to make the damage apparent: imagine the difference between a world of artists each motivated by their own aesthetic sensibility versus a world of artists each motivated to satisfy the largest number of their fellow artists.12) These three arguments approach, from different angles, the same central idea: that Twitter’s scoring mechanisms offer a simplified rendition of the rich plurality of our values. They refract our interests through the particular prism of Twitter’s information collection system, and then average the result. And, insofar as this simplification comes in an attempt to re-optimize the activity for pleasure, we should expect it to reduce that activity’s capacities to perform its other functions. At most, we might have hoped for a compromise between pleasure and the ori ginal goals of the activity; but such a compromise would require a careful, intentional design effort. And we have little evidence that Twitter’s design for gamification arose from any such careful attempt to support the plurality of communicative values. We have, instead, plenty of reason to think that its design features were heavily driven by an interest in increasing user engagement for the sake of profit. Here’s an analogy. Products that seem good and exciting in the store often turn out to be quite poor in quality. This superficiality is no accident; it is the result of systematic pressures. The function of these objects has drifted due to market forces. Where once the function of a shoe was to help us walk, now, so often, the function of the shoe is to get bought.13 And what gets a shoe bought is not its actual long-term quality, but the short-term appearance of quality. I am sug11 Hong and Page (2001, 2004) have famously demonstrated that cognitive diversity trumps cognitive ability in groups of deliberating individuals. For a rich application of these results to political communities, see Hélène Landemore’s (2013, 89–117) discussion of inclusive deliberation. 12 For a further discussion of the relationship between loyalty to personal aesthetic sensibility and a resulting landscape of creative diversity, see Nguyen (forthcoming). 13 This excellent formulation suggested by Alison Rieheld.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
How Twitter Gamifies Communication 419 gesting that something similar happens with gamified discourse on Twitter. Gamification changes discourse from serving the long-term values of communication to serving the function of gathering the most Likes and Retweets. Of course, gamification might not be dangerous if it is managed properly. Here, then, is a more sophisticated defense for the gamification optimist. Perhaps the simplified, gamified value isn’t actually replacing our original value, but simply functioning as a short-term heuristic for that value. Cognitively limited beings like us often need to focus on a short-term proxy for a complex value—like, say, using one’s increased running mileage as a proxy for health, or using one’s grades as a proxy for educational success. Managed properly, such heuristics can serve as an efficient and motivating proxy for some deeper and more complex value. It’s much easier, on a day-to-day basis, to aim at increasing mileage than it is to think about my health as a whole. But proper management is key. Heuristics, after all, are simplifications of the real thing. They are good heuristics insofar as they remain properly tethered to our deeper values. The successful use of a heuristic involves a complex process of management. We need to step back and reflect on whether using the heuristic is actually helping to achieve the underlying values. Increasing your running mileage might sometimes be a good proxy for fitness, but not when it brings irreversible knee damage. We need to adjust our heuristics when they drift. But Twitter’s design for gamification discourage appropriate management. First, Twitter makes the scoring system pervasive and highly salient through its user interface. The ready availability of this pre-fabricated, neatly packaged evaluation system may, by itself, discourage further reflection on one’s values. More importantly, Twitter’s metric is hard-wired into the system. Even if we managed to discover that we did, in fact, want to adjust the heuristic, that adjustment is hard to do on our own—because the scoring system is embedded in an externally controlled technology. Gamification works on us, in part, because of the ready availability of those quantified evaluations.14 In order to get the full pleasure on offer, we must assent to the particular measures that have been baked into the system. So, unless Twitter’s gamified metrics just happen to be the right heuristics to achieve our particular values in communication, and to do so in perpetuity, then taking on those metrics will impede our capacity to manage our proxy targets in light of our real values. And everybody who uses Twitter is pressured to take on precisely the same heuristic, with little room for personal tailoring. Twitter’s interface comes with a pre-fabricated, hard-wired measure, which points its users firmly in the direction of popularity—rather than allowing the user to search out the heuristic that best matches their own interests. Of course, a user could, conceivably, resist the pull of Twitter’s design for gamification and impose 14 See McGonigal’s (2011, especially 52–63) discussion of World of Warcraft and the importance of the visible and steady trickle of points and rewards.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
420 C. Thi nguyen their own self-created and self-managed heuristics on their tweeting. Twitter isn’t actually forcing a value change on us. The system design is seductive, but not compulsory. But Twitter offers us an intoxicating hedonic reward for changing our values along its pre-arranged lines. These thoughts about pre-fabrication point the way to another worry. To provide the kind of carefully engineered, automated, steady feedback that McGonigal praises, we usually need help from large institutions—like corporations and governments. Our choice of activities—and the way we engage in those activities— will then depend deeply on the fabrication efforts of large-scale institutions. Suppose that you find yourself with a strong preference for gamified activities over ungamified ones. When you engage in a particular gamified activity, you will, indeed, find yourself more motivated and more engaged. But you will also be restricted to choosing from the list of activities that institutions have chosen to gamify for you. Right now, for example, there are popular gamifications available for language learning, increasing your step counts, and tracking your weight loss. But there aren’t good gamifications for learning to appreciate complex poetry or becoming a better and more empathetic listener. And even if we think that the choice of gamifications isn’t insidious or manipulative, even if the institutions are well intentioned and just trying to help us lead our best lives, we will still find that the range of activities available to us will be sharply curtailed. Institutions will tend to produce gamifications of activities that more easily admit of technologized measurement. It is easier to gamify weight loss than it is to gamify deep aesthetic appreciation, because the former is easier to measure in an automated way. And when institutions gamify activities with more subtle and complex aims, like communication, then they will tend to tend to change those activities to make the aims more amenable to automated measurement. So, a life of gamification will tend to draw us towards those activities which have clearly measurable goals or those that can be transformed into something with clearly measurable goals. When we demand the pleasures of gamification in our activities, then the range of activities available to us diminishes—and the degrees of freedom we have within the activity also diminishes. Ironically, if we took the spirit of play to involve something like some kind of freedom or spontaneity with respect to one’s values and activities, then gamification turns out to be the opposite of play.15
3. How Gamification Changes Us So far, we’ve discussed how gamification can change the goals of the activity and so change how we conduct the activity. There is now a further question: how 15 This view of play is fairly common in the literature, but this precise articulation is from Maria Lugones (1987).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
How Twitter Gamifies Communication 421 might gamification change the users themselves? How might they transform the users’ lasting values? Much depends here on how the users motivationally interact with the scores. There are several different ways that interaction could go. First, users can treat Twitter as a game proper, taking on its goals temporarily for the sake of the pleasure during the activity. Second, they can internalize those scores and transform their long-term goals for communication. Third, they could keep the scores at motivational arm’s length, treating them only as a measure of some useful resource, but not permitting the scores to function directly in their motivation in any way. Let’s look at these various possibilities one by one. First, suppose one treats Twitter as a game proper. Let’s call such a person a game-playing user. Such a user temporarily adopts Twitter’s scores as their goal while they play Twitter, and then puts those goals away afterwards. In that case, their local adoption of those game-goals wouldn’t count as a long-term change in their value. And this practice would be perfectly harmless, if Twitter were really a game through and through—but Twitter is not. Real games have special properties. As Johann Huizinga famously put it, a game occurs in a separated place (a place he called “the magic circle”), where we take on alternate roles, and our actions took on alternate meanings (Huizinga 1971). Or, as Annika Waern puts it, games take place inside an interpretive frame, where we agree to reinterpret the meanings of the acts inside the game (Waern 2012). These are philosophically rich descriptions of a familiar phenomenon. Actions in games are screened off, in important ways, from ordinary life. When we are playing basketball, and you block my pass, I do not take this to be a sign of your long-term hostility towards me. When we are playing at having an insult contest, we don’t take each other’s speech to be indicative of our actual attitudes or beliefs about the world.16 And there are, in fact, conversational practices that are games, through and through. These are explicit, temporary practices where we conduct conversation while taking on specific goals, obeying specific obstructions, and taking on specific roles. There are structured games of deceit, intended to be played at parties or as tabletop games, like Mafia, Werewolf, The Resistance: Avalon, and Spyfall. There are also informal conversational games, like when we sit around and try to come up with the best insult about each other’s mothers. What makes the deceit in true games morally permissible is that we all know, going in, not to take the
16 The “magic circle” notion has come under significant fire, which I believe can be located in a famously overstated version presented in Rules of Play, an influential early game-design textbook. According to that textbook (at least, according to some readers), magic circles were impermeable membranes for meaning, across which no moral judgment or consequence could cross (Salen and Zimmerman, 95–7). I am relying here on what I take to be a more minimal and defensible version of the magic circle. For various defenses of more reasonable accountings of the magic circle, see Stenros (2012); Waern (2012); Nguyen (2020, 177–80).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
422 C. Thi nguyen in-game speech seriously.17 I don’t actually take your “Yo mama” insults to be presented as reliable testimony about the state of the world. Such games involve the voluntary and consensual entrance, by all the players, into an alternative game-space, where the players know to interpret the actions and communications inside the game under a special light—to not treat them as ordinary, real-world actions. But those aren’t the normative conventions around most of Twitter. The majority of Twitter presents itself as, and is taken to be, ordinary discourse. For the most part, we think that people on Twitter are representing their real beliefs and trying to make claims about the actual world.18 A user who approaches Twitter as a literal game, then, runs the risk of undermining the epistemic goods available to the other users. Suppose I’m on Twitter to actually communicate about ideas, and you’re playing a game with Twitter—saying whatever it takes to get the most Likes and Retweets for the sheer fun of it. If I don’t realize you’re playing a game, then I will be profoundly misinformed by your tweets. Those who approach Twitter explicitly as a game, but don’t clearly mark themselves as game players, are conversing in bad faith. They are presenting themselves as engaging in a discursive, epistemic practice while actually being guided by non-epistemic motives. And, insofar as other Twitter users take game-playing users as serious participants in sincere discourse, then these others will be mistaking gaming talk for serious testimony.19 Second, users could internalize the scores of Twitter, permitting their enduring goals to be influenced by Twitter’s scoring mechanism. Twitter makes this easy by making those scores so prominent and so pervasive. Twitter is a part of a larger phenomenon here, which we can call value capture.20 Value capture occurs when: (1) our natural values are rich, subtle, and hard-to-express;
17 For further discussion of the moral transition into game-life, see Weimer (2012); Kretchmar (2012); Nguyen (2017). 18 An exception is so-called Weird Twitter, which is a subnetwork devoted to irony and verbal game playing, largely its own, largely segregated network. But that is, of course, a specific exception— and not a particularly risky one, since Weird Twitter tweets are so bizarre and incomprehensible, that they are not likely to be mistaken for ordinary discourse. Weird Twitter is a game proper, with clear indicators that the users as just playing around. But things are different elsewhere on Twitter. I suspect that many other people are playing a game on the main stage of Twitter with political speech, which is a far more dangerous affair. 19 To put this into the technical language of the epistemology of testimony: I take Twitter to be a context in which most tweets are treated as “assertions.” To use Elizabeth Fricker’s account: when S asserts that P to an audience H, they thereby vouch for the truth of P to H, presenting P as being so, such that H can form belief that P on S’s say-so. An assertion that P represents the asserter as knowing P (Fricker 2006). Game-playing users, then, are presenting non-assertions into an assertoric context, where others can be reasonably expected to treat them as assertions. 20 I introduced the notion of value capture in Nguyen (2020, 189–215), though my views have, I think, matured since that earlier sketch.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
How Twitter Gamifies Communication 423 (2) we are placed in a social or institutional setting which presents s implified, typically quantified, versions of our values back to ourselves; (3) the simplified versions take over in our motivation and deliberation. Some examples include: starting to exercise for the sake of your health, then getting captured by FitBit and coming to just care about your daily step-counts; going to school for the sake of a good education and coming out obsessed with your grade point average (GPA); becoming a pre-law for the sake of public interest and legal activism, and then coming to care more about getting admitted to the best law school according to the US News & World Report’s law school rankings.21 And, of course, going onto Twitter for the sake of communication, connection, and shared understanding—and coming out obsessed with maximizing Likes, Retweets, and Follower counts. And, obviously, a high step-count isn’t the same as good health; a high GPA isn’t the same as a good education; and high Twitter Likes aren’t the same as connection or collective understanding. Value capture occurs when our values undergo a long-term and enduring simplification, as guided by the external metrics provided by institutions and technologies. The worry here isn’t that our values couldn’t ever be expressed in quantified form, in principle. Rather, it’s that the kind of metrics, measures, and gamified scoring that we typically encounter in our life with bureaucracies, institutions, and corporations are almost always radical simplifications of the values they claim to be measuring. Those simplifications may have certain uses in administration, management, or large-scale scientific data collection. But what makes them useful for those functions is, in fact, their very simplification. It’s useful here to borrow from a nearby discussion: that of the simplifications involved in bureaucratic quantifications. As Theodore Porter puts it, institutional quantification is driven by an interest in making information highly portable. Rich, nuanced, qualitative information is difficult to manage from any sort of informational center. We need to strip out the context-sensitive details and nuance in order to transmit it easily between contexts (Porter 1996). This is why such quantification is beloved of centralized bureaucracies, which need to pass information to distant managers and up many levels in the hierarchy of administration (Scott 1998).22 This context-stripping standardization also allows us to aggregate the information arithmetically. Think, for example, of how teachers assess students. Teachers could offer each student rich and individualized commentary about the strengths and weaknesses of their academic work. Such individualized commentary would
21 Wendy Espeland and Michael Sauder’s Engines of Anxiety is a thorough—and deeply alarming— account of how the USN&WR’s law school rankings have profoundly changed student motivations (Espeland and Sauder 2016). 22 I am also influenced here by Perrow (2014) and Merry (2016).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
424 C. Thi nguyen be vastly useful to the students. But such individualized commentary is incredibly hard to aggregate and manage by upper-level administrators. How is an administrator supposed to compare the portfolio evaluations of the art department with the mathematical performance of the statistics students? So, teachers are asked to provide quantified grades for their students, which can then easily be averaged across classes for one student, and across all the students in a department, university, or school district. Quantified grades strip out much of the most important information. But that context stripping renders information into a standardized form that can be operated upon arithmetically. This allows managers of massive, sprawling institutions to bring their entire domain into view—by putting information in standardized form and then aggregating it. Institutional life exerts a pressure on information, pushing it towards quantified, aggregable form. Notice that these forces do not typically arise to support our individual interests, but instead the interests of management and large-scale administration. But, problematically, those quantifications appeal to our motivations precisely because of their apparent clarity. And once we offer simple, quantified metrics for success, those metrics take over in the motivations of so many people.23 A similar pressure occurs with overt gamifications, especially ones in an automated, technological context—though the motivations for simplifying may be slightly different. In order to create the motivational rewards of gamification, we need to provide a score. In order to provide that score, we need to offer a reliable scoring mechanism. And in a large-scale, technologized context like Twitter, that scoring mechanism needs to function automatically. Twitter can’t offer a score based on quality of engagement, empathy, or depth of thought. It can only score us on what is easily legible to its systems: like whether or not somebody clicks on Like. Such scores present a game-like motivational lure. But because they can also be integrated arithmetically, they can be used to generate more gamifications down the line. Consider, for example, how FitBit’s scores nest. FitBit provides me with a daily score for my walking. But that score can be averaged, so I can also get a score for my walking success each week, and each month. Since other people’s scores can also be averaged, the system can automatically generate rankings and leaderboards, each of which provides another game-like motivational boost. Similarly, Twitter’s scores, once rendered into quantified form, can be extracted and used to generate other scores.24 For example, the explicitly gamified social 23 For excellent case studies into the motivational pull of quantifications, see Sally Engle Merry’s (2016) study of the use of simplified metrics and indicators in motivating political action; and Espeland and Sauder’s (2016) study of the effects of law school rankings on the motivations of students and donors. This brief sketch of value capture, quantified bureaucracy, and seductive clarity here will be developed in future work. 24 Lupton and Smith’s (2017) recent study of quantified self-tracking shows that many self-trackers are extremely interested in the exportability of self-tracking data—of the ability to send FitBit step- counts to a spreadsheet or more macroscopic health-tracking programs. Their approach, however, is
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
How Twitter Gamifies Communication 425 network Empire.kred creates a second-order game out of social media scores. Empire.kred is a virtual stock market, where individuals are the stocks. Individuals can invest in each other using the game’s virtual currency, $Eaves. Their stock value is based on their social media power, as modified through investments. Empire.kred harvests various scores from other social media networks, like Twitter and Facebook, and then aggregates those scores to drive its virtual stock market.25 This is possible because those social media networks have already digested evaluations of user success into a standardized and portable format: a numerical score. The problem with value capture can be put most clearly if we help ourselves to an assumption. There is, it is often thought, a natural aim to belief. Belief aims at the truth.26 We can be tempted by other motivations to abandon that aim: to believe what will feel pleasant or make things easy for us. But to do so is to abandon the natural aim of belief; it is to subvert the activity of believing. The activity of earnest discourse also seems to have a natural aim, which is the collective pursuit of truth. We aim to express what we think of as true, and to question and challenge each other’s expressions, as part of our quest to understand the world. But gamification tempts us to change our goals—to aim at expressions which maximize our score, rather than those which aid our collective understanding. And it promises to reward us for that change with pleasure. Twitter tempts us to subvert the activity of earnest conversation for hedonistic reasons. Besides game-playing users and value-captured users, there is a third possibility: that users could treat the scores of Twitter as simple reports of some instrumental resource, useful for the pursuit of further ends. They treat Twitter’s numbers, not as setting a goal, but merely as useful data. Let’s call such person a value-independent user. Such a user has avoided internalizing the scores of Twitter in any way. They have avoided gamification. Here’s what that might look like. Suppose one wanted public influence. A resource for public influence is having a Twitter account with a wide number of followers—and tweets which are heavily retweeted will reach a large number of people. So, one could aim for high scores simply as an approximate measure of that instrumental resource. Such a value-independent user wouldn’t have any form of change of value or goals, either short-term, or long-term. They also wouldn’t be subject to the motivational boosts that arise from more fully inhabiting the scores of Twitter. They would be holding those scores at phenomenal arm’s length. Such a user, then, would be free of the more pernicious effects of value capture and game playing. They have resisted Twitter’s invitation to gamification. more tuned to the data-gathering side and pays less attention to the motivational possibilities of gamification. 25 See https://play.empire.kred. 26 This idea has its most influential statement with Williams (1970). For more recent discussion, see Velleman (2000); Wedgewood (2002).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
426 C. Thi nguyen Thinking about the value-independent user helps us get clearer on what’s wrong with the value-captured user. The value-independent user manages the scores, where the value-captured user is driven by the scores. Consider, by way of analogy, two relationships you could have with money. First, you could view it as an instrumental resource, to be collected in pursuit of some other value. Second, you could treat it as an enduring end, to be pursued for its own sake. Somebody who sought money as an instrumental resource would manage their pursuit of money in view of their larger ends. Somebody who pursued money as an instrumental resource to happiness wouldn’t take that high-paying job that would destroy their happiness. They would manage their pursuit of money, making sure to pursue money only to the extent that it actually helped their happiness. The person who pursues money for its own sake, however, has no such guiding purpose with which to manage their pursuit of the greatest pile.27 Similarly, consider a user who comes to Twitter for the sake of, say, social progress, and seeks Followers and Retweets simply as an instrumental resource for their mission. They have an external standpoint from which to manage their pursuit of Followers and Retweets. They wouldn’t say anything in order to go viral, for many such things they could say would likely undermine their larger purpose. But the person who has been fully value-captured by Twitter’s scores has no such limitation. They will be driven to say whatever it takes to go viral and get those points. For the value-independent user, Twitter’s scores are merely a means. But for the value-captured user, Twitter’s scores have become the end. The act of communication itself has been instrumentalized to the end of Twitter scores. Rather than using Twitter scores to advance their independent values in communication, they have changed the nature of their communication to advance their pursuit of Twitter scores.
4. Changing Values The key notion here is the idea that gamification problematically instrumentalizes our goals. The notion of instrumentalization will be useful in understanding some of the socially toxic behavior which seems to bloom on Twitter. But first, we’ll need a clearer picture of the notion of instrumentalization. For that, we’ll need to look in greater depth at how and why we fashion new goals for ourselves in games and gamification. My account of games shows that the player has a rather extraordinary form of agential fluidity. During a game, a player takes on an alternate agency with
27 This paragraph arises from discussions with Aaron James.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
How Twitter Gamifies Communication 427 alternate goals. That agency has been engineered to provide satisfaction for the player who adopts it. Gamification works in a similar way—it offers us various satisfactions, in exchange for shifting our goals along its engineered lines. Both games and gamification involve instrumentalizing our goals. This is unproblematic in games, but deeply problematic in gamification. Why? Because games are a very peculiar and distinctive sort of activity, and gamification doesn’t share in some of the most important features. The best account of the special nature of games comes from Bernard Suits’ marvelous attempt to define ‘game’. Suits says that to play a game is to voluntarily take on obstacles to make possible the activity of struggling to overcome them. In other words, in a game, the obstacles are much of the point. We try to run a marathon, and what it is to run a marathon is to try to get to a certain place while submitting to various restrictions. We must run by our own power only—no short-cuts, no taxis. Those restrictions help constitute various obstacles for our efforts. But, says Suits, our devotion to these restrictions shows that we are not motivated simply by the independent value of crossing the finish line. If we just cared about being at that particular point in space, in and of itself, we would take the most efficient means to that end—like a taxi. The fact that we are willing to place extra, unnecessary inefficiencies in our way indicates that our interest is not in actually achieving the goal in and of itself, but in achieving it inside certain specified restrictions. Our interest is to achieve the goal by way of a particular, constructed form of activity. As Suits puts it, in a game, the restrictions help constitute the very activity we are interested in performing. What it is to run a marathon is to run a certain distance under one’s own power. If we took a taxi, we wouldn’t be running a marathon at all (Suits and Hurka 2014).28 As I have argued elsewhere, Suits’ account reveals the possibility of a very peculiar motivational structure.29 There are two different motivational structures for playing a game. One could be an achievement player, who plays the game for the value of winning. Or one could be a striving player, who takes on a temporary interest in winning for the sake of engagement in a struggle. (One could also play for both motivations, in varying proportions.) Striving play is a very special motivational structure; it involves a motivational inversion from ordinary life. In ordinary life, we take the means for the sake of achieving the ends. But in game
28 I don’t think gamified activities count as games proper for reasons that are tangential to the t opics for this chapter. Briefly, according to Suits’ definition, which I largely endorse, games are activities where the goal of the game is partially constituted by the designated restrictions on that goal. What it is to make a basket in basketball is, in part, constituted by the player’s having obeyed the dribbling restriction. For a further discussion of this point, see Nguyen (2020, 27–73). The goals in gamified activities are not restriction-constituted in this way. 29 This is a very brief presentation of one argument for the existence of striving play, among several I have offered elsewhere. The most detailed version of this analysis occurs in Nguyen (2020, 27–73).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
428 C. Thi nguyen life, we select the ends for the sake of the means. We take on a temporary end, and we submerge ourselves in it.30 When my spouse and I play games, we want to both have a good time, so we look for games that we’re both relatively good at. We can see the fact that we’re both striving players by how we manipulate our capacity to win in the long term. Suppose that we have found a game at which we are perfectly matched and are having a lovely set of intense gaming sessions with. Suppose one of us finds a strategy guide to that game. If that person were to read it by themselves, they would pull ahead and start winning. If we were achievement players, then we each should want to read that guide. But we don’t, and it is perfectly reasonable that we don’t. We are willing to suppress our capacity to win in the long term— even though we try, with all our might, to win during the game. Our extra-game behavior reveals that we aren’t actually interested in winning in any enduring sense. Our interest in winning is merely something we temporarily adopt, in order to create the experience of that delicious struggle. And the goal we pursue in the game is often disconnected from our enduring goals and ends—at least, disconnected in the usual linear sense. In many games, our real purpose is to have fun, but we can only have fun by trying to win. But we don’t really care about winning; we just adopt a temporary interest in winning so that we can engage in the fun activity of trying. But after the game is through, we can dispense with that interest in winning. For example: I can start a game of Charades at a party for fun. In order to have fun, I have to genuinely try to achieve the goals of the game—to communicate concepts through gestures, without speech. But after the game, I discard that desire. After all, if I lost at Charades, but we all had a good time together, then I have achieved my true purpose. Only an especially poor sport would think the whole enterprise a failure because they had lost at Charades. So, when we justify our game goals in striving play, we do not do so in reference to the value of the goal itself, or to what follows from it. We justify the game’s goals by pointing to the value of the activity of pursuing those goals. Thus, striving play instrumentalizes our adoption of goals. In striving play, we adopt a goal not for its own value. Our adoption of a game-goal is justified in terms of the activity of pursuit that goal structures. Here, then, is a key difference between games proper and the gamification of non-game life. In striving games, the goals of games are temporary. More import antly, they are disconnected from the network of our enduring ends. In striving play, my in-game goal is winning, but I don’t actually care about winning in the long term. I achieve my real purpose—fun, satisfaction, exercise—by pursuing 30 Some philosophers may protest that I have posited the impossible: that we can desire at will. Please see Nguyen (2019, 451–5) for my argument that striving play reveals that we can, in fact, desire at will.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
How Twitter Gamifies Communication 429 the win, and not by actually winning. And this is why it is perfectly permissible for game designers to change the goals of game-activity. Game-goals can be made as simple and narrowed as is convenient because they aren’t directly attached to our enduring ends. Game designers are changing the play-goals that guide an artificial activity, which has been screened off from many real-world consequences. But gamification is an entirely different matter. In gamification, the designers are instrumentalizing the goals of our real-life activities. FitBit, by gamifying exercise, invites us to change our goals for our health and fitness. And Twitter, by gamifying discourse, invites us to change our goals for conversation, communication, and declaration. Instrumentalizing one’s goals is fine in striving games, because the goals in games were never valuable, in and of themselves, in the first place. But in real-life activity, the goals are often independently valuable. So, when we gamify those activities and instrumentalize those ends for the sake of pleasure, we risk losing sight of the real importance of the activity. Twitter’s gamification changes our communicative goals away from understanding, connection, and the collective pursuit of truth and bends them towards something much more impoverished.
