Groups are said to have many different capacities. They are often described as acting in various ways, such as the U.S. Senate deliberating about a current case, Northwestern saying that it has a new dean of the college, and BP lying about the safety of the dispersants used in cleaning up the oil spill in the Gulf of Mexico. Groups are also characterised as having different kinds of cognitive states, such as PETA believing that factory farms should be abolished and the Bush Administration knowing that Iraq did not have weapons of mass destruction. But how, exactly, should we make sense out of these sorts of claims?
In the philosophical literature, there has been far more attention paid to the action side of groups than to the cognitive side. Of course, group actions can have massive, far-reaching consequences. Groups fight wars, manufacture life-saving medications, and issue death penalty verdicts. And it goes without saying that we want to be able to hold groups responsible for their actions, both morally and legally. So the importance of understanding group action is absolutely clear.
But group action itself depends upon groups having cognitive states, as does holding them fully responsible for such actions. For instance, if BP really did lie when it said that the dispersants used to clean up the oil spill were safe, then this requires that BP in fact believed that they were unsafe. In other words, group lies cannot be understood, nor properly ascribed, in the absence of grasping group belief. And if, for example, the Bush Administration is responsible for illegitimately invading Iraq, then it is crucial that it knew, or at least should have known, that Iraq did not have weapons of mass destruction. A full attribution of responsibility here, then, requires an understanding of group knowledge. Yet surprisingly, there has been very little philosophical work on group belief, and even less on questions more specifically in collective epistemology, such as what group knowledge or group justification is.
The work that has been done on these issues can be broadly divided into two different camps. On the one hand, there are those who favour an inflationary approach, where groups are treated as entities with “minds of their own.” Groups on this model are something more than the mere collection of their members, and group states are distinct from the states of individuals. Such an inflationary approach is typically supported through divergence arguments, which purport to show that a phenomenon can be had at the group level despite the complete absence of it at the individual level. To illustrate this form of argument, let’s focus on a key concept in collective epistemology: group justification, which is the main ingredient that converts true belief into knowledge for groups. According to an inflationary view of group justification, a group can believe a proposition with enough justification for knowledge, i.e. justifiedly believe this proposition, even though no member of the group does. Consider the following:
different evidence: A jury is deliberating about whether the defendant in a murder trial is innocent or guilty. Each member of the jury is privy to evidence that the defendant was seen fleeing the scene of the crime with blood spatter on his clothes, but it is grounded in hearsay that, though reliable, was ruled as inadmissible by the judge. Given only the admissible evidence, the jury as a group justifiedly believes that the defendant is innocent, but not a single juror justifiedly believes this proposition because it is defeated for each of them as individuals by the relevant reliable hearsay evidence.
Juries are governed by special epistemic standards, such as the exclusion of hearsay in a court of law. These special standards prohibit the jury in different evidence from taking into account the evidence about the defendant fleeing the scene of the crime with blood spatter on his clothes. Without this crucial testimony, inflationary theorists argue, the jury as a whole has enough evidence to justifiedly believe the defendant is innocent. But since the jurors qua individuals are not governed by these special standards of evidence, they each believe the defendant is guilty on the basis of the hearsay evidence. Thus, we reach the conclusion of the divergence argument: the jury as a whole justifiedly believes the defendant is innocent despite the fact that no single juror justifiedly holds this belief.
This purported divergence between what happens at the group level and at the level of the individual members leads most philosophers to endorse an inflationary view of group justification, with the most widely accepted being a version of the joint acceptance account. On this view, whether something counts as a reason possessed by a group is determined by its members jointly accepting, or being such that the group would jointly accept, it as the group’s reason. While acceptance does not require belief, it brings with it a commitment to act as if the accepted proposition is true. The epistemic goodness or badness of the group’s reason is then fleshed out in terms of traditional justification-conferring features, such as being produced by a reliable process, being grounded in adequate evidence, and so on.
One of the central virtues of the joint acceptance account is its ability to account for how groups can justifiedly hold beliefs that no single member justifiedly believes. For instance, in different evidence, we can easily explain how the jury justifiedly believes the defendant is innocent without a single juror justifiedly believing this: as a group, the members of the jury accept only the admissible evidence, and thus justifiedly believe the defendant is innocent. However, the individual members not only do not believe that the defendant is innocent, they also don’t have justification for believing it, as the hearsay evidence provides them with counterevidence.
There is, however, a problem with the joint acceptance account that cuts to the heart of the view. Consider the following:
ignoring evidence: Phillip Morris is one of the largest tobacco companies in the world, and each of its members is individually aware of the massive amounts of scientific evidence revealing the links it has with lung cancer. Moreover, each individual member believes that the dangers of smoking give the company a reason to believe that warning labels should be placed on cigarette boxes. However, because of what is at stake financially and legally, none of these members would accept that the dangers of smoking give Phillip Morris a reason to believe that it should put warning labels on cigarette boxes.
