Overview

The Argumentative Artifact Examination Framework is the territory framework for T1 — the operations that take an existing argument, claim-set, position, or text-as-argument and evaluate it for internal soundness, framing structure, rhetorical mechanisms, or propaganda function. Where the depth-molecular Argument Audit Analysis composes coherence and frame audits into a single integrated pass, this framework is the territory map: it names the four atomic and molecular modes that live in T1, the disambiguation question that selects among them, and the boundary lines that mark when the question is no longer T1 work but belongs to T2 (interest and power), T5 (hypothesis evaluation), T9 (paradigm examination), T10 (conceptual clarification), or T15 (artifact evaluation by stance).

The framework hosts four modes. Coherence-audit is the atomic neutral-stance assessment of whether premises support conclusion — Toulmin reconstruction (claim, data, warrant, backing, qualifier, rebuttal) per inferential move, charitable reconstruction first, named fallacies with quoted text and reason-it-fails-here, structural failures beyond the named-fallacy taxonomy. Frame-audit is the atomic stance-suspending surfacing of how the artifact frames the issue — Lakoff metaphor inventory, Goffman primary framework and keyings, Entman’s four functions per frame (problem definition, causal interpretation, moral evaluation, treatment recommendation), selection-and-salience inventory including what is silent, counterframe construction. Propaganda-audit is the atomic adversarial-stance specificity-specialized variant for artifacts suspected of propaganda function — Stanley’s supporting/undermining distinction, flawed-ideology test, not-at-issue content inventory. Argument-audit is the depth-molecular composition that runs frame-audit and coherence-audit in full and synthesizes their outputs into cross-cutting findings neither pass would catch alone.

The framework’s load-bearing intellectual content is the boundary discipline that keeps T1 distinct from adjacent territories. T1 evaluates an argument as an argument for soundness; T15 evaluates it as a proposal by adopting a defined stance (steelman, balanced critique, red team). The disambiguation is sharper than it looks. “Audit this argument’s structure” lands in T1; “make the strongest case for this proposal” lands in T15 even when the proposal happens to be an argument. The Steelman cross-territory case is handled explicitly: Steelman Construction’s home is T15, with T1 cross-reference activated when the artifact under steelmanning is itself an argument and argument-coherence considerations should inform the reconstruction.

The framework also carries the three open debates that T1 modes engage analytically without adjudication. D1 — Is fallacy a property of the argument, or of the dialogue? The classical Aristotelian-Copi tradition treats fallacy as inferential failure independent of dialogue context; pragma-dialectics (van Eemeren and Grootendorst) treats it as violation of rules for critical discussion; Walton’s pragmatic theory occupies a middle position with dialogue-type sensitivity built into argumentation-scheme critical questions. T1’s coherence-audit operates compatibly with all three. D2 — Motte-and-bailey: fallacy or doctrine? Shackel’s preferred usage frames it as a doctrinal characterization requiring multi-claim alternation evidence; popular usage has expanded it into a single-move fallacy label. T1’s argument-audit enforces the structural specificity. D5 — Is Stanley’s propaganda apparatus politically neutral or directional? Sympathetic readings treat Stanley’s diagnostic machinery as neutral; skeptical readings argue the case-base inflects it asymmetrically. T1’s propaganda-audit applies the apparatus while remaining agnostic on the debate, with symmetry guardrails (motive-attribution-without-evidence, propaganda-charge-as-refutation) named as failure modes.

The framework is honest about what T1 does not do. It does not adjudicate the conclusion’s truth (that requires independent grounds outside the audit). It does not evaluate proposals (that is T15). It does not compare paradigms (that is T9). It does not surface who benefits from the argument’s acceptance (that is T2). When the question shifts mid-conversation, the framework’s cross-territory adjacency notes route to the right territory rather than stretching T1 modes beyond their lane.

Systemic context

The Argumentative Artifact Examination Framework is the territory framework for T1 in Ora’s territory map. It composes downward into four mode files (Modes/coherence-audit.md, Modes/frame-audit.md, Modes/propaganda-audit.md, Modes/argument-audit.md) and the depth-molecular paired execution framework Framework — Argument Audit Analysis.md. It composes outward through five cross-territory adjacencies — T1↔T2 (argument soundness vs. interest pattern), T1↔T5 (argument-as-artifact vs. proposition-against-evidence), T1↔T9 (single-artifact frame vs. multi-paradigm comparison), T1↔T10 (argument coherence vs. concept precision), T1↔T15 (soundness audit vs. stance-bearing evaluation, with the Steelman cross-territory case handled explicitly). The four-stage pre-routing pipeline classifies prompts into T1 when the work in front of the user is an existing argument and the question is whether it holds up structurally, framingly, or as a piece of persuasion. Forward-link companion: see [[Paper — Argument Audit Analysis Framework]] for the depth-molecular operation that composes T1’s atomic modes into an integrated pass.

