Epistemic Status Labeling¶
Communication standard — instantiates Knowledge-Warrant Audit
A language convention that marks claims as observed fact, inference, assumption, hypothesis, expert judgment, or unresolved unknown.
Epistemic Status Labeling is a shared vocabulary that travels with the claim into whatever the audience reads, so a downstream reader can tell — without re-running the audit — whether a sentence is an observed fact, an inference, an assumption, a hypothesis, an expert judgment, or an unresolved unknown. Its defining move is that it is a communication convention, not an analytic one: it does not decide how strong a warrant is or whether confidence is justified; it standardizes how the already-determined status is marked in the prose so the distinction survives being copied into a slide, forwarded in an email, or quoted six months later. Where an audit's judgments normally evaporate the moment the meeting ends, a labeling standard freezes them into the language itself.
Example¶
A company's analytics team publishes a weekly business review that leadership reads fast and acts on. Historically every line arrived in the same flat declarative voice — "churn is up because of the pricing change" sat in the same font as "revenue was $4.2M" (illustrative), and readers couldn't tell the measured number from the causal guess. The team adopts an epistemic-status convention: each claim carries a tag. "Revenue was $4.2M" is marked [measured]. "Churn rose to 6.1%" is [measured]. "…because of the March pricing change" is marked [inference — untested]. "Enterprise demand will soften next quarter" is [assumption]. "Whether the pricing change hit new vs. existing customers" is left as [unknown — investigating].
Nothing about the underlying analysis changed; what changed is that a VP skimming the review can now see instantly that the alarming causal story is an untested inference riding alongside two hard measurements. The pricing decision that follows is made knowing which parts of the brief are load-bearing fact and which are the analyst's plausible guess — a distinction the old uniform prose actively hid.
How it works¶
- Fix the label set. Agree a small, mutually-exclusive vocabulary — observed fact, inference, assumption, hypothesis, expert judgment, unknown — and publish definitions so the tags mean the same thing to writer and reader.
- Tag at the claim grain. Attach a status to each claim (often splitting a sentence, so the measured part and the inferred part get different tags) rather than labeling a whole document at once.
- Separate knowledge from assumption in the marking. The convention's sharpest cut is forcing "assumption" and "unknown" to be visibly different words from "fact," so nothing merely assumed can wear a fact's plain voice.
- Enforce it downstream. The tags are only useful if they survive the copy-paste; the standard specifies that a claim carries its label wherever it is quoted, summarized, or escalated.
The distinguishing discipline is portability: the label is a property of the sentence, designed to move with it, not a note that lives back in the analysis.
Tuning parameters¶
- Label granularity — how many distinct statuses the vocabulary carries. Few labels are easy to adopt but blur real distinctions; many labels are precise but get misapplied and ignored.
- Marking overhead — how heavy each tag is, from a one-word prefix to a footnoted rationale. Light tags get used; heavy ones get dropped under deadline.
- Default status — what an unlabeled claim is presumed to be. Defaulting unlabeled claims to "assumption" pressures writers to earn "fact"; defaulting to "fact" is easier but re-opens the door the standard exists to close.
- Scope of enforcement — whether the convention is mandatory in formal artifacts or a norm in everyday writing. Broad enforcement changes the culture but risks tag-fatigue; narrow enforcement is easy but leaks in casual channels.
When it helps, and when it misleads¶
Its strength is that it makes warrant legible at the speed of reading — the downstream user distinguishes evidence-backed knowledge from assumption without access to the audit, which is exactly the archetype's final step. It descends from the tradition of standardized estimative language, where analysts fixed the meanings of words like "assess" and "likely" so that confidence and warrant could not be smuggled through ambiguous prose.[n1]
It misleads when the label is applied by tone rather than by warrant — a writer who tags a favored guess "expert judgment" to borrow its authority has weaponized the very standard meant to prevent that. Tags can also become ritual noise if everything is labeled and nothing is read, or false comfort if a confident [fact] tag is trusted without anyone checking the fact was ever warranted. The guarding discipline is to keep the label tied to an actual warrant determination (the standard communicates a status; it must not manufacture one), audit a sample of tags against their evidence, and treat a mislabel as a real error rather than a style quibble.
How it implements the components¶
Epistemic Status Labeling realizes the communication-and-marking components — the audit's downstream, reader-facing side:
communication_labeling_standard— it is the standard: a fixed, published vocabulary for marking each claim's epistemic status in the prose readers receive.uncertainty_and_unknown_marker— its "unknown / investigating" tags give residual uncertainty an explicit word in the text rather than leaving it to silence.assumption_vs_knowledge_separator— by forcing "assumption" to be a visibly different label from "fact," it keeps the two from speaking in the same voice on the page.
It does not rank warrant types by evidential strength (warrant_type_taxonomy, support_strength_rating — that is its nearest twin Evidence Ladder Labeling; the difference is that the ladder *ranks how strong a warrant is, while this standard marks what kind of claim it is for the reader), nor adjudicate whether the stated confidence is earned (confidence_alignment_rule — that is Claim-Confidence Warrant Review).*
Related¶
- Instantiates: Knowledge-Warrant Audit — this standard is the audit's final "communicate the status" step baked into the language.
- Sibling mechanisms: Assumption Conversion Prompt · Belief-Warrant Matrix · Claim-Confidence Warrant Review · Evidence Ladder Labeling · Source-Independence Cross-Check · Unsupported-Certainty Red Flag · Update-Trigger Checkpoint · Warrant-Decay Review
Editorial Notes¶
Form Classification¶
Form family: Rule, Policy & Commitment
Rationale: The convention imposes a standing vocabulary and claim-level tagging obligation that prevents assumptions, hypotheses, judgments, and unknowns from appearing as unmarked facts.
Nearest alternative: Communication, Facilitation & Learning — The tags communicate warrant to readers, but the operative form is the enforceable labeling standard governing how future claims must be written.
Review outcome: Adjudicated after independent review; high confidence.
Origin Attribution¶
Primary origin: Security Studies & Intelligence Analysis
Origin pattern: Convergent development
Present-day reach: Multi-domain
Rationale: Structured analytic tradecraft cohered explicit marking of observation, inference, assumption, judgment, and uncertainty so consumers can distinguish evidence from interpretation.
Related originating lineages:
- Communication & Media Studies — Scientific and risk communication independently developed calibrated language for evidentiary status.
- Philosophy — Epistemology supplies the underlying categories of warrant and belief status.
Review resolution: The current reviewers agree that security_intelligence is primary. For the reported differences (reported_ambiguity, alternate_origin_disagreement, origin_mode_disagreement), the evidence supports convergent, multi_domain, and communication_media_studies, philosophy; these choices preserve materially formative origins without conflating later domain reach.
Attribution caveat: No single tradition owns generic inline epistemic-status vocabularies.
Review outcome: Reconciled after independent review; medium confidence.
Notes¶
[n1] Sherman Kent's "words of estimative probability" — a proposal to standardize the meaning of uncertainty terms in intelligence writing, so that "probably" or "we assess" carried a fixed, shared sense rather than each writer's private one. Epistemic status labeling applies the same idea to kind of claim, not just degree of likelihood. ↩