Introduction
In 2005, Volkswagen of America learned that its diesel vehicles could not meet emissions standards in the United States. Rather than lower the actual emissions levels, the auto manufacturer inserted software that reported substantially lower emissions levels during testing than were possible when the vehicles were on the road. Before a team from West Virginia University uncovered this deception, these “defeat devices” were installed in 11 million diesel cars sold worldwide between 2008 and 2015. The result was that the fraudulent test results met emission requirements, but the vehicles “spewed as much as 40 times more pollution from tailpipes than allowed by the U.S. Environmental Protection Agency.”1 Since then, scientists at MIT have found that the excess emissions will cause 60 premature deaths across the United States and 1200 in Europe, with Germany, Poland, France, and the Czech Republic being hit the hardest.2
Michael Horn, who was the CEO of Volkswagen Group of America at the time, testified before the House Energy and Commerce Committee’s oversight and investigations panel in October of 2015 about this emissions-test cheating scandal. In response to challenges from lawmakers, Horn said, “This was a couple of software engineers who put this in for whatever reason To my understanding, this was not a corporate decision This was something individuals did ”3 Horn went on to explain that three Volkswagen employees had been suspended as a result of the software that led to the fraudulent test results.
In response to Horn’s testimony, Rep. Chris Collins (R-N.Y.), who is himself an engineer, said, “I cannot accept VW’s portrayal of this as something by a couple of rogue software engineers Suspending three folks it goes way, way higher than that ”4 Collins continued: “Either your entire organization is incompetent when it comes to trying to come up with intellectual property, and I don’t believe that for a second, or they are complicit at the highest levels in a massive cover-up that continues today.”5
Volkswagen’s response to this scandal lies at one end of the spectrum regarding collective responsibility: the blame is entirely the result of a few individual employees, with none attaching to the corporation itself.
At the other end of the spectrum lies a case like this: on March 6, 1984, the United States Department of Defense charged that between 1978 and 1981, the company National Semiconductor had sold them 26 million computer chips that had not been properly tested and then had falsified their records to conceal the fraud.6 These potentially defective chips had been used in airplane guidance systems, nuclear weapons systems, guided missiles, rocket launchers, and other sensitive military equipment. Highlighting the gravity of the situation, a government official noted that if one of these computer chips malfunctioned, “You could have a missile that would end up in Cleveland instead of the intended target.”
Officials at National Semiconductor admitted to both the omission of required tests and the falsification of relevant documentation and agreed to pay $1.75 million in penalties for defrauding the government However, the company refused to provide the names of any of the
individuals who had participated in the decision to omit the tests and falsify the documents, or any who had been involved in carrying out these tasks. The legal counsel for the Department of Defense objected, arguing that “a corporation acts only through its employees and officers” and thus the government would have no assurance that National Semiconductor would not engage in the fraud again. In response, the CEO of National Semiconductor said, “We totally disagree with the Defense Department’s proposal We have repeatedly stated that we accept responsibility as a company and we steadfastly continue to stand by that statement.” A spokesperson for National Semiconductor later reiterated this position: “We will see [that our individual people] are not harmed. We feel it’s a company responsibility, [and this is] a matter of ethics.” National Semiconductor prevailed: no individual employee was ever held criminally or civilly liable for the crime Only the company qua company was penalized
In contrast to Volkswagen’s approach, then, National Semiconductor took complete responsibility for defrauding the government at the level of the company and denied that any individual employees were deserving of blame.
These two different ways of characterizing collective responsibility are closely connected to a central debate in the literature on the epistemology of groups. On the one hand, deflationary theorists hold that group phenomena, such as group beliefs, can be understood entirely in terms of individual members and their states. According to this approach, the states of collectives are not interestingly different than those of individual knowers, and thus collective epistemology turns out to be largely or completely reducible to individual epistemology. Inflationary theorists, on the other hand, hold that group phenomena are importantly over and above, or otherwise distinct from, individual members and their states In this way, groups are often said to crucially have “minds of their own ”7 Settling some of the issues in this debate lies at the heart of making sense of attributions of collective responsibility. If National Semiconductor, for instance, is to be treated as an entity over and above its members in bearing responsibility for defrauding the government, then it is essential to determine whether the company believed or knew that required tests were being omitted and relevant documentation falsified and whether it lied to the government about its fraudulent behavior In the absence of plausible readings of these collective states, it simply wouldn’t make sense to say that National Semiconductor as a corporation bears full responsibility for its actions
A central aim of this book is to make progress in understanding these crucial notions in collective epistemology group belief, justified group belief, group knowledge, group assertion, and group lies so as to shed light on whether it is groups, their individual members, or both who ought to be held responsible for collective actions.
0.1 On the Very Existence of Group Beliefs
One fundamental type of objection I have frequently heard to a project on the epistemology of groups goes like this: groups are not the proper bearers of epistemic states. Being a knower, or justifiedly believing a proposition, requires belief. Belief is a mental state, and groups don’t have mental states in any robust sense. To the extent that we attribute states of belief or knowledge to groups, such talk is loose or metaphorical Anthony Quinton, for instance, writes:
We do, of course, speak freely of the mental properties and acts of a group in the way we do of individual people Groups are said to have beliefs, emotions, and attitudes and to take decisions and make promises But these ways of
speaking are plainly metaphorical To ascribe mental predicates to a group is always an indirect way of ascribing such predicates to its members With such mental states as beliefs and attitudes, the ascriptions are of what I have called a summative kind To say that the industrial working class is determined to resist anti-trade union laws is to say that all or most industrial workers are so minded (Quinton 1975/1976, p 17)
There are two different views of group belief suggested in this passage On the one hand, there is the eliminativist view, according to which it is literally false that groups believe things and hence group belief attributions are simply metaphorical On the other hand, there is the deflationary view mentioned above, according to which it is literally true that groups believe things, but such claims are made true entirely by individual members of the groups believing things. On either reading, we might say that a book on the epistemology of groups is misguided. Groups are not epistemic agents in their own right because they don’t have proper beliefs and, thus, they don’t have justified beliefs or knowledge Talk of group beliefs is either metaphorical or fully reducible to the beliefs of individuals.
