Claim Tagging Matrix¶
Template — instantiates Appearance vs. Reality Distinction Audit
Lists claims in rows and tags each by source, claim type, warrant level, bridge assumptions, and calibrated wording.
Claim Tagging Matrix is the workhorse container of the audit: a grid whose rows are the individual claims pulled out of a document and whose columns are the tags that fix each claim's evidential footing — where it came from, what kind of claim it is, how strongly it is warranted, what bridge would upgrade it, and how it should be worded. Its defining move is breadth over a set: it takes a whole tangle of assertions and lays them out in parallel so mismatches between wording and evidence become visible at a glance. It does not rewrite a single claim into its testable conditions, and it does not lock in a fixed set of domain lanes; it is a general-purpose register that can hold any claims from any source, one per row, so a reader can scan a column and ask "why is this sentence phrased as certainty when its source is a single self-report?" The warrant and bridge columns exist on the grid but are filled by dedicated siblings — the matrix owns the catalogue, not those specialist cells.
Example¶
An investigative desk is fact-checking a long feature before publication. The draft asserts, among many things, "the contractor's crews were untrained," "residents were sickened by the runoff," and "the agency ignored three complaints." Left as prose, these read as equally solid. The editor builds a Claim Tagging Matrix: each sentence that carries weight becomes a row. For "residents were sickened by the runoff," the source cell reads "12 resident interviews"; the type cell reads "first-person report + causal inference"; the warrant cell reads "strong for reported illness, weak for runoff as cause"; the calibrated wording cell rewrites it to "residents report illness they attribute to the runoff; the causal link is not yet established."
Scanning the finished grid, the desk sees that three of the story's punchiest lines share the same weakness — a first-person source carrying an object-level causal claim in its wording. The story ships with those three sentences recalibrated and one flagged for an added document. Nothing was deleted; the matrix simply made the whole argument's evidential shape legible in a way the flowing prose had hidden.
How it works¶
- Extract claims into rows. Lift the actual load-bearing sentences verbatim before editing them, one claim per row; split compound sentences that carry two claims.
- Trace the immediate source. In the source column, name what the claim rests on right now — an interview, a log line, a document, a sensor, a model output — not the conclusion it is used to support.
- Classify the claim type. Tag each row as report, observation, measurement, inference, social/institutional fact, or a mix; the mix tag is common and is not a failure.
- Recalibrate the wording. In the last column, rewrite the sentence so its form matches its warrant — "appears," "is reported," "is consistent with," "is established."
- Import the specialist cells. The warrant-level and bridge-assumption cells are filled by consuming the dedicated siblings rather than being invented inside the grid.
Tuning parameters¶
- Column set — the minimal grid is type + calibrated wording; a high-stakes grid adds source, warrant, bridge, and action. More columns catch more, but a wide matrix nobody finishes is worse than a narrow one that gets filled.
- Row granularity — how finely claims are split. Splitting a compound sentence surfaces a buried inference but multiplies rows; over-splitting drowns the signal.
- Warrant vocabulary — the words allowed in the warrant column, borrowed from the Observation Warrant Ladder; a shared vocabulary makes rows comparable, free text does not.
- Ownership — one editor fills every row, or each claim's originator fills their own; distributed filling scales but drifts without a shared rubric.
- Refresh cadence — a one-time pre-publication pass or a living document re-scanned as sources firm up.
When it helps, and when it misleads¶
Its strength is comparison: laid side by side, a set of claims reveals patterns of overreach that any single sentence hides — the same soft source propping up three different hard conclusions. It turns "does this feel solid?" into a column you can actually read down.
Its failure mode is that a filled grid feels like a finished judgment. A neat table of tags can lend false authority to claims that are still weak — the map is not the territory[1], and a claim does not become better warranted because someone typed a confident word in its warrant cell. The classic misuse is treating tagging as resolving: the row is marked "inference," everyone nods, and no one does the work the tag was pointing at. The guarding discipline is to read every tag as a question posted, not an answer filed — a "weak warrant" cell is an instruction to go get more evidence or soften the claim, not a box now safely ticked.
How it implements the components¶
Claim Tagging Matrix realizes the audit's cataloguing spine — the components that gather and sort a claim set, not the ones that transform or corroborate it:
claim_inventory— the rows are the inventory: the verbatim claims collected before cleanup.appearance_reality_classifier— the type column assigns each claim its report / measurement / inference / social-fact / mixed category.evidential_source_trace— the source column records each claim's immediate grounding, keeping equally important claims from being treated as equally licensed.language_calibration_rule— the calibrated-wording column rewrites each sentence so its form matches its warrant.
It does not annotate the leap inside an inference — it imports bridge_assumption_register from Bridge Assumption Annotation Protocol. It does not rewrite a claim into its testable experience_condition_set — that's Sense-Condition Rewrite Template — and it keeps no fixed clinical lanes with a phenomenological_report_channel, measurement_corroboration_plan, and ontology_commitment_map — that's Symptom–Biomarker Crosswalk.
Related¶
- Instantiates: Appearance vs. Reality Distinction Audit — the matrix is the general register the whole audit hangs on.
- Consumes: Observation Warrant Ladder supplies the warrant-column vocabulary; Bridge Assumption Annotation Protocol fills the bridge cells.
- Sibling mechanisms: Appearance/Reality Audit Checklist · Observation Warrant Ladder · Symptom–Biomarker Crosswalk · Perceived-vs-Measured Performance Dashboard · Sense-Condition Rewrite Template
Editorial Notes¶
Form Classification¶
Form family: Assessment, Review & Assurance
Rationale: Lists claims in rows and tags each by source, claim type, warrant level, bridge assumptions, and calibrated wording, making its operative form a bounded evaluation of existing evidence or work that produces a finding or disposition.
Independent corroboration: The frozen evidence defines Claim Tagging Matrix as 'Lists claims in rows and tags each by source, claim type, warrant level, bridge assumptions, and calibrated wording', so its operative form is Assessment, Review & Assurance.
Nearest alternative: Interface, Display & Cue — Although a fixed grid structures the work, its substantive operation is reviewing warrants and recalibrating each claim’s wording.
Review outcome: Independent reviewer agreement; medium confidence.
Origin Attribution¶
Primary origin: Rhetoric
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Argument-analysis practice established decomposing claims into sources, evidence, warrants, bridge assumptions, qualifiers, and rebuttals so evidential mismatch can be inspected.
Related originating lineages:
- Communication & Media Studies — Editorial fact-checking contributes source tracing and wording calibrated for publication.
- Philosophy — Epistemology supplies the warrant, evidence, and appearance-versus-reality distinctions.
Review resolution: Toulmin's argumentation model supplies the closest primary ancestor for claim, evidence, warrant, assumption, qualifier, and rebuttal fields. Philosophy supplies the epistemic distinctions and editorial communication contributes fact-checking and calibrated wording; the row-wise matrix is an Encyclopedia synthesis.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Researched adjudication after independent review; high confidence.
Sources consulted:
Notes¶
The matrix is deliberately a container, not a solver. The most useful discipline is to keep its specialist columns — warrant, bridge — populated by the dedicated siblings rather than filled in freehand inside the grid; that keeps the matrix broad and scannable instead of collapsing into one overloaded document that pretends to do every job at once.
References¶
[1] Korzybski, A. Science and Sanity: An Introduction to Non-Aristotelian Systems and General Semantics. International Non-Aristotelian Library Publishing Company (1933). Supports the map-territory distinction only; the tag-table and confident-label applications require separate support. registry ↩