Ora’s output

Example question: I have a 4,000-word think-piece a colleague sent me. He thinks it is brilliant; I think something is wrong but I don’t know whether the issue is the framing, the logic, or both. How do I figure out which kind of audit I need before committing to running one?

Ora’s response (running the territory framework’s disambiguation protocol):

The right move is to disambiguate before running any of T1’s modes — a coherence-audit and a frame-audit are different kinds of work and answer different kinds of question. I will walk you through T1’s disambiguation question and then surface the four routing options.

T1’s central disambiguation question. “Is the question about whether the argument holds together internally, or about the frame or lens it is using to see the issue, or about both at once?” Three plain-English answers route differently.

Answer 1 — internal logic. If your reaction is “the conclusion does not follow from the premises” or “the inferential moves are doing something that does not check out,” that is coherence territory. Route to coherence-audit (Tier-2). The audit performs charitable reconstruction first, breaks each inferential move into its Toulmin components (claim, data, warrant, backing, qualifier, rebuttal), names fallacies with quoted text plus the inferential move plus the violated principle plus the reason it fails here, and surfaces structural failures beyond the named-fallacy taxonomy. The verdict is “this argument as given does not establish its conclusion” — which is a different verdict from “the conclusion is false.”

Answer 2 — frame or lens. If your reaction is “the framing here is doing more work than the argument” or “the metaphors are carrying the load” or “this presents a particular framing as the natural way to see the issue,” that is frame territory. Route to frame-audit (Tier-2). The audit names the operative frame in alternative-comparable vocabulary (not the artifact’s own naturalized terms), populates Entman’s four functions per frame (problem definition, causal interpretation, moral evaluation, treatment recommendation), inventories selection-and-salience including what is silent, audits the linguistic mechanisms (Lakoff metaphors, CDA presuppositions, nominalizations) by which the frame travels at the word and grammar level, and constructs at least one counterframe so the operative frame is visible as a frame rather than as the natural description.

Answer 3 — both. If you cannot tell whether the issue is logic or framing, or you suspect both, that is depth-molecular territory. Route to argument-audit (Tier-3). This runs frame-audit and coherence-audit in full and then synthesizes — surfacing arguments whose coherence is frame-conditional, coherence-failures that turn out to be frame-substitutions across claims, and motte-and-bailey-style structures operating across multiple claims. Argument-audit is a heavier pass; the time investment is warranted when both inferential and framing-level issues are plausibly at play. See [[Paper — Argument Audit Analysis Framework]] for the full molecular operation.

Specificity escalation — propaganda. If the artifact reads as propaganda or persuasion-engineered (Stanley-style strategic frame deployment combined with intentional inferential trickery), that is the specificity-specialized adversarial-stance variant. Route to propaganda-audit. The audit names the professed ideal of the artifact (the freedom, fairness, security, or truth it claims to embody), classifies as supporting (non-rational means for worthy ideal) or undermining (presents-as-embodying-ideal-while-eroding-it), identifies the flawed-ideology premises the audience must hold for the contradiction to remain invisible, and inventories the not-at-issue content doing the persuasive work.

Cross-territory boundary check. Three questions help decide whether the work you actually want is T1 at all. (a) Is the question really about who benefits if people accept this argument? That is T2 — interest and power analysis — and T1 modes will not produce that answer. (b) Is the question really about which of several competing hypotheses fits the evidence? That is T5 — hypothesis evaluation — and T1 will treat each competing position as a single argument rather than weighing propositions against evidence. (c) Is the question really “should we accept this proposal?” That is T15 — artifact evaluation by stance — and the Steelman cross-territory case applies (T15 is home; T1 is consulted when the artifact under steelmanning is itself an argument).

For your think-piece specifically. Without seeing the text, my best guess from your description (“something is wrong but I don’t know whether the issue is the framing, the logic, or both”) is that you are in argument-audit territory. The inability to articulate what is off is itself a signal that something cross-cutting may be operating — a frame is doing analytical work that the inferential pass alone would miss, or the coherence is conditional on accepting the frame. Run argument-audit. If the molecular pass reveals the issue is purely on one side, you will get the atomic-mode findings as part of the molecular output and can proceed accordingly.

That is what the territory framework does: it surfaces the four modes, surfaces the disambiguation question that selects among them, and surfaces the cross-territory boundaries so you do not run a T1 mode on a question that lives elsewhere.

Commercial AI comparison

Comparison content auto-populates when the comparison-refresh framework runs against this question. Drafters do not author this section.

Brief comparison commentary

Auto-populates with the comparison content above.

