Modern Theories of Justification: Foundationalism, Coherentism, and More

Every belief you hold has a history. You believed something, and you believed it because of something else, and that in turn rested on yet another belief – and so the chain goes. But where does it end? Or does it end at all? This is the central puzzle driving theories of epistemic justification: the question of what ultimately makes a belief rational or warranted. Modern epistemology has produced four influential answers – foundationalism, coherentism, contextualism, and reliabilism – each resolving the problem of the inferential chain in a fundamentally different way.

The problem each theory is trying to solve 🔗

Before diving into the theories themselves, it helps to understand the shared challenge they address. When you justify a belief, you appeal to another belief. That belief, in turn, needs justification, and so on. This chain of reasons creates what philosophers call the epistemic regress problem. As the Internet Encyclopedia of Philosophy explains, the central aim of theories of epistemic justification is to explain how a person’s beliefs come to be justified in a way that resolves this regress. The options are limited: either the chain terminates somewhere, it circles back on itself, or it continues indefinitely. Each major theory takes a different stance on which of these options is defensible.

Foundationalism: where the chain stops 🔗

Foundationalism is the oldest and most intuitive response to the regress. It holds that the chain of justification must eventually terminate in a set of basic beliefs – beliefs that are not themselves inferred from other beliefs but that can support everything else. These foundational beliefs are typically described as self-evident, incorrigible, or directly grounded in experience.

As the Internet Encyclopedia of Philosophy notes, Aristotle himself employed the regress argument to prove that justification requires basic beliefs – beliefs that are not supported by any other beliefs but are capable of supporting further beliefs. Descartes’ famous “I think, therefore I am” is perhaps the most celebrated example: a belief so indubitable that it requires no further support.

Strong and modest foundationalism 🔗

Not all foundationalists agree on how strong these basic beliefs need to be. Strong foundationalism demands that basic beliefs be absolutely certain and infallible – a high bar that very few beliefs can clear. Modest foundationalism, by contrast, allows basic beliefs to be fallible but still epistemically privileged – for example, a perceptual experience that gives you prima facie justification to believe what it seems to present, even without certainty. Rebus Community’s Introduction to Philosophy: Epistemology lists strong foundationalism, modest foundationalism, and a range of related positions as among the most notable approaches in the contemporary literature.

The problem with foundationalism 🔗

The main challenge foundationalism faces is identifying which beliefs genuinely qualify as basic. Critics argue that even seemingly self-evident beliefs – like direct perceptual reports – depend on background assumptions about the reliability of our senses. According to the Stanford Encyclopedia of Philosophy, some epistemologists challenge the idea that any belief is justified entirely independently of other beliefs we hold about the reliability of our perceptual faculties or memory. If basic beliefs secretly depend on other beliefs, foundationalism’s neat structure begins to unravel.

Coherentism: justification through mutual support 🔗

Rather than locating justification in a privileged foundation, coherentism distributes it across an entire web of beliefs. On this view, no single belief has a special epistemic status. Instead, a belief is justified to the extent that it fits coherently within a broader system of beliefs – supporting and being supported by others.

The Stanford Encyclopedia of Philosophy describes the coherentist picture as holistic: rather than thinking of knowledge like Euclidean geometry – with basic axioms and derived theorems – coherentists treat all beliefs as equal members of a “web of belief,” where what justifies any single belief is ultimately the way in which the whole system hangs together. The beliefs mutually support one another, much like the poles in a tepee.

How coherentism works in practice 🔗

Consider scientific belief. Your belief that the Earth orbits the Sun is not justified by a single foundational certainty; it coheres with your beliefs about gravity, planetary motion, astronomical observations, and the reliability of scientific methodology. Each belief reinforces the others. When a new belief fits neatly into this web – when it increases overall coherence – it earns justification from the system as a whole.

The coherentist’s challenge: the isolation objection 🔗

Coherentism faces a serious objection. As the Routledge Encyclopedia of Philosophy points out, coherentists must explain what makes one coherent set of beliefs more acceptable than a rival coherent set. A person could, in principle, hold a perfectly coherent web of beliefs that is entirely disconnected from reality – an elaborate fiction that hangs together internally but says nothing true about the world. Coherentism alone cannot guarantee that coherent beliefs are true beliefs, which raises questions about whether coherence is really sufficient for justification.

Contextualism: justification is context-sensitive 🔗

Contextualism takes a different approach entirely. Rather than asking what the absolute conditions for justified belief are, it argues that the standards for justification – and for knowledge – vary with the context in which beliefs are attributed or evaluated.

According to the Internet Encyclopedia of Philosophy, epistemological contextualism maintains that whether one knows is somehow relative to context, and that certain features of contexts – such as the intentions and presuppositions of conversational participants – shape the standards one must meet for beliefs to count as knowledge. In some contexts, epistemic standards are unusually high; in most everyday contexts, they are comparatively low.

High-stakes and low-stakes contexts 🔗

A practical way to see this: if someone asks whether you “know” the bus will arrive on time, and you’re simply planning an errand, a quick check of the timetable may be enough. But if missing the bus will cause you to miss a critical medical appointment, the standards for “knowing” the bus will arrive rise considerably. The Stanford Encyclopedia of Philosophy notes that contextualists invariably maintain that epistemic standards do in fact vary from context to context – and that this variability explains why we judge in most ordinary situations that we have knowledge, while in philosophical or high-stakes settings we begin to doubt it.