Does Phillip Morris have a reason to believe that it should put warning labels on cigarette boxes? Clearly, yes. Every member of this group is aware of the scientific evidence showing the dangers of smoking and believes that warning labels should be put on cigarette boxes. The mere fact that the company illegitimately ignores relevant evidence through dogmatically refusing to jointly accept what is not to its liking should not result in its not having this reason, too. This conclusion is supported by noticing that we would surely hold Phillip Morris responsible for the ill effects caused by smoking precisely because we take it to have a good reason to warn people about the dangers of cigarettes. Yet, according to the joint acceptance account, Phillip Morris does not have a reason to put warning labels on cigarette boxes. Indeed, were the company to do so, it would be acting without a reason.
It is just a small step from here to show that the joint acceptance account also leads to problematic results regarding the epistemic justification of group beliefs. Consider ignoring evidence, again: given that all of the evidence showing that smoking is dangerous is not available to the group because of the members’ refusal to jointly accept it, none of it is part of the justificatory basis of the group’s belief. It is, then, not at all difficult to imagine scenarios in which the remaining evidence leaves the group justifiedly believing that smoking does not pose any health hazards. For instance, the group might have access to some studies that, though reliably conducted, had a very limited sample of subjects, none of whom happened to develop lung cancer despite years of smoking. In this case, Phillip Morris’s “belief” that smoking is not unhealthy would be reliably formed and, given the total evidence available, well-grounded, thereby being epistemically justified. But this result is absurd.
The upshot of these considerations is that joint acceptance cannot ground the justification of group beliefs. In particular, ignoring evidence makes clear that group justification cannot be wholly determined by factors over which the members of the group have direct voluntary control. For it is this voluntary control that enables the members of Phillip Morris to simply decide to not jointly accept what they should. Because of this, joint acceptance can be guided by factors that are utterly disconnected from the truth, such as the economic and legal goals of a company.
Thus, any account of group justification that relies entirely on joint acceptance succumbs to what I call the Illegitimate Manipulation of Evidence Problem: If the justification of group beliefs can be achieved through wholly voluntary means, then the evidence available to the group can be illegitimately manipulated, thereby severing the connection between group epistemic justification and truth-conduciveness. Given that the joint acceptance account clearly faces this problem, we need to look elsewhere for an account of the justification of group beliefs.
In light of this serious problem facing an inflationary approach to understanding the justification of group beliefs, a natural response is to move toward a deflationary one, according to which groups do not have minds of their own and group states are nothing more than collections of individual states. The most widely accepted deflationary view is what has come to be known as summativism, according to which the justification of a group’s belief is understood simply in terms of the justification of the individual members’ beliefs. More precisely, a group’s justifiedly believing a proposition is understood in terms of some or all of the group’s members justifiedly believing that proposition.
Deflationary summativism draws inspiration from a judgement aggregation framework. Aggregation procedures are mechanisms that groups with multiple members can use to aggregate the individual beliefs of group members into collective beliefs. For instance, a majority procedure understands group belief in terms of the beliefs of a majority of its individual members. This framework for aggregating member beliefs into collective ones can then be extended to justification – for instance, group justification can be understood in terms of the majority of the group’s members justifiedly believing a proposition. More generally, the deflationary summativist argues that the greater the proportion of members who justifiedly believe a proposition, the greater the group’s level of justifiedness in believing this proposition.
Despite the intuitive plausibility of this deflationary approach to understanding group justification, there is a significant problem, which can be seen by considering the following:
conflicting bases: G is a group whose members consist of 99 guards at Stateville Correctional Centre, M1 – M99, each of whom justifiedly believes that an escape is being planned by only one of three possible inmates – Alan, Bart, and Chris. Each of the first 33 guards, M1 – M33, justifiedly believes that only inmate Alan is planning the escape (= A). By deduction from A, each of them infers that there is an inmate who is planning such an escape (= T). The remaining 66 guards don’t believe and aren’t justified in believing A. Each of the second 33 guards, M34 – M66, justifiedly believes that only Bart is planning the escape (= B), and infers T from B. The other 66 guards don’t believe and aren’t justified in believing B. The final group of 33 members, M67 – M99, justifiedly believes that only inmate Chris is planning the escape (= C) and infers T from C. The 66 others don’t believe and aren’t justified in believing C. So, 99 members of G justifiedly believe T by deduction from some premise he/she justifiedly believes.
In conflicting bases, 99 out of 99 members of G justifiedly believe that there is an inmate who is planning an escape at Stateville and, thus, the group clearly justifiedly believes this proposition. But let us take a closer look at the details of this sort of case.
Each of the first 33 guards, M1 – M33, justifiedly believes that only inmate Alan is planning the escape. Given this, combined with the fact that all of the guards are aware that Alan, Bart, and Chris are the only possible escapees, each of these 33 guards also justifiedly believes that Bart and Chris are not planning the escape. Each of the second 33 guards, M34 – M66, justifiedly believes that only Bart is planning the escape and, given their other background beliefs, also justifiedly believes that Alan and Chris are not planning it. And similar considerations apply with respect to the final subgroup: each of M67 – M99 justifiedly believes that only Chris is planning the escape and also justifiedly believes that Alan and Bart are not.
