Two rival theories of what justifies a belief quietly shape how people build vaults. Foundationalism says some beliefs are basic and others derive from them; justification flows upward from a solid base. Coherentism says no belief is basic; beliefs justify each other through mutual fit within a web. Most PKM architectures lean one way or the other without ever naming the commitment.
The Regress Problem
Both theories are responses to a puzzle. Every justified belief needs a reason. What justifies that reason? Another reason. What justifies that reason? You face three options: the chain goes on forever (infinitism), it loops back on itself (coherentism), or it stops somewhere (foundationalism). Most epistemologists reject infinitism as psychologically implausible, leaving foundationalism and coherentism as the major rival frameworks.
Foundationalism in Brief
Foundationalism posits basic beliefs — beliefs that are justified without deriving from other beliefs. Traditional foundationalists (Descartes, classical empiricists) claimed these include introspective states ("I seem to see red") and self-evident logical truths. Moderate foundationalists admit more basic beliefs (perception under normal conditions, memory reports) with weaker but still non-inferential justification. All other beliefs derive from the basic ones through inference.
The image is architectural: a building with load-bearing foundations supporting upper-story inferences. The base must be solid; the upper stories are only as reliable as the inference chains connecting them to the base.
Coherentism in Brief
Coherentism denies that any belief is basic. Justification is a property of a system of beliefs, not of any single belief in isolation. A belief is justified if it fits into a coherent web — one whose members are mutually supporting, logically consistent, and explanatorily connected. The more integrated a belief is with the surrounding web, the more justified it is.
The image is a web or a net. No single strand holds the structure; the structure holds because every strand connects to many others. Remove any single strand and the web still holds. But a web disconnected from empirical reality is just a consistent fantasy, which is the classical objection to pure coherentism.
Two Vault Architectures
These two theories correspond closely to two real PKM styles.
Foundationalist vaults are built on clear primary sources. Literature notes contain quotes and summaries tied to specific books, papers, and articles. Permanent notes derive from literature notes, with citations. Maps of Content derive from permanent notes. The chain from any claim to its source is traceable. When someone asks "why do you think this?", the vault can answer. Research workflows, academic knowledge bases, and the classical Zettelkasten lean foundationalist.
Coherentist vaults are dense networks of mutually-linked notes with less emphasis on source-to-claim tracing. A note feels true because it connects to many other notes you also take to be true. Bidirectional linking, emergent structure, and the dictum "connections over categories" all favor coherence. Many Obsidian users, especially those following the Linking Your Thinking tradition, produce coherentist vaults.
Failure Modes
Each architecture has a characteristic failure mode.
Foundationalist vaults drift toward brittleness. The base is only as good as your initial source curation. If a foundational source turns out to be unreliable, everything built on it is compromised. Worse, foundationalist vaults often have weak links between the basic layer (literature notes) and the upper layer (permanent notes) — the derivations are implicit, not explicit. When a source is retracted, the derivatives do not automatically update. See also Epistemic Hygiene.
Coherentist vaults drift toward groupthink with the self. The web becomes self-supporting regardless of whether it connects to reality. Dense linking makes every claim feel justified; high-connectivity notes feel most true. Contradictions within the vault get smoothed over by selective linking. The vault becomes an echo chamber with elaborate internal consistency and no anchoring to primary sources. See also PKM Anti-Patterns.
The Hybrid in Practice
Most serious PKM practitioners run hybrids, often without realizing it. Literature notes and source summaries provide a foundationalist base. Permanent notes and Maps of Content form a coherentist layer on top. Dense bidirectional linking creates web-like structure, but the links into literature notes anchor that web to primary sources. The architecture gets its stability from the base and its generativity from the network.
Making this hybrid explicit is a design choice worth thinking about. Which notes serve as foundation (anchored to sources, treated as relatively stable)? Which serve as webwork (connectors, synthesizers, living theory)? Which are basic in the foundationalist sense (typically nothing, in a well-run vault — even "obvious" claims benefit from sourcing)? See also Atomic Notes, Maps of Content, Networked Thought.
The Modern Debate
Contemporary epistemology has softened both positions. Moderate foundationalism accepts that basic beliefs have only prima facie justification, defeasible by the coherence (or incoherence) of the wider web. Moderate coherentism admits that some beliefs (perceptions, say) enter the web with extra initial weight, not treated as equal to pure inferences. The dichotomy has fuzzier edges than the original debate, but the architectural choice in PKM remains real.
Key Points
- Foundationalism: some beliefs are basic; others derive from them; justification flows upward from a solid base
- Coherentism: no belief is basic; justification is a property of a mutually-supporting web
- Both respond to the regress problem (what justifies the justifier?)
- Foundationalist vaults use literature notes + derivation chains; coherentist vaults use dense bidirectional linking
- Failure modes differ: foundationalist brittleness when sources fail; coherentist echo chambers when the web floats free of reality
- Most serious vaults are hybrids: foundationalist base, coherentist upper layer; making this explicit sharpens the architecture
- Modern epistemology accepts moderate versions of both; the dichotomy has fuzzier edges than in classical debates
Open Questions
- Is there tooling that can visualize a vault's foundationalist anchoring — which permanent notes trace back to literature notes, which do not?
- What is the right balance of foundation-to-web for a solo knowledge worker vs a research team?
- Can graph metrics (centrality, cluster density, source-link ratios) diagnose which mode a vault has drifted into?
References
- Stanford Encyclopedia of Philosophy — "Foundationalist Theories of Epistemic Justification"
- Stanford Encyclopedia of Philosophy — "Coherentist Theories of Epistemic Justification"
- BonJour, L. (1985). The Structure of Empirical Knowledge
- Sellars, W. (1956). "Empiricism and the Philosophy of Mind"