5. Twitter and Toxicity I’ve discussed elsewhere two problematic social phenomena associated with polarized discourse: echo chambers and moral outrage porn. Both of these phenomena seem to flourish on social media. We are now in the position to offer the beginnings of an explanation for this relationship. Gamification, echo chambers, and moral outrage porn all share a common central thread: a willingness to instrumentalize what ought not be instrumentalized. Let’s first get clearer on these other phenomena. First: as I’ve argued elsewhere, echo chambers are best understood as structures of manipulated trust. Echo chamber members have been systematically taught to distrust everybody on the outside (Nguyen 2018). To put it more formally, an echo chamber is a social structure in which: (1) one must subscribe to a certain belief system to be a member; (2) that belief system includes the belief that all non-members are untrust worthy, and all members trustworthy. Thus, echo chambers inculcate a radical trust disparity between members and non-members. The belief system includes some explanation for why everybody on the outside is untrustworthy. In the modern landscape, those explanations often take the form of conspiracy theories—such as “The liberal media is in
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
430 C. Thi nguyen the grip of George Soros and totally corrupt.” And the trust disparity is selfreinforcing. The more you trust your fellow echo chamber members, the more their agreement will confirm your shared belief system. And the more you confirm that belief system, the more you will trust your fellow members and distrust outsiders. Compare echo chambers to a nearby phenomenon: that of epistemic bubbles. An epistemic bubble is a social structure where insiders aren’t exposed to views on the outside. Despite the superficial similarity, epistemic bubbles and echo chambers work through entirely different mechanisms. In an echo chamber, inside members may have plenty of exposure to outside views, but outside voices have been undermined. Epistemic bubbles are structures of bad connectivity; echo chambers are structures of manipulated credence. In an epistemic bubble, outside voices aren’t heard; in an echo chamber, outside voices have been systematically discredited. Importantly, I’ve argued, many communities with problematic belief systems have been misdiagnosed as epistemic bubbles. But actually, they are mostly the result of echo chambers. It isn’t that climate change deniers, for example, are simply unaware of what climate change scientist think, or the standard publicly available arguments for climate change. They are, for the most part, quite well acquainted with those arguments and conclusions. It is that they think that the institutions of climate change science have been systematically corrupted and are untrustworthy. This helps to explain the intractability of climate change denialists. Since an epistemic bubble works through simply omitting outside voices, we should be able to shatter one simply by exposing an insider to more voices and more viewpoints. We should expect epistemic bubbles to go down with the first contact with the missing evidence. But echo chamber members are pre-prepared for encounters with external viewpoints and armed with explanatory mechanisms to dismiss those other voices. Echo chambers are far more robust. Why might one enter into an echo chamber? In my earlier discussion, I focused on the possibility that one might be raised in an echo chamber, and, through no fault of one’s own, been trapped in an errant system of trust. But here, I would like to focus on another possibility: that some people choose to enter echo chambers because being in an echo chamber is more comfortable and more pleasurable. Life outside of an echo chamber is full of all kinds of cognitive difficulties. We must constantly struggle with conflicting evidence and unexplained phenomena. And we are confronted, over and over again, with evidence of our own cognitive fallibility. These confrontations humble us—which is good for us, but also quite painful. Echo chambers banish all that epistemic friction.31 They remove, through distrust, the impact of disagreeing voices. Instead of having to cope with new 31 I am influenced here by José Medina’s (2012) account of epistemic resistance, though I emphasize the idea that the experience of and the processing of epistemic resistance is comfortable, and synthetic epistemic environments engineered to be resistance-less are quite pleasurable.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
How Twitter Gamifies Communication 431 e vidence, echo chambers typically present their members with clear, coherent stories about the world. Instead of the humbling confrontation with the evidence of one’s errors, echo chambers offer their members the joys of unanimity and uninterrupted confidence. And notice: these joys are very much akin to the joys of value clarity that we found in games. And both forms of joy emerge from similar engineered conditions. Games involve re-designing the agent’s goals and abilities for pleasure. Echo chambers involve re-engineering their members’ belief system and trust settings for pleasure. And echo chambers are dangerous because they re-engineer, not some temporary and segregated belief system, but real-life belief systems which govern real-life action. We can now see a higher-level similarity between gamification and echo chambers. In gamification, we instrumentalize our real-life goals. In particular, the gamification of Twitter involves instrumentalizing the goals associated with discourse. Gamification involves, to a significant degree, abandoning the aim of truth and understanding, and taking on a simpler goal—where that goal was engineered for the sake of pleasures of value clarity. Echo chambers also involve instrumentalizing our belief systems, abandoning the aims of having the beliefs that are true, and trusting the people that are reliable. And, in exchange for abandoning these epistemic aims, echo chambers offer their members the pleasures of confidence, simple coherence, and unity. There is an interesting complexity in the instrumentalization here. There are two levels of explanation for these simplifying re-designs. It seems plausible, for both Twitter and for many real-world echo chambers, that they are intentionally designed by an external agent. The re-engineering involves instrumentalization at two different levels: at the level of design, and at the level of adoption. Plausibly, Twitter’s makers consciously designed it for pleasure and addictiveness, for the sake of profit. So, there are two instrumentalizations here. First, Twitter’s makers are designing for gamification for the sake of profit, which they pursue by making their design seductively pleasurable to its end users. And second, those users are accepting the seduction, and gamifying their discourse for the sake of pleasure. At both levels, we find people willing to forsake the original goals of discourse for some other end. Similarly, many echo chambers are plausibly designed for political control.32 To that end, designers have a reason to engineer their belief system to be as pleasurable as possible. Once again, there are two instrumentalizations: designers create a belief system for the sake of political control, which involves designing them
32 This view might strike some as cynical. This is, however, the picture offered by Jamieson and Cappella (2010) in their meticulously researched account of Rush Limbuagh and Fox News’s intentional construction of an echo chamber. Their book, Echo Chamber: Rush Limbaugh and the Conservative Media Establishment, is one of the best early analyses of echo-chamber structures, and is the source for my own account of echo chambers.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
432 C. Thi nguyen to be pleasurable to their users. Then users accept those belief systems for the sake of that engineered pleasure. And, once again, at both levels, we find people willing to create or adopt belief systems for reasons that bear, not on their relationship to truth, but to some other end. Let’s turn now to the second toxic phenomenon: moral outrage porn. In earlier work, Bekka Williams and I offer an account of “porn” in the generic sense. We mean to describe the new, modern usage, which includes things like “food porn,” “real estate porn,” and “closet porn.” We propose that a representation is used as generic porn when it is engaged with for the sake of a gratifying reaction, freed from the usual costs and consequences of engaging with the represented content. For example: food porn is pictures of food which people look at to get immediate gratification, while avoiding the calories, cost, and hassle of eating the depicted food. Real estate porn is pictures of expensive, well-maintained homes, which people look at for immediate gratification, while avoiding the costs and hassle of buying and maintaining those actual homes. This account helps us get a grip on an important phenomenon: moral outrage porn. Moral outrage porn is representations of moral outrage, which people engage with for the immediate gratifications of feelings of moral outrage—for the pleasures of feeling smug, secure, and confident in the total wrongness of the other side. And they do so while avoiding the costs and consequences of genuine moral engagement: like the pains of struggling to be morally sensitive, the efforts of seeking the right moral beliefs, and the exhaustion of real moral action. We think it quite clear that social media is suffused with moral outrage porn. And moral outrage porn is quite dangerous. If one is interested in using moral outrage porn for pleasure, one will have an incentive to adopt, not the right moral system, but the one that is easiest to crank for pleasure. One will likely be tempted to, say, adopt a simple and absolute moral system, that will give one the easiest access to the pleasures of smug condemnation. Crucially, our claim isn’t that moral outrage is bad. Real moral outrage is crucial. Moral outrage, when it emerges from a well-tuned moral sensibility, helps us to register injustice and motivates us to end it. The very problem is that moral outrage porn threatens to corrupt the real thing. The proper target of moral outrage is the genuinely outrageous. But when we use moral outrage porn, we use our own moral outrage for pleasure. And so, we are incentivized to change our moral belief system—to ignore the truth and adopt those beliefs that will give us the most pleasurable outrage. Moral outrage porn invites us to instrumentalize our moral beliefs (Nguyen and Williams 2020). So, moral outrage porn and echo chambers often occur together, and they both seem to flourish on social media. Why might that be? We now have the beginnings of an explanation. All of these phenomena involve hedonistic instrumentalization, where we take an attitude or mental state and modify it away from its appropriate target in exchange for pleasure.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
How Twitter Gamifies Communication 433 Why might a similarity of motivational structure lead to frequent co ccurrence? I suggest that these various hedonistic instrumentalizations occur o together because they appeal to the same sorts of motives. In other words, the co-occurrence of gamification, echo chambers, and moral outrage porn are not best explained by features of the individual phenomena themselves, but in terms of the character of their likely adopters. In all of these cases, maintaining the attitudes towards their appropriate aim takes work. Somebody willing to abandon an attitude’s appropriate aim and instrumentalize it for pleasure in one place is likely to do it in another. To put it another way, gamification, echo chambers, and moral outrage porn go together like junk food. Different kinds of junk food are unhealthy in different ways—some are too high in salt, some too high in fat, some too high in sugar. But the reason they are often consumed together is that they are all likely to be consumed by somebody who is willing to trade off health and nutrition in return for a certain kind of quick pleasure. The same is true of gamification, moral outrage porn, and echo chambers. They are all readily available sources of a certain quick and easy pleasure, available to anybody willing to relax with their moral and epi stemic standards. Next, think about things from the point of view of the system designer. Imagine yourself into the shoes of a hostile manipulator. Let’s say you wanted to get people under your political sway. You’d want to design a belief system that was as max imally catchy and sticky as possible. Here’s one way you could do it. First, you could design a belief system that included provisions to distrust all outsiders who didn’t share the belief system. You could make that belief system utterly clear and coherent, all the better to please its adopters. In other words, you’d design an echo chamber. Second, you could rig the belief system with the appropriate amount of moral certainty and superiority over outsiders, so as to provide all the pleasures of moral condemnation. In other words, you’d fill it with moral outrage porn. Third, if it were available, you’d want to entrench that belief system in a communication platform that awarded its users plenty of clear, direct affirmation for agreeing with each other. For that, the gamified setting of Twitter will do quite nicely. Echo chambers instrumentalize our trust; moral outrage porn instrumentalizes our morality; and gamification instrumentalizes our goals.
References Bogost, Ian (2011). “ ‘Gamification Is Bullshit.’ ” The Atlantic (August 9, 2011), www. theatlantic.com/technology/archive/2011/08/gamification-is-bullshit/243338. Brady, William J., Julian A. Wills, John T. Jost, Joshua A. Tucker, and Jay J. Van Bavel (2017). “Emotion shapes the diffusion of moralized content in social networks.” PNAS 114(28): 7313–8.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
434 C. Thi nguyen Chou, Yu-kai (2015). Actionable Gamification: Beyond Points, Badges and Leaderboards. Fremont, CA: CreateSpace Independent Publishing Platform. Espeland, Wendy Nelson and Michael Sauder (2016). Engines of Anxiety: Academic Rankings, Reputation, and Accountability. New York: Russell Sage Foundation. Frankfurt, Harry G. (2005). On Bullshit. Princeton, NJ: Princeton University Press. Fricker, Elizabeth (2006). “Second-Hand Knowledge.” Philosophy and Phenomenological Research 73(3): 592–618. Frost-Arnold, Karen (2014). “Trustworthiness and Truth: The Epistemic Pitfalls of Internet Accountability.” Episteme 11(1): 63–81. Gabrielle, Vincent (2018). “How Employers Have Gamified Work for Maximum Profit.” Aeon Magazine (October 10, 2018), https://aeon.co/essays/how-employershave-gamified-work-for-maximum-profit. Hong, Lu and Scott Page (2001). “Problem Solving by Heterogenous Agents.” Journal of Economic Theory 97(1): 123–63. Hong, Lu and Scott Page (2004). “Groups of Diverse Problem Solvers Can Outperform Groups of High-Ability Problem Solvers.” Proceedings of the National Acdaemy of Sciences of the United States 101(46): 16385–9. Huizinga, Johan (1971). Homo Ludens: A Study of the Play-Element in Culture, reprint edn. Boston, MA: Beacon Press. Jamieson, Kathleen Hall and Joseph Cappella (2010). Echo Chamber: Rush Limbaugh and the Conservative Media Establishment. Oxford: Oxford University Press. Kretchmar, Scott (2012). “Competition, Redemption, and Hope.” Journal of the Philosophy of Sport 39(1): 101–16. Landemore, Hélène (2013). Democratic Reason. Princeton, NJ: Princeton University Press. Lugones, Maria (1987). “Playfulness, ‘World’-Travelling, and Loving Attention.” Hypatia 2(2): 3–19. Lupton, Deborah and Gavin J. D. Smith (2017). “‘A Much Better Person’: The Agential Capacities of Self-Tracking Practices.” SSRN Scholarly Paper ID 3085751. Rochester, NY: Social Science Research Network (December 11, 2017), https://papers.ssrn. com/abstract=3085751. Madrigal, Alexis C. (2013). “The Machine Zone: This Is Where You Go When You Just Can’t Stop Looking at Pictures on Facebook.” The Atlantic (July 31, 2013), www. theatlantic.com/technology/archive/2013/07/the-machine-zone-this-is-where-you-gowhen-you-just-cant-stop-looking-at-pictures-on-facebook/278185. Marwick, Alice and danah boyd (2011). “To See and Be Seen.” Convergence: The International Journal of Research into Media Technologies 17(2): 139–58. McGonigal, Jane (2011). Reality Is Broken: Why Games Make Us Better and How They Can Change the World. New York: Penguin Books. McLuhan, Marshall (1964). Understanding Media: The Extensions of Man. New York: Signet Books.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
How Twitter Gamifies Communication 435 Medina, José (2012). The Epistemology of Resistance: Gender and Racial Oppression, Epistemic Injustice, and Resistant Imaginations. Oxford/New York: Oxford University Press. Merry, Sally Engle (2016). The Seductions of Quantification: Measuring Human Rights, Gender Violence, and Sex Trafficking. Chicago, IL: University of Chicago Press. Miller, Boaz and Isaac Record (2013). “Justified Belief in a Digital Age: On the Epistemic Implications of Secret Internet Technologies.” Episteme 10(2): 117–34. National Public Radio (2014). “Stuck in The Machine Zone: Your Sweet Tooth For ‘Candy Crush.’ ” NPR.Org (June 7, 2014), www.npr.org/sections/alltechconsidered/ 2014/06/07/319560646/stuck-in-the-machine-zone-your-sweet-tooth-for-candycrush. Nguyen, C. Thi (2017). “Competition as Cooperation.” Journal of the Philosophy of Sport 44(1): 123–37. Nguyen, C. Thi (2018). “Echo Chambers and Epistemic Bubbles.” Episteme 17(2): 1–21, https://doi.org/10.1017/epi.2018.32. Nguyen, C. Thi (2020). Games: Agency as Art. New York: Oxford University Press. Nguyen, C. Thi (2019). “Games and the Art of Agency.” Philosophical Review 128(4): 423–62. Nguyen, C. Thi (forthcoming). “Trust and Sincerity in Art”. Ergo. Nguyen, C. Thi and Bekka Williams (2020). “Moral Outrage Porn.” Journal of Ethics and Social Philosophy 18(2): 147–72. Perrow, Charles (2014). Complex Organizations: A Critical Essay. Brattleboro, VT: Echo Point Books and Media. Porter, Theodore (1996). Trust in Numbers. Princeton, NJ: Princeton University Press, https://press.princeton.edu/titles/5653.html. Rini, Regina (2017). “Fake News and Partisan Epistemology.” Kennedy Institute of Ethics Journal 27(S2): 43–64, https://doi.org/10.1353/ken.2017.0025. Salen, Katie and Eric Zimmerman (2004). Rules of Play: Game Design Fundamentals. Cambridge, MA: The MIT Press. Schull, Natasha Dow (2012). Addiction by Design. Princeton, NJ: Princeton University Press. Scott, James C. (1998). Seeing Like a State: How Certain Schemes to Improve the Human Condition Have Failed. New Haven, CT: Yale University Press. Seymour, Richard (2019). “The Machine Always Wins: What Drives Our Addiction to Social Media.” The Guardian (August 23, 2019), sec. Technology, www.theguardian. com/technology/2019/aug/23/social-media-addiction-gambling. Stenros, Jaakko (2012). “In Defence of a Magic Circle: The Social and Mental Boundaries of Play.” www.digra.org/wp-content/uploads/digital-library/12168. 43543.pdf. Strohl, Matt (2017). “Against Rotten Tomatoes.” Aesthetics for Birds (September 21, 2017), https://aestheticsforbirds.com/2017/09/21/against-rotten-tomatoes.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
436 C. Thi nguyen Suits, Bernard and Thomas Hurka (2014). The Grasshopper—Third Edition: Games, Life and Utopia. Peterborough, Ontario: Broadview Press. Sunstein, Cass R. (2009). Republic.Com 2.0. Princeton, NJ: Princeton University Press. Tufekci, Zeynep (2018). “It’s the (Democracy-Poisoning) Golden Age of Free Speech.” Wired, https://www.wired.com/story/free-speech-issue-tech-turmoil-new-censorship. Tufekci, Zeynep (2017). Twitter and Tear Gas: The Power and Fragility of Networked Protest. New Haven, CT: Yale University Press. Velleman, David (2000). “On the Aim of Belief,” in The Possibility of Practical Reason. Oxford: Oxford University Press, 244–81. Walz, Steffan P. and Sebastian Deterding (2015). The Gameful World:Approaches, Issues, Applicaitons. Cambridge, MA: The MIT Press. Waern, Annika (2012). “Framing Games.” DiGRA Nordic ’12: Proceedings of 2012 International DiGRA Nordic Conference 10, www.digra.org/wp-content/uploads/ digital-library/12168.20295.pdf. Wedgwood, Ralph (2002). “The Aim of Belief.” Philosophical Perspectives 16: 267–97. Weimer, Steven (2012). “Consent and Right Action in Sport.” Journal of the Philosophy of Sport 39(1): 11–31. Williams, Bernard (1970). “Deciding to Believe,” in Problems of the Self. Cambridge: Cambridge University Press, 136–51.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 19/03/21, SPi
18
The Epistemic Dangers of Context Collapse Online Karen Frost-Arnold*
1. Introduction One of the important questions for applied social epistemology of the internet is to what extent the internet creates new epistemological questions and challenges. Social epistemologists have investigated the new modes of journalism in the blogosphere (e.g. Coady 2012; Goldman 2008), collaborative knowledge production on Wikipedia (e.g. de Laat 2010, 2015; Fallis 2011; Frost-Arnold 2014a, 2018; Magnus 2009; Simon 2010), and algorithmic knowledge dissemination through web search (e.g. Miller and Record 2013), among many other topics. In this chapter, I analyze a feature of the internet that has not previously captured the attention of social epistemologists—context collapse. I argue that context collapse presents some important epistemic challenges and I sketch a virtue epistemology that can provide a normative framework for addressing some of the problems context collapse poses. Context collapse is the blurring or merging of multiple contexts or audiences into one (boyd 2008; Marwick and boyd 2011). Internet theorist danah boyd is often credited with first observing that context collapse is one of the key features of the internet by noting that “the lack of spatial, social, and temporal boundaries makes it difficult to maintain distinct social contexts” (boyd 2011, 49). Facebook users may intentionally collapse multiple contexts of social relations into one audience by sharing information with all their Facebook friends, which may include family members, close friends, acquaintances, and work colleagues. Context collapse can also occur without one’s consent, a phenomenon which Davis and Jurgenson call ‘context collision’ (cf. Davis and Jurgenson 2014). For instance, a Facebook post that I imagine will only be read by my Facebook friends can be captured as a screenshot and shared via blogs, Twitter, or Reddit to millions of strangers.
* I am grateful to the audience at the 2017 Penn-Rutgers-Princeton Social Epistemology workshop for discussion of an earlier draft of this chapter. Section 4 is particularly indebted to their input.
Karen Frost-Arnold, The Epistemic Dangers of Context Collapse Online In: Applied Epistemology. Edited by: Jennifer Lackey, Oxford University Press (2021). © Karen Frost-Arnold. DOI: 10.1093/oso/9780198833659.003.0018
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
438 Karen Frost-Arnold It is worth clarifying what I mean by claiming that context collapse is a feature of the internet that poses new epistemic challenges. I do not mean that context collapse did not exist before the internet. As boyd herself points out, context collapse is a regular part of many offline social interactions; for example, weddings often bring together multiple different social groups (boyd 2011, 51). Additionally, Meyrowitz analyzes how broadcast media facilitates context collapse for individ uals who appear on television (Meyrowitz 1985). What boyd argues is new about context collapse via social network sites is its scope—before the internet, context collapse was “never so generally experienced” (boyd 2011, 49). I argue that all of us in societies with widespread1 use of digital media are at any minute potentially subject to context collapse. Think about the last conversation you had with someone in a public space. How many of the people around you had access to a cell phone with a camera, Twitter, or Facebook on it? Someone could have surreptitiously livestreamed you on the internet, or an eavesdropper could have live tweeted your conversation. You may have thought you were talking to one person, but your words could have been heard or read by many more people. Boyd and others use the term ‘imagined audience’ to refer to the audience the speaker believes that she is addressing. Your imagined audience was the person you were addressing, but your audience may actually have been much larger. Consider another example. I gave a lecture to my college class last week. I imagined that the audience of my lecture was the group of students in the room, but it is entirely possible that my words were (or will be) heard by people in an entirely different audience via a student livestreaming the class. These examples make it clear that context collapse is something that can happen to us without our awareness. The possibility and actuality of context collapse is now a pervasive feature of our lives. I will show that it is an epistemically important feature of our lives: it shapes what is known about us and our world and it shapes our epistemic practices and habits. Before I address the epistemic problems that context collapse creates, I should briefly acknowledge that the widespread context collapse facilitated by the internet also has many epistemic benefits. Social scientists have analyzed how social media users gain access to useful information through the blurring of their social networks (Ellison et al. 2011). Furthermore, the ways in which the internet facilitates hacking and leaking can have many epistemic benefits as well (although, of course, these also pose their own epistemic challenges). So, I am not arguing that online context collapse is a net epistemic negative; there is not enough space to argue for that all-things-considered evaluation. Instead, I will simply attempt to make some progress on identifying and understanding the epistemic pitfalls of
1 While use of digital media is widespread, it is by no means universal, given the persistence of the digital divide (cf. Hargittai and Hsieh 2013).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemic Dangers of Context Collapse Online 439 context collapse and suggest some ways we can avoid them.2 I leave it to others to expand on this analysis to consider whether (or when and how) these epistemic costs are outweighed by the benefits of context collapse.
