Loading field map
You and a friend, equally intelligent and equally informed, look at the same evidence and reach opposite conclusions. What is the rational response? Should you reduce confidence in your own view, or can you reasonably hold your ground? This puzzle—the peer disagreement problem—has driven a rich and rapidly evolving debate in epistemology since the 1990s. The central question is not merely about stubbornness or open-mindedness; it is about the nature of evidence, the structure of justification, and the limits of rational consensus.
The earliest systematic frameworks in the epistemology of disagreement formed a direct opposition. Conciliationism holds that when you discover that an epistemic peer—someone equally competent and with access to the same evidence—disagrees with you, you should significantly reduce your confidence in your original belief. The disagreement itself functions as higher-order evidence, a reason to think that you may have made a mistake. Conciliationists argue that failing to adjust is epistemically arrogant.
Steadfast Views reject this conclusion. According to steadfast theorists, you may rationally maintain your original belief even after learning of peer disagreement, provided you have independent reasons for that belief. The mere fact of disagreement does not automatically override those reasons. Steadfast views do not deny that disagreement is epistemically relevant; they deny that it carries the decisive weight that conciliationists assign to it. The two frameworks thus disagree about the evidential significance of the disagreement itself.
The Conciliationism–Steadfast debate quickly revealed a deeper disagreement about the nature of evidence. The Uniqueness Thesis states that a given body of evidence supports exactly one rational doxastic attitude (e.g., belief, disbelief, or suspension). If the Uniqueness Thesis is true, then when two peers disagree, at least one of them must be irrational. This conclusion pressures conciliation: the rational response is to suspect that you are the one who has misjudged the evidence.
Permissivism denies the Uniqueness Thesis. It holds that the same evidence can rationally support multiple, incompatible doxastic attitudes. If Permissivism is true, then two peers can both be rational despite disagreeing. This provides a foundation for Steadfast Views: you can maintain your belief because the evidence permits it, even if your peer reaches a different permitted conclusion. The Uniqueness–Permissivism axis is not a side issue; it is the meta-level debate that shapes the entire field. A commitment to Uniqueness strengthens the case for Conciliationism, while Permissivism offers a principled way to resist conciliatory pressure.
As the debate matured, philosophers turned to more precise tools. Bayesian Models of Disagreement apply the formal machinery of Bayesian epistemology to model how rational agents should update their beliefs in light of peer disagreement. These models treat degrees of belief as probabilities and use conditionalization to calculate the impact of learning that a peer holds a different credence. Bayesian approaches test the informal claims of Conciliationism and Steadfast Views by making them mathematically explicit. For example, some Bayesian results suggest that under certain assumptions, peer disagreement should lead to substantial convergence, lending support to Conciliationism. Other models show that if agents have different prior probabilities, disagreement can persist rationally, aligning with Steadfast Views. Bayesian Models thus provide a formal infrastructure for exploring the conditions under which each earlier framework holds.
The Justificationist Approach reorients the discussion in a different way. Instead of focusing primarily on what doxastic attitude to adopt after disagreement, it asks about the justification of the original belief itself. The key move is to examine whether the belief was properly justified before the disagreement arose. If it was, then the mere fact of disagreement may not undermine that justification; if it was not, then disagreement reveals a pre-existing flaw. This approach shifts attention from the higher-order evidence of disagreement to the first-order justification of the belief, offering a way to adjudicate between Conciliationism and Steadfast Views by looking at the epistemic pedigree of the contested claim. It does not replace the earlier frameworks but adds a new dimension to the analysis.
Today, all six frameworks remain active and in live disagreement. Conciliationism and Steadfast Views continue to be the two poles, but few philosophers hold either in its purest form. Most positions are nuanced: some defend a moderate conciliationism that applies only in certain conditions, while others defend a steadfast view that allows for occasional revision. The Uniqueness Thesis and Permissivism are now recognized as the deeper battleground; much of the recent literature focuses on whether evidence is permissive or unique. Bayesian Models have become a standard tool for testing and refining arguments, though they are not universally accepted as settling the debate. The Justificationist Approach has gained traction as a way to integrate the epistemology of disagreement with broader theories of justification.
What do the leading frameworks agree on? Nearly all accept that peer disagreement is epistemically significant—it cannot be dismissed as irrelevant. They also agree that the rationality of a response depends on the details of the case, including the nature of the evidence and the relationship between the peers. Where they disagree is on the weight of the disagreement itself: does it function as a defeater that requires belief revision, or is it merely a reason to re-examine one's evidence? This disagreement remains unresolved, and it connects to neighboring subfields such as the epistemology of testimony (how much should we trust others' judgments?) and epistemic injustice (how do power imbalances affect peer status?). The epistemology of disagreement thus continues to be a vibrant area where fundamental questions about evidence, rationality, and social epistemology converge.