Since any reader of this book who is sympathetic with this line of thought will most likely see no point in moving forward, let me offer a very brief argument right at the outset for rejecting it.
I take as a starting position that groups lie. I will say more about this throughout the book, especially in Chapters 1 and 5, but this does not seem particularly contentious. Google “Facebook lies” and a litany of articles comes up. Corporations have been forced to pay literally billions of dollars for lying. After a 2009 headline in Business Insider that reads “Pfizer to Pay $2.3 Billion in Biggest Fine Ever for Deceitful Drug Marketing,” the first line says, “There is not always truth in advertising, but when you really lie, you really pay especially if you happen to be an enormous drug company.”8 And on a more personal level, imagine that your employer promised to give you research funds while recruiting you, but you then learned after they never materialized that the university said this to you while knowing full well that they did not have the resources to follow through. It seems quite natural for you to say, not at all metaphorically, “My university lied to me.”
With this in mind, here is an argument:
1. Groups lie.
2. Group lies cannot be understood without groups having genuine beliefs.
3. Therefore, groups have genuine beliefs.
I will discuss group lies in far more detail in Chapters 1 and 5. But very briefly, a lie whether offered by an individual or a group just is an assertion that one does not believe oneself that is made with the intention to be deceptive. Indeed, even within debates about the details over how to understand lying, there is consensus that it crucially involves the absence of belief on the part of the liar. Premise (1) is thus widely supported by our social practices, including our notions of moral and criminal responsibility, and (2) follows from every major account of the nature of lying Finally, all I mean by “genuine” is that such beliefs are neither loose talk nor entirely reducible to the beliefs of individuals. Rather, there is a robust sense in which groups are believers in their own right
Of course, I don’t expect this argument to be fully satisfying at this point. But what I hope it does is get the skeptical reader on board with thinking that a project on the epistemology of
groups is worth exploring.
0.2 The Nature of Groups
Another initial question that might be raised about a project on the epistemology of groups is what kind of collective entities will be at issue. Groups obviously come in a variety of forms. At one end of the spectrum, there are highly structured groups with policies, procedures, and robust forms of interaction among the members, such as corporations, universities, juries, and boards; at the other end, there are collections of individuals with no formal structure or interaction among group members, such as left-handed Northwestern students and red-haired New Yorkers And in between these two ends are groups with varying degrees of structure and interaction, such as governments, scientific communities, Americans, and women.
Accounts of group phenomena are often directly shaped by which groups are taken to be paradigmatic. For instance, those who are drawn to restrictive views typically focus on highly structured groups with regular interaction among the members. A particularly clear example of this can be seen in the work of Frederick F. Schmitt who, following Margaret Gilbert, argues that “a set of individuals forms a group just in case the members of the set each openly expresses his or her willingness to act jointly with the other members of the set” (Schmitt 1994, p. 260). This conception of a group requires a high level of conscious interaction among the members, and thus rules out as groups all but very formalized collections It is, therefore, not surprising that Schmitt frequently uses a jury as a classic example of a group. Moreover, this starting point directly impacts his view of other collective phenomena, such as justified group belief, where he relies crucially on the notion of joint acceptance. Clearly, there is a sense in which many groups, such as the Democratic Party or even Northwestern University, are not even properly positioned for joint acceptance, as they are large, dispersed, and made up of members with varying levels of authority 9
In contrast, those who are interested in more permissive accounts of collective phenomena typically focus on large, unstructured groups, such as those that have information distributed throughout their members. Alexander Bird, for instance, takes the scientific community to be a paradigmatic group in his argument on behalf of a phenomenon that he calls “social knowing,” where group states do not even supervene on the mental states of the individual members.10 Similarly, Søren Harnow Klausen recently argued that:
We should allow that the factors which, together with truth (or, in the case of knowing how, some sort of adequacy to the task in question), are necessary and jointly sufficient for group knowledge, can be distributed among the members of the group A well-known example of genuinely distributed cognition has been provided by Hutchins (1995), who describes how a navy vessel crew is able to navigate successfully through the concerted efforts of many individuals, each of whom carries out a very specialized task and does not necessarily have any knowledge of the contributions of others, nor of the more general tasks or the ways in which the different contributions are merged. I suggest that this kind of example, rather than that of a jury or a board of directors facing a specific decision, should serve as a paradigm of collective knowledge (Klausen 2015, p 823)
Here, Klausen begins with a paradigm of a group that not only fails to engage in any sort of collective deliberation, as a jury does, but is also such that the members are wholly unaware both of the actions of the others and of the collective goals. If this is the starting point, then any account that requires joint awareness or activity of any sort will immediately be ruled out.
