Organizations as Wrongdoers: From Ontology to Morality 0192870432, 9780192870438

Citation preview

Organizations as Wrongdoers

Organizations as Wrongdoers From Ontology to Morality STEPHANIE COLLINS

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 © Stephanie Collins 2023 The moral rights of the author have been asserted First Edition published in 2023 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: 2022945694 ISBN 978–0–19–287043–8 DOI: 10.1093/oso/9780192870438.001.0001 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.

Contents Acknowledgements

vii

PART I. METAPHYSICS 1. The Reality of Organizations 1.1 Overview 1.2 Organizations’ Attributes 1.3 Realism about Organizations 1.4 The Questions

3 3 8 14 24

2. Towards a Metaphysics of Membership 2.1 Introduction 2.2 Causation, Supervenience, Anchoring 2.3 Grounding and Parthood 2.4 Members as Parts-Plus: Existing Accounts 2.5 Conclusion

27 27 30 36 42 48

3. A Metaphysics of Membership 3.1 Introduction 3.2 Members: Parts with Commitments, Inputs, and Enactment 3.3 Objections and Replies 3.4 Members of Corporations 3.5 Members of Liberal-Democratic States 3.6 Conclusion

50 50 51 59 65 75 79

PART II. BLAMEWORTHINESS 4. Varieties of Organizational Blameworthiness 4.1 Introduction 4.2 Volitionism for Organizations 4.3 Attributivism for Organizations 4.4 Aretaicism for Organizations 4.5 Conclusion

83 83 86 92 97 105

vi



5. Organizations’ Moral Self-Awareness 5.1 Introduction 5.2 Robotic Blameworthiness 5.3 Why Moral Self-Awareness Matters 5.4 Plural and We-mode Self-Awareness 5.5 Organization-level Moral Self-Awareness 5.6 Conclusion

107 107 108 113 120 122 130

PART III. MEMBERS 6. How Members Are Implicated 6.1 Introduction 6.2 Enactors 6.3 Endorsers 6.4 Omitters 6.5 Conclusion

133 133 136 142 149 155

7. Apportioning Members’ Burdens 7.1 Introduction 7.2 Sources of Cost 7.3 The Source-Tracking Model 7.4 Responding to Shortfalls 7.5 Conclusion

158 158 161 163 171 183

Conclusions References Index

185 189 199

Acknowledgements The ideas in this book have been gestating for more than a decade and have followed me to five universities in three countries. I first started thinking about group agents in 2010 as a PhD student at the Australian National University, under the primary supervision of Bob Goodin and Nic Southwood. I took these thoughts to my first academic job at the University of Manchester, where in 2016 I wrote the article on which Chapter 7 is based and in 2017 wrote the article on which Section 6.3 is based. In 2019, I wrote the article on which Chapter 5 is based while I was a visiting research professor at the University of Vienna. The first full typescript was written in 2020–2021, while I was at the Australian Catholic University. Final revisions were done from Monash University in 2022. I thank each of these organizations—and, more importantly to me, their members—for their support and encouragement. Financial support was provided by the above universities, but also by external funding bodies. This research was funded by the European Union’s Horizon 2020 Research and Innovation programme under grant agreement No. 740922, ERC Advanced Grant ‘The Normative and Moral Foundations of Group Agency’. I subsequently worked on the book while holding an Australian Research Council Discovery Early Career Researcher Award for the project ‘Organisations’ Wrongdoing: From Metaphysics to Practice’ (project number DE200101413). For helpful comments on presentations of the ideas, I thank audiences at the following events: the University of Auckland Philosophy Seminar, a workshop on Acts and Omissions at the Australian Catholic University, the Australian Catholic University’s Dianoia Institute of Philosophy Workshop, a conference on Collectivity at the University of Bristol, the Social Ontology 2021 Conference at the University of California at San Diego, the Deakin University Philosophy Seminar, the Collective Intentionality X Conference at Delft University of Technology, the TU Eindhoven Philosophy and Ethics Seminar, a workshop on Collective Responsibility at University of Jyväskylä, the University of Leeds Ethics and Metaethics Seminar, the London School of Economics Political Philosophy Seminar, the Macquarie University Philosophy Seminar, a

viii



workshop on Collective and Shared Responsibility at the 2019 Mancept Workshops in Political Theory at the University of Manchester, the Monash University Philosophy Seminar, a workshop on Role Ethics at the Open University, the University of New South Wales Philosophy Seminar, the Otago University Philosophy Seminar, a workshop on States as Collective Agents at Princeton University, a workshop on Group Agency and Belief at St Andrews University, the Practical Philosophy Workshop at Saarland University, the University of Sydney Philosophy Seminar (twice), the Undergraduate Philosophy Journal of Australasia’s 2020 Online Conference, the Australasian Association of Philosophy’s 2021 Conference at the University of Waikato, the University of Wollongong Philosophy Seminar, and the 2020 Social Ontology Zoom Family Reunion. For comments on written drafts of individual chapters or the articles on which they’re based, I thank Nevin Climenhaga, Jonathan Farrell, Luara Ferracioli, Simon Goldstein, John Hawthorne, Frank Hindriks, Violetta Igneski, Ole Koksvik, Holly Lawford-Smith, Avia Pasternak, James Pattison, Richard Rowland, Liam Shields, Anna Stilz, David Strohmaier, and Hannah Tierney. I thank Herlinde Pauer-Studer for several deeply enriching conversations about Chapter 5. For comments on full drafts of the book, I thank two anonymous reviewers at Oxford University Press and a reading group at Monash University constituted by Will Cailes, Andy McKilliam, James Molesworth, Andris Raivars, and Simon van Baal. Chapter 5, Section 6.3, and Chapter 7 each closely follow the arguments of my previously published work. I thank the publishers for permission to reprint substantial tracts of the below articles: Collins, Stephanie. 2016. ‘Distributing States’ Duties’. Journal of Political Philosophy, 24/3: 344–66. Collins, Stephanie. 2017. ‘Filling Collective Duty Gaps’. Journal of Philosophy, 114/11: 573–91. Collins, Stephanie. 2022. ‘I, Volkswagen’. Philosophical Quarterly, 72/2: 283–304.

PART I

METAPHYSICS

1 The Reality of Organizations 1.1 Overview Organizations do moral wrong. States pursue unjust wars, businesses avoid tax, and charities misdirect funds. Our social, political, and legal responses require guidance. We need to know what the thing is we’re responding to and how we should respond to it. We need a metaphysical and moral theory of wrongful organizations. This book provides a new such theory, paying particular attention to questions that have been underexplored in existing debates. The underexplored questions include: what’s the metaphysical relation between organizations and their members, and how can this help us to make sense of organizations’ blameworthiness? Can organizations be blameworthy in all the myriad ways ‘blameworthy’ is understood in contemporary moral theory? And what about feelings of guilt, remorse, and shame—can organizations experience these feelings in a phenomenological sense (not just their ‘functional equivalents’)—and why does this matter for organizations’ blameworthiness? And if organizations are wrongdoers ‘in their own right,’ then what does this imply for members? How and why are members implicated in organizations’ wrongs, and how should reparative costs be apportioned among members? This book will answer these and other questions. In doing so, I aim to bridge a divide. On one side of the chasm, recent work in social ontology has aimed to account for the metaphysics of organizations. But this work has not considered how the metaphysics can address the distinctive moral issues raised by organizations’ wrongdoing. On the other side of the chasm, recent work has taken groups’ moral blameworthiness seriously, but without fully establishing the underlying metaphysics and exploring how that metaphysics can inform our moral conclusions.¹ In short: the metaphysical and

¹ For example, List and Pettit 2011 and Tollefsen 2015 address group responsibility with a focus on the underlying philosophy of mind; Isaacs 2011 with an emphasis on philosophy of action. Organizations as Wrongdoers: From Ontology to Morality. Stephanie Collins, Oxford University Press. © Stephanie Collins 2023. DOI: 10.1093/oso/9780192870438.003.0001

4

  

moral treatment of organizations have become detached in the literature. This book aims to attach them and to demonstrate why that attachment matters. The book’s primary metaphysical thesis is that members are material parts of organizations—much as the engine, wheels, and so on are the material parts of a motorcycle. I argue for this thesis in Part I (Chapters 1, 2, and 3). In Part II (Chapters 4 and 5), I demonstrate that this metaphysical thesis yields important moral conclusions: organizations can be blameworthy for a wide range of actions, attitudes, and character traits (including those of members qua members) and organizations can literally feel guilt (when members feel guilt qua members). In Part III (Chapters 6 and 7), I explore the implications of organizations’ wrongdoing for members. I catalogue the ways in which members-as-material-parts can be implicated in organizations’ wrongdoing. As we will see, not all members are equally implicated in each wrongdoing of each organization. There are some organizational wrongdoings in which only a few members are implicated, despite the fact that all members are material parts of the organization that has done wrong. I argue that the costs of organizations’ wrongdoing should be distributed to members in proportion to members’ level of implication in the wrongdoing, rather than being distributed equally on the basis of membership-as-parthood alone. In these ways, I aim to connect the metaphysics and morality of wrongdoing organizations. Of course, not all readers will be interested in this connection. Those who are interested only in the metaphysics might stop reading after Chapter 3. I hope that Chapters 2 and 3 will appeal to anyone interested in the metaphysics of ordinary objects—not only those interested in the metaphysics of social objects. Meanwhile, those who are interested only in the moral and political implications can skip Chapters 2 and 3, without much loss of understanding (though such readers will, I think, be interested in Sections 3.4 and 3.5). Such readers are asked simply to grant me the assumption that members are material parts of organizations. I hope that Part III (Chapters 6 and 7) will engage anyone with an interest in historical injustice and reparations, regardless of their interest in metaphysics. Still, the book’s arguments are best understood as forming a coherent whole: the metaphysical picture is an important premise in the moral arguments, while the moral arguments explain why the metaphysical picture matters. To settle ideas and begin to demonstrate the practical importance of this inquiry, let me lay out two examples.

   

5

1.1.1 Australian Banking Royal Commission In 2019, Australia received the final report of its Royal Commission into Misconduct in the Banking, Superannuation and Financial Services Industry. In Australia, Royal Commissions engage in inquiry and factfinding. They’re less formal than a court, but they have the power to refer entities to courts for prosecution. Perhaps most importantly, their policy recommendations are taken seriously by government and the media. The Banking Royal Commission (as I’ll call it) was set up to inquire whether there was ‘misconduct, or conduct falling short of community standards and expectations, in Australian financial services entities’. If so, the Commission was asked to report on whether that conduct was ‘attributable to the particular culture and governance practices of a financial services entity or broader cultural or governance practices in the relevant industry or relevant subsector’ and whether the conduct ‘result[ed] from other practices, including risk management, recruitment and remuneration practices of a financial services entity, or in the relevant industry or relevant subsector’ (2018, 2). The Commission revealed that Australia’s biggest banks had engaged in widespread and pervasive dishonest practices. These included charging fees when no service had been provided (including to deceased customers), lying to customers, forging customers’ signatures, impersonating customers, falsely witnessing documents, transferring customers’ funds to advisors’ personal bank accounts, underpaying interest on term deposits, and on and on. The Commission’s report attacked the ‘culture’ of ‘dishonesty and greed’ in Australia’s largest banking and finance corporations (Royal Commission 2018, 73; 2019, 138). The Commission’s final report asserted that the corporations themselves were culpable for the misconduct the Commission uncovered (2019, 4). Yet the Commission didn’t explicitly endorse the idea that the banks were blameworthy in a manner ‘irreducible’ to the blameworthiness of the individuals involved. Instead, the Commission simply raised the question: ‘[s]hould there be more focus on criminal proceedings against [banks and financial institutions] rather than individual advisors?’ (Royal Commission 2018, 158). It left answering this question to the politicians and bureaucrats who were the primary audience of its report. In this, the Commission arguably submitted to a long tradition in Australia of viewing corporations as ‘legal fictions’, with individuals (such as financial advisors) being the ones

6

  

who are chased by the legal system. The Commission left it up to regulators whether this tradition should be overthrown. You might think the Commission’s question is a legal one, not a philosophical one. But Australian law gives a philosophically contentious answer. Australia’s Criminal Code ‘applies to bodies corporate in the same way as it applies to individuals’—yet it allows for prosecution of a corporation only when ‘an offence is committed by an employee, agent or officer . . . within the actual or apparent scope of his or her employment, or . . . authority’ (Criminal Code Act 1995 (Cth), Part 2.5, Sec. 12.1, emphasis added). Crimes can also be attributed to a ‘corporate culture’, understood as ‘an attitude, policy, rule, course of conduct or practice existing within the body corporate generally or in the part of the body corporate in which the relevant activities [of employees, agents, or officers] take place’ (Criminal Code Act 1995 (Cth), Part 2.5, Sec. 12.4). But even within this ‘corporate culture’ provision, identifiable wrongful activities of individuals are what’s attributed to the corporate culture. And the corporate culture provision has rarely been used, so there is little clarity over how to interpret and apply it (Hill 2003; Adams et al. 2017). The upshot: legally, in Australia at least, corporate criminal wrongdoings generally exist only when individual criminal wrongdoings exist. In recent years, several philosophers have provided pictures of collective wrongdoing that allow us to argue that this legal orthodoxy is wrong-headed (e.g., French 1984; Pettit 2007; Isaacs 2011; Tollefsen 2015). However, as I aim to show, our current philosophical picture of corporate wrongdoing is incomplete. Therefore, our potential to provide a full and defensible answer to the Commission’s question is also incomplete. This book will fill the gaps, providing the theoretical tools for us to answer the Commission’s question of whether it is desirable—or even defensible, or even possible—to hold organizations themselves culpable. What’s true of legal wrongdoing is also true of legal metaphysics: the picture that prevails is an individualist picture. The above legal approach to corporate criminal wrongdoing reflects the fact that much of Australia’s corporate law seems to employ a shareholder-based metaphysics of corporations, in which the corporation is identified with its shareholders (Hill 2003; this view has recently received philosophical defence in Ludwig 2017a)—or, only slightly less reductively, with a ‘nexus of contracts’ that includes employees, suppliers, customers, and so on (Jensen and Meckling 1976). The nexus-of-contracts model arguably encourages the reduction of the corporation and its wrongdoing to individuals and their wrongdoing: on this

   

7

model, if we can find out what those in the nexus have done, then we will have found out what the organization has done. But organizational wrongdoing is both narrower and broader than this. It’s narrower insofar as many in the contractual nexus—such as suppliers and customers—should not have their actions attributable to the corporation. And it’s broader insofar as wrongs sometimes seem attributable to corporations’ cultures (as the Banking Royal Commission claimed). It’s unclear where to locate a ‘culture’ within a nexus-of-contracts metaphysics. All of this suggests that philosophy can help to answer the Commission’s question—in both its moral and metaphysical dimensions.

1.1.2 The Polluter Pays Principle A second example is the role of the Polluter Pays Principle in distributing the costs of mitigating the devastating effects of climate change—and, in particular, in implementing the 2015 Paris Agreement under the United Nations Framework Convention on Climate Change. The Polluter Pays Principle asserts that industrialized states should set high emissions reduction targets under the Paris Agreement, because these states are culpable for wrongfully high historical emissions. Disagreement over the truth of the Polluter Pays Principle (among many other things) has prevented states from setting sufficiently ambitious emissions reduction targets under the Paris Agreement. Predictably, developing states argue in favour of the principle while developed state resist it (Stavins 2018). If we are to vindicate the Polluter Pays Principle, we need a conception of states as unitary actors, who nonetheless relate to their citizens in such a way that (i) citizens’ polluting actions are attributable to states and (ii) costs attributed to polluting states can justly be passed on to a wide diversity citizens—including passed on to citizens who are not themselves wrongful polluters. Parts I and II of this book vindicate point (i), while Part III vindicates point (ii). As with Australia’s Banking Royal Commission, the law gets us only so far in sorting out these issues. Brownlie’s Principles of Public International Law (Crawford 2012) is the canon on states’ legal treatment. Here, states are metaphysically conceptualized as unitary actors rather than a ‘nexus of contracts’ or a collection of individuals. But in the place of the nexus of contracts, there’s an empty hole: states’ relation to individuals is left entirely untheorized. It’s as if states are completely separate from their individual

8

  

constituents—whether these constituents are conceptualized so as to include ordinary citizens, or not. Part I of this book aims to fill the hole. And needless to say, the force of international law has not been used against states who are high emitters. None of them have been condemned by the United Nations Security Council or referred to the International Court of Justice, for example. Philosophy can help us to work out whether that would be morally appropriate. Although Chapters 2 and 3 will argue that citizens are the material parts of their states, Chapter 4 will demonstrate that many of the moral failings of organizations—including states—cannot be located in the wrongful actions of individuals. Organizations’ moral failings often lie in organizations’ evaluative attitudes or character traits, where these organizational attitudes and traits are not straightforwardly understandable in terms of the wrongful evaluative attitudes or wrongful character traits of members. This analysis reveals the shortcomings of existing treatments of organizations’ wrongdoing, which have tended to focus on organizations’ blameworthiness for their irreducibly group-level choices or decisions. If we can hold organizations blameworthy for their evaluative attitudes and their character traits (not just their choices or decisions), we potentially get many more environmentally negligent organizations on the hook. These two examples will recur throughout the book’s discussion, helping to illustrate and problematize various claims that I’ll consider and defend. But this is a book of philosophy. My arguments will remain theoretical and general, with the details sketched primarily via these two examples. To introduce the book’s philosophical approach, the rest of this introductory chapter turns to theoretical issues. I will first outline how I conceptualize organizations. I’ll then briefly explain why we should be realists about them. Finally, I will give an overview of the arguments to come in the rest of the book.

1.2 Organizations’ Attributes In outlining the Banking Royal Commission and the Polluter Pays Principle, I’ve made an assumption that may surprise some readers. I’ve assumed states are organizations. Often, when we hear ‘organization’, we think of large corporations and charities, but not states.² My usage is broader. As I’ll ² Though I’m certainly not the first to argue that states are collective moral agents (see e.g., Erskine 2001; Erskine 2003; Wendt 2004).

   

9

use the term, an ‘organization’ is a collective agent that involves a large number of people who realize a structure that coordinates divided labour via rules and hierarchical command relations, guided by a collective decisionmaking procedure. This definition is intended to be ecumenical between many different characterizations found across the academic literature. It follows a long tradition in sociology, starting with Max Weber’s characterization of bureaucracies as involving rules and regulations, a division of labour and responsibility, and hierarchical authority structures (1968, vol. I, 223ff.; 1968, vol. III, 956ff.). More recently, sociologists have given similar analyses of institutions, such as Rom Harre’s view of an institution as ‘an interlocking double-structure of persons-as-role-holders or office-bearers and the like, and of social practices involving both expressive and practical aims and outcomes’ (1979, 98) or Jonathan Turner’s definition of an institution as ‘a complex of positions, roles, norms and values lodged in particular types of social structures and organizing relatively stable patterns of human activity’ (1997, 6). My use of ‘organization’ rather than ‘bureaucracy’ or ‘institution’ reflects recent philosophical usage, in which ‘bureaucracy’ has fallen out of the lexicon, while ‘institution’ is often used to capture a wider range of phenomena—including institutions such as marriage or promise-keeping, which are not ‘organizations’ in any sense. My characterization of organizations also roughly matches recent management theory, particularly Geoffrey M. Hodgson’s distinction between social structures, institutions, and organizations. According to Hodgson, ‘[s]ocial structures include all sets of social relations, including the episodic and those without rules, as well as social institutions. Institutions are systems of established and embedded social rules that structure social interactions. Organizations are special institutions that involve (a) criteria to establish their boundaries and to distinguish their members from nonmembers, (b) principles of sovereignty concerning who is in charge and (c) chains of command delineating responsibilities within the organization’ (Hodgson 2007). In addition to sociology and management theory, my characterization resonates with related work in philosophy. Thus my ‘organizations’ include Peter French’s ‘conglomerates’, which have (i) internal organization or decision procedures for choosing courses of action, (ii) enforceable codes of conduct for members, and (iii) members who replaceably occupy roles that provide them with power over other members (French 1984, 13ff.). My view likewise dovetails with Raimo Tuomela’s view of a group agent as ‘an

10

  

interactive social system . . . that consists of interrelated individuals such that this system is, through them, capable of producing uniform actions and outcomes’ (2013, 21). Because my definition encompasses the above definitions, I hope that my metaphysical and moral claims will appeal to a range of different theorists, who might subscribe to different of the above-cited definitions and characterizations of organizations. That is to say, my aim here is not to improve on any of the definitions quoted above from other theorists. Rather, I aim for my definition to be consistent with—and neutral between—all the above definitions. That said, my definition raises a question: what makes organizations (as I’ve characterized them) worthy of a unitary analysis as wrongdoers, despite the vast differences between various types of organization—states, forprofits, and non-profits, for example? After all, many philosophers differentiate organizations according to those organizations’ ends (e.g., French 1984; Rovane 1998; Miller 2010). One might think that organizations with different ends should be conceptualized differently as wrongdoers. On the contrary, the various components of my definition demonstrate that organizations warrant treatment as a class. First, consider organizations’ large size. By ‘large’, I mean ‘large enough that not all members are known to one another personally’. The numerical cut-off here will differ for different organizations, depending on how channels of communication are set up. Definitionally, all organizations are large enough that members cannot have pairwise in-depth conversations about the organization’s plans and policies (at least, not consistently with having the time to ensure those plans and policies are enacted). This is significant, because it means that it is extremely unlikely for there to be perfect uniformity amongst members’ reasons for supporting or opposing particular plans or policies, or even their interpretations of the content of plans and policies. The members cannot gather around a table to reach a consensus view on the whys and hows of organizational practices and policies. This mandates heterogeneity in how we treat members of wrongdoing organizations, as Part III of this book will explain. Such heterogeneity is common to organizations, while not always being true of smaller, more collaborative, collective agents. The lack of cohesion entailed by organizations’ large size also means that organizations’ members cannot simply be picked out by the fact that they all share identical goals, visions, or values. A more complicated account of membership is required, as Chapters 2 and 3 will explain. Second, the members occupy places in a structure. A structure is a collection of roles that stand in relations. In an organization, the roles are

   

11

jobs or tasks and the relations are usually reporting and delegation lines. The roles and their relations can be highly diverse. Both the roles and the relations are representable in an organization chart (that depicts the roles as ‘nodes’ and the relations as ‘edges’). Organizations are composed of people who ‘realize a structure’—that is, occupy the roles in the structure. A structure is an abstract representation of various roles and of the relations between those roles. As an analogy, we can think of a recipe. A recipe is an abstract representation of various roles (the onion role, the garlic role, and so on) and of how the occupants of those roles must be related to one another in order to realize a certain dish. For the dish to be realized, we must find particular role-occupants (particular onions, particular bulbs of garlic, and so on) and we must make it the case that those role-occupants are related in such a way as to follow the recipe.³ Following Katherine Ritchie, saying that an organization is composed of people who realize a structure allows us to say that an organization is a realized structure, where a structure is ‘realized’ when enough people occupy the structure’s roles. (How many is ‘enough’? That will be dictated by the organization’s decision-making procedure—something I will characterize shortly.) This idea of structure gives us an explanation of how an organization is both ‘one’ and ‘many’. It is ‘one’ because it is a realized structure—a single thing. It is ‘many’ because it is the many individuals who occupy its roles.⁴ Chapters 2 and 3 will say much more about the metaphysics here. Specifically, I will argue that members are physical parts of an organization, and that an organization is located wherever its members are. This will incorporate organizations into a naturalist, materialist view of the world. For now, the important point is this: since all organizations are realized structures, the just-mentioned metaphysical views can apply in the same way to different types of organizations (states, for-profits, non-profits, and so on). This vindicates the project of developing a unified theory of organizations as wrongdoers.

³ The recipe analogy comes from Kathrin Koslicki’s (2008) discussion of ordinary objects. ⁴ See Ritchie (2013). Harris (2020, 360–1) objects to Ritchie’s particular version of the organizations-as-structure view: structures do not allow for changes in members, roles, or relations; clearly, organizations can change members, roles, and relations while remaining the same organization. We can respond to this by saying that two realized structures, which exist at different times, are the same organization just in case the later structure is descended in the right way from the earlier structure.

12

  

Third, an organization has a division of labour. Members don’t know all the details of what all the others are doing vis-à-vis the organization’s plans and policies. This complicates the extent to which each can be held blameworthy for what the others (fail to) do. Diminished individual blameworthiness has two implications. First, it generates an imperative that we have a sound theory of distinctively organization-level wrong, since often we won’t find individual-level wrongs that are proportionate to an organization-level wrong. Part II of this book (Chapters 4 and 5) theorizes distinctively organization-level wrongs, including distinctively organizationlevel guilt. Second, diminished individual blameworthiness implies that we need to take care when spelling out the implications of organizations’ wrongdoings for members, since those implications will not be as uniform as they would be in a group with less informational asymmetry between members. Those implications for members are spelled out in Part III (Chapters 6 and 7). Fourth, an organization has rules and hierarchical command relations via which it pursues its plans and policies. These matter because—similar to the informational asymmetries that result from the division of labour—they sometimes attenuate individuals’ blameworthiness. If an individual has perfectly good reason to trust the system, then she may be perfectly within reason to obey the rules and/or her superiors. Yet if the rules themselves, or the command relations, are morally wrong, then this can lead to the organization doing wrong. This phenomenon is common across a wide range of organizations, even if the precise details vary. Worse, because the rules and command relations are themselves designed through a process with a division of labour (including through legislation and regulation that is imposed upon organizations from the outside), there may be no individual who is blameworthy for the fact that the rules and command relations are as they are. This relates to Anthony Downs’ ‘Law of Imperfect Control’, according to which no one individual can fully control a large organization—as well as his ‘Law of Diminishing Control’, according to which the more effort at control is made at the top, the more subordinates will try to evade or counteract such control (Downs 1967, 143–50). Thus, rules and hierarchical command relationships can relieve both superiors and subordinates from blame. Of course, this is not to say individuals within organizations are never blameworthy: Chapter 6 will spell out many ways in which individuals can be implicated in organizations’ wrongdoing, and Chapter 7 will justify individuals bearing costs to repair their organizations’ wrong. Again, though, these phenomena are common across a wide range of

   

13

organizations, vindicating the project of developing a unified theory of organizational wrongdoing. Thus, organizations are distinctive by virtue of being composed of a large number of people who are occupy places in a structure that coordinates divided labour via rules and hierarchical command relations. This makes them worthy of a unified analysis. And there are good reasons to want a unified analysis. By aiming at unity, our theory is forced to fit with a wider range of case studies and examples. All else being equal, this makes our theory better. That is: if we developed a theory of wrongdoing for (say) corporations but not states, then we’d need to develop a separate account of states as wrongdoers. Our over-arching theory would therefore be less unified and less simple. If we can develop a theory that applies to all organizations in a unified way (as I suggest we can), then this is preferable to going back to the drawing board for each and every different type of organization. At the same time, though, it’s worth theorizing about organizations specifically, rather than about collective agents more generally. This is because organizations raise their own problems, as just summarized. While the theory developed in this book might apply to collective agents more generally, I want to make sure we get organizations right. What are ‘collective agents more generally’? I said earlier that organizations are collective agents that are guided by a collective decision-making procedure. In my view, all collective agents are guided by a collective decision-making procedure (Collins 2019a, ch. 6). Specifically, a collective agent is composed of agents united under a rationally operated group-level decision-making procedure that can attend to moral considerations. Thus, each collective agent has its own decision-making procedure. Importantly for organizations specifically, a collective’s decision-making procedure includes the informal, tacit, and vague procedures that we might subsume under the notion of an organization’s ‘culture’ (Weick and Sutcliffe 2001, ch. 3; Herzog 2018). Why include ‘culture’ within a collective agent’s ‘procedure’? Because the norms, practices, and traditions that constitute a culture often hugely constrain the real-world instantiation of any abstractly defined ‘formal procedure’. Think, for instance, of a committee whose formal procedure is ‘egalitarian discussion amongst committee members’. When this formal procedure is actually instantiated, culture will inevitably have a strong influence. For example, tradition and norms might informally decree that older male members speak first, or for longer, or with more authority. Such cultural aspects of an actually-instantiated procedure will

14

  

often be crucial for understanding the procedure’s operation and outputs. Chapter 3 will say more about organizational membership and how members relate to an organization’s decision-making procedure. My account of collective agency produces a broad church. Two friends deciding where to go for lunch are a collective agent, just as much as BP, France, or Oxfam. Thus, many collective agents are far more unstructured, undivided, discretionary, small, and egalitarian than organizations. For the purposes of analysing collective agents’ wrongdoing, it is useful to study organizations specifically, rather than the broad church of collective agents more generally (as was done by, for example, Isaacs 2011; List and Pettit 2011; Tollefsen 2015). This is because of the factors mentioned earlier: organizations are (i) composed of many people, (ii) who realize a structure, (iii) with a division of labour, (iv) governed by rules and hierarchical command relations. These factors produce the complications sketched above, which will be discussed as the book proceeds.

1.3 Realism about Organizations So far, we have two examples where organizations’ wrongdoing is at stake but contestable. And we have a conception of organizations that renders them a distinct category. There is one more preliminary to establish in this chapter: realism about organizations. This is the view that organizations ‘are entities over which we quantify in the set of our best descriptions and explanations of the social world’ (Sheehy 2006, 132). As with any philosophical position, realism about organizations is contested—not just in philosophy, but also in organization theory, business ethics, sociology, and economics.⁵ It is not hard to see why it is contested. We don’t seem to bump into organizations in the street. We might bump into their buildings or billboards, but these things are buildings or billboards—they are not, themselves, organizations. It is not clear where organizations are. (I will defend a view of where they are in Chapters 2 and 3.) And if it’s not clear where they are, then it’s not clear how they could possibly be part of the causal pathways of the natural world. Those causal pathways require a spatiotemporal location, or so it seems. Furthermore, when we look out into the ⁵ Recent sceptics include Heugens, Kaptein, and van Oosterhout 2008, Rönnegard 2015, and Ludwig 2017a—tracing back to Popper 1945, Hayek 1948, Weber 1968, Weick 1979, and Elster 1989.

   

15

social and political world, what we do see is individual humans. Our society and politics seems to be made up of humans. While we might sometimes refer to groups of humans, it is tempting to view this as ‘mere shorthand’ for humans themselves. Despite these criticisms, realism is widely enough accepted—and the reasons for it have been widely enough circulated—that my defence here will be brief, schematic, and familiar to many readers.⁶ This truncated defence will enable the remaining chapters to forge new ground. There are three broad reasons for realism about organizations: their multiple realizability, their inward and outward causal-explanatory power, and (what I’ll call) their functionalist-interpretivist agency.

1.3.1 Multiple Realizability There are multiple ways organization-level facts, properties, events, and explanations—as well as organizations themselves, as objects—can be realized by individual-level facts, properties, events, explanations, and objects. (The reader might doubt that organizations are objects; I will render this plausible in Chapters 2 and 3.) Consider an event: the event of Australia permitting a large new coalmine to be built. We could instead have considered the fact that Australia has permitted the mine, or Australia’s property of having permitted the mine, or an explanation of the mine that includes Australia’s having permitted it, or the entity that is Australia—all of these are ‘multiply realizable’ in relevant individual-level phenomena. By ‘multiply realizable’, I just mean that individual-level phenomena could have been a variety of different ways, consistent with the organization-level phenomenon remaining exactly the same. I focus on phenomena that are events, just to settle ideas. (For discussion of the differences between the

⁶ See, e.g., Goodpaster and Mathews 1982, Ruben 1985, Pettit 1996, Hatch 1997, Reed 2001, Fairclough 2005, Hess 2013, and Thomasson 2019. In her influential overview of organization theory, Hatch (1997) divides organization theory into ‘modern’, ‘symbolic’, and ‘postmodern’ perspectives. In the sense I’m using here, all three perspectives are ‘realist’. The modern perspective analyses organizations as agents of rational control; the symbolic perspective views organizations as sites of cultural meaning and symbolism; and the postmodern perspective views organizations as ideologically driven forces of hegemonic power. My arguments can be wielded from any of these perspectives. More generally, the metaphysical and moral issues addressed in this book cross-cut debates in organization theory about organizations’ interests, values, and meanings.

16

  

multiple realizability of social-level properties, facts, events, and so on, see List and Spiekermann 2013.) The individual-level events that constitute the Australia-level event most prominently include actions on the part of state bureaucrats. And each of these individual-level events could have gone various ways: individuals could work more or less reluctantly, callously, carefully, obediently, etc. Each of these possible realizations is ‘Australia permitting the mine,’ just so long as two conditions are met. First, in each realization, the individual actions are performed by members within the constraints of, and because of, their role in Australia; or, as I will equivalently say, the actions are performed by members while they are performing their role. This is what makes the event one of Australia permitting the mine, rather than one of some private individuals abusing Australia’s procedures to produce a permission for the mine. Second, the permission of the mine results from each realization. This is what makes the event one of Australia permitting the mine, rather than merely trying to permit the mine. If these two conditions are met, then each realization is a realization of Australia’s permitting the mine. The realizations are different from each other, since they each involve individuals performing different actions. Thus, the realizations are not identical to each other. It’s also not true that all the realizations are identical to the permitting of the mine—if they were, and if we accept transitivity of identity, then the realizations would also have to be identical to one another (which we’ve already ruled out, because they are discernible from one another). We also cannot arbitrarily choose just one realization, and say that the permitting of the mine is identical to that realization but not any of the others: if we did that, then the permitting of the mine would not occur if one of the other realizations happened. So, we should say that the permitting of the mine is something distinct from, but that unites, the distinct realizations (Jackson and Pettit 1992; List and Menzies 2009). This is just to say that we should be realists about the organization-level event. Importantly, if the multiple realizability of some organization-level phenomenon is to render it distinct from its individual-level realizers, then the different potential individual-level realizers must be of different types, not just different tokens of the same type (Shapiro 2000; Couch 2004). To see this, consider the analogous case in philosophy of mind. Suppose my pain is always realized by C-fibres firing in my brain, though different C-fibres fire on different pain-occasions (because C-fibres degenerate and are

   

17

replaced by new ones). Then, pain is equivalent to C-fibres firing. For pain to be non-equivalent to (i.e., distinct from) its neural realizers, there must be different types of potential neural realizers. This holds in our example: Australia’s act of permitting the mine could have been secured by legislation, or by a judicial ruling, or by the run-of-the-mill implementation of some existing policy, or in some other way. These are all different types of realizers of the organization-level event. Other philosophers have similarly used multiple realizability to argue for organizations’ existence, but those arguments have tended to focus on how an organization manifests itself across time (e.g., List and Pettit 2011; List and Spiekermann 2013). Such cross-temporal arguments says that an organization persists through time even while its members alter. For example, an organization persists even though some members leave while other members join. Such persistence across changes in membership provides reason to view the organization as a metaphysically distinct entity. But such crosstemporal change is not the only relevant kind of multiple realizability. Instead of tracking the organization across time, we can track the organization across different possible worlds: even if members never change in the actual world, we should be realists about the organization if it’s true that the organization would persist were members to change. The counterfactual (rather than cross-temporal) version of multiple realizability is worthwhile for a few reasons. First, it enables realism about a broader pool of organizations than the cross-temporal version of the argument. Second (and more importantly), it enables us to assess just how robust the organizational entity is. We can ask across how many changes in membership (or across what kinds of changes in membership) the organization would remain. Of course, any survival across changes in membership gives us reason to posit the organization. But a more robust entity—an entity that would survive a higher or more diverse range of changes to its membership—is likely to have more explanatory power, and so more reason to be posited. This brings us to the second reason for realism about organizations.

1.3.2 Inward and Outward Causal-Explanatory Power Above, I mentioned that one might question organizations’ explanatory power, because one might question where they are placed in the natural world. I will address that issue at length in Chapters 2 and 3. Bracketing that issue for now, we can see that organizations have causal-explanatory power

18

  

in two directions. They have inward causal-explanatory power insofar as organizational facts, properties, events, and entities can causally explain facts, properties, events, and entities internal to the organization. And organizations have outward causal-explanatory power insofar as organizational facts, properties, events, and entities can causally explain facts, properties, events, and entities external to the organization. Inward causalexplanatory power implies that (an event, fact, property, etc. that involves) the organization sometimes causes (an event, fact, property, etc. that involves) its components being certain ways. Outward causal-explanatory power occurs when (an event, fact, property, etc. that involves) an organization causally explains (an event, fact, property, etc. that involves) phenomena external to the organizations. Outward causal-explanatory power demonstrates that organization-level explanations are different from an aggregation of explanations that are about the organization’s components. To intuitively see inward causal-explanatory power, consider again the case of Australia permitting a new coal mine. To adequately explain or describe the members’ actions that constitute the organization’s action—to capture members’ motivations, constraints, goals, etc.—we need to refer to the organization’s distinct features (e.g., judgements, goals, intentions, procedures). We can imagine a bureaucrat who is entirely opposed to permitting the mine—who even secretly campaigned against it on weekends—yet who does her role in implementing Australia’s decision. To fully explain and motivate this individual’s actions, we need to refer not just to facts about her, but also to facts about Australia—specifically, that Australia decided to permit the mine. Her action is, in a sense, not her own (at least, it’s not just her own). The locus of agency behind her action is (also) Australia’s. This is the same sense in which the wrongdoings uncovered by the Banking Royal Commissions were problems with the banks, not just with individuals within the banks, and provides a rationale for why the Polluter Pays Principle targets historically emitting states, not just collectives or individuals governed by those states. (For similar points on inward explanatory power, see Elder-Vass 2007, 32–4.) To intuitively see outward causal-explanatory power, consider that organizations can cause effects their members could not cause, were those members to lack the properties in virtue of which they realize the organization’s structure. The coal mine—situated on public land in Australia—could not be permitted by one person acting alone (not even the Prime Minister, without violating due process and therefore not acting as Prime Minister). The coal mine couldn’t even be permitted by many Australian people

   

19

together, if they were not acting within the structure governed by the organization-level decision-making procedure of Australia. Of course, the mine might end up on the land by sheer fluke, without any action by the Australian state, but then we would not describe the mine as having been permitted by anyone. The ‘permitting’ in that kind of case is not an intentional action, but rather an accident or omission. By contrast, if the mine is permitted by an edict produced by Australia’s organization-level decisionmaking procedure, then the permission is reliably and intentionally controlled; therefore, the permission is an action. Yet no individual can successfully produce the permission on their own: the action must be attributed to the state. To these intuitive demonstrations, we can add that some of the leading theories of causal explanation are amenable to including organizations in causal explanations. This is most obviously true on counterfactual theories of causal explanation (see, e.g., List and Menzies 2009; List and Spiekermann 2013). According to these theories, an outcome is explained by the facts that hold across all (or most) of the possible counterfactual worlds where that outcome obtains. Consider Australia’s permitting of the coal mine. This outcome doesn’t co-vary with the specifics of particular individuals: the individuals are more or less replaceable when it comes to securing the outcome. Instead, it co-varies with organization-level facts (events, properties, entities) like ‘Australia’s belief that coal creates jobs and boosts GDP’. Such organization-level phenomena also feature in causal explanations that use a probabilistic theory (on which causes make effects more likely) or a robustness theory (on which causes make effects more secure). (See similarly List and Spiekermann 2013. I say more about social-level phenomena as causes in Collins 2019b; Collins 2019a, ch. 3.) Organizations can also be part of causal explanations if we use a process theory of causal explanation. On these theories, the causes of some effect are the internal elements of the process that pre-empted the event—where that process might be a flow of energy, or might be the transmission of a modification in an otherwise consistent structure, or might be the conservation of a quantity (like mass-energy, linear momentum, or charge) within a given object over time (Schaffer 2016). For organizations to be potential causes on the process view, organizations would have to be concrete particulars: things through which energy can flow, modifications can be transmitted, or mass-energy can be conserved. In Chapters 2 and 3, I will develop and defend a view on which organizations are concrete particulars that have members as material parts. Lines can be drawn around them via their

20

  

structures, as outlined in Section 1.2. They are thus amenable to analysis as causes, from within the process theory of causes. Naturally, both the counterfactual and process theories of causation are open to objections. But they demonstrate the breadth of causal-explanatory theories that accommodate organization-level phenomena as causes.

1.3.3 Functionalist Interpretivism A final reason for realism about organizations is that they are agents. I’ll assume that if we should be realists about anything, we should be realists about agents. I have argued elsewhere that collectives are agents—where ‘collectives’ are to be understood as entities composed of agents united under a group-level decision-making procedure that can process moral reasons (Collins 2019a, ch. 6; see similarly List and Pettit 2011; Isaacs 2011; Hess 2018b; Hindriks 2018). As already explained, organizations are a type of collective agent: they are large, structured, labour-divided, governed, hierarchical collective agents. So, if my arguments elsewhere are correct, then it follows that organizations are agents. It is worth briefly recapitulating the reasons for believing in organizations’ agency, for two reasons. First, many working on collective agency (including my earlier self) do not sufficiently distinguish between functionalist and interpretivist reasons for believing in collectives’ agency. This distinction is important, because philosophers of mind (and others) tend to fall on different sides of the functionalist-interpretivist divide. If organizations were agents in an interpretivist sense but not a functionalist sense, then any argument for collective agency that relied on interpretivism would alienate those who have independent reasons for believing in functionalism—and vice versa, if organizations were functionalist agents but not interpretivist ones. Thus, my aim here is to show that both accounts of agency render organizations agents. The second reason for recapitulating reasons for believing in collectives’ agency is that organizations specifically are particularly clear cases of collective agents. Thus, even if one doubts that collectives in general are agents (for example, if one was unconvinced by my or others’ earlier arguments), then one may still be convincible that organizations specifically are agents. That’s what I aim to do here. What are the functionalist and interpretivist accounts of agency? I will start with functionalism. Within social ontology, functionalism about agency has been defended by Bryce Huebner (2014), Gunnar Björnsson

   

21

and Kendy Hess (2016), and David Strohmaier (2020). It’s a popular account of mental states in general—and attitudes towards propositions in particular—throughout philosophy of mind. The characteristic feature of functionalism is that propositional attitudes are characterized by their functional profile. That is: attitudes such as belief, desire, hope, regret, and so on, when taken towards propositions, are characterized by the function that those attitudes play within some system. This system is the agent or entity that bears the attitude. The function of an attitude includes causing, grounding, or blocking other attitudes and behavioural dispositions. For example, we attribute to me the attitude of belief towards the proposition ‘the ice cream is in the freezer’, I have an attitude that causes me to perform the behaviour of going to the freezer when I want ice cream. In this way, propositional attitudes are characterized by their relations to one another, and to behaviour. By contrast, interpretivism has been defended within social ontology by Christian List and Philip Pettit (2011)⁷ and Deborah Tollefsen (2015). Under this approach, the key question for agency is whether it’s possible (or rational, or reasonable) for an external observer to adopt the ‘intentional stance’ towards an entity. The intentional stance is a mindset in which one views the entity’s behaviours as explained by its beliefs, desires, hopes, fears, regrets, and so on. This stance is rational if it does a good job of explaining and predicting the observed behaviour of the entity. For example, if an observer sees me take the ice cream from the freezer, they might interpret me as desiring ice cream and believing it’s in the freezer. After observing me for a few days, they might predict that I’ll perform this behaviour in all future days after dinner. The (rationality of) ascriptions determine or even constitute my desiring ice cream after dinner and my believing it’s in the fridge. A key distinction between functionalism and interpretivism, then, is that the functionalist focuses on mechanisms internal to the agent, while the interpretivist focuses on behaviours that are observable from external to the agent. The functionalist also includes behaviours in their characterization of ⁷ List and Pettit appeal to interpretivism at some points (e.g., 2011, 11, 13, 23) and to functionalism at other points (e.g., 2011, 171). However, at one point they say ‘the performance [behaviour] itself should dictate the representations and motivations we ascribe to the agent’ (2011, 28). This places them more strongly in the interpretivist camp. I thank David Strohmaier for convincing me of this; see also Strohmaier 2020. Interpretivism differs from behaviourism: I take behaviourism to explain mentality just in terms of an entity’s behaviours, while interpretivism explains mentality in terms of how an idealized observer would interpret those behaviours.

22

  

an attitude’s functional role, but the primary focus is on an attitude’s relation to other attitudes, and in any case the behaviour is taken to be internal to the system, with no observer necessary. As I hope is clear, both functionalism and interpretivism have their merits.⁸ In the case of organizations, functionalism and interpretivism usually rise and fall together: an organization that’s an agent by functionalist lights will almost always happen to be an agent by interpretivist lights, and vice versa. Thus, for it to be sensible for an observer to adopt the intentional stance towards an organization over time, it will usually be the case that the organization has certain internal mechanisms (such as mechanisms that ensure the organization behaves in ways that are roughly consistent across time, as emphasized by List and Pettit (2011)). Meanwhile, an organization’s operation of rationality-producing internal mechanisms will almost always happen to result in outward organizational behaviour towards which it is appropriate for an observer to take the intentional stance. For purposes of figuring out whether real-world organizations are agents, then, we often don’t need to pick between functionalism and interpretivism. Partly because of this, I will make things difficult for myself: I’ll assume that an organization must meet both criteria in order to count as an agent. There’s another reason to insist that both functions and interpretations are important for organizations’ agency: in the case of organizations, it’s rather unclear where internal mechanisms end and observable behaviour begins (Strohmaier 2020). For example, are discussions between board members an internal mechanism, or an observable behaviour? If we endorse a combined ‘functionalist-interpretivist’ account (which demands an agent has both internal mechanisms and outward behaviour), then we don’t need to draw a sharp line between internal mechanisms and observable behaviour. What’s more, we should require both outward behaviour and internal mechanisms because doing so gives us a stronger argument against the sceptic of organizational agency. The higher a bar we force organizational agents to meet, the less room there is for complaint that we are being too ⁸ Both also have problems. Arguably, each theory countenances too many agents. For functionalism, collective agents are sometimes taken to be counterexamples (Block 1978); another would be complex robots. For interpretivism, it seems odd that my mental states are determined by, or even constituted by, someone else’s ascriptions. It seems there must be constraints on the interpreter’s ascriptions, if the ascriptions are to determine my agency. I will assume such problems don’t undermine understanding organisations as agents, though in Chapters 4 and 5 I will address nearby worries about organizations’ blameworthiness, namely, the worry that organizations lack the phenomenology necessary for blameworthiness.

   

23

permissive by including organizations in the category ‘agent’. At the same time, requiring organizations to meet both the functionalist requirement and the interpretivist requirement enables us to retain a clean theory of organizational agency. The two requirements do not contradict one another, and nor are we treating them as disjuncts (which would result in a somewhat disunified theory, with some organizations being ‘functionalist agents’ and others being ‘interpretivist agents’). Instead, the two requirements are layered neatly on top of each other. For all these reasons, I’ll assume an organizational agent must have both internal mechanisms and rationalizable external behaviour. On the functionalist picture, the features of organizations that render them agents are that their structure coordinates divided labour via rules and hierarchical command relations, guided by a collective decision-making procedure. While other kinds of collectives have the last of these features (a decision-making procedure), they do not necessarily coordinate divided labour via rules and hierarchical command relations. Arguably, the decision-making procedure is enough to produce collective-level agency (assuming the procedure abides by rationality constraints). But the other three features (coordinating divided labour, rules, hierarchies) are not present in all collective agents, and they boost the case for the agency of organizations specifically. These features ensure that—by and large, and about as reliably as individual agents—organizations have large webs of mutually-consistent beliefs, preferences, and other propositional attitudes. Such large webs of consistency are needed on the functionalist picture, because this picture characterizes particular propositional attitudes via their causal (and other) relations to other propositional attitudes. Organizations’ coordinated division of labour, rules, and hierarchies ensure that relations of consistency and mutual inter-definability hold between the organizations’ various propositional attitudes. An organization’s decision-making procedures, labour divisions, rules, and hierarchies are the mechanisms by which the web of mental states are produced and revised. What about interpretivism? Notice that organizations (perhaps unlike other collectives) are composed of people who realize a structure. This allows us to be clear about which behaviour belongs to the people qua private individuals, and which behaviour belongs to the organization— which ultimately allows us to attribute actions to the organization. The structure enables us to do this for two reasons. First, the structure provides boundaries to the organization: either a person is a role bearer or she is not;

24

  

only the former’s behaviour should be attributed to the organizational entity towards which we adopt the intentional stance. Second, the structure contains roles. It is only when a person is not just a role-bearer, but also a rolebear who is performing her role, that her behaviour should be attributed to the organizational entity towards which we adopt the intentional stance. Thus, organizations’ nature helps us to pick out the entity towards which we adopt the intentional stance. Most importantly, when we adopt the intentional stance to an organization, we find behaviour that is understandable and predictable: it’s understandable and predictable that banks will push legal limits for a profit and that states will permit environmentally dubious behaviour that boosts the short-term economy. Like under functionalism, the organization’s decisionmaking procedures, labour divisions, rules, and hierarchies loom large here: these are the mechanisms that produce understandability and predictability in the organization’s behaviour. Thanks to these mechanisms of organizations, we find that organizations behave in ways that make it rational for us to adopt the intentional stance towards them on a very regular basis. If interpretivism is true, this makes them agents. So we should view them as real.

