E.17:5.1f - Shared anti-overread invariants for publication-facing units

Use the FPF pattern that defines, constrains, or tests the claim being made or effect under use. Keep any local review field local, preserve reduced bounded use, and address only the unsupported wider claim or effect through the source relation it requires.

Source-relation minimality. Name the smallest direct relation sufficient for the live use. A source reference, publication occurrence, evidence path, engineering-justification record, gate decision, and release decision are different objects or relations; choosing one licenses none of the others. Do not apply A.10, B.3, A.20, or A.21 when the use needs only source-finding, orientation, or inspection of an existing source episteme, publication occurrence, or status-register entry.

Local repair vs publication redesign. A local epistemic precision repair is enough only when it can preserve the current publication face or PublicationUnit while fixing one head, boundary, source relation, bounded use, explanation class, or unsupported downstream claim. If layout, grouping, visual emphasis, comparison arrangement, generated explanation, hidden source limitation, or mixed EntityOfConcern packaging still induces overread after the local relation is repaired, create a redesigned publication face or PublicationUnit instead of adding warning text around the misleading form.

Most-likely careful interpretation constraint. Design and word a publication-facing unit so its most likely careful interpretation does not exceed its named source relation and bounded use. A visible Approved head needs a visible GateDecision or a different head; sorted output needs its comparator or sorting relation visible if no recommendation is intended; generated explanation separates inferred links from pinned source claims by wording, label, or source reference.

Visual cue claim pressure. Layout, order, color, prominence, icon, grouping, and proximity can imply evidence, readiness, preference, equivalence, approval, or verification. Green can suggest readiness; top position preference; grouping equivalence; proximity to evidence an evidence relation; a badge approval; and a lock or checkmark verification. If that implication would change the next action, recover the exact evidence, assurance, gate, decision, recommendation, bridge, approval, or other record or relation that actually carries it, or redesign the face so the unsupported overread is no longer invited.

Extraction survival. When a PublicationUnit is excerpted, quoted, screenshotted, summarized, copied into a tutorial, retold by a generator, or moved to a slide, it keeps only the claims, source pins, boundary line, references named by value, and bounded use carried in that extracted unit. Any use that depended on hidden neighboring context is lost unless that context is carried by source pins, a boundary line, or a reference named by value. A dashboard screenshot does not carry the underlying gate record, a quoted comparison row does not carry the full comparator or sorting relation unless that relation is included or referenced, a copied explanation paragraph does not carry source pins unless pins remain recoverable, and a pattern excerpt does not carry the whole pattern boundary unless the excerpt states or cites it.

No-extra-pattern case. If a publication-facing unit has bounded use only for ordinary orientation, learning, source-finding, review, comparison, or planning preparation, and no operative work or reliance, evidence, gate, assurance, bridge, source-dispute, or release claim is present, keep the existing publication source relation and proceed with ordinary use. The visible closure is: no operative work or reliance, evidence, gate, assurance, bridge, source-dispute, release, durable naming, or project-side source-relation claim recovered; ordinary publication wording remains bounded to the current use.

Pattern-inflation anti-pattern. Do not apply a neighboring pattern merely because the publication-facing unit resembles a worked example. Apply the neighboring pattern only when a claim being made or effect changes the next available project move.

Strategic overread invariant. Apply the same anti-overread rules whether the misleading interpretation is accidental, conventional, incentive-driven, or intentionally induced by publication design. Green status color without GateDecisionRef, reviewed-looking wording without approval, selective source links without operative-claim source relation, comparison ordering without selection decision, hidden caveats behind a source link, or pins for trivial claims beside unpinned causal linkage do not create evidence, gate, decision, assurance, work, release, or bridge relation by design pressure.

Carrier-travel invariant. A copied, exported, screenshotted, summarized, generated, translated, or re-rendered face can carry orientation or source-finding cues. It carries no evidence, authority, gate, approval, engineering justification, work, currentness, or release relation unless the exact corresponding object and relation remain recoverable for that use.

Derivative-chain decay. A second-order rendering inherits at most the bounded use that is explicitly carried from the prior source relation. It does not inherit source faithfulness, evidence relation, currentness relation, authority-reference relation, gate decision, work relation, or reliance relation by default.

Publication-face snapshot and refresh identity. A face can keep the same layout, name, or carrier while its source pins, data window, source-relation status, currentness, EditionId, or bounded use changes. Visual sameness is not source, evidence, or use-boundary sameness. Beyond orientation, identify the face edition or snapshot, the source pins or data window that still carry the claim, and any changed bounded use. If those cannot be recovered, use the face only for orientation/source-finding or reissue it from the source under E.17 and the concrete downstream rule that the new use needs.