Attempts have been made in the literature to distinguish these different kinds of groups in ways that have theoretical or practical significance. One common distinction that is drawn here is between established groups and non-established groups, where it is standard to rely on examples of each kind rather than definitions or criteria. Margaret Gilbert, for instance, writes:
There are ascriptions of cognitive states to two or more people who are understood to constitute an established group of a specific kind such as a union, court, discussion group, family, and so on There are also ascriptions of cognitive states to two or more people without any presumption that they constitute an already established group (Gilbert 2004, p 96)11
While there are certainly important differences between, say, a court and a collection of lefthanded Northwestern students, it is doubtful that “being established” is the most theoretically or practically relevant feature to focus on. Suppose, for instance, that unbeknownst to the lefthanded students at Northwestern, I fill out all of the necessary paper work for them to have official status as a club on campus. Left-handed Northwestern students are now recognized by the university and thus constitute an established group. Surely, however, this official status by itself does not change the group in any deeply significant way. So, the distinction between being established and non-established doesn’t seem to mark a particularly meaningful difference between these two kinds of groups.
Christian List and Philip Pettit (2011) have argued that while a corporation is a group, lefthanded Northwestern students are a mere collection, and this difference is grounded in whether the collection in question can survive changes of membership. They write:
Collections of individuals come in many forms. Some change identity with any change of membership. An example is the collection of people in a given room or subway carriage Other collections have an identity that can survive changes of membership Examples are the collections of people constituting a nation, a university or a purposive organization
We call the former “mere collections”, the latter “groups” Our focus here is on groups (List and Pettit 2011, p 44)
Once again, however, it is not clear that an important distinction has been highlighted. On the one hand, a company may not survive the firing of its CEO, a cult may not survive the death of its leader, and a political group may not survive its dictator being overthrown, but there is nonetheless a clear difference between these groups and left-handed Northwestern students. On the other hand, a collection of individuals trying to save a beached whale over the course of 36 hours12 may survive a number of changes in membership,13 yet there may be significant asymmetries between this type of group and a medical association or academic department. So, it seems as though there can be paradigmatic groups unable to survive changes in membership, and classic instances of mere collections that are able to do so. Thus, the ability to survive changes in membership also fails to track a theoretically substantive difference among collective entities.
One distinction that has clear significance along a number of dimensions such as epistemic, moral, legal, and practical is the group’s ability, or lack thereof, to engage in collective deliberation or reasoning. Such reasoning involves, at a minimum, a sensitivity to evidence, the capacity to engage in belief revision, and being the proper subject of normative evaluation For instance, school boards often meet to discuss issues relevant to their goals, bouncing ideas off of one another, considering and responding to objections as a group, weighing various pieces of evidence, and revising their collective beliefs as necessary. In contrast, in the absence of unusual circumstances, left-handed Northwestern students simply do not engage in these sorts of activities for several reasons. First, collections of individuals such as left-handed Northwestern
students often simply don’t conceive of themselves as a group and so wouldn’t consider engaging in any sort of reasoning with other members. Second, these sorts of collections are sometimes widely dispersed across even larger groups of individuals, rendering it practically impossible to collectively reason, especially when there aren’t any formal meetings or events at which they can do so. Finally, these groups of individuals are frequently united by features, such as left-handedness, that have very little general interest or value
For ease of expression, let us call those groups capable of engaging in collective reasoning deliberative groups and those that are not non-deliberative groups.
There are two points about non-deliberative groups that should be emphasized here. First, that this sort of group is incapable of engaging in collective reasoning is simply a contingent feature that arises in the circumstances in which it is found, one that can certainly change from one minute to the next. For instance, Northwestern administrators may request that all lefthanded students gather in an auditorium to discuss whether the campus is suitably sensitive to their needs. Prior to this gathering, the group was incapable of engaging in collective reasoning and was therefore non-deliberative But as soon as they find themselves in an auditorium identified as sharing a certain salient property, they might then reason together as a group and thus become deliberative For instance, each student may now share with the others evidence about how many classrooms do not have desks for left-handed students, how many times each student has been inconvenienced because of his or her left-handedness, and so on, and they may decide as a group what their position is regarding whether the campus is appropriately set-up for their needs. So, being non-deliberative is not necessarily an enduring property of a group.14
Second, despite the fact that non-deliberative groups cannot engage in collective reasoning, there is a clear sense in which they can nonetheless have group beliefs, albeit ones that are different from those had by their deliberative cousins For instance, suppose that I am in charge of safety at Northwestern and I wish to determine whether left-handed students regard the campus as properly designed for their needs. To this end, I send out a survey for all left-handed students to fill out and, after receiving the results, I aggregate their judgments via a supermajority procedure. I then report on this basis that left-handed Northwestern students believe that the campus is not suitably sensitive to their particular needs It is not uncommon to think that there is nothing strained or mistaken about this belief attribution as a group, left-handed Northwestern students do hold this belief, just not in the same way that a group capable of collective reasoning might do so.
While deliberative groups depend for their existence on features such as an appropriate structure, a set of constitutive rules, accepted social integration, and so on, non-deliberative groups can simply be brought into existence through internal or external interest. For instance, someone who is either herself left-handed or is simply interested in those who are may be inspired to survey Northwestern students with this feature, aggregate their responses, and reveal their belief as a non-deliberative group. The same can be said for other collections of individuals, such as red-haired New Yorkers This interest brings the group into existence, and the surveying and aggregating reveals their group belief.