1.4 The Questions Having introduced organizations, their wrongdoing, and their reality, the stage is set for the book’s substantive argument. As mentioned above, the book is divided into three parts: Part I is ‘Metaphysics’ (Chapters 1 to 3), Part II is ‘Blameworthiness’ (Chapters 4 and 5), and Part III is ‘Members’ (Chapters 6 and 7). Parts I and II treat organizations as unified wholes—just as this chapter has done in providing a philosophical characterization of organizations and three reasons for including organizations in our social ontology. Part III will consider the implications of organizations’ wrongdoing for variously-placed members of organizations. The rest of Part I is devoted to conceptualizing organizations as material objects that are dependent upon members. I develop an account of the organization-member relation that doesn’t reduce organizations to members, but that also acknowledges that members are crucial to organizations’ performances. The aim is to produce an account that makes organizations concrete material particulars, so that we can more readily incorporate their causal powers into a naturalistic understanding of the world. We want a

   

25

metaphysical account that will enable us to hold organizations blameworthy in their own right, while producing the right kinds of moral implications for members. Chapter 2 will consider various ways of doing this that don’t quite work. This will pave the way for Chapter 3’s positive proposal, which walks the tightrope between the various pictures rejected in Chapter 2. On my proposed account, organizations are material objects that have members as material parts. Applied to the Banking Royal Commission, members of financial corporations include directors, non-manager employees (such as financial advisors), and shareholders; but excludes intermediaries such as mortgage brokers (who offer products from more than one bank). Applied to the liberal democracies at issue in the Polluter Pays Principle, members include legislators, public servants, judges, and those with the right to vote in legislative elections; but excludes residents who lack the right to vote. On my picture, these individuals are the flesh-and-blood of organizations, even though organizations are not reducible to them. Part II, ‘Blameworthiness,’ begins in Chapter 4 by distinguishing recent moral-philosophical lenses on blameworthiness: the volitionist lens, the attributivist lens, and the aretaic lens. I explain how these lenses each apply to organizations. Most who have written on collectives’ blameworthiness have presumed the volitionist lens is the only relevant lens. I will argue that social ontologists can draw more widely and liberally from moral philosophers’ toolkits when it comes to our treatment of organizations. We can use the attributivist and aretaic lenses, as well as the volitionist lens. Drawing on Part I’s metaphysical picture, I will argue that organizations should be attributed both the ‘structural’ wrongs that inhere in the organization as a whole and the ‘individual’ wrongs that are enacted by members-qua-members. Chapter 5 responds to an important objection to the idea that organizations can be wrongdoers, under any of the volitionist, attributivist, or aretaic lenses. The objection is that organizations lack the capacity for moral self-awareness, that is, awareness of their own wrongdoing. Unlike other philosophers who have defended collective blameworthiness, I believe the capacity for moral self-awareness is indeed presupposed by attributions of blameworthiness—and that this undermines a simple functionalistinterpretivist account of wrongdoers (even though I maintain that, as argued in this chapter, the functionalist-interpretivist account works for mere agents). I give a novel account of how organizations have the capacity for moral self-awareness. This account relies on the metaphysical story of

26

  

Chapter 3. To foreshadow: the key innovation is that when members are performing their roles, certain pieces of their phenomenology are the organization’s phenomenology. The organization ‘inherits’ phenomenology from its physical parts (members), just as a motorbike ‘inherits’ colour from its physical parts (engine, wheels, and so on). This allows organizations to have the phenomenology of moral self-awareness. Part III, ‘Members’, explains the implications of organizations’ wrongs for members. Chapter 6 catalogues three ways members can be implicated in any given organizational wrong: a member can be an ‘enactor’, an ‘endorser’, or an ‘omitter’. I explain how divisions of labour and hierarchical relationships can affect the type and degree of an individual’s implication via each of these three mechanisms. Chapter 7 provides a justification for members’ reparative burdens when organizations are blameworthy for wrong. Importantly, Chapter 7 develops an account on which different members are liable to different shares of the reparative costs of their organizations’ wrongs. Members who are implicated in the wrong (in one of the ways analysed in Chapter 6) are liable to a larger portion of the costs than non-implicated members. However, nonimplicated members—including all present-day members of organizations that are blameworthy for historical injustice—can be liable to cost in virtue of having the capacity to repair relationships that were damaged by the organization’s wrongdoing, or in virtue of their having benefited from the organization’s wrongdoing.

2 Towards a Metaphysics of Membership 2.1 Introduction In Chapter 1, I characterized organizations as realized structures. I said that structures become realized when enough of their roles are occupied. Those roles are occupied by an organization’s members. This makes organizations dependent, in some sense, on their members. Nonetheless, if we are to hold a bank blameworthy for misconduct, then the bank must not be identical to a mere collection of members. It must be something we quantify over—as I argued we should over organizations in Section 1.3. If organizations at once depend upon, yet are distinct from, their members, then this raises the question: just what is the metaphysical relation between an organization and its members? We’ve seen that it cannot be a relation of identity. So what relation is it? Options include supervenience, causation, anchoring, grounding, and parthood. It matters which option we choose: these options give different results for the extent to which, and way in which, members are implicated in organizations’ wrongdoing. The options also ‘pick out’ different individuals as members. As we’ll see, if members are those upon whom organizations supervene, then this produces an implausibly vast number of members. Conversely, if members are those involved in an organization’s anchoring, then this will sometimes produce an implausibly narrow range of members. Getting the relation right matters for another reason. To incorporate organizations into a naturalistic picture of the world, it’s preferable for organizations to be material objects. If organizations are causally efficacious and we fail to account for their materiality, then we’d have to explain how an entirely nonmaterial entity can be causally efficacious. Even worse, we’d have to explain how there can be entirely nonmaterial agents in our world.

Organizations as Wrongdoers: From Ontology to Morality. Stephanie Collins, Oxford University Press. © Stephanie Collins 2023. DOI: 10.1093/oso/9780192870438.003.0002

28

  

That’s a large task, and one it’d be best to avoid—or so I will assume.¹ As Kendy Hess puts it, ‘[w]ithout this justified claim to material existence, the corporate agent would be exactly the ghostly, mysterious, incorporeal, bizarre, hovering entity mocked by the critics’ (Hess 2018a, 51, emphasis added). For an organization to be a material object, it must be an ‘individual entity with a unity of form and causal capacity through which it can be individuated and located in terms of its spatio-temporal coordinates’ (Sheehy 2006, 138). But where is an organization’s location? And what is its materiality? How can we draw lines around organizations as objects in the world? In this chapter and the next, I aim to answer these questions by analysing the relation between organizations and their members. The goal of this chapter and the next, then, is to answer two related questions. First, given that organizations are real yet dependent on members, just how do they relate to their members? Second, how can we account for organizations’ existence as material objects? As we’ll see, my answer to the second question will fall out of my answer to the first question. In this chapter, I argue that we should view members as material parts of an organization. In Chapter 3, I’ll provide a more specific analysis, using Kathrin Koslicki’s (2008) theory of parthood. Koslicki develops her theory in order to account for ordinary objects like motorcycles.² I will apply Koslicki’s theory to organizations in general, and then to corporations and liberal-democratic states specifically. Before Chapter 3 provides that positive proposal, the present chapter rules out various alternative proposals. Section 2.2 argues that the relations of supervenience, causation, and anchoring cannot capture the relation between an organization and its members. I examine these relations closely before rejecting them, for two reasons. First, they feature prominently in debates about individualism versus holism in the social sciences. From reading the literature on ontological individualism versus holism, one might get the impression that causation and supervenience are all that

¹ Some philosophers might find nonmaterial agents entirely acceptable. My aim in this chapter and the next is to give an account of organizations’ materiality, rather than to convince these philosophers. ² On Koslicki’s theory, an object’s material parts together constitute the object. So Koslicki’s theory—and my application of it to organizations—uses the notion of constitution, as well as that of parthood. But I emphasize the ‘parthood’ aspect, because parthood captures the thought that an organization’s members are themselves individualizable material objects, in the manner just quoted from Sheehy. Chapter 3’s use of Koslicki explains why I exclude ‘constitution’ from the above list of relations: on Koslicki’s view, which I follow, constitution is a kind of parthood; specifically, constitution is material parthood.

    

29

matter for capturing the way in which social phenomena relate to individual phenomena. These relations therefore deserve to be taken seriously. Second, these relations hold between organizations and various entities, where those entities are blameworthy in virtue of that relation. That is: we hold often hold various individual and collective agents blameworthy, for the fact that they subvene under, cause, or anchor organizations. Thus, these relations should be included with our full picture of blameworthiness when organizations do wrong. The relations simply do not capture how organizations are themselves material objects that do wrong, or how organizations relate to their members—or so Section 2.2 will argue. Section 2.3 turns to grounding and parthood. I argue that members are most centrally parts of organizations, even though it’s true that some facts about members ground some facts about organizations. This is important for organizational wrongdoing: Part II will argue that various properties of individuals are (under certain circumstances) also properties of the organization of which they are members. The ‘transfer’ of these properties from members to organizations will be important for incorporating wrongful organizations into our socio-political blame practices. The ‘transfer’ of members’ relevant properties will not work if individuals are merely causes, or supervenience bases, or anchors, or grounds, of organizations. The transfer will rely on the idea that members are parts of organizations. Once we know that members are parts of organizations, we are left with the question: how do we pick out who the members are? In Section 2.4, I’ll consider other social ontologists’ answers to this question, where those social ontologists take members to be parts of organizations, as I do. As we’ll see, none of their answers quite work. But consideration of their views will produce several desiderata on a full account of the organization–member relation. These desiderata will guide my own full account of the relation, which I will present in Chapter 3. In this way, the arguments of this chapter are largely negative. I will proceed by examining various possible characterizations of the relation between organizations and their members. I will argue that none of those characterizations is satisfactory. But these negative arguments will have important payoffs, since they will guide and constrain my positive proposal in Chapter 3. Before beginning, I should admit that the ‘realized structures’ view of organizations is not the only game in town, though it is the picture I presume in this book. Perhaps organizations are instead pluralities, fusions, or sets (Uzquiano 2018; Wilhelm 2020; Horden and López de Sa 2020;

30

  

Effingham 2010). If so, then the relation between organizations and members might not be one of parthood. My arguments for the parthood view (in this chapter) and for my specific version of the parthood view (in the next chapter) will assume that organizations are realized structures. I favour this view because it emphasizes the importance of roles and intra-organization relations in the make-up of an organization. Roles and intra-organization relations are crucial to organizations’ explanatory and normative power. What’s more, as Parts II and III will show, the idea that ‘organizations as concrete material particulars are realized structures of which members are material parts’ can do useful work in moral and political philosophy. For that reason alone, it is a view worth defending.

2.2 Causation, Supervenience, Anchoring 2.2.1 Causation In Section 1.3, I alluded to two prominent theories of causation—the ‘counterfactual’ and ‘process’ theories. I emphasized that, on both theories, causation can hold in the organization-to-components direction (facts internal to organizations causally depend on facts about organizations) and in the organization-to-world direction (facts external to organizations causally depend on facts about organizations). Causation also holds between members and organizations: facts about an organization sometimes causally depend on facts about members; and facts about members sometimes causally depend on facts about organizations. This seems true on both the counterfactual and process theories. So you might think causation captures how organizations relate to their members. The most obvious examples of member-to-organization causation are cases where members cause organizations’ decisions. For example, what caused Australian banks to approve loans that customers were clearly unable to service? Many factors had a causal influence, but a key factor was individuals’ actions. As one informant of the Banking Royal Commission put it, ‘mortgage brokers were more likely to say to a client “this is the estimate monthly expenditure for a family your size; does that sound right to you” than ask open-ended questions designed to identify actual expenses or consider bills and bank statements to test the reality of the client’s estimates’ (Royal Commission 2018, 53). Here, the words of mortgage brokers made a causal difference to the bank’s decision about whether to grant a loan—

    

31

according to both the process and counterfactual theories of causation. This causal influence isn’t diminished even if we also acknowledge that causal influence was wielded by the customer (in giving the answer they did) or by banks’ norms and expectations (for example, that brokers receive higher commission if the customer borrows more money). Things are more obvious in the other direction: organizations can cause their members to behave, think, and even feel certain ways. Organizations’ formal policies of remuneration, promotion, and misconduct can constrain and guide members’ behaviour. Harder to track, but no less important, is how organizations’ cultural norms, implicit expectations, and on-the-ground practices of subtle reward and punishment can affect what members do. Causation runs both ways between organizations and members. Yet to view causation as the essence of the member–organization relation is to understate the internality of the relation between an organization and its members. The example just used demonstrates this: brokers are not usually taken to be members of banks, partly because they usually offer products from several different banks.³ Causation can hold between an organization and just about any individual or object. Causation is not particular to members. Causation can hold between entities that are wholly separate from one another. So, to construe the member–organization relation as one of mere causation misses the internality of the member– organization relation. While causation often holds between an organization and its members (in both the organization-to-member direction and the member-to-organization direction), the same is true of numerous nonmembers. Causation cannot not the heart of the matter.

2.2.2 Supervenience Many social ontologists focus on supervenience (e.g., Macdonald and Pettit 1981; Kincaid 1986; Sawyer 2002; List and Spiekermann 2013). This is understandable, since many social ontologists are concerned to show that we should be realists about collective phenomena, while at the same time showing that realism does not make collective phenomena into mysterious forces in the world. One popular way to achieve non-mysterious realism is to say that organizations supervene on individuals (more precisely, to say that ³ Chapter 3’s account of membership will explain in more detail why brokers should be excluded.

32

  

facts about organizations supervene on facts about individuals: supervenience is usually taken to hold between facts or properties, not objects). In general, facts about A supervene on facts about B if and only if any change in any fact about A implies a change in facts about B. Another way to put this is: once we have settled the B-facts, we have settled the A-facts; or, two worlds that are indiscernible with regards to B-facts are also indiscernible with regards to A-facts. In philosophy of mind and philosophy of social science, if B-phenomena (such as individuals) are not mysterious, then the A-phenomena (such as organizations) are usually also taken to be (in some sense) not mysterious. Yet supervenience is compatible with realism about the A-phenomena, since A-facts’ supervenience on B-facts is compatible with the A-phenomena being multiply realizable, causally efficacious, and agential (as discussed in Section 1.3). Because the supervenience of organizations on members can produce non-mysterious realism about organizations, supervenience is often taken to be an important component of the organization-member relation. But is supervenience the core of the organization–member relation? Are an organization’s members exactly those entities such that any change in any fact about an organization implies a change in facts about the entities? The most plausible route to a ‘yes’ answer would focus on global supervenience (on which see Currie 1984). According to this possible view, any change in any fact about the organization implies a change in some fact or other about members, not necessarily a change in facts about members where those member-facts are somehow ‘about’ or ‘related to’ the changed organizationfact. That is, the most plausible supervenience view would say that if we have settled all the facts about members—including all facts about their behaviours, mental states, histories, dispositions, and so on—then this will determine all facts about the organization. Yet even if we take in as many facts about members as we possibly can, these facts will not be enough to fill in the supervenience base of facts about organizations. That is: the supervenience base of facts about organizations clearly includes facts about entities that are not members. Indeed, Epstein (2015, esp. 46) compellingly argues that even individuals-at-large are not the entirety of the supervenience base for facts about organizations. He gives the example of a coffee company that goes insolvent overnight because all the fridges break. Here, there has been a change in the facts about the company—the company has gone from being solvent to insolvent—without any changes in facts about any individuals, let alone members. Individuals (let alone members) might not know, believe, or suspect that the company

    

33

has gone insolvent: the company can go from solvent to insolvent while all individuals in the world sleep. Now, one might object that for a coffee company to go insolvent overnight, there needs to be a background of social principles that determine what it means for a company to be or become insolvent. These ‘insolvency principles’ are, perhaps, set up or underpinned by individuals. But this doesn’t mean individuals—let alone members—are the entirety of the supervenience base of the insolvency: we can imagine two possible worlds, in which all the insolvency principles and all the facts about the company’s members are exactly the same, yet in one world the company’s fridges break and in the other the fridges keep working. There is a difference in the company-level facts of these two worlds (a difference in the solvency facts), without any difference in any facts about individuals—let alone a difference about members specifically. In this way, supervenience is too broad for analysing the relationship between organizations and their members. Just like with causation, a huge number of individuals and objects are within the supervenience base for facts about organizations. Of course, this may prove useful for real-world blame practices: if an agent (individual or collective) is such that changes in certain facts about them would change certain facts about an organization, then that agent may well be blameworthy for certain facts about the organization. In a way, then, supervenience may turn out to be extremely useful for responding to the world at large when organizations do wrong. It’s just not useful for understanding the internal composition (or material reality) of the organization itself as a wrongdoer. This obvious point is worth stressing, for two reasons. The first is the prevalence of supervenience in social ontological debates. It’s been prevalent for a good reason (namely, its role in establishing the non-mysterious irreducibility of social facts and properties)—but that doesn’t mean it captures the most interesting and important relations between individuals and organizations, such as the relation of membership. The second reason for discussing supervenience is that the limits of supervenience are more apparent for organizations specifically, rather than for collective agents or human groups at large. For some human groups—such as two individuals painting a house together or two individuals going for a walk (examples discussed by Bratman (2014) and Gilbert (1989), respectively)—it might be plausible that facts about members are the entirety of the supervenience base for facts about the group. Those groups are fairly small, self-contained, short-lived, simple, and unregulated by outsiders. But supervenience fails

34

  

to capture the relation members stand in to organizations, even if it captures the relation members stand in to other kinds of groups or collective agents.

2.2.3 Anchoring Epstein (2015) develops the concept of ‘anchoring’ to capture one way in which social facts are set up. It’s a way that was alluded to in my discussion of supervenience. Roughly, X anchors Y just in case X sets up the principles that state the conditions under which some Z grounds some Y. (I’ll discuss grounding soon.) Epstein explains anchoring via John Searle’s view of money. According to Searle, certain pieces of paper ‘count as’ money (or, in Epstein’s framework, ‘ground’ money) when those pieces of paper are printed by a relevant authority. So, we have a ‘money’ principle: under the conditions of being printed by a relevant authority, pieces of paper ground money.⁴ But what underpins this principle? This is the question: what anchors money? According to Searle, the anchor of money is collective acceptance: collective acceptance sets up the money principle. Epstein problematizes Searle’s ‘collective acceptance’ theory, but Epstein is agnostic on whether facts about individuals (different from or in addition to facts about individuals’ collective acceptance) anchor social facts (2015, 106). So, organizations might be anchored by facts about individuals. And one might think that they are anchored by their members specifically, not by individuals at large. It’s important not to confuse anchoring with causation. To illustrate via Searle’s view of money: in Searle’s view, it’s not that collective acceptance causes the money principle to hold. That is, it’s not that collective acceptance causes it to be the case that, under the conditions of being printed by a relevant authority, pieces of paper ground money. The relationship is more intimate than that: on the Searlean picture, collective acceptance is the constant or ongoing metaphysical explanation for the money principle. In Epstein’s terms, the anchors ‘put in place’ or ‘set up’ principles that state the conditions under which social phenomena are grounded (2015, esp. 81). On the Searlean picture, money is anchored in collective acceptance because ⁴ Strictly speaking, this is the ‘paper money’ principle. Electronic money has grounds other than paper, so electronic money will require a different principle, which will in turn be anchored by potentially different agents or phenomena. I use the money example simply to introduce the anchoring relation. I don’t take a stand on how money in general is grounded or anchored. I thank an anonymous reviewer for pressing the importance of electronic money.

    

35

the money principle depends upon collective acceptance in an ongoing way—much as a boat’s location depends upon its anchor in an ongoing way. The ongoing nature of anchoring makes anchoring different from causation. The illustrative metaphor of an ‘anchor’ might seem apt for the member– organization relation. But it isn’t. The anchors of organizations are numerous, varied, and complex—they include much else in addition to facts about members. The anchors include facts about laws, norms, practices, physical infrastructure, and so on. As Epstein says, ‘[t]hings like corporations and money are devilishly complicated, much more so than Searle and others suggest. They are anchored by elaborate social structures, and their grounding conditions alone would fill books’ (2015, 130). Take a bank, for example. What is it that ‘sets up’ or ‘underpins’ the principles that state the conditions under which a bank exists or is grounded? Foremost amongst these is surely the law. But laws about banks are not facts about, nor are they set up by, banks’ members. I mention anchoring primarily to acknowledge, again, that things other than members have a crucial role to play in underpinning organizations’ existence and in determining specific facts about organizations. Things other than members are an organization’s anchors, and we cannot lose sight of the anchors when fitting organizations into our social, political, and legal practices—including practices of blame. Sometimes, the anchors will be blameworthy for the principles that govern the existence of organizations. But precisely because anchors include phenomena as diverse as laws and physical infrastructure, this role does not pick out the way in which organizations relate to their members. There is another reason for mentioning anchoring, which is that it helps to further distinguish organizations from other types of collective agent. As with supervenience, anchoring might capture the relation between members and simple, small, short-term collective agents—such as two people painting a house or going for a walk. I am neutral on whether this is correct. Whether it is or not, the anchors for organizations are far more complex, historically protracted, and socially extended. The anchors for organizations go well beyond members. So anchoring won’t work for capturing the relation between members and organizations, even if it does capture the relation between members and some other groups. Two themes emerge from my discussion of causation, supervenience, and anchoring. First, these relations are candidates for capturing (what’s important about) the member–collective relation in simple, small, short-term collective agents. But they don’t work for organizations. So, these relations

36

  

demonstrate one way in which organizations are worth handling on their own terms, separate from simple, small, short-term collective agents. Second, causation, supervenience, and anchoring help to explain how social phenomena exist. These social phenomena include organizations. What’s more, these relations are plausibly sometimes a basis for wrongdoing. Perhaps we should sometimes blame the causes, subveners, or anchors of organizations—instead of, or additionally to, organizations themselves. But this book is about how to conceptualize organizations themselves as wrongdoers.

2.3 Grounding and Parthood 2.3.1 Grounding Given the foregoing discussion of anchoring, it might seem natural to turn to grounding. In the Searlean picture, grounding is the relation that holds between facts about certain pieces of paper and facts about money. In the most general terms, grounding is the ‘in virtue of ’ relation. One thing grounds another when it is ‘the metaphysical reason’ for the other or that ‘in virtue of which’ the other holds. Grounding is generally taken to be irreflexive (nothing grounds itself), asymmetric (if A grounds B, then B doesn’t ground A), and transitive (if A grounds B and B grounds C, then A grounds C). In its most straightforward application, grounding is a relation between facts—not objects, properties, events, etc. For example, the fact that there is $10 in my wallet is grounded in the fact that a certain kind of piece of paper is in my wallet. $10 is in my wallet in virtue of a certain kind of piece of paper being in my wallet. Grounding gives us the following way of understanding the organization– member relation: certain facts about members are the metaphysical reason why certain facts about organizations obtain (and perhaps vice versa). Of course, facts about all sorts of things ground facts about all sorts of other things. And we can ask ‘what are the grounds?’ for specific social facts (such as ‘this bank lied to customers’) or more general social facts (such as ‘organizations exist’). For grounding to give us a substantive proposal about organizations’ relation to their members, we need to know which facts about members ground which facts about organizations (and vice versa). Only then will grounding enable us to distinguish the members from the non-members. We might then be able to draw lines around

    

37

organizations as objects in the world and account for organizations’ material existence. Epstein makes such a proposal. According to him, members constitute groups,⁵ where ‘constitution . . . is about the stuff of one thing being part of the metaphysical reason that another thing is made up of the stuff it is’ (2015, 149, emphasis added). That is, for Epstein, constitution is about grounding. Using this notion of constitution, Epstein argues that a set of people—a set labelled ‘M’—constitutes a group. (I’ll ignore the strangeness of saying that a concrete material object is constituted by an abstract object like a set.) Epstein uses the example of the Massachusetts Department of Transport (MassDOT) board. He proposes that ‘[f]or M to constitute the MassDOT board at t is for M to meet two conditions: M needs to coincide with the MassDOT board at t, and certain facts about M need to partially ground the fact that the board has the membership it does, at t’ (2015, 149). That is, members are those who partially coincide with the group and who partially determine the fact the group has the membership it does. However, as Epstein notes, a huge number of people are such that facts about them are part of the metaphysical reason why a group has the membership it does. The second part of Epstein’s proposal doesn’t help to pick out members specifically. This suggests that the first part of Epstein’s proposal—that members coincide with the group—does most of the work, for the purpose of establishing how members in particular relate to the group. But ‘coincidence’ doesn’t help us with actually picking out who the members are. This condition doesn’t tell us who coincides with the group. Instead, it assumes we already know which individuals coincide with the group.⁶ But this is precisely what our account of the member–organization relation should enable us to figure out. Aside from the specifics of Epstein’s proposal, there are two general problems with cashing out the member–organization relation in terms of grounding. First, if a ground is a ‘metaphysical reason’, then grounding can’t explain the materiality and physical location of organizations. That is, grounding can’t help us to ‘draw lines’ around organizations, as I’ve been

⁵ Epstein talks about groups rather than organizations. I take groups to be a family, of which collective agents are a genus, of which organizations are a species. (I thank an anonymous reviewer for this taxonomizing language.) ⁶ Epstein might reply that the anchors tell us who coincides with the group. But as I’ll argue in Section 2.4, depending on what the anchors are, it might be that we shouldn’t defer to anchors in this way. This point will become the ‘non-conventionality’ desideratum by the end of this chapter.

38

  

putting it. Grounding can’t help us to point to organizations as objects in the world, since it’s a relation between facts, not objects. That said, grounding could perhaps feature in explaining organizations’ materiality and locations—for example, perhaps facts about organizations’ materiality and locations are grounded in other facts. But in that case, those latter facts would be doing all the work in allowing us to conceptualize organizations as concrete material objects around which we can draw lines. Second, grounding doesn’t capture the distinctive relation that members stand in to organizations. As I noted above, facts about all sorts of individuals are part of the metaphysical reason why organizations have the members they do. Some of these individuals are not members. This seems likely to be true for any fact about an organization: non-members will be part of the metaphysical reason for that fact. Thus, grounding does not give us an account of organizations as objects in the world and does not give us an account of the organization–member relation.

2.3.2 Classical Parthood Parthood is traditionally understood in an extensional sense: wholes are individuated by their proper parts. No two objects of the same type can be distinct from each other if there are no differences in any of their parts. Thus, if two drystone walls share all the same bricks, then those walls are one and the same wall. Koslicki (2008) calls this view ‘classical extensional mereology’. If the relationship between organizations and their members were one of classical parthood, then the following would hold: if two organizations share all the same members, then those organizations are one and the same organization. For this reason, classical parthood won’t quite do for capturing the relationship between organizations and their members. This is because two organizations can share all the same members while being distinct organizations. We can imagine a large and complex history club, that happens to have all the same members as an equally large and complex philosophy club. The two clubs are both organizations. They are distinct, despite sharing all the same members. This problem—of how to differentiate such ‘overlapping groups’—has been noticed by numerous social ontologists (e.g., Sheehy 2006; Uzquiano 2004; Effingham 2010; Epstein 2015; Ritchie 2015; Hawley 2017; Strohmaier 2018). You might think that classical parthood is problematic only if one thinks that members are the only parts of organizations. What if one thinks that

    

39

members are some of the parts of organizations, while holding that organizations also have other parts? For example, suppose you think organizations have buildings, machines, and paper amongst their parts. We might then be able to distinguish the history club from the philosophy club, by pointing out that they have different buildings, machines, and paper from each other. In this way, perhaps we can maintain that the relation between organizations and members is one of classical parthood, without implausibly saying that the history club and the philosophy club are identical. However, we can imagine that the history club and the philosophy club share all their buildings, machines, paper, and so on in common, yet are distinct organizations. Perhaps the history club uses the office space (with its accompanying machines and so on) on Mondays and Tuesdays, while the philosophy club uses the office space on Thursdays and Fridays. In that case, the two clubs share all the same parts—including non-member parts. Thus, classical parthood is inadequate even on this more expansive view of organizations’ parts. Of course, one can bite the bullet: perhaps the history club and the philosophy club are the very same organization. They share all the same parts, they are located in the same place, and are one and the same concrete material particular. Katherine Hawley (2017) defends this position. In its defence, she provides an analogy. She points out that one person can be both the Mayor of London and a Member of Parliament (as Boris Johnson once was). In this case, the Mayor and the MP are the same person. There is just one person: Boris Johnson. According to Hawley, this person sometimes acts-as-mayor and sometimes acts-as-MP. Likewise, Hawley suggests, we should embrace the conclusion that overlapping organizations are simply one and the same organization, just as the mayor and the MP are one and the same person. In our example, there is just one organization—where that organization sometimes acts-as-history-club and sometimes acts-as-philosophy-club. We just have to be careful to distinguish when we are dealing with the-organization-qua-history-club and when we are dealing with the-organization-qua-philosophy-club. But it’s the same organization either way, just as Boris Johnson is the same person regardless of whether he’s playing the role of mayor or MP. Thus, members are ‘parts’ of organizations in the classical mereological sense of ‘parts’. I have doubts about the analogy between individuals occupying social roles and organizations being themselves. When an individual such as Boris Johnson occupies a social role, this is a matter of an individual gaining a property. By contrast, when an organization is a history club, this is a matter

40

  

of a new object coming to exist: a history club. It doesn’t make sense to talk of the organization independent from, or prior to, its being a history club or a philosophy club. By contrast, it does make sense to talk of Boris Johnson independent from, or prior to, his being mayor or MP. So, Hawley’s argument by analogy seems doubtful.⁷ But suppose we accept Hawley’s proposed analogy between social roles and organizations. Still, the classical parthood view just pushes the interesting question back. If the history club and the philosophy club are one and the same organization, then the interesting question becomes: how are we to distinguish when this organization is acting-qua-history-club from when this organization is acting-qua-philosophy-club? After all, no one can deny that there is a meaningful distinction between the two. If the organization does wrong, then it will be important to know whether the organization did wrong qua-history-club or qua-philosophy-club. After all, we will likely hold the organization to different moral standards, depending on whether it is wearing its history ‘hat’ or its philosophy ‘hat’. (Similarly, we would hold Johnson to different standards, depending on whether he was wearing his mayor ‘hat’ or his MP ‘hat’.) Our social ontology should make the distinction. The traditional parthood relation doesn’t enable us to make the distinction. Furthermore, we would still need to know what it takes for any given agent to be a member of the history-cum-philosophy club. I aim to answer this ‘what it takes’ question in Chapter 3. Paul Sheehy gets closer to our goal. Like Hawley, he paints a picture on which members are parts of organizations, understood in the classical mereological sense. Like Hawley, he accepts that the history club and the philosophy club are spatially coincident, co-located, and have all the same material parts (2006, 131, 140). Yet, Sheehy insists that coinciding groups ‘remain distinct entities in virtue of the different ways in which the parts are organised and related’ (2006, 140). He elaborates: two groups can be in the same place at the same time because the individuals constituting them can be organised in distinct ways. Sameness of parts does not entail sameness of fundamental structure. . . . individuals organised according to distinct sets of relations are constitutive of distinct groups possessed of their own properties and powers. (2006, 135)

⁷ Horden and López de Sa (2020) rely on this same analogy in their argument that groups are pluralities; my doubts expressed here extend to their argument.

    

41

As Sheehy notes (2006, 127), this account gives up classical parthood, according to which two objects with all the same parts are the same object. On his view, the history club and the philosophy club are to be differentiated by their structures, where a group’s structure is not a part of it. I believe Sheehy’s focus on structure is a step in the right direction. The idea of an organization’s ‘structure’ will importantly re-emerge in Chapter 3. But Sheehy’s view still doesn’t give us a substantive account of who is a member. It doesn’t allow us to distinguish the parts that are members from the parts that are non-members (if indeed there are any parts that are nonmembers, such as buildings, machines, and paper). Or, if all parts are members (or parts of members), then the view doesn’t give us a way of distinguishing the parts from the non-parts. Although Sheehy provides a coherent account of organizational parthood, he doesn’t provide a full account of organizational membership. He doesn’t tell us which individuals are the parts. This means his account doesn’t enable us to draw lines around organizations as objects in the world, to locate them in spacetime, or to pick them out as causally efficacious. In any case, there is another reason why classical parthood doesn’t fully capture the relation between organizations and members. As mereologists understand classical parthood, that relation is transitive. Thus, if I’m a part of New Zealand and if New Zealand is a part of the UN, then I’m a part of the UN. But it seems I am a member of New Zealand, and that New Zealand is a member of the UN, while I am not a member of the UN. This suggests that ‘part’ and ‘member’ cannot be perfectly interchangeable. Instead, our ordinary use of the concept of ‘membership’ seems to evoke some kind of involvement that a member has in an organization, where this involvement is not transitive. I have this kind of involvement in New Zealand, but I do not have this kind of involvement in the UN; so I am a member of New Zealand but not the UN. Because parthood is transitive while membership is not, membership cannot be reduced to parthood. Given all these twists and turns, I suggest we conclude as follows: members are parts of groups, but they are not merely parts of groups. Members are parts-plus: they are parts that also have some further property. This further property is the kind of ‘involvement’ that I have in New Zealand, but that I do not have in the UN. Thus, I am a part of the UN, but I am not a member of the UN, because I do not have the further property. I am a part, but I am not a part-plus. This raises the question: what is the ‘plus’? What is the further property, in virtue of which a part is a member? Hawley defers to lawyers or to practices on this issue, stating that ‘the membership conditions of a given group are primarily a matter for

42

  

empirical or legal investigation’ (2017, 403). On this view, all the interesting ontological work is done once we have endorsed the classical extensional mereological view. The rest is law or sociology. I find this dissatisfying, because we have to choose which laws or practices to take as authoritative. Suppose, for example, that the UN simply declared that I was a member— without changing anything about my rights or obligations vis-à-vis the UN. Or suppose that the history club and the philosophy club simply declared that they are one organization—without doing anything to unite their structures, constitutions, or procedures. These organizations would be getting something wrong: despite the declarations, it would remain true that I am not really a member of the UN and the two clubs are not really identical. The declarations are not enough to determine the facts of membership or the facts about organizations’ distinctness from each other. We don’t defer to the law (or to the prevalent views in society) when drawing the lines around individual agents, so we shouldn’t defer to these things when drawing the lines around organizational agents, either. To be clear, I do not doubt that the conditions for organizational membership are grounded in human practices: organizations and membership are socially constructed. But those conditions have been socially constructed in such a way that they can be captured and analysed by philosophy. To do that would be to give a metaphysics of membership. What’s more, as I’ll go on to show in Parts II and III, our metaphysics of membership has moral upshots. What we say about organizations’ metaphysical relationship to members has implications for the blameworthiness, reactive attitudes, and reparative burdens of both organizations and members. It’s important that we get it right.

2.4 Members as Parts-Plus: Existing Accounts Let’s take stock. I have suggested that members are parts of organizations. But they are not just any old parts (as I am to the UN). Members are those amongst an organization’s parts that have some special features. Members are ‘parts-plus’. I will lay out my own proposal for the ‘plus’ in Chapter 3. In the final section of this chapter, I will examine other parthoodplus proposals. As I mentioned in Section 2.1, examining these other proposals will yield several desiderata on an account of the organization– member relation. These desiderata will guide my positive account in

    

43

Chapter 3. I examine others’ proposals here so that the desiderata can be understood and justified.⁸ First, one might turn to Margaret Gilbert’s influential work. For Gilbert, two or more individuals constitute a ‘plural subject’ when ‘they are jointly committed to intending as a body to do A’ (Gilbert 2000, 22). The ‘as a body’ construction might be taken to imply that the individuals are thereby parts of the plural subject. And one might think that the ‘joint commitment’ is the relevant ‘plus’. For Gilbert, the joint commitment is not a commitment of each member, but rather a commitment that is irreducibly held by the members together. That is, perhaps the members of an organization are all and only those entities that are parties to a joint commitment to act as a body to do some organization-level action, where being a party to a joint commitment makes one a member of the relevant organization.⁹ The problem here is that many organizations exist without any joint commitments, and certainly exist without joint commitments that have all members as parties. This point has been made by Kendy Hess (2013; 2018b). She focuses on shared intentions rather than joint commitments, but her point applies likewise to joint commitments. The point simply is that many members of organizations play their roles without thinking about the organization’s actions, still less by being party to a joint commitment together with all other members to perform those organization-level actions as a body. As Hess puts it, ‘[i]f you can do your job without knowing or caring about the ultimate end goal of the corporation – which would seem to be the norm – then the corporation is going to continue to function just as intentionally and rationally as ever, even if none of its members actually intend that it do so or care if it does’ (2013, 64). Joint commitments to organizational goals cannot, it seems, be necessary for membership.¹⁰ As ever, one might bite the bullet here. Perhaps many employees in large organizations simply are not members of those organizations. Perhaps membership in an organization is a relatively rare phenomenon, with the

⁸ One might wonder whether all parts of organizations are ‘parts-plus’, because the organization’s structure can always be used to explain why or how something is a part. This explanation would be a ‘plus’. I thank an anonymous reviewer for this point, which is I think is broadly correct. However, the ‘plus’ for members is particularly important, since members are agentially implicated in organizations. In Chapter 3, Section 3.2, I give my view of what the ‘plus’ is for members. ⁹ I thank an anonymous reviewer for suggesting Gilbert’s account. ¹⁰ In Chapter 3, my positive proposal will require minimal commitments from members— but these will not be joint commitments to group goals, which are what makes Gilbert’s account distinctive, and distinctively demanding.

44

  

members being only those few founders or senior managers who really are jointly committed to perform certain organization-level actions as a body. However, there may be some organizations that have no parties to any such joint commitment. Here we can think of organizations that have existed for a long time (such that the founders are off the scene) and that are run for a profit (such that individualism and competition rule the day, with few participants thinking of themselves as part of a collective body). These organizations are not conducive to the intimate picture of joint commitments as painted by Gilbert. Certainly, it seems preferable for our theory of organizations as wrongdoers to include such organizations—and to include, amongst the actions for which organizations might be blameworthy, actions that are performed by employees who may not be party to a joint commitment. Another possible view is parthood-plus-designation-as-member. That is, perhaps the members are those entities that are both parts of a group and are designated as its members. This view avoids the problems of overlapping groups and transitivity. The history club and the philosophy club become distinct entities, because each organization has performed its own distinct acts of designating—even though all the parts of one are also parts of the other, such that the two groups are perfectly co-located. And, on this view, I can be a member of New Zealand without being a member of the UN, because one can be designated a member of New Zealand without being designated as a member of the UN. It’s still true that members are parts: they’re just parts with a special designated status. So we have solved the problems of where the organization is located (it is located where its parts are located) and how it is a concrete material object (its parts are its matter), while distinguishing two groups via their different designated memberships. This view is unsatisfactory for a few reasons. First, ‘member’ occurs on both the left-hand and right-hand side of the analysis: to be a member is to be a part that’s been designated as a member. This leaves unclear what it means to be designated as a member, rather than designated as anything else. Second, the concept of ‘designation’ needs analysis, or else it opens the door to just the kind of problematic deference that we found in Hawley’s account: designations can get it wrong, in the sense that they can designate as members parts of the organization that are not really members (such as a judge’s nose, which is a part of a court but not a member), or they can fail to designate as members parts of the organization that really are members (such as a judge, which is a part of a court and a member). David Strohmaier (2018) makes an alternative suggestion: the members are the parts that are agents appropriately designated to contribute to the organization’s functioning (2018, 132). This seems to solve both the

    

45

problems mentioned above: ‘member’ occurs on the left-hand side, but not the right-hand side, so we avoid circularity. And the account doesn’t purely defer to designators: it says the designators get it wrong when they designate as members those that do not contribute to the organization’s functioning. Yet Strohmaier’s account makes short-term one-off contractors into members, if they are ‘designated’ to contribute to the organization’s functioning. After all, these contractors surely do, in fact, contribute to an organization’s functioning. So all his account requires is that they are furthermore ‘designated’ to do so. Thus, while this account doesn’t entirely defer to designators, it defers to them too much. In response, Strohmaier could say that contractors are not parts. But then we need an account of what makes an entity a part, in order to have a full analysis of membership. So the problem would just be moved around within the analysis.¹¹ It’s important to guard against false positives, such as contractors being members. Why? As I’ll discuss in Chapter 7, organizations’ wrongdoing has important implications for members: it generates costs for them. And as I’ll argue in Chapter 5, members’ capacities have important implications for organizations’ status as wrongdoers: members’ capacities for phenomenology make organizations capable of moral self-awareness, where moral selfawareness is a condition on blameworthiness. Our account of membership should enable us to distribute these implications appropriately, meaning it should guard against false positives. We could remove ‘designation’, then, and simply say that the members are the parts that play a functional role within the organization. Ritchie’s account is like this. She proposes that ‘[s]ome things (X) are the members of a group with a structure S at time t and world w if, and only if, together X occupy the nodes of S at t at w (i.e., X are functionally related, or at least normatively bound, in the ways required by S at t at w)’ (2018, 23). We saw in the previous chapter that I endorse Ritchie’s view that organizations are realized structures, in which members occupy roles. And above, I endorsed Sheehy’s move to place some emphasis on structure when distinguishing between organizations with the same members. So, Ritchie’s account of membership might seem amenable to my characterization of organizations. However, objects other than members can occupy roles in an organization’s structure. Buildings, machines, paper, and so on play functional roles in organizations, which might be expressly laid out in something like an organizational chart. Yet these objects are not members, even if they are

¹¹ I thank David Strohmaier for this reply and rejoinder.

46

  

parts. We can fix this by saying that the members are all and only the agents that play functional roles. But this still includes short-term contractors. Members are more than merely ‘functionally related’ to the organization. Membership even requires more than being normatively bound by one’s relationship to the organization, since, again, contractors and others with tenuous relationships to organizations can be normatively bound to the organization, in virtue of their contract or other promissory relationship. Ritchie might reply that non-agents and contractors are not usually included in an organization’s structure or role-chart. But then again this makes the account too contingent on conventional practice: if we discovered an organization that did include its computers and contractors in its organizational chart, then suddenly those computers and contractors would count as ‘members’. When we are picking out organizations as wrongdoers, we should not allow for such contingencies. Another kink in Ritchie’s account concerns the concept of ‘occupying’ a role: do I ‘occupy’ a role when I hold a role in an organization, or only when I’m performing that role (however imperfectly)? Take my relationship to my university. Did I start occupying a role when I signed my employment contract? Or on the day my employment started? Or only when I first showed up for work? Most importantly: do I continue to occupy my role on the weekend, or while I’m away on holiday? The most natural interpretation is that my role-occupancy began on the day I started employment and that my role-occupancy continues on weekends and holidays, until my contract is terminated. But this raises a dilemma about organizations’ location, materiality, and existence. If members are material parts of organizations by virtue of occupying roles in the structure, and if they occupy those roles from the moment they start their ‘job’—including throughout weekends and holidays—then is my university partially located in Fiji when I go on holiday to Fiji? Ritchie’s answer, it seems, must be ‘yes’ (the same goes for Harris (2020)). This counterintuitive result is the first horn of the location dilemma. In Chapter 3, I will indeed grasp this horn—but I will argue for doing so. On the other horn of the location dilemma, if Ritchie says that the structure is instantiated or realized only when people are performing their roles, then we don’t get the counterintuitive result that my university is located at my holiday destination. Instead, we get the counterintuitive result that the organization’s material existence rises and falls with work hours: the organization pops in and out of material existence as people clock in and out of work. This is the second horn of the location dilemma.

    

47

There’s a way out of the dilemma: perhaps organizations are material objects that have members as material parts, yet organizations are not located where their members are located. Along these lines, Hindriks points to an ambiguity in ‘location’. He notes that, in the US at least, limited liability companies are required to have a designated office. The office’s address enables government departments to communicate with the company and determines which laws apply to the company. Hindriks calls the designated office’s location the institutional location of an organization (2013, 418). In addition, the organization has a physical location. This is the location of its members. He argues that ‘some of our intuitions about the location of an organization pertain to its institutional location, whereas others pertain to the location of its members [i.e., its physical location]’ (2013, 418; similarly Hindriks 2021). This raises the question: if the location of the members doesn’t determine the organization’s institutional location, then in what sense does the institutional location provide a material basis for the organization’s existence? The organization’s institutional address might be nothing more than a mailbox, a simple hole in the wall. Recall that one of the goals of making organizations into concrete material objects was to incorporate their causality into a naturalistic picture of the world. The location dilemma matters because of the concerns of methodological individualists, as laid out in Section 1.3. Those concerns were about how to incorporate organizations into the natural world in a way that is both non-redundant and nonmysterious. An important part of such an incorporation is giving organizations some sort of spatiotemporal base that allows the organization to have causal influence. When we are treating an organization as a wrongdoer, we furthermore want that causal influence to be agential. This kind of influence can arise only via (what Hindriks calls) organizations’ physical location. The institutional location is no help with making organizations into material objects that have causally efficacious agency. But if we are emphasizing an organization’s physical location—understood as the location of its members—then the questions that I raised for Ritchie need to be answered. I will answer those questions in Chapter 3.¹²

¹² In a recent footnote, Hindriks (2021, fn. 6) states that organizations go on holiday with their members, but are ‘asleep’ while there. That is, he grasps the first horn of the location dilemma. In Chapter 3, I will propose something similar, using the notion of ‘dormancy’ rather than sleep. I will also elaborate on what this means and why we should accept it. I thank Frank Hindriks for discussion.

48

  

2.5 Conclusion The results of this chapter have largely been negative: I have surveyed many relations that fail to fully capture organizational membership and that fail to explain how organizations are material objects. But each of these relations does capture something. Members are undoubtedly amongst the causes, subveners, anchors, and grounds of (various facts about) organizations. Members are simply much more than that. What’s more, each of these relations captures the ‘involvement’ that various entities have with organizations, where those entities are not members. Those entities will sometimes be blameworthy for that involvement. So these relations capture much that is important, but they don’t tell us what makes up organizations, as concrete material objects, and how members fit into that picture. The discussion of these failed options has revealed several desiderata on an account of organizations’ relation to their members. These desiderata will guide Chapter 3: Materiality. Our analysis of the organization–member relation should render organizations material objects locatable in the natural world. (Supervenience, causation, anchoring, and grounding fail at this.) Internality. Our analysis of the organization–member relation should render members internal to organizations (in a loose and intuitive sense of ‘internal’), such that we don’t count as a ‘member’ everyone who interacts with an organization. (Supervenience, causation, anchoring, and grounding fail at this.) Overlap. Our analysis of the organization–member relation should allow that distinct organizations can ‘overlap’, in sense of having exactly the same members. (As revealed by the discussion of the history club and the philosophy club.) Non-transitivity. Our analysis of the organization–member relation shouldn’t make membership transitive. (It should be possible that I’m a member of New Zealand, and that New Zealand is a member of the UN, even while I’m not a member of the UN.) Non-circularity. Our analysis of the organization–member relation shouldn’t define membership in terms of membership. (This avoids an uninformative circularity.) Agency. Our analysis of the organization–member relation should require that members are agents. (This enables us to exclude buildings, machines,

    

49

pieces of paper, and so on as members, even if such objects are parts of organizations.) Non-conventionality. Our analysis of membership should not be ‘whatever the law or social practices or organizational charter says membership is’. (This enables us to explain how these conventions can get membership wrong.)

In the next chapter, I lay out an account that satisfies all these desiderata. The result will be a picture of organizations that makes them material objects, with locations and physical matter, around which we can ‘draw lines’. These objects will be made of and dependent on (yet distinct from) members. They will be liable for analysis as wrongdoers in Part II of this book.

3 A Metaphysics of Membership 3.1 Introduction As mentioned in Chapter 2, my account of the organization–member relation is a ‘parthood-plus’ account: the members of an organization are material parts of that organization, but they are not just any old material parts. They are material parts that have certain special features. There are three special features: members are those material parts that have committed to abide by the organization’s decisions, that have inputs into the organizations’ decisions, and that have the ability to enact at least some of the organization’s decisions. Any entity that meets these three conditions is a ‘member’ of an organization. On this account, I am a member of my university, but my nose is not; New Zealand is a member of the UN, but I am not. However, my nose is a part of my university and I am a part of the UN. This is because, on the proposed view, the material parts of an organization include not just (i) members (i.e., entities with the three features just listed) but also (ii) parts of members. This allows us to retain much of the classical mereological account of parthood, such as parthood’s transitivity, while denying the transitivity of membership. As we’ll see, on my account, an organization is (partly) located wherever its members are located. The organization follows its members around, going to whatever location they go to. This applies even when members are ‘off the clock’, for example on holiday. When members are on holiday, the organization is partly present wherever the holidaying member is, but the organization is dormant at that location. I explain these and other upshots in Sections 3.2 and 3.3. In Sections 3.4 and 3.5, I apply my account of membership to two kinds of organization: corporations and liberal-democratic states. We get the result that directors are members of corporations, as are shareholders and many non-management employees. But intermediaries (for example, many brokers) are not members. Meanwhile, members of liberal-democratic states include all individuals in the executive, legislature, and judiciary, alongside all individuals who are eligible to (register to) vote (at not disproportionate Organizations as Wrongdoers: From Ontology to Morality. Stephanie Collins, Oxford University Press. © Stephanie Collins 2023. DOI: 10.1093/oso/9780192870438.003.0003

   

51

cost), and who enjoy other basic civil and political liberties, within the state. This conception of state membership excludes many permanent residents and others who are disenfranchised. As we’ll see in Part II, this attractively excludes powerless individuals from being counted as part of the entity that did wrong, in cases of organizational wrongdoing. And as Chapter 7 will explain, even the members should not bear costs equally when their organization does wrong. Less powerful members will often be justified in demanded a less weighty share of the costs of organizations’ wrongdoing, compared to more powerful members. Still, the present chapter will argue, more and less powerful members are all members. This means that they are the material basis for the organization’s presence in the world.