Claim-level source relation only. Do not assign one whole-unit source-relation status unless every operative claim in that publication-facing unit has the same source relation named by value for the same use and unsupported downstream uses are explicit.

Modality and deontic-force preservation. Publication-facing transformations preserve possibility, obligation, permission, recommendation status, decision status, confidence, scope, and temporal window when those values change the claim or use. If one changes, narrow the bounded use or apply the concrete definition, constraint, decision, evidence, work, gate, or authority rule that carries it. Comparison does not become recommendation or decision; explanation does not become evidence; a face does not become authority; a publication unit does not smuggle a downstream effect; source-linked does not mean source-available for reliance; ready-looking does not mean gate-passed.

This preservation rule also applies across extraction, translation, screenshotting, summary, and generated retelling. A translated permission is not wider permission, a screenshot of approval-looking display is not an approval record, a summary of evidence is not an evidence path, and a generated retelling of a decision is not the decision record unless the source relation that makes the operative claim recoverable by value and source pins survive in the new publication-facing unit.

Reader position is not a project system-role kind or assignment. Reader position, audience, target user model, verifier position, review-reader position, and learner position do not become project system-role kinds, U.SystemRoleAssignment occurrences, decision authority, gate authority, issuer relations, responsibility relations, or Work contexts by publication. If any of those values is current, cite its typed project-side reference and direct predicate separately; otherwise record the exact missing governor rather than inferring it from a reader label.

Source-gap states. When the source relation is missing, say which source gap is present: source not named; source named but unavailable; source available but not used; source used but insufficient; source stale or outside its window; source contradicted; or mismatch among the source-maintenance System, any maintenance Work whose exact performer is recovered through A.13 and which A.15.1 admits independently, the status register, and a separately established responsibility relation. Add F.6 only when the source-maintenance comparison expressly consumes precise assignment-bound attribution; its failure leaves the maintenance Work intact. Assignment establishes neither source-maintenance responsibility, classification, nor status-register authority. Block only the unsupported effect and keep any reduced bounded use available.

Measure and display overread. Before relying on a displayed number, score, percentage, color, rank, confidence, similarity, or status, recover the exact claim and its source. Retain the characteristic or criterion, scale, scope, window, aggregation rule, calibration, or evidence basis only where the applicable pattern requires it or it changes the value’s interpretation. A status or ordinal judgment does not become a measurement by display. If the basis required for the intended use is missing, keep only the supported orientation or source-finding use. Use A.10 for evidence reliance, B.3 for assurance, A.20/A.21 for an actual constraint or gate claim, and F.9 only for a separately claimed Bridge or its bounded use; F.9.1 may add a stance note about an already constituted claim.

World-contact stop. When a source update, revocation, policy change, holon-state change, incident, model update, environmental change, or new observation changes a premise required by a downstream use, revalidate that affected use before it continues. Refresh the source, reissue the publication, or recover a changed project-side result as that use requires. Reuse unaffected results; a changed label or edition alone does not establish changed meaning or invalidate every publication use.

Functional-description boundary. A functional, architectural, descriptive, representational, or explanatory fit claim creates no permission, obligation, approval, gate passage, release relation, performed-work evidence, or engineering justification. Those uses need the exact separate work, authority, evidence, decision, gate, release, or assurance object and relation that carry the claim.

Mixed bundle no-shared-evidence-relation rule. A bundle with source-pinned, reduced-use, speculative, didactic, comparison, and evidence-facing parts is not interpreted under one shared evidence relation or use-boundary value borrowed from another member. Each operative claim keeps its own source relation and unsupported downstream use.

Educational usefulness. Didactic, onboarding, tutorial, and workshop usefulness is real orientation aid. It is not evidence, gate passage, approval, work occurrence, engineering justification, release permission, or bridge relation.

Comparison exposes conflict; it does not adjudicate it. A comparison note can expose contradiction, asymmetry, different foregrounding, or residue. It does not select an option, approve release, pass a gate, or create bridge or substitution relation unless the corresponding C.11, A.20/A.21, F.9, or other exact decision relation carries that result.

Same publication-facing unit, multiple interpretations. A green release dashboard can be one MVPK face for source-finding, an A.10 evidence/provenance path that cites a G.11 currentness result when the source query is recoverable, an A.21 gate-decision view when the GateDecisionRef is recoverable, or an unsupported release cue when those sources are missing. A generated comparative explanation can be an E.17.EFP explanation-use case, an E.17.ID.CR comparison case, a A.6.3.CR generated-summary case, or source-finding only; it is never all of those under one shared evidence-relation class or bounded-use value by fluency alone.