2. Veritistic Framework Before I argue that context collapse can have negative epistemic consequences, I offer a brief overview of the epistemic framework that I use to evaluate this feature of the internet. This is a veritistic, systems-oriented social-epistemic framework. Systems-oriented social epistemology evaluates the epistemic outcomes of epistemic systems. Goldman defines an ‘epistemic system’ as “a social system that houses social practices, procedures, institutions, and/or patterns of interpersonal influence that affect the epistemic outcomes” of their members (Goldman 2011, 18). Systems-oriented social epistemology is not merely a descriptive enterprise. It is a consequentialist epistemic framework that evaluates epistemic systems according to their positive or negative epistemic results for agents affected. But what is the epistemic good that epistemic systems ought to produce for the agents affected? For veritistic social epistemologists, the fundamental epistemic good is true belief. I follow Goldman in often referring to true belief as ‘knowledge’ in a weak sense. Thus, the veritistic systems-oriented social epistemology that I use in this chapter evaluates the social systems of the internet in terms of its effects on the formation and dissemination of knowledge (true belief) within the community. To the extent that the internet facilitates forms of context collapse that result in false beliefs propagating through the community, the internet (as currently structured) is epistemically deficient. Veritistic, systems-oriented social epistem ology is a useful epistemic framework that helps us see ways in which online context collapse produces these epistemic harms.
3.1 Epistemic Challenge 1: Context Collapse Facilitates Harassment The first epistemic challenge is that context collapse facilitates online harassment, and harassment does epistemic harm by decreasing the diversity of epistemic communities. I begin with the first premise in this argument: context collapse facilitates online harassment (McMillan Cottom 2016). The following example illustrates the connection between context collapse and harassment. It is a 2 Since this volume focuses on applied epistemology, I will not address the ethical pitfalls of context collapse. Context collapse can cause moral as well as epistemic harm; e.g. the moral harms of privacy violations. Thus, one might argue that we need to avoid context collapse to protect individual liberty and social justice. This chapter will address the less discussed epistemic consequences of context collapse.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
440 Karen Frost-Arnold hypothetical example, based loosely on several actual examples of harassment of professors in the United States.3 Professor X is giving a lecture in her feminist philosophy class one day when she makes a comment that one of her students (call him student Y) interprets as an attack on men. This student had surreptitiously recorded the lecture, contrary to university policies that prohibit recordings of class without permission. Student Y shares the video on Twitter, where it is picked up by right-wing media organizations and quickly goes viral. Professor X receives harassing emails, voicemails, and death threats. A campaign is organized to contact the administration of Professor X’s institution to push for her dismissal. This incident is a dramatic example of context collapse. Professor X’s imagined audience for her lecture was the group of students in her class who had grown to know her over the course of the semester, but in fact the audience ultimately included people all over the world who knew nothing about her, other than the short video clip that they watched. This is an example of the type of context collapse that Davis and Jurgenson call context collision, which occurs “when different social environments unintentionally and unexpectedly come crashing into each other” (Davis and Jurgenson 2014, 480). Professor X’s classroom environment did not just come crashing into the environments of regular followers of her student’s Twitter account; it was picked up by right-wing media outlets and entered the world of online trolls who look for opportunities to harass people in their online and offline lives. It is important to understand just how pervasive and powerful the phenom enon of online trolling and harassment has become. I argue that it is impossible to understand the epistemology of the internet without recognizing that large numbers of loosely organized and skilled trolls are a formidable part of the online ecosystem. Some people think of trolling as just the ill-mannered comments of people who happen to come across something online that they disagree with, but as sociologist Tressie McMillan Cottom describes, trolls are increasingly active and looking for prey (McMillan Cottom 2016). In the context of Twitter, she notes, “They search terms to find victims. They target users. And, they leverage network ties to force context collapse” (McMillan Cottom 2016). Trolls who find a potential victim often organize with other trolls on sites such as 4chan to share the material that has exposed the victim, to coordinate attacks, and to share information (such as the victim’s phone number, home address, and contact information for relatives of the victim) for doxing and harassment. McMillan Cottom argues that forcing context collapse helps trolls mobilize others to harass a victim and, in some cases, to garner the attention of mainstream media, which can dramatically increase the victim’s exposure (McMillan Cottom 2016).
3 See Markham (2012) for a methodological discussion of using composite accounts of actual cases as a privacy-protective methodology for internet studies scholarship.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemic Dangers of Context Collapse Online 441 Context collapse can be a tool of online harassment, but why is this epistemically damaging? Online harassment is epistemically harmful because it disproportionately targets members of marginalized groups (e.g. women and people of color) and thereby decreases the diversity of online epistemic communities. The Pew Research Center survey found that people of color are more likely to experience online harassment (“Some 51% of African-American internet users and 54% of Hispanic internet users said they had experienced at least one of the six [types of harassment measured], compared with 34% of white internet users”) (Pew Research Center 2014, 15). Consider the data on gender: 60–70 per cent of cyberstalking victims are women, and one report found that 72.5 per cent of cyber harassment victims between 2000 and 2011 were women (Jeong 2015, 39–40).4 This is obviously a social and ethical problem of injustice, but why is it an epi stemic problem? Here I draw on an argument that I have made elsewhere that epistemic diversity is veritistically valuable (Frost-Arnold 2014a). Members of oppressed groups contribute novel true beliefs to the epistemic community as a whole, and they also provide valuable transformative criticisms of dominant beliefs; thus, oppressed groups help weed out false beliefs (Frost-Arnold 2014a, 69–72; cf. Intemann 2010; Longino 1990, 2002). Given that members of oppressed groups contribute valuable true beliefs and help weed out errors, it is epistemic ally damaging if they withdraw from epistemic communities or smother their testimony (cf. Dotson 2011). But this is precisely what online harassment causes. The literature on online harassment is full of examples of victims of harassment who stop writing their blog, or delete their Twitter and Facebook accounts, or stop teaching their classes to avoid harassment (cf. Citron 2014). Additionally, the threat of online harassment has a deterrent silencing effect as others preemptively avoid engaging online to avoid the harassment that they see others facing. When marginalized people withdraw from epistemic communities, the community loses the epistemic value of both the true beliefs they share and the false beliefs
4 For a summary of multiple studies showing that women are disproportionately targeted for online harassment, see Citron (2014, 14–19). The research in this area is complex, given that there are many different types of behaviors with different degrees and types of harm that fall under the umbrella term of ‘online harassment’ (e.g. name calling, sexual harassment, stalking, threats of violence, non- consensual porn). Additionally, different online spaces have different norms of acceptability for behaviors such as name calling (e.g. certain types of name calling may be more common and less distressing in online gaming than they would be on social media). Given these complexities, it is hard to compare different studies. So, while Jeong (2015) and Citron (2014) cite many studies showing disproportionate harassment of women, there are some studies that do not follow this trend. For example, the Pew Research Center found that men are somewhat more likely than women to be victims of online harassment (“Some 44% of men and 37% of women have experienced at least one of the six types of harassment”, Pew Research Center 2014, 13). This study used a broad definition of ‘online harassment’ that included name calling as well as a broad array of online spaces that included online gaming, social media, email, and newspaper comment spaces, among others.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
442 Karen Frost-Arnold they help remove from circulation. Thus, to the extent that context collapse is used by harassers to target marginalized victims, it is epistemically damaging.5
3.2 Epistemic Challenge 2: Context Collapse Disrupts Epistemic Communities for Marginalized Groups The second epistemic problem presented by context collapse is that it threatens the integrity of marginalized epistemic communities in which some types of true belief flourish. As Alison Jaggar has argued, some types of knowledge production flourish when agents can produce knowledge within somewhat circumscribed communities (Jaggar 2004).6 This is particularly true for marginalized communities. For example, marginalized people may need to create their own language in which to describe their oppression (Fricker 2007; Jaggar 2004; Pohlhaus 2012; Spivak 1988). These groups are often vulnerable to epistemic injustice and epi stemic violence in communities in which they are minority members (Dotson 2011; Fricker 2007; Jaggar 2004). Thus, there are good reasons why members of marginalized groups create spaces and communities in which to have conversations amongst themselves.7 5 Note that my account of epistemic diversity focuses on the epistemic value that marginalized agents bring to communities. In this respect, the argument takes a position on the debate between feminist standpoint epistemologists and feminist empiricists regarding the epistemic value of domin ant or oppressive agents and perspectives. While feminist empiricists, such as Helen Longino (1990, 2002), argue that (tempered) equality of intellectual authority requires that all qualified agents be granted epistemic authority, feminist standpoint epistemologists argue that those who have achieved the standpoint of the oppressed (who often tend to belong to oppressed groups) be granted authority above those who lack such insights or who perpetuate bias and ignorance. I side with the standpoint epistemologists for the reasons given by Kristen Intemann: “It appears that feminist empiricists do not have the resources to exclude certain values and interests from epistemic communities because all are instrumentally valuable to contributing to the scrutiny of background assumptions” (Intemann 2010, 792). To give a concrete example, for feminist standpoint epistemologists, it is not clearly a loss of important epistemic diversity for harassing white supremacists to be banned from Twitter, while a feminist empiricist may be forced to grant the white supremacist equality of intellectual authority and view banning them as a loss of diversity. I maintain that in a racist society, white supremacist views already dominate the epistemic landscape, so for some white supremacists to be blocked would not lead to a damaging loss of diversity. Thus, my argument that loss of epistemic diversity caused by context collapse is epistemically harmful should be read in light of standpoint theory. 6 See also Neil Richards’ argument that protecting private social spaces “buttresses free thought and our other processes of belief formation” (Richards 2008, 413). 7 It is worth noting that Jaggar argues that while there is epistemic value for closed communities, she also argues that closed communities can also be subject to epistemic dangers such as groupthink and echo chambers. Ultimately, she argues that the best epistemic solution may be periods of closed conversation followed by careful engagement with outsiders who may recognize biases (Jaggar 2004, 241). There is not room in this chapter to engage with the details of when and how to temporarily open a closed community. My argument in this section is simply that context collapse can be epistemically damaging in foreclosing the possibility of sustaining even temporarily circumscribed communities.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemic Dangers of Context Collapse Online 443 Context collapse can disrupt such epistemic communities, thereby hindering the production of knowledge. Consider the practice of live tweeting talks at a conference. Many people think that this is an innocuous practice that simply helps the author more widely disseminate their arguments. But live tweeting talks at some conferences (or particular conference sessions) that function as epistemic communities for marginalized members of a profession is potentially damaging. There are two issues to disentangle here. First, live tweeting a talk can be problematic if done without consent of the speaker and audience, because it can be a violation of trust and privacy, which can be disruptive to the functioning of the community. But is live tweeting epistemically harmless if done with consent? I argue that the answer is no, because the second problem with live tweeting is that it collapses the context and removes the epistemic benefits of communities dom in ated by marginalized people. Consider the following thought experiment. Suppose all feminist philosophers agreed that all talks at feminist philosophy conferences would be simultaneously live tweeted to the public. There would be no violation of trust or felt privacy violation if all members agreed and felt comfortable with this new policy. But would something of epistemic value be lost? Yes, because feminist philosophy conferences often function as spaces where marginalized members of the profession can gather to produce knowledge together. Participants will often share personal stories or ideas at conferences that they would not in other contexts. For example, it is not uncommon for survivors of sexual assault to discuss their assault in feminist spaces, when they might not elsewhere. Some of this is due to the fact that feminist spaces tend to have a common understanding of some practices that avoid gaslighting, testimonial injust ice, and testimonial smothering, which are common epistemic harms perpetuated against survivors of sexual assault (cf. Dotson 2011; Fricker 2007; McKinnon 2017). These disclosures and the knowledge that is produced is veritistically valu able because participants can share true beliefs that they would not otherwise. Also, the disclosures made in circumscribed communities of trust can help members weed out their own biases and false beliefs, which is also veritistically valu able. If, however, participants knew their words would be broadcast to the world via Twitter, they may choose the lowest common denominator tactic for dealing with context collapse in which speakers disclose only what will be appropriate for all members of their audience, including every member of the public (cf. Davis and Jurgenson 2014, 482). Useful knowledge production and error detection may be lost. Thus, context collapse potentially disrupts the very thing that makes marginalized communities epistemically valuable—it can prevent marginalized members from producing true beliefs and detecting errors. This is just one example of the “chilling effect” on knowledge production and acquisition that loss of privacy through context collapse can create (cf. Fallis 2013, 154).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
444 Karen Frost-Arnold
3.3 Epistemic Challenge 3: Context Collapse Promotes Misunderstanding A final problem with context collapse rests on the context- sensitivity of understanding. As Andrew Peet puts it, “To assign content to many utterances audiences must rely on their contextual knowledge” (Peet 2015, 29). If the speaker and hearer do not share some background knowledge about the setting of the conversation, the speaker and hearer, past conversations, the goals of the conversation, Gricean norms and more, then the hearer may misinterpret the speaker’s utterance (Fricker 2012, 66; Peet 2015, 31). Following Elizabeth Fricker, I call the participants’ representation of these factors the knowledge context. In many cases of context collapse, the speaker and overhearing audience do not share a know ledge context. When speaker and audience do not share enough of a knowledge context, miscommunication is likely, i.e. the audience will not recover the propos ition that the speaker intended to convey to their imagined audience. This is epi stemically damaging because often the misunderstanding will cause the audience to form false beliefs. This is particularly true when utterances are spread by third parties in ways designed to produce misunderstandings. For example, in context collision third parties may share a snippet of a conversation or utterance and add some information when they repost or retweet it. This practice can provide misleading contextual information that causes new audience members to form false beliefs about what was said. To connect this third epistemic problem to the first problem, promoting misunderstandings is a common tool of online harassers who want to incite an online mob against the victim. By sharing the victim’s words with misleading contextual information, the harassers can cast the victim in an especially negative light that is likely to draw widespread condemnation. Consider a concrete example from academia. On December 24, 2016, George Ciccariello-Maher, Associate Professor of Politics and Global Studies at Drexel University, tweeted “All I Want for Christmas is White Genocide” (Blumenstyk 2017).8 Following the tweet, Ciccariello-Maher and his friends and family became victims of an organized campaign of online harassment, including death threats (Blumenstyk 2017). Ciccariello-Maher says that his tweet was supposed to be a satirical joke mocking the white supremacist use of the term ‘white genocide’ to attack interracial marriage and diversity efforts. In order to understand this intention of his tweet, the audience would need to have this background understanding of the white supremacist use of this term, as well as perhaps the current events context in which there was a recent white supremacist backlash to a State Farm Insurance ad showing a black man proposing to a white woman, in which many racists were using the hashtag #WhiteGenocide (Blumenstyk 2017). Clearly, one
8 I thank George Ciccariello-Maher for permission to discuss this case.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemic Dangers of Context Collapse Online 445 would need to have the background knowledge that white genocide is an imaginary threat created by white supremacists to stoke fear in whites to see that Ciccariello-Maher’s tweet is satirical. But the tweet was shared to many Twitter users (and television viewers) who did not have this background. The ensuing outrage suggests that many believed that Ciccariello-Maher sincerely stated that he wanted a genocide of white people.9 This is just one example of how failure to share a knowledge context can create misunderstandings. Context collapse encourages this kind of misunderstanding as information is shared beyond an intended audience which would have the necessary knowledge context. This is veritistically damaging when these misunderstandings prompt the audience to form false beliefs, such as the belief that college professors “call for” genocide (e.g. in a television interview with Ciccariello- Maher, Fox News anchor Tucker Carlson repeatedly claimed that Ciccariello-Maher had “called for” white genocide) (cf. Ciccariello-Maher 2017). Context collapse can, thus, be a tool for spreading misunderstanding through thinly veiled misinformation campaigns.
4. Epistemic Virtues to Prevent Problems of Online Context Collapse Many tactics are required to address the problems of context collapse. Some writers advocate improvements in platform design to give users more tools to control the spread of their content. Hull et al. (2011) discuss ways that Facebook could change its design to remind users of the possibility of context collapse. For example, users add many friends on Facebook, but tend to only post status updates with a smaller subset of their friends in mind (i.e. their imagined audience) (Hull et al. 2011, 298). Facebook could add alerts and reminders of who is reading one’s updates (e.g. “Mary has just looked at all of your updates”). This would help readers keep in mind the larger audience who is potentially reading their status updates, thereby potentially avoiding unintentionally harmful context collapses. McMillan Cottom (2016) critiques the re-design of Twitter’s @replies that was introduced to some users in fall 2016 and to all users in early 2017. As I mentioned in section 2, McMillan Cottom argues that context collapse is a common tool for online harassers. She discusses one strategy, “selective visibility,” that she uses to avoid this problem: by selectively replying to users, she avoids engaging with trolls who may try to interact with her simply as a tentative lure to gauge her vulnerability. Unfortunately, the re-design of the @replies made it harder for users to determine to whom they were replying, which undermines 9 Of course, much of the expressed outrage was also faux outrage spread by agents who were well aware that the tweet was satirical, but who were happy to spread the misinterpretation for political purposes.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
446 Karen Frost-Arnold this strategy for avoiding harassment. Both McMillan Cottom’s critique of Twitter and Hull et al.’s suggestions for Facebook share an important approach to dealing with the problems of context collapse—careful attention to the design of social media platforms. By giving users more control and raising their awareness of the possibility for context collapse, some of the epistemic harms I have outlined may be avoided. Applied epistemologists should pay careful attention to such design solutions to epistemic problems. However, we should be wary of adopting a dogmatic faith that a technological solution is available to every problem with the internet. In contrast, internet theorists are increasingly skeptical that better coding and design is all we need (boyd 2017). Even if social media platforms are both carefully designed to make the threat of context collapse apparent to users and designed to give them tools to carefully circumscribe their audience, users can still do harm by habitually ignoring these tools or thoughtlessly (or malevolently) sharing information with collapsed audiences. Part of the solution has to be that we change our habits and re-evaluate our practices of sharing content online. This is where I argue that a virtue epistemological approach is helpful. I do not argue that virtue epistemology is the only epistemological framework useful for analyzing these issues, nor will I argue that it is superior to alternatives. My goal in this chapter is simply to show that virtue epistemology illuminates important features of the epistemology of context collapse and that it suggests some epistemically beneficial habits that we can adopt to address our epistemic predicament. As I will argue, given the speed with which we confront decisions about whether to share content online, developing epistemically virtuous habitual responses is particularly valuable. Before I sketch the virtue epistemology and its benefits, some clarifications are required. First, I want to flag that I am not trying to use virtue epistemology for some of the purposes for which it has traditionally been found useful.10 For example, my goal is not to solve the Gettier problem or to explain why some reliable processes do not yield knowledge (cf. Lepock 2014; Zagzebski 1996). I do not use the concept of virtue here to give an account of what it means for S to know that p in terms of S’s epistemic virtue. Instead, what I am interested in is whether the concept of epistemic virtue can be helpful in systems-oriented social epistem ology. I think it can, in the following way. Systems-oriented social epistemology examines a system “to see whether its mode of operation is genuinely conducive to the specified epistemic ends” and it also identifies “alternative organizational structures that might be epistemically superior to the existing systems” (Goldman 2011, 19). Realistic work in this area (e.g. non-toy models) can only be 10 Thus, my project does not fall under what Jason Baehr calls “conservative virtue epistemology” which maintains that “the concept of intellectual virtue is useful for addressing one or more problems in traditional epistemology” (Baehr 2011, 12).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemic Dangers of Context Collapse Online 447 done with a conception of what the epistemic agents who interact in the epistemic system are like; for example, what faculties they tend to have, what character traits they exhibit, etc. In other words, we need to know what epistemic virtues the agents have in order to know what consequences will result from them interacting with each other within the epistemic system. Additionally, the organizational structures and virtues of the agents are not independent, but instead often interact in feedback loops such that certain kinds of social structures encourage agents to develop certain habits, and agents with certain habits are more likely to design systems with certain features. So, investigating whether alternative epistemic systems might have superior epistemic outcomes sometimes requires investigating how they change the habits of the agents involved. Conversely, systems-oriented social epistemologists should also study how changes in epistemic habits change the outcomes of epistemic systems. It is for this reason that I investigate the epi stemic virtues that might solve some of the problems of context collapse. I cannot address all of these interactions between virtues and the social system of the internet in this chapter. I will focus on the question of whether the internet as a social system might have different (and perhaps superior) epistemic results if the agents involved were to cultivate certain epistemic virtues. A second way in which this project departs from much of traditional virtue epistemology is that the virtues I focus on are other-regarding epistemic virtues that produce true belief (and avoid error) in others. Virtue epistemology has tended to focus on the self-regarding epistemic virtues that help S attain know ledge for S. For example, intellectual courage can motivate an agent to doggedly pursue true beliefs for herself, even in the face of potential risks. In contrast, Jason Kawall describes other-regarding epistemic virtues as follows, “Such virtues are epistemic as they produce knowledge, but other-regarding as the knowledge is created in others, rather than in the agent herself ” (Kawall 2002, 257). Kawall maintains that the epistemic virtues can be divided into those that are selfregarding, those that are other-regarding, and those that can be both self and other-regarding (e.g. the intellectually courageous scientist may doggedly pursue truth both so that she can know the truth but also so that she can contribute to human knowledge). In contrast, Jason Baehr (2011) argues that all the epistemic virtues can be other-regarding. While some tokens of epistemic virtues may only be self-regarding, “a person who possesses the relevant intellectual virtues completely or in their fullness will also be concerned, not just with her own share in the epistemic goods, but also with the share of others” (Baehr 2011, 219). I tend to agree with Baehr that the epistemic virtues are all potentially other-regarding,11 but that point is not critical for the ensuing discussion. Regardless, otherregarding epistemic virtues are central to thinking about the internet. The social 11 In siding with Baehr on this issue, I also reject Driver’s argument that the epistemic virtues are not other-regarding at all (Driver 2003).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
448 Karen Frost-Arnold epistemology of the internet has to include analysis of other-regarding epistemic virtues, because if we constrain our focus to self-regarding virtues that help agents attain knowledge themselves from what they read online, we miss much of the epistemic work that agents are doing on the internet. Sharing, contributing content, and moderating online spaces are all practices shot through with epistemic significance. By engaging in these practices in epistemically virtuous ways, agents can reliably help members of their community attain knowledge, and their epi stemic vices can be causes of false beliefs propagating throughout the community. Having clarified the kind of virtue epistemology I have in mind, I now argue that cultivating habits of trustworthiness and discretion with online content can prevent some of the epistemic harms of context collapse. The first epistemic virtue relevant to preventing some of the problems of context collapse is trustworthiness. When we are trustworthy, we are motivated to avoid betraying others’ trust in us—to avoid taking advantage of the vulnerability of those who trust us. In epistemic trustworthiness, the motivation is to avoid betraying others’ trust in us to take care of their epistemic goods. Annette Baier models trust as a three-place relation in which the trustor (A) trusts the trustee (B) with valued good (C) (Baier 1996, 101). In extending this kind of epistemic trust, we place our epistemic goods (e.g. our possession of the truth, or our epi stemic communities) in the hands of those we trust (Zagzebski 1996, 161). In doing so, we make ourselves vulnerable to being betrayed and our epistemic goods harmed by those we trusted to care for them in various ways. The epistem ically trustworthy person is motivated to avoid taking advantage of that vulner abil ity and to care for the epistemic goods entrusted to them. Epistemic trustworthiness is an other-regarding epistemic virtue in that it aims at the epi stemic flourishing of others, but it can also have benefits to the bearer. For example, living up to others’ trust in you to not disrupt the online spaces for dialogue that produce knowledge certainly aims at the epistemic flourishing of others who use those spaces, but it also promotes one’s own epistemic flourishing if one also uses those epistemic spaces to obtain knowledge. In sum, this kind of epistemic trustworthiness will motivate people to avoid the second problem of context collapse; the epistemically trustworthy person will avoid undermining epistemic communities for marginalized people by betraying the norms community members are expected to uphold in order to protect the epistemic status of the community. Additionally, it is worth noting that the moral12 virtue of trustworthiness can have epistemic benefits here. The moral virtue of trustworthiness can be helpful 12 The use of the term ‘moral virtue of trustworthiness’ is slightly philosophically awkward here, since I follow Baehr in viewing the epistemic virtues as a subset of the moral virtues (Baehr 2011, 206–22). I use the term in contrast to ‘epistemic trustworthiness’ to flag that the kind of trustworthiness at issue is non-epistemic, in that the other-regarding goods that the morally trustworthy person aims to promote are not the other’s epistemic goods (as in epistemic trustworthiness).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemic Dangers of Context Collapse Online 449 in preventing some instances of context collision. Many cases of context collision are clear violations of privacy norms that speakers were trusting others to uphold.13 Helen Nissenbaum’s account of privacy as contextual integrity is helpful here (2004; 2010). Nissenbaum argues that our expectations of privacy are highly contextual. Different social contexts are characterized by different norms of what information it is appropriate to reveal within that context and how information should be transferred from one party to another (Nissenbaum 2004, 138–9) Privacy is violated when the norms of a context are violated. One of the implications of Nissenbaum’s account is that information is always tagged to a context, and it is a violation of privacy to share information outside that context when doing so flouts the norms of information transfer constitutive of that context (Nissenbaum 2004, 142). People trust each other to follow the norms of information sharing relevant to the context of their interaction. Sharing information outside that context in ways that violate those norms is a betrayal of trust and is a privacy violation. Hence the connection to trustworthiness as a virtue is made clear. When others are trusting a trustworthy agent to not share their speech outside of certain contexts, that agent will be motivated to avoid betraying that trust. As a consequence, they will be ceteris paribus less likely to perpetuate context collision and its attendant epistemic harms, because they will not want to betray the trust placed in them. Of course, it will depend on the particular situation, as sometimes betraying expectations of privacy is epistemically virtuous, such as in cases of whistleblowing. I expand on this issue in section 5. The second relevant epistemic virtue is what I am calling discretion, in the sense of being discreet.14 Part of the OED definition for the adjective ‘discreet’ reads as follows: “Possessing or exhibiting sound judgement in speech or action, esp. in such a way as to avoid one’s own or another’s disgrace or embarrassment; . . . tactful” (OED Online 2017). So, discretion in this sense involves good judgment in sharing speech, especially in a way that is attentive to the costs to others. On my account, to exhibit the other-regarding virtue of discretion in this sense on the internet is to habitually weigh carefully the risks and benefits of sharing others’ information online, especially with regard to avoiding harassment and misunderstanding of others’ speech. Someone with this virtue is not focusing on whether the speaker trusts them to not share the information; that is the focus of 13 It is beyond the scope of this chapter to fully discuss the ways in which context collapses constitute violations of privacy, since that determination depends on which account of privacy one adopts. In this section, I draw on Nissenbaum’s influential account of privacy as contextual integrity to illustrate one way to analyze the privacy implications of context collapse. Readers interested in other frameworks for assessing privacy and context collapse may find the following references useful: (Moor 1997; Richards 2008; Fallis 2013). I thank an anonymous reviewer for suggesting these resources. 14 In another sense of the term, ‘discretion’ refers to a kind of good judgment, akin to practical wisdom. This is not the sense I am interested in here. I discuss the need for practical wisdom with regard to decisions about context collapse in section 5.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
450 Karen Frost-Arnold someone with the virtue of trustworthiness. Instead, this epistemic virtue addresses our need to develop habits of carefulness in sharing information, even if we would not betray anyone’s trust by sharing it. This virtue involves respecting the boundaries of epistemic communities and contexts and making tactful decisions about when to share information from an epistemic community to outsiders. How might these virtues be part of addressing the three challenges of context collapse? First, consider online harassment and reducing the diversity of epi stemic communities. The epistemically virtuous agent will think twice before sharing someone’s speech that is likely to expose its author to online harassment, especially if the author is a member of a marginalized group who is both more likely to be targeted and at risk of withdrawing in the face of harassment. The moral virtue of trustworthiness is helpful here because a trustworthy person will be motivated to avoid sharing others’ speech in violation of their trust. For example, this could prevent Professor X’s student from sharing the surreptitious video of her classroom. Her university has policies against recording classes, so Professor X is entitled to trust her students not to do so. A trustworthy student would be motivated to avoid betraying that trust. Additionally, the virtue of discretion is helpful in avoiding the problems of online harassment because we need to be careful about sharing information even if the speaker makes their speech completely public. For example, imagine that I make all my tweets public, and I tend to tweet on feminist topics. The discreet reader of my tweets will still not share my tweets willy nilly. For example, they might have the good judgment not to share some of my tweets on 4chan or the subreddits where men’s rights activists organize to harass feminist scholars. The epistemically discreet person is motivated to avoid actions that will expose marginalized agents to online harassment that will tend to silence them. Second, how will these virtues help prevent agents from undermining epi stemic communities for marginalized people? To the extent that members of such communities trust participants not to share information beyond the boundaries of the community, the trustworthy agent will avoid doing so. Additionally, even if the community lacks such expectations, the discreet agent will still pay attention to whether damage would be done by sharing information. For example, even if a scholarly community has norms that permit live tweeting, the discreet person will not share those tweets or conference conversations with any and all audiences. The discreet person will consider whether sharing some kinds of information might expose the community to attack by outsiders. Finally, these virtues can be part of the solution to the problem of contextdriven misunderstandings. As I have mentioned, much of the information that is shared out of context, thereby leading to misunderstandings, is done in violation of the privacy expectations of the speaker. Thus, trustworthiness as a virtue
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemic Dangers of Context Collapse Online 451 should avoid some of those misunderstandings as fewer privacy violations should occur. Additionally, the epistemically discreet agent will consider whether a tweet or snippet of a conversation will be misunderstood out of context. If sharing a piece of speech, such as a public tweet, out of context is likely to cause misunderstandings, the discreet person will think twice about sharing it.