3.2 Members: Parts with Commitments, Inputs, and Enactment My proposal uses Kathrin Koslicki’s mereological analysis of constitution. Koslicki developed this analysis to deal with ordinary objects, such as motorcycles and tables. On Koslicki’s view, each object has both material components and formal components. Chapter 1 alluded to a similar view, when I endorsed Ritchie’s account of organizations as structures. Roughly, we can view Koslicki’s ‘formal components’ as analogous to Ritchie’s ‘structure’, and Koslicki’s ‘material components’ as analogous to Ritchie’s ‘realizers’. I take Koslicki’s view to be a more detailed picture that accords with the broad outline provided by Ritchie—and, to some extent, with Sheehy’s view discussed in Section 2.3. For my purposes, the important detail in Koslicki’s view is that an object’s formal components are themselves parts of the object. But they are not material parts. Instead, objects have nonmaterial parts. Koslicki summarizes her view as follows: Some objects, m₁, . . . , mn, constitute an object, O, just in case m1, . . . , mn are O’s material components, i.e., m₁, . . . , mn are those among O’s proper parts which satisfy the constraints dictated by O’s formal components, f₁, . . . , fn. (2008, 185, emphasis original)

Here, we see how Koslicki’s view combines constitution and parthood. Constitutors differ from parts in that parts can be either material or nonmaterial, while the constitutors must be material. The nonmaterial parts are

52

  

the structural components. The material parts are the things that realize the structure. Koslicki uses the analogy of a ‘recipe’ to explain the formal components: the formal components specify the arrangement, variety, and sometimes the number of the material components (constituents). Thus, the structure of a table or a motorcycle specifies the ways in which the material parts of that object must be arranged, in order for there to be a table or motorcycle. If the material parts are arranged in a way that does not accord with the structure, then we do not have a table or motorcycle. (Imagine a motorcycle disassembled on a garage floor.) The material parts must be arranged in conformity with the structural part, if the object is to be realized. On Koslicki’s picture, then, ‘constitution’ is a species of the genus that is ‘parthood’ or ‘composition’. The relation of constitution is ‘the relation which a whole bears to certain specific ones among its proper parts, viz., its material components; the relation of composition, on the other hand, holds more generally between a whole and all of its parts, including its formal components: all of its proper parts [material and formal] together compose a whole, but only its material parts constitute it’ (2008, 185, emphasis original). To keep the terminology consistent, I will usually talk about ‘material parts’ and ‘nonmaterial parts’, leaving the language of constitution and composition to one side. My proposal applies Koslicki’s analysis of ordinary objects to organizations, such that all and only the members (and the members’ parts) are the material parts of organizations, while the role structure and decision-making procedure are the organization’s nonmaterial parts. An organization’s nonmaterial parts differ somewhat from a motorcycle’s nonmaterial parts: a motorcycle’s structure specifies a spatial arrangement of the material parts, while an organization’s structure and decision-making procedure usually specify a functional arrangement of the material parts. That is, an organization’s material parts do not usually have to be situated in some particular spatial configuration (although sometimes they do, if the members’ functions require the use of certain tools or materials). Instead, an organization’s parts have to be situated in a functional relation to each other: they must have certain causal interactions with one another, rather than (necessarily) certain spatial relations to one another. Still, in the case of both motorcycles and organizations, the object’s nonmaterial part specifies the ‘arrangement’ of the material parts, in some loose sense of ‘arrangement’.¹ ¹ Asya Passinsky (2021) also draws on Koslicki to analyze social objects. Unlike my account, Passinsky’s (1) does not analyze ‘member’ (so does not give us the resources to ‘draw lines’

   

53

This proposal attractively gives an ontologically significant role to the organization’s structure and decision-making procedures. Such ontological significance is attractive for two reasons. First, an organization’s structures and procedures are important for capturing how the organization has downward and outward causal-explanatory powers, as outlined in Section 1.3. Second, by giving the structure ontological significance (by viewing it as a part of the organization), we can account for our appropriate responses to organizational wrongdoing: sometimes, we respond by demanding changes in the structure, not just changes in which individuals happen to realize that structure. For these two reasons, it’s important that organizations are not just their members. Organizations are their members structured in a certain way, that is, structured according to the formal components—which includes the organization’s role chart and decisionmaking procedure. Furthermore, by applying Koslicki to organizations, we are already able to satisfy three of the seven desiderata that were laid out at the end of Chapter 2: Materiality, Internality, and Overlap. Materiality required that we analyse the organization–member relation so as to render organizations material objects locatable in the natural world. If wholes are made up of material parts, then wholes are rendered (at least partly) material. This allows us to locate organizations in the world: the organization is wherever its members are. Internality said our analysis should render members internal to organizations (in a loose and intuitive sense of ‘internal’). Parthood is perhaps the paradigmatic internal relation: parts are internal to their wholes. So, by applying Koslicki to organizations, we satisfy Internality. (That said, one idea behind Internality was that we shouldn’t count as a ‘member’ absolutely everyone who interacts with an organization. This idea will be accommodated below, where I give three conditions for being a member of an organization.) Finally, Overlap required that we should allow distinct organizations to ‘overlap’, in the sense of having exactly the same members. Koslicki’s approach satisfies this, since organizations with exactly the same members are distinct via their nonmaterial, formal parts.² We can therefore somewhat endorse classical parthood, on which two objects that share all the same parts are identical. But we add to around all organizations) and (2) renders all formal components of social objects ‘normative’ (which I believe is too restrictive). Additionally, she does not locate organizations where their members are (2021, 14). ² Here, my view is similar to that of Sheehy, as described in Chapter 2, Section 2.3. Soon, I will add substantive conditions on membership that take my account beyond Sheehy’s.

54

  

classical parthood that an organization’s structure is a (nonmaterial) part of it, such that two organizations that have all the same members (and therefore all the same material parts) do not have all the same parts in total. The remaining four desiderata are not straightforwardly satisfied simply by using Koslicki’s metaphysics. These desiderata require us to produce a theory on which membership is non-transitive, non-circular, agential, and non-conventional. These desiderata require not just that we distinguish an organization’s material parts from its nonmaterial parts, but that we explain what it takes for a material object to be material part of an organization. Only then will we be able to draw lines around each organization in our world. I will propose three conditions that a material object must meet to count as a member of an organization. I will then suggest that all material parts are either members or material parts of members. At that point, we will be able to pick out everything that is a material part of an organization. I propose the conditions as individually necessary and jointly sufficient for membership, though I accept that the three conditions admit shades of grey and that there may be vague or borderline cases. My argument for these three conditions consists in the fact that they allow us to satisfy all seven desiderata.³ First, there is the commitment condition: to be a member, an entity⁴ must be psychologically committed (even if only tacitly and pro tanto) to abide by the results of the organization’s decision-making procedure, if and when these results are relevant to her actions. This commitment is usually overridable, but is nonetheless presumptively decisive in the entity’s decisionmaking. To use Australia as an example: the average citizen’s behaviour largely reflects a commitment to abide by Australia’s decisions as those decisions bear on the citizen’s behaviour. This manifests as a commitment to abide by Australia’s laws. This doesn’t entail that the citizens together have ‘shared intentions’, ‘joint commitments’, or any other kind of proorganization mental states. For example, an Australian citizen might be committed to the law simply because she’s incentivized by the threat of punishment. And, of course, citizens sometimes do break laws. But as long

³ I originally proposed these conditions for membership in my account of collective agents in Collins 2019a, but without the connection to mereology (or metaphysics more generally). I thank Andy McKilliam for pressing the possibility of vague boundaries. ⁴ I use ‘entity’ to avoid the circularity of using ‘member’ (which would define ‘membership’ in terms of ‘member’), to allow that some members are collective agents rather than individual agents, and so as not to pre-judge the question of whether my account of membership will satisfy the Agency desideratum.

   

55

as occasional law-breaking is the overriding of a general presumption of abidance on a member’s part, the member is committed. It would take a particularly staunch and isolationist kind of anarchism for a democratic citizen to fail to meet the commitment condition. What’s more, the commitment might be forced. It need not be free from duress, coercion, social pressure, and so on. The commitment condition does not require that abidance by the organization’s decisions is not too costly, or that the entity would remain committed even if abidance were not too costly.⁵ The commitment condition requires only that the member is committed, given the actual situation she finds herself in. This might seem morally problematic: are coerced members blameworthy for their organizations’ wrongs? No. They are part of an entity that is blameworthy, but this does not mean they are blameworthy in their private (non-organizational) capacity. Part III will spell out the implications of organizations’ wrongdoing for members, to produce morally palatable implications even for coerced members. The accommodation of coerced membership is important for capturing how many organizations actually operate: if coerced members were excluded, then we would have a hard time accounting for how many modern-day corporations operate. When a coerced worker performs an organizational task, that performance should be attributed to the organization. This is especially true if the performance is wrongful: the organization should be blamed for the member’s coerced performance, which means that the performance should be attributed to the organization. This means we should view the coerced performance as an instance of the organization materially operating in the world, and so we should view the coerced member as a material part of the organization. Yet this does not automatically place the coerced worker themselves on the moral hook for that particular action, as Part III will explain. In general, there is a distinction between (i) whether one is agentially bound up with an organization when that organization is considered as an entity independent of any particular action, that is, when it is considered as an entity that generally operates in the world and (ii) whether one is blameworthy for one’s involvement in a

⁵ In this, my account of democratic citizens’ commitments differs from that of Avia Pasternak (2021), according to whom ‘genuine intentional citizens’ are those who participate in their state and who would continue to participate even if the costs of refraining from participating were not unreasonable.

56

  

particular organizational act. The present conditions relate to (i), while Part III will relate to (ii). Second, there is the inputs condition on membership: for an entity to be a member of an organization, the organization must be systematically and intentionally set up such that: the beliefs and preferences that the organization’s procedure aims to take as inputs, and the way the organization aims to process those inputs to form organization-level decisions, can be influenced by the behaviour (e.g., deference, votes, meeting contributions, etc.) of the entity, even while the organization’s inputs and processes are operationally distinct from the inputs and processes that the entity uses when deciding for herself. These aspects are ‘operationally distinct’ insofar as: first, the organization has the capacity to take in different considerations when forming its decisions than each member takes in when forming her own decisions; second, its method for processing those considerations is different from the method of any one member when deciding for herself; and third, the organization’s decisions are not the straightforward conjunction of members’ decisions. This second condition makes the organization’s agency distinct, while still reliant on the agency of members. To again use Australia as an example: Australia is systematically and intentionally set up such that its beliefs, preferences, and decisions are influenceable by the manifestos of political parties, the directives of ministers, the judgements of juries, and so on. The first two of these (at least) are systematically and intentionally influenceable by the votes, letters, petitions, and so on of ordinary citizens. So, Australia’s decisions—including the inputs into those decisions and the processes by which those decisions are reached—are systematically and intentionally influenceable by the behaviour of all of these entities. What matters for the inputs condition is that an entity has the capacity to behave in ways that systematically and intentionally influence the organization’s decisions. She need not exercise this capacity. This is just as it’s my capacity for agency that makes me an individual agent, even if I never exercise that capacity. At the same time, Australia’s beliefs, desires, and decisions are operationally distinct from those of members, in three ways. First, when arriving at its beliefs, desires, and decisions, Australia takes in different inputs than any given citizen (legislator, executive, judge) does when arriving at her own beliefs, desires, and decisions (for example, because Australia aims to take in the inputs of all voters). Second, Australia’s methods for processing these inputs to arrive at beliefs, desires, and decisions include parliamentary debates, Cabinet meetings, and so on—these are decidedly not

   

57

the procedures that members use when processing their own beliefs and preferences to make their personal decisions. Third, the decisions of Australia are not just ‘the Prime Minister decides this and the Treasurer decides that and . . .’ Instead, Australia resolves differences in members’ decisions to arrive at an organizational position. Thus, the organization’s procedure is operationally distinct from those of members. Third and finally, there is the enactment condition on membership: for an entity to be a member, it must be that the enactment of some of the organization’s feasible decisions requires actions from that entity, where those actions are attributable to the organization. When a member enacts the organization’s decision, they are acting as a constituent, organ, or part of the organization. Without referring to the organization, we cannot locate the decision-maker that is behind the member’s action. To be clear, this condition requires that each member has open to her at least one feasible action which is such that: if the member performed that action, then that action would be properly attributable to the organization, as well as (perhaps) attributable to the member. As applied to citizens of liberal-democratic states, the most prominent such actions include voting in a new government or financing social services. Both of these actions are attributable to the state yet enacted by ordinary citizens. With the commitment, inputs, and enactment conditions on table, we can satisfy the remaining four desiderata. First, these conditions render the analysis non-transitive. The members are the parts who meet the commitment, inputs, and enactment conditions. And the members are parts in virtue of meeting these three conditions. But the members’ parts might not meet the three conditions. Thus, organizations do have some material parts—namely, some of members’ material parts—that are not members. Consider me, New Zealand, and the UN. I can be committed to abide by New Zealand’s decisions, but not committed to abide by the UN’s decisions. New Zealand can systematically and intentionally make decisions in ways that derive from my actions, without the UN systematically and intentionally making decisions in ways that derive from my actions. This can be true even if the UN does systematically and intentionally make decisions in ways that derive from New Zealand’s actions: the UN might be indifferent to whether or not New Zealand takes me into account (which it is: the UN does not require New Zealand to be a democracy). And some of New Zealand’s feasible decisions might require actions from me (for example, actions of paying tax), without any of the UN’s feasible decisions requiring any actions from me.

58

  

Second, the conditions are non-circular: they do not define membership in terms of membership. Each of the three conditions is stated such that any entity can satisfy the condition. The conditions themselves do not refer to a pre-existing notion of membership. Third, the only kinds of entities that can satisfy the three conditions are agents. This is because the conditions require psychological commitments (in the commitment condition) and actions (in the inputs and enactment conditions)—which, I assume, can be had only by agents. Fourth and finally, the three conditions are non-conventional: they do not defer to laws, social practices, or organizations’ charters. Of course, an entity can satisfy the three conditions only if an organization’s practices and charters are a certain way (namely, a way that is appropriately related to her commitment, inputs, and enactments). But laws’, practices’, and charters’ understandings of membership might depart dramatically from what is laid out in the three conditions. The three conditions explain how laws, practices, and charters can be wrong about who is a member. Laws, practices, and charters get it wrong if they attribute membership to agents that do not meet all three conditions, or if they fail to attribute membership to agents that do meet all three conditions. A final point. On my view, all material parts are either members or material parts of members. One might worry that this ignores other parts of organizations: parts like buildings, machines, and paper. But when we consider organizations as agents who can do wrong, it seems that this more expansive view includes too many things as the parts of organizations. Organizations do not automatically do wrong whenever their buildings cause harm. Instead, their buildings’ harms amount to organizational wrongdoing only if the buildings’ harms are appropriately connected (e.g., via intent, recklessness, negligence, etc.) to organizations’ structures, procedures, or members. If we included buildings, machines, paper, and so on as amongst the parts of organizations when considering organizations as agents of wrongdoing, then it seems we would be conflating the parts of an organization with its property. This is why—when it comes to the specific purpose of picking out organizations as wrongdoers—I think it is a mistake to include such non-agential materials amongst organizations’ material parts. Instead, I propose viewing an organization’s buildings and so on in the following way: these objects stand to the organization as my house, car, and so on stand to me. These objects are something over which the organization has property rights, but the objects not a part of the organization when we conceive of the organization as a material agent capable of wrongdoing. To be sure, an organization can be blameworthy when its buildings cause harm,

   

59

just as I can be blameworthy when my house causes harm. But to include the organization’s building (or my house) within the agent that is blameworthy is, I think, to misconstrue the material base of the agent. When my house harms someone, we turn to blame me. When an organization’s building harms someone, we turn to blame the organization. The house and the building are not included within the blameworthy agent. At least, this is the view I will assume. But it is not essential for the arguments that follow. If the reader wishes to include buildings and so on as amongst material parts of organizations, then the arguments of Parts II and III will go ahead regardless. What really matters for what follows is that material parts of organizations include all those entities who meet the commitments, inputs, and enactment conditions. If other non-agential entities are also included, the upcoming arguments are unaffected.⁶

3.3 Objections and Replies At this point, one might object that my account fails to satisfactorily resolve the location dilemma. Recall the dilemma: on the first horn, organizations are wherever their members are, including when members are ‘off the clock’ (i.e., not playing their organizational roles). This means organizations follow their members to members’ holiday destinations, weekend activities, and bedrooms. On the second horn, organizations are wherever their members are, but only when members are playing their organizational roles. This means organizations lose their material existence when members are off the clock. Organizations pop in and out of material existence with members’ working hours. Neither horn is entirely comfortable and each has something going for it. Let us consider the second horn first. There is perhaps something intuitive about the idea that organizations have no spatio-temporal location at some times, such as when all members are on leave or asleep. On this kind of view, an organization would be materially composed not of members, but rather of stages of members. At any given moment, we would have to look out into the world, see which members are playing their organizational role, and locate the organization wherever those members are. The organization would be materially composed of any given member only for as long as

⁶ I thank Andy McKilliam for pressing the possibility of non-member components.

60

  

the ‘stage’ (temporal duration) during which that member was playing her organizational role. This might seem particularly attractive for sporadic and secretive organizations, such as an underground gambling organization that meets just once a month. It might seem correct to say that the gambling organization is nowhere in between its meetings.⁷ This option worries me mainly because of its implications for holding organizations accountable. Suppose the underground gambling organization arranges for the assassination of the family of someone who owes it money. It arranges for the assassination during one of its monthly meetings, by paying off an external hitman. The meeting then disbands. The assassination happens a week later. But the assassination doesn’t go to plan: there is a witness, or a passer-by is caught in the crossfire. As a result, the police and various vigilantes start trying to pin down the organization. So, each member of the organization individually and privately decides never to go back to another meeting. As a result, the organization is never again materially constituted—never again has a material presence. What can be said about the agent to which the police and others can take their claims? The police can of course go after the members in their capacity as private individuals. But we could set up the case such that no individual member is all-things-considered blameworthy for their role in hiring the assassin. The possibility of such ‘gaps’ are one of the key insights of recent work in collective responsibility (see, e.g., French 1984; Copp 2007; Pettit 2007; Collins 2019b). It might be that the organization is the only agent that can appropriately be held to account. To get our hands on the organization— including its money for financial recompense, its spokespersons for apologies, and its charter for restructuring—we often must target the members qua members, not qua private individuals. Although the organization is dormant at the locations of off-duty members, it should still be possible to causally interact with the organization by interacting with those off-duty members, for example by compelling those members to take up their roles and thereby awaken the organization from its slumber.⁸ Our practices of holding organizations blameworthy are safer if organizations persist materially even while members are off-duty. Does this make our metaphysics problematically beholden to our moral commitments? I don’t think so. The goal of this book is to provide a conceptualization of organizations that enables us to fit them into our moral, social, and political practices of responding to wrongdoing. The ⁷ I thank David Braddon-Mitchell for pressing this. ⁸ I thank Sam Baron for questioning interactions with dormant organizations.

   

61

aim is to provide a kind of ‘possibility proof ’ for such a conceptualization. It can be no objection to a possibility proof that it is designed specifically to make possible what it aims to make possible. For these reasons, I advocate the first horn of the dilemma: organizations follow members wherever members go, including to holidays, weekend activities, and bedrooms. The organization follows all and only the entities that meet the three membership conditions. Two points make this horn more attractive than it might appear. First, we should distinguish between something’s fully versus partly present. On my view, my university is partly present in Fiji when I go on holiday to Fiji—but only partly, since not all of its material parts are in Fiji. Second, we can distinguish between something’s being present versus active. Even though my university is (partly) present in Fiji, it is not active there (not even partly). This is because I do not play my role while I am in Fiji. We can phrase this by saying the organization is ‘partly present but dormant’ in those locales where some (but not all) members are present but not performing their organizational roles. An organization can also be ‘fully present but dormant’, if all members are at some locale but none of them are performing their roles. For an organization to be fully present and active in any locale, it needs to be that all members are present there and that all members are performing their roles. It is thus a rarity that an organization is fully present and active anywhere, since it will almost always be the case that at least some members are in different locales from others. The most common situation is one in which an organization is partly present and active in a few locales, and partly present but dormant in several other locales. We can capture all these different circumstances by talking about organizations being more or less present and more or less active in various locales. An organization is more present somewhere the more members are present there, and it is more active there the more the present members are performing their organizational roles while there. Crucially, we should talk not just about whether the organization is present (and to what extent it is present, that is, what proportion of members are present), but also about whether the organization is active as compared with dormant (that is, how many of the present members are performing their roles). On this view, then, organizational presence and activity is not an ‘all-or-nothing’ affair. These points should not strike us as strange. Analogous points can be made about other material objects, such as humans.⁹ Consider the question:

⁹ I thank David Strohmaier for questioning this.

62

  

am I under my desk as I type this paragraph? Parts of me are: my legs and feet. Parts of me are not: my torso, arms, and head. It would not be felicitous to say that I am under my desk. But I am partly under my desk. Likewise: am I physically active as I type this paragraph? My fingers are active, typing on the keyboard. But my legs and feet are relatively inactive. They are relatively dormant. So, I am partly under my desk and I am fairly dormant insofar as I’m under the desk. The relative dormancy of my under-the-desk parts likely explains why it is infelicitous to say that I am under my desk. Something similar applies to my organization when I go on holiday to Fiji. My university is partly in Fiji, but it’s dormant insofar as it’s in Fiji. So, it’s not felicitous to say that my organization is in Fiji. The under-the-table analogy softens the strangeness of saying that my university goes to Fiji with me.¹⁰ The concepts of ‘more or less present’ and ‘more or less dormant’ also help us to account for overlapping organizations. Recall the history club and the philosophy club, who have all the same material parts but different formal parts. Suppose all members agree to attend the annual national history convention in order to promote the history club’s activities. Is the history club present at the convention? Intuitively, yes. If a few members were sick, we’d say that most of the club was present, or that the club was mostly present. My proposal agrees with these intuitive verdicts. Is the philosophy club present at the history convention? My account says yes: the philosophy club is present (maybe wholly materially present, if all members are there). It’s simply not at all active. I have two things to say to make this result more palatable. First, the material parts of the philosophy club are present, but its nonmaterial (formal) parts are not present. So, the philosophy club is not entirely present, even though it is entirely materially ¹⁰ I’m not the first to accept that organizations might be present on their members’ holidays. This is also a result of Hess’s (2018a) view of organizations, which she presents through the metaphysics of Lynne Rudder Baker and of Peter van Inwagen (without explicitly stating whether the relation is parthood or not). Hess sidesteps the holidays issue by noting that the organization acts only ‘when the members act from their positions in the corporate agent on the basis of corporate commitments’ (2018a, 45). This focus on action suppresses the distinct issue of location. As Hess admits in a footnote, her account of action prevents my university from performing the action of going to Fiji, but it doesn’t prevent my university from being present in Fiji. In response to this concern, Hess raises two options: either we can ‘resist the suggestion that this is implausible’ or we can ‘suggest that [the organization] is present only when the members are acting from the [organization’s rational point of view]’ (Hess 2018a, 55). She doesn’t pick between these options. Neither hits the mark: the second option allows the organization to pop in and out of material existence as the work day starts and finishes. The first option seems to fly in the face of the reflective equilibrium methodology. My proposal of organizations being partly present, and/or being present-but-dormant, brings the result into better reflective equilibrium with our other intuitions, without denying its strangeness.

   

63

present. The philosophy club is not formally present. (It is analogous to a motorbike that is disassembled on the garage floor.) Meanwhile, the history club is entirely present both materially and formally. So my account does allow us to distinguish between the two clubs: one is formally absent; the other is formally present. Second, if it still seems odd that philosophy club is entirely materially present (even having noted that it’s not at all formally present), we should note that the philosophy club’s material parts are dormant rather than active. This is precisely because of the absence of the philosophy club’s formal parts. The location dilemma is just one objection to my view. Another objection was stated by David-Hillel Ruben as early as 1983. To introduce the objection, consider the following example: Imagine a tour of a country with no national Red Cross (Albania, perhaps) by a group of individuals who stand in any relation whatever to the International Red Cross—the Red Cross workers, or officials, or executives, etc. Let the tour even be in the official Red Cross capacity of those individuals. Locating those individuals in that country for however long a time does not bring it about that the Red Cross is now located there. (Ruben 1983, 225)

Contra Ruben, on my view, the Red Cross is (partly and actively) located in Albania for the duration of the tour. In defence of this, observe that we would readily say ‘the Red Cross visited Albania’ or ‘the Red Cross is touring Albania’. We could imagine an Albanian official asking his aide ‘I’m meant to meet the Red Cross now. Where are they?’ And the aide might answer ‘Here’s the Red Cross now’ as the tour party walks through the door. All of this implies that the Red Cross is in Albania. Now, perhaps we would refrain from saying ‘the Red Cross is located in Albania’. But this is simply because ‘located’ conversationally implies a fixed, stable, or long-term location. Saying that an organization is ‘located’ somewhere is a bit like saying an individual ‘lives’ somewhere. When I go on holiday to Fiji, we would not say I am ‘living’ in Fiji. But we would say I’m locatable there. Likewise, we might not say the Red Cross is ‘located’ in Albania. But we would—should—say it is locatable there. Conversely, Ruben argues not just that members can be where the organization is not, but also that the organization can be where the members are not. He writes: ‘It is possible for an organization to have an inactive branch or affiliate at some place, and hence be at that place, where it has no

64

  

members at all’ (1983, 225). Similarly to Hindriks (discussed in Chapter 2), Ruben aims to make something’s address into its location. He suggests that our thoughts about the Red Cross mandate such a move. But we should resist this move. We can say that I have an address at some place, without that place ever being where I am. For example, if I use an address for tax purposes but never go there, then I am not located there. Likewise for office buildings at which organizations have no members.¹¹ So members’ and organizations’ locations do not come apart in the way Ruben suggests. Ruben provides yet another objection to my view, this time drawing on overlapping organizations. He imagines France’s territory completely overlaps with the territory of a diocese of the Catholic Church. He correctly notes that this wouldn’t cause the two organizations to ‘compete for room’. He concludes from this that organizations are not composed of material parts at all (whether those parts are members or anything else). This is because two objects that had all the same material parts would, he thinks, ‘compete for room’. Once again, on my view, this point confuses an entity’s parts with its property. On my view, neither France nor the Catholic Church have geographic territory as parts. Those organizations have authority or property rights over geographic territories, while having members (and members’ parts) as parts. France and Catholic Church would, presumably, compete for authority or property rights over the territory, in Ruben’s thought experiment. So they would compete. But they would not compete ‘for room’, because they are materially composed of members, not of territory. Furthermore, on my view, unlike on Ruben’s, two overlapping organizations need not even ‘compete for room’ over their material parts (members). Thus, if the reader believes that geographical territory is a material part of France or the Catholic Church, this is in no deep conflict with my view. A ‘competition for room’ can be avoided by judicious specification of the respective organizations’ formal parts. These formal parts dictate the conditions under which some material part (whether member, territory, building, or whatever) is to be considered under the auspices of the organization. When we consider the members of the history club and philosophy club, we see that these members are active for the history club when they play their role in that organization, and they’re active for the philosophy club when

¹¹ Note that this is a different point from the one about whether buildings, machines, and so on should be counted amongst an organization’s material parts. That debate is most plausibly construed as a debate about the buildings, machines, and so on that are used by members. In contrast, Ruben’s point is about entirely unused buildings.

   

65

they play their role in that organization. The clubs’ respective formal parts lay down the rules for when the agents are under their respective auspices, so as to avoid competing over members’ bodies. Likewise, the Catholic Church and France has jurisdiction over different goings-on within the territory, thereby each specifying in what respects the territory should be considered under the auspices of each organization. Of course, organizations’ structures do not always specify this perfectly, as conflicts-of-interest attest: sometimes, a member is pulled into acting for two competing organizations at once, and sometimes one aspect of a territory is claimed for jurisdiction by two organizations. But in those cases, the organizations do indeed compete for room, contra Ruben. In saying all this, I should admit that my account bites a significant bullet: an organization is partly and dormantly located wherever its members are, even if those members are performing their roles in another organization or are on holiday. I accept that this result is a cost in the reflective equilibrium. But it is a smaller cost than those paid by the alternatives. It is also a smaller cost than is faced by Ruben’s own view, which is that ‘some social entities’ are ‘spatially locatable non-material entities’ (Ruben 1983, 233). He thinks these entities are spatially locatable because he locates them at the territory over which they have rights. Yet they are non-material. This makes organizations’ agency difficult to integrate into naturalist ontology, since it is clear that a mere piece of territory does not act as an agent, even if one thinks it is a non-agential part of an organization. To further demonstrate that my account gives the right results, I will round off this chapter by applying my account of membership to the two organization-types mentioned in Chapter 1: corporations (with a focus on the Banking Royal Commission) and liberal-democratic states (with a focus on the Polluter Pays Principle).

3.4 Members of Corporations Talk of the ‘members’ of corporations is bound to be controversial. After all, individuals tend to find corporations alienating—more so than clubs, nonprofits, or democratic states. Few individuals feel the warm glow of ‘membership’ when it comes to the corporations in our lives. This is true not just for low-ranking employees when considering the corporation that employs them, but also for corporations’ shareholders and directors, who are almost always ‘in it for the money’ and may feel personal disinterest or even

66

  

antagonism towards the goals of the organization through which they make money. And all this is true before we’ve considered the contractors, casuals, and customers that are more tenuously connected to organizations. A related issue arises in economic theories of the firm. Those theories seek to distinguish between activities that occur within the boundaries of firms and activities that occur outside the boundaries of firms. The question for those theories, in effect, is where to draw lines around the firm for economic purposes (e.g., Coase 1937; Holmstrom and Roberts 1998). My approach is different: I seek to characterize all organizations (not just firms) and for the purpose of attributing wrongdoing (rather than economic activity). It’s possible that a focus on economic activity—rather than wrongdoing, or agency more generally—would produce a different account of membership than the one I presented in Section 3.2. My aim here is to apply my account of membership to corporations, for the purpose of picking out the organizational agent of wrongdoing. My question is therefore: just how low-ranking, alienated, short-term, or uncommitted might an individual be in relation to a corporation, while satisfying the three conditions I’ve proposed for being a member of it? There’s no way to put a numerical requirement on this. And, as I noted above, all the conditions admit of gradations and borderline cases, with the result that there are sometimes borderline cases of members. It’s best to proceed with illustrative examples. To keep the theory somewhat grounded in practice, I’ll take my examples from Australia’s Banking Royal Commission. As mentioned in Chapter 1, the Banking Royal Commission was set up to inquire whether there was misconduct, or conduct falling short of community standards and expectations, in Australian financial services entities. Its focus was primarily on organizations themselves—and organizations’ cultures— rather than particular categories of individual. Nonetheless, certain rolebearers were inevitably discussed at some length in the Commission’s reports.

3.4.1 Directors Predictably, directors were singled out. The Commission went so far as to say: ‘[t]here can be no doubt that the primary responsibility for misconduct in the financial services industry lies with the entities concerned [i.e. the organizations] and those who managed and controlled those entities: their boards and senior management’ (2019, 4, emphasis added). It went on: ‘[d]irectors must exercise their powers and discharge their duties in good

   

67

faith in the best interests of the corporation, and for a proper purpose. That is, it is the corporation that is the focus of their duties. And that demands consideration of more than the financial returns that will be available to shareholders in any particular period’ (2019, 402, emphasis original). The Commission implied that these duties were not always fulfilled in the entities it investigated, with some directors instead giving priority to ‘whatever person or body may have nominated the director to serve in that office’ (2019, 244). These comments imply that directors are at the heart of organizations. After all, the Commission attributes to directors obligations to focus squarely on the corporation’s interests, as distinct from the interests of shareholders or even directors themselves. But are directors material parts of a corporation? My account of membership vindicates the common-sense view that directors are indeed material parts of corporations. First, directors are committed to abide by corporations’ decisions—both contractually and as a matter of practice, since they oversee the implementation of those decisions. Second, directors obviously have inputs into corporations’ decisions: their main role is that of deciding a corporation’s overall direction. And third, some actions of directors should be attributed to the organization: actions like firing an executive or revising shareholder dividends. Corporate directors are a clear case of ‘members’ in my sense. One might, however, doubt that directors meet the commitment condition. Those at the top of an organization sometimes view themselves as above the structure, in the sense of not being beholden to the internal norms or regulations. Here we can imagine directors that misuse company funds for personal aims and projects, despite implementing strict policies that prevent those lower down from likewise misusing funds. Do such ‘rogue’ directors really have a commitment to abide by the organization’s decisions? I believe the answer is ‘yes,’ even though this pro tanto commitment is often overridden by self-interest in the directors’ all-things-considered reasoning. We can see their commitment by asking: if the director were pressed (say, by a legal regulator), would the director try to justify their decisions as somehow being in accordance with corporate policy? Does the corporate policy figure as a disincentive, or a reason against, the director’s nefarious practices—even if that disincentive or reason is ultimately defeated in the director’s reasoning? If so, the director does hold a pro tanto and tacit commitment. Furthermore, when we consider whether the director is committed to the ‘outputs of the organization’s decision-making procedure’, we should

68

  

consider not just the organization’s official policies, but also its informal norms, practices, and expectations. These implicit and diffuse aspects of a corporation often do permit those at the top to ‘get away’ with more than those lower down can get away with. And perhaps the nefarious director is indeed committed to the informal norms, practices, and expectations, where these permit the director’s misuse of funds, even though the director is not committed to the official policy. The de facto decisions of an organization do not always reflect the decisions as written in corporate policy. The director might be committed to abide by the former but not the latter; this is sufficient for meeting the commitment condition. But what if the director pays absolutely no mind to corporate policy, whether de facto or de jure? In that case, we should accept that the director is indeed more like an external agent that wields power over the organization, rather than an internal part of the organization. When it comes to wrongdoing, this is to say that the organization itself should not be held blameworthy for what this rogue director does. Instead, the rogue director should be held blameworthy in their own right, including for the effects they have on the organization (here, thinking of the organization as an agent entirely external to the director). Perhaps the organization has an obligation to try to ‘reign in’ the director—and the organization itself would be blameworthy if it did not perform that obligation. And perhaps the director has an obligation to become committed to the organization’s policies. But if the director and the corporation are really operating to such entirely different agendas (including when we take into mind the informal aspects of the corporation’s culture, and ask how the director might try to justify their actions), then we should view them as agents external to each other.

3.4.2 Non-Manager Employees Low-ranking employees are a second example. Compared with directors, low-ranking employees have less decision-making power and therefore might be expected to be more alienated from the corporation. In the context of Australia’s Banking Royal Commission, financial advisors are a good test case. Advisors are often not managers, meaning they lack decision-making power over others’ contributions to corporate ends. Yet they sat at the heart of the Commission’s concerns, primarily because they were often remunerated based on how much profit they brought in. This led to a conflict: advisors’ role was to advise customers on which product best fit the

   

69

customers’ needs, yet advisors were rewarded for selling customers products that lead to high profits for the corporation. The Commission found that this incentive structure led, amongst other things, to advisors doing inadequate research, giving clearly bad advice, and claiming to have given advice that was never given. The Commission treated advisors as internal to the corporation. This treatment is demonstrated in the Commission’s use of the concept of ‘vertical integration’: a situation in which banks are at once the producers, retailers, and advisors on their financial products. The Commission noted that this ‘ “one stop shop” model creates a bias towards promoting the owner’s products above others, even where they may not be ideal for the customer’ (Royal Commission 2019, 126). Notice the Commission’s assumption here: vertical integration creates a conflict of interest for the corporation. This assumption demonstrates the Commission’s presupposition that advisors are internal to the corporation. This picture of advisors as internal is only partly vindicated by my account. Advisors are committed to abide by the corporation’s decisions: this is a matter of their employment contracts. And they act as the corporation: their actions of advising customers are the corporation’s actions of advising customers. But do they have inputs into the corporation’s decisions? Here we uncover some ambiguities in my account. First, there’s a trivial sense in which all agents who can act as the corporation have inputs into corporate decision-making: when such an agent decides to perform a corporate action quickly rather than slowly (or vice versa), then this can be viewed as an ‘input’ into—i.e., a cause of—the corporation’s subsequent decision to perform some different action sooner rather than later. We can resolve this by saying that a member’s input must not just systematically and intentionally cause an organization’s decision, but that a member’s input must inform the organization’s later decision, in the sense of being explicitly weighed up by the organization. However, there’s an even more trivial sense in which all agents who can act as the corporation have inputs into corporate decision-making: the agent’s decision to act quickly rather than slowly can be viewed as an input into—because fully constitutive of—the corporation’s decision to act quickly rather than slowly. Does the enactment condition thereby render the inputs condition redundant? I don’t think so. We can resolve this by saying that the member’s input must be non-overlapping with the decision into which it is an input—that is, an ‘input’ to a decision cannot fully constitute the decision to which it is

70

  

an input. The question then becomes: to what extent do non-manager employees have open to them actions, where corporate decisions are systematically and intentionally set up to be influenced by those actions, and where the corporate decision isn’t constituted by the employee’s decision to perform the action? The question invites two sub-questions. First, how much discretion does the employee have to plan how they will perform their role in the future? If they have discretion now that influences how they do their job in the future, where the corporation systematically and intentionally gives them this discretion, then they have appropriate inputs into later corporate decisions (where those later corporate decisions are made by the employee themselves). Second, to what extent is the employee consulted by managers about decisions the managers make? If employees influence managers’ decisions, where the corporation systematically and intentionally gives them this influence, then the employee again meets the inputs condition. In this way, employees might have inputs at two places: employees might have inputs into decisions the employee themselves will make in the future, and they might have inputs into decisions managers make in the present or the future. The result is that relatively autonomous and consulted employees do count as corporate members, while relatively controlled and ignored employees do not count as corporate members. Membership status will differ for different employees of different corporations. Interestingly, this also opens up space for a graded notion of membership. So far, I have been working with a binary notion of membership: an entity either meets the commitment, inputs, and enactment conditions, or it does not. And I believe this is indeed the right way to think about organizations for the purpose of picking them out as material objects, notwithstanding that there can be borderline cases. But once we have established which entities satisfy the binary notion of membership, we can then start grading those entities. (This is akin to giving grades to all students who pass a test.) Entities that have more inputs can be considered fuller members or more central members—they can be considered to have a higher ‘grade’ of membership.¹² The higher a member’s grade, the more likely they are to be implicated in the organization’s wrongdoings (Chapter 6) and the more likely they are to be an appropriate bearer of reparative burdens (Chapter 7).

¹² I thank Glenda Satne and Heikki Ikäheimo for pressing degrees of membership.

   

71

What about employees who do not have sufficient inputs to count as members? My account might seem to downplay the agency and rights of these employees. But in fact it’s the opposite: by treating these employees as non-members, we shield them from the implications that corporate wrongdoing has for members. (Of course, such controlled and ignored and employees can still be individual wrongdoers, in their own right, and they still have claims against the organization of which they are not a member.) I will lay out these implications in Part III, especially in Chapter 7. We therefore protect non-member employees’ agency and rights, by ensuring that they are not adversely affected by the wrongdoing of corporations from which they are alienated. Conversely, if some non-manager employees are members of the corporations for which they work, then this has important practical implications for the organization. For example, the Australian Securities and Investment Commission (ASIC) has the task of penalizing banks and financial services entities that break the law. According to the Banking Royal Commission, ASIC’s criminal prosecutions have almost always been brought against individuals within organizations—not against organizations themselves (2018, 271, 275–6). From 2008 to 2018, ASIC banned 229 individuals from providing certain financial services (2018, 148). Contrast this with a mere six civil penalty proceedings against financial service organizations themselves during the same period (2018, 275–6). If many of these individuals are internal material parts of the organizations for which they work, and if they commit offences while performing their organizational role, then organizations’ role in these ‘individual’ offences should also be examined. In Australia at least, it seems there is a habit of treating non-manager employees as individual wrongdoers that are separate from the organization. If my account of membership is correct, then this is a myopic approach that should be broadened to include the corporation in our full account of members’ wrongdoing. (At least, we should do this if organizations can be wrongdoers in their own right—a conclusion for which I will argue in Part II.)

3.4.3 Shareholders A third example of corporate membership is shareholders. In the common imagination, shareholders are entirely distinct from the corporations that they own: corporations’ primary legal obligation is to pursue profit for the

72

  

good of shareholders, but this is an obligation held by one party to another, not by one party to a material part of itself (see, e.g., Bakan 2004, 6). The Banking Royal Commission mentioned shareholders several times. Shareholders were seen as potential sources of conflicting obligations for corporations, since corporations or their directors may have obligations to shareholders as well as to customers. Most prominently, organizations’ obligation to secure immediate shareholder profit might compete with organizations’ obligation of risk management and long-term customer care (Royal Commission 2019, 228, 235, 243, 246, 353, 355–7). Shareholders were also mentioned as possible bearers of obligations to corporations’ clients—but immediately upon suggesting this, the Commission suggested that asserting such obligations would make no difference to client satisfaction (2019, 236). Finally, shareholders were mentioned as agents that might intervene to prevent organizations from performing their obligations to customers or to wider society in the long-term (2019, 247, 356, 402). Again, such intervention was implicitly framed as one agent acting upon another, not one agent acting upon a broader entity of which the intervener is a material part. The Commission, then, painted shareholders as external to the corporation. Most often, shareholders were presented as opposed to customers, with the two groups being gravitational forces that pull the corporation in different directions. This suggests that shareholders are as external to the corporation as customers are: each group holds particular rights against the corporation, and can use their leverage to pull the organization in a certain direction. My account of membership straightforwardly opposes this. Shareholders are internal material parts of the corporations they own. After all, shareholders are committed to abide by the corporation’s decisions, for example the corporation’s decision of which executive remuneration packages to put forward for shareholder approval. And shareholders have inputs into the corporation’s decision-making. For example, in Australia’s big banks, shareholders vote for the board of directors and vote to accept or reject executive remuneration packages. Finally, some actions of shareholders are properly attributable to the corporation itself. We might naturally say ‘CommBank voted in a new board of directors’ or ‘CommBank rejected executive bonuses’—where these are actions of shareholders. At this point, there is an interesting objection. The objection concerns shareholders as members of corporations. The objection derives from the fact that most modern-day shareholders are themselves organizations and

   

73

two organizations can each own shares in the other.¹³ The phenomenon of reciprocal shareholding is well-documented in, for example, Japanese ‘Keiretsu’ networks. These are collections of firms that are connected by reciprocal equity ownership, director transfers, and other ties (see, e.g., Lincoln, Gerlach, and Ahmadjian 1996). In a Keiretsu network or other case of reciprocal shareholding, my analysis gives the result that Company A is a member of Company B, and that Company B is a member of Company A. This violates the popular principle of ‘mereological antisymmetry’: two distinct things cannot be parts of each other.¹⁴ This should not worry us, however, since Koslicki’s view is that the formal aspects of objects are themselves parts of that object. It is plausible that wholes do not straightforwardly acquire all the structural aspects of their parts. For example, my lung has cells as material parts and a lung-structure as its formal part. My lung is a part of me. But I do not have a lung-structure as my formal part. My structure might be indifferent to whether I have organic lungs or iron lungs (where the latter are structurally and materially different from the former). In the case of organizations that are a part of other organizations, this is even clearer: the subsuming organization might have a structure that is entirely indifferent to the structures of the subsumed organizations that compose it. For example, the UN is indifferent to the structure of the states that compose it (with regard to such issues as democracy, monarchy, and so on). In this way, when two corporations own shares in each other, they do not wholly become parts of one another. This is because their structural aspects do not become parts of one another. Instead, there are some material components (the members) that are parts of both corporations. This is no violation of mereological antisymmetry.¹⁵ I am not the only social ontologist to include shareholders in corporations. Kirk Ludwig argues that ‘[t]he corporation is, in fact, literally its shareholders’ (2017a, 239). In my view, Ludwig’s view is rather too reductionist, since it does not give sufficient ontological significance the organization’s structure or to other role-bearers. It is important to give significance to the structure, so as to retain mereological antisymmetry—and, perhaps more importantly,

¹³ I thank David Strohmaier for raising this objection. ¹⁴ It is possible to deny mereological antisymmetry. But I will stick to the mereological orthodoxy, since this better serves my purposes in Chapter 5, where I’ll argue that phenomenology can percolate upwards from members to organizations. I thank Sam Baron for discussion. ¹⁵ For the same reason, when I say ‘New Zealand is part of the UN’ strictly speaking I mean ‘the material parts of New Zealand are parts of the UN’. I thank Sam Baron and John Hawthorne for discussion.

74

  

to capture how organizations act. But I agree with Ludwig that completely separating shareholders from corporations, as the Banking Royal Commission did, is a mistake. Moreover, this metaphysical mistake has moral implications. If shareholders are amongst organizations’ material parts, then they are amongst the objects in the world to which we can bring complaints against organizations. And when they perform their shareholder role (for example, by voting in shareholder meetings), they are acting as the corporation and they should act with a view to the moral standards to which the corporation is subject.¹⁶ Shareholders acquire the kinds of obligations that the Banking Royal Commission attributed to directors. Their inclusion helps to illustrate the moral importance of this chapter’s metaphysical position.

3.4.4 Brokers Having considered directors, non-manager employees, and shareholders, a fourth and final example of corporate membership is intermediaries. In banking, brokers are a key type of intermediary. Like advisors, brokers are typically paid based on the value and volume of business they bring to the corporation. Like advisors, brokers’ controversial remuneration schemes made them a core interest of Australia’s Banking Royal Commission. But the Commission treated brokers differently from advisors: brokers were treated not as internal to banks, but instead as intermediaries between a bank and its customers. Yet brokers’ pay and conditions were deemed within the remit of banks’ wrongdoing: their incentive packages were an ‘important contributor to misconduct and conduct falling short of community standards and expectations and poor customer outcomes’ (Royal Commission 2018, 62). So there’s a question of whether brokers should, in fact, be viewed as members. Here, my account of membership sides with the Royal Commission. Brokers meet only one of my three conditions for membership in banks. They are committed to abide by those outputs of the banks’ procedures that are relevant to them. But they lack systematic and intentional inputs into the bank’s decisions and they cannot act as the bank. They lack inputs insofar as they engage with the bank as a partner of exchange in the marketplace. ¹⁶ That is, shareholders have what I have elsewhere called ‘membership duties’ (Collins 2017b; Collins 2019a, ch. 7).

   

75

Banks’ decision-making procedure are not intended to systematically take inputs from brokers. And brokers lack enactments. It is never apt to say, for example, ‘the bank offered a low interest rate’ when a broker offers such a rate to a customer: instead, the bank’s offering of the low interest rate occurred when the bank offered that rate to the broker (not when the broker offered that rate to the customer). In this regard, then, my account vindicates some common-sense intuitions about membership. As these four examples show, determining the lines of membership in corporations is not straightforward. When it comes to the organization’s status as a blameworthy agent, we cannot simply read the list of members off whatever membership rules the organization itself happens to have in place. Determining which agents meet the three membership conditions is primarily an empirical task, which will vary not just for different types of corporation, but also for different token corporations and even for different individuals who have the same job title within the same corporation. Nonetheless, I hope this sketch has provided some guidance on how we might draw lines around the materiality of corporations.

3.5 Members of Liberal-Democratic States If I’m right about the metaphysics of organizations—if organizations are material objects that have members as material parts—then this requires some revision of the mainstream legal conception of statehood. In his classic text of international law, James Crawford asserts that ‘[a] State is not a fact in the sense that a chair is a fact, it is a fact in [the sense in] which a treaty is a fact: that is: a legal status attaching to a certain state of affairs’ (2007, 5). My somewhat revisionary proposal is that a state is indeed an object in just the same way a chair is an object: both are a combination of material and formal components. The question is: who are the material components of states? In answering this, I will focus on liberal-democratic states, for two reasons. First, in these states we get the somewhat surprising result that those who are eligible to vote are material components of states. I sketched this result above, but I will now explain it in-depth. This interesting result obviously does not extend to non-democratic states. Second, I focus on liberal-democratic states since they are the states primarily at issue in the Polluter Pays Principle. States’ blameworthiness for historical emissions is the example I will focus on when discussing how to apportion burdens to members (Chapter 7). It’s therefore worthwhile drawing the boundaries of liberal-democratic states

76

  

specifically. (For a more general discussion of collective agency and states, not drawing on Part I’s metaphysics, see Collins and Lawford-Smith 2017.)

3.5.1 Legislature, Executive, and Judiciary To warm up, consider the ‘easy cases’: the legislature, executive, and judiciary. The individuals that fill these roles are members of states, according to my account of membership. These agents are committed to abide by the state’s decisions, as a result of the incentives provided in their employment contracts if nothing else. (Although some do ‘go rogue,’ in parallel to the earlier discussion of rogue directors of corporations.) These role-bearers have inputs into the state’s decisions, be it through legislative debate, regulatory drafting, or judicial opinions. And they act as the state when they pass law, enact regulations, or hand down sentences. There may be room for debate about, for example, the government’s opposition. The opposition sits in the legislative chamber, but tends to be outvoted whenever they oppose a potential law. So you might think the opposition doesn’t systematically inform the state’s decision-making in ways intended by the procedure. But this takes an overly narrow view of legislative decision-making: through debate, committees, and campaigning, the opposition legislators do systematically inform the overall direction in which the state goes over periods of months or years (even if the opposition is voted down on particular pieces of legislation). And the procedure is set up with the intention of their doing this. Verdicts about membership are trickier when we consider those who live within a state’s territory, but who are neither legislators, nor public servants, nor judges. I’ll consider just two examples, which fall on different sides of the membership line: eligible voters and non-voting permanent residents. These two cases both sit extremely close to the membership line. There is much room for reasonable dispute about my analysis of these cases and both cases may be viewed as borderline. For exactly that reason, it is instructive to consider them.