Archetypal publication-use cases. Use these as quick recognition slices, not as a closed taxonomy:

• Green dashboard tile. A tile says Model ready. Treat the tile as the PublicationUnit when that tile carries the present release overread. Recover the exact A.21 gate result and the source basis required by its profile application before using the tile as gate passage. A release claim also needs its own decision or authority basis. Without those, retain source-finding and status orientation only; green color or placement supplies neither result.
• Generated explanation with source links. A generated text explains a method and cites sources. The explanation rendering is not source replacement. Source links carry only the pinned operative claims they actually carry. If work or reliance is present, use A.10 for the evidence path named by value or keep the rendering as reader help; if the rendering is deliberately reduced-use, use A.6.3.CSC.
• Comparison table. A table compares two methods and places one first. Ordering is not selection. The comparator or sorting relation, source references, shared review frame, and unsupported downstream claim remain visible. Choice or decision needs C.11; equivalence or a Bridge needs F.9, while F.9.1 may add only an optional stance note about an established bounded-use claim.
• Unrecovered source wording. A draft uses source-object wording, undeclared interpretive-view shorthand, or generic unit wording without naming the FPF kind. Recover the FPF kind and relation positions instead of minting source-relation pseudo-kinds or undeclared interpretive-view pseudo-kinds. Use PublicationUnit only when a bounded reader-inspected unit inside a publication is present; otherwise use the exact episteme, view, publication, carrier relation, section of a named non-pattern FPF publication form whose reader-help function and reference are recoverable, A.6.P relation claim, or typed project-side FPF kind and reference named by value.
• Translated tutorial. A translated tutorial can improve reader access to an FPF pattern. It is a derivative rendering, not the original source. Operative claims need source mapping for reliance, translated heads can need E.17.AUD.LHR or C.2.P, and F.18 is present only when durable naming, UTS, Core-facing, or cross-context naming work is intended.

Practical harm prevented by neighboring pattern. Use this map when the reader asks what the discipline buys in practice:

Blocked overread with useful publication use remaining.

• A comparison table appears to select option B. Block the selection interpretation when no C.11 ChoiceResult, decision record, or visible selection relation exists. Useful publication use remains: use the table as a bounded comparison under E.17.ID.CR, or apply C.11 when selection is intended.

• A green dashboard tile appears to permit release. Block that interpretation when the exact gate result, its applicable profile and required check results, or the release decision or authority basis is not recoverable. Useful publication use remains: source-finding and status orientation, followed by inspection of the missing source when release is intended.

• A generated explanation appears to prove a causal relation. Block the evidence or assurance interpretation when source pins and evidence path are absent or insufficient. Useful publication use remains: use the explanation as reader help or source-finding, then use A.10 for the evidence path or B.3 for the engineering-justification claim.

• C.2.P prevents the wrong object from being treated as source, the wrong relation from being treated as source relation, and a loose phrase from being treated as an FPF kind.

• E.17.AUD and E.17.AUD.OOTD prevent action on a publication unit whose primary subject, carried publication move, or outside boundary shifted silently.

• E.17.ID.CR prevents a comparison unit from being used as decision, equivalence, bridge, evidence, or release source relation.

• E.17.EFP prevents fluent explanation from laundering unsupported claims into reliance, assurance, gate, or evidence use.

• E.17 MVPK prevents a readable publication face from being treated as evidence, gate, work, authority, or release source relation by display quality.

• F.18 prevents a local name from becoming global identity without context, kind, lineage, and bridge or cross-context naming relation.

Anti-escalation examples. Do not apply a neighboring pattern when its claim being made is absent:

• Do not apply F.18 when a one-off local phrase repair restores the local kind, relation, and bounded use without minting a durable reusable name.
• Do not apply A.10 when the publication-facing unit is not being used for reliance, evidence, provenance, currentness, or claim-bound evidence relation.
• Do not apply A.21 when a dashboard tile is merely status orientation and no GateDecisionRef or gate profile is present.
• Do not apply F.9 when a comparison does not claim sameness, substitution, bridge relation, or cross-context equivalence.
• Do not apply E.17.EFP when the text is only a same-entity rewrite or representation change under A.6.3.CR or A.6.3.RT.

Concrete reopen trigger. Name the condition and the nearest source-bearing side or the concrete definition, constraint, test, decision, evidence, work, or authority relation to revisit. A vague reopen if needed does not preserve the source relation.