Epistemology — PKM Wiki

Every PKM system encodes an implicit theory of knowledge. What counts as a fact, what counts as belief, what deserves a note, what deserves to be linked — these are epistemological choices, even when they happen silently. Making the epistemology explicit is what separates a filing cabinet from a thinking environment.

What Epistemology Is

Epistemology is the branch of philosophy that asks: what is knowledge, how do we acquire it, and what makes a belief justified? The classical definition from Plato's Theaetetus — knowledge as "justified true belief" — has structured 2500 years of debate. Later philosophers added fallibilism (all claims are provisional), coherentism (beliefs justify each other in webs), foundationalism (some beliefs are basic and others derive from them), and reliabilism (beliefs formed by reliable processes count as knowledge). A PKM system quietly commits to positions on all of these.

PKM Is Applied Epistemology

When you capture a quote, you make an epistemological claim: this is worth preserving as information. When you link it to another note, you assert a coherence relation: these ideas support, contradict, or qualify each other. When you evergreen a note into a permanent statement, you commit to it as a stable belief. When you delete or retire a note, you retract it from your knowledge base. A vault is not a neutral archive — it is a running ledger of what you currently treat as true, plausible, or suspect.

Most practitioners ignore this dimension. They capture aggressively, link intuitively, and never revisit. The result is an epistemically flat vault where a Wikipedia skim and a peer-reviewed study carry equal weight because both are just notes with backlinks.

Foundationalism vs Coherentism in Notes

Two dominant PKM architectures reflect two epistemologies. Foundationalist vaults build upward from trusted primary sources — literature notes feed permanent notes, which feed maps of content. Provenance chains are explicit; the base is load-bearing. Coherentist vaults rely on dense bidirectional linking — a note feels true because many other notes reference it, and the network as a whole supports itself. Zettelkasten practice leans coherentist; classical research workflows lean foundationalist. Most real vaults are hybrids, and the tension between the two shows up as friction: the Zettelkasten feels liberating until you realize you can't trace claims back to sources, and the research vault feels rigorous until you realize nothing connects.

Tacit and Explicit Knowledge

Polanyi's distinction — "we know more than we can tell" — is a hard limit on what any PKM system can capture. Skill, intuition, taste, and pattern recognition resist externalization. A vault can hold explicit propositions, quotes, and summaries, but the tacit layer that makes knowledge usable lives only in the knower. This is why notes from someone else's vault rarely feel useful: you inherit their explicit traces without their tacit scaffolding. It is also why rereading your own old notes sometimes fails: the tacit context has decayed faster than the explicit text.

Fallibilism and Evergreen Notes

Popper's fallibilism — the view that all knowledge is conjectural and subject to revision — should shape how evergreen notes are maintained. An evergreen is not a permanent truth; it is a current best conjecture, stable enough to build on but never beyond revision. Without this mindset, evergreen notes calcify into personal dogmas. Epistemically healthy vaults track revision history, flag superseded claims, and treat long-dormant notes as candidates for re-evaluation rather than archaeological relics. Confidence markers, dated claims, and review cadences operationalize fallibilism.

Epistemology Under LLM Mediation

The AI era introduces a new epistemic class: LLM-generated content, which is neither primary observation nor cited testimony but probabilistic synthesis from training data. This content can be confidently wrong in ways that feel authoritative. When LLM output enters a vault, it needs its own epistemic category — not "source," not "my thinking," but "AI synthesis, confidence X, verified against Y." Without explicit marking, AI-laundered claims compound through linking and become indistinguishable from verified knowledge. This is the defining epistemological problem of contemporary PKM.

The Calibration Imperative

A well-calibrated vault tracks not only what you believe but how strongly you believe it and why. Confidence markers ("high / medium / low / speculative"), source-quality tiers (primary / secondary / testimony / folk), and explicit uncertainty statements ("I'm not sure, but...") make the epistemic structure visible. The goal is not certainty but calibration: beliefs weighted in proportion to evidence, refuseable under new information, and traceable to their origins.

Key Points

PKM systems are implicit epistemologies — capture, linking, evergreening, and deletion are epistemic acts
The classical "justified true belief" frame raises operational questions: what justifies a note, what makes it true, what makes it yours
Foundationalist and coherentist architectures both work but produce different vaults and different failure modes
Tacit knowledge is a hard limit — vaults externalize only what can be told, never the full scaffolding that makes knowledge usable
Fallibilism should shape evergreen maintenance: all claims are provisional, revision-worthy, and datable
LLM-generated content needs its own epistemic category to prevent synthesis laundering
Calibration (confidence + provenance) beats certainty as the epistemic ideal for a vault

Open Questions

Can a vault surface its own epistemic structure — which notes are foundational, which are coherentist hubs, which have no provenance?
What is the right granularity for confidence markers: per note, per claim, per sentence?
How should a vault handle belief change over time without erasing the revision trail?
Does dense linking produce genuine coherence or just the feeling of coherence?

References

Plato, Theaetetus — the justified-true-belief tradition
Polanyi, M. (1966). The Tacit Dimension
Popper, K. (1963). Conjectures and Refutations
Gettier, E. (1963). "Is Justified True Belief Knowledge?" — the modern challenge to the classical definition
Stanford Encyclopedia of Philosophy — Epistemology entry (overview of foundationalism, coherentism, reliabilism)