3.5.2 Voters I’ll understand ‘voters’ as those who are eligible to register to vote at not disproportionate cost, and who enjoy other basic civil and political liberties,

   

77

within a particular state, in which they have a (perhaps forced, implicit, and pro tanto) commitment to being basically law-abiding (see similarly the discussion of this group in Collins and Lawford-Smith 2021). Under my account, thus-defined voters are material parts of their state. Each aspect of the ‘voter’ definition helps to explain why voters are members—material parts—of liberal-democratic states. First, it’s important that voters (in my sense) are those who are eligible to register to vote. They need not actually vote. After all, it is the capacity for inputs into organizational decisions that is one of the three conditions on organizational membership. The capacity need not be exercised. Yet in order for someone to truly have the capacity to provide inputs, it must not be prohibitively expensive for them to provide their inputs. Thus, voters who are systematically suppressed—who legally may register to vote, but who face large obstacles to doing so—do not count as members. As with low-level employees, this is as it should be: these people are shielded from the individual implications of organizational wrongdoing, which I’ll explain in Part III. The basic civil and political liberties—such as freedom of speech, association, conscience, press, and so on—ensure that voters’ inputs into the organizational procedure are truly their own. In these ways, voters meet the inputs condition on membership. The commitment condition on membership is met via voters’ commitment to basic law-abidingness. Someone has this commitment if they go about their lives with a general but overridable presumption that they will obey the law. A voter might occasionally decide to break a law, but this decision is the overriding of a general presumption. Thus, committed anarchists are not members of their states, in my sense, even when those states are liberal democracies. Does this let petulant citizens off the membership hook too easily? I don’t think so. It takes a lot to be a truly committed anarchist: one must be committed not just to the abstract idea that there is no ‘content-independent reason’ to obey the law (which is how ‘anarchism’ is often understood in contemporary philosophy). One must also lack the practical commitment, in one’s everyday life, to generally obeying laws even as a way of approximating moral requirements, or of seeking self-interest, or of getting by without punishment. Finally, do voters meet the enactment condition: do they have feasible actions open to them that, if performed, would amount to the voters actingas-the-state? They certainly have fewer such actions than legislators, executives, or judges. But only one such potential action is needed, in order to meet the enactment condition on membership. For voters, a paradigm such

78

  

action is voting. Here there is a close analogy between voters in states and shareholders in corporations. Shareholders can act-as-the-corporation only occasionally, at regular shareholders’ meetings, where they might vote on a board of directors or an executive remuneration package. Voters likewise act-as-the-state when they vote: we might see the headline ‘Australia elects new government’ or ‘New Zealand backs criminal justice referendum,’ without any hint of a misattribution of action: voters have acted as the state.¹⁷ Another act of voters that is attributable to the state is the act of funding public goods, which is typically done through paying taxes. To say that education is ‘state-funded’ and to say that it is ‘voter-funded’ is to take the voters’ act of funding and attribute that action to the state. Of course, voters are far from the only ones who pay taxes and thereby fund public services. But non-voter taxpayers who thereby meet the enactment condition typically do not meet the inputs condition.

3.5.3 Permanent Residents What about non-voting permanent residents? These are people who live in a state’s territory—and perhaps plan to do so for the rest of their lives—yet who lack the eligibility to register to vote at not disproportionate cost, or who lack some basic civil and political liberties. These people usually meet the commitment condition on membership: they have a standing practical commitment to abide by the state’s laws (either because those are the laws, or because the laws well-enough approximate morality, or because of selfinterest). Yet these people do not have the right kinds of inputs into the state’s decision-making procedures, so they do not meet the inputs condition. To see that permanent residents don’t meet the inputs condition, recall that the decisions of organizations must be systematically and intentionally informed by behaviours of members. Non-voting permanent residents might be able to systematically and reliably influence a state’s decision. For example, they might be able to attend protests, write letters to legislators, or engage in debate on social media that ultimately contributes (marginally)

¹⁷ There’s a sleight-of-hand here, insofar as it is voters plural who elect a government or pass a referendum. Can my solo action of voting be attributed to the state? I think the answer is yes: when I vote for a certain party, my state partly votes for that party. For dissent to the idea of citizens as members, see Lawford-Smith 2019; Collins and Lawford-Smith 2021.

   

79

to public opinion. But the state’s procedure is not set up with the intention of processing the inputs of these people, if they are not granted the opportunity to register to vote at not-disproportionate cost. As for the enactment condition, permanent residents meet the enactment condition via the action of funding public services. But it’s worth noting that they meet the enactment condition to a lesser degree than voters, and therefore receive a lower ‘grade’ on this condition. Consider the action of protesting or petition, which is an action of individuals that we might attribute to the state. Residents might attend a mass demonstration or participate in a letter-writing campaign. And again, when such events happen, we might reasonably expect headlines that attribute the action to the state, such as ‘Britain says Black Lives Matter!’ or ‘Canada demands legislative reform!’ But if it was discovered that the vast majority of participants in such actions were not voters, we would pause before attributing that action to the state. To identify a civil society movement with the state requires that the movement holds a certain level of internal legitimacy—a level of internal legitimacy held by voters but denied (often unjustly) to nonvoter permanent residents. Permanent residents therefore meet the enactment condition to a lesser extent.

3.6 Conclusion By now, we have a detailed picture of how organizations are material objects. The starting point for this is found in the well-worn reasons for realism about organizations, as outlined in Chapter 1. But—more innovatively—we now also have a picture on which organizations as real objects are not abstract real objects. They are instead material objects, which have members as physical parts. Organizations are different objects from the humans that compose them—but this is just as a solar system is a different system from the system of each planet that makes up that solar system, or a forest is a different system from the trees that compose it. The rest of this book will take seriously the idea that humans are to organizations as planets are to solar systems and as trees are to forests. (Hess (2018a, 41) proposes the solar system and forest analogies for corporations, for different purposes.) As mentioned in Chapter 2, all sorts of entities (inanimate objects, individual agents, and collective agents) stand in all sorts of metaphysically interesting and important relations to organizations. These include the relations of supervenience, causation, anchoring, and grounding. In

80

  

Chapter 2 we also saw that there are numerous ways to posit members as parts of organizations, but many of those ways don’t tell us who the members are, or don’t get the members quite right. By detailing commitments, inputs, and enactments as three individually necessary and jointly sufficient conditions on organizational membership, I hope that I have resolved the problems I posed for the various views considered in Chapter 2. These conditions allow us to say which agents are the members of organizations. On my view, these agents (and their parts) exhaust the material parts of an organization. Once we combine these members with the structural (nonmaterial) parts of an organization—the roles connected by relations and governed by a decision-making procedure—we have a full formula for individuating an organization as a material object. Part II of this book explains how this object can be blameworthy, and how the vision of members as material parts helps us to conceptualize and justify that blameworthiness. To be clear, my focus in Chapters 2 and 3 has been on establishing organizations’ materiality by explicating their relationship to the agents that compose them. But while I have focused on the material components of organizations, the nonmaterial components matter hugely. An organization’s structures and procedures make them something ‘over and above’ their material constituents, just as a motorcycle’s structure makes it something ‘over and above’ a collection of motorcycle parts. In the case of organizations, the structures and procedures make it the case that the organization can do wrong without any individual doing wrong: the structures and procedures can be at fault without any individuals being at fault, for instance when the structures and procedures are beyond the control of any individual. Despite the importance of structures and procedures, I have focused on organizations’ material components in order to demonstrate that organizations are not mere agents (as others have shown before), but that they are embodied agents. Their embodiment will matter in the arguments to come.

PART II

BLAMEWORTHINESS

4 Varieties of Organizational Blameworthiness 4.1 Introduction This book is titled Organizations as Wrongdoers. So far, I’ve said a lot about organizations, but not much about wrongdoing. The rest of the book changes that. The present chapter defends organizations’ blameworthiness through three different lenses: the volitionist lens, the attributivist lens, and the aretaic lens. Each lens offers a way to view an entity as wrongful, by focusing on a different aspect of the entity (see, e.g., Watson 2004; Smith 2005; Levy 2005; Shoemaker 2011). The volitionist lens focuses on the will, choices, or intentions of the entity. The attributivist lens focuses on the entity’s evaluative attitudes, over which the entity might never have made a choice. The aretaic lens focuses not on what an entity chooses or values, but at what an entity is—specifically, on an entity’s character flaws or vices. These lenses have been developed with individual humans in mind. This chapter will show that they apply just as well to organizations. Philosophers working on the blameworthiness of group agents have tended to ignore the variety of these lenses. Instead, they’ve tended to employ the volitionist lens, seemingly assuming that volitionism is the only game in town (e.g., French 1984; Isaacs 2011, ch. 2; List and Pettit 2011, ch. 7; Tollefsen 2015, ch. 6; Collins 2019a, ch. 6). Such an approach unduly limits the scope of organizations’ wrongdoing, since, as I’ll argue, each of the three lenses enables us to view organizations as wrongdoers, though each focuses on slightly different aspects of an organization. My aim in this chapter is to demonstrate that organizations can be blameworthy not just for what they choose, but also for what they value and for what they are. As we’ll see, my discussion of each lens will rely on the metaphysical picture presented in Part I, according to which members are material parts of organizations whose role performances are attributable to the organization.

Organizations as Wrongdoers: From Ontology to Morality. Stephanie Collins, Oxford University Press. © Stephanie Collins 2023. DOI: 10.1093/oso/9780192870438.003.0004

84

  

In laying out these three lenses on what organizations can be blameworthy for, I will assume the following: if it can be shown that organizations make morally significant choices, have morally significant evaluative attitudes, and embody morally significant character traits, then organizations can be blameworthy for these features. The chapter argues for the antecedent of this conditional. Of course, one might doubt that the conditional is true. The conditional is an assumption of my argument. This assumption is supported by our everyday practices of holding organizations as blameworthy, which is part of the interpretivist stand of organizations’ agency (see Section 1.3). And in Chapter 5, I will respond to an important objection to this assumption: the objection that organizations cannot be blameworthy because they lack phenomenology. In this chapter’s examination of what organizations can be blameworthy for, I remain neutral on what it is to be blameworthy. There are at least three views on what it is to be blameworthy. First, following in the tradition of Peter Strawson (1962), blameworthiness could be a matter of being a fitting target of certain backward-looking reactive attitudes, such anger, resentment, indignation, and disdain. To be a ‘fitting’ target of negative reactive attitudes is different from it being all-things-considered justifiable for others to hold these attitudes. For example, it might be justifiable to have these reactive towards a child as a way of educating them, even though the attitudes are not really fitting. Conversely, it could be fitting to blame someone even though blame is not all-things-considered justifiable, for example because all potential blamers lack standing to blame, or because blame would have negative consequences. A second theory of blameworthiness, roughly following T.M. Scanlon (2008, ch. 4), says that to be blameworthy is to have performed in ways that impair one’s relationship with others. I’ll assume such ‘performances’ can be choices, evaluative attitudes, or character traits—even though Scanlon himself focuses on blameworthiness for actions that reflect attitudes. On this view, to blame is to ‘judge him or her to be blameworthy [i.e., to judge him or her to have performed in ways that impair the relationship] and to take your relationship with him or her to be modified in a way that this judgement of impaired relations holds to be appropriate’ (2008, 128–9). Unlike the Strawsonian ‘reactive attitudes’ theory, this ‘relationship impairment’ theory says that being blameworthy is a matter of being liable to others modifying their relations with you in a way that reflects the impairment you have imposed upon the relationship. Such modifications need not involve any reactive attitudes such as anger, resentment, indignation, or disdain.

   

85

Third and finally, following Angela Smith (2015), being blameworthy could be a matter of being in principle liable to requests to answer for some aspect of oneself. Again, I’ll assume that ‘aspect’ could be one’s choices, one’s evaluative attitudes, or one’s character traits. To answer for an aspect of oneself is to explain it, excuse it, defend it, denounce it, or so on. In this ‘answerability’ theory, blameworthiness involves in the ‘in principle’ validity of a request for an answer. The request is only ‘in principle’ valid, because an entity could be blameworthy even though no one has the standing to ask the entity to answer for anything. For example, perhaps it would be hypocritical of anyone to ask the entity to answer for themselves. In that case, the entity is not in fact liable to answer for themselves—but they are still in principle liable to answer for themselves (in the sense that the entity would become liable, if someone with standing were to come along). Because the entity is in principle liable to the request, the entity is blameworthy. A request for answerability could arise, or could be legitimate, without any Strawsonian reactive attitudes arising (or being legitimate), and without any Scanlonian modifications of relations arising (or being legitimate). Thus, these three theories of what it is to be blameworthy come apart from one another. I take these three theories to be mutually compatible: perhaps to be blameworthy is to be liable to reactive attitudes, or to be liable to a change in relationship, or to be liable to a request to answer for oneself. That is, perhaps any one of the three conditions is sufficient for blameworthiness, without any one of the three conditions being necessary. Perhaps blameworthiness simply increases if the entity satisfies two or three of the conditions, rather than merely satisfying one of them. Or perhaps different of these three theories are apt in different cases. Be that as it may, I take these theories of what blameworthiness is to run orthogonal to theories of what an entity can be blameworthy for. In Table 4.1, I demonstrate that the questions are orthogonal. In what follows, my aim is to show that organizations can fall into any of the three columns of Table 4.1: organizations perform morally assessable actions, bear morally assessable attitudes, and have morally assessable character traits. In making this argument, I will assume that organizations can fall into any of the three rows of Table 4.1. I will assume that it is possible to hold reactive attitudes to organizations—such as anger, resentment, and indignation. I will assume that it is possible to alter one’s relation to an organization—such as by avoiding it, distancing oneself from it (whether psychologically, politically, or economically), or ceasing to admire it. And I will assume that it is possible to demand an organization provides

86

  

Table 4.1 What blame is versus the targets of blame Volitionism: blame targets volitions

Attributivism: blame targets attitudes

Aretaicism: blame targets character

Blame is reactive attitudes Blame is altered relations

E.g., anger over someone’s choices

E.g., resentment over someone’s values

E.g., avoiding someone over their decisions

Blame is demands for answers

E.g., demanding an explanation of someone’s intentions

E.g., distancing oneself from someone over their outlook E.g., demanding a disavowal of someone’s thoughts

E.g., indignation over someone’s vices E.g., ceasing to admire someone over their dispositions E.g., demanding a reason for someone’s practices

answers—such as explanations, disavowals, or reasons—by demanding such answers from role-bearers qua role-bearers. That such blame is possible I take to be demonstrated by reading any news website: people regularly do hold reactive attitudes to organizations, alter their relations to organizations, and demand answers from organizations. I seek to vindicate these practices, by showing that the practices have apt targets: by showing that organizations have morally assessable volitions, attitudes, and character traits. I assume that having any one of wrongful actions, wrongful attitudes, or wrongful traits is sufficient to render one blameworthy, while none of these three features is necessary for being blameworthy. In this way, my argument is maximally ecumenical, both with respect to what blame is and what blame’s targets are.

4.2 Volitionism for Organizations Volitionism has received a lot of air time in the literature on the blameworthiness of group agents. This goes back at least to Virginia Held’s (1970) discussion of blaming an ‘organized group or collectivity’, by which she meant a group ‘distinguishable not only by certain characteristics that delimit its membership from other persons, but especially by its possession of a decision method for action’ (1970, 471). Held assumes the following picture:

   

87

To hold an individual responsible for an action requires that he be aware of the nature of the action, in the sense that he is not doing A in the belief that he is doing B, and it requires the judgement ‘He could have done something else’ be valid of the action he has performed. To hold an individual morally responsible for an action requires that he be aware of . . . the moral nature of the action, which can be understood as either the moral import of the kind of action of which it is an instance or as the moral value of the consequences it may produce. These requirements . . . require that he be aware of the kind of action he is performing, but not that he know everything about either its empirical or moral aspects. (1970, 472)

The context makes clear that Held assumes ‘morally responsible’ to mean ‘praiseworthy or blameworthy’. The idea, then, is that blameworthiness requires the agent to be aware of their action and of that action’s moral nature. Held’s examples—which include throwing an explosive, flipping a light switch, and stealing a car—suggest she means for blameworthiness to range over actions, which are performed knowingly and intentionally. Because she focuses on actions that are chosen knowingly, she sits within the ‘volitionist’ camp: entities are blameworthy for willed actions. Held asserts: ‘If these requirements for responsibility and moral responsibility can be met by individual persons, they can, I think, be met by collectivities’ (1970, 473). More recently, volitionism is assumed in List and Pettit’s influential discussion of the blameworthiness of group agents (2011). List and Pettit distinguish moral responsibility—which, again, they use synonymously with ‘praiseworthiness or blameworthiness’—from mere causal responsibility and from strict liability (such as the kind of liability a parent has for the harmful behaviours of her teenage children). Within moral responsibility, they presume a volitionist model. They write: a group agent is fit to be held responsible for doing something to the extent it satisfies these requirements: First requirement. The group agent faces a normatively significant choice, involving the possibility of doing something good or bad, right or wrong. Second requirement. The group agent has the understanding and access to evidence required for making normative judgements about the options. Third requirement. The group agent has the control required for choosing between the options.

88

  

These conditions differ slightly from Held’s, demonstrating the variability within volitionist views. Whereas Held focused on the entity’s awareness of the nature (including the moral nature) of the action, List and Pettit focus on the entity’s understanding, access, and control over its options. But the broad idea is the same: an entity is blameworthy for what it does knowingly, intentionally, or wilfully. Like Held, List and Pettit argue that groups with decision-making procedures can meet their three conditions. Similar arguments are provided by French (1984), Isaacs (2011), and Tollefsen (2015), amongst others. The volitionist lens is more-or-less taken for granted by social ontologists such as Held or List and Pettit. But it is prominently defended for individual agents by Neil Levy (2005). According to Levy, blameworthiness requires that an entity control for a certain state of affairs. What’s more: [i]t is not sufficient for an agent to be able to cause an alteration in a state of affairs, or even actually to cause such an alteration, for that agent to control that state of affairs. If I do not know either that I cause such change, or how I alter the state of affairs, then I do not control it. (2005, 5)

Levy’s main argument for this picture is that it enables us to distinguish the blameworthy from the merely bad. He considers a serial killer who ‘may not have had a genuine chance to become a better person’ and suggests that this person is merely bad, not blameworthy—after all, they do not control for their malevolence. Entities must have volitions if they are to be blameworthy. My view of organizations allows for Held’s, List and Pettit’s, and Levy’s conditions to apply to organizations. Organizations are realized structures that are governed by decision-making procedures. Decision-making procedures enable organizations to form decisions about how to act. I illustrated this in Chapter 1, Section 1.3, with the example of Australia’s action of permitting a mine. Decision-making procedures also enable organizations to make decisions about what to believe. In the case of individuals, the notion of ‘deciding’ what to believe is somewhat alien: we normally feel compelled to believe. By contrast, in organizations, the process of deciding to believe is often quite explicit, as when a Board of Directors draws up a Statement of Values, or when a panel of judges discusses what precedent the court should believe was set by an earlier judgement. Such processes of belief-formation enable organizations to have the kind of ‘awareness’ required on Held’s and Levy’s pictures of blameworthiness—an awareness

   

89

in which one performs an action while believing one is performing that action. The ‘belief ’ here is a simple propositional attitude, which can be characterized in the functionalist-interpretivist way outlined in Chapter 1, Section 1.3. (I will consider phenomenologically richer kinds of ‘awareness’ in Chapter 5.) Organizations can also satisfy List and Pettit’s conditions. Organizations constantly face morally significant choices. They have the understanding and evidence required to make such choices, insofar as they can believe propositions such as ‘doing X would be right’ and members can gather the evidence needed to justify such beliefs before presenting that evidence to the organization’s decision-making procedure. And organizations can have control over morally significant choices, in the sense that the organization can determine that some member or other will act within their role to ensure the organization enacts one choice rather than another (List and Pettit 2011, 158–63). At first glance, it might be surprising that volitionism has been the framework of choice for those working on collective blameworthiness. After all, many of the harms done by organizations do not seem deliberate, controlled, or chosen. Instead, organizational wrongdoing is often the result of sheer indifference; an utterly unthinking ignorance about the interests of others. The banks that were examined by Australia’s Banking Royal Commission didn’t intentionally harm customers; they simply failed to consider customers’ interests while going about the ardent pursuit of profit (and while bank members went about the ardent pursuit of higher remuneration). But volitionism doesn’t require that blameworthy harms are intended. It merely requires that blameworthy harms are controlled. One can control for a harm that one causes as a side-effect of one’s intentional action. One can control for what one does out of recklessness or negligence, as well as wilfully. One can even control for what one does in complete ignorance, as long as one’s ignorance is traceable back to a moment in time in which one knowingly had a choice about whether to become ignorant. There’s another way in which volitionism casts a wider net than first meets the eye, at least within organizations. In much discussion of group agents’ blameworthiness, there is an implicit assumption that we are considering blameworthiness only for official corporate policy. For example, in Pettit’s (2007, 197) discussion (on which List and Pettit’s (2011) is based), we are asked to imagine a company in which employees are given the opportunity to vote on whether to forgo a pay rise in order to increase

90

  

workplace safety. The votes on various aspects of the decision are tallied, such that the group arrives at an official group position. The thought is that we can hold the group blameworthy for this decision: this is a decision over which the group per se has control, even though no member controls for it. This raises the question: if official corporate policies are the paradigm phenomenon over which organizations wield control, then how far does corporate control extend? What about the informal norms, expectations, and unspoken but sacrosanct requirements within an organization? Here we can think about things like whether customers are viewed as humans with legitimate interests, or as fodder for the maximization of profit. It’s likely that no official organizational document would ever take a stand on this issue. Is it, then, a matter of organization-level volition, given that it’s not a matter of official corporate policy? Can organizations be blameworthy for such norms, under the volitionist lens? Yes, for two reasons. First, the organization’s structure and decisionmaking procedure gives the organization control over such issues, even if no policy explicitly mentions them. For example, consider the widespread banking policy of remunerating advisors in accordance with the business they bring in. This organization-level policy controls for advisors’ treating customers as mere sources of business—rather than, as the term ‘advisor’ might suggest, treating customers as persons with legitimate interests who deserve advice on how best to fulfil those interests. Insofar as the organization knows—or could easily have known—about the incentives the policy creates for advisors, it controls for how advisors treat customers, without needing to officially say anything. Importantly, the organization can be blameworthy for giving advisors these incentives, even if the advisors are not blameworthy for acting on those incentives. Perhaps it is too much to ask an advisor not to take up the incentives. In this case, the organizations can be volitionally blameworthy for how members are situated within the organization’s structure, where the members themselves are not blameworthy for being thus situated. On the volitionist lens currently under examination, the organization’s blameworthiness arises because the organization’s decision-making procedure gives the organization control over how the members are situated in the structure. Second, even if corporate policy doesn’t knowingly incentivize a certain attitude towards customers, we can use Part I’s metaphysics of membership to attribute control over such treatment to the organization. Recall from Chapter 3: members are material parts of organizations. Organizations are present and active in those contexts where members are present and

   

91

performing their organizational role. Now imagine a workplace culture in which—via no piece of corporate policy, not even as a foreseeable sideeffect—banking advisors come to treat customers as profit-fodder. Perhaps this culture arises via the macho norms of wider society (which are not under the organization’s control), combined with a team of advisors that happens to be entirely male (again, not under the organization’s control: suppose all applicants for advisor roles just happened to be male). The macho norms create a competitive culture in which customers are instrumentalized by advisors who hope to ‘win’ the macho competition over who can bring in the most business. In this circumstance, we can probably attribute some volitional control to the organization simply via the route of official organizational policy. For example, we might say that the organization did wrong by not issuing a policy of customer-focused care, or by not issuing a policy of affirmative action in hiring. Yet we might imagine that either of these policies would have been too much to ask. And even if the organization is volitionally blameworthy for (because controlling of) the absence of these policies, there is still more that we can say about the organization’s volitional blameworthiness for the advisors’ actions and attitudes in this case, using Part I’s metaphysics. What we can say is this. When the advisors are playing their role in the organization, their actions and choices are attributable to the organization— such that whatever an advisor controls for is something that the organization also controls for. On this picture, the organization controls for more than just its corporate policies and the foreseeable effects of those policies. It also controls for anything the members control for when playing their role, precisely because the organization is present and active whenever the member is playing their role. The organization ‘inherits’ the blameworthiness from its material part. To use an example close to home: when a lecturer chooses to treat her students as customers rather than as mentees, then her university also makes precisely that choice—because it is a choice she makes while performing her role in the university. In this way, volitionism—combined with Part I’s metaphysics—gives us a vindication of Australia’s criminal code in this arena. Recall from Chapter 1 that Australia’s criminal code attributes an offence to a corporation when ‘an offence is committed by an employee, agent or officer . . . within the actual or apparent scope of his or her employment, or . . . authority’ (Criminal Code Act 1995 (Cth), Part 2.5, Sec. 12.1, emphasis added). This makes sense if members are material parts of organizations, whose actions are attributable

92

  

to the organization whenever members are performing their role. If members weren’t material parts of organizations (with the organization being active whenever members are performing their role), then it would be puzzling why each role performance should be attributable to the organization itself. At the same time, volitionism (combined with Chapter 1’s account of organizations-as-agents) explains why Australia’s criminal code is incomplete. Australia’s criminal code neglects organization-level decisions that are not made by any one member, but that rather result from the inter-member operationalization of the organization’s decision-making procedure—including in cases where members’ own contributions to this operationalization might not be at all individually volitionally blameworthy. Thus, the volitionist lens gets us an awful lot of blameworthiness for organizations. By using this lens to examine what organizations control (and what members control qua material parts of organizations), we can see organizations as fitting targets of reactive attitudes, as having impaired their relationships with others, or as liable to calls to answer for their conduct. However, volitionism does not capture everything for which we might want to hold an organization blameworthy.

4.3 Attributivism for Organizations Consider again List and Pettit’s three conditions for blameworthiness. Their first condition requires that the entity faces a choice. But what about blameworthiness for things that didn’t involve choice? Their second condition requires that the entity understands the moral significance of its choices. But what about organizations that have been designed to be blind to certain considerations, such that they cannot see the moral significance of certain choices? Finally, List and Pettit’s third condition demands that the entity can control for choosing the morally right option. But what about organizations that cannot control for choosing the morally right option as the morally right option, because the organization has not (and could not have) adopted a moralized interpretation of the situation at hand? The attributivist lens answers these questions. It countenances blameworthiness over aspects of an entity’s life that don’t involve any choice or volition. Even when it regards choices, the attributivist lens countenances blameworthiness over choices whose full moral significance the entity does not understand. And it allows entities to be blameworthy for choices that the entity may not conceptualise as moral choices, perhaps because the entity’s

   

93

background commitments or constitution rule out such a conceptualization. Obviously, attributivism greatly expands the scope of outcomes and behaviours for which entities are blameworthy. As Robert Adams put it in an early statement of an attributivist-like view for individuals: ‘No matter how he came by them, his evil beliefs are part of who he is, morally, and make him a fitting object of reproach’ (1983, 19). To see attributivism in individuals outside organizations, consider Angela Smith’s (2005) discussion of forgetting her friend’s birthday. Clearly, forgetting her friend’s birthday wasn’t an intentional action. Still less was it something Smith was aware of (as Held requires), or something she chooses with understanding, access to evidence, and control (as List and Pettit require). Nonetheless, Smith seems blameworthy for forgetting her friend’s birthday—even if the blameworthiness in question is relatively mild. Now, it might be that if we look closely, we will find some prior action or omission (such as choosing not to put the birthday in her calendar), for which Smith is blameworthy in the volitionist sense. But, Smith suggests, we can imagine a case in which there was no such prior action or omission, yet Smith is still blameworthy for the failure to remember. The alternative is concluding that her forgetting the birthday was tragic outcome. But that fails to ring true to the practices of both Smith and her friend. Smith is blameworthy for the forgetting in an attributivist sense. According to Smith, ‘the kind of activity implied by our moral practices is not the activity of choice, but the activity of evaluative judgement’ (2005, 237). We are therefore blameworthy for any action or outcome that ‘reflects our own evaluative judgements or appraisals’ (2005, 237)—even if these judgements are implicit, unreflective, unconscious, unchosen, and uncontrolled. Our evaluative judgements themselves are often not something we have wielded control over, and those judgement often go on to affect our thoughts, feelings, and dispositions in ways that are likewise uncontrolled. Our evaluative judgements are not things we choose, or even cause, or even maintain—they are, instead, things we ‘inhabit’ or that ‘make up the basic evaluative framework through which we view the world’ (2005, 251). At the same time, evaluative judgements are ‘in principle answerable to a person’s judgement’, such that it is appropriate to ask the person to ‘acknowledge and to defend or disavow the judgements implicit in her responses to the world around her’ (2005, 256). Levy (2005) suggests that these features bring evaluative judgements under volitional control, and therefore suck attributivism of its claim to be distinct from volitionism. But I think there’s a different way to understand Smith’s claim: our evaluative judgements are

94

  

things over which we might not have been able to wield control in the past, but over which we can try to wield control in the future. Nonetheless, when we are blameworthy for them, we are blameworthy for their past manifestations—thus, being blameworthy for things over which we lacked volitional control at the period for which we are blameworthy. Of course, one can debate whether blame is apt, if the entity lacked the opportunity to exercise the kind of control and knowledge emphasized by Held or List and Pettit. Perhaps the possibility of future control is not sufficient to motivate past-directed blameworthiness. But I think Smith (2005), amongst others, makes a strong case that volitionism doesn’t cover all cases of blameworthiness. For example, Gary Watson points out that unchosen evaluative attitudes do not amount to excuses. As Watson puts it, ‘we would never be inclined to say “he should not be held responsible” of a person whose failure was caused by an inconsiderate evaluative attitude: we wouldn’t say “he should not be held responsible; he’s simply an unthinking racist” ’ (Watson 1996, 236; similarly Scanlon 2008, 193–8; Vargas 2005). In any case, I want to leave aside the debate between volitionists and attributivists. That is, for the sake of argument, I want to grant to attributivists their premise that individuals can be blameworthy for what they value (even if those valuings were beyond their control). I want to ask how this extends to organizations. I mentioned above that some organizations are designed to be blind to certain considerations. The clearest example is a corporation which is designed to be blind to any consideration other than maximizing profit for shareholders. It might seem doubtful that such attitudes are open to attributivist blameworthiness. After all, if an organization is set up to pursue certain ends, then those ends look like things the organization cannot even try to defend or disavow in the future, thus failing to satisfy Smith’s futureoriented condition that the evaluative attitude is ‘in principle answerable to a person’s judgement’. Meanwhile, all other goals of an organization might seem to be chosen—and so, covered by volitionism. It might seem that there’s nothing left in an organization’s make-up for attributivism to lay claim to. This isn’t quite right. It’s true that most organizations are oriented around particular goals. These might be goals the organization cannot but pursue, at least within a particular time period, and that the organization has never had opportunity or occasion to choose or control for—even though the goal is, in the long run, amenable to revision, defence, or disavowal. (I elaborate on these in Collins 2020a.) In a corporation’s day-to-day operations, it might be taken for granted that profit is the goal; everything else is a means to that

   

95

goal. This might be taken for granted to such an extent that it counts as a value the organization inhabits, rather than something it currently controls for. Indeed, insofar as a corporation was set up to pursue profit, the pursuit of profit may never have been on the agenda as something over which the organization could make a choice. As Smith puts it when discussing individuals: ‘In order to regard an attitude as attributable to a person, and as a legitimate basis for moral appraisal, we need not also claim that a person is responsible for becoming the sort of person who holds such an attitude. . . . What matters . . . is that the attitude is in principle dependent upon and sensitive to the person’s evaluative judgements’ (2005, 267–8). Fortunately for those who want to hold organizations blameworthy, a goal like ‘maximize profit’ is amenable to evaluation, and even revision, by the organization in the long-term. Such a goal is also amenable to defence or disavowal. Such goals of an organization are a key target of attributivist forms of organizational blameworthiness: perhaps a bank never chose to pursue profit at the expense of all else—perhaps it even never had the clear opportunity to make that choice—but now that that’s the evaluative judgement the bank unreflectively acts on, the bank can be subject to reactive attitudes for it, can be called upon to defend or disavow it, and can have others alter the shape of their relation to the bank on its basis. Indeed, this is exactly what the Banking Royal Commission did. When we consider evaluative attitudes such as ‘this bank should maximize profit’ or ‘this state should prioritise growth in GDP’, we can see why the attributivist lens is such an important lens to have in our toolkit, alongside the more traditional volitionist lens. In addition to an organization’s goals, there’s another key type of evaluative judgement for which organizations have attributivist blameworthiness. To see this, we once again need to use Part I’s metaphysics: we must view members as the material parts of organizations such that the organization is present and active whenever a member is performing their role. With this in mind, notice that members can play their roles in ways that reflect certain unchosen evaluative judgements, where those evaluative judgements are nonetheless open to defence or disavowal from within the member’s organizational role. Here there is a parallel to my above discussion of volitionism: just as members’ choices while performing their roles can render an organization blameworthy, so too members’ evaluative judgements while performing their roles can render an organization blameworthy. For example, consider again the imagined group of macho financial advisors within an organization. Each of these advisors makes choices that

96

  

treat customers as profit-fodder, but not as a result of any choice that was controlled by the organization. This time, additionally, suppose that there was no choice that was controlled by any of the individuals. Perhaps the macho culture is so deeply embedded in their society that the question has never arisen for any of them about whether or not he should treat customers awfully in pursuit of competitive advantage over other advisors. (These advisors are structurally similar to a person raised in a deeply racist community, to whom Smith extends attributivist blameworthiness for racist evaluative judgements (2005, 267).) Imagine that these macho advisors—who have never had any control over being such, and whose organization has never had any control over their being such—end up inadvertently acting in ways that reflect their underlying evaluative judgement that customers are mere profit-fodder. In their unchosen and unreflective rush to win the profit-producing competition, each advisor engages in various unchosen slights: they forget to call customers when they say they will; they misplace important documents; they fail to notice when customers have given them incomplete information. They perform their roles in ways that reflect problematic unchosen evaluative judgements about the importance of meeting customers’ needs relative to the importance of maximising corporate profit (as a result of maximizing their own individual remuneration). In this, they perform their role of ‘financial advisor’ much as Smith performs her role of ‘friend’ when she forgets her friend’s birthday. As long as these evaluative attitudes are permitted by, and held while performing, their role in the organization, these evaluative attitudes are attributable to the organization. They are not merely evaluative attitudes of the individuals. As Miranda Fricker puts it: the individual features in question . . . are not found at the level of individuals considered independently from the group, for the individuals only have that feature if they are wearing their group-member hat. Some practical identities of individual members are thus intrinsically groupinvolving, and in such cases there is no lower level of group-independent features to which the higher-level [organization-level] features can be reduced. Any attempted reduction of the group to a sum of uncommitted non-group identified individuals would literally change the subject, and so fail. (Fricker 2010, 238–9)¹ ¹ Three comments on Fricker. First, Fricker unpacks such collective-level attributes in terms of Margaret Gilbert’s notion of ‘joint commitment’. I have already said that organizations’

   

97

That is: although the advisors’ evaluative judgements are neither controlled for by them, nor controlled for by the organization, those evaluative judgements are attributable to the organization, as well as to the members. This is because those judgements are inhabited only while performing the organizational role: outside the organizational role, the macho financial advisors have no occasion to view customers as mere profit-fodder.

4.4 Aretaicism for Organizations The final lens that I wish to apply to organizations is the aretaic lens: the lens that focuses on virtue and vice. Because my subject is organizations’ wrongdoing, I will focus on vice rather than virtue. Having a vice is different from making a wrongful choice, or even from having a morally bad or wrongful evaluative judgement. A vice is a stable character trait—a multi-faceted property consisting of a range of interlocking judgements, behaviours, emotions, dispositions, habits, and so on. Unlike volitions, vices certainly need not involve opportunities for conscious choice or control. Unlike evaluative judgements, vices are not just about one’s (perhaps implicit) assessment of a matter of fact or value, but rather are a more holistic assessment of one’s orientation to the world that cannot be pinned down to any one judgement, however implicit, wide-ranging, or long-run. Just like a virtue, a vice is ‘a disposition, well entrenched in its possessor—something that, as we say, goes all the way down, unlike a habit such as being a tea-drinker—to notice, expect, value, feel, desire, choose, act, and react in certain characteristic ways’ (Hursthouse and Pettigrove 2016). To have a vice is ‘to be a certain sort of person with a certain complex mindset’ (ibid.). If volitionism is about one’s choices and attributivism is about one’s judgements, then aretaicism is about one’s character.

members may have no joint commitments—certainly there may be no joint commitment to problematic evaluative attitudes—so I don’t view Fricker’s analysis of collective-level evaluative judgements as applicable to organizations; I’m here endorsing only her rejection of individualist readings of the relevant evaluative judgements. Second, the above Fricker quote comes from her discussion of group virtue—but the examples she discusses are best understood as group evaluative attitudes, rather than group character traits. I distinguish these in the next section. Third, Sean Cordell (2017) argues that Fricker’s account of group-level evaluative commitments can be understood in reductively individualist terms. However, once we view members as the physical parts of organizations—whose agential features are applicable to the group when members are performing their role—then the application of (certain of) members’ features to organizations is apt.

98

  

To see that character traits are different from evaluative attitudes, consider that the former include habits, dispositions, and practices that may not be a reflection of just one—or even a combination of—evaluative attitude(s). Furthermore, in Smith’s presentation, attributivist blameworthiness rests partly on the fact that an entity can be called upon to either defend or disavow their evaluative attitudes (even if they never had past control over adopting those evaluative attitudes). Evaluative attitudes, then, can be examined by their bearer in a piecemeal, one-by-one, fashion: a single evaluative attitude can be held up to the light and rejected or defended, presumably in light of the bearer’s other evaluative attitudes. By contrast, character traits are so widely embedded across one’s life that to hold up such a trait for examination would seem to require adopting a ‘view from nowhere’ upon one’s self. Nonetheless, other agents are capable of assessing one’s character traits and holding one blameworthy for them.², ³ Where do organizations fit in here? Organizations may be more amenable to character-based assessments than individuals. We can see this by thinking about a large objection to virtue ethics: the situationist objection (Doris 1998; Harman 2003). According to this objection, individuals don’t possess stable character traits, as can be seen from the fact that small changes in individuals’ situations can alter how they act. Arguably, the objection depends upon an overly simplistic picture of character traits, which ignores their multifaceted and multi-track nature (Sreenivasan 2002). Be that as it may, organizations are arguably less prone to situation-based fickleness than human beings. This is because organizations have policies, rules, hierarchies, and so on that are precisely intended to produce stability and predictability in organizations’ dispositions (French 1995, 80). Individuals have no such

² Scanlon (2008, 127) suggests that assessment of others’ characters amounts to a ‘pointless assignment of moral “grades” ’. As stated in Section 4.1, I’m open to the possibility that holding an entity blameworthy for its character amounts to changing one’s relationship with that entity in certain ways—in which case, the assignment of ‘grades’ for character would be far from pointless (as Scanlon notes (2008, 153)). But I also believe we sometimes rightly have reactive attitudes to vicious characters, even if we have no relation to the vicious entity. ³ In his seminal ‘Two Faces of Responsibility’, Watson (1996) distinguishes the volitional from the aretaic. My reading is that Watson’s ‘aretaic’ elides the attributivist and aretaic lenses, because it elides evaluative attitudes and character traits. Watson says, for example, that ‘[r]esponsibility is important to issues about what it is to lead a life, indeed about what it is to have a life in the biographical sense, and about the quality and character of that life’ (1996, 229). This implies an aretaic approach. Later, Watson says ‘if what I do flows from my values and ends, there is a . . . sense in which my activities are inescapably my own: I am committed to them’ (1996, 233). This implies an attributivist approach. I therefore draw on the virtue ethics tradition (rather than Watson’s discussion) when characterizing character traits.

   

99

procedural producers of situational consistency. Organizations are therefore ripe for character-based assessment.⁴ Indeed, vices are regularly attributed to organizations—particularly corporations. Multiple times, Australia’s Banking Royal Commission asserted that ‘greed’ or ‘avarice’ was the source of the numerous layers of transgression it uncovered (2018, xix, 73, 74, 122, 131; 2019, 138). And in response to findings like those of the Royal Commission, various regulatory bodies in Australia have recommended the targeting of corporations’ culture or character. Thus the Australian Prudential Regulatory Authority (APRA) introduced new standards around ‘risk culture’, requiring boards of directors to: specify the quality and character of the culture that they seek to attain within their organizations (typically done in terms of specifying the organization’s core values and principles); measure the extent to which the actual culture aligns with the ideal; and develop and implement measures to close any identified gaps between the actual and the ideal. APRA’s suggestion that culture or character is a regulative ideal—something always to be striven for but perhaps never reached—reflects the widely-accepted view within virtue ethics that traits of character are ongoing or insatiable goals, unlike the making of a particular choice in a particular circumstance. So organizational character or vice is already present in nonphilosophical discourse. Indeed, some theorists have already addressed the virtue of businesses or corporations. For example, Geoff Moore (2005) has applied Alastair MacIntyre’s (1984) virtue ethics to businesses. For MacIntyre, virtue is grounded in existing practices. Moore therefore argues that businesses should develop the virtues internal to existing business practice—specifically, the virtue of ‘craftmanship’. Likewise, Karl Schudt (2000) has defined corporate virtue as ‘a habit of corporate action that is conducive to the achievement of the corporate good’, where the corporate good is characterised by ‘sustainable profit’ (2000, 713). Yet perhaps businesses—and organizations more generally—should aim for more generally required virtues, such as care, beneficence, and humility; not just the business-focused virtues of craftmanship or profit (however ⁴ Will Cailes pointed out to me that organizations’ being composed of many agents might undermine organizations’ rationality, which might undermine their possession of stable character traits (not to mention their ability to form volitions). It’s an empirical question whether organizations are more rational than individuals, whom we hold blameworthy despite their occasional (or not-so-occasional) lapses in rationality. We should be careful not to hold organizations to a higher bar of rationality than we hold individuals. In any case, organizations’ procedures and structures render them rational enough to have character traits, even if organizations’ internal heterogeneity creates some roadblocks to rationality.

100

  

sustainable that profit may be). What’s more, the recommendation of a specific virtue doesn’t tell us the mechanisms by which, in general, an organization might have any given virtue or vice. Other business ethicists have considered how businesses might help their members to flourish (Hartman 2015; Solomon 1993)—‘flourishing’ being a central concept of Aristotelean virtue ethics. Yet it’s unclear why businesses should focus on their members’ flourishing, rather than on the organization itself being virtuous, or at least non-vicious. Again, we want to know how the concepts of ‘virtue’ and (especially) ‘vice’ might apply to organizations themselves (not to their members), in a way that is neutral on which virtues and vices are to be cultivated. To achieve this, I’ll outline a few mechanisms via which organizational vice might arise, where these mechanisms are not best understood as reducible to the volitionist and attributivist mechanisms outlined above, and where these mechanisms are open-minded about which precise virtues organizations should embody. In doing this, it’s important to appreciate that vices (like virtues) are multi-track dispositions: they are dispositions to act in a range of different ways across a range of different circumstances. This means they cannot be applied to an organization on the basis of one action or policy. My examples will be ‘one-shot’ actions, but the reader should remember that for an organization to be truly vicious it would have to exhibit several such actions, reliably and over time. Importantly, both my examples will be of cases where an organization has some character trait even though no individual member has that same character trait. The possibility of such discontinuity between the organizations’ trait and the members’ traits is an important plank in any argument for collectives’ irreducible blameworthiness. In providing such examples, my argument therefore departs from those who analyse organizations’ virtues as a matter of the right kinds of relations between individuals. For example, Anita Konzelmann Ziv suggests that organizations’ virtues are largely a matter of how members’ virtues ‘aggregate and converge’ (2012, 96). I think organizations’ character traits can be less member-dependent than this, as I hope to show in what follows. Following Aristotle’s framework for individual virtue, we can separate three ways in which an agent might fail to be virtuous: they might be continent (acting as the virtuous person would, but with difficulty); they might be akratic (acting as the vicious person would, against the agent’s better judgement); or they might be vicious (acting as the vicious person would, in accordance with the agent’s judgement). I believe organizations can manifest all three flavours of aretaic failing.

   

101

To see how organizations can be continent, imagine a bank that performs in a way reflecting the right volition or the right evaluative judgement, but where the bank finds it difficult to do so. Specifically, imagine a remuneration committee debating the principles the bank will use for remuneration of financial advisors. The committee acknowledges that the right thing to do would be to remunerate advisors according to customer satisfaction, or according to the extent to which the advisors provide advice that best fits the customers’ needs. The committee ends up choosing to do this. It thus performs in a way that reflects morally good choices and good evaluative judgements. But the committee finds this decision difficult. The committee finds the decision difficult not because the decision is pragmatically unwieldy to implement, or because the decision is so costly that it risks making the bank go bust. Instead, the committee finds it difficult because the bank has a long-standing habit, track record, or disposition of doing whatever it can to maximize profit. In this case, the bank exhibits a character trait: the trait of being continently considerate. It does not have the full virtue of considerateness, even though its practices end up passing the volitionist and attributivist standards for non-blameworthy behaviour. The bank is not virtuous, because it does the right thing only reluctantly or with difficulty. Virtue ethicists capture this with the concept of ‘continence’: the continent agent has strength of will, but the fact that the agent requires this strength of will to do the right thing demonstrates an underlying character flaw (Foot 1978, 11–14). In this case, to assess the bank as not virtuous, we must look internally to the instantiation of the processes by which the bank arrived at its choice or evaluative judgement—we cannot simply look at the choice or evaluative judgement itself. Only the complex push-and-pull of internal wrangling demonstrates the bank’s character trait—in the example, its lack of virtue.⁵ What about akrasia? To see organizations’ akrasia, imagine a tweak on the foregoing example. Suppose the organization has recently committed to a statement of values that puts customers’ needs first. A commitment to ⁵ This example replies to Cordell’s scepticism about collective virtue, which arises because he finds it ‘difficult to see how the group agent could deliberate about or reflect upon the value of [a virtue], or how its value would figure in the group agent’s motives’ (2017, 53). Committee discussions provide a concrete example of how such deliberation can occur. Note that committee deliberations are not situations in which each individual committee member is deliberating about how he or she will behave; they are situations in which the committee is deliberating about how it will behave. By using the example of continence, I also part ways with Karl Schudt (2000, 713), who argued that in applying Aristotelean virtue ethics to corporations ‘[t]here will be no need for the doctrine of the Golden Mean, since corporations do not have appetites’. Their appetites, I suggest, can be found in their long-standing dispositions.

102

  

fulfilling customers’ needs is, thus, the organization’s ‘better judgement’. But the organization also has a long-standing habit of putting profit ahead of customers’ needs. In this way, the organization is analogous to a human who has recently committed to giving up smoking. The human’s better judgement mandates that they decline the offered cigarette. But the human’s longstanding habit pushes in a different direction. The human is akratic if their long-standing habit trumps their better judgement, resulting in the human accepting the cigarette. Likewise, we can imagine the remuneration committee being akratic: we can imagine that the long-standing corporate culture of pursuing profit plays trumps over the recently-endorsed statement of values, resulting in the committee enacting a remuneration policy that rewards profit-maximizing. Again, this complex interplay of judgements, habits, and actions is not explicable in terms of any one volition or attitude. It is a character flaw—akrasia—of the organization. The continence and akrasia examples demonstrate an organization that is not virtuous—but is not yet fully vicious. To see how viciousness itself can occur in organizations, consider its antithesis within Aristotelean virtue ethics: eudaimonia. To have eudaimonia is to flourish or be excellent. It is to have a deep-seated and well-entrenched practice of doing the right things for the right reasons. To be vicious, then, is to have a deep-seated and wellentrenched practice of doing the wrong things for the wrong reasons. But what is ‘the right thing for the right reasons’, by reference to which we can characterize ‘the wrong thing for the wrong reasons’? Many virtue ethicists agree (following Aristotle) that eudaimonia is type-relative: what it is for a human being to flourish is different from what it is for a dog or a lily to flourish. To provide a possibility-proof of organizational vice, I will assume this Aristotelean picture. When it comes to organizations, then, we can develop the notion of viciousness by asking ‘what is it for an entity of this type to fail?’ The answer is likely to be different for different types of organizations. One might be tempted to say that the flourishing corporation just is the corporation that maximizes shareholder profit. Maximizing shareholder profit is, on this view, the paragon of corporation virtue. This position was famously endorsed by Milton Friedman (1970), from within a utilitarian perspective rather than a virtue-ethical one. But this is like saying the flourishing human is the human who gets the most for themselves, with no concern for others. Following in the Aristotelean tradition, I will assume this is not the correct view of human or organizational flourishing.

   

103

The suggestion, then, is this. As with human flourishing, the right way to approach organizational flourishing is to ask about the proper functioning of an organization of that type. Crucially, for this exercise we do not sort organizations into ‘types’ according to whether or not they are run for a profit—or, more generally, by examining their goals, as that term was used in Section 4.3. Instead, we sort organizations into types according to the human purpose that the organization achieves or is oriented towards. My use of ‘purpose’ here differs from the use of ‘goal’ employed in Section 4.3’s discussion of organizations’ evaluative attitudes: goals are acknowledged by organizations (even if implicitly); purposes need not be. The ‘purpose’ of an organization is what it does for humans. An organization’s purpose is the reason why we do—or, more to the point, should—keep the organization going. The purpose need not be recognized by the organization itself or even by its members or founders. The virtuous organization acts for its purpose because of its purpose. In contrast to this, the vicious organization acts antithetically to its purpose because of reasons that are antithetical to its purpose. I assume all organizations have a purpose in the intended sense. That purpose might be to govern a territory, as in the case of states, or to provide financial products and services, as in the case of banks. As Peter French put it, all for-profit corporations have a purpose akin to ‘make automobiles, airplanes, breakfast cereals, dog food, etc.’ (1984, 102). I assume such purposes apply also to organizations that evolve organically out of human practices or background conventions, such that the organization’s purpose was never explicitly entertained by anyone. Human purposes are even held by organizations that seem, on the surface, to be rotten to the core. Here, we might think about a criminal organizations. Perhaps a criminal organization has the purpose of pursuing the safety or security of its members. A criminal organization’s pursuit of that purpose might be incredibly vicious, insofar as it violates the rights of non-members. But the purpose is a standard by which we can assess the organization. We can hold the organization to that standard. By taking this purpose-driven view on the ‘species’ to which an organization belongs, we can judge an organization’s character by the extent to which it succeeds or fails at fulfilling that purpose for the sake of that purpose. One might doubt that all organizations can be construed as having a legitimate human purpose. If so, then there’s an even broader way in which organizations’ flourishing can be conceptualized. On this approach to

104

  

organizations’ virtue (and vice), we begin by situating an organization within the wider moral community. We then ask how the organization can flourish as the kind of entity it is, taking as fixed its being embedded within and dependent upon that wider moral community. (Here my strategy is similar to Smith’s (2018) strategy for analysing corporate rights, which assigns corporations rights according to their relations with others; my suggestion is that we can do likewise with assessments of corporate virtue and vice.) The idea is that an organization is a type of entity that depends on others’ ongoing support and goodwill—and this is true even if the organization’s sole purpose and true ‘type’ is inherently evil. Thus, when an organization rides roughshod over the broader community upon which it depends, it fails as the kind of entity it is, because it undermines the conditions for its own success. The fact that organizations depend upon others in the moral community in this way can be seen from the discussion in Chapter 2, where I emphasized that a wide range of individuals, objects, and practices are the supervenience base, causes, anchors, and grounds (not to mention the physical parts) of organizations. Without these various other entities, the organization would fail to be the kind of entity it is—because it would fail to exist at all. The vicious organization, then, is the organization that has a deep-seated and long-run disposition to undermine its very foundations— where those foundations reside in the moral community’s ongoing support. In this way, the virtue ethical concept of flourishing gets us a new perspective on organizational wrongdoing. This perspective does not readily fall out of the volitionist or attributivist lenses. The use of ‘flourishing’ also allows us the following nice result: what character traits lead an organization to flourish (or be vicious) is different from what character traits lead an individual human to flourish (or be vicious). As a result, different character traits are virtues and vices for organizations, as compared with human beings. It might also be that different character traits are virtues and vices for different ‘types’ of organization. Perhaps, for example, temperance is a virtue of humans but not of any organizations. And perhaps humility is a virtue for some organizations, but not for others. Thus, there might be something worth pursuing in Moore’s suggestion of ‘craftmanship’ or Schudt’s suggestion of ‘sustainable profit’ as paradigm corporate virtues. I’m neutral on these issues here, since my aim is not to give a full account of exactly which character traits are virtues and vices for exactly which types of organization. My aim has been to demonstrate how the aretaic lens can generally apply to organizations.

