The Epistemology of Groups

Abstract

Jennifer Lackey’s excellent new book is very ambitious, as the title suggests: it covers all of collective epistemology (group justification, group knowledge, group speech acts), but also related issues in ethics (group responsibility), language (group assertion and lying), and mind (group belief).I don’t have the space here to do justice to all the excellent philosophy in this book. I will therefore focus on the topics that I take to be central to collective epistemology—namely, group justification and knowledge.I will also not go into much detailed discussion of Lackey’s attacks on deflationary views, as I take her arguments to flawlessly show that there are very serious problems with conceiving group epistemic states along such lines. Rather, this piece’s focus will be twofold: I will first put forth a worry for Lackey’s preferred account of group justification. Second, I will look at Lackey’s arguments against inflationism, and identify ways for my favorite variety of inflationism to resist the attack.Lackey agrees with inflationism that groups are epistemic agents in their own right, in that the epistemic status of their doxastic attitudes does not fully reduce to the epistemic status of members’ doxastic attitudes. Nevertheless, Lackey thinks, group epistemic status depends in crucial ways on members’ epistemic status:The Group Epistemic Agent Account (GEAA): A group, G, justifiedly believes that p if and only if: (1) A significant percentage of the operative members of G (a) justifiedly believe that p, and (b) are such that adding together the bases of their justified beliefs that p yields a belief set that is coherent. (2) Full disclosure of the evidence relevant to the proposition that p, accompanied by rational deliberation about that evidence among the members of G in accordance with their individual and group epistemic normative requirements, would not result in further evidence that, when added to the bases of G’s members’ beliefs that p, yields a total belief set that fails to make sufficiently probable that p. (96)GEAA has important advantages over its main competitors: unlike inflationism, it does not fully divorce group epistemic status from members’ epistemic status, and thus it does not shoulder the burden to explain independently floating doxastic phenomena. Unlike deflationism, however, it enjoys good extensional adequacy.My main worry for the account concerns the centrality given to the epistemic status of operative members’ doxastic states. For a worry for the sufficiency claim, consider the following case:WITCHES: Almost the entire population of Darmania falsely and unjustifiedly believes witches exist. They have no evidence whatsoever to this effect (indeed, not even testimonial evidence—the topic is taboo). The king and his government, which consists of three members, are the only ones that know that witches do not exist.Intuitively, it seems right to say that Darmanians believe witches exist. GEAA, however, will predict that Darmanians justifiedly believe witches do not exist.1Conversely, consider, against the necessity claim:WITCHES*: Almost the entire population of Darmania knows that witches do not exist. The superstitious king and his government, which consists of three members, are the only ones who believe witches exist, without any justification whatsoever.It seems as though, contra GEAA, it is right to say things like ‘Darmanians know witches do not exist, it’s just their superstitious leadership that believes they do, for no good reason’.If I am right, this seems to suggest that epistemic status of the beliefs of operative members need not be intimately related to group-level epistemic status.According to Inflationism, group knowledge and justification need not depend on any members’ epistemic states, but rather float freely at group level.Lackey takes the main motivation for inflationism to come from ‘divergence cases’ like the following:2DIFFERENT 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.The standard interpretation of this case is that, while the jury justifiedly believes that the defendant is innocent, none of the jurors justifiedly believes this because their justification is defeated by the relevant reliable hearsay evidence—which is legally inadmissible for the jury. According to Lackey, however, we have little reason to think the notion of justification at stake when it comes to the jury’s belief is genuinely epistemic. That is because, Lackey argues, legally admissible evidence does not coincide with epistemically admissible evidence in this case. If so, the jury, just like its members, is not epistemically justified to believe that the defendant is innocent, in virtue of epistemic defeat. The jury, compatibly, is legally justified to believe that the defendant is innocent.I agree with Lackey that there is a group of which the jurors are members and about which it is true that they are epistemically unjustified to believe that the defendant is innocent: it’s the unstructured group ‘the jurors’. Under this description, the group has no epistemic reason to dismiss the defeating testimonial evidence. However, the same does not hold when we consider the same individuals under the description ‘the jury’. And here is why: juries, much like city halls, games, and money, are social entities, social-norm-constituted: what it is to be a jury is to be