Unlike the features discussed earlier such as whether a group is established or can survive changes in membership the ability to engage in collective reasoning has clear significance. For instance, a collection that is capable of weighing evidence and revising its beliefs as a group can be evaluated as rational and irrational in a way that one that is not so capable cannot. Sure, after conducting the survey of left-handed Northwestern students, the administration may have evidence that these students have beliefs that collectively conflict with one another, and they may
assess them on this basis But this is an assessment of the beliefs of a collection of individuals rather than of a collective entity. Corresponding to the evaluation of a group being appropriately judged irrational is the responsibility that such an entity might collectively bear for this shortcoming. If a school board is irrational in its belief that the honors program at a local high school ought to be discontinued, then it can be held responsible as a group for this epistemic shortcoming Relatedly, the school board can then act as a collective agent by deciding to discontinue the honors program and, accordingly, can be held morally and legally responsible for some of the effects that this move has on the community. For instance, the school board can be appropriately deemed self-serving or callous if it is discovered that the decision was made simply for political benefits. Or the school board can be sued if parents regard the cessation of the honors program as failing to provide their children with the educational opportunities that are required by the state. Thus, there is obvious epistemological, metaphysical, moral, legal, and practical value that comes with the ability to engage in collective reasoning
In addition to deliberative and non-deliberative groups, there are what we may call mere collections, which are sets of individuals that share a common feature though one for which there is no interest, either internal or external. This is currently the status of left-handed Northwestern students since neither the members of this collection nor those outside of it have any interest in left-handed students qua being left handed. As we have seen, this can certainly change. Lefthanded students may organize themselves into a deliberative group by meeting weekly to discuss their plight, formalizing rules for making decisions among themselves, and electing a board to officially represent their interests on campus. Alternatively, a right-handed member of the community worried about issues of fairness may convert left-handed students into a nondeliberative group through her interest in them.
There is, however, a feature that is even more general one that may even cut across the deliberative/non-deliberative distinction15 that captures the kinds of groups that will be the focus of this book; namely, being subject to normative evaluation. In particular, I will be interested in those groups that are properly subject to normative assessment, such as praise and blame, along both epistemic and moral dimensions, and the corresponding attributions of responsibility, accountability, and so on. Put succinctly, if we can properly hold a group, G, responsible for φ-ing, then this is sufficient for regarding G as a group in the sense relevant for this project.
0.3 Chapter Overviews
This book is divided into five chapters. In Chapter 1, I take up the question: how should we understand the sense in which groups have beliefs? In stark contrast to the quote from Anthony Quinton earlier, the received view in collective epistemology is that group belief must be understood in non-summative or inflationary terms. Such views are motivated by cases that purport to show that a group can be said to properly believe that p, despite the fact that none of its individual member believes that p. If this is true, then group belief cannot be understood, even in part, in terms of the beliefs of individual group members This is the negative claim of the non-summative view. The positive claim is that group belief should be characterized in terms of something that the members do, and this is typically identified as joint acceptance. Very roughly, group belief is the result of the members jointly agreeing to accept a given proposition as the group’s, even if no member believes it herself.
In this chapter, I challenge this orthodoxy by raising an entirely new objection to this general approach to understanding group belief. I show that joint acceptance accounts crucially lack the resources to be able to explain how groups can lie and bullshit, and, more generally, I argue that group belief cannot be determined by states or processes that are under the direct voluntary control of the members. I also show that such non-summative views countenance as group beliefs states that are riddled with incoherence among the bases of the beliefs of the group members. This leaves group belief without an appropriate mind-to-world direction of fit and renders it unsuitable for proper epistemic evaluation and collective deliberation. In addition, I show that the original cases used to motivate non-summative views can be fully explained without the need to posit group belief.
I then go on to develop and defend a new view, which I call the Group Agent Account: group belief is determined in part by relations among the bases of the beliefs of members, where these relations arise only at the collective level, and are crucial especially insofar as the group is able to function as an agent. At the same time, group belief is partly constituted by the individual beliefs of members. In this way, the resulting view is neither strictly summative nor nonsummative.
In Chapter 2, I turn to the question of justified group belief, which has received surprisingly little attention in the literature. Mirroring the debate regarding group belief, there are those who favor an inflationary approach, where the justificatory status of group belief involves only actions or features that take place at the group level, such as the joint acceptance of reasons. On the other hand, there are those who endorse a deflationary approach, where justified group belief is understood as nothing more than the aggregation of the justified beliefs of the group’s members.
In this chapter, I raise new objections to both of these approaches. Against inflationary views, I show that they face what I call the Illegitimate Manipulation of Evidence Problem, according to which accounts that allow the justification of group beliefs to be achieved through wholly voluntary means also permit the evidence available to the group to be illegitimately manipulated, thereby severing the connection between group epistemic justification and truthconduciveness. Against deflationary views, I argue, that they lead to the Group Justification Paradox in which a group ends up counting as justifiedly believing both that p and that not-p
Finally, I develop and defend a positive view of justified group belief that parallels my account of group belief in Chapter 1 in critical respects, which I call the Group Epistemic Agent Account: groups are understood as epistemic agents in their own right, ones that have evidential and normative constraints that arise only at the group level, such as a sensitivity to the relations among the evidence possessed by group members and the epistemic obligations that arise via membership in the group. These constraints bear significantly on whether groups have justified belief. At the same time, however, group justifiedness on the Group Epistemic Agent Account is still largely a matter of member justifiedness, where the latter is understood as involving both beliefs and their bases. The result is a view that neither inflates nor deflates group epistemology, but instead recognizes that a group’s justified beliefs are constrained by, but are not ultimately reducible to, members’ justified beliefs.
