Loading field map
What makes a belief justified? This question has driven one of epistemology's most persistent debates. At its core lies the regress problem: if a belief is justified only when supported by another justified belief, we seem to face an infinite chain, a circle, or an arbitrary stopping point. The frameworks that make up justification theory are, in large part, competing answers to this puzzle.
Foundationalism, the oldest systematic framework, proposes a two-tier structure. Some beliefs—basic beliefs—are justified without needing support from other beliefs. They might be self-evident, grounded in sensory experience, or indubitable. All other beliefs derive their justification by being properly connected to these foundations. This approach, visible in Aristotle and later in Descartes, offers a clean solution to the regress: the chain ends with basic beliefs. But foundationalism has always faced a challenge: what makes a basic belief justified? Critics argue that candidates for basic beliefs are either too thin to support much else or themselves require justification, threatening the whole structure.
Pragmatist epistemology, emerging in the late nineteenth century with Charles Sanders Peirce, William James, and John Dewey, rejected the very idea of fixed, indubitable foundations. For pragmatists, justification is tied to practical success and the ongoing process of inquiry. A belief is justified if it works in guiding action and survives testing against experience. This framework does not replace foundationalism so much as shift the goal: instead of seeking a permanent grounding, it treats justification as provisional and fallible. Pragmatist epistemology thus coexists with foundationalism as a fundamentally different approach to the regress problem—one that accepts circularity as long as the circle is wide enough to be useful.
Coherentism, which gained prominence in the early twentieth century, offers a more direct rival to foundationalism. On a coherentist view, no belief is basic. Justification is a property of the entire belief system: a belief is justified to the extent that it coheres with the rest of what one believes. This holistic structure avoids the regress by making justification a matter of mutual support rather than linear dependence. Early coherentists like Brand Blanshard and later Keith Lehrer argued that coherence includes logical consistency, explanatory power, and comprehensiveness. The main objection is that coherence alone seems insufficient to guarantee a connection to truth—a coherent fiction could be perfectly justified on coherentist grounds. This tension between coherence and truth-conduciveness would later shape the internalism-externalism debate.
By the 1960s, epistemology faced a new pressure: the Gettier problem showed that justified true belief could still fall short of knowledge. This forced a reexamination of what justification itself requires. The resulting split between internalism and externalism became a central axis of justification theory. Internalism holds that justification depends solely on factors accessible from the subject's perspective—one's evidence, reasons, or mental states. Externalism, by contrast, allows that factors outside the subject's awareness, such as the reliability of the belief-forming process, can determine justification. This is not a single framework but a deep disagreement that cuts across many later theories.
Reliabilism, developed by Alvin Goldman in the 1970s, is a paradigmatic externalist framework. It answers the Gettier challenge by tying justification directly to truth-conduciveness: a belief is justified if it is produced by a reliable cognitive process—one that yields a high proportion of true beliefs. Reliabilism thus absorbs the externalist insight that justification need not be transparent to the believer. It narrows the focus from the subject's perspective to the objective track record of the process. Critics worry that reliabilism cannot account for cases where a subject has no access to the process's reliability, and that it struggles with new, untested processes.
Evidentialism, articulated by Earl Conee and Richard Feldman in the 1980s, offers a refined internalist alternative. On this view, a belief is justified if it fits the subject's evidence. Evidence includes experiences, memories, and other beliefs, all of which are internal to the subject's perspective. Evidentialism preserves the deontological intuition that we ought to believe in accordance with our evidence, and it directly opposes reliabilism's externalism. The two frameworks remain in living disagreement: reliabilists argue that evidentialism cannot explain why evidence itself must be truth-conducive, while evidentialists counter that reliabilism misses the normative character of justification.
Virtue epistemology, which emerged in the 1980s with Ernest Sosa and later Linda Zagzebski, attempts to integrate the strengths of both internalism and externalism. It shifts the focus from individual beliefs or processes to the intellectual character of the agent. A belief is justified if it results from the exercise of epistemic virtues—stable traits like open-mindedness, intellectual courage, and perceptual carefulness. Virtue epistemology can be seen as a transformation of reliabilism: instead of reliable processes, it emphasizes reliable agents. At the same time, it accommodates internalist concerns by requiring that the agent's virtuous character be reflectively accessible in some way. This framework has become one of the leading approaches today because it offers a unified account of justification, knowledge, and epistemic value.
Social epistemology, which also took shape in the 1980s, challenges the individualistic assumptions shared by all the frameworks so far. Justification, on this view, is not merely a property of individual believers but emerges from social processes—testimony, peer review, collective deliberation, and institutional practices. Alvin Goldman's social epistemology and Helen Longino's critical contextualism both argue that the reliability of belief formation depends on social structures. Social epistemology does not replace earlier frameworks but adds a new layer: it shows that even reliabilism and virtue epistemology must account for the social dimensions of reliability and virtue. It coexists with them, often hybridizing with virtue epistemology to produce a socially extended virtue theory.
Bayesian epistemology, which became prominent in the 1990s, brings mathematical rigor to justification theory. It models degrees of belief as probabilities and defines justification in terms of coherence with the probability calculus and updating via conditionalization. A belief is justified if it is assigned a high probability given the evidence, and the updating process is rational. Bayesianism offers a precise, quantitative alternative to the qualitative debates of earlier frameworks. It narrows the focus to probabilistic coherence and evidential support, but it also raises questions about the interpretation of probability and the role of prior beliefs. Bayesian epistemology is now a leading framework in formal epistemology, often used alongside reliabilism and evidentialism to model specific problems.
Today, no single framework dominates justification theory. The leading approaches—virtue epistemology, social epistemology, and Bayesian epistemology—each address different aspects of justification. They agree that justification is a normative concept tied to truth-conduciveness, but they disagree on the primary locus of that normativity: the agent's character, the social context, or the formal structure of belief. Foundationalism and coherentism remain active, often hybridized with other views (e.g., foundationalist elements in virtue epistemology, coherentist elements in Bayesianism). Reliabilism and evidentialism continue to be defended and refined, with the internalism-externalism debate still unresolved. The field is marked by pluralism: different frameworks are better suited to different domains—perceptual justification, testimonial justification, scientific reasoning, and everyday belief. The central tension that opened this story—the regress problem—has not been definitively solved, but the frameworks that have emerged in response have deepened our understanding of what it means to have reasons for belief.