Diversity of opinion both presents problems and affords opportunities. Differences of opinion can stand in the way of reaching an agreement within a group on what decisions to take. But at the same time, the fact that the differences in question could derive from access to different information or from the exercise of different judgemental skills means that they present individuals with the opportunity to improve their own opinions. This paper explores the implications for solutions to the former (aggregation) problem of supposing that individuals exploit these opportunities. In particular, it argues that rational individual revision of opinion implies that aggregation problems are unstable in a certain sense and that solving them by exploiting the information embedded in individual opinion has profound implications for the conditions that we should impose on aggregation procedures.

The standard image of how consensus can be achieved is by pooling evidence and reducing if not eliminating disagreements. But rather than just pooling substantive evidence on a certain question, why not also take into account the formal, testimonial evidence provided by the fact that a majority of the group adopt a particular answer? Shouldn't we be reinforced by the discovery that we are on that majority side, and undermined by the discovery that we are not? Shouldn't this be so, in particular, when Condorcet's jury theorem applies? It turns out not. There are serious problems attending any strategy of majoritarian deference.

We know that we can learn much from the reports of multiple competent, independent, unbiased observers. There are also things we can learn from the reports of competent but biased observers. Specifically, when reports go against the grain of an agent’s known biases, we can be relatively confident in the veracity of those reports. Triangulating on the truth via that mechanism requires a multiplicity of observers with distinct biases, each of whose reports might be one-way decisive in that fashion. It also presupposes that all observers share the same fundamental epistemic standards.

It is often assumed that democracies can make good use of the epistemic benefits of diversity among their citizenry, but difficult to show why this is the case. In a deliberative democracy, epistemically relevant diversity has three aspects: the diversity of opinions, values, and perspectives. Deliberative democrats generally argue for an epistemic form of Rawls' difference principle: that good deliberative practice ought to maximize deliberative inputs, whatever they are, so as to benefit all deliberators, including the least effective. The proper maximandum of such a principle is not the pool of reasons, but rather the availability of perspectives. This sort of diversity makes robustness across different perspectives the proper epistemic aim of deliberative processes. Robustness also offers a measure of success for those democratic practices of inquiry based on the deliberation of all citizens.

For multiple reasons, deliberating groups often converge on falsehood rather than truth. Individual errors may be amplified rather than cured. Group members may fall victim to a bad cascade, either informational or reputational. Deliberators may emphasize shared information at the expense of uniquely held information. Finally, group polarization may lead even rational people to unjustified extremism. By contrast, prediction markets often produce accurate results, because they create strong incentives for revelation of privately held knowledge and succeed in aggregating widely dispersed information. The success of prediction markets offers a set of lessons for increasing the likelihood that groups can obtain the information that their members have.

Classical Athens provides a historical case study of effective joint action by a democratic community, at scale, over time, and across a socially and epistemically diverse population. Athens was concerned both with aggregating diverse knowledge for decision-making and with building common knowledge for coordinated joint action. A preserved prosecution speech delivered in an Athenian treason trial reveals how common knowledge was generated by democratic institutions and employed in legal arguments. Common knowledge facilitated effective coordination among citizens through productive alignment cascades. Yet Athenian lawcourt procedures rendered cascading difficult, pointing to an institutionalized awareness that coordination among jurors might violate fairness as equality before the law; such violations threatened democratic values and the equilibrium between elite and non-elite citizens. Common knowledge was promoted by repeated public rituals, public monuments, and public architecture that allowed for the interpresence of large numbers of deliberative and intervisible citizens.

Much liberal theorizing of the past twenty years has been built around a conception of neutrality and an accompanying virtue of reasonableness according to which citizens ought to be able to view public policy debates from a perspective detached from their comprehensive conceptions of the good. The view of “justificatory neutrality” that emerges from this view is discussed and rejected as embodying controversial views about the relationship of individuals to their conceptions of the good. It is shown to be based upon a “protestant” assumption according to which conceptions of the good can be cashed out in terms of propositional beliefs. An alternative conception of reasonableness, grounded in the stable disposition of individuals to prefer social peace over conflict is described. It is argued that it better satisfies the neutralist requirement than do theories of justificatory neutrality.

Although John Stuart Mill places considerable emphasis on three information signalling devices – debate, votes and prices – he remains curiously sceptical about the prospects of institutional or social epistemology. In this paper, I explore Mill's modest scepticism about institutional epistemology and compare and contrast that with the attitudes of liberal theorists such as F. A. Hayek and John Dewey who are much more enthusiastic about the prospects of social epistemology as part of their defences of liberalism. The paper examines the extent to which Hayek and Dewey ignore concerns originally raised by Mill. I conclude that Mill's modest scepticism is reflected in the epistemological abstinence of contemporary liberal philosophers such as John Rawls, and that his elevation of philosophy over democracy remains a challenge to contemporary defenders of the political value of social or institutional epistemology.