5. Objections and Replies I began by framing this chapter in terms of the social epistemologists’ challenge of finding the new questions posed by the internet. At this point, one might object that my analysis fails this challenge. According to this objection, I have shown that the epistemic challenges posed by the internet are simply the ones we face in offline life, and further that the solutions are similarly familiar. We already knew that indiscreet gossip or privacy violations were problems, and we already knew that trustworthiness and discretion were valuable virtues. So, what is new here? In response, it is certainly correct that these are to some extent perennial problems and that these are traditional virtues. But this does not mean that there are no new issues here, and it does not mean that the virtues are unaffected. Context sensitivity is important here; what may be virtuous in one environment may not be in another. The internet changes our epistemic landscape, making the virtues of trustworthiness and discretion increasingly valuable and relevant in new contexts. To support this, I want to expand on my earlier point that what is new about the internet is the pervasiveness of context collapse in an online world. While the mere expansion of context collapse into more areas of our lives and the increased chance of non-celebrities experiencing massive context collapse make context collapse a more pressing problem, there are three additional features of the internet that make the virtues increasingly important. First, the size of audience provided by the internet potentially gives each of us a megaphone that can reach millions. In a pre-internet era, I might have been able to surreptitiously taperecord my teacher, but unless I had access to newspaper, radio, or TV outlets, it was very difficult for me to share her speech with a large audience. Second, the ease with which I can share others’ speech has been dramatically changed by the internet. It takes practically no effort to click the ‘share’ button on Facebook or the ‘retweet’ button on Twitter, it is very easy to upload videos or livestream via Facebook Live, and live tweeting is similarly simple. Third, the speed with which context collapse can occur is increased in an online world. It can take just a few hours for context collapse to occur on a massive scale. These features of the internet mean that it is much easier for us to do significant harm through context collapse, and that harms can be perpetrated much faster than before. So, while the virtues of trustworthiness and discretion may have been valuable before the
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
452 Karen Frost-Arnold internet, they are increasingly important at a time when, if we do not have habits in place to avoid doing harm, we may be tempted to take actions that collapse contexts for others and cause significant harm before we even realize that we have done so. This is one of the reasons why I find virtue epistemology helpful for internet epistemology—the speed with which harm can be done makes habitual, instinctive, virtues especially valuable in an online world. If the previous response is persuasive, one might agree that my analysis raises newly pressing issues, but nonetheless disagree with the details of the argument. One objection maintains that the virtues I have identified (trustworthiness and discretion) are potentially epistemically damaging because they push agents away from sharing information and sharing information on the internet is epistemic ally beneficial. As I have outlined, trustworthiness and discretion mitigate the problems of context collapse by motivating agents to avoid sharing other people’s speech. The objector worries that this could undermine much valuable sharing on the internet, particularly sharing of speech that may be viewed by the speaker as damaging to them and which they trust others not to share. Examples include sharing videos of police brutality, whistleblowing, and leaking of government, institutional, or corporate information, which the objector thinks may carry much epistemic value to the public. Thus, on this objection, my argument is misguided in that it lauds habits of respecting the trust of others and trying to avoid harm to them, when, in fact, these habits help perpetuate a problematic culture of secrecy. This objection is uncharitable because it ignores two features of my argument so far. First, it ignores my explicit acknowledgement that the argument is not an all-things-considered evaluation of context collapse. I grant that context collapse may have many epistemic benefits, and to the extent that trustworthiness and discretion prevent context collapse, they may make access to these benefits more difficult. So, I am not giving an all-things-considered argument in favor of trustworthiness and discretion, nor am I arguing that they are the only virtues relevant to achieving true belief and avoiding error. Sometimes other virtues, such as courage, may be needed; for example, in pursuing a whistleblowing campaign to expose previously hidden information publicly. Second, the objection uncharit ably reads my argument as calling for dogmatic reflexes of trustworthiness and discretion, which are insensitive to the specifics of the situation. I have not argued that all agents should do what others trust them to do regardless of the situation, nor have I argued that agents should always discreetly avoiding sharing speech that could be damaging to the speaker. Earlier in the chapter, I claimed that sometimes expectations of privacy that others trust us to uphold may need to be betrayed, such as when whistleblowing is necessary. Let me expand upon this point. Elsewhere, I have argued that trustworthiness is a virtue that follows the doctrine of the mean (Frost-Arnold 2014b). This means that one can be
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemic Dangers of Context Collapse Online 453 both excessively trustworthy and insufficiently trustworthy. We are excessively trustworthy when we do what someone trusts one do to in circumstances that call for violating those expectations. Sometimes virtue demands that we betray o thers’ trust in us. This is one of the reasons why practical wisdom (phronesis) is needed in our epistemic lives. We need practical wisdom to help us develop habits of sharing information “at the right time . . . [to] the right people, for the right reason, and in the right manner” to quote Aristotle roughly (Aristotle 1962, 1106b21–23). The goal of this chapter has been to make the epistemic harms of context collapse salient and to suggest some virtues we might cultivate to help avoid those harms. To apply these tools, the person of practical wisdom may have to balance competing virtues in deciding whether and/or how to share information online.
6. Conclusion I have argued that context collapse presents some epistemic challenges to which we are all increasingly exposed. Context collapse can be veritistically harmful, as a tool of online harassment that damages the diversity of epistemic communities, as a way of disrupting epistemically valuable communities for marginalized people, and as a means of spreading misunderstandings. I have also argued that one part of dealing with these challenges involves individual agents cultivating epistemically beneficial virtues of trustworthiness and discretion. If we work on developing habits of being more careful in sharing information online, we can be part of addressing the epistemic problems created by context collapse.
References Aristotle (1962). Nicomachean Ethics, trans. Martin Ostwald. Englewood Cliffs, NJ: Prentice Hall. Baehr, Jason (2011). The Inquiring Mind: On Intellectual Virtues & Virtue Epistemology. New York: Oxford University Press. Baier, Annette (1996). Moral Prejudices: Essays on Ethics. Cambridge, MA: Harvard University Press. Blumenstyk, Goldie (2017). “The Far Right’s ‘New Offensive against Academia.’ ” The Chronicle of Higher Education (April 1, 2017), https://www.chronicle.com/article/ the-far-rights-new-offensive-against-academia/. boyd, danah (2017). “Google and Facebook Can’t Just Make Fake News Disappear.” Backchannel (May 29, 2017), https://backchannel.com/google-and-facebookcant-just-make-fake-news-disappear-48f4b4e5fbe8.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
454 Karen Frost-Arnold boyd, danah (2011). “Social Network Sites as Networked Publics: Affordances, Dynamics, and Implications,” in Zizi Papacharissi (ed.), A Networked Self: Identity, Community, and Culture on Social Network Sites. New York: Routledge, 39–58. boyd, danah (2008). “Taken Out of Context: American Teen Sociality in Networked Publics.” Ph.D. thesis, University of California, Berkeley, CA. Ciccariello-Maher, George (2017). Tucker Carlson Tonight, interview by Tucker Carlson (March 30, 2017). Washington, DC: Fox News. Citron, Danielle (2014). Hate Crimes in Cyberspace. Cambridge, MA: Harvard University Press. Coady, David (2012). What to Believe Now: Applying Epistemology to Contemporary Issues. Malden, MA: Wiley-Blackwell. Davis, Jenny and Nathan Jurgenson (2014). “Context Collapse: Theorizing Context Collusions and Collisions,” Information, Communication & Society 17(4): 476–85. de Laat, Paul B. (2010). “How Can Contributors to Open-Source Communities Be Trusted?” Ethics and Information Technology 12(4): 327–41. de Laat, Paul B. (2015). “The Use of Software Tools and Autonomous Bots against Vandalism: Eroding Wikipedia’s Moral Order?” Ethics and Information Technology 17(3): 175–88. Dotson, Kristie (2011). “Tracking Epistemic Violence, Tracking Practices of Silencing.” Hypatia 26(2): 236–57. Driver, Julia (2003). “The Conflation of Moral and Epistemic Virtue.” Metaphilosophy 34(3): 367–83. Ellison, Nicole, Charles Steinfield, and Cliff Lampe (2011). “Connection Strategies: Social Capital Implications of Facebook-Enabled Communication Practices.” New Media Society 13(6): 873–92. Fallis, Don (2013). “Privacy and Lack of Knowledge.” Episteme 10(2): 153–66. Fallis, Don (2011). “Wikipistemology,” in Alvin I. Goldman and Dennis Whitcomb (eds), Social Epistemology: Essential Readings. New York: Oxford University Press, 297–313. Fricker, Elizabeth (2012). “Stating and Insinuating.” Proceedings of the Aristotelian Society Supplementary Volume 86(1): 61–94. Fricker, Miranda (2007). Epistemic Injustice: Power and Ethics in Knowing. New York: Oxford University Press. Frost-Arnold, Karen (2014a). “Trustworthiness and Truth: The Epistemic Pitfalls of Internet Accountability.” Episteme 11(1): 63–81. Frost-Arnold, Karen (2014b). “Imposters, Tricksters, and Trustworthiness as an Epistemic Virtue.” Hypatia 29(4): 790–807. Frost-Arnold, Karen (2018). “Wikipedia,” in David Coady and James Chase (eds), The Routledge Handbook of Applied Epistemology. New York: Routledge, 28–40. Goldman, Alvin (2011). ‘A Guide to Social Epistemology,’ in Alvin I. Goldman and Dennis Whitcomb (eds), Social Epistemology: Essential Readings. New York: Oxford University Press, 11–37.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemic Dangers of Context Collapse Online 455 Goldman, Alvin (2008). ‘The Social Epistemology of Blogging,’ in Jeroen van den Hoven and John Weckert (eds), Information Technology and Moral Philosophy. New York: Cambridge University Press, 111–22. Hargittai, Eszter and Hsieh, Yuli P. (2013). “Digital Inequality,” in William H. Dutton (ed.), The Oxford Handbook of Internet Studies. New York: Oxford University Press, 129–50. Hull, Gordon, Heather R. Lipford, and Celine Latulipe (2011). “Contextual Gaps: Privacy Issues on Facebook.” Ethics and Information Technology 13(4): 289–302. Intemann, Kristen (2010). “25 Years of Feminist Empiricism and Standpoint Theory: Where Are We Now?” Hypatia 25(4): 778–96. Jaggar, Alison (2004). “Globalizing Feminist Ethics,” in Cheshire Calhoun (ed), Setting the Moral Compass: Essays by Women Philosophers. New York: Oxford University Press, 233–55. Jeong, Sarah (2015). The Internet of Garbage. Jersey City, NJ: Forbes. Kawall, Jason (2002). “Other-Regarding Epistemic Virtues.” Ratio, 15(3): 257–75. Lepock, Christopher (2014). “Metacognition and Intellectual Virtue,” in Abrol Fairweather (ed.), Virtue Epistemology Naturalized: Bridges between Virtue Epistemology and Philosophy of Science. New York: Springer, 33–48. Longino, Helen (1990). Science as Social Knowledge. Princeton, NJ: Princeton University Press. Longino, Helen (2002). The Fate of Knowledge. Princeton, NJ: Princeton University Press. Magnus, P. D. (2009). “On Trusting Wikipedia.” Episteme 6(1): 74–90. Markham, Annette (2012). “Fabrication as Ethical Practice: Qualitative Inquiry in Ambiguous Internet Contexts.” Information, Communication & Society 15(3): 334–53. Marwick, Alice E. and boyd, danah (2011). “I Tweet Honestly, I Tweet Passionately: Twitter Users, Context Collapse, and the Imagined Audience.” New Media & Society 13(1): 114–33. McKinnon, Rachel (2017). “Allies Behaving Badly: Gaslighting as Epistemic Injustice,” in Ian J. Kidd, José Medina, and Gaile Pohlhaus, Jr (eds), The Routledge Handbook of Epistemic Injustice. New York: Routledge, 167–74. McMillan Cottom, Tressie (2016). “Twitter’s New @Replies Re-Design Isn’t Just Stupid; It’s Really Stupid.” Medium (October 30, 2016), https://medium.com/@ tressiemcphd/twitters-new-replies-re-design-isn-t-just-stupid-it-s-really-stupidc471ca254f0a. Meyrowitz, Joshua (1985). No Sense of Place: The Impact of Electronic Media on Social Behavior. New York: Oxford University Press. Miller, Boaz and Isaac Record (2013). “Justified Belief in a Digital Age: On the Epistemic Implications of Secret Internet Technologies.” Episteme 10(2): 117–34. Moor, James H. (1997). “Towards a Theory of Privacy in the Information Age.” ACM Sigcas Computers and Society 27(3): 27–32.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
456 Karen Frost-Arnold Nissenbaum, Helen (2004). “Privacy as Contextual Integrity.” Washington Law Review 79: 119–57. Nissenbaum, Helen (2010). Privacy in Context: Technology, Policy, and the Integrity of Social Life. Stanford, CA: Stanford Law Books. OED Online (2017). “discreet, adj., n., and adv.,” OED Online. Oxford: Oxford University Press (April 14, 2017), www.oed.com/view/Entry/54025. Peet, Andrew (2015). “Testimony, Pragmatics, and Plausible Deniability.” Episteme 12(1): 29–51. Pew Research Center (2014). “Online Harassment.” (October 22, 2014), www. pewinternet.org/2014/10/22/online-harassment. Pohlhaus Jr, Gaile (2012). “Relational Knowing and Epistemic Injustice: Toward a Theory of Willful Hermeneutical Ignorance.” Hypatia 27(4): 715–35. Richards, Neil M. (2008). “Intellectual Privacy.” Texas Law Review 87(2): 387–445. Simon, Judith (2010). “The Entanglement of Trust and Knowledge on the Web.” Ethics and Information Technology 12(4): 343–55. Spivak, Gayatri C. (1988). “Can the Subaltern Speak?,” in Cary Nelson (ed.), Marxism and the Interpretation of Culture. Urbana, IL: University of Illinois Press, 271–316. Zagzebski, Linda T. (1996). Virtues of the Mind: An Inquiry into the Nature of Virtue and the Ethical Foundations of Knowledge. New York: Cambridge University Press.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 19/03/21, SPi
19
‘Yikkity Yak, Who Said That?’ The Epistemology of Anonymous Assertions Veronica Ivy
Most of us know that we shouldn’t believe everything that we read online, particularly those things said anonymously. But should we believe any of it? This is the question at the heart of a debate in contemporary epistemology on the epis temology of anonymous assertions. Some such as Sanford Goldberg (2013, 2015) argues, that it is epistemically inappropriate to believe anonymous assertions. In essence, as audience members, we are in an epistemically impoverished position from which to judge the veracity of the statements and the reliability of the speakers. Implicitly or explicitly, those such as Goldberg adopt what we may call a punitive model of speaker responsibility. The idea is that speakers (tend to) attempt to assert well only because hearers have the ability to hold the speaker to account for having asserted what they did and in some sense to bring punish ments to bear on the speaker for speaking improperly. And since these account ability mechanisms are absent in instances of anonymous assertion, speakers lack the sorts of incentives to be honest and reliable (or lack the disincentives against being dishonest and unreliable). Consequently, there are (nearly) omnipresent defeaters when we receive anonymous assertions. Therefore, except in relatively rare circumstances, we are not justified in believing anonymous assertions and, there fore, we do not acquire knowledge through anonymous assertions. Anonymous assertions, then, are an epistemically defective source of knowledge. In this chapter, I argue that the punitive model is mistaken. Moreover, once we abandon it in favor of a pro-social trust-based model, I argue that we are suffi ciently epistemically well positioned to form justified beliefs (and thus know ledge) on the basis of anonymous assertions. I use a recent social media platform, built around anonymous assertions, known as Yik Yak, to make my case. I finish by connecting this discussion to issues of silencing and testimonial injustice. The worry is that if we treat anonymous assertions as unworthy of justified belief, then we will contribute to epistemic injustice by causing credibility deficits, particu larly for a number of disadvantaged voices, and thus treat such speakers with insufficient epistemic authority. We thereby unjustly exclude such people as knowers. Veronica Ivy, ‘Yikkity Yak, Who Said That?’ The Epistemology of Anonymous Assertions In: Applied Epistemology. Edited by: Jennifer Lackey, Oxford University Press (2021). © Veronica Ivy. DOI: 10.1093/oso/9780198833659.003.0019
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
458 Veronica Ivy
1. The Credibility Assessment Task Our first task is to get some handle on what we mean by “anonymous” when we talk about the epistemology of anonymous assertions. Intuitively, an assertion is anonymous when we don’t know the identity of the speaker. Goldberg (2015, 2016) gives the following definition: ANON Given a speaker/writer S who produces a message m, S is anonymous (with respect to m) relative to an audience A when (i) no one among the mem bers of A is currently in a position to acquire knowledge to the effect that S pro duced m, (ii) it is reasonable (for S) to expect that no one among the members of A will come to know something to this effect, and (iii) the explanation for both of the foregoing is that S’s identity, qua producer of m, is in some way “hidden” from the members of A.
Similarly, Karen Frost-Arnold (2014b, 64) gives the following definition: I take anonymity to be a relation of non-coordinatability of traits, such that aspects of an anonymous person’s identity are not coordinatable with traits known to others . . . Anonymity thus encompasses pseudonymity, and a user can be anonymous even though she may have a comprehensive and accurate online pseudonymous identity (when the pseudonym prevents others from coordinat ing the online persona’s traits known to them with aspects of the real-world identity, such as the real-world name).