How to use this framework

You can run the Argumentative Artifact Examination pattern with any AI of your choice. The composition is single-pass against the territory disambiguation protocol followed by dispatch to the selected mode.

The prompt:

[Paste the framework specification]

Run T1 disambiguation on this question.

Question: [What you want to know about the argumentative artifact.]

Artifact (optional, attach if available): [Paste or describe the artifact.]

Why I think this needs auditing (optional): [What feels off, or what specific aspect you want examined.]

The AI runs T1’s central disambiguation question, identifies the right mode (coherence-audit, frame-audit, propaganda-audit, or argument-audit), and either dispatches to the selected mode or returns a routing recommendation if the question turns out to belong to an adjacent territory (T2, T5, T9, T10, or T15).

For best results:

  1. State your reaction in plain English first. “The conclusion does not follow” routes differently from “the framing is doing too much work.” Naming your reaction lets the disambiguation question fire correctly.
  2. Do not pre-select the mode. The framework’s value is the disambiguation discipline. If you walk in saying “run a frame-audit” when your actual question is coherence territory, you will get a frame-audit that does not answer your question.
  3. Provide the full artifact when you run a mode. All four T1 modes work better with the full text than with excerpts. Argument-audit in particular will fail to detect motte-and-bailey or frame-substitution patterns from excerpts that omit the multi-claim alternation.
  4. Accept the boundary checks. If the framework routes you to T2, T5, T9, T10, or T15 rather than to a T1 mode, that is the framework working as designed. Pushing a T1 mode against a non-T1 question produces the appearance of analysis without the substance.

The framework is deliberately tool-agnostic. The four-mode taxonomy, the disambiguation question, the cross-territory boundary discipline, and the open-debates engagement-without-adjudication all survive the lift to any environment.

Other examples

  • A campaign ad with explicit persuasive intent. The artifact’s persuasion-engineering puts it in propaganda territory immediately; the disambiguation routes to propaganda-audit rather than to coherence-audit or frame-audit. The audit names the professed ideal (security, fairness, freedom), classifies as supporting or undermining, identifies the flawed-ideology premises, and inventories the not-at-issue content. Demonstrates the specificity-specialized routing when the artifact’s category is recognizable up front.
  • A philosophy paper arguing for a particular metaethical position. The disambiguation surfaces both coherence and frame considerations. The user routes to argument-audit (molecular). The audit’s cross-cutting integration finds the inferential structure mostly holds, but the frame the paper imports (the specific metaethical vocabulary) is doing analytical work that the coherence pass alone would treat as background. The motte-and-bailey check is negative for this artifact (consistent frame across claims). Demonstrates the molecular routing for genuinely cross-cutting cases.
  • A debate transcript with two opposing arguments. The disambiguation surfaces that the user’s actual question may be T9 (paradigm examination) rather than T1 — when both sides have internally coherent arguments within incommensurable frames, the dispute is at the frame level rather than at the inferential level, and frame-comparison or worldview-cartography in T9 is the right tool. The framework routes accordingly rather than running coherence-audit on each side and producing analyses that miss the actual disagreement. Demonstrates the cross-territory boundary check working correctly.

Citations

The Argumentative Artifact Examination Framework’s four-mode structure draws from three convergent traditions. Argumentation theory: Toulmin’s The Uses of Argument (1958/2003) for the six-component decomposition; Walton, Reed, and Macagno’s Argumentation Schemes (2008) for the dialectical-classification overlay; Hamblin’s Fallacies (1970) and the post-Hamblin theoretical critique of the textbook taxonomy. Frame analysis and cognitive linguistics: Lakoff and Johnson’s Metaphors We Live By (1980); Goffman’s Frame Analysis (1974); Entman’s 1993 four-function operationalization; the Critical Discourse Analysis tradition (Fairclough, van Dijk, Wodak). Propaganda theory: Stanley’s How Propaganda Works (2015) and How Fascism Works (2018) for the supporting/undermining distinction and flawed-ideology mechanism; Bernays, Ellul, Herman and Chomsky for the structural-context lineages.

The territory’s disambiguation question and cross-territory adjacency map are internal to Ora and emerged from observing that conventional AI deployment on argumentative artifacts conflates audit-the-argument work with evaluate-the-proposal work, producing analyses that mix soundness assessment with stance-bearing critique in ways that satisfy no purpose well. The territory framework was compiled 2026-05-01 from the territory entry, member mode specs, lens dependencies, and open debates; v1.0 is the current version.

Downloads

  • Framework specification (PDF) — link to ora-ai.org canonical artifact when published
  • Framework specification (plain text) — link to ora-ai.org canonical artifact when published
  • Full white paper (PDF) — link when published