We are now in a position to see the following problem unfold: for each of the three possible candidates of the escape in question, 66 out of 99 guards justifiedly believe that he is not planning it. On nearly every judgement aggregation procedure – for example, a majority or supermajority one – it follows from this that the group, G, also justifiedly believes that each of the three possible candidates is not planning the escape. Since the group justifiedly recognises that Alan, Bart, and Chris are the only possible candidates for planning the escape, the group justifiedly believes that no one is planning it. But remember that G also justifiedly believes that someone is planning an escape at Stateville Correctional Centre since 99 members justifiedly believe this. Deflationary summativism thus leads to what I call the Group Justification Paradox: G ends up justifiedly believing both that no one is planning an escape and that someone is planning an escape.
We have seen that the paradigmatic versions of both inflationary and deflationary views of justification suffer from debilitating objections. The joint acceptance account treats groups as epistemic entities that can float freely of the evidential profiles of their individual members. For instance, as ignoring evidence reveals, even if every member of Phillip Morris possesses massive amounts of scientific evidence revealing the links between smoking and lung cancer, such a view permits groups to choose not to accept this evidence and thereby end up justifiedly believing that smoking does not pose a health hazard. But groups cannot pick and choose what evidence is available to them – they are constrained by the evidence possessed by their individual members. This is the central lesson of the Illegitimate Manipulation of Evidence Problem.
In contrast, the aggregative account altogether avoids concerns associated with this problem by securing a close dependence of group justifiedness on member justifiedness. But this is done at the cost of failing to appreciate the distinctive epistemic issues that arise at the group level. For the proponent of the aggregative account, group justifiedness is a simple “justified belief in/justified belief out” matter. Yet we have seen that this model ignores the complexity of justified belief at the group level, particularly the evidential relations that exist between members’ beliefs and the bases, or grounds, for such beliefs.
Thus, the justified beliefs of groups should be treated neither as states that can float freely of the evidence possessed by their individual members, nor as nothing more than the aggregation of the justified beliefs of their members. Instead, groups should be understood as epistemic agents in their own right, though ones whose justified beliefs are constrained by the epistemic statuses of their individual members.
As a start, group justification should be grounded in the justification that at least some of the group’s members have for the proposition in question – perhaps a significant percentage of those members who have decision-making authority in the domain in question. A group’s belief, then, will inherit a strong, positive epistemic status from the members’ justified beliefs in which it is based. These member beliefs will, in turn, be justified by whatever features are required at the individual level, such as that they are produced by reliable processes or grounded in adequate evidence.
In addition, group justification should depend, not only on the justification of the individual members’ beliefs, but also on the bases for those beliefs. This is what was learned from conflicting bases: member justifiedness does not necessarily transmit smoothly to group justifiedness, since the bases of the members’ beliefs might be wildly conflicting. When this happens, such a view faces the Group Justification Paradox. To avoid this problem, it is clear that group justification is possible only when the members’ bases for their individual beliefs can be combined in a coherent belief set.
It should be clear that the inclusion of the first requirement altogether avoids the Illegitimate Manipulation of Evidence Problem afflicting inflationary views. One of the features that make the joint acceptance account susceptible to this objection is that the evidence available to the group can end up being a matter of choice. But if the justification of group beliefs is necessarily a matter of the justification of the beliefs of individual members, and the evidence that is available to individual subjects is not a matter of choice, then there is no worry that epistemic justification for group beliefs can be achieved through the illegitimate manipulation of evidence.
At the same time, it should also be apparent that including the second requirement denies justified group belief in conflicting bases and so avoids the Group Justification Paradox. In particular, adding together the bases of the Stateville guards’ beliefs yields an incoherent set of beliefs. This results in the group failing to justifiedly believe that someone is planning an escape, despite the fact that all 99 members of the group justifiedly believe this. Since it was the attribution of justified belief to the group that generated the problem stemming from conflicting bases in the first place, it simply doesn’t arise once this requirement is in place.
This, of course, is just a start to understanding group justification and, more broadly, when groups have knowledge. But it is an important start. For including these two requirements – individual justification and the coherence of individual bases for belief – handles the problems facing the two dominant approaches to the justification of group beliefs. While inflationary theorists focus entirely on what a group does, i.e. whether its members engage in joint acceptance or not, deflationary ones focus exclusively on what a group has, i.e. whether its members’ beliefs are individually justified. What is needed, however, is a view that incorporates both components. Groups should be understood as epistemic agents in their own right, ones that have evidential constraints that arise only at the group level, but group justifiedness should still significantly depend on member justifiedness. Walking this middle ground between inflating and deflating group epistemology promises to carve out space for both groups, and the individual members that make them up, to shoulder the responsibility of group actions.