In Chapter 3, I take up two quite influential kinds of purported group knowledge that are inflationary and non-summative in nature, and that pose direct challenges to the account of justified group belief developed in Chapter 2. The first, developed and defended in most detail by Alexander Bird, is often referred to as “social knowledge.” A paradigmatic instance of social knowing is taken to be the so-called knowledge possessed by the scientific community, where no
single individual knows a given proposition, but the information plays a particular functional role in the community. The second is “collective knowledge,” which occupies an important place in United States law. According to the “collective knowledge doctrine,” knowledge may be imputed to a group by aggregating bits of information had by its individual members. If these are correct, then my view of justified group belief is false, particularly the requirement that some of the members of a group have the relevant justified beliefs themselves However, I show in this chapter that both social knowledge and collective knowledge sever the crucial connection between knowledge and action, and open the door to serious abuses, not only epistemically, but morally and legally as well. Bits of information that are merely accessible to group members, or individual instances of knowledge that are aggregated with no communication, do not amount to group knowledge in any robust sense I conclude, then, that neither social knowledge nor collective knowledge is genuinely group knowledge, and thus neither poses a challenge to my Group Epistemic Agent Account
I turn, in Chapter 4, to understanding what it means for a group to assert a proposition. It is especially crucial to get a grip on this notion for being in a position to hold collectives, such as corporations, morally and legally responsible for what they say. I begin by distinguishing between two kinds of group assertion coordinated and authority-based and I argue that authority-based group assertion is the core notion. I then show that a deflationary view of group assertion, according to which a group’s asserting is understood in terms of individual assertions, is misguided. This is the case because a group can clearly assert a proposition even when no individual does. I then develop a positive inflationary view of group assertion according to which it is the group itself that is the asserter, even though this standardly occurs through a spokesperson(s) or other proxy agent(s) having the authority to speak on behalf of the group. This is supported by the fact that paradigmatic features of assertion apply only at the level of the group. A central virtue of my account is that it provides the framework for distinguishing when responsibility for an assertion lies at the collective level and when it should be shouldered by an individual simply speaking for herself.
In the final chapter of this book, I take up group lies. Despite the prevalence of group lies and their often far-reaching effects, such as those seen in both the Volkswagen and National Semiconductor cases, there has never before been a philosophical treatment of group lies 16 This chapter begins the process of filling this surprising gap in the literature by focusing on the question of what a group lie is. After providing an account of how to understand individual lies, I consider, first, whether group lies can be understood in terms of the lies of the group’s members and, second, whether group lies can be characterized in terms of joint agreement by the group’s members to lie After showing both views to be misguided, I offer my own account of group lying, according to which it crucially involves the group offering a statement. In particular, because what a group says can come apart from what its individual members say, I argue that a group might lie when no individual member lies, and a group might fail to lie even though every individual member does. A central virtue of my account is that it captures the often subtle and complex relationship that can exist between most groups and their spokespersons. In this way, my view provides the basis for understanding how groups are responsible for their lies, as well as for determining when it is appropriate to trace this responsibility to the individual members of the group and the spokespersons who represent them.
0.4 The Bigger Picture
Let’s return to the two cases we discussed at the start of this Introduction. Straightforward deflationary views of collective phenomena have the resources for holding only individual members of groups responsible for their actions. After all, according to such accounts, only individuals believe, know, assert, lie and so on, and so there quite literally are no states or actions of collective entities to bear attributions of praise and blame. Such a framework accords very well with the way the Volkswagen of America CEO tried handling the emissions-test cheating scandal. There were a few rogue software engineers on this reading, with no responsibility attaching to the corporation itself, and so the suspension of these employees fully handled the matter. But as Rep. Chris Collins (R-N.Y.) and other lawmakers made clear, this response is not only deeply unsatisfying, it is also wildly implausible. Deception of this magnitude undoubtedly involves some degree of culpability at the level of the corporation itself In particular, installing a device explicitly designed to produce fraudulent test results in 11 million diesel cars sold worldwide cannot occur without complicity or, at the very least, negligence at the highest levels of leadership and oversight. Thus, purely deflationary views of collective epistemic phenomena lack the resources both for providing the correct diagnosis in a case such as this and for holding the relevant parties accountable.
Purely inflationary views of collective phenomena, on the other hand, have the resources for holding only groups responsible for their actions, allowing individual members to go entirely or largely blameless For, on such views, activity that takes place at the collective level determines whether groups believe, know, assert, lie and so on, and so there quite literally are no individual states or actions of this kind that are part of the analysis to shoulder praise and blame. This framework very nicely captures the way that National Semiconductor approached its omission of required tests and the falsification of relevant documentation. The company took full responsibility for defrauding the government and agreed to pay $1.75 million in penalties, but vehemently denied that any blame belonged to individual employees. As the legal counsel for the Department of Defense made clear, however, corporations act only through their employees and officers, and so there is no way that this level of illegal activity occurred without knowledge and involvement on the part of individuals. Hence, strictly inflationary views of collective epistemic phenomena lack the resources both for providing a diagnosis that includes culpable activity on the part of individuals and for holding all of the relevant parties accountable.
To be sure, group members might be held accountable for jointly accepting the proposition in question on an inflationary view. But there are at least two ways in which this is inadequate. First, this would distribute responsibility equally, as there is no interesting sense in which some might have jointly accepted more than others. In actual cases, however, it is clear that members often have significantly different roles in group actions and, accordingly, bear different degrees of praise and blame Second, it is important to have a framework that can account for the responsibility members shoulder in collective action that go beyond joint acceptance. Groups often have complicated structures, where those at the highest levels of leadership may avoid joint acceptance altogether but might nonetheless be involved in collective action through more nuanced and harmful ways.