   

105

4.5 Conclusion The aim of this chapter has been to demonstrate the range of features for which organizations can be blameworthy. Organizations can be blameworthy for their choices, whether these are reflected in official policy or in the choices of members qua members. The literature on group blameworthiness has tended to focus only on choices, and indeed only on choices that are reflected in official organizational policy. But organizations can also be blameworthy for their unchosen evaluative attitudes. These evaluative attitudes might be reflected in the goals the organization pursues, even if the organization has never had the occasion to consider, defend, or denounce those goals and the attitudes they reflect. Organizations’ evaluative attitudes can also be exhibited in the attitudes of members-qua-members, where these might not have been subject to either organization-level or individual-level control. Finally, organizations can be blameworthy for character traits. Character might be a matter of the way in which an organization makes moral choices (or comes to endorse evaluative attitudes), for example if the organization makes choices in a ‘continent’ or ‘akratic’ way. Or, worse, organizations’ vices might be exhibited through their failing to flourish as the type of creature they are—which, for organizations, would mean failing to fulfil their justifiable human purpose for the sake of that purpose, or failing to show proper regard for those on whom the organization depends. My discussion has shown that organizations can be blameworthy even when no member is blameworthy: an organization can make organizationlevel choices that no member individually makes; an organization can inhabit organization-level evaluative attitudes (embedded in its goals) that no member individually inhabits; an organization can exhibit the character traits of continence or akrasia or vice, even when no member has that trait as an individual. Such discontinuity between organizations and members is important for the idea that organizations’ blameworthiness is irreducible to members’ blameworthiness. My discussion has also shown that Part I’s metaphysics has moral implications. In discussing volitionism, I explained that wrongful member-level choices ‘percolate up’ to the organization, at least if those choices are made while the member is performing their role, since members are the material parts of organizations. And in discussing attributivism, I explained that bad member-level evaluative attitudes likewise percolate up to the organization, for the same reason. In such cases, the object of blameworthiness (whether a

106

  

choice or an evaluative attitude) inheres in the members. But the choices or evaluative attitudes of members-qua-members should be attributed to organizations, because the organization is physically present and active when members are performing their roles. This chapter has demonstrated that organizations’ blameworthiness is wide-ranging. Yet there is one large objection to organizations’ blameworthiness. This objection applies to all three objects of blame (the volitionist, the attributivist, and the aretaic) and to all three theories of blameworthiness (the reactive attitudes theory, the impaired relations theory, and the answerability theory). The objection is that an entity’s being blameworthy entails that the entity has certain kinds of mental states—where organizations seem incapable of such mental states, because the mental states require phenomenal consciousness, which organizations lack. It’s easy to see how such a thought arises from each of the views considered in this chapter. Perhaps liability to reactive attitudes, changes in relations, or requests for answers entails that the liable entity is themselves capable of reactive attitudes, interpersonal relations, or demands for answers. One might think organizations are incapable of these things. Scanlon writes that ‘[i]t makes no sense to attribute feelings or emotions to countries or other collective agents, unless what is meant is just that most of the individual members have these feelings’ (2008, 162). Under Scanlon’s theory of blameworthiness, this means organizations cannot be held blameworthy for their attitudes to others, though they are blameworthy for their responsiveness to reasons. Likewise, one might think an organization’s being blameworthy for its volitions, evaluative attitudes, and character traits entails that the organization has the phenomenology of volition, the feeling of evaluative attitudes, and the emotionality that’s part and parcel of some character traits. Organizations seem incapable of such phenomenology, feeling, or emotionality. In the next chapter, I develop, defend, and respond to a version of this objection to organizations’ blameworthiness.

5 Organizations’ Moral Self-Awareness 5.1 Introduction Chapter 1 conceptualized organizations using a functionalist-interpretivist approach to agency. Functionalism and interpretivism are by far the dominant approaches amongst social ontologists who endorse group agency and blameworthiness. Yet functionalist-interpretivism suggests that a robot is an agent. In Section 5.2, I argue that this is deeply morally puzzling. Chapter 4’s three theories of what blameworthiness is, and the three lenses on the targets of blame, seem clearly inapt for robots. Yet robots are relevantly similar to organizations. If we are to vindicate holding organizations blameworthy, we need a principle that can separate organizations’ blameworthiness from robots’ (lack of) blameworthiness. I consider three such principles: that blameworthy entities must be capable of reasoning, that blameworthy entities must be capable of giving reasons, and that blameworthy entities must be capable of being responsive to reasons. Unfortunately, these principles do not suffice to explain why robots lack blameworthiness while organizations have blameworthiness. It will remain somewhat mysterious how Chapter 4’s theories and lenses of blameworthiness can possibly apply to organizations, if we assume that those theories and lenses do not apply to robots. Section 5.3 proposes a principle that solves the mystery. The aim is to vindicate holding organizations blameworthy, even though robots are not blameworthy. The principle states that entities that are blameworthy (under any of Chapter 4’s theories and lenses) must be capable of moral selfawareness. That is, blameworthy entities must be capable of grasping that they have done wrong. I give five reasons in favour of this principle. The challenge for organizations’ blameworthiness, then, is to explain how organizations have the capacity for moral self-awareness. In Section 5.4, I consider and reject two mechanisms for producing organizations’ capacity for moral self-awareness. These mechanisms respectively draw on Hans Bernhard Schmid’s (2014, 2018) notion of ‘plural

Organizations as Wrongdoers: From Ontology to Morality. Stephanie Collins, Oxford University Press. © Stephanie Collins 2023. DOI: 10.1093/oso/9780192870438.003.0005

108

  

self-awareness’ and Raimo Tuomela’s (2006, 2013) notion of ‘we-mode awareness’. I argue that neither mechanism produces quite the organizational self-awareness that we need, if we are to vindicate organizations’ blameworthiness. In Section 5.5, I propose that organizations’ capacity for moral selfawareness is best conceptualized as a capacity of organizations’ membersqua-members—following Part I’s metaphysics on which organizations are present and active wherever members are present and performing their roles. The idea is that the ‘awareness’ part of organizations’ self-awareness resides in members, while the ‘moral’ and ‘self ’ parts reside at the organization level. The proposal will be that—when members are performing their roles and have certain feelings about the organization—the phenomenology or feelings of members are the phenomenology or feelings of organizations. More specifically, I endorse the following view: Organizational Moral Self-awareness: An organization is morally selfaware if, and because, a member is aware that ‘this first-personal agent did (or is doing, or will do) wrong’ where that awareness is had (1) while performing their role in the organization, and (2) while adopting the rational point-of-view of the organization.

Thus, an organization is morally self-aware when a member is self-aware within their role—much as a table is green when one of its legs is painted green while attached to the table. Organizations are capable of the feelings necessary for blameworthiness, if we use Part I’s metaphysics of membership.¹

5.2 Robotic Blameworthiness On the functionalist-interpretivist picture of agency, a robot can be an agent. This hasn’t gone unnoticed in the social ontology literature. In fact, List and Pettit embrace robotic agency. They go so far as to use an example of a simple robot when introducing their theory of agency (2011, 19–20). They claim an agent ‘is a system with . . . representational states [that depict the environment], motivational states [that specify how the agent requires the environment to be], and a capacity to process them and to act on their basis’

¹ This chapter presents an edited version of the argument in Collins 2022a.

 ’  -

109

(2011, 20). They describe a small robot moving around a table-top, on which there are cylinders. Some cylinders are upright while others are on their sides. The robot has arm-like limbs, which it uses to place the sideways cylinders upright. The robot has representational states (that depict certain cylinders as upright and others as sideways) and motivational states (that require all cylinders to be upright). The robot acts on these states by setting upright the sideways cylinders—bringing the world to conform with the motivational states, by relying on the representational states. If this makes for agency, then organizations (as I characterized them) more than qualify: an organization has a complex web of representational and motivational states. It acts on these via its members, as the robot acts via its limbs. This functionalist picture faces problems with blameworthiness. As in Chapter 4, I’ll be neutral and pluralistic about the nature of blameworthiness: perhaps to be blameworthy is to be a fitting target of reactive attitudes (Strawson 1962); perhaps it is to have impaired one’s relationship with others in a way that warrants certain alterations in the relationship from them (Scanlon 2008); perhaps it is to be in principle liable to calls for answer for some aspect of oneself (Smith 2015). Or perhaps some combination of these. I’ll likewise be neutral and pluralistic about the target of blameworthiness: perhaps blame appropriately targets actions that an entity controls, as under the volitionist lens; perhaps it appropriately targets evaluative attitudes that an entity holds, as under the attributivist lens; perhaps it appropriately targets character traits that an entity embodies, as under the aretaic lens. Or perhaps—as I’ve suggested—all three of these. While being neutral on this, I’ll assume blameworthiness is reactive and backward-looking: blame might be fitting even if there are consequentialist reasons not to blame (the deceased can be fitting targets of private blame, though such blame is pointless); conversely, there might be consequentialist reasons to blame an entity, though that entity is not blameworthy (we might blame a toddler as a way of teaching them about blameworthiness). With these various ways of conceptualising blameworthiness in hand, consider List and Pettit’s robot. Suppose some calamity will ensue unless the robot sets the cylinders upright. Suppose the robot is programmed to predict and avoid the calamity, yet it fails to set the cylinders upright. Is the robot blameworthy? No. It is not fitting to react with hostility, anger, resentment, and indignation towards the robot. It’s not fitting to act as an aggrieved party to a relationship. It’s not fitting to demand answers from the robot. If we had such reactions, we’d quickly feel silly. Instead, we would direct our blame at those who engineered the robot—or perhaps we would write off the event as

110

  

a tragedy. And it’s not just that this particular robot is relatively simple. Consider self-driving cars, self-taught algorithms, or other real-life learning machines. These agential systems are highly complex, yet blame is inapt. Even if an algorithm or machine is self-taught—its programmers don’t dictate its beliefs or preferences—we still blame the environment, not the machine, if the machine produces morally bad results. To see this, consider the algorithms used to make parole decisions within the US criminal justice system. They learn by being fed data about released prisoners in many historical cases, including whether those prisoners were re-arrested. The algorithm’s task is to interpret the data and predict whether a current prisoner is likely to be re-arrested if released. The programmers do not control which factors the algorithm reads off the historical cases, or the weight it gives those factors when making predictions about future cases. We can think of the algorithm as having the representational state ‘X is likely to be re-arrested’ and the motivational state ‘X is not released’. Some such algorithms make predictions that are, many argue, biased against black people (Angwin et al. 2016). If we want to attribute blame, we blame the algorithm’s environment or its set-up. We do not blame the algorithm itself, despite its being a self-taught, highly complex, self-shaping (functionalistinterpretivist) agent. Yet if a human judge were similarly biased, we would presumably blame the judge—even though the judge is shaped by their environment and (neural or cognitive) set-up. To be clear, my claim here about the inaptness of blaming machines applies only to present-day machines. We can imagine science fiction scenarios in which entities with artificial intelligence become conscious. By the arguments of this chapter—on which the capacity for moral self-awareness is the crucial necessary condition for blameworthiness—such science fiction entities would be blameworthy. But such science fiction entities would be very different from any present-day machines. Such science fiction entities would also be very different from present-day organizations. They would be far more complex, for a start. So, when I discuss machines, I limit myself to present-day machines. The question is why we don’t blame them—and, more importantly, whether we can answer that question in a way that vindicates blaming organizations.² Why don’t we blame present-day machines? Pettit implies one answer: the cylinder-straightening robot can’t reason. It’s unable to ask itself ² I thank Andy McKilliam for pressing this distinction between present-day and possible robots.

 ’  -

111

‘questions about connections between propositions, say about whether they are consistent or inconsistent, and then do something – pay attention to the inter-propositional relations – out of a desire to have a belief form one way or the other’ (2007, 499). To perform such assessment, the robot would require ‘meta-propositional beliefs’, that is, beliefs about relations between propositions. A meta-propositional belief about propositions A and B might be that A and B stand (or don’t stand) in relations of support, consistency, entailment, and so on. When we reason, we use meta-propositional beliefs to revise our first-order attitudes towards propositions. The revision is aimed at enhancing the relations of support, consistency, and so on, between our first-order attitudes. Pettit suggests such reasoning separates robots from humans, and from organizations (2007, 499–501; similarly List and Pettit 2011, 177–8). Perhaps the incapacity for reasoning drives our reluctance to blame robots, while the capacity for reasoning vindicates blaming organizations. But robots—like humans, like organizations—can be (and often are) reasoners. (List and Pettit deny this of their simple robot (2011, 188), but that’s a stipulation that they could have made false.) Machines can apply modus ponens to the propositional objects of their first-order representational or motivational states; they can revisit their evidence and use this to question their beliefs; and so on. This is how learning works in parole algorithms: the machine is given a data set, which it analyses to reach conclusions. It tests those conclusions by running back through the data, updating the conclusions as need be. The system must decide whether to reject some data as anomalous or adjust the conclusion to accommodate that data. This ‘recursive learning’ process considers the relation between the data and the conclusion. The propositions in the data and conclusion are entertained meta-propositionally. In that sense, machines reason. But they’re not blameworthy. The capacity for reasoning cannot be crucial attribute. Maybe it’s not reasoning, but reasons. Humans can give (moral, justifying) reasons for their behaviour; present-day machines cannot. Now, machines can be programmed to do the right thing—as when Google produces the Samaritans phone number for the search ‘how to kill yourself ’. And if we feed machines only data in which agents act morally, then they can predict what those agents will do in a new situation—thus making (what we might recognize as) a moral judgement about the new situation. But present-day machines are not very good at giving (what we would recognize as) moral justifications for their predictions or performances. Their decision-making is a black box (Krishnan 2020). If the capacity to

112

  

provide reasons is necessary for blameworthiness, and if organizations can provide reasons while machines cannot, then we have a vindication of organizations’ blameworthiness that doesn’t extend problematically to machines. There are three problems with this suggestion. First, if a reason is a proposition that supports an action, then present-day machines can give reasons. My laptop regularly tells me it will shut down ‘so that updates can be installed’. Second, humans—even ethicists, even judges—often can’t provide justifications for their decisions. We go off the feel of a case, or the judgement an ‘ordinary person’ would likely reach, even when we can’t justify that feeling, or don’t know how the ordinary person would justify that judgement. When we do justify our decisions, those justifications are liable to evolutionary debunking, social conformity explanations, and so on. This is particularly true of justifications for blameworthy decisions. So, machines sometimes give reasons, and humans sometimes can’t give (accurate) reasons. Third, we’re ultimately concerned with organizations. The capacity for reason-giving doesn’t look likely to vindicate organizations’ blameworthiness. Organizations often cannot provide reasons for their decisions, because their members endorse the organization’s decision (if they do) for different reasons. I discussed this when characterizing organizations in Section 1.2, since it’s one of the ways organizations should be held apart from smaller and more consensual collective agents. An organization’s decision is often a ‘modus vivendi’, an ‘overlapping consensus’, or an ‘incompletely theorized agreement’ (borrowing terminology from Rawls (1993) and Sunstein (1995)). Thus, when we ask the organization to justify its decision, we’ll often get an answer that appeals to the decision-making procedure—for example, ‘that’s what most shareholders voted for’. After all, if the shareholders voted for mutually contradictory reasons, the organization cannot coherently provide all those reasons. But ‘that’s what shareholders voted for’ doesn’t make a decision justified; it makes a decision legitimate. It’s the equivalent of asking me why I did something, and receiving the answer ‘because that’s what my balance of reasons supported’. The substantive justifying reasons for an organization’s decision are the reasons why shareholders voted for the decision, not the fact that the shareholders did vote for it. But the organization cannot provide those substantive reasons without self-contradiction. So, organizations often cannot give justifying reasons for their decisions. Yet we blame them. One might say: so much the worse for

 ’  -

113

organizations’ blameworthiness. But, as I mentioned, many humans also often cannot give reasons for their immoral actions, yet we blame them. This suggests reason-giving is not the heart of the matter. Third and finally, perhaps an entity is blameworthy only if it has the capacity to respond to reasons (rather than give reasons).³ On one popular view, an entity lacks the capacity to respond to reasons just if ‘his behaviour would be the same, no matter what reasons there were’ (Fischer and Ravizza 1998, 37). That is, responsiveness-to-reasons requires that one has the capacity to make (at least some of) one’s behaviour counterfactually sensitive to reasons. It also requires that ‘the mechanism that actually issues in the action is [one’s] own, reasons-responsive mechanism’ (ibid., 39). Organizations are responsive to reasons (Björnsson and Hess 2016). So, this responsiveness-to-reasons approach would render organizations blameworthy. The problem is: algorithms are also responsive to reasons (ibid., 294). Perhaps algorithms do not conceptualise reasons as reasons, but their counterfactual behaviour varies with reasons, via a mechanism that is their own. Therefore, responsiveness-to-reasons also doesn’t appropriately divide entities along blameworthiness lines.

5.3 Why Moral Self-Awareness Matters We seek a condition on blameworthiness that appropriately renders machines non-blameworthy, while appropriately rendering most adult humans blameworthy. The subsequent task will be to argue that organizations can satisfy that condition. In this section, I argue that the relevant condition is the capacity for moral self-awareness. An entity has moral self-awareness when it has a phenomenal belief-like attitude (‘awareness’) to the proposition ‘I will do wrong’, ‘I have done wrong’, or ‘I am doing wrong’. There are three components to moral selfawareness: a wrongness component (the ‘moral’ part), a first-personal selfidentifying component (the ‘self ’ part), and the phenomenal belief-like

³ According to O’Connor and Franklin (2018), reasons-responsiveness is one of the two main compatibilist accounts of free will. I therefore take the present reasons-responsiveness proposal to subsume the possibility that free will is the crucial necessary condition on blameworthiness that separates machines from humans. The other main compatibilist account of free will is the identification account, which again seems potentially satisfiable by machines. I thank Simon van Baal for pressing the possibility that free will is the relevant dividing line between humans, machines, and groups.

114

  

component that is best captured by the notions of awareness, grasping, feeling, or acquaintance (the ‘awareness’ part). Present-day machines have the capacity for the first two of these, but not the last. This explains why they cannot be blameworthy. Most adult humans have the capacity for all three. This explains why they can be blameworthy. In Section 5.5, I will argue that organizations, too, can have all three. In this section, I’ll explain the three components and motivate their importance for blameworthiness.⁴ Before doing so, I should emphasize that moral self-awareness is not a problem for functionalist-interpretivism as an account of humans’ mentality. It’s a problem for functionalist-interpretivism as it has been applied to collective agents, including organizations—and specifically it is a problem for these entities’ moral blameworthiness. Functionalist-interpretivism may not be an issue for aspects of agency other than moral blameworthiness. Indeed, many philosophers of mind are functionalists or interpretivists about humans’ agency. But they provide functionalist or interpretivist theories that account for humans’ phenomenal experiences, such as moral self-awareness. To account for humans’ phenomenal experiences within functionalist-interpretivism, philosophers rely on human neurology (as overviewed by Wu 2018), or on human brains’ functional complexity, the likes of which is not exhibited by any present-day organizations (List 2018). So existing functionalist or interpretivist theories of mind can (arguably) give us humans’ phenomenology, including humans’ moral self-awareness. But they do so with mechanisms that are a poor fit for organizations. Section 5.5 will give a theory of moral self-awareness that’s tailor-made for organizations. This theory can be endorsed whatever theory one holds about humans’ mental states (including if one rejects functionalist or interpretivist theories of humans’ mental states). So: what is moral self-awareness? And why is the capacity for it necessary for blameworthiness? The ‘self ’ component is familiar from discussions of ‘de se’ beliefs, originating in John Perry (1979). To represent content that’s de se—‘about the self ’—List and Pettit’s robot must be capable of representing not just ‘whichever robot is on the table is about to do wrong’ (a ‘de

⁴ Relatedly, Baddorf (2017) argues that the capacity for phenomenology is necessary for blameworthiness. But his argument relies on the controversial ‘phenomenal intentionality research programme’. My argument has no such reliance and is thereby much more ecumenical. More importantly, Baddorf argues (briefly) that collective agents lack the capacity for phenomenology. I’ll argue organizations have the capacity for moral self-awareness, via the capacities of their material parts. I thereby vindicate organizational blameworthiness, while Baddorf claims to refute it.

 ’  -

115

dicto’ belief), nor even ‘this particular robot, which happens to be on the table, is about to do wrong’ (a ‘de re’ belief). Rather, the robot must be capable of representing ‘whichever robot is on the table is about to do wrong and I am the robot on the table’ or ‘this particular robot, which happens to be on the table, is about to do wrong and I am that robot’. The robot can have such representations in a functionalist sense: it can move to correct itself when it’s about to do wrong, it can say ‘I’m sorry’ when it has done wrong, and it can have the functionally characterized mental states that are entailed by ‘I am a wrongdoer’, such as ‘I am less good than I could have been’ (List and Pettit 2011, 187–8). The ‘moral’ component of moral self-awareness requires that the entity has the concept of ‘wrongness’. Again, a simple robot can satisfy this. As I mentioned above, machines can abide by hard constraints. There is no roadblock to them labelling those constraints ‘what morality requires’ or ‘constraints the violation of which would be wrong’. In this sense, they can possess the concept of wrongness. Satisfying the moral component doesn’t require that the entity has the correct theory of morality, or even that its moral beliefs are ever true: a blameworthy entity might be systematically incorrect about whether it has done wrong, as callous humans are. Indeed, some callous humans never even bother to operationalize a flawed concept ‘wrongness’. Thankfully for their blameworthiness, my claim is that blameworthiness presupposes the capacity to operationalize the concept of wrongness. So the moral component is relatively easy to satisfy. The awareness component requires the entity to have a belief-like grasp, feel, or acquaintance towards the proposition ‘I have done wrong’ (or ‘I am doing wrong’ or ‘I will do wrong’). The awareness, grasping, feeling, or acquaintance is ‘belief-like’ in that it has a mind-to-world direction of fit. But unlike the notion of ‘belief ’ I have been employing so far, this ‘belieflike’ state has a particular phenomenology. The phenomenology is perhaps impossible to analyse precisely, but there are evocative examples. Consider David Bourget’s comments on the size of the Sun: I can draw numerous inferences from the proposition : I can infer that the Sun has a volume greater than 1.3  10¹⁸, that the Sun has a radius of more than 100,000 km, and so on. I can also draw non-deductive inferences. For example, I can infer that the Sun is larger than my house, that I could not eat the Sun, and that the Sun would make a bad tennis ball. The relevant deductive and nondeductive inferences that I am in a position to make can be multiplied

116

  

ad nauseum. I am also able to behave as required by such reasoning when relevant. My behavior and inferences with regard to the proposition that the volume of the Sun is 1.412  10¹⁸ km³ are about as rational as can be expected of anyone towards any empirical proposition they believe, yet, intuitively, I do not really grasp how big the Sun is. (2017, 300)

Bourget is adept with the size of the Sun, just as a robot can be adept with its wrongdoing. A robot can draw all sorts of inferences, and perform all sorts of behaviours, based the belief ‘I have done wrong.’ So it can believe ‘I have done wrong’, in a non-phenomenal sense of ‘believe’. This gives an entity a lot, in terms of its relation to that proposition. But there is something missing. The size of the Sun is beyond Bourget’s awareness, grasp, feel, or acquaintance. Likewise, a present-day robot’s wrongful agency is beyond the robot’s awareness, grasp, feel, or acquaintance. This is because present-day robots cannot be aware of, grasp, feel, or be acquainted with any proposition. Here, I’m using ‘acquaintance’ to evoke the relation Bertrand Russell (1912) thought connected us to qualia or sense data. That is: there is something distinctively phenomenological or consciousness-based about awareness. At its most intuitive, moral selfawareness can be thought of as a component (though probably not the entirety) of self-directed reactive emotions like remorse, guilt, or regret. If one conjures up the feeling of these emotions, one will conjure up awareness, grasping, feeling, or acquaintance towards ‘I have done wrong.’⁵ Three clarifications will be helpful at this point. First, my suggestion is not that an entity must have the capacity for moral self-awareness in order to master moral concepts. Perhaps present-day robots can master moral concepts, because perhaps mastery is characterizable in functionalist or interpretivist terms. Instead, my suggestion is that an entity must have the capacity for moral self-awareness in order to be apt for blame. There may be entities that have mastery of moral concepts but that are not apt for blame (such as machines). Second, my suggestion is that the capacity for moral self-awareness is necessary (not sufficient) for blameworthiness. Perhaps blame-apt creatures need to grasp not just their own wrongdoing, but also moral reasons or wrongness. If so, we have every reason to think that organizations’ grasping of moral reasons or wrongness will work in the way I outline for ⁵ Hess (2010) argues organizations can satisfy the ‘self ’ component of moral self-awareness. But it’s the ‘awareness’ part (not the ‘self ’ part) that really causes trouble.

 ’  -

117

organizations’ moral self-awareness in Section 5.5. But I focus on moral self-awareness—rather than awareness of moral reasons or wrongness— because self-awareness requires reflexivity (the ‘self ’ component), whereas awareness of moral reasons or wrongness does not. As Section 5.5 will explain, this reflexivity produces some strangeness in organizations’ moral self-awareness: members have to be aware as the organization. Thus, organizations’ moral self-awareness is more difficult to argue for than their awareness of moral reasons or wrongness. This is why I focus on the former, even though my account is extendable to the latter. Third, an entity can be blameworthy in a context without being morally self-aware in that context. Instead, the capacity for moral self-awareness is presupposed by blameworthiness, because that capacity is necessary for being the kind of entity that can be blameworthy. The claim is not that an entity is blameworthy only when that entity in fact has the feeling and belief that it has done wrong. That claim is obviously false. Instead, the claim is that we are blameworthy only if we have the capacity for the feeling and belief that we have done wrong. This will be particularly important for the below discussion of organizations’ capacity for moral self-awareness: an organization might have this capacity—in virtue of the fact that at least one member has the capacity to instantiate the organization’s self-awareness—despite not exercising the capacity on a given occasion. On such an occasion, the organization would still be blameworthy, on my account, even though it is not self-aware on that occasion. It would be blameworthy in virtue of its capacity. There are at least five reasons for believing that blameworthiness presupposes the capacity for moral self-awareness. First, imagine a science fiction scenario in which the table-top robot acquired the capacity for moral selfawareness: suppose it could feel ‘I have done wrong.’ And suppose it could have awareness of moral reasons and wrongness. Suppose a moral calamity would ensue unless the robot set the cylinders upright, where the robot had been programmed to detect and avoid the calamity. Then, I contend, we’d blame the robot if it failed to set up the cylinders. In this way, the thesis that ‘blameworthiness presupposes the capacity for moral self-awareness’ doesn’t rely on moral self-awareness being impossible in robots. It merely relies on the idea that if robots became morally self-aware, then they would meet one of the presuppositions of blameworthiness. Numerous science fiction works explore this idea. If blameworthiness presupposed the capacity for moral self-awareness, then that’d give us a good explanation of our hesitancy around robots.

118

  

Second, if blameworthiness presupposed the capacity for moral selfawareness, that would explain why some philosophers think blameworthy organizations must be ‘mysterious’, ‘incorporeal’, ‘ghostly’ entities, ‘hovering’ over and above their human constituents. (The quoted words come respectively from Rönnegard 2015, Rönnegard and Velasquez 2017, Velasquez 2003, and Ludwig 2017b; all quoted in Hess 2018a, 36.) The explanation would be: these philosophers infer moral self-awareness from blameworthiness, and what really seems mysterious is organizations’ moral self-awareness. The only way organizations could have the capacity for moral self-awareness (so we commonly think) is by being mysteriously ghostly and hovering. After all, organizations lack organization-level neural and physiological systems. It’s human-level neural and physiological systems that allow for non-mysterious, naturalistic explanations of humans’ moral self-awareness (whether those explanations are functionalist, interpretivist, or use some other theory of mind). So, organizations’ moral selfawareness seems mysterious. If such self-awareness were presupposed by blameworthiness, that would explain why blameworthy organizations might be thought to be mysterious. Third and conversely, sometimes our insistence that organizations not be mysterious causes us to consider retracting our blame. This is perhaps what the First Baron Thurlow was getting at when he famously said that groups cannot be punished, because they have ‘no soul to be damned and no body to be kicked’ (quoted in Coffee 1981). A more contemporary version of Thurlow’s point is: collectives lack the neural and physiological systems that provide naturalistic bases for humans’ moral self-awareness (whether those explanations are functionalist or non-functionalist); so, they cannot be appropriately blamed or punished. When attempting to blame entities like banks, someone might reasonably think ‘I could blame Westpac Bank for the wrongs uncovered by Australia’s Banking Royal Commission, but where is Westpac? There’s nothing there. Blaming Westpac is like blaming a ghost. Westpac cannot look me the eye and self-consciously say “I did wrong.” The CEO can do that—but the CEO of Westpac is not Westpac.’ This Thurlow-style reasoning would make sense if blameworthiness presupposed the capacity for moral self-awareness.⁶

⁶ One might object: we blame unstructured groups, such as lynch mobs or misogynists, without the shadow of a presupposition that the group has the capacity for moral selfawareness. Two responses. First, blame might be a mistake when it targets unstructured groups

 ’  -

119

Fourth, as I mentioned above, moral self-awareness can be conceptualized as a component of attitudes like remorse, guilt, and regret. Thus, if blameworthiness presupposes the capacity for moral self-awareness, that would gel with views on which the capacity for remorse, guilt, or regret is necessary for blameworthiness (e.g., Shoemaker 2011). Indeed, David Shoemaker (2019) argues that sincerely felt remorse is a psychological precondition to forgiveness, and so that sincerely felt remorse is necessary for blameworthiness. Of course, some philosophers give non-phenomenal accounts of remorse, guilt, or regret—allowing organizations to have these attitudes (e.g., Gilbert 2002; Huebner 2011; Björnsson and Hess 2016). But these accounts are highly revisionary. In Section 5.5, I’ll suggest that we needn’t be revisionary. We can accommodate organizations within the standard picture of emotions, on which emotions involve phenomenology (Scarantino and de Sousa 2018). The point here is that, if moral emotions involve phenomenology, then this section’s thesis dovetails with a view on which the capacity for moral emotions is necessary for blameworthiness. Fifth, consider the function of blame. According to popular accounts (e.g., Macnamara 2011, esp. 90–1; Shoemaker 2015, 110, 171; Fricker 2016), one important function of blame is to induce the target to feel remorse. This might be so, regardless of whether blame involves reactive attitudes, changed relationships, or demands for answers. This view on the function of blame doesn’t imply that an attitude is blame only if it succeeds in fulfilling this function. Instead, the idea is that an important function of blame is to induce felt remorse, analogously to how an important function of knives is to cut. If an important function of knives is to cut, then the apt target of a knife is something that has the capacity to be cut. Air is not an apt target for a knife. Likewise, if an important function of blame is to induce felt remorse, then an apt target of blame is something that has the capacity to feel remorse. Again, it’s plausible that part (though perhaps not all) of what it means to feel remorse is to have moral self-awareness: to grasp ‘I have done wrong.’ In this way, my view dovetails with these accounts of blame, as well as the above-mentioned accounts of blameworthiness. I don’t claim to have proven beyond doubt that blameworthiness presupposes the capacity for moral self-awareness. But the five reasons just given place that thesis on a firmer footing—vis-à-vis robots and humans—than as a distinct entities, rather than each member individually (for discussion, see Feinberg 1968; Held 1970). Second, the proposals in Section 5.4 (which I reject for organizations) might capture the capacity for moral self-awareness of unstructured groups.

120

  

any of the three proposals considered in Section 5.2 (the reasoning, reasonproviding, and reason-responsiveness proposals). If blameworthiness presupposes the capacity for moral self-awareness, then the question arises: do collectives have the capacity for moral self-awareness?

5.4 Plural and We-mode Self-Awareness In a series of papers, Hans Bernhard Schmid has developed the concept of plural pre-reflective self-awareness in cases of uncoordinated groups, such as a group of passengers on a train (esp. Schmid 2014; Pettit (2018) uses Schmid’s account to argue for corporations’ indexicality). Plural prereflective self-awareness is not held by any individual, or by the group as a singular entity. Instead, it’s held by individuals as a plurality. The subject and object of awareness is ‘we’ (plural), not ‘I’ (singular), where that ‘I’ might be individual or collective. The awareness is pre-reflective: noninferential, non-observational, direct, and immediate. And the awareness is self-awareness: awareness by us, of us; the subject and object are identical. If the members of Westpac Bank have the capacity for plural pre-reflective self-awareness, combined with some shared conception of wrongness derived from Westpac’s rational point-of-view, then this might vindicate our blame of Westpac. Thus Schmid (2018, 101) imagines a mugging on a train in which passengers gaze out the window. The passengers see the mugging reflected in the window. Each gazer falsely believes that the reflection is happening in a train on the parallel track. Each thinks ‘those people should act to prevent that mugging’. Suddenly, each grasps that they are looking at a reflection. It dawns on each passenger that those people are us. Schmid contends that this is different from each passenger grasping ‘I am in the mugging train.’ It is different because each lacks the power to stop the mugging. Only the plurality can stop the mugging—so, it is only when each realises that those people are us that each can reasonably feel the moral imperative to play a part in stopping the mugging. Because each feels this imperative, the imperative must derive from the plural perspective, which means ‘that’s us!’ must be felt from the plural perspective. If the mugging were to continue, perhaps this would transmute into plural moral self-awareness, in which the plurality grasps ‘we did wrong by inaction’. Somewhat differently, Raimo Tuomela has long-defended the idea that individuals are sometimes self-aware in the we-mode (e.g., Tuomela 2006).

 ’  -

121

Tuomela distinguishes the subject, mode, and content of any given mental state. Suppose I believe I did wrong. The subject of this belief is me, the belief is in the I-mode, and belief ’s content refers to me. Suppose I believe we did wrong. Then the subject remains me, and the mode remains I-mode, but the belief has we-content: it’s a belief about us. Suppose we believe that I did wrong: here, the subject is us, the content contains me, but the mode becomes ambiguous. We each might I-mode believe that I did wrong—we each believe this from our own individual perspective. By contrast, we each might we-mode believe that I did wrong—we each believe this from the group’s perspective. Perhaps we-mode beliefs can produce collectives’ moral self-awareness: although ‘we did wrong’ is grasped by each individual (the individual is the subject), this content is grasped by each individual from the group’s perspective, that is, in the we-mode. Unfortunately, plural pre-reflective self-awareness and we-mode beliefs share a problem: the subject of the awareness is either the plurality (in plural pre-reflective self-awareness) or the individual (in we-mode beliefs). But the subject of an organization’s moral self-awareness needs to be the organization that outlives these specific members—not me, not us, and not me in the organization mode. The special features of organizations (as compared with other collective agents) demonstrate why this is important. First, the organization’s identity stays constant as the membership changes. But as the membership changes, there will be changes in whether it’s appropriate for members to have plural self-awareness of us as wrongdoers, or appropriate for members to be the subject of a we-mode belief that ‘we’ did wrong. New members (who were not around when the wrong was done) were not part of the plurality that did wrong. So it would be irrational for them to have the kind of realization the train passengers have (in plural self-awareness), or to be the subject of a group-mode belief where the group in question did not include them (in we-mode beliefs). Yet we want to hold the organization blameworthy for what it did before the change in membership. It’s true that new members could irrationally have plural selfawareness or we-mode beliefs. But our practices of holding organizations blameworthy shouldn’t depend on members’ capacities to be irrational. Second, consider public discourse around wrongdoing organizations. The public often doesn’t care whether these people acknowledge their wrong: the public wants the organization to acknowledge its wrong. This becomes clearest when considering apologies for historical wrongdoing. Historical wrongdoings of states will be the focus of Chapter 7, where I will focus on organizations’ blameworthiness for historical greenhouse gas emissions. But

122

  

there has been little public discourse around that example. So, to see that public discourse often demands that the organization acknowledge its wrong, consider the language of Australia’s former Prime Minister, Kevin Rudd, when he apologized to the generations of Indigenous Australians that were forcibly taken from their families as children: as Prime Minister of Australia, I am sorry. On behalf of the government of Australia, I am sorry. On behalf of the parliament of Australia, I am sorry. I offer you this apology without qualification. We apologise for the hurt, the pain and suffering that we, the parliament, have caused you by the laws that previous parliaments have enacted. We apologise for the indignity, the degradation and the humiliation these laws embodied. We offer this apology to the mothers, the fathers, the brothers, the sisters, the families and the communities whose lives were ripped apart by the actions of successive governments under successive parliaments. (Rudd 2008, emphasis added)

Although Rudd used the first-person plural ‘we’, it is clear that the entity he took to be giving the apology was the state of Australia—or, at a minimum, the parliament or government of Australia. The statement ‘we apologize’ is syntactically plural but semantically singular, just like the terms ‘sunglasses’ and ‘scissors’. This distinctive organization-level acknowledgement of wrongdoing—and the demand for such acknowledgement, in the face of wrongdoing—is not well-captured by the potential for moral self-awareness of pluralities (Schmid), or the potential for moral self-awareness of individuals in a certain mode (Tuomela). Apologies like Rudd’s are about an organization that acknowledges its wrong. Schmid’s and Tuomela’s accounts give us moral self-awareness with the wrong subject.

5.5 Organization-level Moral Self-Awareness Present-day robots don’t have awareness (grasping, feeling, acquaintance) of any proposition. Organizations might seem to be in the same boat. After all, awareness is phenomenological. But things are not so simple. To see this, we need to take seriously the idea—defended in Chapters 2 and 3—that organizations (unlike robots) have humans as material parts, such that humans are to organizations as planets are to solar systems and as trees are to forests. This picture allows us to view members as the realizers of the organization’s self-awareness, where that awareness is nonetheless properly attributed to the

 ’  -

123

organization. This is an intuitive idea: if a tree contains green, the forest contains green; if a planet supports life, the solar system supports life. Of course, not just any property ‘transfers’ from the constituent to the whole: a given tree might be one metre tall, while the forest is more than one metre tall (because the tallest tree is 20 metres tall); a planet might have a mean temperature of 15 degrees Celsius, while its solar system has a mean temperature much lower than that. For any given property, we need a story about how and why it is transposed from the material part to the material whole. To begin the story, notice that at least some attributes transfer from members to organizations. Consider a job search committee within an organization such as a university. When the chair of the search committee—a human—makes a job offer that she is authorized to make, the university makes a job offer. The chair acts as a limb or organ of the university; her action is ‘incorporated’, in French’s (1984) phrase. In this case, the attribute that percolates up is an action. But the same is true of some mental states, such as beliefs. The search committee has been authorized to find the best candidate for the job. So, when the committee believes a particular candidate is the best candidate, the university believes this. If the committee has just one member, then this is belief percolates from a singular human to the university. But the university doesn’t believe everything the hirer believes. We must place conditions on when the mental states of a member are also mental states of their organization, just as we place conditions on when the physical properties of trees are also physical properties of their forest. These conditions might be different for different mental states. We are concerned with moral self-awareness. And we are looking for a possibility proof of organizations’ capacity for moral self-awareness, because we are looking for a possibility proof of their blameworthiness. So, we are looking for merely jointly sufficient conditions for a member rationally realizing an organization’s self-awareness (‘rationally’ because we don’t want to rely on irrationality, like the appeal to plural pre-reflective self-awareness). That said, we want our jointly-sufficient conditions to capture a good number of cases, so it would be good if something in the vicinity of our jointly-sufficient conditions were necessary. And it would be good if our conditions were somewhat explanatory. I propose the following conditions: Organizational Moral Self-awareness: An organization is morally selfaware if, and because, a member is aware that ‘this first-personal agent did (or is doing, or will do) wrong’ where that awareness is had (1) while

124

  

performing their role in the organization, and (2) while adopting the rational point-of-view of the organization.

This account is designed to ensure that not just any member’s grasping of an organization’s wrong is a grasping by the organization. As explained when I introduced role performance in Section 1.3, an individual ‘performs’ their role when they meet two conditions: first, they perform in ways that are permitted by the role; second, they perform in those ways because of the role. The ‘performances’ here might include mental performances, such as grasping or being aware. Only when the grasping is permitted by, and done because of, the member’s role is the member’s grasping authorized by the collective—just as a search committee’s actions and beliefs are authorized by a university. Of course, the relevant permission (to grasp the organization’s wrong from the organization’s point of view) is unlikely to be written into any member’s contract. More often, the permission will be inferable from the norms, ethos, or culture of the organization. These ‘informal’ role elements are as real as formal ones. And we should require that organizational moral self-awareness might merely be permitted—rather than mandated—by a member’s role, because otherwise it would be too easy for an organization to avoid blameworthiness by failing to mandate any member’s moral selfawareness. The permission is much harder for an organization to avoid, consistent with the freedom that is necessary for high-ranking members (at least) to do their job. If moral self-awareness is had while performing one’s role, then that moral self-awareness is had because of one’s role. This condition can be broken down into two components: the member’s role is part of the correct psychological explanation of the member’s mental state and the member’s role is part of the justification the member would give for their mental state if asked. These two components can come apart, since sometimes the true psychological explanations of our action are hidden from us, even after reflection. The member ‘adopts the organization’s rational point-of-view’ when the organization’s bundle of beliefs, preferences, and other mental episodes is the bundle from which the member reasons. Following Carol Rovane (1998), an organization’s rational point-of-view is made up of all the organization’s beliefs, preferences, and other mental states (where those mental states are characterized in functionalist-interpretivist terms). But to adopt an organization’s point-of-view, the member needn’t have access to all the organization’s mental states. This would be too tall an order. Instead, the

 ’  -

125

member need only be able to adopt enough of the organization’s point-ofview for her to reason as that point-of-view requires. Specifically, a member who houses the collective’s moral self-awareness must have access to the collective’s beliefs about the wrongful action in question, including the collective’s belief that the action is wrong. Both conditions (1) and (2) admit of a ‘subjective’ and an ‘objective’ reading. That is, an individual can believe she acts while performing her role and believe she is adopting the organization’s point-of-view. Or an individual can act in a way that is objectively permitted by, and because of, her role (the latter because we can be mistaken about our true motivations) where she has adopted the organization’s point-of-view without realizing it. The member’s moral self-awareness is the organization’s moral self-awareness when both the subjective and objective readings of the two conditions are true. This makes the conditions somewhat difficult to satisfy—but, again, the conditions are intended as jointly sufficient.⁷ Both of the conditions rely upon the member’s relation to the organization’s structure: the member’s role is a matter of her place in the structure, and she can adopt (enough of) the organization’s point-of-view only if she is privy to the right parts of the organization’s structure. In this way, the transfer of moral self-awareness from a member to an organization is interestingly disanalogous to the transfer of colour from a tree to a forest. The transfer of colour from tree to forest does not depend on the structure of the forest or on the tree’s particular relation to the structure of the forest (other than that the tree is a part of the forest).⁸ This disanalogy demonstrates that I have set a relatively high bar for members to have to meet, in order for them to instantiate the organization’s self-awareness. This should increase our confidence in the account, because it does not allow too many mental states to percolate from members to organizations. Conditions (1) and (2) have some precedent in the literature. Condition (2) is similar to what List and Pettit (2011, 191–2) briefly discuss as identification with a group agent, though they do not suggest anything like a transfer of mental states. Tollefsen (2006, 235, fn. 12) perhaps comes closest to what I suggest, saying collective guilt is ‘felt “through” . . . members. But insofar as it is produced by collective acts, results in ⁷ The subjective components do not make self-awareness voluntary, that is, they do not imply that a member can instantiate organizational self-awareness entirely at will. The subjective components require that the member has certain beliefs. I assume occurrent beliefs are not voluntarily chosen. ⁸ I thank Simon van Baal, Will Cailes, and Andy McKilliam for discussion.

126

  

assessment of the group, and it is the group which has the ability to get rid of such guilt (via reparations, collective apology, and so on), the guilt “belongs” to the group itself.’ Yet Tollefsen says this briefly in a footnote, and her latter three conditions don’t make the feeling of guilt (roughly equivalent to the grasping of moral agency) group-level rather than individual-level. My account of members as material parts renders the feeling group-level, just as a tree’s greenness is forest-level. Also, the three properties that Tollefsen says make the guilt ‘belong’ to a collective are non-phenomenal properties. Indeed, Tollefsen suggests a non-phenomenal view of guilt, which robots are capable of, when she says ‘[c]ollective guilt is functionally similar to individual guilt and thus is a genuine form of guilt’ (2006, 234). Elsewhere, she considers (without defending) a similar approach, which again ‘does not have the group itself feeling the emotion’ (2015, 132). In these ways, my account differs from—and is more radical than—Tollefsen’s.⁹ Delicacy is required in characterizing organizational moral selfawareness. We need to take particular care with characterizing who ‘this first-personal agent’ is, in the member’s awareness that ‘this first-personal agent did wrong’. Is the first-personal agent the member, in which case, how can this realize the organization’s being aware of itself? Or is the firstpersonal agent the organization, in which case, how is the realizer the member’s moral self-awareness? The second option is correct: the member is aware of the organization, and so the realizer of the organization’s moral self-awareness is not properly construed as the member’s individual moral self-awareness. Instead, the member is morally self-aware as the organization. Regarding the three parts of moral self-awareness: the ‘moral’ part is the organization’s functionally characterized concept of wrongness, as determined by its rational point-of-view; the ‘self ’ is the organization; and only the grasped, felt, or acquainted ‘awareness’ inheres in the member. More concretely, when a member grasps ‘we did wrong’ while conditions (1) and (2) are met (in both the subjective and objective readings), then the ‘we’ refers to the collective as a singular entity. In English, this grasping is expressed using ‘we’ rather than ‘I’—but the ‘we’ is the first-person singular. This makes sense of language such as former Prime Minister Rudd’s, when he says ‘we, the parliament’. The parliament is a singular entity, yet the firstperson plural happens to be used to refer to it.

⁹ I provided an account more like Tollefsen’s—on which members feel only ‘membership emotions’—in Collins 2018b. The present chapter revises that view.

 ’  -

127

Thus, if a member of Westpac Bank is aware of (feels, grasps, is acquainted with) ‘we did wrong in how we incentivized our financial advisors’—and if this awareness is had both while performing the member’s role and while adopting Westpac’s rational point-of-view—then Westpac grasps ‘I, Westpac, did wrong in how I incentivized my financial advisors.’ Importantly, this mental state has a phenomenal component. Westpac’s mental state is realized in the human member, which gives us a physiological and neural basis for this phenomenology. Functionalists or interpretivists about human phenomenology can breathe easy, while those who hold other views of human phenomenology can account for the human member’s phenomenology in their usual way. If Westpac weren’t made of people, its moral self-awareness would be mysterious, ghostly, and hovering. But since Westpac is made of people, nothing so strange is happening when Westpac is aware of, grasps, or feels that ‘I, Westpac, did wrong.’ Assuming apologies such as Rudd’s are sincere, what we see in such statements is an individual being the realizer of an organization’s awareness that ‘I, Australia, did wrong.’ (Of course, ‘we’ is the word used; again, my suggestion is that ‘we’ is sometimes semantically singular.) Rudd—we hope—grasps, feels, or is aware of Australia’s wrongdoing. What’s more, he is aware of this as Australia, because he uses the first-personal pronoun of the collective while performing his role and while adopting Australia’s rational point-of-view. Taking the apology at face value, Australia has been morally self-aware.¹⁰ As emphasized in Section 5.3, blameworthiness presupposes that an entity has the capacity for moral self-awareness—not that the entity actually is morally self-aware in each and every context where it does wrong. But so far, I’ve only explained the conditions under which an organization actually is morally self-aware. When does an organization have the capacity for moral self-awareness? Simply, it has that capacity just if two conditions are met: first, it bears (in a functionalist or interpretivist way) the concepts of ‘wrongness’ and ‘self ’; second, at least one member has the capacity to be the realizer of the organization’s moral self-awareness. Of course, organizations are large, so it will be unlikely that all members realize the organization’s moral self-awareness at a given time, or even that all members have the capacity to do so. This might seem contradictory: the

¹⁰ Being the realizer of an organization’s moral self-awareness is orthogonal to being the proxy or spokesperson of the organization, as that notion is analysed by (e.g.) Ludwig 2014.