a collection of individuals that is essentially governed by a set of constitutive norms. Sometimes (often) some of these constitutive norms are epistemic norms. Indeed, it is extremely plausible that juries will (also) be constituted by epistemic norms, given that their main task is to evaluate evidence. The norm forbidding the jury qua jury from taking hearsay into account in their assessment of the evidence is an epistemic norm that constitutively governs what it is to be a jury to begin with. If so, ‘the jury’ will be justified in believing that the defendant is innocent, although each juror, as well as the group ‘the jurors,’ are unjustified to hold this belief.Lackey argues—to my mind, conclusively—that one of the historically most popular versions of inflationism—the Joint Acceptance account—is untenable. I am in full agreement.I will turn, however, to Lackey’s case against my preferred inflationism – the distributive account defended most prominently in (Bird 2010; see also Simion, Kelp, and Carter 2022). Accounts like these license ‘social knowledge’ (hereafter SK)—a paradigmatic instance of which is the knowledge possessed by the scientific community, where no single individual knows a proposition, but the information plays a functional role in the community. On this view, information that p need not be accessed by any of the group members for the group to know that p: that the information is accessible to the group members is enough.Lackey argues that endorsing SK leads to two serious epistemological problems, one pertaining to the connection between knowledge and action, and another having to do with epistemic defeat. I will look at these in turn.First, Lackey argues that there will be cases of SK that do not license group action, which comes in conflict with the intuitive sufficiency of knowledge that p for acting on p (KNA). That is because, according to Lackey, groups can only act through their individual members. But if individuals don’t know that p, they cannot permissibly act on p (by KNA; Lackey gives the example of a scientist not being able to permissibly assert that p on behalf of the scientific community based on mere SK that p). As such, the thought goes, instances of SK are instances in which groups cannot act although they know—thereby coming in conflict with KNA.I am unconvinced by this argument at two junctures: first, it’s not clear that groups can only act through their individual members. For instance, it seems as though groups can endorse some claim in virtue, indeed, of their individual members doing nothing at all, rather just staying silent.Second, note that Lackey’s argument merely shows that groups with SK cannot act through one member (or a subset thereof) that does not know that p: compatibly, acting as a group permissibly may be slightly more complicated than that, in that members only partially contribute to the group action (which Lackey herself allows for). Recall Edwin Hutchins’s (1995) case of the ship safely navigating the way to port: each crew member is responsible for tracking and recording the location of a different landmark, which is then entered into a system that determines the ship’s position and course. The crew as a whole knows, for instance, that they are traveling north at twelve miles per hour, even though no single crew member does. The ship’s behavior as it safely travels into the port is clearly well informed and deliberate, and based on the crew’s knowledge of its whereabouts. Compatibly, the crew cannot act on its SK through any individual member.How about Lackey’s defeat-based case against SK accounts? Lackey considers a case in which, while p is accessible to the scientific community but not accessed, not-p comes to be believed by a significant number of scientists. It would seem as though this is a case of SK defeat. But, Lackey argues, it seems epistemically arbitrary to maintain that the mental states of individual members of the group can contribute negatively to social knowing, but not positively. Second, if the scenario described features a defeater, Lackey thinks we will get the result that there will be considerably less social knowledge than we might have thought.A couple of things by way of response on behalf of the SK inflationist: first, the view merely takes it that group knowledge can be distributed in this way, not that it needs to: indeed, the heart and soul of a functionalist view like this is multiple realizability. If so, the view does not exclude cases in which members’ mental states do contribute positively to the epistemic status of the group mental state; it merely does not require that they do.Second, the defeat prediction need not deliver widely spread skepticism about social knowledge: after all, defeat can and often is partial defeat: the justificatory force of the defeater can bring the justificatory force of the evidence below knowledge threshold, but it need not.Setting aside the three criticisms raised, The Epistemology of Groups, to reiterate, is a fantastic book. It is a model of thorough epistemology, and it synthesizes a comprehensive and sophisticated framework. It’s a must-read for anybody working in social epistemology, but also for philosophers interested in group ethics, the philosophy of science, political philosophy, and economic ethics. I expect it will be the main point of reference in collective epistemology for years to come.