I take these to be commensurable explications of a fairly common-sense under standing of anonymity. Turning to the epistemology of anonymous assertions, then, Goldberg (2013, 2015) offers what I take to be the most cogent argument against treating anonymous assertions as a generally suitable source of justified beliefs (and thus knowledge). There are three primary problems with the epistemic status of anonymous assertions. First, hearers are unable to access the normal sorts of information on which we base the requisite credibility assessment of speakers. Second, because of this lack of information with which to judge the credibility of the speaker, hearers lack the sort of counterfactual sensitivity to defeaters that responsible hearers require. Third, since it is mutually manifest to hearer and speaker that the hearer can’t hold the speaker responsible for epistemically faulty assertions, speakers lose their motivation to avoid making false, or otherwise epistemically defective, assertions.1 Consequently, speakers are more likely, ex 1 In Goldberg (2015), he connects this to anonymous assertions violating his version of the norm of assertion (the mutually manifest epistemic norm of assertion: MMENA), whereas this isn’t a factor
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Anonymous Assertions 459 hypothesi, to make false or epistemically defective assertions. In this section, I focus on Goldberg’s first two arguments. In section 2, I take up the third, which I connect to what we can call the “punitive model” of assertoric speech. Anonymous assertions quite easily fit within the broader topic of the epistem ology of testimony. People have a variety of different views on what makes a testimony-based belief justified. Roughly speaking, one can be some variety of reductionist.2 On such views, testimony itself isn’t sufficient for forming a justi fied belief; instead, one must have some good reason to think that the speaker (or testimony in general) is reliable. Alternatively, one could be some variety of anti-reductionist.3 On such views, testimony itself is sufficient for forming a prima facie justified belief, provided that there aren’t any defeaters present. Goldberg argues that his view on the epistemology of anonymous assertions is neutral with respect to the reductionism/anti-reductionism debate. In short, he argues that his view works on either a reductionist or an anti-reductionist frame work. Let’s start with assuming a reductionist account. He argues that hearers lack the regular sources of information in order to have adequate positive reasons to think that the speaker is reliable in their assertion. Hearers tend to rely on four sources of information in judging the content of an assertion as a candidate for justified belief (or knowledge): information regarding the speaker, information regarding the act of assertion (or communication) itself, information regarding the content of the assertion, and information regarding the context of the asser tion. I will take these briefly in turn. First, information regarding the speaker has to do with a hearer’s knowledge of a speaker’s truth-telling track record. How well do I know the speaker? Are they a generally reliable speaker? Can I trust them? Are they knowledgeable? Are they authoritative? If I know, for example, that my friend is an expert on medical ethics, then when she asserts something in her field of expertise, my knowledge about her expertise gives me good reasons to believe what she says. Alternatively, given the content of her assertion, is the speaker likely to have motives to lie, or to tell the truth? Suppose that I ask a friend how much she’s had to drink tonight, and she tells me, ‘Two drinks.’ If I know from past experience with her that in this context she is likely to lie, then my knowledge of her informs my decision not to believe her. Second, information regarding the act of assertion has to do with a hearer’s knowledge of how a speaker asserted what they did. Was it said with confidence, or were they nervous and displaying signs of lying? Were they in a rush, and did they blurt out the first thing that came to mind, or did they seem to consider their in Goldberg (2013). I think that this is a relatively minor factor in his arguments regarding the epis temology of anonymous assertions, and so I won’t focus on it here. 2 For instance, see Fricker (1994, 1995, 2002, 2004, 2006a, 2006b); Coady (1973, 1992). 3 For instance, see Lackey (2003, 2006a, 2006b); Goldberg (2006, 2008, 2013); Ivy [McKinnon] (2015a, 2017).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
460 Veronica Ivy answer before speaking? Assertoric contexts tend to give us a great deal of information and reliability signals.4 These first two sources of information thus focus on our assessment of the speaker herself, rather than on the content of her assertion per se. Third, information regarding the content of the assertion has to do with a hear er’s background knowledge about the content of the assertion. For example, I’m an expert in badminton. If someone asserts, for example, that a new racquet can be strung at 50lbs of tension (assuming that I know nothing else about the person speaking) and given my background knowledge about the topic of the assertion (string tension capabilities of racquets), then I can use my knowledge to judge the veracity of the statement (almost certainly false in this case!) and the reliability of the speaker. I may judge, for example, that the speaker doesn’t know anything about badminton racquets. Furthermore, is the content of the assertion the sort of content about which people are inclined to lie or mislead? For example, we might think that an asser tion about what one had for breakfast in a low-stakes conversational context isn’t the sort of content one would lie about; however, we might think that gossip about someone’s sexual history is the sort of content about which one might lie or mislead. We can thus use what we know about the relationship between the con tent of an assertion and a general theory of speaker honesty or dishonesty to help judge the reliability of a speaker’s assertion. Fourth, and finally, information regarding the context of the assertion has to do with things like the medium of the content. For example, was the assertion com municated in person, on a reputable television show, or in a reputable publica tion? Or was it an anonymous website known for troll comments, where people are known to lie for the fun of it? This can also include the situation under which the assertion was made. For example, was the person under duress (say torture), or do they have disincentives for lying?5 These last two sources of information thus focus on our assessment of the content of the assertion, and where it was asserted, rather than on the speaker herself. So how does Goldberg’s argument work assuming a reductionist account of testimony? In order for a hearer to justifiably believe p on the basis of S’s assertion “p,” a hearer must have positive reasons to judge S reliable about, and in their reporting of, p. However, Goldberg argues that the sorts of information on which we tend to rely for such judgments—the four sources just canvassed—are absent, or severely lacking, in instances of anonymous assertion. 4 Kenyon (2013) nicely makes this point. 5 It’s not entirely clear how to categorize the ‘context’ of an assertion. Goldberg seems to include a speaker having a motivation to lie about a particular topic as part of the first category: information about the speaker. But he also seems to include some information about the context of the assertion as part of the fourth category: information about the context of assertion. For a lengthy discussion of assertoric ‘context,’ see McKinnon (2013, 2015b).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Anonymous Assertions 461 We certainly lack information on the identity of the speaker, by definition of it being anonymous. Anonymity often requires methods of assertion that block our ability to assess the act of assertion itself: it may be online, and thus typed, or might be voice processed through digital effects, such as in video interviews. Certainly, not all anonymous assertions are such that the receiver lacks all know ledge of the content of the assertion, but what matters is that there will be at least some, if not many. And similarly, there will be some anonymous assertions expressed in contexts where we lack information about the context itself, espe cially if it’s a new medium. Thus, there are at least some anonymous assertions that lack all four sources of reliability indicators, which are Goldberg’s focus. Even if there are some reliability indicators present, what information is left is generally insufficient: we’re left with crumbs, as it were. Therefore, in general, one cannot form justified beliefs (and thus knowledge) via anonymous assertions. Now suppose that one prefers an anti-reductionist account of testimony: one has a default, prima facie, justification to believe what one is told, provided that there aren’t epistemic defeaters present. Take a typical non-anonymous assertion. I’m visiting your town and I ask you, a stranger, which way to a well-known local landmark. You, knowing the way, tell me honestly how to get there, and I straight forwardly believe you.6 You gave me no signs to doubt you. But let’s counterfactu ally suppose that you don’t know. Plausibly, had you tried to lie to me, I would have picked up on some signals of deceit. For example, you probably would have given off signals that you weren’t confident in your answer, or that you otherwise didn’t know. So, when you told me the truth, I was counterfactually sensitive to the presence of defeaters. But what about anonymous assertions? Goldberg notes a second mark against anonymous assertions: hearers lack the sort of counterfactual sensitivity to defeat ers that they tend to have in non-anonymous assertoric contexts. He argues that: the following seems reasonable: a hearer is not entitled to regard the speaker in this way unless the hearer herself is counterfactually sensitive to the presence of defeaters, such that, had there been any relevant defeaters, she should have attended to them, and so would have refrained from so regarding the speaker.7
But the sorts of information we depend on in order to be counterfactually sensi tive to defeaters are mostly absent in instances of anonymous assertion—again, principally, the four sources of information canvassed above. Therefore, on an anti-reductionist account of testimony, hearers are generally not justified in believing anonymous assertions. 6 Of course, as Kenyon (2013) rightly argues, I don’t “straightforwardly” believe you. But let’s set that aside. See also McKinnon (2015b, 2016a). 7 Goldberg (2015, 209).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
462 Veronica Ivy In sum, Goldberg’s view doesn’t depend on taking a stance on the reductionism/ anti-reductionism debate. His argument that the hearer is unable to perform the credibility assessment task can be incorporated into either a reductionist or an anti-reductionist framework. If one is a reductionist, then Goldberg’s argu ment is that one is unable to gain enough information on which to judge the reliability of the speaker in order to satisfy the positive reasons requirement. If one is an anti-reductionist, Goldberg’s argument is that the paucity of informa tion on which to perform (much of) any credibility assessment means that one cannot adequately, counterfactually, monitor for defeaters. Moreover, as I discuss in section 3, he argues that such contexts provide (near) omnipresent defeaters. So, on either framework of the epistemology of testimony, hearers do not form justified beliefs, and thus knowledge, based on anonymous assertions.
2. The Punitive Model of Assertoric Speech Goldberg’s first two arguments against the general acceptability of believing anonymous assertions focus on problems for the hearer in performing the requis ite credibility assessment task. The first argument targets those inclined towards reductionist accounts of the epistemology of testimony, while the second targets those inclined towards anti-reductionist accounts. His third argument focuses more directly on the speaker. In short, he argues that in contexts where a hearer can’t hold a speaker responsible—such as anonymous assertions—for epistemic ally defective speech (whether false, misleading, or deliberately lying), speakers are less likely to conform to social norms of truth telling. Speakers in such con texts are thus less reliable sources of information than in face-to-face assertoric contexts. Therefore, there are nearly omnipresent defeaters for hearers receiving anonymous assertions, and hearers thereby cannot form justified beliefs based on anonymous assertions. Critical to Goldberg’s view is the idea that speakers (tend to) attempt to assert well only because hearers have the ability to hold the speaker to account for hav ing asserted what they did and in some sense to bring punishments to bear on the speaker for speaking improperly. Call this the punitive model of assertoric behavior. Goldman (2008) argues, for example, that blogging has less epistemic value than traditional (print) journalism because speakers in the latter context may face economic sanctions (such as losing their job as a journalist) that the former do not. The threat of punishment or reprisal for poor reporting or false claims thus incentivizes the print journalists to assert well, or else.8 Another way of putting 8 Of course, perhaps we should recognize that this doesn’t seem to be true. Some news organiza tions thrive even though they regularly report falsely. I’ll leave it as an exercise for the reader to think of examples.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Anonymous Assertions 463 the same point is that the print journalists are disincentivized to lie or otherwise report poorly. Here is how Frost-Arnold (2014a, 65) characterizes the punitive model of assertoric behavior: Accountability mechanisms are attempts to increase the trustworthiness of agents grounded in a Hobbesian or economic view of human nature and the solution to its ills. On this Homo economicus view, humans are self-interested, rational, utility-maximizing agents. Under certain conditions . . . it is in the self-interest of unfettered rational egoists to engage in antisocial and untrustworthy behavior. But punitive systems create disincentives for such antisocial behavior . . . Mechanisms to detect and sanction untrustworthy behavior hold members accountable and provide a disincentive for betrayal. Accountability mechanisms in epistemic com munities increase detection and punishment mechanisms for betraying epistemic norms (e.g. norms of honesty).
I like to refer to this Homo economicus view of assertoric practice, where we only assert truthfully for rational self-interested reasons, as assertoric egoism. In simi lar fashion, Goldberg (2015, 210) writes: . . . the audience might assess such things as whether the speaker had any present motive for insincerity, as well as what ‘pressures’ there are on the speaker to speak responsibly . . . These ‘pressures’ derive from what I will describe as the mechanisms for the ‘policing’ of assertoric responsibility. By this I include all the practices and institutions that encourage speakers to live up to their assertoric responsibilities and/or enforce the repercussions for those who do not do so. These ‘pressures’ can take many forms, but I will focus primarily on those that involve negative consequence the speaker would (likely) face if she is found to be speaking irresponsibly.
Goldberg thus argues that we need accountability mechanisms in order for hearers to rationally come to believe assertions, which effectively rules out most anonymous assertions, or at least places anonymous assertions at a considerable disadvantage. And when speakers are operating in contexts where those punitive measures are absent, speakers will tend not to assert truthfully or reliably. In such contexts, then, hearers have (near-) omnipresent defeaters for such assertions. And since punitive measures are almost always completely absent in anonymous assertions, hearers cannot form justified testimony-based beliefs from anonymous assertions.9 9 There are some caveats, to be sure. For example, if the source of the assertions is vetted, then although it’s anonymous, the hearer has adequate reason to believe the assertion to form a justified belief (and even knowledge).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
464 Veronica Ivy
3. Giving Directions to Avoid the Punitive Model Here are two cases where I think most people (though certainly not all), in the absence of accountability mechanisms, will answer honestly and sincerely (or decline to answer if they don’t think they know) and will feel bad about speaking falsely, even though it would have no negative effects on, for example, their reputation. As I will argue, it seems that we’re motivated in these cases more by pro-social aims, rather than by avoiding punitive accountability mechanisms. I take these two cases to provide strong prima facie evidence that the punitive model (of assertoric practice) is false. First case: think back to the last time someone you don’t know asked you for directions.10 Did you try your best to answer honestly? If you didn’t know, did you admit that and say that you can’t help them (perhaps apologizing)? Call this case HONEST TRUTH. Now, on Goldberg’s view, these aren’t anonymous asser tions. But did you give the person your name and contact information? Was there any way for this person to hold you accountable if your directions turned out wrong? Now, I’m willing to bet that you answered honestly and sincerely (or told the person that you didn’t know, if you didn’t know), even though there were effectively no accountability mechanisms to punish you had you answered dishonestly.11 If you were to have given the wrong directions, it’s not as if the stranger could track you down, or even be able to write a social media post naming you as giving bad directions. Moreover, it’s unlikely that anyone in your social circle would know that you had given bad directions, unless you chose to disclose that. Effectively, then, there are no accountability mechanisms at work, and yet you chose to answer honestly. The punitive model, though, predicts that you would only tend to assert honestly if there is a sufficient disincentive, through accountability mechanisms, for lying. But even in the absence of accountability mechanisms, you asserted truthfully (or at least what you took to be true). Second case: now think back to the last time someone you don’t know asked you for directions, you answered honestly, but you later realized that you gave the wrong directions by mistake. Call this case HONEST MISTAKE. Again, was there any way for the person to hold you accountable for your directions being wrong?
10 Goldberg also recognizes cases of “stranger assertion” where we lack a great deal of information on which to perform the credibility assessment task. He argues that we tend to have enough, though, to form justified beliefs. However, he doesn’t provide much argument for this. Moreover, there’s nearly no argument for why the shift from stranger assertion to anonymous assertion takes us from just enough information to not enough information. Consequently, I suspect that the punitive model he adopts is doing the heavy lifting in his view. 11 Certainly, there are some local practices of dishonesty. For example, my students have informed me that it’s a practice in Charleston that when someone asks for a recommendation for a good night club, they send them to Pantheon (the main gay nightclub). So, while directions to, say, Patriots Point are likely to be reliable, directions to a good nightclub (where it’s implied that the request is not for a gay nightclub) are not—if one is asking college-age people.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Anonymous Assertions 465 Unlikely. But here’s the important question: did you feel bad that you had given the wrong directions? I’m willing to bet that you did (and that most people experience this). Moreover, I bet you experienced this feeling (maybe of regret or remorse) even if no one else knew that you gave incorrect directions. And this is important: even though accountability mechanisms were not plausibly at work, you experienced negative affect for unintentionally asserting falsely. The punitive model seems to predict that you wouldn’t feel bad for having given bad directions, though. Since there is no way for you to suffer the ill effects of asserting falsely, you shouldn’t experience negative affect.12 But you do. Why? I suggest that it’s more typical for speakers to exhibit pro-social behaviors: we tend to experience some sort of internal duty of honesty and truthfulness, and this drives our assertoric practices. Now, I’m not suggesting that punitive disincen tives play no role in our assertoric practices. Of course, they do! My argument is that the punitive model is wrong. We assert truthfully when there are no public, punitive disincentives for lying, and we feel badly for asserting falsely (some times) even when there are no public, punitive consequences. The punitive model predicts wrongly for both HONEST TRUTH and HONEST MISTAKE. Moreover, Frost-Arnold (2014a) argues that a focus on policing and account ability mechanisms “can both damage epistemic communities’ ability to dissem inate true beliefs and undermine their ability to weed out error” (64). She further notes that people are not motivated only by narrow self-interest. Instead, we are often motivated by “altruistic and pro-social concerns” (77). For example, there is a rapidly expanding body of research into socio- and behavioral economics sup porting this.13 I think that we should place more stock in taking seriously the view that speakers experience some fiduciary duties to their audience, even in cases of anonymous (internet) assertions. The upshot of this is that if we take ser iously the idea that speakers will tend to assert appropriately, on their own accord and not because of potential external punishments for asserting inappropriately, even in online anonymous contexts, then hearers can often be suitably epistemic ally positioned to form justified beliefs and knowledge. Furthermore, while the punitive model gets some key cases wrong about our assertoric practices, taking our assertoric practices to be guided primarily by punitive concerns leads to add itional problems. As I argue in section 6, it contributes to epistemic injustices.
12 One might think that while we can’t face any social punishment for giving bad directions, a speaker might feel bad for personal moral or religious reasons. That’s certainly true. But the punitive model argues that we aim to speak truthfully only (or chiefly) because we fear the repercussions of being caught speaking falsely. If one’s religion includes a deity that punishes false statements, then that is a form of punishment that would fit in with the punitive model. But such external sources of pun ishment don’t factor into atheists, or those whose religion does not include such punishments for speaking falsely. 13 For a small sample, see Batson and Powell (2003) and Mazar and Ariely (2006).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
466 Veronica Ivy
4. Epistemic Trust So, what should we do if we reject the punitive model of assertoric practice? The punitive model suggests that we should adopt a default attitude of skepticism with respect to anonymous assertions. On this view, one needs strong reasons to over come the skepticism.14 However, these would be exceptions to the rule. For my part, I suggest that we should adopt a default attitude of trust towards assertions, including anonymous assertions. However, this trust is defeasible: if we have strong positive reasons to doubt the assertion, then these reasons serve as epi stemic defeaters. But contra Goldberg, anonymity itself is not a defeater. On my view, doubting anonymous assertions would be the exception, rather than the rule. What would epistemic trust look like, though? Linda Zagzebski (2012) pro vides a useful characterization. Following Annette Baier (1995), trust is “accept ance of vulnerability to [epistemic] harm that others could inflict, but which we judge that they will not in fact inflict” (Zagzebski 2012, 152). When someone gives us directions, we epistemically trust them when we accept that in trusting them, we open ourselves to the possibility that what they say is false (and thus will lead us astray of our epistemic and practical aims), but we judge—whether accur ately or not—that they will not say something false or misleading. And when we think of anonymous assertions, when someone asserts anonymously, we trust that what they say is true, while simultaneously recognizing that it’s possible that it is false. But that’s the cost of trust. We trust strangers on the street to give us accurate directions (and to demur if they do not know). We do so while recogniz ing that their directions may lead us astray. Zagzebski argues that we should adopt a presumptive, prima facie attitude of epistemic trust to others. I agree: this is the attitude we ought to adopt with respect to other people’s assertions, even anonymous ones. The default position should be to trust assertions, and only to doubt when we have positive reasons to doubt. But anonymity alone is not a positive reason to doubt.15 The primary motivation for treating anonymity itself as a positive reason to doubt seems to assume a punitive model of assertoric practice. But once we abandon the latter, we lose the justification to treat anonymity itself as a reason to doubt an anonym ous assertion. In order to retain the view that we are not generally warranted in believing anonymous assertions, one would have to adopt some version of
14 And to be fair to Goldberg, he thinks that some cases (such as posts on the What It’s Like to be a Woman in Philosophy blog, and cases of whistleblowing where the costs of non-anonymity are very high) satisfy this requirement. 15 My view is thus in line with anti-reductionist accounts such as Weiner (2003). Interestingly, Michaelian (2010) argues that there’s empirical evidence that ‘blind trust’ does as well, epistemically speaking, as a reductionist sort of ‘monitoring’ for trustworthiness.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Anonymous Assertions 467 reductionism about testimony. Zagzebski’s view of trust is very close to—if not indistinguishable from—most forms of anti-reductionism. Why should we trust anonymous speakers, though? One reason is to reflect on our own behaviors in such contexts. Many of us will observe that our own online behaviors are such that when we write anonymous comments, we tend to aim at asserting truthfully.16 One example is participation in online forums, such as those dedicated to various hobbies or sports.17 One popular forum for triathletes and endurance athletes is Slowtwitch.com. It’s easy to find hundreds of threads where people pose questions, and anonymous users post honest, truthful answers and opinions. While this is an empirical question, in my experience in such spaces, intentionally false assertions are a tiny minority. On Zagzebski’s view of epistemic trust, reflecting on our own epistemic behavior gives us some reason to trust the anonymous assertions of others. While her arguments don’t broach anonymous online assertions, I think it applies just as well to anonymous asser tions as to non-anonymous assertions. Her argument is that we ought to treat like cases alike. I observe in my own anonymous online behaviors that I tend to assert honestly and truthfully. I have reason to think that other people are roughly like myself in relevant respects. Therefore, I’m prima facie warranted in trusting the anonymous assertions of others. It’s important that the warrant—or epistemic justification, whichever you prefer—is prima facie, though: it’s defeasible and subject to defeaters. At a minimum, my reading an anonymous online assertion that p provides me with a pro tanto reason to believe that p. However, as Goldberg successfully identifies, I might have information about the context of the assertion (i.e. where it was anonym ously asserted—for instance, if it’s a site known for a high ratio of troll comments such as 4chan.org), or background information about p (e.g. an anonymous assertion of p when I know full well that p is false) that defeats the prima facie justification that the assertion gives me to believe that p. One might wonder how strong Zagzebski’s argument is, though: it seems to depend on whether—and to what degree—I am justified in believing that others’ online anonymous assertoric behavior resembles my own. And one might think that even if one is a well-behaved anonymous online asserter, one might have good reason to think that one is exceptional in that regard. That is, we should recognize that there’s so much poor online anonymous assertoric behavior that we’re not sufficiently justified in believing that others are sufficiently like us. And to an extent I agree: Zagzebski’s argument only goes so far. However, I conjecture that on the whole, most online anonymous assertoric speech sufficiently
16 For instance, when I reflect on my own online anonymous commenting practices, I only assert anonymously when I fear that my truthful assertions may result in negative consequences for me, given the circumstances. 17 I thank Jonathan Shaffer for suggesting this example to me.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
468 Veronica Ivy resembles face- to- face, non- anonymous speech. Moreover, the prevalence of epistemically defective online anonymous assertions seems not obviously worse than non-anonymous face-to-face assertions. People often note—rightly, I think—that online discourse tends to be less kind, more hostile than offline discourse. Insults, misunderstanding, and straightfor ward vitriol seems far more prevalent online than offline. However, I notice many commentators attach this to anonymity of many online spaces, rather than to its being online. Goldberg, for example, writes, “It is perhaps not particularly con troversial to think that the anonymity that the online world provides is partially responsible for the quality here.”18 But this strikes me as a misdiagnosis. Online communication is more plausibly what produces these features, but that’s not an important point of debate. Consider the hostility and incivility ubiquitous in comment sections of news articles written by people under their real names (often using their Facebook accounts). In fact, Facebook is a very useful source of this evidence: simply look at the comment section of a controversial news article. False assertions, misleading information, even propaganda, are everywhere. The common feature, then, is not that they are anonymous (since the latter are not); the common feature is that they are online. While online assertoric speech is plausibly less polite, and there are distinct cultures of trolling, there doesn’t seem to be the sort of sufficiently widespread lying and misleading that would warrant the sort of default skepticism that those like Goldberg advocate.19 Perhaps this is fertile ground for some experimental work, but in my experience inhabiting a range of online assertoric contexts— forums, Twitter, blogs and news comment sections, and so on—while there may be many false statements, I see people generally at least aiming at asserting truth fully. That is, the false statements are offered as sincere assertions. Finally adopting a non-reductionist position of default trust (while still deploying screening procedures) doesn’t amount to gullibility. This is important. Too often ‘trust’ is inappropriately equated with gullibility. But gullibility connotes a kind of uncritical trust: trust even in the face of strong reasons not to believe. This is not what I advocate. Trust along the lines of “critical openness,”20 and through the deployment of screening procedures, where we accept a piece of testimony unless we have sufficiently strong reasons to doubt, is more active and critical than gullibility. Fricker (1994) argues that an anti-reductionist view along the lines I have sketched doesn’t amount to much difference from a standard reductionist account where the hearer must assess the speaker for trustworthiness. I disagree. Consider 18 Goldberg (2015, 204). 19 Certainly, with the evolution of Twitter and social media, it seems that online ‘discourse’ has continued to deteriorate. And while there is certainly widespread lying and trolling, it’s an empirical claim about how prevalent it is. I have thus tried to appropriately qualify my claims. 20 Fricker (2007) advocates for critical openness in our role as recipients of testimony.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Anonymous Assertions 469 an analogy. A nightclub could adopt either of two different screening procedures for its patrons. On the one hand, it could check identification at the entrance. People admitted into the club are deemed acceptable through a screening process, but once inside, they’re left largely unmonitored. On the other hand, it could admit everyone, but monitor those inside for signals of being underage (or other wise worthy of being excluded from the club). The former resembles a reduction ist account of testimony: we look for signs of reliability or trustworthiness and, if present, we accept the testimony. The latter resembles an anti- reductionist account of testimony: we presume trustworthiness unless we clearly see signs of untrustworthiness. These procedures differ substantially. Fricker’s claim that they do not does not have much support. Monitoring for trustworthiness (as the reductionist does) is quite different from monitoring (however passively) for untrustworthiness.