128

  

organization is at once morally self-aware and not morally self-aware; or capable of it and not capable of it. But there is no contradiction here. We can usefully view the members of an organization at one time as analogous to the temporal parts of an individual across time (Dietz 2020). There are times when I am explicitly morally self-aware, times when I am implicitly morally self-aware, and times when I am not morally self-aware at all. This is a ‘contradiction’ only if we insist on asking whether I am morally self-aware in some time-neutral sense. Likewise, there is a contradiction within organizations only if we insist that all co-temporal parts of an organization are perfectly in-sync. But the various parts of any agent are almost never insync. One member realizing the organization’s moral self-awareness is enough for the organization’s moral self-awareness. This unlocks two virtues of my analysis of Organizational Moral SelfAwareness. First, we can easily give a gradable account of organizations’ moral self-awareness: an organization is more morally self-aware if a higher proportion of members realise its moral self-awareness. And an organization is more capable of moral self-awareness if a higher proportion of its members are capable of realizing its moral self-awareness. All else being equal, an organization is a more open to blame when it does wrong if it has a higher proportion of members that are capable of realizing its moral self-awareness. The idea of being ‘more open to blame when one does wrong’ is familiar from our treatment of each other: the average fifteenyear-old is more open to blame when she does wrong than is the average ten-year-old. The second virtue concerns the differential capacities of differentlyranked members. When we blame organizations, we often want highlyranked members to realize the organization’s self-awareness. My conditions explain this. Recall that the ‘moral’ and ‘self ’ components of moral selfawareness are given a functionalist-interpretivist reading: these components are characterized by their functional role in the organizational agential system, including their role of causing the organization to be disposed to certain behaviours (such as apologies, reparations, and preventions of future wrongdoings), alongside the interpretations that others would make of the organization. In many organizations, only highly-ranked members have the capacity to produce the relevant functional role and interpretations, so only highly-ranked members are capable of realising the ‘moral’ and ‘self ’ components of moral self-awareness. When we demand moral self-awareness from an entity like Australia or Westpac, we don’t just want the phenomenal awareness of ‘I, Australia, have done wrong.’ We also want the functional

 ’  -

129

and interpretive profile that comes with moral self-awareness. So we go after the Prime Minister, not the average citizen on the street.¹¹ One might object that I have set the bar too high, if my aim is to provide sufficient conditions that are explanatory and that approximate necessity. Specifically, one might worry that organizations will easily be able to set things up such that no member has the capacity to realize the organization’s moral self-awareness, so that the organization lacks the capacity for moral self-awareness, so that the organization cannot be blameworthy. Most straightforwardly, perhaps an organization can do this by never introducing the concept of wrongness into its rational point-of-view. Perhaps criminal organizations do this, or cut-throat for-profits. Yet surely such organizations are blameworthy. So my account must be wrong. My answer is two-fold. First, organizations that lack the capacity for moral self-awareness are extremely rare in the real world, perhaps to the point of being a mere theoretical possibility. Even the Mafia has definite views about moral right and wrong, so it has the capacity for moral selfawareness (even if those beliefs are systematically mistaken). And almost all corporations pay lip service to values such as human rights, seemingly committing to a functionally characterized belief that human rights should be upheld. As thin as such commitments are, they let the concept of wrongness into the organization’s rational point-of-view. So almost all real-world organizations have the concept of wrongness. As explained in Section 4.3, an organization’s evaluative attitudes can exist even if no member has voluntarily, consciously, or culpably chosen to act on those evaluative attitudes. Second, for the few organizations that genuinely lack the capacity for moral self-awareness, I bite the bullet: they are apt targets of containment, deterrence, and restraint—but not of blame. Crucially, though, there will almost always be blame in their vicinity. It’s plausible that agents have obligations to teach organizations the capacity for moral self-awareness. These agents include founders, managers, regulators, shareholders, and stakeholders—all when acting outside their organizational roles, if they have such a role, since (ex hypothesi) moral self-awareness cannot get a foothold from within their roles. If those agents have reneged on those

¹¹ Importantly, this does not mean that organizations’ blameworthiness is under the voluntary control of such high-ranking members, since what matters is the members’ capacity to realize the organization’s self-awareness, and members do not straightforwardly control for whether or not they have this capacity.

130

  

obligations, then those agents are blameworthy. This is analogous to parents being blameworthy for not teaching their children the capacity for moral self-awareness. I do not theorize the blame of these supporting agents in this book, just as I do not theorize the blame of those who cause, subvene under, anchor, or ground organizations. But as I hope to have made clear throughout, such external-yet-supportive agents will often be on the hook for their distinctive relationships to organizational wrongs.

5.6 Conclusion Philosophers who endorse the blameworthiness of collective agents uniformly insist that the capacity for phenomenology is not necessary for blameworthiness. In this chapter, I have argued—contrary to this literature—that a functionalist-interpretivist picture of moral agency is inadequate for capturing blameworthy agents, since present-day robots and self-taught machines can satisfy the functionalist-interpretivist picture, yet they are (I assume) not blameworthy. I have argued instead that blameworthiness presupposes the capacity for moral self-awareness—the capacity to grasp one’s wrongful agency. I gave five reasons for believing that blameworthiness presupposes the capacity for moral self-awareness. If this is correct, it explains why we do not blame present-day machines. However, this raised a thornier question: how do organizations have the capacity for moral self-awareness? I argued that existing concepts in the literature—namely, plural prereflective self-awareness and we-mode self-awareness—are not adequate to capture organizations’ moral self-awareness. I proposed a new route to organizations’ moral self-awareness: via members that grasp content while performing their roles, and while adopting the organization’s perspective, organizations can grasp their own wrongful agency. The blameworthiness of organizations is thereby vindicated.

PART III

MEMBERS

6 How Members Are Implicated 6.1 Introduction The argument so far has conceptualized organizations as unified wholes, with members as material parts. On this picture, when an organization does wrong, members are part of a thing that has done wrong. This is enough to ‘implicate’ a member in an organization’s wrong, in a broad and weak sense of ‘implicate’. If your agency is bound up with the agency of a wrongdoing organization—via your membership in that organization—then any number of responses might be appropriate on your part. First, it might be appropriate for you to feel something like Bernard Williams’ (1981) ‘agent-regret’ with regard to the organization’s wrongdoing. You are not straightforwardly blameworthy for your organization’s wrongdoing; your organization’s wrongdoing is not your fault. But you are like a truck driver who accidentally kills a young child while driving conscientiously, to use Williams’ example. Your agency is bound up with a terrible event—in this case, your organization’s wrongdoing. This warrants distinctive reactive attitudes, despite the fact that blame might not be apt. Second and relatedly, it might be appropriate for you to instantiate the organization’s moral self-awareness, as that was theorized in Chapter 5. Indeed, your membership in an organization is what makes it possible for you to instantiate the organization’s moral self-awareness, where such instantiation is impossible for non-members. Third, it might be appropriate for outsiders to come to you with questions, complaints, and demands for explanation, where it would not be appropriate for them to bring those complaints to other outsiders. After all, you are a material part of the organization: where else should they take their complaints? Such interrogative interactions with outsiders are likely familiar to anyone who has ever travelled abroad and faced questions about their state’s policies. Of course, such interrogation sometimes goes too far—especially if one is not a member of one’s state, for example because one’s state is not properly democratic

Organizations as Wrongdoers: From Ontology to Morality. Stephanie Collins, Oxford University Press. © Stephanie Collins 2023. DOI: 10.1093/oso/9780192870438.003.0006

134

  

(as discussed in Section 3.5). Still, the picture of organizations, membership, and wrongdoing from Parts I and II does vindicate some extent of this practice. The previous paragraph applies to all members of wrongdoing organizations. But for many organizational wrongs, we want to let many members off the hook as individuals: these members did nothing to contribute to the wrong, or they didn’t support the wrong, or they couldn’t have done anything to distance themselves from it. Such members are—as I’ll use the term—not ‘implicated’ in the organization’s wrong, even though they are a material part of an organization that did wrong and even though they might appropriately have the responses of agent regret, of organizational moral self-awareness, and of receiving complaints from outsiders. In general, then, it is important to distinguish (i) the general question of whether one’s agency is bound up with the organization so as to make one a member from (ii) the specific question of whether one’s agency is implicated in some specific wrongdoing perpetrated by the organization (where that wrongdoing might be an action, an evaluative attitude, or a failure of virtue—as explained in Chapter 4). For an example of a non-implicated member, we might think of cashiers in banks—or even managers of cashiers—when it comes to wrongs that are primarily operationalized in the parts of the bank that deal with financial advice or home loans. Or we might think of present-day green-lifestyle citizens who are part of a state that bears attributivist blameworthiness for historical greenhouse gas emissions. In Chapter 7, I will argue that many such individuals are in fact liable to bear reparative burdens. Someone can be liable to reparative burdens for all sorts of reasons—for example, because they benefitted from a wrong or have the capacity to repair a wrong. Chapter 7 will argue for those justifications of members’ reparative burdens. But those justifications do not render an individual implicated in the wrong itself, as this chapter will develop that concept. In this chapter, I provide a taxonomy of the ways in which members can be implicated in the wrongs of their organizations—where this means something more than merely being a member of an organization that did wrong. I’ve already outlined a few ways in which implication can happen. In Chapter Four, I explained how members’ wrongful choices and wrongful evaluative attitudes are attributable to the organization, when members are performing their roles. I gave the example of financial advisors who make choices or have evaluative attitudes (while performing their roles) that treat customers as profit-fodder. Such members are implicated in organizational

   

135

wrong in a quite direct sense. But members can be implicated in more ways than this, or so this chapter will argue. I begin in Section 6.2 by analysing how members can be the enactors of an organization’s wrong. Implication as a principal enactor is the most obvious way in which a member can be implicated in an organization’s wrong. Principal enactors include the imagined macho financial advisors discussed in Chapter 4. When a member is a principal enactor of an organization’s wrong, the member makes choices, inhabits evaluative attitudes, or bears character traits while performing their role, such that those choices, evaluative attitudes, or character traits are wrongful and attributable to the organization. There is also another, less obvious, mode of enactment. This is complicitous enactment, in which a member performs their role in such a way as to be complicit with the principal enactors. I canvass various types of complicitous enactment. Section 6.3 argues for an even less obvious kind of implication: implication via endorsement. Here, a member has tacitly exchanged commitments (either with another member, or with the organization itself) to a certain organizational goal, where a wrong is enacted as part of the organization’s reasonably predictable pursuit of that goal. As we’ll see, endorsement differs from enactment in that endorsement often isn’t a matter of how one performs one’s organizational role. Instead, a member often tacitly exchanges commitments—thereby becoming an endorser—by entering a role, rather than by performing that role. Finally, Section 6.4 argues that even members who are not implicated as enactors or endorsers can be implicated as omitters. These are members who feasibly could have performed their role in the organization in a way that took steps against an organizational wrong, but who failed to perform their role in this way. One way to think about the distinctive status of omitters is the following: while enactors and endorsers take active steps in favour of an organizational wrong, omitters fail to take active steps in opposition to an organizational wrong. Omitters fail to do what they ought to have done to make themselves complicit in, or to endorse, rightfulness (rather than wrongfulness) in their organization. Of course, these categories do not exhaust the wrongs of organizations. Following Part I’s metaphysics, the actions, evaluative attitudes, and character traits of principal enactors, complicitous enactors, endorsers, and omitters are all properly attributed to the organization (if those actions, attitudes, and traits are manifested while the members are performing their roles). These wrongful attributes are held both by the member and by the

136

  

organization. But organizations’ wrongs also go beyond these four categories of member-manifested wrongs. Organizations’ wrongs include the actions, evaluative attitudes, and character traits that can be attributed only to the material structure as a whole, not to any individual members. I discussed these wrongs in Chapter 4. This chapter is focused on members, but this focus should not mislead us into thinking that wrongful organizations are nothing more than a bundle of implicated members. My view—laid out in Parts I and II—is that organizations are structured wholes that can have morally wrongful features, even without any members having morally wrongful features.¹ That remains true, even as we now turn our focus to how members can do wrong.

6.2 Enactors Christian List (2021) distinguishes three ways in which members may be assigned responsibility for the actions of a group agent. It’s worth noting that List uses a volitional model, which ignores groups’ attitudes and groups’ character traits. This produces a relatively narrow range of enactors, since he risks overlooking the members that are implicated in groups’ attitudes or character traits. Conversely, List is concerned with group agents in general, rather than organizations in particular. This produces a potentially broader range of enactors than one might expect to find in organizations. Despite these disanalogies between List’s purposes and mine, List’s distinctions are a useful starting point for thinking about enactment. List suggests that: the members can be held responsible for a group agent’s actions only to the extent that they have played certain normatively relevant roles: an enacting role, for instance as a knowing, willing, and uncoerced manager, official, or representative of the group agent, an authorizing role, for instance as a director, board member, owner, share-holder, or regulator, or an organizational-design role, for instance as a founder, framer, policy maker, or institutional designer. (2021, 1224–5)

¹ A now well-worn example of organization-level wrong without member-level wrong is the ‘discursive dilemma’ (discussed by Pettit 2007). See also Copp 2007.

   

137

Clearly, although List aims to capture all groups agents, he has corporations specifically in mind. Even within this narrow focus, there are several ways in which Part I’s ontology splices members’ implication in a different way than List suggests. First, for List, the ‘enactors’ are paradigmatically managers, officials, or representatives. I argued in Section 3.3 that non-manager employees are sometimes members of organizations, so I would include those individuals as potential knowing, willing, and uncoerced enactors of organizations’ actions. What’s more, I included directors, board members, owners, and shareholders as members on a par with other kinds of members, since these people meet the commitment, inputs, and enactment conditions outlined in Section 3.2. Therefore, I would not draw a sharp line between List’s ‘enactors’ and ‘authorizers’. While it’s true that List’s authorizers might make higher-level decisions than List’s enactors, all of these role-occupants can perform in ways that are attributable to the organization. These performances include actions, evaluative attitudes, and character traits. All the role occupants that List takes to be enactors—and all (except regulators) that he takes to be endorsers—are potential enactors, according to my account. This is because they are all members of organizations, on my account. In my view, regulators are not members of organizations: they do not meet the commitment or enactment conditions on membership, even if a case can be made for their meeting the inputs condition. Thus, regulators are not members, let alone members that are implicated as enactors of organizational wrongs. Regulators can, of course, be blameworthy for the regulations that they implement. But in that case, they are blameworthy for the implementation of regulations—not directly for what an organization does on the basis of those regulations. Regulators are not implicated as authors of organizations’ wrongs. Those who are implicated as authors must be internal to the organization, which regulators are not. To see that implication-as-an-author requires internality, consider that my parents and friends are not the authors of my actions, no matter how much influence they have over me. They are not authors of my actions because they are not internal to me. Similarly, regulators are not implicated as enactor-authors, because they are not internal to the organization. Again, this is not to say they are not blameworthy. But their blameworthiness is not the concern of this book.² ² I thank Simon van Baal, Will Cailes, Andy McKilliam, and Andris Raivers for pressing the blameworthiness of regulators and other supportive outsiders.

138

  

The same goes for those List assigns an ‘organizational design’ role. Founders, framers, organizational policy makers, and organizational designers might be members—but they are not members simply in virtue of being founders, framers, policy makers, or designers. These roles alone do not implicate them as author-enactors in the organization’s wrongdoing. At most, those roles implicate them in the founding, framing, policy making, and designing of organizations from which they are wholly separate. Here, it’s important to recall Section 2.2’s discussion of causation, supervenience, and anchoring. Many regulators, framers, policy makers, and designers are causes, supervenience bases, and anchors of organizations. They can be blameworthy for how they perform those roles in relation to organizations. But those are external roles—roles that do not render their performers members. External performances are not attributable to the organization itself, however much they constrain or guide the organization. Again, these agents are important but beyond the scope of this book. In short, those who enact an organization’s wrong are members who perform their roles in wrongful ways. Recall that members perform their roles just when they meet two conditions. First, they act, or hold an attitude, or exhibit a character trait, within their roles. That is, the organization’s structure and decision-making procedure permits someone with that role to have that action (attitude, trait), where the organization permits it because the action (attitude, trait) mirrors, promotes, fulfils, or respects (or so on) the organization’s rational point of view. Second, the act (attitude, trait exhibition) is held by the member because of their role. That is, the member’s role is at least some small part of the actual psychological explanation for why the member performed the wrongful action (or inhabited the wrongful evaluative attitude or exhibited the wrongful character trait). At this point, I should note a few things about this characterization of ‘performing’ a role. Notice that an organizational role can permit an action (or attitude or trait) even if that trait is never the subject of explicit or implicit instruction from the member’s superiors within the organization. Notice also that a member’s action (or attitude or trait) can be performed partly because of the member’s role, even if the member doesn’t know it. Psychological explanations are often beyond the access of those about whom they are explanations. Thus, many actions, attitudes, and traits will count as members ‘performing’ their role, and will therefore amount to enactments that are attributable to the organization, even if the organization or member has never explicitly considered the reasons for the performance, or its attributability to the organization.

   

139

All role performances by members are enactments that are properly assigned to the organization. All role performances implicate members as enactors of organizations’ performances. But which role performances are appropriately viewed as enactments of organizational wrongdoing? To answer this, consider an example: the forging of a customer’s signature by a financial advisor, in order to change the client’s investment from a low-risk to a high-risk product. Clearly, the advisor themselves is an enactor of this wrong. But many other enactments (role performances) are related to this wrong. As Australia’s Banking Royal Commission put it when discussing cases of this kind, That generally similar conduct occurred in all of the major entities [banks and financial institutions] suggests that the conduct cannot be explained as ‘a few bad apples’. That characterisation serves to contain allegations of misconduct and distance the entity from responsibility. . . . The misconduct acknowledged by the major entities gives rise to broader questions than those answered by the ‘few bad apples’ response. (2018, 87–8)

At the same time, it’s implausible that (say) a bank cashier should count as an ‘enactor’ of signature forgery, if the forgery was committed in a completely separate division of a huge organization. The question becomes: if the enactment of wrongdoing is not just a matter of ‘a few bad apples’, but if not all members are enactors of a given wrong, then how do we draw the lines around which role performances count as enactments of the wrongdoing? We have to begin by drawing careful lines around the wrongdoing itself. What wrongdoing are we concerned with? The answer to this will enable us to pick out the enactors of that wrongdoing. If the wrongdoing is something as precise as ‘forging this particular customer’s signature’, then there really is only one enactor of the wrongdoing: the individual who performed the forgery. But we are usually not concerned only with such specific wrongdoings. We are concerned with broader wrongdoings—still internal to the organization, still committed by the organization—that are (in some sense) related to these specific wrongdoings. Such broader wrongdoings get us a broader pool of enactors. We need to know how to draw a line around the enactors of these broader, related wrongdoings. These are the enactors the Royal Commission was alluding to when it resisted the ‘few bad apples’ approach.

140

  

I suggest we draw this line using the concept of complicity. Here I draw upon Chiara Lepora and Robert E. Goodin’s (2013) analysis.³ As Lepora and Goodin define it, ‘complicity’ is a matter of standing in a certain relation to the principal author of a wrong. For our purposes, the ‘principal’ wrongdoer is the enactor of the specifically circumscribed wrong—something like forging a specific signature. So: what makes one member complicit in another member’s principal enactment of an organizational wrong, such that the former as well as the latter is an enactor of an organizational wrong that’s relevantly related to the principal wrongdoing? Lepora and Goodin begin by distinguishing ‘complicity’ from ‘joint authorship’. Joint authorship comes in four species: full joint wrongdoing (in which multiple parties make identical contributions to wrongdoing), conspiracy (in which parties together agree to act wrongly), cooperation (in which parties share the planning and execution of wrongdoing), and collusion (cooperation which is secretive) (2013, 36–41). In all these species, joint authorship implies two or more agents who work together in the execution of some wrongdoing. But this is just to say that we have two or more enactors of the principal wrongdoing—for example, two or more forgers of a signature, or two or more collaborators in a scheme of signature forgery. So, the concept of joint authorship isn’t right for capturing the broader pool of members whose enactments render the organization’s wrong of forging signatures more than just ‘a few bad apples’. In addition to joint authorship, Goodin and Lepora distinguish four ‘conceptual cousins’ of complicity: conniving, condoning, consorting, and contiguity (2013, 44–51). Conniving involves deliberately ignoring, or tacitly consenting to, a wrong. Condoning involves forgiving or pardoning a wrong. Consorting involves habitually spending time with those who do wrong. And contiguity involves proximity to the wrongdoer, where this is indicative of a bad character or of tacit assent to the wrongdoing. All four complicity-cousins are ways in which members can perform their organizational roles, so as to be enactors of an organization-level wrong that is relevantly related to the principal wrong. Which role-bearers are likely to be complicit in a wrong? The managers of the principal enactors are the most likely candidates for conniving or condoning: superiors have the power to ignore, consent to, forgive, or pardon a wrong. Meanwhile, the immediate colleagues of the principal enactors are better candidates for consorting or

³ A different—and more monolithic—conception of complicity is proposed by Christopher Kutz (2000). I have critiqued Kutz’s account elsewhere (Collins 2017a; Collins and de Haan 2021).

   

141

contiguity, since they are likely to spend more time with the principal wrongdoer or be more proximate to them. Importantly, only one of the four complicity-cousins is readily enacted by those who are underlings of the principal enactors. Underlings cannot meaningfully ‘ignore’ or ‘consent to’ a wrong by their superior (whistleblowing notwithstanding—I’ll discuss this briefly in Section 6.4). Underlings cannot meaningfully ‘forgive’ or ‘pardon’ a wrongdoing of their superior. And underlings’ proximity to a superior’s wrongdoing is unlikely to be indicative of the underling’s bad character. Usually, the only way an underling can engage in a complicity-cousin (vis-à-vis a wrongdoing that is principally enacted by the underling’s superior) is by ‘consorting’ with that superior. But even here, consorting is wrongful only if it is not justified or excused (the same is true of conniving, condoning, and contiguity). Underlings will often be justified or excused for their consorting. Of course, if the underlings have a reasonable option to leave the organization, then they might count as connivers, condoners, or being wrongfully contiguous. But leaving the organization will often be too much to ask—in which case, underlings’ enactment will be excused and therefore not a type of implication. Thus, the four conceptual cousins of complicity often rightly get lowranking members off the hook for being implicated in wrongs that are principally enacted by their superiors. Last but not least, we should consider complicity simpliciter, as distinct from joint wrongdoing and as distinct from the four conceptual cousins of complicity. According to Lepora and Goodin, complicity simpliciter exists only when none of the four conceptual cousins exist and when there is not joint authorship. Those who are complicit know (or should have known) that by acting (or holding an attitude or character trait) they further the intentions of the principal wrongdoer. This ‘furthering’ happens insofar as the complicitous actors ‘contribute causally (or could have been expected to contribute causally) to the implementation of the principal wrongdoing . . . often perform contributory acts that “give access” to the principal wrongdoing, facilitating it or perhaps even making it possible. . . . induce or incentivise the wrongdoing . . . or encourage it . . . or make it easier to perform’ (Lepora and Goodin 2013, 41–2). Complicity is important for organizational wrongdoing, particularly in the example of Australia’s Banking Royal Commission. Recall from Section 3.3 that a core concern of the Commission was the remuneration of advisors, brokers, and so on. Advisors were often paid more if their advisees took actions that brought business to the organization. The architects of such remuneration policies are prime targets for complicity. Theirs is

142

  

a clear case of inducing or incentivizing a principal wrongdoing, such as forging signatures. Financial incentives—in the form of an increased paycheck—is an obvious way to induce or incentivize an action. Thus, complicity simpliciter helpfully picks out members who are more directly causally linked to the principal wrongdoing than are the connivers, condoners, consorters, or contiguous members. In this way, complicity and its conceptual cousins pick out a broad range of enactors, who each enact part of a broad organizational wrong. That organizational wrong can be characterized as the organization’s supporting the principal wrongdoing, by having material parts that perform their roles in ways that support the principal wrongdoing (where the principal wrongdoing is also the performance of an organizational role). Yet complicity and its cousins are also noteworthy for the range of members they exclude. In any wrongdoing committed by a large organization, there will be many members who are not joint authors; nor connivers, condoners, consorters, or contiguous members; nor complicitous. These members are not properly viewed as enactors—either of the organization’s principal wrongdoing or of the organization’s wrong of supporting that principal wrongdoing. In sum, the enactors of organizational wrongdoing are those members who perform their organizational roles so as to (i) commit a principal wrongdoing, or (ii) be complicit (including the complicity-cousins) in a principal wrongdoing that another member commits via role performance. From this, it follows that a member enacts an organizational wrong (either as principal or as complicit) only if the member performs their role. Role performance is crucial for enactment, because members’ role performances are the only actions of members that are attributable to the organization. So it is only those wrongs that members commit via role performance that are enactments of organization-level wrongdoing (where that organization-level wrongdoing might be the principal wrongdoing itself, or might be the support of that principal wrongdoing via complicity and its cousins). However, members can be implicated in organizations’ wrongdoing without being enactors of that wrongdoing. The rest of this chapter discusses the two other modes of members’ implication in organizations’ wrongs: endorsement and omission.

6.3 Endorsers Endorsement differs from enactment primarily in the following two features: endorsement can occur when one takes up a role in an organization

   

143

(rather than just when one performs a role one already occupies) and endorsement can occur any length of time after a wrong has been enacted (rather than proceeding or being proximate to a wrong). In general, then, endorsement involves a less direct connection to the organization’s wrong, as compared with principal enactment or even complicitous enactment. Nonetheless, I’ll now argue, members who are endorsers of specific organizational wrongs are implicated as agents in those very wrongs.⁴ In the most general terms, endorsement occurs when two or more entities exchange commitments to one another to respond positively to a permissible goal. Thus, commitment to a goal is a three-place relation: an agent, A, commits to another agent, B, to a goal, G. Commitments to another to a goal are ‘exchanged’ when A commits to B to G and B commits to A to G.⁵ As I’ll use the term, ‘endorsement of wrongdoing’ occurs when reasonably foreseeable wrong arises from one or more of the agents’ positive responses to the goal over which they have exchanged commitments. Positive responses to a goal include responses such as realising, pursuing, endorsing, or maintaining the goal. I will be interested in cases where A is a member and B is an organization. But the notion of ‘endorsement’ is a general one: endorsement can hold between two individuals, between two organizations, or indeed between a non-member individual and an organization. Although endorsement can occur between all these different agents, I think endorsement importantly captures the manner of implication-in-wrongdoing that is often acquired when an individual takes up membership in an organization (even though membership is not always voluntary, such as membership in nation-states). Endorsement is therefore worth spelling out at length, if we are interested in cataloguing the ways members can become implicated in specific wrongs of organizations. Endorsement is a fairly minimal notion. If A and B exchange commitments to one another to a goal, then this does not entail that A and B have committed to respond positively to the goal together, or with we-intentions (Tuomela 2007), or as a plural subject (Gilbert 1989), or as a joint action (Bratman 2014), or with participatory intentions (Kutz 2000)—though it doesn’t exclude these things, either. Endorsement also does not mean that

⁴ What follows is an edited version of the view presented in Collins (2017a, 585–91). ⁵ My particular use of ‘commitment’ rules out commitments to oneself, because this use is engineered to explain how we can become implicated in others’ wrongs by endorsing those wrongs.

144

  

either A or B cause the other to respond positively to the goal. Each agent’s committing need not even cause the goal itself. Instead, if A commits to B to G, then (constitutively) A has assured B that, or pledged to B that, or promised to B that, or invited B to rely on or trust in the fact that, she is disposed to respond positively to the goal. This is neutral on the relation between such notions as promising, relying, assuring, raising expectations, and trust (as examined by, e.g., Scanlon 1998, 295–327). Such debates notwithstanding, the crucial point is this: by exchanging commitments to the goal, each agent gives the other agent standing to hold her to account if she does not respond positively to the goal in some way or other. The precise ‘some way or other’ will vary depending on the agents, their goal, and their context. Some goals are to be realized, others pursued, others recommended, others (for example, impossible goals) merely desired, and so on. For some goals, any such response is appropriate and the nature of the positive response may be a matter of discretion. By committing, the agent may ‘sign up’ to one or more of these responses (perhaps without determining which). When A and B have exchanged commitments to one another to G, then they are each permitted to rely on, or trust that, or consider themselves assured (as appropriate) that each is disposed to respond positively to G. I use the notion of ‘commitment’ rather than ‘promise’ to keep open the range of ways in which one might commit to another to a goal. The exchanged commitment need not be as explicit as promises typically are. Depending on context, a commitment might arise with a mere glance or might arise gradually over time.⁶ It might arise by cheering on, or expressing solidarity with, a cause. In some contexts, we can commit simply by doing precisely the things that committing commits us to do in the future. In organizations, we most often exchange commitments to goals by joining the organization—paradigmatically, by signing an employment contract— where it is common knowledge in the context that the organization responds positively to the goal. For example, if an academic signs a teaching-andresearch contract with a university, then she and the university have exchanged commitments to the pursuit of certain generally specified goals, such as the goals of research and teaching. ⁶ Here I follow Gilbert (1989). My proposal differs from Gilbert’s notion of joint commitment, in that if A and B each commit to the other to G, then (in my view) there are two commitments. In contrast, for Gilbert, if A and B jointly commit to G, then there is just one commitment, held by the plural subject, to do something as a plural subject. My account is thus thinner and perhaps easier to satisfy.

   

145

To see how the exchange of commitments implicates us in one another’s wrongs, take an example where both A and B are individual agents. Suppose it’s new year’s eve and we are discussing new year’s resolutions. I say to you, firmly, ‘I’m going to aim at cutting emissions this year. That’s a commitment.’ You say the same to me. We have now exchanged commitments. Given our exchanged commitments, I may rely on you to pursue emissions reduction to at least some extent and in some ways. (Let’s assume pursuit is the appropriate positive response to this particular goal in our context.) I may hold you to account if you entirely fail to do so. You are in a state of accountability to me. You are in my pocket and I am in yours. We are on the hook to each other for our respective pursuit of this goal. An attractive flipside of this mutual vulnerability-to-reproach is mutual claim-to-support. If you take up cycling as a way of reducing emissions, but then get a tyre puncture, then I should help you repair it insofar as I can. If I must decide how to limit my attendance at overseas conferences, in order to fly less, then you should offer to help me decide which conferences to forgo. If you find vegetarianism horrible, then I should listen sympathetically and make suggestions where I can. The prospect of such support is one reason why we exchange commitments to others to goals in the first place. If I don’t offer such support, then you can question whether my commitment to emissions reduction was genuine in the first place (and you can beret me for it not being). This mutual claim-to-support arises in cases where the commitments are exchanged (i.e., from A to B and from B to A), even if it does not arise in unilateral commitments (i.e., cases where A commits to B but not vice versa). When we exchange commitments to each other to a goal, you and I compose a coalition that is unified by that goal (in the terms of Collins (2019a)). You and I are a coalition in the sense that both (1) you and I are commonly known to each hold a particular goal and (2) you and I are disposed to work with each other to realize that goal. I’ve argued elsewhere that members of coalitions are rationally permitted—and sometimes, morally required—to engage in a form of ‘we-reasoning’ with respect to their shared goal (Collins 2019a, ch. 5). When someone we-reasons, he ‘considers which combination of actions by members of the team would best promote the team’s objective, and then performs his part of that combination’ (Sugden 2003, 167). In these ways, if you and I have exchanged commitments to a goal, then you and I hold a weak type of proto shared agency. This proto shared agency arises from: our common goal, our common dispositions to predict, rely

146

  

upon, and reinforce each other’s responses to that goal, and our common rational availability of we-reasoning decision-making processes. The combination of proto shared agency, mutual vulnerability-to-reproach, and mutual claim-to-support, provides a rational and normative architecture that involves our agency in one another’s positive responses to the goal. With this in place, suppose that you do wrong while responding positively to our shared goal, in ways that I could reasonably have been expected to foresee. Then, I am implicated as an endorser of your wrong. My agency’s relationship to your wrongdoing is not simply a matter of bad luck, nor is my relationship to your wrongdoing straightforwardly causal (as in complicity). Even the label of ‘endorsement’ is slightly misleading: it is not your wrongdoing that I endorsed; what I endorsed was the goal in relation to which you did wrong. Still, my agency is connected to your wrongdoing in a way that implicates me. The idea of reasonable foreseeability is important here. I am implicated as an endorser of your wrongdoing only if your wrongdoing was reasonably foreseeable to me at the time that we exchanged commitments to the relevant goal. Take the example in which you and I exchange commitments to reduce emissions. Suppose that, in order to reduce emissions, you drive to visit your family (rather flying, which would be more convenient). You uncharacteristically drive recklessly, as a result of which you run over a small child. In that case, your wrong (driving recklessly) was not reasonably foreseeable to me at the time of our exchanged commitments. By stipulation, your recklessness was uncharacteristic. In that circumstance, I am not implicated in your wrong. It must have been reasonably foreseeable to me that you would do wrong in your positive response to our shared goal. How does this apply to members of organizations? As I mentioned above, perhaps the most common way of exchanging commitments with an organization is to take up a job at that organization. In taking up a job, a person often—though not always—exchanges commitments with the organization to some goal or goals. The goal might be something like providing financial products and services, or (in the case of civil servants) governing a territory for the good of residents. If the organization—which is one of the parties to this exchanged commitment—reasonably foreseeably commits wrong in the course of its positive response to the goal, then the other party to the exchanged commitment (the employee) becomes implicated as an endorser. Importantly, the ‘goals’ at issue here are different from organizations’ ‘purposes’, as those were characterized in Section 4.4’s discussion of organizations’ virtue and flourishing. Recall that an organization’s purpose is what

   

147

it does for humans. It’s the reason why we do or should keep the organization going. The purpose need not be recognized by the organization itself or indeed by its members or founders. In Section 4.4 I noted that organizations’ goals are acknowledged by organizations (even if implicitly), while purposes need not be. Endorsement concerns organizations’ goals, not their purposes. To see how implication-via-endorsement operates in organizational contexts, consider another example from Williams (1973, 97). Williams asks us to imagine George, who has recently graduated with a PhD in chemistry. Should George take up a job working in biological and chemical warfare? If George doesn’t take this job, then someone else will—so it’s not plausible that George would thereby become complicit simpliciter in the wrongs of warfare. We can also imagine that the warfare organization is so large that George would not be conniving, condoning, consorting, or contiguous with the members who are principally enacting the organization’s wrongs— George doesn’t satisfy any of the conceptual cousins of complicity. In this way, George is not an enactor of the organizations’ wrongs. If George is implicated in the organization’s wrongs, then he is not implicated as an enactor. There seems to be something wrong with George’s taking the job. Williams says that the problem is that ‘each of us is specially responsible for what he does, rather than for what other people do. This is an idea closely connected with the value of integrity’ (1973, 99). But, of course, what George does in taking the job is precisely to connect himself problematically to what other people do. The problem, I suggest, is that George’s taking the job involves George’s exchanging commitments with the organization to the goals of biological and chemical warfare, where he can reasonably foresee that the other party to the exchange of commitments (i.e., the organization) will respond to those goals in a way that includes wrongdoing. This is just to say that George is implicated as an endorser in the wrongs. As a vivid real-world example, consider the practice of ‘earning to give’. This is a practice in which someone who wants to do a much good as possible takes a lucrative job at (for example) an investment bank. The person then donates a large portion of their high salary to povertyalleviating charities. The justification for this practice is that it is sometimes the most effective way in which a person can do good. By taking a highpaying job at a morally dubious company, this ‘effective altruist’ does much more good—improves others’ lives much more—than they would have if they had taken a job at a poverty-alleviating charity, or if they had volunteered for such a charity on weekends (Todd 2014).

148

  

One intuitive objection to ‘earning to give’ runs as follows. Those who earn to give are somehow implicated in the wrongs that are perpetrated by the organizations for which they work. These altruists are part of the problem of poverty, rather than being part of the solution. Their actions do not challenge the powerful forces that create poverty, but rather ‘align themselves’ with precisely those forces (Srinivasan 2015). The idea of implication-via-endorsement gives us a way to make this intuitive objection more precise. I suggest that implication-via-endorsement captures the way in which those who earn-to-give are implicated in the wrongs of the organizations for which they work—even if those who earnto-give do not themselves enact any wrongs in their performance of their organizational roles. By taking a job at a morally dubious company, the altruist has exchanged commitments with the company to a certain end (such as investment). They are thereby implicated in the wrongs that it was reasonably foreseeable the company would commit in the company’s positive response to that end. Am I suggesting that all employees are implicated as endorsers of all organization-level wrongs? No. An organization like a bank might have numerous goals, many of which conflict with one another, and many of which the bank has exchanged commitments to with only some of its members. For example, it has the goals of maximizing profits and of keeping customers happy. Someone who works as a night security guard might have committed to neither of these goals. A customer service representative might have committed to only the customer satisfaction goal. An executive board member might treat customer satisfaction as purely instrumental to the goal of maximizing profits. Are all of these individuals implicated as endorsers when the bank misleads customers? No. Exactly what it takes to exchange commitments to a particular goal is a highly context-dependent and contestable matter. And it’s an empirical question exactly which goal a bank was responding to when it misled a customer. Roughly, we can judge whether commitments have been exchanged to a particular end by asking whether (i) a culturally informed reasonable person, or an ideal observer, or an idealized inter-subjective deliberation process, would reach the conclusion that commitmentexchange has occurred; and (ii) if the agents were to honestly and coldly reflect (assuming this were possible), they would acknowledge the goal as one to which they have exchanged commitments. Merely taking up a job will not always be sufficient to endorse all organizational goals, especially in tight job markets. (In the above discussion of earning-to-give, I was assuming the

   

149

employee could have found reasonable work elsewhere.) And we can judge which goal an organization was responding to (when it performed a particular wrong) only by looking at the actual mechanisms that lead to that wrong, alongside assessing counterfactuals about the circumstances in which the wrong would have been avoided. Often, the goal that the organization was pursuing when it did wrong will be one that is narrower than— or anyway different from—the goals that many employees endorsed when they started their jobs.

6.4 Omitters There is a final kind of implication that I wish to analyse: implication-viaomission. As the label implies, implication-via-omission is relevant to (some) members who do nothing in relation to an organization’s wrong. Omitters are often too distant from an organizational wrong to be its principal enactors or even its complicitous enactors. And they haven’t performed the behaviours necessary for implication via endorsement. Indeed, they haven’t performed any behaviours at all. They have stood by while wrong happened within the organization. In discussing enactors and endorsers, I focused on employees at corporations. In discussing omitters, I will focus on citizens in liberal-democratic states. Together, I hope this spread of examples demonstrates the breadth of this chapter’s account of members’ implication in organizations’ wrongs. In what follows, I will use ‘citizen’ as shorthand for ‘person who is eligible to register to vote at not-disproportionate cost’. Following the arguments of Section 3.5, I will assume that citizens meet the conditions for being members of their states. Apathetic or indifferent citizens are almost always too distant from a state’s wrongdoing to count as its principal enactors or even as its complicitous enactors. And these citizens do not engage in actions that might produce implication-via-endorsement: these citizens do not vote, or petition, or otherwise indicate support of any particular state goals or policies. Nor do most citizens ever ‘join’ the organization, so they do not exchange commitments by that route. (The exception is naturalized citizens, but most citizens are not naturalized.) Nonetheless, I’ll argue, apathetic or indifferent citizens are often implicated in their states’ wrongs by omission. To see this, consider an election in which a terrible candidate is elected to lead a liberal-democratic state. For our purposes, there are three relevant

150

  

citizens—they are Voti the voter, Inti the intentional non-voter, and Uni the unintentional non-voter. Voti votes for the terrible candidate. Inti intentionally omits to vote, on the considered basis that politics is a waste of time. Uni unintentionally omits to vote, simply because it doesn’t occur to her, though her social-political context was such that it easily could have occurred to her. Assume that both Inti and Uni would have voted for a good candidate, if they had voted at all. And all three citizens could easily have reasoned their way into voting for the good candidate. In this particular election—like in almost all real-world elections—no individual’s actions or omissions are but-for causes of the wrongful outcome, that is, no individual’s action or omission is necessary for the terrible candidate being elected; still less are any of their actions or omissions necessary for the terrible candidate’s being the principal enactor of organizational wrongs. Voti is an easy case: he’s a complicitous enactor of the wrongs that are principally enacted by the terrible leader.⁷ To see this, recall Lepora and Goodin’s account of complicity. Citizens such as Voti ‘contribute causally (or could have been expected to contribute causally) to the implementation of the principal wrongdoing [enacted by the leader] . . . perform contributory acts that “give access” to the principal wrongdoing, facilitating it or perhaps even making it possible. . . . induce or incentivise the wrongdoing . . . or encourage it . . . or make it easier to perform’ (Lepora and Goodin 2013, 41–2). None of this requires that Voti’s vote is a but-for cause of the election result. An organization member can be a causal contributor, facilitator, inducer, incentivizer, encourager, or cost-reducer for a wrongful outcome that is principally enacted by another member—and thereby be complicit in that wrongful outcome—even if the complicitous member’s action makes only the slightest expected difference to the probability, or ease, of the principal enactment.⁸ What about Inti and Uni, each of whom omit to vote? Are they complicit in the organizational wrongs that are principally enacted by the leader? It’s not clear that complicity can operate by omission. After all, contributing,

⁷ Beerbohm (2012) and Kutz (2002) each argue similarly, using different notions of complicity. ⁸ Here I am side-stepping the growing literature on ‘no-difference’ cases: cases where it seems like a group of people (such as voters) cause an outcome even though no individual in the group makes any causal difference to the outcome (for an overview, see Nefsky 2019). I assume that an appropriately massaged account of causation will include all the people in the group as causal contributors, facilitators, inducers, or so on. Thanks to Will Cailes for raising the relevance of this literature.

   

151

facilitating, inducing, incentivizing, encouraging, and cost-reducing all look like actions. At the very least, these forms of complicity look like behaviours that are not omissions—even if they are unintentional behaviours, and therefore not actions. We need an argument for why, and when, organizations’ members who omit to perform actions related to an organization-level wrong are implicated in that very wrong. To start, what are omissions? According to Carolina Sartorio’s (2009) ‘causal’ theory of action, actions are caused by intentions, while omissions are caused by an absence of an intention. Sartorio provides the following example. Suppose I am sitting on the edge of a pond in which a young child is drowning. If I save the child, then my saving of the child happens because I intended to save him: the action is caused by my intention. By contrast, if I eat an ice-cream rather than save the child, then what’s the causal explanation of my letting the child die? My letting the child die happens because I failed to intend to save him. My letting the child die didn’t happen because I intended to eat the ice cream. Sartorio argues that ‘she ate an ice cream’ is too specific as an explanation of why I let the child die. After all, the child would have died if I had read a book, taken a walk, had a nap, and so on. All these possibilities are united by the fact that, in all of them, I fail to intend to save the child. So, to give an adequately general and robust causal explanation of my letting the child die, it is appropriate to refer to the absence of my intention to save him—rather than referring to the presence of my intention to eat ice cream, which isn’t a general or robust causal explanation of why I let the child die. This account of omissions has the following crucial implication: when we are determining whether someone omitted to do something, it doesn’t matter why the person omitted to do that thing. All that matters is that there was an absence of an intention on the person’s part to do the thing. For example, when we ask whether I omitted to save the child, all that matters is that I failed to intend to save the child. It doesn’t matter that I instead ate an ice cream, or read a book, or so on. In particular, Sartorio’s account implies that an absence of an intention can itself be intentional or unintentional: one can intentionally omit to do something (as Inti does) or one can unintentionally omit to do something (as Uni does). Both Inti’s and Uni’s failures to vote are explained by the absence of an intention to vote—even though Inti had an intentional absence of an intention to vote, while Uni had an unintentional absence of an intention to vote. Most importantly, Sartorio’s account of omissions implies that all citizens who fail to vote are implicated in the wrongs that flows from the election (assuming they had

152

  

reasonable opportunity to vote, such that it was their failure to intend that truly explained their failure to vote). Sartorio’s account gives us a broad and inclusive pool of members who are implicated via omission. But to what extent are the different citizens implicated? Suppose Voti, Inti, and Uni all have exactly the same capacity and opportunity to vote. Despite this similarity, it might seem that Voti is more implicated in the leader’s actions than either Inti or Uni (because Voti is implicated-viacomplicity), while Inti is more implicated than Uni (because Inti’s implication-via-omission is intentional). To see this intuition in a simpler example, imagine three ways in which a person might be related to a drowning child: the person might have pushed the child into the water; or the person might have seen the child drowning and decided not to save them; or the person might have been inattentive to the drowning. Surely, the first person is more implicated than the second, and, surely, the second person is more implicated than the third. Likewise, one might think, those who are complicit in their organization’s wrongs (like Voti) are more implicated than those who intentionally omit to take action to avoid complicity (like Inti). And those who intentionally omit to take action against complicity (like Inti) are more implicated than those who unintentionally omit to take action against complicity (like Uni). The analogy to the pond seems apt, at least upon first glance. This perhaps natural intuition is somewhat supported by results in experimental philosophy. In surveys on people’s reactions to trolley problems, it has been found that people prefer harms by omission over harms by action (Cushman, Young, and Hauser 2006). And in surveys on vaccination, it’s likewise been found that people prefer to cause harm by omitting to vaccinate children and thereby causing disease, rather than causing harm by vaccinating children with vaccines that are themselves harmful (Ritov and Baron 1990; Baron and Ritov 2004). If the participants in these surveys are to be trusted, then Voti is more implicated than their Inti or Uni. As for the comparison between Inti and Uni, we might note that the law generally treats intended harms as worse than unintended harms. In that’s correct, then Inti is more implicated than Uni. Despite these perhaps natural intuitions, I suggest that Voti, Inti, and Uni are equally implicated in the wrongs of their organization. This is because all three culpably fail to take action against complicity with the principal enactor of the organizational wrongs. To see the argument for this, let us assume that Voti is implicated, and then consider whether there are any morally significant differences between Voti, Inti, and Uni. If there are no

   

153

morally significant differences, then, I suggest, all three are equally implicated in the wrongs that flow from the election. First, let’s take the comparison between Inti and Uni: why think there are no morally significant differences between them? Randolph Clarke has put the answer well. Here’s what he would say about Uni, who unintentionally omits to vote: Provided that the agent has the capacities which make her a morally responsible agent, she is blameworthy for [an omission that isn’t intentional and of which she is unaware] if she is free in failing [to do] the thing in question and if her lack of awareness of her obligation to do it—and of the fact that she isn’t doing it—falls below a cognitive standard that applies to her, given her cognitive and volitional abilities and the situation she is in. (Clarke 2014, 167)

That is, Uni could easily have thought about voting and could easily have voted well, given her cognitive and volitional abilities. She has no excuse for not having done so. She could at least have thought about things to the extent of Inti, who considers her options, weighs them up, and ultimately decides to omit to vote because politics is pointless. A different way to put this is: at least Inti troubled to think about the issues. Uni hasn’t even bothered to do that! Inti might be blameworthy for having thought about it badly, but likewise Uni is blameworthy for not having thought about it at all. Both agents are blameworthy for something. Given that they had equal capacity and opportunity to vote well, their respective blameworthiness for these different things are on a par. Indeed, Uni should remind us of Angela Smith from Section 4.3, who negligently failed to remember her friend’s birthday. Under the attributivist lens on blameworthiness, such failures can be just as blameworthy as intentional actions or intentional omissions. What about the comparison between Inti and Uni on the one hand, and Voti on the other hand? This distinction—between actors and omitters—is probably morally significant in single-agent cases. Compare someone who pushes a child into a pond with someone who omits to rescue the child from the pond (whether intentionally or unintentionally). In such single-agent cases, actions and omissions play different causal roles. The action of pushing a child into a pond plays the causal role of initiating or triggering a threat. By contrast, the omission to rescue the child plays the causal role of facilitating, or enabling, or failing to intervene upon, an existing threat (Sartorio 2005; Lombard 1990). These are quite different causal roles.

154

  

But these differences in causal roles don’t map onto cases in which the actor is merely complicit in the principal wrong, while the omitters fail to take actions against complicity. That is: Voti isn’t well-understood as a threat-initiator or as a threat-trigger, analogous to a person who pushes a child into a pond. Voti is merely complicit. Remember, his vote is not a butfor cause of the election result, so he does not initiate or trigger the wrongs that are principally enacted by the election winner. Rather than being analogous to someone who pushes a child into a lake, Voti is more analogous to someone who cheers on the pusher from the sidelines. Meanwhile, Inti’s and Uni’s omissions facilitate or enable or fail to intervene upon the election outcome, just as Voti’s action does. Inti and Uni are not analogous to people who fail to rescue a child from a pond (intentionally or unintentionally), because the election result is overdetermined by the pattern of votes. Because none of the voters’ actions or omissions stand in but-for relations to the election result, we should resist drawing sharp analogies to the pond case. This leaves us without positive reason for thinking that Voti is more implicated than Inti and Uni. But we are still without positive reason for thinking that all three are equally implicated. We need some story about what grounds Inti’s and Uni’s implication, so as to vindicate the presumption that all three are equally implicated. The story I propose is as follows: Inti and Uni failed to fulfil obligations to perform their roles in such a way as to be complicit in opposition to organization-level wrong. They violate obligations to act against organization-level wrongdoing; rather than, like Voti, having positively acted in favour of organization-level wrongdoing. When the stakes are the same—as they are in this case, since each vote counts for one and none for more than one—these two forms of implication are equally morally weighty. This suggestion does not completely deny the moral significance of the distinction between actions and omissions. Instead, it says that membership in an organization is a special context that generates duties to act. When a member omits to discharge these duties, her omission implicates her in the associated wrongdoing, just as it would have if she had performed an equally efficacious action in favour of that wrongdoing. More generally, then, my suggestion on implication-via-omission is this. Members have obligations to—insofar as possible, and within reasonable cost—perform their roles in an organization so as push against complicity in, or endorsement of, wrongdoing. ‘Pushing against complicity’ means taking active steps to: avoid causally contributing to the enactment of a principal wrongdoing; avoid performing contributory acts that ‘give access’ to the principal wrongdoing; avoid facilitating the principal wrongdoing or making it possible; avoid inducing or encouraging the principal wrongdoing;

   

155

and avoid making it easier to perform. ‘Pushing against endorsement of wrongdoing’ means exchanging commitments—with the organization, with other members, or even with outsiders—to goals that run counter to organizational wrongdoing. The failure to push against wrongdoing is an omission that implicates one those wrongdoings, in the specific case of one being a member of the wrongdoing organization. As with enactors and endorsers, there are caveats to place on the extent to which omitters are implicated. First, it’s important that omitters are implicated only if they could have pushed against organizational wrongdoing within reasonable cost. In real life, it will not always be possible to push against wrongdoing within reasonable cost. Consider citizens in liberal-democratic states who lack the reasonable means to vote. Or consider whistle-blowers in all sorts of organizations, including states, corporations, and non-profits. Whistle-blowers take on huge personal cost to reveal wrongdoing within their organizations. Some have argued that there is an institutional duty to blow the whistle—a duty grounded in organizations’ practices of public accountability (Ceva and Bocchiola 2019). Any such duty must be proportionate to the cost involved. Likewise, those who omit to discharge the duty to whistle-blow are implicated in the resulting wrongs only to the extent to which they could have blown the whistle within reasonable cost. Often, this will not have been possible, simply because they could not even have come to know, within reasonable cost, about the organizational wrongdoing. Second, it’s important that many low-ranking members of organizations will often have no reasonable opportunity to take positive steps against wrongdoing. They might work in a section of the organization that is wholly shielded from the wrongdoing; they might lack the ability to signal dissent by leaving the organization (this is particularly clear in the case of states); they might lack knowledge of the wrongdoing; they might lack knowledge of how they can use their role to push against the wrongdoing. Thus, although implication-viaomission includes many more members than merely those who have implication-via-enactment and implication-via-endorsement, it does not open the floodgates to bringing all members on the hook. Implication-viaomission depends on what options a member has—within their role and within reasonable cost—to push against organizational wrongdoing.