The views that I develop and defend in this book avoid all of these pitfalls. Because I argue that group belief and group justified belief involve both the beliefs and epistemic statuses of individual members, and also relations and normative requirements that hold only at the level of the collective, my views are neither purely inflationary nor deflationary. Rather, I provide a framework for distributing responsibility across groups and their individual members. I regard this as a central virtue of this book. Neither Volkswagen nor National Semiconductor got things
right in the attributions of responsibility, and this led to significant pushback and dissatisfaction among those impacted by their responses. We need a framework for fully understanding accountability in such cases, and my views provide this.
Moreover, while my views of group assertion and group lies are robustly inflationary, they also have the resources for the proper distribution of responsibility between collectives and their individual members. For example, spokespersons have the authority to speak on behalf of groups, and when they assert or lie in their official capacity as a representative of the group, it is the group itself that asserts or lies. Similar considerations apply to collective action more broadly that is performed through a proxy agent. It is then, quite straightforward, how responsibility attaches to groups on my view. But this relationship between groups and spokespersons provides the resources for also holding individual members accountable. Spokespersons are often chosen by members of the group who bear responsibility for not only ensuring that what is conveyed on its behalf is accurate and well-supported, but also for keeping the spokespersons properly informed. When things go awry along any of these dimensions, there are often particular individuals clearly deserving of blame. In addition, spokespersons are frequently (though crucially not always) members of the groups they represent. The Chair of my Department, for instance, is both a member of our group and our spokesperson when it comes to conversations with the Dean about hiring decisions. We can certainly imagine situations in which my Department is on the hook for a group decision we made but where the Chair, as our spokesperson, deserves greater, or a different sort of, blame. Perhaps she withheld crucial information in our decision-making or inaccurately conveyed our position in conversation with the Dean. On my view, the Department would shoulder the responsibility of its assertion conveyed through the spokesperson, but the Chair could still be uniquely blamed for her role in the process.
One final issue that is worth addressing at the outset is this: I argue in the first three chapters of this book on behalf of views of collective phenomena, such as group belief and justified group belief, that include as an epistemic anchor, so to speak, the states of individual members. So, for instance, my accounts of group belief and of justified group belief both require that at least some of the individual members of the group instantiate the states in question Yet in the last two chapters, I defend views of both group assertion and group lies that are robustly inflationary. In other words, I show that groups can assert and lie when no member of the groups is even aware of the proposition in question. What explains this asymmetry in a way that is not ad hoc?
Robustly inflationary views should be adopted only where a group is capable of granting authority to another agent or agent-like entity to do something on its behalf. So, for instance, a group can grant authority to a lawyer to speak on its behalf, to lie on its behalf, to bullshit on its behalf, and to act on its behalf. In all of these cases, then, it will be possible for the group’s actions to be constituted by the actions of another, even when the group itself is entirely ignorant of the matter. Thus, accounts of all of these phenomena will be robustly inflationary in nature. In contrast, a group cannot grant authority to another to believe on its behalf, to desire on its behalf, to justifiedly believe on its behalf, or to know on its behalf. Accordingly, accounts of all of these phenomena will require that at least some of the individual members of the group instantiate the states in question
The Epistemology of Groups Jennifer Lackey, Oxford University Press (2021) © Jennifer Lackey DOI: 10 1093/oso/9780199656608 003 0001
1 https://www chicagotribune com/news/sns-bc-us--volkswagen-emissions-scheme-20150921-story html, accessed August 8, 2019.
2 http://news mit edu/2017/volkswagen-emissions-premature-deaths-europe-0303, accessed August 8, 2019
3 https://www.latimes.com/business/autos/la-fi-hy-vw-hearing-20151009-story.html, accessed August 7, 2019.
4 https://www latimes com/business/autos/la-fi-hy-vw-hearing-20151009-story html, accessed August 7, 2019
5 https://www.washingtonpost.com/news/the-switch/wp/2015/10/07/volkswagens-pulling-the-plug-on-its-2016-americandiesel-cars/, accessed August 7, 2019
6 This case, including all of the quotations, are discussed in Velasquez (2003).
7 See Pettit (2003)
8 https://www.businessinsider.com/pfizer-to-pay-23-billion-in-biggest-fine-every-for-deceitful-advertising-2009-9, accessed August 7, 2019.
9 Others distinguish between collections of individuals and group agents, with very strong requirements to qualify as the latter For instance, according to Jesper Kallestrup, “A collective is a group agent only if (i) its individual members intend that the collective act and form attitudes together, i.e. each of these individuals must intend that they together enact the joint performance and come to a group attitude. Moreover, (ii) each must intend to do their part, and (iii) intend to do so because of their belief that others intend to do their bit A different but related set of constraints concerns the office of the collective as fixed by its charter A collective is a group agent only if its (founding) members jointly set up common goals and agree on how to proceed in order to meet them Both the ends and the means, which are carried out for the purpose of achieving them, are captured by the group’s charter, which is sometimes formally enshrined in a system of laws, other times its existence is evidenced by the practice of the group and its members When these two sets of constraints are met, a collection of individuals unites in forming a rational agent in its own right” (Kallestrup 2016) It should be clear that many collectives that act together will fail these conditions, such as a group of strangers who work together to save a drowning swimmer. In at least some sense, this collection surely might properly be regarded as a group agent, despite failing Kallestrup’s demanding conditions.