5. The Epistemology of Anonymous Assertions Now, although Goldberg’s view on anonymous assertions is driven largely by adopting the punitive model of assertoric practice, which I have argued is a poor account of our practices, that was merely the third prong of his argument. What about his first two arguments about the epistemic impoverishment on the hearer’s behalf in their credibility assessment task of the speaker? As noted in section 1, anonymous assertions quite easily fit within the broader topic of the epistemology of testimony. But what view should we take on the epistemology of testimony? Should one be reductionist, anti-reductionist, or perhaps something else? While my leanings are anti-reductionist, my view on the epistemology of anonymous assertions is neutral on this question. My view is that hearers can readily form justified testimony-based beliefs from anonymous assertions on reductionist and anti-reductionist accounts. In order for my argument to proceed, though, it’s critical to get the cases and examples right. Moreover, it’s methodologically important to have a good variety of cases. So, let’s think about the sorts of anonymous assertions that Goldberg argues fail to provide us justified testimony-based beliefs. Recall the four sources of information on which speakers perform the requisite credibility assessment task of speakers’ claims. It is fairly easy to think of cases where we lack the first two kinds of information, but we have some amount of information regarding the latter two kinds. For example, an anonymous assertion on a well-known website with a good reputation, on a topic that I know something about, can still permit me to form a justified belief about the asserted content. So, these aren’t the sorts of anonymous assertions that Goldberg is concerned with. Unfortunately, I note, he’s not very specific about the cases he’s thinking of. But we can flesh that out
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
470 Veronica Ivy easily enough. Part of my view is that when we carefully consider real examples, it seems implausible that we don’t regularly form justified testimony-based beliefs. I call those assertions for which we lack significant levels of information of all four kinds listed above fullY Anonymous unKnown assertionS, or YAKS. A recent social media platform was effectively designed to allow people to easily assert YAKS, called Yik Yak.21 For clarity, I’ll use ‘yak(s)’ to refer to posts on Yik Yak, and YAK(S) to refer to fully anonymous unknown assertions. People install the app on their phone, and based on their GPS location, they only see yaks local to them (although one can search other locations if one wants). Your own yaks will also only appear to those local to you (unless someone specifically searches for yaks from your location). The yaks are always anonymous (unless one chooses to use a pseudonym, or their real name, if they really wanted to). Supposing that we’ve just learned about the app, for us it doesn’t have a general reputation for being reliable or unreliable, so we lack the fourth kind of information. We don’t have information on the speakers or about the act of their assertion, so we lack the first two kinds of information as well. Now all we need for a YAK is a yak on a topic on which we don’t have any background information. Here’s an example of a real Yak written in the vicinity of my college, the College of Charleston: GLASS “Woah, someone busted the glass to the side door of Maybank Hall.” Suppose that as a reader, you are roughly aware of where Maybank Hall is (it’s one of the main classroom buildings on campus), but you haven’t been near it lately to know whether this yak is accurate. And the yak’s content isn’t about something you can have enough background knowledge to use to judge the reliability of the speaker or the assertion. You don’t know who the speaker is. You don’t particu larly know how likely it is for people to break in the glass of a door to a campus building. So, we have a YAK. On Goldberg’s view, one would not be justified in believing that the glass in the side door of Maybank Hall is broken.22 Let’s consider some more examples. Goldberg notes in a footnote (2015, 204, fn. 1) that another recent app called Secret—which is similar in some respects to Yik Yak—has had the sort of results that “are perhaps what one would expect,” citing a New York Times article on the app.23 But there’s no mention of a rash of false or epistemically defective assertions. Rather, the focus is on people using the app for gossip. What is gossip, though? Often, gossip involves a mix of true claims about people that are otherwise inappropriate to assert publicly, especially in 21 Indeed, the inspiration for calling fully anonymous unknown assertions “YAKS” comes from Yik Yak’s name. To be fair, though, I don’t think that true YAKS exist, where we completely lack all kinds of information on which to perform even the slightest of credibility assessment of a speaker’s asser tion. I agree with Kenyon (2013): testimonial contexts, even anonymous ones, are “informationally rich,” even if the information is implicit. 22 I want to make clear that Goldberg does think that there can be contexts in which we’re justified in believing anonymous assertions, but he takes them to be the exception to the rule. I’m arguing, rather, that doubting anonymous assertions should be the exception to the rule of trusting them. 23 New York Times (2014).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Anonymous Assertions 471 polite company. Jorg Bergmann (1993), for example, partly characterizes gossip thus: The relationship between the gossip producer and the gossip recipient that is based on mutual acquaintanceship is specifically characterized by the special kind of information that is transferred in their interaction. Information about another’s private affairs is morally contaminated information and thereby places those who exchange it in a relationship of co-informers. The gossip recipient finds [the]mself, as it were, in the situation of one who accepts a gift that [they] as well as the giver knows is stolen. This co-informership binds the gossipers together in [fraternity] and affects the ability of their relation to last.
Gossip is thus socially “contaminated,” but not epistemically. In an important sense, it’s socially contaminated because it’s not epistemically tainted: gossip involves the sharing of true information that one is not (socially) warranted in sharing.24 Sometimes, though, gossip is the spreading of rumors (which are often epi stemically defective—either false or lacking epistemic justification).25 But the worry with Secret, however, seems to be the true gossip, where this gossip is closely associated with sharing secrets. And while social norms, with a number of caveats, speak against gossiping or the sharing of secrets, these norms are in force because of the social effects of gossiping, and not because such assertions are epi stemically defective. In fact, the worries about negative social effects of gossip are generally due to the epistemic merit that such claims tend to have. Here are a few of the examples of assertions listed in the New York Times article on Secret. “I really hope the King I.P.O. is a flop.” “Battery put us through the wringer, but they never had any intention to invest.” “Going through a merger is like going on a double date as your friend’s wingman.” Some liken the behavior on Secret (and Yik Yak) to scribblings on bathroom stalls. But look at the assertions offered as typical. The first expresses what seems to be a genuine attitude of the author: that they hope an IPO fails. This strikes me as a good candidate for justified belief on our part as audience members. The second expresses the speaker’s belief that a company never had any intention of 24 I discuss the complicated interplay between epistemic and social norms of assertion in McKinnon (2015b). 25 Coady (2012) has a discussion of rumors.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
472 Veronica Ivy investing. Since it’s about someone’s reading of another party’s intentions, a subject that we tend to know is difficult to reliably form true beliefs on, we are unlikely to believe the speaker merely on the speaker’s say-so. And the third is offered as a humorous simile for going through a business merger. When we dig into the examples, Goldberg’s pointing to Secret at most seems to support the claim that online (and anonymous) assertions tend to be uncivil but doesn’t speak at all to his claim that such assertions tend to be epistemically defective. We monitor for defeaters—as much as we can—and none arise for the first assertion. It seems to be a clear candidate for justified belief on our part: I (justifiably) believe that the author hopes that the King IPO fails. In the second assertion, we notice a defeater: assertions about other parties’ intentions are notoriously unreliable. So, we don’t form a justified belief about the second message. In the third, it’s not entirely clear what proposition we’re evaluating. Is it that going through a merger is awkward? (That seems to be the intended content of the analogy.) Maybe this is a candidate for justified belief on our part: there are no obvious defeaters for the belief that going through a merger feels awkward. But since it’s not clear what the asserted content is, it’s not a clean example for analysis. Goldberg (2015, 214) recognizes that even in cases of anonymous assertion, we can have information about categories (1) and (2). Some online assertions involve very poor spelling and grammar. They may involve excessive use of expletives and insults. They may inappropriately abuse caps lock. Such messages can give us information about the speaker’s emotional or mental state of mind. Extended messages or exchanges can be assessed for internal coherence. But, he argues, in well composed messages, the audience has little information to go on in perform ing their credibility assessment.26 However, note that Goldberg has admitted that there are many cases where we are sensitive to the presence of defeaters! We’re often able to spot evidence of intemperance, even in online anonymous asser tions. So, his argument against the non-reductionist is diminished. We may not be robustly counterfactually sensitive to the presence of defeaters, but we are counterfactually sensitive to many kinds of defeaters that exhibit themselves, even in anonymous assertions. Moreover, we are not robustly counterfactually sensi tive to defeaters in non-anonymous assertions: otherwise, lying and misleading
26 I worry that Goldberg’s argument seems to admit that in many cases of anonymous assertion, we’re counterfactually sensitive to the presence of defeaters—and so in such cases we seem able to form justified beliefs—but in other cases we are not. He thus focuses only on the cases where we’re not counterfactually sensitive and concludes that these should stand as the paradigm for anonymous assertions: we’re generally not justified in believing anonymous assertions. But this is odd: he seems to define himself into his conclusion. Why take such cases as representative? Rather, he could have argued for a principle that one is justified in believing an anonymous assertion provided that one is counterfactually sensitive (to a sufficient degree) to defeaters. I suspect that, in most cases of anonym ous assertion, we satisfy that (admittedly externalist) requirement.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Anonymous Assertions 473 wouldn’t be such a widely successful practice. Indeed, our practices of lying depend on and exploit our lack of counterfactual sensitivity to defeaters.27 Simply put, his requirement of strong counterfactual sensitivity to defeaters is implausibly strong. We rarely, if ever, satisfy it when we form testimonial beliefs. At a minimum, it needs to be qualified. Consider the hearer in HONEST MISTAKE. The speaker is confident that they know the correct directions and asserts accordingly. But they’re wrong. The directions are incorrect. The speaker’s being wrong is a defeater for the hearer’s justification. But the hearer is not counterfactually sensitive to this defeater. On Goldberg’s (presentation of the anti-reductionist) view, the hearer is not justified in their belief. But that’s the wrong result. I worry that Goldberg’s requirement for a strong counterfactual sensitivity to defeaters entails that we cannot form justified false testimony-based beliefs. For those who do not subscribe to a factive account of epistemic justification, this would be a serious mark against Goldberg’s view.28 But let’s return to our analysis of GLASS. GLASS is well composed. The gram mar is fine, it doesn’t seem intemperate, and it doesn’t flout norms like using all capitals. While we’re not robustly counterfactually sensitive to defeaters, I have suggested that this doesn’t preclude us from forming justified testimony-based beliefs. On an anti-reductionist account of testimony, what’s stopping us from forming a justified belief? I submit that the answer is “Nothing.” We read this yak and can readily form a justified belief that the glass on the door to Maybank Hall is broken. And this strikes me as the correct result. On an anti-reductionist account of testimony, then, we form a justified belief about the broken glass, and this generalizes to many anonymous assertions, even YAKS. What is the epistemic status, though, of YAKS on a reductionist account of testimony? One might think that it would be difficult to accept anonymous asser tions as a source of knowledge on reductionist accounts, particularly on account of the positive reasons requirement. However, even in the barest of contexts of YAKS, the hearer has at least some information on which to perform the credibil ity assessment task required by the reductionist (including, I think, Goldberg). Kenyon (2013) argues that testimonial contexts are “informationally rich.” While we are not robustly sensitive to the presence of epistemic defeaters in testimonial contexts, we are relatively well attuned to a variety of sources of information. For example, if you want to know where a local landmark is, you don’t just ask anyone; that is, you don’t ask someone who looks just as lost as you. Or if you ask someone for the time, the way they answer matters: did they blurt out a time without looking at a watch or timepiece? Does their answer sufficiently accord
27 I would argue that our practice of bullshitting also depends on, and exploits, our not being robustly counterfactually sensitive to defeaters. For some discussion, see McKinnon (2015a) and (forthcoming b). 28 Littlejohn (2012) is one prominent example of a factive account of epistemic justification.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
474 Veronica Ivy with your background information about the time of day (i.e. did they say 9:18 a.m. when you’re confident that it’s still somewhere between 8 and 10 a.m.? Or did they say 9:18 a.m. when you’re confident that it’s the late afternoon?). Kenyon’s argument is that there is plenty of information to use to satisfy the “positive reasons” requirement of many reductionist accounts of testimony. One provocative option is Jennifer Lackey’s “dualist” position, which I take to be a version of reductionism. In short, one is justified in believing a speaker’s tes timony that p if and only if: (1) one believes p on the basis of the speaker’s testimony that p; (2) the speaker’s testimony that p is reliable or otherwise truth conducive; and (3) one has appropriate positive reasons for accepting the speaker’s testimony that p.29 Lackey thus builds in speaker reliability as a necessary condition for justified testimony-based beliefs. Her view is thus clearly an externalist account of testi monial justification. So how does it work? Suppose I ask you the time and you tell me that it’s 9:18 a.m. In order for me to form a justified belief that the time is 9:18 a.m., based on your testimony, on Lackey’s view I must form the belief based on your say-so, you must be reliable in your reporting the time (at least in this instance), and I have to have some positive reasons to judge you credible or reli able. I saw you look at your watch, and I know that it’s sometime in the morning, so I have some positive reasons for believing your report. So, on Lackey’s view, I’m justified in believing that it’s 9:18 a.m. Let’s now apply Lackey’s dualism to anonymous assertions like GLASS. First, suppose you read this and form the belief that the glass of the side door to Maybank Hall is currently broken, and you form this belief on the basis of read ing the yak. Next, let’s assume that what the speaker says is true, and they are reliable in their reporting. We’ve satisfied the first two conditions of Lackey’s dualist view of testimonial justification. All that’s left is whether you, as the audi ence of the message, have enough information to constitute appropriate positive reasons to regard the speaker’s testimony. Following Kenyon (2013), testimonial contexts—even anonymous online ones!—are informationally rich. You note that the grammar and spelling of the message are appropriate. It doesn’t show signs of intemperance. You also have some background information on its being relatively unlikely that someone would lie about this sort of message. Whether some door on campus has broken glass isn’t the sort of thing that people would be prone to
29 Paraphrasing a little from Lackey (2006a, 170).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Anonymous Assertions 475 lie about, let alone take time to yak about, unless it were true. This matters: this is information that contributes to your having positive reasons to believe the speaker’s assertion. I submit that audience members of this yak—and, indeed, YAK—are sufficiently well positioned to have appropriate positive reasons to believe the speaker, and so satisfy the final requirement of Lackey’s dualism. Therefore, you form a justified belief about the door. I think that this is how it goes for most anonymous online assertions. When the speaker is reliable, or otherwise truth conducive in their asserting that p, hearers have access to enough positive reasons to believe the speaker, and thereby form justified beliefs. By adding the speaker reliability requirement, Lackey’s view allows many anonymous assertions to permit justified beliefs in message recipi ents. Recall that Goldberg argues that we systematically lack the sorts of informa tion on which to do the requisite credibility assessment task—that is, to have sufficient positive reasons to accept a speaker’s testimony. But when we look at a range of examples and given the informational richness of testimonial contexts (even anonymous online ones), I have argued that hearers are not epistemically impoverished in anonymous (online) contexts. On the contrary, hearers quite readily have enough information on which to form justified beliefs about such assertions. Therefore, on either an anti-reductionist or reductionist account of the epistemology of testimony, hearers readily form justified beliefs in (online) anonymous assertion contexts.
6. Silencing and Epistemic Injustice I close with considering an additional worry attached to the punitive model of assertoric practice along with a default skepticism about anonymous (online) assertions. Namely, if our default position is to doubt anonymous assertions, unless we have overwhelming reasons to trust the source (which is essentially Goldberg’s view), then this will contribute to silencing and epistemic injustice. Following Fricker (2007), epistemic injustice—of the testimonial sort—occurs when a speaker suffers a credibility deficit due to identity biases or prejudices. There are a number of topics, though, where speakers are inclined to assert only if they can do so anonymously. One prominent example in the world of professional philosophy is the What It’s Like to Be a Woman in Philosophy blog, https:// beingawomaninphilosophy.wordpress.com . However, I have observed over and again that many people doubt the stories on the blog merely because they’re anonymous. Ironically, these expressions of doubt for anonymous assertions hap pen both non-anonymously and in anonymous online gossip forums. If speakers are suffering a systematic credibility deficit due to the manner of their assertion
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
476 Veronica Ivy (in this case, anonymous online assertions), this is properly understood of as an instance of epistemic injustice.30 Now, doubting an anonymous assertion doesn’t clearly fit Fricker’s mold of tes timonial injustice because the credibility deficit isn’t due to an identity prejudice or stereotype.31 However, it’s a credibility deficit due to inappropriate reasons nonetheless. People who are, in a sense, structurally forced to assert anonymously, if they are to assert (about particular topics, such as sexual or gender harassment) at all, will be excluded from being treated as the epistemic authorities that they are. This is textbook testimonial injustice, then. Moreover, if vulnerable speakers know that they’ll only be treated as an epistemic authority if they assert nonanonymously, but the risks and costs associated with non-anonymous assertion are too high, then they will likely simply remain silent.32 So, if it becomes a stand ard norm that we ought not to believe anonymous assertions (except in some rare contexts where the omnipresent defeaters are defeated by such considerations), then these sorts of people will likely stop speaking. Since they’ll (rightly) think that their anonymous assertions won’t secure adequate uptake and that it’s unsafe to non-anonymously assert, they’ll choose to stay silent. This is exactly what Kristie Dotson (2011) refers to as testimonial smothering. When an agent believes that her audience is unlikely to give her adequate uptake (and, indeed, to be hos tile to her assertion), she’ll respond by remaining silent. Critically, as Dotson argues, this is an instance of epistemic injustice. Moreover, it’s a primary form of epistemic oppression. The stakes here are often quite high. And the systematic doubting of anonymous speakers—and their exclusion as knowers—constitutes what Dotson refers to as testimonial quieting: the speakers suffer such a heavy credibility deficit that it’s as if the speakers didn’t assert at all. I take the worries of increased epistemic injustice and silencing as simply add itional reasons to reject the punitive model. So, if my arguments in section 2 con stitute strong prima facie reason to reject the model, I take these additional worries to provide additional support beyond merely prima facie support. Moreover, the remedy isn’t searching for additional reasons to believe these speakers: ex hypothesi, since they’re anonymous, such reasons are absent. Rather, the remedy is to adopt a default attitude of trust towards the speakers, and only begin to doubt when we have positive reasons to doubt (where anonymity itself is not such a reason). 30 For a survey on recent research on epistemic injustice, see McKinnon (2016b) and Kidd et al. (2017). 31 Although, I think a case could be made that anonymity itself is an identity in the relevant sense, such that people have an identity prejudice against anonymous agents that results in a pattern of cred ibility deficits. 32 See also Frost-Arnold (2014a, 2014b) and (forthcoming).
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Anonymous Assertions 477
References Baier, Annett (1995). “Trust and Anti-Trust,” in Moral Prejudices: Essays on Ethics. Cambridge, MA: Harvard University Press. Batson, Daniel C. and Adam A. Powell, (2003). “Altruism and Prosocial Behavior,” in Theodore Millon and Melvin J. Lerner (eds), Handbook of Psychology, Volume 5: Personality and Social Psychology. Hoboken, NJ: Wiley, 463–84. Bergmann, Jorg (1993). Discreet Indiscretions: The Social Organization of Gossip, trans. John Bednarz, Jr. New York: Aldine de Gruyter. Coady, C. A. J. (1973). “Testimony and Observation.” American Philosophical Quarterly 10: 149–55. Coady, C. A. J. (1992). Testimony: A Philosophical Study. Oxford: Clarendon Press. Coady, David (2012). What to Believe Now. Hoboken, NJ: Wiley-Blackwell. Dotson, Kristie (2011). “Tracking Epistemic Violence, Tracking Practices of Silencing.” Hypatia 26(2): 236–57. Fricker, Elizabeth (1994). “Against Gullibility,” in Bimal K. Matilal and Arindam Chakrabarti (eds), Knowing from Words: Western and Indian Philosophical Analysis of Understanding and Testimony. Amsterdam: Kluwer, 125–61. Fricker, Elizabeth (1995). “Telling and Trusting: Reductionism and AntiReductionism in the Epistemology of Testimony.” Mind 104: 393–411 (critical notice of Coady 1992). Fricker, Elizabeth (2002). “Trusting Others in the Sciences: a priori or Empirical Warrant?” Studies in History and Philosophy of Science 33: 373–83. Fricker, Elizabeth (2004). “Testimony: Knowing Through Being Told,” in I. Niiniluoto, Matti Sintonen, and J. Wolenski (eds), Handbook of Epistemology. New York: Springer, 109–130. Fricker, Elizabeth (2006a). “Testimony and Epistemic Autonomy,” in Jennifer Lackey and Ernest Sosa (eds), The Epistemology of Testimony. Oxford: Oxford University Press, 160–89. Fricker, Elizabeth (2006b). “Second-Hand Phenomenological Research 73: 592–618.
Knowledge.”