6.5 Conclusion I started this chapter by noticing that we often want to let many members off the hook when an organization does wrong. That desire has been partly

156

  

vindicated by the arguments I have provided. For example, I started this chapter by suggesting that cashiers in banks—or even managers of cashiers—are not implicated in wrongs that are primarily operationalized in parts of the bank that deal with financial advice or home loans. And I suggested that present-day green-lifestyle citizens are not implicated in their state’s blameworthiness for historically distant injustices. Both of these groups remain off the hook by the arguments I have given for implication-viaenactment, implication-via-endorsement, and implication-via-omission. Let me close by saying something about these examples. Consider how cashiers relate to the organizational wrong of giving poor financial advice. Cashiers are not the principal enactors of wrong, since their job is not to give advice. Nor do cashiers enact any of the cousins of complicity: cashiers do not connive with the members who give bad advice, or condone those members, or consort with them, or stand contiguous with them. And cashiers are not complicit in wrongful financial advice, insofar as they do not contribute causally to the enactment of it, or perform contributory acts that ‘give access’ to the principal enactors (most people do not reach advisors by walking into a bank and talking to a cashier); nor do they facilitate the principal wrongdoing, make it possible, induce it, encourage it, or make it easier to perform. Because they do not instantiate principal enactment, or a complicity-cousin, or complicity itself, cashiers are not blameworthy as enactors. The same goes for green-lifestyle citizens of democratic states, when we are considering wrongs that were committed even before those citizens were born—such as the historical emissions at issue in the Polluter Pays Principle. Often, cahiers also do not satisfy the conditions for endorsing the wrong of bad advice. Endorsement is likely blocked at one or both of two places. First, in signing their employment contract with the bank, the cashier might have had little other choice; they lacked the reasonable possibility of avoiding exchanging commitments with the bank to the goal of (say) providing financial services and products.⁹ Second, even supposing a cashier did have the possibility to avoid exchanging commitments with the bank—that is, supposing the cashier took the job in a employment market where they could easily enough have taken a job at a different organization—still, it might not have been reasonably foreseeable to the cashier that the ⁹ In Section 4.4, I suggested that ‘providing financial services and products’ is the purpose of banks. But something’s purpose can also be (one of) its goals; indeed, one might hope that it would be.

   

157

organization would do wrong on its way to pursuing the goal of providing financial services and products. Who would believe that their colleagues would forge signatures? And, as for green-lifestyle citizens’ relationship to historical emissions, endorsement hasn’t occurred, because the relevant exchange of commitments has not occurred. Some citizens will have used their votes (or petitions, protests, or so on) to exchange commitments to climate-destroying goals—but not, by definition, those who have ‘green lifestyles’. Finally, what about omissions? Cashiers are not implicated as omitters in the wrong of bad financial advice, when and insofar as there is nothing cashiers could have done to actively push against the wrongdoing. Implication-via-omission implies that the member has failed to take active steps to avoid either implication-via-enactment or implication-via-endorsement. But if one could not have taken any such active steps, then one’s failure to take them is not explained by one’s absence of intention—and so, one didn’t really ‘omit’ to take the steps. The same goes for present-day green-lifestyle citizens’ relationship to the historical emissions for which their states bear blameworthiness. Enactment, endorsement, and omission together implicate many members in organizations’ wrongdoing. Yet, as several philosophers have noted, the implication of members is not always ‘enough’ to capture the full extent of organizational wrongdoing. When an organizational wrong is particularly egregious—or, particularly, when an organizational wrong occurred in the past—there might be few, or even no, present-day members who are implicated via enactment, endorsement, or omission (Pettit 2007; Copp 2007; Braham and van Hees 2011; Collins 2017a; Collins 2019b). The result is a dearth of members’ implication in organizations’ wrongdoing: a so-called ‘gap’ between the extent to which we want to hold an organization blameworthy and the extent to which we can hold members as implicated. For many philosophers, such gaps are part of what makes organizations’ wrongdoing appealing: if we can’t find enough (or any) individual members who are implicated, then it’s appealing that we can blame the organization itself. I argued in Parts I and II that organizations can, indeed, be blamed as entities distinct from (while composed of) their members. However, the imposition of reparative burdens upon an organization often entails costs for members. There is a big question about how to justify this distribution of costs, if we cannot find many implicated members. The final chapter turns to this question.

7 Apportioning Members’ Burdens 7.1 Introduction When organizations do wrong, the correct response is often the imposition of burdens. In the case of corporations, such burdens are sometimes imposed by the state, for example in the form of fines. But burdens can also be imposed on corporations by non-state actors—in the simplest case, by non-member individuals, via consumer boycotts. Likewise, when a state does wrong, a burden might be imposed by an international body such as the United Nations Security Council or North Atlantic Treaty Organisation (often in the form of sanctions, no-fly zones, and so on). But as with corporations, states’ burdens might also be imposed by ‘ordinary’ individuals, for example in the form of artistic or cultural boycotts. This chapter addresses the implications such burdens have for members. Almost all forms of organizational burden impose costs on members. For example, fines eat into corporations’ proceeds, which reduces value for shareholders and inhibits the organization’s ability to provide benefits for employees. Other burdens on wrongdoing corporations—such as forced restructuring—consume the time and energy of members, which could have been spent on activities that serve the benefit of the corporation and its members. Meanwhile, sanctions on states impose material costs on governments that reverberate through to national budgets. Burdens imposed upon states also create more intangible costs, such as a loss of cultural or national respect—which, again, percolate through to members. Organizations are distinct from—though composed of—individuals; just as forests are distinct from trees and solar systems are distinct from planets. How might it be justifiable to impose burdens—punishments, sanctions, reparations, and so on—on organizations, given the effects these burdens have on members? As I hope is clear, enactors, endorsers, and omitters are candidates for the imposition of burdens, given that their enactment, endorsement, and omission cannot be justified or excused. But many members are not enactors, endorsers or omitters—insofar as actions that might appear to be enactments, endorsements, or omissions are, in fact, justified or Organizations as Wrongdoers: From Ontology to Morality. Stephanie Collins, Oxford University Press. © Stephanie Collins 2023. DOI: 10.1093/oso/9780192870438.003.0007

  ’ 

159

excused by the member’s lack of reasonable alternatives. What’s more, some members might do everything they can to distance themselves from organizational wrongs. Consider, for example, a citizen of a liberal-democratic state who uses his role to protest, petition, and vote against a wrong perpetrated by his state. Are these unimplicated members liable to burdens? This question is brought into sharp focus by historically distinct wrongs, such as the historical emissions at issue in the Polluter Pay Principle. Present-day states bear blameworthiness for historical emissions, insofar as the present-day state is the same entity as the historically emitting state.¹ But what about present-day members? It’s possible that no presentday members are implicated in historical wrongs. Yet present-day members must bear the cost of any present-day burdens. Thus, members’ implication, of the kind analysed in Chapter 6, will not always be enough to justify the imposition of burdens on members. Some authors have argued that member burdens often cannot be justified, and that in such cases we must not sanction organizations when they are blameworthy for wrong (Hasnas 2009; for an analysis of what’s wrong with punishing citizens by punishing states, see Erskine (2010)). This chapter aims to do better. I develop a framework for apportioning organizations’ burdens amongst members. This framework produces high cost for some members, no cost for others, and a middling cost for many. I aim to situate organizations’ wrongdoing-based burdens amongst the other kinds of burden we might impose on organizations and their members.² The argument begins in Section 7.2 by outlining a view on which moral agents—whether individual or collective—are liable to cost on the basis of numerous and varied ‘sources’. These sources include wrongdoing—but they also include causing harm, having the capacity to bring about some good, benefitting from harm, being in an associative relationship, and so on. Different sources have different levels of importance in different contexts,

¹ There is an interesting question of when a present-day state should be treated as the ‘same entity’ as a historically distant state. The rough answer, I believe, is: when there is continuity of the state’s structure and decision-making procedures, such that the present-day state’s structure and decision-making procedures arose primarily out of the internal exercise of the structure and decision-making procedures of the earlier state, rather than through the external imposition of revolutions, conquests, and the like. This issue deserves further exploration. In any case, I assume large historical emitters—such as states in Western Europe and North America—are properly continuous with their historically emitting incarnations, at least over the last several decades. ² This chapter presents an edited version of the argument in Collins 2016.

160

  

and often several sources combine to generate a justification for imposing a certain level of cost upon an agent. Section 7.3 uses this framework to answer the question of how costs should be apportioned within small-scale collective agents, in which it’s relatively straightforward to implement a differentiated apportioning of costs amongst members.³ I advocate the following: the apportioning of costs amongst members should, insofar as possible, track the moral source(s) of the cost. As a first step, the cost to which the collective is liable should be ‘earmarked’ to the different cost sources, in proportion to those sources’ importance in generating the collective’s liability to cost. As a second step, the costs that have been earmarked to each cost-source should be divided amongst members in proportion to members’ instantiation of that source of cost. I call this the ‘source-tracking’ model for apportioning costs amongst members. I argue that it applies as much to organizations as to small-scale collective agents. Section 7.4 attends to the ‘shortfall’ problem, which is rife in organizations specifically. The shortfall problem arises when the organization’s liability to cost derives from source(s) that aren’t instantiated by any member. This is most acute for costs that are sourced in historical wrongdoing. If an organization did wrong decades ago, it’ll often be the case that no current member was implicated in that wrong, in the ways outlined in Chapter 6. Here, the organization is blameworthy for wrong, but no member is blameworthy for being implicated in that wrong. So, an important source of the organization’s liability to cost—perhaps the only source of that liability— isn’t instantiated by any member. In these situations, the organization’s costs cannot be distributed in a source-tracking way. The result is a shortfall: the organization is liable to cost as a result of its wrongdoing, but these costs cannot justifiably be distributed to any members in a source-tracking way (because no members were implicated in the organization’s wrongdoing). Shortfalls have led some authors to reject source-tracking distributions of states’ wrongdoing-based liabilities to cost (Stilz 2011, 193–4; Pasternak 2013, 365).⁴

³ I assume small-scale collective agents can have obligations, do wrong, and justly be lumped with reparative burdens. For an account of their moral agency, see Collins 2019a, ch. 6. ⁴ Recently, Pasternak (2021) has rejected source-tracking for a different reason: that a source-tracking apportioning of wrongdoing-based costs is often impractical, for three reasons: it is unlikely that enough individuals are blameworthy enough to fully compensate the harm; blame has negative social and political repercussions; and implementing a blame-tracking

  ’ 

161

I address the shortfall problem by asking when and why shortfalls matter. I suggest that shortfalls matter only when cost sources other than wrongdoing are operating alongside wrongdoing, to generate moral reasons for the organization to bear cost. To illustrate: suppose a state is liable to cost on the basis of historical wrongdoing, but no present-day members were implicated in that historical wrongdoing as enactors, endorsers, or omitters. The state’s liability to cost cannot, under the source-tracking model, be distributed to members. This shortfall will bother us—I claim—when and because the state’s members nonetheless benefit from the historical wrongdoing, or have an associative relationship with the victims, or have a strong capacity to contribute to providing victims with a specific relational good. These are sources of liability to cost in themselves, which are in operation alongside wrongdoing. My proposal is that some combination of other cost-sources— perhaps benefiting, association, or capacity, amongst others—can be used to ‘top up’ the shortfall, under certain conditions that will be specified. Sourcetracking is thus vindicated as a viable method for distributing the costs of organizations’ wrongs amongst members.

7.2 Sources of Cost There are several ways of connecting a moral agent, A, to a moral person, P, such that A is singled out as being liable to bear cost as a result of A’s relation to P. Some of these are: wrongdoing, where A is blameworthy for P’s situation; causal responsibility, where A contributed to P’s situation in a mere causal way; capacity, where A is especially capable of remedying P’s situation; benefitting, where A benefits from P being in the situation; and community, where A stands in an associative relationship with P (Miller 2007, 100–4). We can view each of these connections as a potential source of A’s liability to cost. Each source may be operative in many different contexts. But only in some contexts will one or more source be weighty enough to trigger a justification for others to impose costs on A—that is, only in some cases will one or more source be significant enough to generate A’s liability to cost. Two observations are important for my purposes. First, different sources are weighty in different contexts. To adapt one of David Miller’s examples:

distribution detracts resources from the state’s other obligations (35–40). All of these issues are dealt with by Section 7.4’s proposal. For discussion of Pasternak’s recent work on these issues, see Collins forthcoming.

162

  

suppose one of a team of mountaineers becomes injured and must be helped down the mountain. If the injury is serious, perhaps the strongest team member should go with him, in case he needs carrying. In that case, capacity is the most important source of liability to cost. If the injury is not serious, perhaps the team member most closely associated with him—his closest friend in the team, say—is the most appropriate bearer of cost. Here, association is weightier than capacity. If someone caused his injury, then the injury-causer might be the most appropriate person to bear cost, as recompense. And so on. Second, multiple sources might apply in one context. If one teammate caused the mountaineer’s injury, and she’s his sister, and she’s very strong, then three sources are at play: causation, association, and capacity. All three sources ‘add up’ to determine—or overdetermine—the fact that she bears liability to cost. The weightiness of each source depends on, inter alia, the extent to which that source is present. For example, if the sister caused the injury, but only in combination with unlikely environmental factors (slippery rocks, say), then causation isn’t so weighty as a source of liability. If she’s his sister, but they have been estranged for many years, then association might have less weight than it otherwise would in determining her liability. And if she’s the most capable, but not by much, then capacity is less weighty than if she were the most capable by far. Even if each source is not very weighty on its own, because each is not present to a very great extent, it’s still true that their combined weight may be large enough to generate a liability for her. It will sometimes—though perhaps only sometimes—be possible to roughly rank the sources in terms of the weight they lend to the production of a liability. This picture is neutral with regard to normative ethics: it doesn’t take a stand on any particular substantive moral theory. Perhaps most naturally, liability sources can be viewed as generating Rossian ‘prima facie’ duties, each of which is morally fundamental and all of which must be relied upon in reaching all-moral-things-considered decisions (Ross 2002 [1930]). But they can also be viewed as rules of rule consequentialism (Hooker 2000), or public rules or private heuristics of other forms of indirect consequentialism (Railton 1984), or considerations that might drive the virtuous agent to action (Hursthouse 1999). Insofar as there are substantive moral claims evoked in this picture, they are substantive moral claims on which there is a wide philosophical consensus. Moreover, I’m somewhat neutral on exactly which sources can produce liability to cost, and on which sources tend to be more weighty than which others. In what follows, I will assume that

  ’ 

163

wrongdoing produces liability to cost. And I will use examples to suggest that capacity, association, and benefit are likewise liability sources.⁵ I remain neutral on the relative importance of these sources and on whether some of them can be ‘reduced’ to each other.

7.3 The Source-Tracking Model With this sketch of liability sources in hand, we can turn to examples where groups are liable to cost and must distribute those costs amongst members. To settle ideas, let’s focus first on a small group: five teenage friends who earn pocket money on weekends, doing odd jobs together for people in the neighbourhood. I will assume the group is a collective agent. The friends have a group-level decision-making procedure, under which group decisions are made via deliberation-based consensus with majoritarian voting if no consensus is reached. This procedure is used to decide which services the group advertises, which jobs they accept, who does which parts of which job, and so on. Of course, this team is not an organization: it doesn’t have the size, hierarchies, and divisions of labour that make organizations particularly complex and fraught bearers of liability to cost. But, like organizations, the team of teenagers is a collective agent that can do wrong and can be liable to cost—or so I shall assume. Its relative simplicity makes it a useful group for illustrating the source-tracking model. Suppose the group is shifting a heavy bookcase. Tim is chatting negligently and not watching his footing. He stumbles in the stairwell and loses his grip. The bookcase slides down the stairs and breaks the foot of a passing child. Plausibly, the team becomes liable to bear cost to remedy this. The question arises: how should the team’s costs be apportioned amongst the members? In the short-term, when the bookcase must be heaved off the child, the answer is easy. A child’s foot is at stake, so each member should simply lift as quickly and forcefully as he or she is able: a capacityrelative distribution is best in the immediate aftermath. But what about the longer-term burdens? How should the child’s medical bill be paid? Who should bear the brunt of the parents’ reprimands? A straightforward answer is that ‘the team as such’ should bear these burdens. This answer is vindicated by Part II’s argument that organizations

⁵ I have argued for the importance of capacity-based duties in Collins 2015.

164

  

(and, let’s assume here, collective agents more generally) can be blameworthy. Team-level burdens might work for reprimands (though even then, some member will bear the burden of responding appropriately). But it won’t work for medical bills. Perhaps there is property that is owned by the team as such—mowers, ladders, and so on—which could be sold to pay the medical bills. However, the team will then have to save for replacement mowers and ladders, or bear the costs of not owning them (such as having to reject certain jobs). Either way, the point is that costs will be imposed on members. The question remains of how these costs should be distributed amongst the members. We might answer that the team may distribute these costs however it likes, as determined by its decision-making procedure. It’s an agent, after all, and it will have preferences over how the costs should be distributed. Yet it is counterintuitive that the team has full reign here. Suppose the decision goes to a vote, and the team decides four-to-one that the poorest member, Gareth, should always pay all medical bills arising from the team’s operations. We would assert there’s something not right about this; that the team is overlooking important moral reasons; that the team’s distribution of burdens should be different. Perhaps, then, we should advise a capacity-relative—or maybe an equal— distribution of both the long-term and short-term burdens. In support of capacity-relative or equal distributions, we could point out that the acquisition of liability to cost is a risk of acting in the world, including acting as a material part of a collective agent. We might recommend that members pool this risk, as insurance against individually costly liabilities. This would result in equal or capacity-relative distribution of the costs of ‘insurance claims’ made by those in the risk pool. Such risk pooling makes sense when agents cannot control the risks. But the acquisition of many liabilities—in particular, many wrongdoing-sourced liabilities—is the result of choice. For these liabilities, it looks less sensible to pool risks. To see this, we might ask: why would diligent Dawn, who is evercareful about her footing in stairwells, want to pool her risk with negligent Tim? As James Crawford and Jeremy Watkins put it (from the perspective of international law), using a capacity-relative or equal distribution for wrongdoing-sourced liabilities would entail distributing costs to members who are, by any standard, morally blameless. Indeed they may be the majority. So far as they are concerned their treatment . . . may seem as unfair and ethically backwards as the treatment meted out under primitive

  ’ 

165

systems of collective responsibility in which whole tribes or nations are subject to reprisals: after all, in both set-ups innocent people are called upon to pay the price for the misdeeds and mistakes of their rulers [or fellow members]. (Crawford and Watkins 2010, 290)

In other words, equal or capacity-relative distributions of cost are insufficiently sensitive to members’ abilities to control their ‘misdeeds and mistakes’ that generated the team’s liability to cost in the first place. Additionally, to distribute the costs of the team’s wrongdoing-based liability in a desert-insensitive way allows individuals to ‘distance themselves morally and psychologically from the wrong in which they are implicated’ (Crawford and Watkins 2010, 291). In my terms: the enactors, endorsers, and omitters have less practical reason to view the wrongdoing as ‘theirs’. This distancing creates a moral hazard: why bother watching your footing in the stairwell, if you don’t take ownership of the results and you know that you personally will be distanced from any nasty consequences, insofar as your mates will bear some of those nasty consequences? One might defend equal or capacity-relative distributions differently, by pointing out that the members are ‘beneficiaries of a common practice in which participants are treated fairly . . . and so they should also be prepared to carry their share of the costs . . .’ (Miller 2004, 25). In the case of states, Avia Pasternak (2011, 199–201) has similarly argued that ‘living in a selfruling political community’ provides net benefits, such that citizens of democracies should accept equal distributions of their state’s liabilities— including even if those liabilities are sourced in the state’s wrongdoing in which present-day citizens are not implicated (in Chapter 6’s sense of ‘implicated’). However, we are trying to work out what it is for members to be treated fairly in a common practice—or at least, what it is for members to be treated fairly in that aspect of the practice that involves apportioning costs amongst members. The practices of the team are not an all-or-nothing affair, such that members must accept the entire package if that package provides them with net benefit. Instead, we should consider whether the benefits of working together (for example, receiving pocket money) can be pragmatically disentangled from the costs (for example, paying for your friend’s negligent stumbling). In the case of the teenagers, these costs can be separated: if the team distributes the costs non-equally or not relative to capacity, it will not bring crashing down the entire package of value contained within the shared practice. If the benefits and costs can be disentangled, then there is no

166

  

justification for appealing to net benefits in order to block members from asking for the benefits but not the costs. A third justification for equal or capacity-relative distributions appeals to group solidarity. Here, one might draw on the work of Larry May.⁶ According to May, group action centres around the presence of solidarity. May follows Jean-Paul Sartre in characterizing solidarity by the presence of a ‘sufficiently strong common interest’, alongside which members come to identify with the group rather than with ‘isolated individual identities’ (1987, 35). I am sceptical that solidarity plays a central role in organizations, considering how many members might be alienated, begrudging, or just ‘in it for the paycheck’. As I mentioned in Section 3.4, many members of corporations do not identify with the group, nor take themselves to have common interests in common with other members. This is increasingly the case in liberal democratic states, too. Moreover, I am sceptical that the mere psychological fact of solidarity would warrant an equal distribution of burdens, even supposing such a fact held true. Some people identify with their country’s rugby team, but this psychological fact hardly warrants imposing costs on the fans when we impose costs on the team (say, if the team is embroiled in a culture of sexual assault). Pasternak likewise employs solidarity, using a slightly different conception than May. According to Pasternak, there are ‘two components of solidary action: first, people who act in solidarity are acting jointly to achieve a worthy social goal [for example, doing jobs for people in the neighbourhood] and, second, they view themselves as equal participants in their joint project’ (Pasternak 2011, 200). Pasternak argues that ‘if . . . solidary action is an intrinsically valuable social good that characterizes certain relationships, then in those relationships the equal sharing of the consequential responsibility for the group’s joint actions is in fact an associative obligation of its participants. Equal sharing is an AO [associative obligation] because it is constitutive of the social good of solidary action’ (Pasternak 2011, 197). Pasternak illustrates this with a group pushing a car out of the mud. When they succeed, they each (rightly, she implies) see themselves as deserving an equal amount of the driver’s thanks, because they were acting in solidarity (Pasternak 2011, 196–7). Pasternak explicitly extends this reasoning to citizens of democratic states, though of course it would have a more tenuous application in corporations. ⁶ Although it is worth noting that May himself does not endorse an equal distribution of costs, at least not within for-profit corporations. See May 1987, 104.

  ’ 

167

I’m not sure solidarity can justify equal or capacity-relative apportionings of cost. Suppose the least-helpful member of Pasternak’s car-pushing team tells the owner that the others did more work than her, and therefore deserve more thanks. In other words, suppose she resists an equal distribution. Would we think that she had failed to discharge an associative obligation to share praise equally? We would not. We might think that she had no obligation not to take an equal share of praise. But Pasternak’s claim is not that solidary actors are permitted to share praise equally. It’s that they ought to do so. Of course, one might think that praise operates differently than blame. If so, suppose the car-pushers shove the car too far, and run over the owner’s cat. If one member was standing in front of the car to check the way was clear, we might think it positively immoral for him to absorb an equal share of the owner’s rage, rather than putting himself forward for the most censure. The point is this: an imperative to apportion costs equally amongst members does not seem to arise automatically out of Pasternak’s solidary action, unless by definitional fiat. This is even before we get to the problem that most organizations aren’t characterized by solidary action. The intuitions that drive these objections suggest the following about our original bookcase example. If the teenagers’ liability to cost is purely wrongdoing-based, and if Tim is the only one who wrongly stumbled, then Tim should pay the entire bill and absorb all of the parents’ reprimands. (The fact that capacity loomed large in the immediate aftermath might suggest that the team’s liability isn’t entirely wrongdoing-based. Later in this section, I discuss a multi-source interpretation of this example. Section 7.4 explains what should happen if Tim cannot pay.) Suppose Tim stumbled, while Dawn took great care to avoid stumbling. Perhaps Dawn did warm-up stretches, wore oppressively sturdy shoes, craned her neck uncomfortably to watch her footing, and so on. Dawn took on costs specifically to avoid stumbling, and succeeded in doing so. She was not implicated in the team’s wrong in any of the ways canvassed in Chapter 6: Dawn was not an enactor (she was not a principal enactor like Tim, nor was she complicit with Tim’s enactment), nor was Dawn an endorser (since Tim’s stumbling wasn’t reasonably foreseeable when Dawn exchanged commitments to the team’s odd-jobs end), nor was Dawn an omitter (since she took active steps to perform her role in ways that pushed against being an enactor or endorser of the wrong). It’s not clear how the group might justify a wrongdoing-insensitive distribution to Dawn. If Dawn was also negligent, then she should bear some costs. In that case, she’d be an enactor alongside Tim. And if Tim’s negligence was reasonably foreseeable,

168

  

then Dawn would be implicated as an endorser. And if Dawn omitted to use her role to (for example) promote norms of careful footing within the group, then Dawn would be implicated as an omitter. But if Dawn was a model removalist—and if the group’s liability to cost was entirely wrongdoingbased—then she should not pay. To extend these judgements to non-wrongdoing sources of liability, suppose the team happens to pass a child trapped under a fallen bookcase. The team thereby acquires a capacity-based liability to take on the cost of lifting the bookcase. Since the team’s liability is capacity-based, a capacityrelative distribution is plausible for both immediate costs and any final costs that might arise. Or imagine the team is moving furniture for Tim and Gareth’s grandparents, whom the other members also know well. The team has an association-based liability to take on cost, by doing the job for free. Although it is the team’s liability, it is fairest for Tim and Gareth to take on the hardest work. After all, they have the strongest association with the parents. Of course, not everyone will share these judgements about fairness. In forwarding them, I am making substantive moral claims. The aim of the rest of this chapter is to see how far these judgements get us when apportioning costs amongst organizations’ members. Specifically, I will be concerned to get a method of cost-apportioning that gives us the right results for organizations’ wrongdoing. The general proposal is: when an organization is liable to costs, these costs should be distributed amongst members in proportion to the members’ respective instantiations of the source(s) of the organization’s liability. As per the previous section’s framework, those sources will be different in different cases and often a combination of sources will contribute to determine to total amount of cost to which an organization is liable. To explain the general proposal: the idea is that the apportioning of costs should proceed in two steps. First, the costs are earmarked amongst liability sources, in accordance with those sources’ weight. Second, the costs earmarked to each source are divided amongst members in proportion to members’ individual instantiations of that source. Cases with one source are easy. If an organization’s liability is based purely in its wrongdoing, then costs should be distributed to members in proportion to their implication in that wrongdoing (as implication was characterized in Chapter 6). Cases with more than one source are trickier: suppose an organization’s liability is generated by both wrongdoing and capacity, with wrongdoing being twice as weighty as capacity in the context. In this case, first, (roughly)

  ’ 

169

67 per cent of costs should be earmarked for wrongdoing and (roughly) 33 per cent of costs should be earmarked for capacity. Second, the wrongdoing-earmarked costs should be divided amongst members in accordance with their varied levels of implication in the wrong. Likewise, the capacity-earmarked costs should be divided amongst members in accordance with their varied levels of capacity to take on cost. This is the ‘source-tracking’ model for distributing liabilities amongst members. Source-tracking is relatively straightforward for our removalist teenagers. Organizations are far more complicated. Consider states: these are organizations that sometimes last for centuries and are usually constituted by millions of people, who vary wildly in their capacities, associations, and receipt of benefits—not to mention in the extent to which they are implicated in their organizations’ wrongs. In states, members are often turned into adversaries by the very decision-making procedure that unites them, some members have much more influence over the organization’s decisions than others, the reasons behind the organization’s decisions change with the winds of political favour, and many decision-making methods are casuistic or not explicit (Collins and Lawford-Smith 2021). These features make source-tracking difficult. Yet, just as for the teenagers, I believe source-tracking captures commonsense judgements about how to apportion costs amongst members in states. Suppose the US is morally liable to take on cost for its failure to seriously reduce emissions, especially since 1980 when the science of climate change has become increasingly clear. Members of the US who have campaigned for climate action could rightly ask why they should have to contribute to the costs of this liability. If the US’s climate liability really was grounded wholly in wrongdoing, then the source-tracking model would uphold these citizens’ complaints and redirect the costs to those members who were implicated (including by omission) in the US’s status as a polluter. But if the US’s climate liability was instead grounded in (say) its capacity to bear the costs of emission reduction, then the members’ question would have a straightforward answer: you should contribute to this liability because you have the capacity to do so. Either way, the members receive an answer that appeals directly to the reality of their lives and their connection to the source of their organization’s liability. This is very appealing. That said, one initial complication concerns the obvious infeasibility of distributing organizations’ liabilities in perfect proportion to each member’s instantiation of the relevant sources. This problem can never fully be addressed, at least for very large organizations such as states. Organizational

170

  

policy is a blunt instrument. But there are feasible ways of approximating the source-tracking ideal. In states, one method is targeted taxes. This arguably already occurs for states’ causation-based liabilities around climate change: some states pass this cost onto consumers of petrol, through higher purchasing taxes on that good. The source-tracking view would justify this, stating that individual causers of climate change should bear the costs of the state’s fulfilment of the causation-based ‘part’ of its liability to bear cost to respond to climate change. More generally for states’ causation-based liabilities, if an activity causes the harm, then that activity can often be taxed. One might think this unpalatably allows past causers to avoid costs, by ceasing to consume the relevant good. But such cessation is itself a cost, justifiably imposed upon past harm-causers in virtue of their causation. Similarly, the capacity-designated percentage of a state’s liability might be paid out of general taxation. The more costs there are that are both earmarked for capacity and paid out of general taxation, the more reason there is for the state to have both higher and more progressive taxation on wealth. Such capacity-based liabilities are sometimes tricky: if Australia has a purely capacity-based liability to provide medical care to climate refugees, should this cost be lumped entirely on doctors? If source-tracking implied so, then it would be unpalatable. Intuitively, taxpayers more generally should bear the costs of paying these doctors well, and so bear costs for the exercise of doctors’ capacities. But this just demonstrates the importance of being clear about which capacity generates the liability: it is Australia’s capacity to adequately train and remunerate doctors that generates its liability, rather than the capacity of particular doctors to provide medical care. Finally, suppose a state’s liability to bear costs in reducing greenhouse gas emissions is based equally on causation, capacity, and benefitting. We can then imagine a third of the relevant taxes targeting users of petrol, a third of them targeting the wealthy, and a third of them targeting oil or transport companies (or whoever benefits from emissions). Inevitably, such policies will generate some false negatives and false positives. But this need not stop us getting as close to a source-tracking distribution as is feasible. Naturally, things will look different in organizations that are not states, not least because such organizations don’t have broad discretion to impose financial liabilities on their members. For non-state organizations, a targeted distribution of costs will often mean impositions on members’ time, energy, and cognitive or physical resources during working hours: those who instantiate

  ’ 

171

the liability source should spend their (work-related) time, energy, and resources addressing the liability.

7.4 Responding to Shortfalls Shortfalls arise in cases where an organization is liable to cost, such that the payment of that cost requires imposing costs upon members, yet it is unreasonable (by the lights of the source-tracking model) to impose (all of) these costs on members. Organizations’ historical wrongs are the starkest—and probably most common—instance of shortfalls. More specifically, states’ liabilities that are sourced in historical wrongdoing serve as a compelling case study. My discussion will focus on these. Suppose it’s the year 2100. Climate change is running wild, but the global economy has marched forward with ever-increasing inequality. It’s been around 120 years since states first acquired blameworthiness for greenhouse gas emissions under the volitionist lens, though blameworthiness arguably extends back even further under the attributivist and aretaic lenses. Plausibly, most affluent citizens in 2100 are not implicated in the historical emissions for which their states bear blameworthiness (by the lights of all three of Chapter 4’s lenses). Yet some people face severe hardship as a result of those historical emissions. Suppose some state—let’s focus on Australia— holds a wrongdoing-based liability to bear cost to compensate those who have been harmed by Australia’s wrongful failure to act on climate change. Yet many Australian citizens in 2100 were not implicated in that historical wrong (because it occurred before those citizens were born). The sourcetracking model is silent on how Australia should pass the cost of its historical wrongdoing-based liability to its members. It seems Australia bears a liability, but no (or not enough) members of Australia should bear costs through its discharge. The liability is ‘shortfallen’. Shortfalls have led some authors to reject source-tracking for wrongdoing-based liabilities. As Anna Stilz argues, ‘[s]queamishness about implicating individuals . . . seems rarefied when we turn our attention to the uncompensated wrongs perpetrated by states in their citizens’ names’ (Stilz 2011, 196–7). Therefore, she suggests, she should use a model other than source-tracking. Likewise, Pasternak entertains the idea of source-tracking the costs of states’ obligations to repair harms. But, she says, this has ‘one obvious drawback: [ . . . it] is likely to create “liability shortfalls”, whenever it would be hard to identify those individuals within the state who are more

172

  

responsible for its misconduct, or to extract the necessary resources from them’ (Pasternak 2013, 365; similarly Pasternak 2021, 13–40). These conclusions are too quick. I suggest we start addressing the shortfall problem by asking why shortfalls matter. For example, why does it matter that Australia-in-2100 bears high cost for compensating the victims of climate change? Australia’s past wrongdoing (its failure to seriously reduce emissions) obviously has something to do with it. But perhaps it is not the end of the story. Consider Leif Wenar’s examples: We do not hear calls for reparations from the Germans for the Allied bombings of Dresden, or from the United States for the Japanese surprise attack on Pearl Harbor, or for that matter from the United States for the British arson of the White House in 1814. These are events which many at least believe were historical injustices, but we do not hear demands for reparations from these believers because there is little sense that those who could potentially claim reparations under backward-looking principles suffer from an unjust distribution of rights or resources at present. (Wenar 2006, 402, emphasis original)

On this basis, Wenar surmises that ‘valid claims for reparation rest at the deepest level . . . on reasons we have to improve our current relations so that we can get along better in the future’ (2006, 396). Several other philosophers likewise use sources other than wrongdoing to ground reparations for historical injustice (e.g., Kumar and Tan 2006; Miller 2007; Thompson 2002; Vernon 2003). These authors’ arguments generally do not appeal to liability shortfalls (although Kukathas (2006, 340–1) alludes to them). Instead, such authors appeal to problems with assessing counterfactuals (such as ‘what would have happened if Australia had seriously reduced emissions?’), problems with identifying victims’ descendants, or problems with retributive justice generally. (Of course, some authors working on historical injustice do believe that wrongdoing is the proper basis for liability to reparative costs (e.g., Sher 2005; Simmons 1995)—these authors will face the shortfall problem.) In contrast to these authors’ arguments, my suggestion, in the spirit of Wenar’s, is that organizations’ liabilities are often not entirely what they seem. This is not as strong as Wenar’s suggestion that non-wrongdoing sources are the only sources of states’ reparative liabilities related to historical injustices. I accept that historical wrongdoing might be a—or the only— source of some liabilities of some organizations, included states. This poses

  ’ 

173

no problems if enough present-day members are implicated in those wrongdoings. But in cases where the liabilities are shortfallen, any demand that members bear costs to make up the shortfall must derive from sources that are not shortfallen. In other words, in order to justify imposing shortfallen costs on organizations’ members, we must assert that non-shortfallen sources are adding some weight to the organization’s liability. We can then justifiably require members to make up the shortfall, on the basis of their individual instantiations of these non-shortfallen sources. The most prominent nonshortfallen sources in this connection are: capacity to repair the wrong, association with the wronged party, and benefits received from the wrong. To use the example of Australia in 2100, the idea would be that Australian citizens who are not implicated in Australia’s failures to act on climate change are nonetheless liable to cost if they have the capacity to bear cost, if they stand in associative relations to the victims of climate change, or if they benefitted from Australia’s historical failure. There are two main conditions on using non-wrongdoing sources to make up shortfalls in wrongdoing-based liabilities. First, the non-shortfallen sources must have genuine weight in generating the organization’s duty. Such sources are likely to do this when and because there is damage to the current (or projected future) relationship between the organization and the would-be recipients of costs. An apt description of such ‘damaged relationships’ is given by Margaret Urban Walker. She describes damaged relationships as those in which participants lack confidence that they share the same standards for reciprocal treatment, or lack trust in one another to abide by those standards, or lack reasons to be hopeful that inadequate relations will be remedied if they come about (Walker 2006a, 384; for in-depth discussion, 2006b). Of course, in some contexts the fact that a relationship is damaged doesn’t generate any liabilities at all. Damaged relationships are most troubling— and most likely to generate liability to cost—when parties have frequent or avoidable interaction, or when confidence, trust, and hope between them are particularly low. In such cases, liability sources such as association, capacity, and benefit will often—perhaps in combination with one another—generate liabilities for an organization and its members to bear cost in the service of repairing the relationship, independently of wrongdoing. Like Wenar, I believe this holds in cases of states’ historical wrongdoing where it is intuitively appropriate that the present-day state bears cost, even if no present-day members are implicated in the wrongdoing. In

174

  

such cases, the state has a damaged relationship with the would-be cost recipients. And sources other than wrongdoing are instantiated by the state: there is an association between the state and its partner in the damaged relationship (such as a post-colonial association), or the state benefits from being in the damaged relationship (say, because it has power over the other party), or the state simply has the capacity to repair the damaged relationship.⁷ An organization’s liability to bear cost to repair its damaged relationship can come from a range of, or combination of, such sources—even if we ignore the cost-generating power of wrongdoing, because not enough members are implicated in the state’s historical wrongdoing. Agreeing with this is not hard: simply imagine a relation between an organization and a population, in which general confidence, trust, and hope are very low, due to exogenous factors. Would we say the organization is liable to reparative (literally, repairing) costs? I think would. If so, then something other than wrongdoing must be generating them. These non-wrongdoing sources can be equally present in cases where the organization has, additionally, done wrong, even if members are not implicated in that wrong. All this suggests the following necessary condition on topping up organizations’ shortfallen wrongdoing-based liabilities (this will be the first condition of three): (i) the organization has a liability to cost, where that liability is grounded in non-wrongdoing sources. I’ve suggested that the liability referred to in (i) will be a liability to repair the damaged relationship between the organization and the would-be costrecipients. The liability to bear cost in repairing this damaged relationship might be sourced in association, benefit, capacity, or some combination of these. I haven’t spent time analysing how organizations can stand in the latter relations to cost-recipients, because this book is focused on ⁷ One might wonder: does it really make sense to think of victims as having a relationship with a wrongdoing state, where that relationship can be damaged through lack of confidence, trust, and hope? Walker’s framework falls within the broadly care ethical tradition, which one might think is not readily applicable to organizations. (I thank an anonymous reviewer for raising this.) Collins (2015) argues that collective agents can indeed be givers of care. Collins and Ferracioli (forthcoming) argue that organizational caregivers are unreliable if they operate for a profit, but that non-profit caregiving organizations (including states and charities) can be valuable participants in meaningful caring relationships. I assume these relationships can be damaged when one loses confidence, trust, and hope in the organization.

  ’ 

175

organizations’ wrongdoing, rather than on organizations’ associations, benefits, or capacities. Still, I assume a story can be told for how organizations instantiate each of these sources of liability, as I told the story for wrongdoing in Part II of this book.⁸ There are two more necessary conditions for topping up shortfalls: (ii) enough of the organization’s members have a damaged relationship with the would-be cost-recipients, where the member-level damaged relationship is causally or constitutively connected to the organization-level damaged relationship. (iii) enough of the members in (ii) instantiate the sources that ground the organization’s liability to cost (from (i)), where the members’ instantiation of these sources is causally or constitutively connected to the organization’s instantiation of those sources. The relevant causal connection might go in either direction: from memberlevel to organization-level, or vice versa. The causal or constitutive connection between member-level facts and organization-level facts mirrors the removalist teenagers’ individual properties, which caused or constituted their team’s instantiation of liability sources. As an example to illustrate conditions (ii) and (iii), take Australia’s liability in 2100 to compensate victims of climate change. (Assume Australia’s mining industry is still strong enough that Australia’s GDP can sustain such burdens.) Relationships between Australian citizens and victims of climate change (including Australian victims) can be predicted to be one of low confidence, trust, and hope. If so, the example would satisfy the first part of (ii): enough of the organization’s members have a damaged relationship with the would-be cost-recipients. One might object that many Australians will know very few of the world’s climate change victims. How can their relationship be damaged, if there is no relationship to speak of? In reply: a relationship is damaged if it lacks confidence, trust, and hope. This doesn’t require interactions. It can be assessed counterfactually, by asking whether a lack of confidence, trust, and hope would be exhibited if the parties were to interact. The second part of condition (ii) can also be seen in this example: the member-level damaged relationship is causally or constitutively connected ⁸ For an argument that collective agents can bear several different sources of liability to cost, see Collins and Lawford-Smith 2016.

176

  

to the organization-level damaged relationship. There is plausibly a causal or constitutive link between the damaged relationship at the individual level and at the state level: the fact that Australian citizens don’t receive confidence, trust, or hope from climate change victims will likely reinforce, and be reinforced by, the fact that Australia itself doesn’t receive such confidence, trust, or hope. What about condition (iii)? The condition requires that enough Australians will instantiate the same liability sources as the Australian state, with regard to their liability to costs that repair the damaged relationship between Australia and victims of climate change. Australians are associated with victims of climate change through economic relations, Australians benefit from the injustice that caused the damage relationship (via the economic boost that coal and gas gave to the Australian economy, which we might assume will reverberate for decades), and we can suppose Australians will retain the capacity (money; political and social power) to repair the relationship. And these individual-level sources plausibly perpetuate, and/or are perpetuated by, the Australian state’s instantiation of the same sources. Indeed, it’s not clear what it would mean for a democratic state to be economically associated, or receive a benefit, or have some capacity, without at least some members also being economically associated, or receiving redounded benefits, or having the capacity to do their part in the exercise of the state’s capacity. The members’ liability sources partly constitute the state’s liability sources, even though latter are not reducible to the former. It is therefore legitimate for the Australian state—even in 2100— to impose costs on Australians in the discharge of its liability to repair its relationship with victims of climate change, even if none of those Australians were implicated in the historical wrongdoing that damaged that relationship. Let me summarize. Ideally, we should implement the source-tracking model, outlined in Section 7.3. However, if some source(s) of an organization’s liability are instantiated by so few members that we cannot impose upon those members the entirety of the costs earmarked for the shortfallen source, then we should impose what we can upon these few members on the basis of that shortfallen source. We should then consider whether conditions (i), (ii), and (iii) are satisfied. If they are, then we should normalize the weightings of the sources in (i)—that is, make them sum to 100 per cent, retaining their relative weightings—and distribute the rest of costs according to these new normalized weightings.

  ’ 

177

I’ll give a simplified example to illustrate how this works. To demonstrate the breadth of the account, the example will involve a corporation. Take Westpac Bank, one of Australia’s largest banks. In September 2020, Westpac agreed to pay a fine of AUD 1.3 billion (USD 0.9 billion) for breaching money laundering laws. Suppose—simply for the sake of argument—that Westpac’s moral liability to this cost is based 50 per cent in wrongdoing, 30 per cent in capacity to repair relations (for example, the damaged relationship between Westpac and its customers), and 20 per cent in benefitting from wrongdoing. Suppose (falsely) that no member enacted wrongdoing as a principal or via complicity, that no member exchanged commitments with Westpac to ends in pursuit of which Westpac could reasonably have been expected to commit this wrong, and that no member omitted to use their role to push against complicity or endorsement. We then ignore wrongdoing when distributing costs to members, and normalize the remaining sources: the liability is now 60 per cent capacity-based and 40 per cent benefit-based. Condition (i) thus holds. We then implement the source-tracking proposal for this new ratio of sources, if conditions (ii) and (iii) are satisfied, and if association and benefit have enough aggregate absolute weight to generate a liability to pay the AUD 1.3 billion independently of wrongdoing. The final ‘if ’ clause prevents over-inflating the newly normalized liability sources, and will be met only when the original liability is overdetermined. We should normalize and re-earmark the full costs only if the nonshortfallen liability sources are strong enough, in aggregate but independently of the shortfallen source, to generate a liability for the organization to pay the shortfallen cost. And even once normalized, the total cost that may be earmarked to any one source is subject to a ceiling, which is determined by the absolute weight of that source. The ‘cost ceiling’ of any given source might not be hit if there are other sources that both generate the organization’s liability and can be used to distribute costs to members. But once some sources are ignored (because they are shortfallen), then the costs that are earmarked to the remaining sources will come closer to— though never exceed—their respective ceilings. The point of this is that normalizing does not enable to us to attribute more liability to members than considerations of fairness amongst members would require them to bear. However, this means that any liability generated by the non-shortfallen sources might demand less than would be demanded if those sources were working in tandem with the shortfallen source. Perhaps capacity and benefit

178

  

together can demand only AUD 0.5 billion from Westpac, rather than the AUD 1.3 billion these sources would demand alongside wrongdoing. If so— and if there really are no Westpac members who are implicated in the wrongdoing—then only this less expensive liability should be paid. Crucially, though, if we balk as such a result in a particular case, then we should consider whether capacity, or benefit, or some other source, is in fact doing more work than we initially thought, and can indeed justify the more expensive imposition. The role of the ‘if ’ clause is to force us to consider this. A slightly different treatment of this example is required if there are some members who are implicated in Westpac’s wrongdoing, but it would be objectionably burdensome to lump these few members with 50 per cent of the total costs. Here, we should impose whatever cost is appropriate on these implicated wrongdoers, in accordance with their respective implication in the wrongdoing. Different implicated members might have different levels of wrongdoing, and so be liable to different level of cost on that basis. The source-tracking proposal is neutral about what costs are appropriate: perhaps, as at law (Cane 2002, 178), the principal enactors should sometimes be severally liable in full for repairing the wrongs they commit with others. If so, the few remaining wrongdoers might be liable for the full 50 per cent of Westpac’s costs. But suppose all the enactors, endorsers, and omitters are such that we should give them altogether only 10 per cent of Westpac’s costs: AUD 0.13 billion. Then, this 10 per cent should be imposed on them. The remaining AUD 1.17 billion should then be re-normalized for capacity and benefitting, and so earmarked 60 per cent to capacity and 40 per cent to benefitting. If there is a strong imperative to see the full AUD 1.3 billion paid, then it will be plausible to let these sources carry this weight. There are two main reasons for responding to shortfalls in this way. One is that it allows the non-shortfallen liability sources to exercise the strength they would anyway have exercised in the absence of the shortfallen source. It is thus properly responsive to those sources. I can see no reason to deny these sources the liability-generating strength they have independently of the shortfallen source, in a situation where that strength allows us to make up shortfalls that are, ex hypothesi, morally troubling. A second reason is that it is fair to the individual cost-bearers. It appeals to their attributes as individuals, and to liability sources that apply to them as individuals, as outlined in Chapter 6. In this way, we can respond to victims’ claims—the claims that ignite our concerns about shortfalls—while also being able to

  ’ 

179

justify the imposition of costs on individual members of the liable organization, by evoking sources they actually instantiate.⁹ Pragmatically, of course, in some contexts only a non-cardinal ordering of sources’ importance may be possible. And, more generally, it will rarely be plausible to produce precise percentages, by asserting that association is, say, twice as important as benefiting. These facts make source-tracking difficult to implement in a highly specific way. Ultimately, contextual debates will have to be had between those with differing views about sources’ weightings. Compromise between these differing views will inevitably have a role in determining final distributions. In some cases, though—such as the liabilities of Australian coal magnates or Westpac’s directors—the same costbearers and rights-bearers will be implicated, regardless of which precise combination of sources we use to top up the shortfall. One objection to this ‘topping up’ method is that the liability the organization ends up discharging is not really a wrongdoing-based liability. There remains a shortfall in some sense, since the organization’s wrongdoingbased liability never gets discharged qua wrongdoing-based liability. The organization discharges a different liability: a capacity-based liability, or a benefit-based liability, or some combination of these—but it never discharges the liability that derived from the shortfallen source. This is true. But in cases where shortfalls bother us (and not all will), the crucial concern is that the wrongdoing’s victims should receive recompense to the full extent feasible. This is what motivates the shortfall problem. And this concern is resolved, regardless of what liability source we use to resolve it. One might respond that, for victims of historical injustice, ‘it will not do simply to lump them in with the poor more generally, for their claim is of a very specific nature’ (Kukathas 2006, 332). Under the source-tracking proposal, if a ‘specific’ (i.e., source-indexed) claim cannot be resolved without imposing unfair costs on others, then that specific claim will indeed go unresolved. Crucially, though, there is still a specific claim that is acknowledged. This is a claim based on a very particular capacity, or particular ⁹ One might worry: what about individuals who are not members? These include the customers of a corporation or the non-citizen residents of a liberal democratic state. These individuals might bear cost as a result of organizations’ wrongdoing burdens, and one might think our distributions should be sensitive to this. My inclination is to factor in such external effects at the initial step of imposing liabilities-to-cost on the organization, rather than at the second step of distributing the cost of those liabilities amongst members. If we are worried about the negative effects on outsiders, then we should reduce the amount the organization has to pay. But this is not itself a problem of apportioning or distributing the organization’s costs. I thank Simon van Baal, Will Cailes, and Andy McKilliam for discussion.