10 See Bird (2010)
11 See also Lahroodi (2007) and Bird (2010)
12 This is a modified example that List and Pettit (2011) use to illustrate a mere collection
13 For instance, a reward that is given in recognition of the efforts of those saving the beached whale would include all of the people who participated in the rescue at any point This group would, then, be the subject of praise and other kinds of normative assessment in ways that mere collections of individuals are not
14 Of course, being deliberative can also go out of existence.
15 I should note that while I am interested in normative evaluation of both deliberative and non-deliberative groups, as a general rule the normative evaluation of deliberative groups tends to be richer and more significant to our broader understanding of the epistemic, moral, and legal landscape.
16 Lackey (2018b) is the only exception
1.1 Summative and Non-Summative Views of Group Belief
Let’s begin with the traditional summative account, according to which a group’s believing that p can be understood in terms of the individual members of the group believing that p. A conservative version of the summative account (CSA) can be formulated in the following way:
CSA: A group G believes that p if and only if all or most of the members of G believe that p
The CSA correctly subsumes some classic instances of group belief. For instance, it is plausible to characterize the Northwestern community’s belief that its institution is in Illinois in terms of all or at least most of its members holding this belief However, other common examples of group belief do not seem to fare as well on such a view. Suppose, for instance, that the President of the university issues a statement saying that Northwestern believes that the quarter system is not beneficial to students’ academic success. Suppose further, however, that this belief is held by only a small, yet powerful constituency of the Northwestern community, such as the administration, or an appointed committee that oversees this matter It may still be appropriate to attribute the belief that the quarter system is not beneficial to students’ academic success to the Northwestern community, despite the fact that neither all nor most of its members holds this belief. Because of cases such as this, summative accounts are typically formulated more liberally (LSA) as follows:
LSA: A group G believes that p if and only if some of the members of G believe that p
On this version of the summative view, it may be appropriate to attribute belief that p to G even if only one of its members believes that p. For instance, perhaps it is sufficient that the President of the university alone believes that the quarter system is not beneficial to students’ academic success to properly attribute this belief to Northwestern. This may also be true when a CEO holds a belief within a company, a leader holds a belief within her cult, and a dictator holds a belief within her nation.
Even with this modification, however, the summative account of group belief is said to suffer from a debilitating objection. In particular, it is argued that a group can be properly said to believe that p, even when not a single of its members believes that p A classic example of this sort of case is where a group decides to let a view “stand” as that of the group’s, despite the fact that none of its members actually holds the view in question. For instance, consider the following:
PHILOSOPHY DEPARTMENT: The philosophy department at a leading university is deliberating about the final candidate to whom it will extend admission to its graduate program. After hours of discussion, all of the members jointly agree that Jane Smith is the most qualified candidate from the pool of applicants. However, not a single member of the department actually believes this; instead, they all think that Jane Smith is the candidate who is most likely to be approved by the administration.
Here, it is argued that the philosophy department believes that Jane Smith is the most qualified candidate for admission, even though none of the members holds this belief This attribution is
supported by the group’s actions: the group asserts that Jane Smith is the most qualified candidate, it defends this position, it heavily recruits her to join the department, and so on. This is taken to show that individual belief that p on the part of even one of the group members is not necessary for the group’s believing that p 8
There are different ways in which the sort of scenario found in PHILOSOPHY DEPARTMENT can come about A standard route is through compromise If, say, half of the members of a group believe that candidate x is the best, and the other half believe candidate y is, they might compromise and put forward candidate z as their top candidate. Or suppose that one member of a company believes that the appropriate minimum age for employment is 18 and another believes it is 16. The group might adopt the position that it is 17. Another common way for a case such as PHILOSOPHY DEPARTMENT to arise is through the following of externally imposed rules. For instance, a jury might come to the conclusion that a defendant is nnocent because it was instructed to exclude all hearsay evidence, but each individual juror might nonetheless believe that he is guilty. Similarly, an evaluating panel might deliver the verdict that a submitted study is unpublishable because it does not rise to the exceedingly high standards of the journal in question, but each member of the group might personally believe that it is. A further way for the scenario in PHILOSOPHY DEPARTMENT to arise is through pragmatic considerations. This is one way to understand the case above, where the members of the philosophy department put forward Jane Smith as the top candidate because they believe she is the most likely to be approved by the administration Or suppose that a group of political leaders puts forward views, not because any of the individuals believe them, but because collectively they regard these positions as increasing their chances of being voted back into office.
Opponents of the summative account of group belief also hold that a group can be properly said to not believe that p, even when every single one of its members believes that p. Hence, it is argued that individual belief that p on the part of all of the members of a group is not sufficient for the group’s believing that p. Consider the following:
PHILOSOPHY DEPARTMENT2: The same philosophy department that is deliberating about the final candidate to whom it will extend admission to its graduate program is also such that every single one of its members believes that the best red pepper hummus in Chicago can be found at Whole Foods.
Despite the unanimity of individual belief in such a case, it is argued that it is not correct to say that the philosophy department believes that the best red pepper hummus in Chicago can be found at Whole Foods. This is because assessment of red pepper hummus is entirely irrelevant to the goals and purposes of the group.9
These problems have motivated the now widely accepted non-summative account of group belief, according to which a group’s believing that p is irreducible to some or all of its members believing that p. Such a view holds that in some very important sense, the group itself believes that p, where this is understood as over and above, or otherwise distinct from, any individual member believing that p.
There are two central versions of non-summativism. The first and perhaps most widely accepted is what we may call the joint acceptance account (hereafter, JAA), a prominent expression of which is offered by Margaret Gilbert in the following passage:
JAA: A group G believes that p if and only if the members of G jointly accept that p.