Philosophy
and
Fricker, Miranda (2007). Epistemic Injustice: Power and the Ethics of Knowing. Oxford: Oxford University Press. Frost-Arnold, Karen (2014a). “Trustworthiness and Truth: The Epistemic Pitfalls of Internet Accountability.” Episteme 11(1): 63–81. Frost-Arnold, Karen (2014b). “Imposters, Tricksters, and Trustworthiness as an Epistemic Virtue.” Hypatia 29(4): 790–807. Frost-Arnold, Karen (forthcoming). Who Should We Be Online? A Social Epistemology for the Internet. Oxford: Oxford University Press.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
478 Veronica Ivy Goldberg, Sanford (2006). “Reductionism and the Distinctiveness of Testimonial Knowledge,” in Jennifer Lackey and Ernest Sosa (eds), The Epistemology of Testimony. Oxford: Oxford University Press, 160–89. Goldberg, Sanford (2008). “Testimonial Knowledge in Early Childhood, Revisited.” Philosophy and Phenomenological Research 76: 1–36. Goldberg, Sanford (2013). “Anonymous Assertions.” Episteme 10(2): 134–51. Goldberg, Sanford (2015). Assertion: On the Philosophical Significance of Assertoric Speech. Oxford: Oxford University Press. Goldman, Alvin (2008). “The Social Epistemology of Blogging,” in Jeroen van den Hoven and John Weckert (eds), Information Technology and Moral Philosophy. New York: Cambridge University Press, 111–22. Kenyon, Timothy (2013). “The Informational Richness of Testimonial Contexts.” Philosophical Quarterly 63(250): 58–80. Kidd, Ian James, José Medina, and Gaile Pohlhaus Jr (2017). Routledge Handbook of Epistemic Injustice. New York: Routledge. Lackey, Jennifer (2003). “A Minimal Expression of Non-Reductionism in the Epistemology of Testimony.” Noûs 37: 706–23. Lackey, Jennifer (2006a). “It Takes Two to Tango: Beyond Reductionism and AntiReductionism in the Epistemology of Testimony,” in Jennifer Lackey and Ernest Sosa (eds), The Epistemology of Testimony. Oxford: Oxford University Press, 160–89. Lackey, Jennifer (2006b). “The Nature of Testimony.” Pacific Philosophical Quarterly 87: 177–97. Lackey, Jennifer (2006c). “Learning from Words.” Philosophy and Phenomenological Research 73: 77–10. Littlejohn, Clayton (2012). Justification and the Truth Connection. Cambridge: Cambridge University Press. Mazar, Nina and Ariely, Dan (2006). “Dishonesty in Everyday Life and Its Policy Implications.” American Marketing Association 25(1): 1–21. McKinnon, Rachel (2013). “The Supportive Reasons Norm of Assertion.” American Philosophical Quarterly 50(2): 121–35. McKinnon, Rachel (2015a). “Propaganda, Lies, and Bullshit in BioShock’s Rapture,” in Luke Cuddy (ed.), BioShock and Philosophy: Irrational Game, Rational Book. Malden, MA: Wiley-Blackwell, 107–113. McKinnon, Rachel (2015b). The Norms of Assertion: Truth, Lies, and Warrant. New York: Palgrave Macmillan. McKinnon, Rachel (2016a). “How to Be an Optimist about Aesthetic Testimony.” Episteme Online (March 1, 2016), 1–10. McKinnon, Rachel (2016b). “Epistemic Injustice.” Philosophy Compass 11(8): 437–46. McKinnon, Rachel (2017). “Allies Behaving Badly: Gaslighting as Epistemic Injustice,” in Ian James Kidd, José Medina, and Gaile Pohlhaus Jr (eds). Routledge Handbook of Epistemic Injustice. New York: Routledge, 167–174.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
The Epistemology of Anonymous Assertions 479 McKinnon, Rachel (2018). “The Epistemology of Propaganda.” Philosophy and Phenomenological Research XCVI(2): 483–89. Michaelian, Kourken (2010). “In Defence of Gullibility: The Epistemology of Testimony and the Psychology of Deception Detection.” Synthese 176: 399–427. New York Times (2014). “Anonymous Social App Offers a Hotbed of Silicon Valley Gossip.” New York Times (March 19, 2014), www.nytimes.com/2014/03/19/ technology/new-social-app-has-juicy-posts-but-no-names.html. Weiner, Matthew (2003). “Accepting Testimony.” The Philosophical Quarterly 53(211): 256–64. Zagzebski, Linda (2012). Epistemic Authority: A Theory of Trust, Authority, and Autonomy in Belief. Oxford: Oxford University Press.
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Index Abramson, Kate 123 n.13, 142, 143 n.39 Adamic, L. A. 398 Adleberg, Toni 228–29 Adler, Jonathan 258 n.19 affirmative consent 321, 322 n.1, 337, 342, 344, 348–9, 351, 373, 374 n.26, 375, 378 n.29, 379 n.29, 381, 383 agentive authority 332–4 Ahrens, Courtney E. 260 n.24 Aikin, Scott F. 403 n.14 Albright, Jonathan 399 Alcoff, Linda Martín 194, 215 n.2, 258 n.21 Alexander, Joshua 61 Alexander, Michelle 205, 306 n.2 Alexander, Larry 334 Allen, Anita 226 Allen, Nicole E. 260 n.24 Alston, William P. 371 n.22, 390 n.2 Alvarez, Lizette 309 n.9, 310 n.11, 311 n.12, 315 n.19 Ambrose, Zandrea 83 n.55 Anderegg, Christopher 81 n.48, 82 n.51, 83 n.54 Anderson, Elizabeth 53, 201, 208–9, 217 n.5 Anderson, Janna 399 Anderson, Scott 307 n.5, 318 n.24 animal experimentation 4, 5, 67–8, 70–1, 72 n.13, 73 n.14, 74, 83–4, 91 n.91, 92 n.92, 94 animal models 5, 72, 78 n.24, 80–3, 84 n.62,65, 85–9, 91 n.91, 92 n.92, 93 anonymous assertions 15, 457, 458 n.1, 459, 461–3, 466–9, 470 n.22, 472 n.26, 473–6 Antony, Louise 100 n.4, 215 n.2, 228, 232 Archard, David 382 n.33 Archibald, Kathy 81 n.48, 82 n.51, 83 n.54 Ariely, Dan 465 n.13 Aristotle 290, 291, 453 Asad, Talal 205 Ashurst, Libby 120 n.8, 123 n.13 Audi, Robert 340 n.36 Baehr, Jason 446 n.10, 447 n.11, 448 n.12 Baier, Annette 448, 466 Bailey, Jarrod 81 n.48, 82 n.51, 83 n.54 Baker, Katie J. M. 339
Baker, Lee D. 205 Baker, Timothy 260 n.24 Bakshy, E. 398 Balcombe, Jonathan 77 n.22 Baptist, Edward E. 206 Barajas, Joshua 132 n.26 Barker, Kathy 120 n.10 Baron, Marcia 382 n.33 Bar On, Bat–Ami 172 n.3, 318 n.22, 25 Bashir, Layli Miller 181–3 Basile, Kathleen C. 256 n.12 Basu, Rima 5, 107 n.25, 108 n.26, 109 n.28, 111 n.35, 350 n.2 Batchelor, Nim 230 Batson, Daniel C. 465 n.13 Baumann, Hansel 46 Beal, Morgan R. 256 n.11 Beebee, Helen 216 Beech, Anthony R. 120 n.3 Begby, Endre 112 n.36 behavioral mechanism 124–5, 129 Belle, Kathryn S. 215 n.2 benefits argument 5, 67, 72 n.13, 73, 76–7 Benford, Robert D. 21 Berenstain, Nora 239 Bergmann, Jorg 471 Bergmann, Michael 340 n.35 Bergner, Raymond 157 Bernard, Claude 73 Berscheid, Ellen 258 n.20 Bettcher, Talia 215 n.2, 3 Biddle, Justin 45 Biddle, Justin B. 43 biomedical research 5, 67, 72, 73 n.14, 76, 77–8, 88 n.82, 91 n.91, 92 n.92 Biss, Eula 318 n.23 Blackburn, Robin 197 Blazer, Dan.G. 255 n.6 Bloch–Schulman, Stephen 230 Blumenstyk, Goldie 444 bodily testimony 7, 8, 171–2, 174–6, 178, 180–1, 183–8 bodily group testimony 8, 184, 185 n.9, 186, 187 n.11, 188 body language 174, 177–8, 180, 182–3, 188, 375 Bogost, Ian 416 n.10
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
482 Index Bolinger, Renee J. 100 n.3, 105 n.20, 106 n.22, 107 n.25, 112 n.9 Bonilla–Silva, Eduardo 211 BonJour, Laurence 232–3, 340 n.34, 36 Borschmann, R. 256 n.12 Botts, Tina F. 215 n.2, 216, 235 n.17 boyd, danah 411, 437, 446 Boykoff, Jules 21, 23 Brabham, D. C. 395 Brackenridge, Celia 120 n.5 Brackenridge, Celia 120 n.5 Brady, Michael 184 Brady, William J. 416 Brandl, Bonnie 120 n.9 Brandom, Robert 165 Breiding, Matthew J. 256 n.12 Brickman, Ellen 260 n.24 Bridges, Ana J. 284 n.5 Bright, Liam Kofi 215 n.2, 216, 235 n.17 Brister, Evelyn 271, 275–6 Bronstein, Carolyn 283 n.2 Brown, Jessica 323 n.5, 366 n.16 Brown, Sarah 120 n.2 Brownmiller, Susan 177, 256 n.7, 283 n.2 Brownstein, Michael 277 n.57 Buckley, Cara 310 n.11, 311 n.12, 315 n.19 Buckwalter, Wesley 228 Buchwald, Emile 256 n.8 Burge, Tyler 114 n.41, 340 n.36 Burnham, M. Audrey 260 n.24 Burt, Martha R. 256 n.8
Cohen, Carl, with T. Regan 70 n.9, 72 n.13, 73, 84 Cohen, Stanley 37 n.1 Cohen, Stewart 253, 268 Cohn, Carol 222 common sense 70, 114, 197, 199, 258, 339 Conee, Earl 100 n.2 consent 3, 10–13, 287, 305, 306 n.2, 3, 307, 308 n.7, 311–13, 314 n.18, 315, 318–19, 321 n.1, 322 n.1, 325 n.13, 326 n.15, 327 n.18, 328, 329 n.19, 330–2, 333 n.23, 334, 335 n.27, 336 n.30–1, 337 n.33, 338–42, 343 n.40, 344, 348–51, 353 n.3, 353 n.6, 354, 355 n.9, 356, 357 n.11, 359–66, 368, 372, 373 n.23, 374 n.26, 375–7, 378 n.29, 379 n.29, 380 n.31, 381 n.32, 382 n.34, 383, 437, 443 context collapse 14, 437–8, 439 n.2, 440–1, 442 n.5, 443–8, 449 n.13, 14, 450, 453 contextualism 10, 255, 267, 268 n.43, 44, 271–7, 350 n.1 conversational contexts 68, 268, 272, 277, 355, 411 Conway, Erick M. 402 n.13 Cote, Ashley M. 260 n.26 Craig, Edward 154 Craven, Samantha 120 n.2 Crawford, Lester 84 n.61, 84 n.64, 92 n.92 Cuenca, Alex G. 88 n.83, 89 n.87, 90 Culp–Ressler, Tara 374–5 Curley, Edwin M. 314 n.17 Curtis, Annaleigh 254 n.3
Cahill, Ann J. 230 Calhoun, Cheshire 215 n.2, 229 Callimachi, Rukmini 130 n.24 carceral logic 4, 29 Carruthers, Peter 70–1 Cavender, Gray 157 Chamberlain, Lorna 91 n.91 Chang, Nancy 32 Cherry, Myisha 215 n.2, 216, 235 n.17 Chisholm, Roderick M. 340 n.36 Chou, Yu–kai 413 n.6 Chung, Nancy P. Y. 83 n.55 Ciccariello–Maher, George 444–5 Cimpian, Andrei 224 n.10 Citron, Danielle 441 n.4 CNN 195 Coady, C. A. J. 459 n.2 Coady, David 408, 437, 471 n.25 Coates, Ta–Nehisi 295 n.25 Code, Lorraine 41 n.2 coercion 6, 11, 120–1, 140, 143, 306, 307 n.5, 308 n.7, 309, 312, 314–15, 318–19, 322 n.1, 335, 354, 356
Daggs, Joy L. 123 n.15 Darby, Derrick 208 Darwall, Stephen L. 123 n.13, 392, 396–7 Daston, Lorraine 59–60 Davis, Angela 19, 20 n.1, 25, 28, 35, 318 n.22, 320 Davis, Emmalon 8, 9, 239 Davis, Jenny 437, 440, 443 Davis, Robert C. 260 n.24 de Laat, Paul B. 437 defeater 12, 15, 40, 75, 321, 339, 340 n.35,36, 341, 457–9, 461–3, 466, 467, 472 n.26, 473 n.27, 476 Dembroff, Robin 106 n.22, 111 n.34, 215 n.2, 218 n.6 Denhollander, Rachel 133–5, 138 DeRose, Keith 42, 162, 269 n.43–4, 271 n.47, 324 n.9–10 Devlin, Shannon 120 n.6 Diaz–Leon, Esa 273–5 Dillon, Robin S. 396, 398 disagreement 53, 54, 61, 153 n.2, 162, 270, 273–5, 288, 290 n.16, 305, 329 Donovan, Roxanne 256 n.9, 318 n.26
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Index 483 Dougherty, Tom 306 n.3, 327 n.17, 335 n.28, 352 n.4, 355 n.9, 373 n.23, 378 n.29, 379 n.29 Douglas, D. M. 101 n.6, 103 n.11–12, 106 Douglas, Heather E. 43, 59 n.12 Dotson, Kristie 3–4, 6 n.2, 19, 29, 136 n.32, 138–9, 167 n.5, 171, 179 n.7, 180 n.8, 217 n.5, 225, 230, 231 n.13–4, 244, 404 n.15, 441–3, 476 doxastic wronging 5, 99, 106 n.23, 107 n.25, 108, 109 n.27, 110 n.31, 111, 112 n.36, 113–15 doxastic voluntarism 140–1, 390 n.2 Dretske, Fred 101 n.7 Driver, Julia 447 n.11 Du Bois, W. E. B. 49, 193, 202, 209–11 Ducharme, Lori 157 Duff, Antony 382 n.34 Du Mont, Janice 255 n.6 Dushku, Eliza 134 n.28, 136 Dutilh Novaes, Catarina 10, 285, 290 n.17, 291 n.19, 292 n.20, 294 n.22 Dworkin, Andrea 283 n.1, 284, 286, 290 n.18, 299 Dworkin, Emily R. 260 n.24 Eaton, A. W. 10, 285 n.8, 288 n.13, 289 n.14, 290 n.17, 291–4, 294 n.23 Ebony 193 echo–chambers 13–4, 398, 400, 429–30, 431 n.32, 432–3, 442 n.7 Ellevold, Barbara L. 123 n.15 Elliott, Ian 120, 123 n.14–6, 124 n.17, 125 n.19, 126 n.20, 127 n.21 Elliott, Marc N. 256 n.10 Ellison, Nicole 438 End Rape on Campus 322 n.1 Engel Jr., Mylan 4–5, 75 n.20 Engel, Pascal 141 n.37, 141 n.38, 196, 390 n.2 Entman, Robert M. 21 epistemic agency 5, 12–3, 44, 122 n.12, 140, 141 n.37–8, 142, 144 n.40, 171–2, 325, 338, 343–4, 368 n.19, 389, 390 n.1,2, 391 n.4, 392–6, 398–401, 402 n.12, 403 n.14, 404–7 responsible epistemic agent 337, 342, 392 reasons–responsiveness accounts 141 collective 144 n.4 epistemic anxiety 261–2 epistemic arrogance 13, 392, 396–9 epistemic authority 135, 143, 332, 334, 335 n.27, 442 n.5, 457, 476 epistemic benefits 5, 68, 73–4, 77, 91–2, 438, 443, 449, 452 epistemic health 130–1 epistemic infringement 5, 6, 122, 133 n.27, 140–2, 144
epistemic injustice 8, 153 n.2, 171 n.1, 172, 178, 187–8, 193–4, 196–8, 201, 207–9, 211–12, 215–17, 218 n.6, 222, 225, 227–9, 232–3, 235, 238, 241, 243, 259, 276 n.56, 442, 457, 475, 476 n.30 epistemic justification 5, 74–5, 78, 115 n.42, 332, 348, 366, 368–9, 372, 467, 471, 473 n.28 epistemic misconduct 121–2, 136, 137 n.34, 139–40, 142 epistemic norms 5, 61, 122, 140, 142, 144, 226, 262, 265, 350, 390, 463 epistemic resistance 171 n.1, 172, 188, 430 n.31 epistemic responsibility 11, 178, 183, 257, 306, 315 epistemic violence 138, 139 Espeland, Wendy Nelson 423 n.21, 424 n.23 Evancic, Christine 256 n.9, 318 n.26 Evans, Jennifer L. 256 n.11 externalists 74, 76 Fallis, Don 437, 443, 449 n.13 false efficacy predictors 84–5, 92 n.92 false inefficacy predictors 85, 92 n.92 Fantl, Jeremy 42, 58, 104 n.17, 253 n.1, 263 n.35, 323 n.6, 325, 336 n.30, 366 n.16 Fasting, Kari 120 n.5 Fay, George 158 FDA 81, 83–7, 91 n.91, 92 n.92 Feagin, Joe R. 193, 211 Fekete, Liz 32 Feldman, Richard 100 n.2, 371 n.22 Fellmeth, Robert 405 n.17 feminist pornography 10–11, 285 n.6, 286, 287 n.11, 288–9, 290 n.17, 291 n.19, 292–3, 294 n.22, 296, 298–9 feminist propaganda 285, 290, 296, 299 Fernbach, Philip M. 396 Ferzan, Kimberley 353 n.6, 381 Figdor, Carrie 235 n.17 filter–bubbles 13, 398, 400 Fine, Cordelia 228 n.11 Fine, Gary 157 Finlayson, Lorna 286, 288 n.12, 289 Fisher, Matthew 396 Flaherty, Colleen 52 Fleming, Alexander 86, 87 n.77 Fletcher, Pamela 256 n.8 Foley, Linda A. 256 n.9, 318 n.26 forensic psychology 6, 139 Fox, Craig R. 396 framework shift 3, 19–20, 26, 28, 30–1 Frankfurt, Harry G. 416 n.10 Freeland, Edward 224 n.10 Fricker, Elizabeth 340 n.36, 404 n.16, 422 n.19, 444, 459 n.2, 468 n.20, 469
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
484 Index Fricker, Miranda 6 n.1, 136 n.30, 137, 139, 151–2, 153 n.2, 154, 155 n.3, 156, 163, 166–8, 171, 176–7, 178 n.6, 184, 194, 196–8, 200–1, 209, 217, 221, 241–3, 244 n.21, 258 n.21, 259, 272 n.49, 276 n.56, 284 n.3, 397–8, 442–3, 475–6 Fritz, James 350 n.2, 367 n.17, 368 Frohmann, Lisa 258 n.20 Frost–Arnold, Karen 14–15, 458, 463, 465, 476 n.32 Fuller, Steve 391 n.4 Gabrielle, Vincent 413 Galison, P. 41, 59–60 Gallagher, James C. 91 n.91 gamification 14, 411–12, 413 n.6, 414–15, 416 n.10, 418–21, 424, 425 n.24, 426–9, 431, 433 Garcia, Jorge L. A. 199 Gardiner, Georgi 101–3, 112 n.36, 350 n.2, 368 n.20 Gardinier, Lori 260 n.26 Garfinkel, Harold 156–7, 159 Garrett, Jeremy 68 n.6 gaslighting 133 n.27, 140–4, 443 Gauker, Christopher 271 n.48, 272 George, Linda K. 255 n.6 George, William H. 256 n.9, 258 n.20 Gendin, Sidney 82 n.50 Gendler, Tamar Szabó 102 n.8, 290 n.16 Gertler, Brie 335 n.29 Gettier, Edmund 261 n.29, 446 Gilchrist, Elizabeth 120 n.2 Gilligan, Carol 228 Gilmore, Leigh 258, 259 Gladwell, Malcolm 119 n.1, 126 Gliddon, George Robins 203 goal–oriented behavioral mechanism; see behavioral mechanism rapport building 124, 125 n.18, 127–32, 134 n.28 incentivization 124, 125 n.18, 129, 131–2, 134 n.28 disinhibition 123, 124, 125 n.18, 127 n.21, 129–133, 134 n.28 security management 124, 125 n.18, 126, 128, 132, 133, 135, 137 n.34 disclosure 123–4, 125 n.18, 126 n.20, 260 Goddu, Mariel K. 396 Goldberg, Sanford 15, 340 n.36, 391 n.4, 457, 458 n.1, 459 n.3, 460 n.5, 461 n.7, 462–3, 464 n.10, 466 n.14, 467, 468 n.18, 469, 470 n.22, 472 n.26, 473, 475
Golding, Jacqueline M. 260 n.24 Goldman, Alvin 74, 331 n.22, 340 n.35–6, 437, 439, 446, 462 Gomes, Anil 224 n.9 Goodchild, Michael 40 Gould, Stephen Jay 196, 203 Gracia, Jorge J. E. 215 n.2 Granucci, Anthony 158 n.4 Gray, Kishonna 157 Greco, David 371 n.22 Greek, C. Ray 78 n.24, 79 n.25, 27, 28, 80 n.30,34, 81 n.41,42,44,46, 82 n.41, 83 n.57, 59, 84 n.60, 85 n.67–8, 86 n.69,71,73,76, 87 n.77–8, 88 n.81, 89 Greek, Jean Swingle 78 n.24, 79 n.25, 27, 28, 80 n.30, 34, 81 n.41, 42, 44, 46, 82 n.41, 83 n.57, 59, 84 n.60, 85 n.67–8, 86 n.69, 71, 73, 76, 87 n.77–8, 88 n.81, 89 Greenough, Patrick 276 n.55 Grosz, Elizabeth 172 n.3 group testimony 8, 172, 184, 185 n.9, 186, 187 n.11, 188, 405 n.18 Guerrero, Alexander A. 12–13, 268 n.44, 338 n.34, 343 n.38, 350 n.1, 354 n.8, 365–6, 369 n.21, 370, 380 n.30 Gunn, Hanna K. 13, 396, 404 Haack, Susan 53 Hacking, Ian 53–4 Hagopian, Amy 120 n.10 Hamilton–Giachritsis, Catherine E. 120 n.3 Hampton, Jean 256 n.7 Haring Hidore, Marilyn 120 n.7 Harding, Sandra G. 44, 45, 172 n.3 Hargittai, Eszter 438 n.1 Harman, Gilbert 70 n.8 Hartsock, Nancy 44, 57 Harvin, Cassandra Byers 179–81, 183 Haslanger, Sally 54, 215 n.2–3, 226, 229, 235 n.17 Hatziioannou, Theodora 83 n.55 Haugeland, John 41 Hawkins, Mike 203 Hawthorne, John 42, 253 n.1, 276 n.55, 322 n.3, 324 n.8, 325, 340 n.36, 366 n.16 Healicon, Alison 259 n.22 Heath, Joseph 53 Hekman, Susan 56 n.10 Hegel, G. W. F. 38, 39, 55 Heine, Steven J. 222 n.8 Heiner, Brady 25, 29 Heisler, Candace J. 120 n.9 Hempel, Carl G. 43 n.3