180

  

benefits, or particular association, or so on. To explain why various sources—like capacity, association, or benefit—are liability generators for particular organizations and their members in particular contexts, we will have to make reference to historical, relational, and material details. Thus, the new claim (once the shortfallen source is removed) will be highly specific. This prevents victims from being ‘lumped in with the poor more generally’—even if the newly-acknowledged specific claim is different from the specific claim of having been wronged. A second objection asks: what if there is a shortfall and the ultimatelyevoked liability sources are instantiated by numerous organizations and their members? This is problematic in those cases where it seems clear that one organization alone bears the liability in the first instance—because of its wrongdoing—with others bearing liabilities only if the first organization reneges. To illustrate, consider again the climate-related burdens of polluting states in the year 2100. Suppose—hypothetically—that a historically low-emissions state (Rwanda, say) ends up with a great capacity to alleviate climate harms and ends up benefitting from climate change. Meanwhile, a state with historically high emissions per capita (such as Australia) also has a great capacity to alleviate climate change and benefits from climate change. How should burdens be divided amongst different states? My approach to shortfalls suggests that Australia has a liability to repair the damaged relationship with climate victims, and (at least some of) its members will have liabilities to bear costs. The collective-level and individual-level duties are based on capacity and benefit—but not wrongdoing, insofar as that is a shortfallen source. This is fine as far as it goes. But it produces strange interactions with the liabilities of Rwanda. After all, if capacity and benefit are the only sources in play (since we have removed the shortfallen source of wrongdoing), then should Rwanda and Rwandans pay as much as Australia and Australians? My proposal says ‘yes’. Intuitively, we want to say ‘no’. We can make the objection even stronger by imagining that Rwanda is by far the most capable of helping climate victims—perhaps because of its relative accessibility to climate refugees. It then appears that, on my account, Rwanda is liable to more cost than Australia, even assuming that both states benefit to the same extent. My reply to this objection emphasizes the intimate connection between being a participant in a damaged relationship and having a capacity to repair that relationship. ‘Capacity to repair a damaged relationship’ does not mean ‘gross domestic product’ or ‘general international political influence’.

  ’ 

181

In organizations’ lives, as in individuals’ lives, outsiders’ capacities for relationship repair tend to be conditional on the willingness of relationship participants to exercise their capacities for relationship repair. This makes it very odd to assert that Rwanda’s capacity to repair the damaged relationship between Australia and climate change victims could possibly be stronger than Australia’s capacity to do so. This is consistent with Rwanda having the strongest capacity to more generally assist climate change victims. If Rwanda has that capacity, then Rwanda is liable to the most cost in generally assisting the victims, while Australia is liable to the most cost in repairing the relationship. These objections highlight two details about the source-tracking proposal for distributing the costs of organizations’ wrongs amongst members. First, we must always be clear about the target of the cost—i.e., which outcome or action the cost is morally meant to serve—since this will affect which liability sources apply to which agent. For example, is cost being borne in order to ‘help climate change victims’, or ‘repair polluters’ relationship to climate change victims’? If the former, Rwanda’s (hypothetical) capacity looms large; if the second, Australia’s capacity and benefit gain more traction as sources of liability. This will affect how we deal with instances of historical wrongdoing by organizations, where contemporary members are not liable to wrongdoing-sourced costs. Second, we must be aware that contextual details can furnish particular agents with highly specific capacities, associations, and benefits, which can act as weighty sources of liability. Liabilities can require agents to exercise specific capacities, act in accordance with specific associations, or disgorge specific benefits. Thus while (hypothetical future) Rwanda might be closely associated with climate change victims, and might be liable to bear high costs with a view to aiding climate change victims on that basis, the association between Rwanda and the victims doesn’t render Rwanda liable to costs to fix others’ relationships. Likewise, while Australia might have benefited the most from the injustice, and so be liable to costs that disgorge or compensate for those particular benefits, other states will have received different benefits, that give them a unique and irreplaceable liability to compensate for those particular benefits. There is one last concern to address. Fittingly, this concern brings us full circle: it relates to Chapter 1’s arguments for realism about organizations. The concern runs as follows: the source-tracking model is concerned with the fair distribution of costs amongst individual members. Is such a concern really consistent with realism about organizations as entities, agents, and

182

  

wrongdoers? This question is motivated by the following: if all the costs imposed on an organization because of its wrongdoing must be justified on the basis of the attributes of individual members, then is there still any need to posit the organization’s liability to cost? The answer, of course, is ‘yes’. The justification for this takes us back to Chapter 1’s three reasons for realism about organizations. First, the fact that the organization bears liability is multiply realizable in facts about members: the organization’s bearing a liability is consistent with members’ liabilities having a wide range of different sources. Take Australia’s present-day liability to take on costs to reduce greenhouse gas emissions, under the Polluter Pays Principle. This liability persists despite members being born and dying, and changing jobs and social roles. Under source-tracking, such changes mean that changes should occur in Australia’s distribution of costs amongst members. This means the organization’s moral liability entails different distributions—is realized in different ways at the individual level—at different times. Yet each distribution is a realization of one overarching liability, which belongs to the organization itself, rather than to its members. Several facts unite these different possible distributions into one overarching organization-level liability: the distributions are enforced by the organization; the distributions include only those who are under the organization’s authority; and the distributions allow the organization to ‘tick off ’ a particular liability on its moral ‘to-do list’. These facts pick out the different cross-temporal distributions as each being a distribution of—a realization of—one and the same liability of the organization.¹⁰ Second, the organization’s liability has explanatory power. It explains why there are costs that need apportioning amongst members in the first place: they are the costs necessary for paying the organization’s liability. The organization’s liability also explains why we must ask whether (and to what extent) members instantiate this particular liability source, rather than that one: we should ask after all and only those sources that generate the organization’s liability. Other sources may generate liabilities for the members as private individuals, but the payment of those liabilities would not contribute to the payment of any liability of the organization. Third and finally, we should posit the organization’s liability because the organization is an agent that can do wrong, have capacities, receive benefits, and so on, in ways that are not reducible to these agential features of the

¹⁰ This explains why organizations’ liabilities’ effects on outsiders are not theorized in this book: unlike effects on members, effects on outsiders do not realize the organization’s paying of its debt. See the previous footnote.

  ’ 

183

organization’s members. Indeed, it is precisely the organization’s status as a wrongdoer—distinctive from members’ status as wrongdoers—that produces the shortfall problem. An organization’s status as a wrongdoer was explored earlier in this book, and that status is not undermined by the fact that an organization’s liabilities have effects on, and are constrained by, members’ liabilities.

7.5 Conclusion When an organization has a liability that requires the distribution of costs amongst members, how should those costs be distributed? I have answered: in proportion to the members’ respective instantiations of the sources of the organization’s liability. In cases where the organization’s liability is sourced in the organization’s having done wrong, members’ ‘instantiation’ of the liability-source will take the form of enactment, endorsement, or omission in relation to the organization’s wrong, as those three notions were developed in Chapter 6. But sometimes it is not possible to distribute the costs of organizations’ liabilities in a way that tracks all sources of the organization’s liability. This is clearest in cases of historical wrongdoing, in which no present-day members are implicated. Here, we should acknowledge that such shortfalls matter in cases that cry out for someone to act, even if no member instantiates the shortfallen source. This suggests that something other than the shortfallen source is doing work; that some other liability source is generating the imperative to make up the shortfall. The proposal in such cases is that we should understand organizations’ seemingly shortfallen liabilities as grounded in sources that are instantiated by a sufficient number of members, and so do not result in shortfalls. If no such sources are weighty enough to generate (enough of) a liability in a given case, then we should consider cost-sharing among many organizations. However, my hunch—not fully defended here—is that sources other than wrongdoing can carry much more weight than we often think, both in generating organizations’ liabilities and in enabling source-tracking distributions of those liabilities’ costs amongst members. If so, then where ever there are victims, there will almost always be organizations and individuals who bear duties. The result of this chapter might be unexpected: in a book devoted to organizations as wrongdoers, I have ended up suggesting the organizations’ status as wrongdoers might be less important—for placing burdens on

184

  

them—than is organizations’ status as capacity-bearers, beneficiaries, associators, or participants in damaged relationships. Yet this just shows that there is much more work to be done regarding the ethics of organizations as distinctive moral agents: wrongdoing is not the totality of normative enquiry. In the coming pages, I consider these and other questions that remain to be explored.

Conclusions I started this book by saying that I would focus on questions that have been underexplored in existing philosophical debates about organizations as wrongdoers. I hope to have demonstrated that those underexplored questions are crucial for gaining a full picture of how organizations can be wrongdoers. Those underexplored questions were varied. They concerned the metaphysical relation between organizations and members; the variety of lenses with which we can examine organizations’ blameworthiness; the role of feelings of guilt, remorse, and shame in being a wrongdoer; the variety of members’ involvement in organizations’ wrongdoing; and the distribution of reparative burdens amongst members. I answered these questions as follows. Organizations are structured wholes that have members as material parts, where members are those who have relevant inputs, commitments, and enactments concerning the organization’s decision-making procedure and actions. In short, members are to organizations as a motorcycle’s parts are to a motorcycle. Organizations can be held blameworthy under all three of the volitionist, attributivist, and aretaic lenses on wrongdoing. This produces far more organizational blameworthiness than the dominant volitionist approaches recognize. The capacity for ‘moral self-awareness’— which is a component of guilt, remorse, and shame—is essential for being held blameworthy as a wrongdoer. But this shouldn’t lead us to us give up on organizations as wrongdoers: organizations can have moral selfawareness—not merely its functional equivalent—when members have moral self-awareness qua members. Members are bound up with organizational wrongdoing simply by the fact of membership, but members are more particularly and more forcefully involved in an organization’s wrongdoing when they enact the organization’s wrong (including by complicity), when they endorse the organization’s wrong, and when they unreasonably omit to push back against the organization’s wrong. And organizations’ wrongs rightly imply costs for these enactors, endorsers, and omitters—though other members may also be on the hook for costs, in virtue of those

Organizations as Wrongdoers: From Ontology to Morality. Stephanie Collins, Oxford University Press. © Stephanie Collins 2023. DOI: 10.1093/oso/9780192870438.003.0008

186

  

members’ capacities, benefits, associations, and other sources of reparative burden. Despite this substantial headway on answering neglected questions, there are still many questions that remain to be explored. Some of these questions concern the details of organizations’ different structures, procedures, purposes, and goals. For example, in this book I have almost always talked about ‘members’ as if all members of organizations were human individuals. In fact, many members of organizations are themselves organizations (or collective agents more generally). This raises the question of at which ‘level’ we should attribute an organizational wrongdoing, in cases where there are multiple organizations or collective agents ‘nested’ within one another. This is particularly problematic in cases where the lower-level organizations receive a large amount of autonomy from the subsuming organization—as in the case of the UN or other intergovernmental organizations, and in the case of some nation-states that have federal systems. Also concerning the details of organizations’ make-up, there is more work to be done in fleshing out the picture of group attributivism and group virtue, which I began to sketch in Chapter 4. Key questions here include: where exactly are a group’s ‘character traits’, and in what exactly do its ‘evaluative attitudes’ inhere, if we resist the idea that these things are purely a matter of facts about members? I suggested that organizations’ character traits might be gleaned from norms and functionings, and that their values might be gleaned from their goals and culture. These concepts—norms, functionings, goals, culture—deserve more analysis in the organizational context, since norms, functionings, goals, and cultures are neither an explicit result of formal organization-level decision-making procedures, nor a matter merely of member actions and attitudes. Likewise, there is more to say about Chapter 4’s suggestion that different kinds of organizations should be held to different standards when assessing their virtue. If virtues are relative to a species, then what are the different ‘species’ of organization and what are the virtues that are particular to each of these ‘species’? Other avenues for future research generalize across all organizations. These include the conditions for organizations’ inheritance of blameworthiness across time. In Chapter 7, I assumed that Australia-in-2100 will be blameworthy for what was done by Australia-in-1980. But what if Western Australia secedes, or Australia becomes a republic, or there is a revolution? In that case, will it still make sense to hold Australia-in-2100 blameworthy for the wrongs of Australia-in-1980? And going backwards in time from the present, the same questions arise with arguably more urgency. Australia was



187

federated in 1901: does this mean that Australia-in-2023 is not morally blameworthy for the wrongs committed by the British colonies that were present in Australia in the nineteenth century? Can we tell a story on which Australia is blameworthy for the colonization of Australia, without making Australia blameworthy for events in sixteenth-century England? Similar issues arise for other kinds of organization, such as corporations. For corporations, the key concept here is ‘phoenixing’. Phoenixing occurs when one organization is disbanded and replaced with an almost-identical organization, for the purposes of avoiding legal liability. Morally, we want to hold the post-phoenixing organization blameworthy for any wrongs committed by the pre-phoenixing organization. It is difficult to explain on what philosophical basis we can do this. Another issue that generalizes across organizations is organizations’ rights: what are they owed, from each other and from us, morally and politically? If we are committed to the idea that organizations are moral agents that can do wrong, it becomes natural to think that wrongdoing and rights-bearing must be two separate avenues of normative inquiry, with different philosophical bases. That is, it becomes natural to want to commit to organizations’ moral agency while stringently avoiding a commitment to organizations’ moral patiency. We may want to assert that wrongdoing and rights-bearing are two entirely different statuses that have different bases. But this may require us to take large stands on debates in moral theory, such as the debate between the ‘will’ and ‘interest’ theories of rights. Perhaps our advocacy of organizations’ agency commits us to the ‘interest’ theory of rights: after all, if we sign up for the ‘will’ theory of rights, then our belief in organizations’ agency seems to commit us to organizations’ rights. But such knee-jerk reactions to long-standing debates might be too quick. It is possible that more nuance and pluralism is required here. Perhaps there is more to irreducible organization-level moral rights—at least for some organizations—than many are inclined to think.¹ Or perhaps there is something particular about organizations’ wills that precludes their rights, even though rights generally are grounded in the bearer’s will. In general, there are all sorts of concepts in moral theory—concepts such as rights—whose extension from humans to organizations deserves more investigation. I hope that the theory developed in this book will aid us in figuring out these other aspects of organizations’ place in our moral, social, and political

¹ I explore issues related to groups’ rights in Collins 2020b and Collins 2022b.

188

  

world. That theory includes the metaphysical picture developed in Part I, according to which organizations are real material agents with individuals as material parts; the moral picture developed in Part II, according to which organizations are blameworthy agents capable of moral self-awareness; and the membership picture developed in Part III, according to which members are morally linked to their organizations in numerous ways. Wrongdoing is not the totality of normative inquiry, but a theory of organizations’ status as wrongdoers paves the way for us to theorize their moral and political status more generally.

References Adams, Michael A., Grace Borsellino, and Angus Young. 2017. ‘Regulating Conduct of Financial Institutions in Australia: Is Culture the New Frontier of Regulation?’ Journal of International Banking Law and Regulation 32(3), 86–92. Adams, Robert Merrihew. 1983. ‘Involuntary Sins’. Philosophical Review 94, 3–31. Angwin, Julia, Jeff Larson, Surya Mattu, and Lauren Kirchner. 2016. ‘Machine Bias’. Pro Publica, 23 May. Available at https://www.propublica.org/article/machinebias-risk-assessments-in-criminal-sentencing Baddorf, Matthew. 2017. ‘Phenomenal Consciousness, Collective Mentality, and Collective Moral Responsibility’. Philosophical Studies 174, 2769–86. Bakan, Joel. 2004. The Corporation: The Pathological Pursuit of Profit and Power. New York: Simon & Schuster. Baker, Lynne Rudder. 2000. Persons and Bodies: A Constitution View. Cambridge: Cambridge University Press. Baron, Jonathan and Ilana Ritov. 2004. ‘Omission Bias, Individual Differences, and Normality’. Organizational Behavior and Human Decision Processes 94, 74–85. Beerbohm, Eric. 2012. In Our Name: The Ethics of Democracy. Princeton: Princeton University Press. Beggs, Donald. 2003. ‘The Idea of Group Moral Virtue’. Journal of Social Philosophy 34(4), 457–74. Björnsson, Gunnar and Kendy Hess. 2016. ‘Corporate Crocodile Tears? On the Reactive Attitudes of Corporate Agents’. Philosophy and Phenomenological Research 94(2), 273–98. Block, Ned. 1978. ‘Troubles with Functionalism’. Minnesota Studies in the Philosophy of Science 9, 261–325. Bourget, David. 2017. ‘The Role of Consciousness in Grasping and Understanding’. Philosophy and Phenomenological Research 95(2), 285–318. Braham, Matthew and Martin van Hees. 2011. ‘Responsibility Voids’. The Philosophical Quarterly 61(242), 6–15. Bratman, Michael. 2014. Shared Agency: A Planning Theory of Acting Together. Oxford: Oxford University Press. Cane, Peter. 2002. Responsibility in Law and Morality. Oxford: Hart Publishing. Ceva, Emanuela and Michele Bocchiola. 2019. Is Whistleblowing a Duty? Polity. Clarke, Randolph. 2014. Omissions. Oxford University Press. Coase, Ronald. 1937. ‘The Nature of the Firm’. Economica 4, 386–405. Coffee, John C Jr.1981. ‘ “No Soul to Damn: No Body to Kick”: An Unscandalized Inquiry into the Problem of Corporate Punishment’. Michigan Law Review 79, 386. Collins, Stephanie. 2015. The Core of Care Ethics. Basingstoke: Palgrave Macmillan.

190



Collins, Stephanie. 2016. ‘Distributing States’ Duties’. Journal of Political Philosophy 24(3), 344–66. Collins, Stephanie. 2017a. ‘Filling Collective Duty Gaps’. Journal of Philosophy 114(1), 573–91. Collins, Stephanie. 2017b. ‘Duties of Group Agents and Group Members’. Journal of Social Philosophy 48(1), 38–57. Collins, Stephanie. 2018. ‘The Government Should Be Ashamed: On the Possibility of Organisations’ Emotional Duties.’ Political Studies 66(4), 813–829. Collins, Stephanie. 2019a. Group Duties: Their Existence and Their Implications for Individuals. Oxford: Oxford University Press. Collins, Stephanie. 2019b. ‘Collective Responsibility Gaps’. Journal of Business Ethics 154(4), 943–54. Collins, Stephanie. 2020a. ‘How Much Can We Ask of Collective Agents?’ 50(7), 815–31. Collins, Stephanie. 2020b. ‘Are Organizations’ Religious Exemptions Democratically Defensible?’ Daedalus 149(3), 105–18. Collins, Stephanie. 2022a. ‘I, Volkswagen.’ Philosophical Quarterly 72(2), 283–304. Collins, Stephanie. 2022b. ‘A Human Right to Relationships?’ In Kimberley Brownlee, David Jenkins, and Adam Neal (eds), Being Social: The Philosophy of Social Human Rights. Oxford: Oxford University Press. Collins, Stephanie. Forthcoming. ‘Responsible Citizens, Irresponsible States by Avia Pasternak’. The Philosophical Review. Collins, Stephanie and Niels de Haan. 2021. ‘Interconnected Blameworthiness’. The Monist 104(2), 195–209. Collins, Stephanie and Luara Ferracioli. Forthcoming. ‘Care for a Profit?’ Perspectives on Politics. Collins, Stephanie and Holly Lawford-Smith. 2016. ‘Collectives’ Obligations and Individuals’ Obligations: A Parity Argument’. Canadian Journal of Philosophy 46(1), 38–58. Collins, Stephanie and Holly Lawford-Smith. 2017. ‘Responsibility for States’ Actions: Normative Issues at the Intersection of Collective Agency and State Responsibility.’ Philosophy Compass 12(11), 1–8. Collins, Stephanie and Holly Lawford-Smith. 2021. ‘We the People: Is the Polity the State?’ Journal of the American Philosophical Association 7(1), 78–97. Copp, David. 2007. ‘The Collective Moral Autonomy Thesis’. Journal of Social Philosophy 38(3), 369–88. Cordell, Sean. 2017. ‘Group Virtues: No Group Leap Forward with Collectivism’. Res Publica 23, 43–59. Couch, Mark. 2004. ‘Discussion: A Defense of Bechtel and Mundale’. Philosophy of Science 71(2), 198–204. Crawford, James. 2007. The Creation of States in International Law. Oxford: Oxford University Press. Crawford, James. 2012. Brownlie’s Principles of Public International Law. 8th edition. Oxford: Oxford University Press. Crawford, James and Jeremy Watkins. 2010. ‘International Responsibility’. In S. Besson and J. Tasioulas (eds), The Philosophy of International Law, Oxford: Oxford University Press, 283–98.



191

Currie, Gregory. 1984. ‘Individualism and Global Supervenience’. British Journal for the Philosophy of Science 35(4), 345–58. Cushman, Fiery, Liane Young, and Marc Hauser. 2006. ‘The Role of Conscious Reasoning and Intuition in Moral Judgements: Testing Three Principles of Harm’. Psychological Science 17, 1082–9. Dietz, Alexander. 2020. ‘Are My Temporal Parts Agents?’ Philosophy and Phenomenological Research 100(2), 362–79. Doris, John. 1998. ‘Persons, Situations, and Virtue Ethics’. Nous 32(4), 504–30. Downs, Anthony. 1967. Inside Bureaucracy. Boston: Little, Brown and Company. Effingham, Nick. 2010. ‘The Metaphysics of Groups’. Philosophical Studies 149, 251–67. Elder-Vass, Dave. 2007. ‘For Emergence: Refining Archer’s Account of Social Structure’. Journal of the Theory of Social Behaviour 37, 25–44. Elster, John. 1989. Nuts and Bolts for the Social Sciences. Cambridge University Press. Epstein, Brian. 2015. The Ant Trap: Rebuilding the Foundations of the Social Sciences. Oxford: Oxford University Press. Erskine, Toni. 2001. ‘Assigning Responsibilities to Institutional Moral Agents: The Case of States and Quasi-states’. Ethics & International Affairs 15(2), 67–89. Erskine, Toni (ed.). 2003. Can Institutions Have Responsibilities? Collective Moral Agency and International Relations. Basingstoke: Palgrave Macmillan. Erskine, Toni. 2010. ‘Kicking Bodies and Damning Souls: The Danger of Harming “Innocent” Individuals While Punishing “Delinquent” States’. Ethics & International Affairs 24(3), 261–85. Fairclough, Norman. 2005. ‘Discourse Analysis in Organization Studies: The Case for Critical Realism’. Organization Studies 26, 915–39. Feinberg, Joel. 1968. ‘Collective Responsibility’. Journal of Philosophy 65(21), 674–88. Fischer, John Martin and Mark Ravizza. 1998. Responsibility and Control. Cambridge: Cambridge University Press. Fodor, Jerry A. 1968. Psychological Explanation: An Introduction to the Philosophy of Psychology. New York: Crown Publishing Group/Random House. Foot, Philippa. 1978. ‘Virtues and Vices’. In Virtues and Vices and Other Essays in Moral Philosophy. Oxford: Oxford University Press. French, Peter. 1984. Collective and Corporate Responsibility. New York: Columbia University Press. French, Peter. 1995. Corporate Ethics. Fort Worth: Harcourt Brace College Publishers. Fricker, Miranda. 2010. ‘Can There Be Institutional Virtues?’ In T.S. Gendler and J. Hawthorne (eds), Oxford Studies in Epistemology 3, Oxford: Oxford University Press, 235–52. Fricker, Miranda. 2016. ‘What’s the Point of Blame? A Paradigm Based Explanation’. Nous 50(1), 165–83. Friedman, Milton. 1970. ‘The Moral Responsibility of Business is to Increase Its Profits’. New York Times Magazine. Gilbert, Margaret. 1989. On Social Facts. Princeton: Princeton University Press. Gilbert, Margaret. 2000. Sociality and Responsibility: New Essays in Plural Subject Theory. Lanham: Rowman and Littlefield.

192



Gilbert, Margaret. 2002. ‘Collective Guilt and Collective Guilt Feelings’. Journal of Ethics 6(2), 115–43. Goodpaster, Kenneth E. and John B. Mathews Jr. 1982. ‘Can a Corporation Have a Conscience?’ Harvard Business Review 60, 132–41. Harman, Gilbert. 2003. ‘No Character or Personality’. Business Ethics Quarterly 13(1), 87–94. Harre, Rom. 1979. Being Social. Oxford: Blackwell. Harris, Keith. 2020. ‘How Individuals Constitute Group Agents’. Canadian Journal of Philosophy 50(3), 350–64. Hartman, Edwin M. 2015. Virtue in Business: Conversations with Aristotle. Cambridge: Cambridge University Press. Haslanger, Sally. 2012. Resisting Reality: Social Construction and Social Critique. New York: Oxford University Press. Hasnas, John. 2009. ‘The Centenary of Mistake: One Hundred Years of Corporate Criminal Liability’. American Criminal Law Review 46, 1329–58. Hatch, Mary Jo. 1997. Organization Theory: Modern, Symbolic, and Postmodern Perspectives. Oxford: Oxford University Press. Hawley, Katherine. 2017. ‘Social Mereology’. Journal of the American Philosophical Association 3(4), 395–411. Hayek, F. 1948. Individualism and Economic Order. Chicago: University of Chicago Press. Held, Virginia. 1970. ‘Can a Random Collection of Individuals be Morally Responsible?’ Journal of Philosophy 67(14), 471–81. Herzog, Lisa. 2018. Reclaiming the System: Moral Responsibility, Divided Labour, and the Role of Organisations in Society. Oxford: Oxford University Press. Hess, Kendy. 2013. ‘Missing the Forest for the Trees: The Theoretical Irrelevance of Shared Intentions’. In The Background of Social Reality. Dordrecht: Springer, 57–75. Hess, Kendy M. 2010. ‘The Modern Corporation as Moral Agent: The Capacity for “Thought” and a “First-Person Perspective” ’. Southwest Philosophy Review 26(1), 51–69. Hess, Kendy M. 2018a. ‘The Peculiar Unity of Corporate Agents’. In Kendy M. Hess, Violetta Igneski, and Tracy Isaacs (eds), Collectivity: Ontology, Ethics, and Social Justice. Lanham, MD: Rowman and Littlefield, 35–60. Hess, Kendy M. 2018b. ‘Does The Machine Need a Ghost? Corporate Agents as Nonconscious Kantian Moral Agents’. Journal of the American Philosophical Association 4(1), 67–86. Heugens, Pursey P.M.A.R., Muel Kaptein, and J. van Oosterhout. 2008. ‘Contracts to Communities: A Processual Model of Organizational Virtue’. Journal of Management Studies 45, 100–21. Hill, Jennifer. 2003. ‘Corporate Criminal Liability in Australia: An Evolving Corporate Governance Technique?’ Journal of Business Law 1, 1–44. Hindriks, Frank. 2013. ‘The Location Problem in Social Ontology’. Synthese 190(3), 413–37. Hindriks, Frank. 2018. ‘Collective Agency: Moral and Amoral’. Dialectica 72(1), 2–23.



193

Hindriks, Frank. 2021. ‘Establishments as Material Rather than Immaterial Objects’. Australasian Journal of Philosophy 99(4), 835–40. Hodgson, Geoffrey M. 2007. ‘Institutions and Individuals: Interaction and Evolution’, Organization Studies 28, 97–116. Holmstrom, B. and A. Roberts. 1998. ‘The Boundaries of the Firm Revisited’. Journal of Economic Perspectives 12, 73–94. Hooker, Brad. 2000. Ideal Code, Real World. Oxford: Oxford University Press. Horden, John and Dan López de Sa. 2020. ‘Groups as Pluralities’. Synthese 198(11), 10237–71. Huebner, Bryce. 2011. ‘Genuinely Collective Emotions’. European Journal for Philosophy of Science 1, 89–118. Huebner, Bryce. 2014. Macrocognition: A Theory of Distributed Minds and Collective Intentionality. New York: Oxford University Press. Hursthouse, Rosalind. 1999. On Virtue Ethics. Oxford: Oxford University Press. Hursthouse, Rosalind and Glen Pettigrove. 2016. ‘Virtue Ethics’. In Stanford Encycloepdia of Philosophy. Isaacs, Tracy. 2011. Moral Responsibility in Collective Contexts. Oxford: Oxford University Press. Jackson, Frank and Philip Pettit. 1992. ‘Structural Explanation in Social Theory’. In D. Charles and K. Lennon (eds), Reduction, Explanation, and Realism. Oxford: Clarendon Press, 97–131. Jensen, Michael C. and William H. Meckling. 1976. ‘Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure’. Journal of Financial Economics 3(4), 305–60. Kim, Jaegwon. 1984. ‘Concepts of Supervenience’. Philosophy and Phenomenological Research 45(2), 153. Kincaid, Harold. 1986. ‘Reduction, Explanation, and Individualism’. Philosophy of Science 53(4), 492–513. Konzelmann Ziv, Anita. 2012. ‘Institutional Virtue: How Consensus Matters’. Philosophical Studies 161, 87–96. Koslicki, Kathrin. 2008. The Structure of Objects. Oxford: Oxford University Press. Krishnan, Maya. 2020. ‘Against Interpretability: A Critical Examination of the Interpretability Problem in Machine Learning’. Philosophy & Technology 33, 487–502. Kumar, Rahul and Kok-Chor Tan, eds. 2006. Reparations. Special Issue of Journal of Social Philosophy 37(3), pp. iv–v, 323–482. Kukathas, Chandran. 2006. ‘Who? Whom? Reparations and the Problem of Agency’. Journal of Social Philosophy 37(3), 330–41. Kutz, Christopher. 2000. Complicity: Ethics and Law for a Collective Age. Cambridge: Cambridge University Press. Kutz, Christopher. 2002. ‘The Collective Work of Citizenship’. Legal Theory 8(4), 471–94. Lawford-Smith, Holly. 2019. Not in Their Name: Are Citizens Culpable for Their States’ Actions? Oxford: Oxford University Press. Lepora, Chiara and Robert E. Goodin. 2013. On Complicity and Compromise. Oxford: Oxford University Press.

194



Levy, Neil. 2005. ‘The Good, the Bad and the Blameworthy’. Journal of Ethics and Social Philosophy 1(2), 2–16. Lincoln, James R., Michael L. Gerlach, and Christina L. Ahmadjian. 1996. ‘Keiretsu Networks and Corporate Performance in Japan’. American Sociological Review 61(1), 67–88. List, Christian. 2018. ‘What Is It Like to Be a Group Agent?’ Nous 52(2), 295–319. List, Christian. 2021. ‘Group Agency and Artificial Intelligence’. Philosophy and Technology 4, 1–30. List, Christian and Peter Menzies. 2009. ‘Non-Reductive Physicalism and the Limits of the Exclusion Principle’. The Journal of Philosophy CVI(9), 475–502. List, Christian and Philip Pettit. 2011. Group Agency: The Possibility, Design, and Status of Corporate Agents. Oxford: Oxford University Press. List, Christian and Kai Spiekermann. 2013. ‘Methodological Individualism and Holism in Political Science: A Reconciliation’. American Political Science Review 107(4), 629–43. Lombard, L.1990. ‘Causes, Enablers, and the Counterfactual Analysis’. Philosophical Studies 59, 195–211. Ludwig, Kirk. 2014. ‘Proxy Agency in Collective Action’. Nous 48(1), 75–105. Ludwig, Kirk. 2017a. From Plural to Institutional Agency: Collective Action II. Oxford: Oxford University Press. Ludwig, Kirk. 2017b. ‘Do Corporations Have Minds of Their Own?’ Philosophical Psychology 30(3), 265–97. Macdonald, Graham and Philip Pettit. 1981. Semantics and Social Science. London: Routledge. MacIntyre, Alasdair. 1984. After Virtue. Notre Dame: University of Notre Dame Press. Macnamara, Coleen. 2011. ‘Holding Others Responsible’. Philosophical Studies 152(1), 81–102. May, Larry. 1987. The Morality of Groups: Collective Responsibility, Group-based Harm, and Corporate Rights. Notre Dame, IN: University of Notre Dame Press. Miller, David. 2004. ‘Holding Nations Responsible’. Ethics 114, 240–68. Miller, David. 2007. National Responsibility and Global Justice. Oxford: Oxford University Press. Miller, Seamus. 2010. The Moral Foundations of Social Institutions: A Philosophical Study. New York: Cambridge University Press. Moore, Geoff. 2005. ‘Humanizing Business: A Modern Virtue Ethics Approach’. Business Ethics Quarterly 15(2), 237–55. Nefsky, Julia. 2019. “Collective Harm and the Inefficacy Problem.” Philosophy Compass 14(4), 1–17. O’Connor, Timothy and Christopher Franklin. 2018. ‘Free Will.’ In Edward N Zalta (ed.), The Stanford Encyclopedia of Philosophy. Fall 2018 Edition. Available at http://plato.stanford.edu/archives/fall2018/entries/freewill/. Passinsky, Asya. 2021. ‘Norm and Object: A Normative Hylomorphic Theory of Social Objects.’ Philosophers’ Imprint 21(25), 1–21.



195

Pasternak, Avia. 2011. ‘Sharing the Costs of Political Injustices’. Politics, Philosophy & Economics 10(2), 188–210. Pasternak, Avia. 2013. ‘Limiting States’ Corporate Liability’. Journal of Political Philosophy 21(4), 361–81. Pasternak, Avia. 2021. Responsible Citizens, Irresponsible States: Should Citizens Pay for Their States’ Wrongdoings? Oxford: Oxford University Press. Perry, John. 1979. ‘The Problem of the Essential Indexical’. Nous 13(1), 3–21. Pettit, Philip. 1996. The Common Mind: An Essay on Psychology, Society, and Politics. Oxford: Oxford University Press. Pettit, Philip. 2007. ‘Responsibility Incorporated’. Ethics 117, 141–201. Pettit, Philip. 2018. ‘Consciousness Incorporated’. Journal of Social Philosophy 49(1), 12–37. Popper, Karl. 1945. The Open Society and Its Enemies. London: Routledge. Railton, Peter. 1984. ‘Alienation, Consequentialism, and the Demands of Morality’. Philosophy and Public Affairs 13(2), 134–71. Rawls, John. 1993. Political Liberalism. New York: Columbia University Press. Reed, Michael I. 2001. ‘Organization, Trust, and Control: A Realist Analysis’. Organization Studies 22, 201–28. Ripstein, Arthur. 2004. ‘Justice and Responsibility’. Canadian Journal of Law & Jurisprudence 17(2), 361–86. Ritchie, Katherine. 2013. ‘What Are Groups?’ Philosophical Studies 166(2), 157–72. Ritchie, Katherine. 2015. ‘The Metaphysics of Social Groups’. Philosophy Compass 10(5), 310–21. Ritchie, Katherine. 2018. ‘Social Creationism and Social Groups’. In Kendy Hess, Tracy Isaacs, and Violetta Igneski (eds), Collectivity: Ontology, Ethics, and Social Justice. Lanham, MN: Rowman and Littlefield. Ritov, Ilana and Jonathan Baron. 1990. ‘Reluctance to Vaccinate: Omission Bias and Ambiguity’. Journal of Behavioral Decision Making 3, 263–77. Rönnegard, David. 2015. The Fallacy of Corporate Moral Agency. New York: Springer. Rönnegard, David and Manuel Velasquez. 2017. ‘On (Not) Attributing Moral Responsibility to Organizations’. In Eric Orts and N. Craig Smith (eds), The Moral Responsibility of Firms. Oxford University Press, 123–42. Ross, W.D. 2002 [1930]. The Right and the Good. Oxford: Oxford University Press. Rovane, Carol. 1998. Bounds of Agency: An Essay in Revisionary Metaphysics. Princeton: Princeton University Press. Royal Commission into Misconduct in the Banking, Superannuation and Financial Services Industry. 2018. Interim Report. Volume 1. Available at https://www. royalcommission.gov.au/banking/interim-report Royal Commission into Misconduct in the Banking, Superannuation and Financial Services Industry. 2019. Final Report. Volume 1. Available at https://www. royalcommission.gov.au/royal-commission-misconduct-banking-superannuationand-financial-services-industry Ruben, David-Hillel. 1983. ‘Social Wholes and Parts’. Mind 92(366), 219–38.

196



Ruben, David-Hillel. 1985. The Metaphysics of the Social World. London: Routledge and Kegan Paul. Rudd, Kevin. 2008. ‘Apology to Australia’s Indigenous Peoples’. House Hansard, Parliament of Australia. Available at https://parlinfo.aph.gov.au/parlInfo/search/ display/display.w3p;db=CHAMBER;id=chamber%2Fhansardr%2F2008-02-13% 2F0003;query=Id%3A%22chamber%2Fhansardr%2F2008-02-13%2F0003%22 Russell, Bertrand. 1912. The Problems of Philosophy. Oxford: Oxford University Press. Sartorio, Carolina. 2005. ‘Causes as Difference-Makers’. Philosophical Studies 123, 71–96. Sartorio, Caroline. 2009. ‘Omissions and Causalism’. Nous 43(3), 513–30. Sawyer, R. Keith. 2002. ‘Durkheim’s Dilemma: Toward a Sociology of Emergence’. Sociological Theory 20(2), 227–47. Scanlon, T.M. 1998. What We Owe to Each Other. Cambridge, MA: Harvard University Press. Scanlon, T.M. 2008. Moral Dimensions: Permissibility, Meaning, Blame. Cambridge, MA: Belknap Press of Harvard University Press. Scarantino, Andrea and de Sousa, Ronald. 2018. ‘Emotion’. Stanford Encyclopedia of Philosophy. Schaffer, Jonathan. 2016. ‘The Metaphysics of Causation’. Stanford Encyclopedia of Philosophy. Schmid, Hans Bernhard. 2014. ‘Plural Self-Awareness’. Phenomenology and the Cognitive Sciences 13(1), 7–24. Schmid, Hans Bernhard. 2018. ‘Collective Responsibilities of Random Collections: Plural Self-Awareness among Strangers’. Journal of Social Philosophy 49(1), 91–105. Schudt, Karl. 2000. ‘Taming the Corporate Monster: An Aristotelean Approach to Corporate Virtue’. Business Ethics Quarterly 10(3), 711–23. Searle, John. 2005. ‘What Is An Institution?’ Journal of Institutional Economics 1(1), 1–22. Shapiro, Lawrence. 2000. ‘Multiple Realizations’. The Journal of Philosophy 97(12), 635–54. Sheehy, Paul. 2006. ‘Sharing Space: The Synchronic Identity of Social Groups’. Philosophy of the Social Sciences 36(2), 131–48. Sher, George. 2005. ‘Transgenerational Compensation’. Philosophy & Public Affairs 33, 181–201. Shoemaker, David. 2011. ‘Attributability, Answerability, and Accountability: Toward a Wider Theory of Moral Responsibility’. Ethics 121(3), 602–32. Shoemaker, David. 2015. Responsibility from the Margins. Oxford: Oxford University Press. Shoemaker, David. 2019. ‘Blameworthy but Unblameable: A Paradox of Corporate Responsibility’. Georgetown Journal of Law & Public Policy 17, 897–917. Silver, Kenneth. 2019. ‘Can a Corporation Be Worthy of Moral Consideration?’ Journal of Business Ethics 159, 253–65.



197

Simmons, A. John. 1995. ‘Historical Rights and Fair Shares’, Law and Philosophy 14, 149–84 Smith, Angela. 2005. ‘Responsibility for Attitudes: Activity and Passivity in Mental Life’. Ethics 115(2), 236–71. Smith, Angela. 2015. ‘Control Responsibility, and Moral Assessment’. Philosophical Studies 138(3), 367–92. Smith, Barry. 2003. ‘John Searle: From Speech Acts to Social Reality’ In John Searle. Cambridge: Cambridge University Press, 1–33. Smith, Leonie. 2018. ‘The Curious Case of Ronald McDonald’s Claim to Rights: An Ontological Account of Differences in Group and Individual Person Rights’. Journal of Social Ontology 4(1), 1–28. Solomon, Robert C. 1993. Ethics and Excellence: Cooperation and Integrity in Business. Oxford: Oxford University Press. Sreenivasan, Gopal. 2002. ‘Errors about Errors: Virtue Theory and Trait Attribution’. Mind 111, 47–68. Srinivasan, Amia. 2015. ‘Stop the Robot Apocalypse’. London Review of Books 37, 3–6. Available at http://www.lrb.co.uk/v37/n18/amia-srinivasan/stop-the-robotapocalypse. Stavins, Robert. 2018. ‘An Economist’s Take on the Poland Climate Conference: The Glass is More than Half Full’. The Conversation. 18 December 2018. Available at https://theconversation.com/an-economists-take-on-the-poland-climate-conferencethe-glass-is-more-than-half-full-108915. Stilz, Anna. 2009. Liberal Loyalty: Freedom, Obligation, and the State. Princeton: Princeton University Press. Stilz, Anna. 2011. ‘Collective Responsibility and the State’. Journal of Political Philosophy 19(2), 190–208. Strand, Anders. 2012. ‘Group Agency, Responsibility, and Control’. Philosophy of the Social Sciences 43(2), 201–24. Strawson, Peter. 1962. ‘Freedom and Resentment’. Proceedings of the British Academy 48, 1–25. Strohmaier, David. 2018. ‘Group Membership and Parthood’. Journal of Social Ontology 4(2), 121–35. Strohmaier, David. 2020. ‘Two Theories of Group Agency’. Philosophical Studies 177(7), 1901–18. Sugden, Robert. 2003. ‘The Logic of Team Reasoning’. Philosophical Explorations 6(3), 165–81. Sunstein, Cass. 1995. ‘Incompletely Theorised Agreements’. Harvard Law Review 108(7), 1733. Thomasson, Amie. 2009. ‘Social Entities’. In Robin Le Poidevin (ed.), Routledge Companion to Metaphysics. London: Routledge. Thomasson, Amie. 2019. ‘The Ontology of Social Groups’. Synthese 196, 4829–45. Thompson, Janna. 2002. Taking Responsibility for the Past: Reparations and Historical Injustice. Cambridge: Polity Press. Todd, Benjamin. 2014. ‘Why and How to Earn to Give’. 80,000 Hours. Available at https://80000hours.org/articles/earning-to-give/

198



Tollefsen, Deborah. 2006. ‘The Rationality of Collective Guilt’. Midwest Studies in Philosophy 30(1), 222–39. Tollefsen, Deborah. 2015. Groups as Agents. Cambridge: Polity Press. Tuomela, Raimo. 2006. ‘Joint Intention, We-Mode and I-Mode’. Midwest Studies in Philosophy 30(1), 35–58. Tuomela, Raimo. 2007. The Philosophy of Sociality: The Shared Point of View, Oxford: Oxford University Press. Tuomela, Raimo. 2013. Social Ontology: Collective Intentionality and Group Agents. Oxford: Oxford University Press. Turner, Jonathan. 1997. The Institutional Order. London: Longman. Uzquiano, Gabriel. 2004. ‘The Supreme Court and the Supreme Court Justices: A Metaphysical Puzzle’. Nous 38(1), 135–53. Uzquiano, Gabriel. 2018. ‘Groups: Toward a Theory of Plural Embodiment’. Journal of Philosophy 115(8), 423–52. Vargas, Manuel. 2005. ‘The Trouble with Tracing’. Midwest Studies in Philosophy XXIX, 269–91. Varzi, Achille. 2016. ‘Mereology’. In Stanford Encyclopedia of Philosophy. Velasquez, Manuel. 2003. ‘Why Corporations Are Not Morally Responsible for Anything They Do’. Business Ethics Quarterly 13(4), 531–62. Vernon, Richard. 2003. ‘Against Restitution’. Political Studies 51, 542–57. Walker, Margaret. 2006a. ‘Restorative Justice and Reparations’. In Kumar, Rahul and Kok-Chor Tan, eds. Reparations, pp. 377–95. Special Issue of Journal of Social Philosophy 37(3), pp. iv–v, 323–482. Walker, Margaret. 2006b. Moral Repair: Reconstructing Moral Relations after Wrongdoing. Cambridge: Cambridge University Press. Watson, Gary. 1996. ‘Two Faces of Responsibility’. Philosophical Topics 24(2), 227–48. Watson, Gary. 2004. Agency and Answerability. Oxford: Oxford University Press. Weber, Max. 1968. Economy and Society: An Outline of Interpretive Sociology. Ed. Guenther Roth. New York: Bedminster Press. Weick, K. 1979. The Social Psychology of Organizing. Redding, MA: Addison-Wesley. Weick, Karl E. and Kathleen M. Sutcliffe. 2001. Managing the Unexpected: Resilient Performance in an Age of Uncertainty. Oxford: Wiley. Wenar, Leif. 2006. ‘Reparations for the Future’. In Kumar, Rahul and Kok-Chor Tan, eds. Reparations, pp. 396–405. Special Issue of Journal of Social Philosophy 37(3), pp. iv–v, 323–482. Wendt, Alexander. 2004. “The State as Person in International Theory.” Review of International Studies 30(2), 289–316. Wilhelm, Isaac. 2020. ‘The Stage Theory of Groups’. Australasian Journal of Philosophy 98, 661–74. Williams, Bernard. 1973. ‘A Critique of Utilitarianism’. In J.J.C. Smart and Bernard Williams, Utilitarianism: For and Against. Cambridge: Cambridge University Press, 77–150. Williams, Bernard. 1981. ‘Moral Luck’. In Williams, Moral Luck. Cambridge: Cambridge University Press, 20–40. Wu, Wayne. 2018. ‘The Neuroscience of Consciousness’. In Stanford Encyclopedia of Philosophy.

Index For the benefit of digital users, indexed terms that span two pages (e.g., 52–53) may, on occasion, appear on only one of those pages. Akrasia 101–2 Anchoring 34–6 Aretaicism 97–104 Attributivism 92–7 Australian Banking Royal Commission, see Royal Commission into Misconduct in the Banking, Superannuation and Financial Services Industry Blameworthiness Of machines 108–13, 117 Of organizations’ members 133–6, 138, 140–2, 146–9, 154–5 Of non-members 33, 35–6, 68, 71, 129–30, 137–8 Of organizations for volitions, see Volitionism Of organizations for attitudes, see Attributivism Of organizations for character traits, see Aretaicism Theories of 84–5 Bourget, David 115–16 Causation 19–20, 30–1, 53 Character traits, see Aretaicism Citizenship, see States Clarke, Randolph 153 Collective agency 13–14 Complicity 140–1, 150 Constitution 37, 51–2 Control 88–90 Corporations 5–7, 65–75, 94–5, 137, 147–8 Corporate culture 6, 13–14, 30–1, 67–8, 90–1, 95–104 De se content 114–15 Division of labour 12–13

Earning to give 147–8 Emotions, See Reactive attitudes, see also Phenomenology Epstein, Brian 32–5, 37 Eudaimonia 102–4 Explanatory power 17–20, 182 French, Peter 9–10, 98–9, 103, 123 Fricker, Miranda 96–7 Functionalism 20–4, 108–13, 127 Fusions 29–30 Gilbert, Margaret 43, 143–4 Goodin, Robert E. 140–2 Grounding 36–8 Hawley, Katherine 39–42 Held, Virginia 86–7 Hess, Kendy 43, 62n.10, 79 Hindriks, Frank 47, 63–4 Historical wrongdoing, see Responsibility gaps Holism, see Realism Individualism, see Realism Intentional stance, see Interpretivism Interpretivism 20–4, 127 Koslicki, Kathrin 11n.3, 38, 51–2 Law 5–8, 41–2, 58, 75–6, 91–2 Lepora, Chiara 140–2 Levy, Neil 88, 93–4 List, Christian 14–17, 19–21, 87–8, 108–9, 125–6, 136–7 Location, see Materiality Ludwig, Kirk 73–4 Materiality 27–8, 37–8, 46–7, 54, 59–61, 79–80 May, Larry 166

200



Membership Account of 51–9, 70, 75 Actions of 57, 91, 123–4, 138, 142 Commitments of 54–6, 67–8, 95–6, 144, 148–9 Inputs of 56–7, 69–70 Desiderata on an account of 48–9 Mereology, see Parthood Miller, David 161–2, 165 Moral self-awareness 113–20, 123–4, 128 Multiple realizability 15–17, 32, 182 Nexus of contracts 6–7 Omissions 89, 149–55 Ontology, see Realism Organizations, definition of 8–14 Parthood 5, 38–42, 51–4 Pasternak, Avia 55n.5, 160n.4, 165–7, 171–2 Pettit, Philip 6, 14, 16–17, 20–1, 60, 87–90, 108–11, 125–6 Phenomenology 113–20, 122–3 Pluralities 29–30 Polluter Pays Principle 7–8, 75–6, 156–7, 159, 169, 171, 173, 175, 180 Reactive attitudes 84, 119, 133–4 Realism 14–24, 28–9, 31–2, 118, 181–2 Reasoning 110–13, 116–17 Responsibility, see Blameworthiness Responsibility gaps 60, 90, 100, 157, 160, 171–83

Ritchie, Katherine 11, 45–6, 51 Rovane, Carol 124–5 Royal Commission into Misconduct in the Banking, Superannuation and Financial Services Industry 5–7, 30–1, 65–75, 99, 139, 141–2, 156–7, 177 Ruben, David-Hillel 63–5 Sartorio, Carolina 151 Scanlon, T.M. 84, 98n.2, 106 Schmid, Hans Bernhard 120 Sets 29–30 Sheehy, Paul 40–1, 51, 53n.2 Smith, Angela 85, 93–5 States 54–7, 75–9, 121–2, 149–50, 169 Strawson, Peter 84 Strohmaier, David 44–5 Source-tracking model 163–71 Stilz, Anna 171–2 Structure 9–11, 21, 23–4, 51–3, 80, 125 Supervenience 31–4 Tollefsen, Deborah 21, 125–6 Tuomela, Raimo 9–10, 120–1, 143–4 Vice and virtue, see Aretaicism Volitionism 86–92 Walker, Margaret Urban 173 Wenar, Leif 172–3 Whistle-blowing 155 Williams, Bernard 133–4, 147 Wrongdoing, see Blameworthiness