Contextualism as a semantic thesis 🔗

It is important to note, as Wikipedia’s entry on contextualism clarifies, that contextualism in epistemology is primarily a semantic thesis about how the word “knows” functions in English: what varies with context is not the quality of the evidence but the standards relative to which knowledge is being attributed or denied. This distinguishes it from purely internalist theories, since it builds the social and situational setting into the very meaning of knowledge claims. Critics, however, argue this makes contextualism a theory about language rather than a genuine theory about what knowledge or justification consist in.

Reliabilism: justification from the outside in 🔗

Reliabilism shifts the focus away from the internal structure of beliefs entirely. Instead of asking whether beliefs cohere, or whether they rest on a foundation, reliabilism asks: how was this belief formed? If the process that produced the belief is reliable – meaning it tends to generate true beliefs – then the belief is justified, regardless of whether the believer can articulate reasons for it.

The Stanford Encyclopedia of Philosophy traces the modern formulation of reliabilism to Alvin Goldman’s landmark 1979 paper “What Is Justified Belief?”, which is widely regarded as a turning point – Michael Williams called it “the Reliabilist Revolution.” Goldman’s central idea is that the justifiedness of a belief depends on the reliability of the cognitive process that caused it.

Process reliabilism explained 🔗

Goldman’s own entry in the Routledge Encyclopedia of Philosophy puts the core claim directly: reliabilism about justification holds that a belief is justified if and only if it is produced by a reliable psychological process – meaning a process that produces a high proportion of true beliefs. Random guessing, for instance, does not systematically yield truths, so beliefs acquired by guessing are not justified. By contrast, ordinary visual perception is generally reliable, so beliefs formed through sight are justified, even if the perceiver cannot explain the underlying cognitive mechanisms.

This makes reliabilism a form of epistemic externalism: what justifies a belief is not something the believer necessarily has access to from the inside. As the Internet Encyclopedia of Philosophy notes, most reliabilists maintain that a belief can be justified even if the believer does not know about or understand the process that makes the belief reliable. The cognitive machinery does the work; the believer does not need to inspect it.

Reliabilism’s advantages and limits 🔗

According to Goldman, one major virtue of reliabilism is its ability to counter scepticism: people do not need infallible or certainty-producing processes to have justified beliefs – only fairly reliable ones. This is a more realistic and pragmatic standard than what classical foundationalism demands. However, reliabilism is not without problems. Critics have raised the “new evil demon” problem: if an evil demon were manipulating our experiences to systematically mislead us, our cognitive processes might seem reliable to us while in fact producing mostly false beliefs. Reliabilism would struggle to say whether such beliefs are justified or not.

Comparing the four theories: where each stands on the regress 🔗

Each theory approaches the regress problem differently. Foundationalism terminates the chain at self-supporting basic beliefs. Coherentism dissolves the linear chain altogether, replacing it with a mutually supporting network where every belief contributes to and draws from every other. Contextualism argues the chain need not be resolved absolutely – what counts as adequate justification depends on the conversational and situational context. Reliabilism sidesteps the internal chain entirely, grounding justification in the external reliability of the processes that produce beliefs.

As the Routledge Encyclopedia of Philosophy observes, most foundationalists and coherentists are internalists – they hold that the conditions determining whether a belief is justified are primarily internal psychological states, such as what beliefs and experiences one has. Reliabilism, by contrast, emerged in the last quarter of the twentieth century as an important externalist alternative, arguing that justification cannot be determined without examining the believer’s relationship to the external world.

Each theory has genuine strengths and genuine vulnerabilities. Foundationalism offers clarity and a definite stopping point, but struggles to identify which beliefs are truly basic. Coherentism captures the interconnected nature of belief systems but cannot guarantee that coherent beliefs track reality. Contextualism explains everyday epistemic practice elegantly but may reduce justification to a matter of conversational standards rather than epistemic substance. Reliabilism grounds justification in the real-world performance of cognitive processes but leaves the believer with no internal access to whether their own beliefs are justified.

What these debates reveal is that justification is not a simple binary property – a belief either has it or it doesn’t. It is a complex, contested, and philosophically rich concept that touches on how we relate to truth, to each other, and to the cognitive processes we often take for granted.

What do you think? If your belief is produced by a reliable cognitive process – but you have no awareness of that process – does it feel genuinely justified to you? And if the standards for justification shift depending on context, can there be any universal account of what it means to truly “know” something?

References

1. https://iep.utm.edu/epi-just/
2. https://iep.utm.edu/foundationalism-in-epistemology/
3. https://press.rebus.community/intro-to-phil-epistemology/chapter/epistemic-justification/
4. https://plato.stanford.edu/entries/justep-foundational/
5. https://plato.stanford.edu/entries/justep-coherence/
6. https://www.rep.routledge.com/articles/thematic/knowledge-concept-of/v-1/sections/foundationalism-and-coherentism
7. https://iep.utm.edu/contextualism-in-epistemology/
8. https://plato.stanford.edu/entries/contextualism-epistemology/
9. https://en.wikipedia.org/wiki/Contextualism
10. https://plato.stanford.edu/entries/reliabilism/
11. https://www.rep.routledge.com/articles/thematic/reliabilism/v-1
12. https://iep.utm.edu/reliabilism/
13. https://en.wikipedia.org/wiki/Reliabilism
14. https://www.rep.routledge.com/articles/thematic/justification-epistemic/v-1/sections/epistemic-and-non-epistemic-justification