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Index 485 Hendee, William R. 73 n.14 Henning, Tempest 7–8, 172 n.3 hermeneutical injustice 137, 138 n.35, 171, 176–7, 188 n.12, 194, 196, 198, 244 n.21 Hess, Amanda 374 n.26 Hieronymi, Pamela 141 n.36, 38 Hill–Collins, Patricia 48, 49, 266 n.40, 318 n.25 Hintikka, Jaakko 371 n.22 Holroyd, Jules 277 n.57 Hong, Kari 382 n.33–5 Hong, Lu 418 n.11 Hookway, Christopher 221 Houghton, Ruth 285 n.7 Hopkins, Lauren 132 n.25 Hornsby, Jennifer 283 Horsman, Reginald 205 Hsieh, Yuli P. 438 n.1 Huebner, Bryce 40 Hull, Gordon 445–6 Hume, David 174, 175 Hund, Wulf D. 203 Hundleby, Catherine E. 265 n.38 Hunter, D. 105 n.21 Hurd, Heidi M. 306 n.2, 334 n.24, 353 n.5 Husak, Douglas 375 n.27, 376 n.28, 377 Hutchinson, Katrina 215 n.2 Hynes, Michelle 256 n.10 Ichikawa, Jonathan Jenkins 9, 10, 42, 215 n.2, 260 n.27, 261 n.27, 262 n.30, 268 n.24, 366 n.15 ideal theory 194 ideology 4, 10–11, 37, 45, 52, 58, 152–3, 155, 196–201, 207, 255 n.5, 257, 285 n.7, 286, 290 n.16, 18, 291–3, 294 n.24, 295 n.25, 296–9 ideological Catch–22 situation 286 inductive risk 43–4, 50, 58, 59 n.12, 104 n.17, 105–7 Intemann, Kristen 57, 441, 442 n.5 internalists 74, 76 n.21 internet 3, 13, 106, 389–97, 398 n.8, 399–400, 401 n.11, 402–7, 410 n.2, 437–9, 440 n.3, 441, 446–9, 451–2, 465 ISIS 130 n.24 isolated secondhand knowledge 325 Jackson, Michael 125 n.8, 128–9 Jaggar, Alison 442 n.7 JAMA 91 n.91 James, Joy 34 n.11 James, V. Denise 226–7 Jamieson, Ruth 26–7, 28 n.5, 30–1
Jarvis, Benjamin W. 262 n.30 Jeffrey, Richard C. 50 Jenkin, Z. 115 n.42 Jenkins, Fiona 215 n.2 Jenkins, Katharine 289 n.15 Jennings, Carolyn Dicey 235 n.17 Jeong, Sarah 441 n.4 Jeppesen, Lars B. 395 Johnson, Antuan M. 254 n.3 Johnson, G. 100 n.4, 101 n.5, 6, 103 n.10–12, 104 n.14, 17, 108 n.26 Jones, Anthony 158 Jones, Karen 181–2, 258 n.21, 327 Jonsson, Patrik 310 n.11, 315 n.19, 317 n.21 Jordan, Jan 259 n.23 Jordan, Winthrop D. 202 Jost, John T. 416 Jurgenson, Nathan 437, 440, 443 Kant, Immanuel 38, 39, 45 n.4, 50, 52, 55, 109–10, 154, 158 Karnik, Karnik 256 n.9, 318 n.26 Kassindja, Fauziya 181–3 Kawall, Jason 447 Keil, Frank C. 396 Kelland, L. 272 n.50 Keller, S. 113, 114 Kenyon, Timothy 460 n.4, 461 n.6, 470 n.21, 473–4 Khalifeh, H. 256 n.12 Kidd, Ian James 194, 476 n.30 Kim, David Haekwon 215 n.2 Kim, Mimi 29 Kim, Minsun 228 n.11 Kindermann, Dirk 276 n.55 King, Janet 256 n.9, 318 n.26 Klaassen, Marleen 284 n.5 Kloess, Juliane 120 n.3 Knoll, James 120 n.4 knowledge norm of practical rationality 322, 323 n.7 knowledge norm of practical reasoning 322, 323 n.7 known unreliability 76, 92 n.92 Kollmeyer, Barbara 80 n.39 Kornblith, Hilary 390 n.2 Koss, Mary P. 255 n.6 Krishnamurthy, Meena 215 n.3 Kuhn, Thomas S. 103 n.11, 195 Kukla, Rebecca 40, 43; see also Kukla, Quill R. Kukla, Quill R. 4, 57, 104 n.17, 374 n.25 Kushel, Margot B. 256 n.11 Kvanig, Jonathan L. 57, 58
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
486 Index Lacey, Hugh 58, 59 Lackey, Jennifer 12, 105 n.18, 106 n.23, 137 n.33, 173–4, 184, 185 n.9, 187 n.11, 258 n.19, 325 n.12, 340 n.35–6, 368 n.19, 375 n.27, 459 n.3, 474–5 Ladner, Joyce 209 LaFollette, Hugh 74, 79 n.28, 80 n.29, 31, 83 n.59, 85 n.68, 86 n.75 Lakhani, Karim R. 395 Lam, Barry 228 n.11 Lancet 79 n.27, 82 n.52 Langton, Rae 109, 283 n.1, 2, 284 n.3, 4, 289, 296 Latulipe, Celine 445–6 law enforcement 9, 253, 254 n.2, 255 n.4, 265, 267–9 Lebron, Chris 237–8, 241 Lederer, Laura 283 n.2 Lee, Carole 215 n.2 Lee, Emily S. 215 n.2 Le Gallois, M. 71 n.12 Lehrer, Keith 75 n.19, 76 Lepock, Christopher 446 Leslie, Sarah–Jane 224 n.10 Levi, I. 105 n.19 Levin, Mark 88 Levinson, Stephen C. 125 n.19 Levy, Neil 265 n.37 Lewis, David 101 n.7, 104 n.15, 258 n.18, 268 n.43, 272 n.51, 273 n.53 Liberman, Rachel 284 n.5 Liberto, Hallie 11, 326–7 Lindemann, Hilde 110 n.30, 32, 111, 223 Lipford, Heather R. 445–6 Lisak, David 260 n.26 Littlejohn, Clayton 58, 262 n.31–2, 473 n.28 Lloyd, Genevieve 224 n.10 Lockard, Claire A. 230 long–con 121 Longino, Helen E. 45, 103 n.12, 104 n.14, 441, 442 n.5 Lopes Cardozo, Barbara 256 n.10 Ludwig, Kirk 326 n.15 Lugones, Maria C. 56 n.10, 57 Lugones, Maria C. 420 n.15 Lynch, Michael P. 13, 396, 404 Machery, Edouard 228, 362 n.13 Mack, Kathy 258 n.20 Mackinnon, Catharine 256 n.13, 257, 265, 283 n.1, 284, 288–9, 290 n.18, 296, 299 n.27 Madden, Vicky 47 n.7 Madva, Alex 215 n.2 Maes, Hans 290 n.17 Magnus, P. D. 437
Maitra, Ishani 137, 155, 164, 167 Maldonado–Torres, Nelson 216 n.3 Mallarangeng, Gunter 215 n.2, 216, 235 n.17 Mallon, Ron 228 manipulation 5–6, 119, 121, 124, 140, 142, 195, 197–8 Manne, Kate 264 n.36, 285 n.7 Markham, Annette 440 n.3 Marlin, Randal 290 Marmor, A. 106 n.22 Marshall, Grant N. 256 n.10 Martin, Brian 157 Martínez, Lorraine J. 256 n.9, 258 n.20 Martínez, Tiffany 47 n.8, 256 n.9 Marušić, Berislav 101, 109, 112–14, 141 n.38 Marwick, Alice E. 411, 437 Mazar, Nina 465 n.13 McAlinden, Anne–Marie 120 n.8, 123 n.13, 15 McAusland, Pamela 255 n.6 McConnachie, Kirsten 26–7, 28 n.5, 30–1 McCormick, Miriam Schleifer 141 n.38 McDowell, John 340 n.36 McEvoy, Kieran 26–7, 28 n.5, 30–1 McGlynn, Aidan 10–11, 284 n.3, 285 n.7, 289 n.15, 294, 296, 299 n.27 McGlynn, Claire 285 n.7 McGrath, Matthew 42, 58, 104 n.17, 253 n.1, 263 n.35, 323 n.6, 325, 336 n.30, 366 n.16 McGregor, Joan 308 n.7, 312 n.14 McHugh, Conor 141 n.37–8, 190 n.2 McKee, Alan 287, 288 n.12, 292 McKinlay, John B. 73 n.14 McKinlay, Sonya M. 73 n.14 McKinnon, Rachel 46 n.6, 443, 459 n.3, 460 n.5, 461 n.6, 471 n.24, 473 n.27, 476 n.30 McMillan Cottom, Tressie 439, 440, 445–6 Medina, José 7–8, 171 n.1, 172 n.2–3, 178 n.6, 180, 194, 217 n.5, 276, 430 n.31, 476 n.30 Meskhidze, Helen 230 Messing, S. 398 Meyer, Meredith 224 n.10 Meyrowitz, Joshua 438 Michaelian, Kourken 466 n.15 microdosing technology 93–4 micro–invalidation 180 n.8, 181 micro–resistance silencing 180–1 Midgley, Mary 71 n.11 Mikkola, Mari 283 n.1, 285 n.6, 286 n.9, 290 n.17, 293 n.21 Miller, Boaz 41, 44, 125 n.19, 391, 398 n.8, 411, 437 Miller, Franklin G. 355 n.10 Miller, Karen–Lee 255 n.6 Miller, Kenneth 157
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Index 487 Miller–Young, Mireille 285 n.6, 287 n.11, 288 Mills, Charles 8, 167 n.5, 188, 195–6, 198–9, 203, 215 n.3, 223, 225, 245, 257 n.17, 265, 266 n.39, 277 n.57 Millstone, Erik 85 n.66, 89 n.84–5, 90 n.88 Mion, Giovanni 271 n.48 Moor, James H. 449 n.13 Moore, Stephen D. 80 n.39, 273–4 moral encroachment 5, 99, 100 n.3, 101, 102 n.8–9, 103, 105 n.18, 20, 106, 107 n.25, 111, 112 n.36, 113–15, 350, 368–9, 372, 383 moral outrage porn 14, 429, 432–3 moral panic 4, 37 n.1, 45, 50–1, 53, 58 moral permissibility 71, 325 n.13, 343–4, 378 Moran, P. 256 n.12 Morris, Aldon 202, 204, 206, 209–11 Morrison, Toni 259 Morton, Jennifer 101 Moss, Sarah 102 n.9, 107 n.25, 272, 273 n.52, 350 n.2 Moulton, Janice 228 Muhammad, Khalil Gibran 205 Mulder, Marianne R. 258 n.20 Myhr, Terri L. 255 n.6 Myrdal, Gunnar 193 Nader, Ralph 405 n.17 Nagel, Jennifer 261 n.29 Nagel, Mechthild 28 n.6, 261 n.29 Nahmias, Eddy 228–9 Narayan, Uma 215 n.2, 266 n.40 Nassar, Larry 6, 119 n.1, 132 n.25, 133, 138–9 National Institute of Allergy and Infectious Diseases 83 n.54 Nelson, Mark T. 101, 262 n.31–2, 317 Neta, Ram 58 New York Times 470 n.23 Nguyen, C. Thi 13–14, 41, 421 n.16, 427 n.28–9, 429 Nichols, Shaun 61, 62 n.13, 228 Nicksa, Sarah C. 260 n.26 Nissenbaum, Helen 449 n.13 No Child Left Behind Act 120 Nott, Josiah Clark 203 Nozick, Robert 340 n.35 Nuffield Council on Bioethics 67 n.1 Ocasio–Cortez, Alexandria 120 n.10 OED Online 449 Olson, Dustin 390 n.2 Olson, Loreen N. 123 n.15 online harassment 14, 439, 441 n.4, 444, 450, 453 Oreskes, Naomi 402 n.13 Orr, Deborah 228
Osborne, R. C. 113 Outlaw Jr, Lucius 215 n.3 Owens, David 355 n.10 Pace, Michael 350 n.2 Page, Scott 418 n.11 Paludi, Michelle A. 120 n.7 Pariser, Eli 398 Parks, Angela 256 n.9, 318 n.26 Parreñas Shimizu, Celine 285 n.6, 287 n.11, 288 Parrott, Matthew 330 n.21, 331–4 Pateman, Carole 195, 263 n.33, 314 n.18 Paul, Sarah 101 Paxton, Molly, 215 n.2, 216, 235 n.17 PCRM 81 n.47 Peels, Rik 141 n.38 Peet, Andrew 444 Penley, Constance 285 n.6, 287 n.11, 288 pernicious ignorance 138–9 Perry, Sharon 256 n.11 Peter, Jochen 284 n.5 Pew Research Center 399, 441 n.4 Pineau, Louis 382 n.33 Pippen, John J. 80 n.32, 37, 81 n.43, 84 n.64, 85 n.68, 86 n.72, 87 n.80 Pittenger, Samantha L. 260 n.24 Plantinga, Alvin 340 n.35 Plunkett, David 274 n.54 Pohlhaus Jr., Gaile 155, 157, 177–8, 194, 201, 217 n.5, 241, 244, 442, 476 n.30 political prisoners 3–4, 20, 25–6, 27 n.4, 28–31, 33 n.10 Pollock, John 75 n.17–8, 76, 340 n.35 Powell, Adam A. 465 n.13 Power, Helen 382 n.33–5 pragmatic encroachment 10, 42–3, 49, 54, 57–8, 104 n.17, 263 n.35, 264–6, 268, 336 n.30 predatory grooming 5, 6, 119–23, 136–7, 138 n.35, 139–44 Generalized Social–Psychological Model 122 groomer 124–5, 127, 129, 137 n.34, 139 groomee 122, 125 n.18, 126–7, 129, 131–2, 139 Preece, Rod 91 n.91 propaganda 10, 140–1, 143–4, 283 n.2, 284–6, 290 n.17, 2912, 292 n.19, 294 n.24, 295, 296 n.26, 297–9, 401, 402 n.12, 405, 468 punitive model of assertoric behavior 15, 463 quid pro quo 121 race 3, 8–9, 31, 51, 56, 151–2, 166, 172, 186–7, 193, 199, 202, 204, 206–10, 215 n.1, 3, 217, 219, 222–6, 234 n.16, 235 n.17, 236–7, 242, 256, 287, 305, 310, 315–17, 319
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
488 Index racialized bodies 7, 186, 188 Rackley, Erika 285 n.7 Rainie, Lee 399 Ramsey, F. 107 rape culture 9, 253 n.1, 254, 256–7, 259, 261, 264–6, 271–2, 274–5, 277 Rawls, John 195, 361 Record, Isaac 41, 125 n.19, 391, 398 n.8, 411, 437 Reed, Baron 105 n.18, 141 n.38, 336 n.30, 340 n.35–6 Reed, Dan 128 n.23 Regan, Tom 68 n.4, 83 n.56–8 Reider, Patrick J. 391 n.4 Rennison, Callie Marie 257 n.16 Resnik, Judith 32–3 Rich, Adrienne 266 n.41 Richards, Neil M. 442 n.6, 449 n.13 Ritter, Nancy 257 n.16 Roberts, Dorothy 318 n.22 Rodriguez, Dylan 31 Rogers, Teddy K. K. 123 n.15 Rogers, Todd 396 Roman, John K. 315 n.20 Rooney, Phyllis 215 n.3, 224 n.10 Rose, Hilary 197 Rose, Steven 197 Ross, Jacob 104 n.17 Ross, Sheryl Tuttle 291 n.19 Roth, Martha 256 n.8 Rothgerber, Hank 222 n.8 Roy, D. 402 n.13 Ruby, Matthew B. 222 n.8 Rudner, Richard 43 n.3, 104 n.16, 105, 112 Rury, John L. 208 Russell, Camisha 215 n.2, 240, 245 Russell, Diana E. H. 283 n.2 Ryder, Richard D. 68 n.5, 82 n.50 Saini, Angela 196 Saint–Croix, C. 106 n.22, 111 n.34, 215 n.2, 218 n.6 Salamon, Gayle 215 n.3 Sandusky, Jerry 6, 119, 126 Sauder, Michael 423 n.21, 424 n.23 Saul, Jennifer M. 114 n.39, 216, 277 n.57, 401 n.11 Schiffer, Stephen 276 n.55 Schnell, Frauke 120 n.6 Schouten, Gina 232 Schroeder, Mark 104 n.17, 108 n.26, 350 n.2 Schwabe, Rudolf 80 n.39 Schwitzgebel, Eric 235 n.17 Sebastiani, Silvia 203 Sedgwick, Eve Kosofsky 183–4
Self–Defense Law 315 self–trust 135–7 Seok, Junhee 88 n.83, 89 n.87, 90 Setiya, Kieran 141 n.37, 38, 390 n.2 sexual assault 9, 11, 138, 253 n.1, 254, 255 n.5, 256–65, 266 n.39, 267–70, 272, 274–5, 277, 307, 312, 321 n.1, 322 n.1, 374 n.26, 381, 443 actus reus 11, 205, 312, 382 mens rea 11, 305, 308, 312, 382 n.35 “Reasonable Expectation from State” standard 11, 306 “Reasonable Man” standard 11, 312 “Reasonable Woman” standard 11, 312–13 sexual coercion 11, 120, 306, 307 n.5, 308 n.7, 309, 314–15, 318–19 sexual consent 3, 11–12, 311, 321–2, 325–9, 332–4, 335 n.27, 336 n.30, 337 n.33, 338–9, 341–3, 344 n.45, 351, 368 n.19, 373, 380 n.31 attitudinal view 12, 334–5, 349, 353 epistemic agency 12–13, 321, 325–6, 329, 332–9, 343–4 performative view 335 n.27, 353–4, 355 n.9 Total Evidence View 12, 321, 339, 342, 344 withdrawal 328, 330 Shanks, Niall 74, 79 n.28, 80 n.29, 31, 83 n.59, 85 n.68, 86 n.75 Sharpe, Robert 79 n.28, 80 n.33, 81 n.40, 82, 83 n.53, 86 n.71, 89 n.84 Sharrer, Erica 284 n.5 Shelby, Tommie 199, 297–8 Siegel, Judith M 260 n.24 Siegel, Susanna 115 n.42 Simon, Judith 437 Sims, Sam 229 Sinan, A. 402 n.13 Singer, Peter 68 n.3, 79 n.28, 81 n.42, 45, 94 situated knowledge 37, 50–1, 54 situatedness 4, 37–8, 40, 57–8, 62 n.13 skepticism 101, 103 n.131, 174, 200, 209, 253 n.1, 258, 261–4, 266, 270, 274–7, 357–8, 405, 466 default skepticism 468, 475 Sloman, Steven A. 396 Smedley, Audrey 202 Smedley, Brian D. 202 Smith, Dorothy E. 44 Smith, Rogers M. 205 Smith, Sharon G. 256 n.12 Snow, David A. 21 Snowdon, Paul 224 n.9 Snyder, Mark 258 n.20 social media 15, 41, 389, 391, 396, 398, 399 n.10, 401–3, 411 n.3, 425, 429, 432, 438, 441 n.4, 446, 457, 464, 468 n.19, 470
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
Index 489 social transformation 3–4, 19–21, 24–6, 34, 35 n.12 sociology 8, 156, 194, 201, 204, 207, 209–11 Solomon, Miriam 45 Sorenson, Susan B. 260 n.24 Sosa, Ernest 340 n.36 Spaulding, Shannon 358, 359 n.12 Spelman, Elizabeth V. 56 n.10, 57 Spencer, Quayshawn 215 n.2, 216, 235 n.17 Spillers, Hortense J. 21 Spivak, Gayatri C. 442 Stahlmann, Ralf 80 n.39 Stalnaker, Robert 160, 167, 371 Standing, Craig 395 n.6 Standing, Susan 395 n.6 standpoint epistemology 44, 55–7 standpoint theory, see: standpoint epistemology 44, 48, 53–7, 266 Stanley, Jason 42, 58, 100 n.4, 104 n.17, 143, 253 n.1, 263 n.35, 276 n.55, 285, 290, 294 n.24, 294–8, 322 n.3, 323, 325, 366 n.16 Steel, Daniel 43 Stein, Judith A. 260 n.24 Steinberg, Stephen 202, 204, 207, 209, 211 Steinfeld, Charles 438 Steup, Matthias 74 n.15, 75, 366 n.15, 390 n.2 Stich, Stephen 61, 62 n.13, 228 Stiegel, Lori A. 120 n.9 Stine, Gail C. 101 n.7, 268 n.43 Stinnett, J. et al. 69 Stoick, Kristie 80 n.32, 37, 81 n.43, 85 n.68, 86 n.72 Strawson, P. 109–10, 224 n.9–10, 225 Stroud, S. 113 Sullivan, Shannon 277 n.57 Sullivan, Rebecca 287, 288 n.12, 292 Sun, Chyng 284 n.5 Sundell, Tim 274 n.54 Sundquist, Eric J. 193 Sunstein, Cass R. 399 n.10, 411 Superson, Anita 128 n.22, 215 n.3 Swain, Stacey 61 Talisse, Robert B. 403 n.14 Tanesini, Alessandra 396 Tang, Kwong–leung 257 n.15 Tanke, Elizabeth D. 258 n.20 Taormino, Tristan 285 n.6, 287 n.11, 288 Tarrant, Shira 288 n.12, 298 testimonial injustice 6–9, 13, 136, 137 n.33–4, 138 n.35, 139, 142, 151, 152 n.1, 153–6, 159, 161, 163–4, 166–8, 171, 176, 216–19, 221 n.7, 222, 225–6, 228, 231–3, 234 n.16, 235–7, 238 n.18, 239–41, 242 n.20, 243, 244 n.21, 245 n.21, 276 n.56, 397, 457, 476
content–based 218–19 degradation 7, 151, 153–4, 156–9, 165–8 derivatization 7, 154, 156 identity–based 218–19 objectification 7, 154, 155 n.3, 156 testimonial smothering 139, 142, 171, 179 n.7, 180, 183, 443, 476 testimonial quieting 139, 476 testimony 7–9, 12–13, 15, 58, 126, 128, 130 n.24, 132 n.25, 133 n.27, 134–5, 137–8, 142, 151–2, 154–8, 163–4, 166, 168, 171–8, 179 n.7, 180–4, 185 n.9, 186, 187 n.11, 188 n.12, 200, 208, 217–18, 254–5, 257–63, 265, 266 n.39, 269–71, 273–6, 288, 309–11, 316, 321, 326, 328–9, 331, 336 n.32, 337–42, 344, 358, 397, 404 n.16, 405 n.18, 406–7, 422 n.19, 441, 459, 462, 467, 468 n.20, 469, 473–5 test–operate–test 6, 125 n.19, 126, 129, 132 Thérèse, Sandrine 157 Thomas, George 375 n.27, 376 n.28, 377 Thompson, Morgan 216 n.4, 228–30 Thomson, J. J. 105–6 Tiberius, Valerie 235 n.17 Title IX 254 Tollefsen, Deborah 41 Tooley, Michael 48 Torrance, John 196 track–record evidence 74 n.15, 75–6, 92 n.92 Tremain, Shelley 215 n.3 trustworthiness 15, 182, 404–5, 448 n.12, 449–53, 463, 466 n.15, 468–9 Tsosie, Rebecca 194 Tucker, Joan S. 256 n.10 Tucker, Joshua A. 416 Turri, John 61, 62 n.13 Twitch.com 120 n.10 Twitter 13–14, 391, 399, 401, 402 n.13, 410 n.1, 411–12, 415–21, 422 n.18, 19, 423–6, 429, 431, 433, 437–8, 440–1, 442 n.5, 443, 445–6, 451, 468 n.19 Tyler, Kimberly A. 256 n.11 Tyson, Sarah 25, 29 Ullman, Sarah E. 260 n.24 undue influence 121 uptake 6–7, 127, 138–9, 142, 172, 175, 178–81, 183, 185, 188, 200, 208, 221, 259, 262, 292–3, 476 Valian, Virginia 223, 229 Van Bavel, Jay J. 416 van Brabandt, Petra 287 n.10 Van Cleve, James 404 n.16 Vann, Beth 256 n.10
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 15/03/21, SPi
490 Index virtue epistemology 14, 437, 446 n.10, 447, 448, 452 Vitalis, Robert 206 Volkwein–Caplan, Karin 120 n.6 von Eschenbach, Andrew C. 84 n.63 Vosoughi, S. 402 n.13 Waldron, Jeremy 158 Walker, Margaret 215 n.3, 227, 240 Wall Street Journal 399 n.9 Ward, Jeanne 256 n.10 Warren, H. Shaw 88 n.83, 89 n.87, 90 Waters, Anne 215 n.2 Weatherson, Brian 49–50, 253 n.1, 263 n.35 Wedgwood, Ralph 253 n.1 Weinberg, Jonathan M. 61, 62 n.13, 228 Weiner, Matthew 160, 466 n.15 Wenzel, Suzanne L. 256 n.10 Wertheimer, Alan 325 n.13, 340 n.37, 343 n.40, 355 n.9, 10, 357 n.11 Westen, Peter 326 n.15, 334 n.25, 353 n.6 Whitcomb, Dennis 218 n.6 White, Stephen 101, 109, 112 Whitman, James Q. 158 n.4 Whitt, Matt 172 n.2
Whyte, Kyle 215 n.2–3 Wilholt, Torsten 43 Williams, Michelle 256 n.9, 318 n.26 Williamson, Stephanie L. 256 n.10 Williamson, Timothy 159–61, 322 n.2, 371 Wills, Julian A. 416 Wilson, Sean 230 Wilson, Yolanda Y. 215 n.2, 226 Winfield, Idee 255 n.6 Winkel, Frans Willem 258 n.20 Winsberg, Eric 40, 43 Wortley, Natalie 305 n.1 Wosnitzer, Robert 284 n.5 Wray, K. Brad 41 Wyatt, Gail Elizabeth 260 n.24–5 Wylie, Alison 44, 49, 56 n.10, 57 Yancy, George 215 n.2–3, 226 Yikkity Yak 15 Yong, Ed 40 Yuan, Yuan 228 n.11 Zagzebski, Linda 446, 448, 466–7 Zheng, Robin 287 n.10, 294 n.23 Ziv, Amalia 286, 287 n.10