The members of G jointly accept that p if and only if it is common knowledge in G that the members of G individually have intentionally and openl…. expressed their willingness jointly to accept that p with the other members of G. (Gilbert 1989, p. 306)10
A key aspect of such an account is that joint acceptance does not require belief on the part of a single member of the group in question. She writes:
It should be understood that: (1) Joint acceptance of a proposition p by a group whose members are X, Y, and Z, does not entail that there is some subset of the set comprising X, Y, and Z such that all the members of that subset individually believe that p. (2) One who participates in joint acceptance of p thereby accepts an obligation to do what he can to bring it about that any joint endeavors…among the members of G be conducted on the assumption that p is true. He is entitled to expect others’ support in bringing this about (3) One does not have to accept an obligation to believe or to try to believe that p. However, (4) if one does believe something that is inconsistent with p, one is required at least not to express that belief baldly. (Gilbert 1989, pp. 306–7)
Thus, according to Gilbert’s non-summative view, so long as a group jointly accepts that p in the way described above, such a group is said to believe that p. 11
On a joint acceptance account of a group’s believing that p, then, it is neither necessary nor sufficient that some of its individual members believe that p. It is not necessary because joint acceptance by the group members does not require individual belief on their part, and it is not sufficient because individual belief by the group members does not involve their joint acceptance of the proposition in question.12 According to this account, then, the philosophy department in the first case above believes that Jane Smith is the most qualified candidate for admission, even though none of the members hold this belief, precisely because they jointly agree to let this position stand as the group’s. Moreover, the philosophy department in the second case does not believe that the best red pepper hummus in Chicago can be found at Whole Foods, even though every single member of the group possesses this belief, because the members never jointly accepted such a claim Thus, such an account delivers the correct intuitive result in both instances 13
There is, however, an immediate problem facing the version of the joint acceptance account proposed by Gilbert: groups are often large, with committees or boards that are appointed to make decisions on behalf of the group as a whole For instance, consider the following:
MEDICAL ASSOCIATION: The Board of Directors of the American Academy of Pediatrics convenes and decides that its official position is that there are significant health benefits to circumcision, which it proceeds to publish in all of its relevant materials.14 Despite this, all of the doctors who are members of the American Academy of Pediatrics recognize that the evidence is inconclusive, and so have some lingering doubts that prevent them from individually holding this belief.
As it stands, the JAA cannot countenance group belief in MEDICAL ASSOCIATION since the members of the American Academy of Pediatrics fail to jointly accept that there are significant health benefits to circumcision. In particular, only a very small percentage of the group’s members namely, the Board of Directors satisfies the requisite joint acceptance condition. This structure
of a collective entity is quite commonplace: groups are often vast, rendering it practically difficult if not impossible to have each member engage in any sort of joint activity. Thus, a smaller, more manageable body is either elected or appointed to represent and make decisions for the larger group. Given that MEDICAL ASSOCIATION is certainly in the spirit of precisely those sorts of cases that the joint acceptance view of group belief was designed to accommodate, the JAA requires modification
Raimo Tuomela proposes a different version of the joint acceptance account that is formulated to avoid exactly the worries found with the JAA. Specifically, he offers the following:
JAA2: G believes that p in the social and normative circumstances C if and only if in C there are operative members Al,…, Am of G in respective positions Pl…, Pm such that: (1’) the agents Al,…, Am, when they are performing their social tasks in their positions Pl…, Pm and due to exercising the relevant authority system of G, (intensionally) jointly accept that p, and because of this exercise of authority system, they ought to continue to accept and positionally believe it;
(2’) there is a mutual belief among the operative members Al,…, Am to the effect that (1’); (3’) because of (1’), the (full-fledged and adequately informed) non-operative members of G tend tacitly to accept or at least ought to accept p, as members of G; and (4’) there is a mutual belief in G to the effect that (3’) (Tuomela 1992, pp 295–6)
Tuomela’s account of group belief is quite similar to Gilbert’s, though crucially he requires only that “operative members” engage in the joint acceptance of the proposition in question. Operative members, according to Tuomela, are those who are responsible for the group belief having the content that it does which, in turn, is determined by the rules and regulations of the group in question. For instance, in the case of a corporation or a large company, the board of directors may be the operative members while the employees who work on the assembly line or in the housekeeping department may be non-operative members. Given this amendment, the JAA2, unlike the JAA, delivers the verdict that the American Academy of Pediatrics believes that there are significant health benefits to circumcision in MEDICAL ASSOCIATION since the Board of Directors is obviously comprised of operative members in the relevant sense.15 In what follows, then, I will take the JAA2 as the paradigmatic joint acceptance account of group belief.16
The second version of non-summativism commonly accepted in the literature is what we might call the premise-based aggregation account (hereafter, PBAA), a central proponent of which is Philip Pettit. Like other non-summativists, Pettit grounds his view in the argument that a group can be properly said to believe that p, even when not a single of its members believes that p. Unlike other views, however, he locates his project within a judgment aggregation framework. “Aggregation procedures are mechanisms a multimember group can use to combine (‘aggregate’) the individual beliefs or judgments held by the group members into collective beliefs or judgments endorsed by the group as a whole” (List 2005, p. 25).17 For instance, a dictatorial procedure, “whereby the collective judgments are always those of some antecedently fixed group member (the ‘dictator’)” (List 2005, p. 28) understands the belief of a group in terms of the beliefs of a single member the dictator. A majority procedure, “whereby a group judges a given proposition to be true whenever a majority of group members judges it to be true,”
