#Controlled Semantic Coarsening

A.6.3.CSC · Stable · Architectural (A) · Normative unless marked informative · Part A - Kernel Architecture Cluster

#About this pattern

This is a generated FPF pattern page projected from the published FPF source. It is canonical FPF content for this ID; it is not a fpf-memory product feature page.

#How to use this pattern

Read the ID, status, type, and normativity first. Use the content for exact wording, the relations for adjacent concepts, and citations to keep active work grounded without pasting the whole specification.

Type: Architectural (A) Status: Stable Normativity: Normative unless marked informative

Placement. Controlled Semantic Coarsening is a specialization under A.6.3 U.EpistemicViewing for same-lineage coarsening from one source-bearing side into one coarsened rendering, whether the coarsening was planned before publication or discovered during review of a target that can be retained only under a narrower-use card. That source-bearing side may be one governed source episteme, source publication, or declared source set with a stable source-set identifier and bounded membership; it is not an open corpus.

Builds on. A.6.3, A.6.3.CR, A.6.3.RT, E.17.EFP, A.6.P, E.8, E.10, E.19, and F.18.

Coordinates with. E.17.ID.CR, F.9, F.9.1, A.15, A.6.4, A.20, and A.21.

Use this when. A summary, briefing, redaction, dashboard tile, lookup handle, didactic compression, or other readable target coarsens one source-bearing side by dropping or narrowing distinctions, recoverability, reliability transport, or admissible-use posture, or when review discovers that the target can be retained only as a coarsened rendering.

#Keywords

• controlled semantic coarsening
• source-bearing episteme or source publication
• coarsened rendering
• narrower admissible use
• non-admissible downstream use
• reopen trigger
• redaction
• dashboard tile
• lookup handle
• state-representation shortcut.

#Relations

#Content

#Problem frame

Use this when. A summary, briefing, redaction, dashboard tile, lookup handle, didactic compression, or other readable target coarsens one source-bearing side by dropping or narrowing distinctions, recoverability, reliability transport, or admissible-use posture, or when review discovers that the target can be retained only as a coarsened rendering.

Plain recognition line. A short version is useful only while the reader can still see what it came from, what it leaves out, and when to go back.

Controlled Semantic Coarsening governs one coarsened rendering that remains useful only because the source-bearing side stays identifiable, the admissible use is narrower, downstream use is non-admissible from the coarsened rendering alone, and escalation reopens that source-bearing side. It is the FPF governing pattern for that source-to-rendering relation. It is not a tag, token, U.* kind, publication face, carrier, bridge card, stance overlay, work plan, approval, or gate.

Start here when. Your first honest publication unit is a small controlled-coarsening card: source-bearing side, coarsened rendering, narrower admissible use, declared source-loss mode, non-admissible downstream use, and reopen trigger. Read orientation use, reliance use, operative claim, non-admissible downstream use, and reopen trigger through the shared E.17:5.1c terms; use E.17:5.1d when the primary live question may be ordinary rewrite, representation change, explanation, comparison, bridge or substitution, work or reliance, gate, evidence, assurance, retargeting, or carrier or front-end work instead of coarsening.

Neighboring project records and governing patterns. Ordinary same-entity wording belongs under A.6.3.CR; representation-scheme change belongs under A.6.3.RT; explanation-facing class discipline belongs under E.17.EFP; bounded comparison belongs under E.17.ID.CR; bridge or substitution use belongs under F.9 or F.9.1; changed described entity belongs under A.6.4; work authority requires A.15-governed selected method, U.WorkPlan, performed U.Work, work-result record, or result-measurement record; gate or adjudication authority requires A.20 or A.21-governed project records.

What goes wrong if missed. A helpful coarsened rendering starts acting like the source-bearing side: a summary becomes evidence, a redaction becomes accountability closure, a dashboard tile becomes a causal verdict, a comparison note starts supporting bridge or substitution use, or a briefing becomes work authority.

What this buys. FPF users get a cheap admissible way to publish coarsened renderings without hiding declared loss, overclaiming authority, or forcing every ordinary summary through a full assurance record. This is the positive path for bounded dashboard tiles, redactions, partner notes, lookup handles, workshop simplifications, and didactic compressions that help work without pretending to be the source-bearing side.

Working action spine. A coarsened rendering is useful for a narrower use but cannot carry the source-bearing side -> separate source-bearing side, coarsened rendering, narrower admissible use, declared source-loss mode, non-admissible downstream use, and reopen trigger -> use the coarsened rendering for orientation, triage, disclosure, retrieval, comparison, or planning preparation -> output the six-row mini-card -> reopen or hand off if reuse, reliance, citation, dispute, bridge, work, gate, privacy, or engineering-justification demand appears.

Ordinary use. If the coarsened rendering is admissible only for orientation, bounded disclosure, retrieval, workshop framing, preliminary triage, comparison, or planning preparation, use the six-row mini-card and stop there.

Load-bearing use. Open the load-bearing coarsening record only when the coarsened rendering will be externally relied on, disputed, cited, used across context, policy-bearing, bridge-adjacent, work-adjacent, gate-adjacent, privacy-sensitive, or engineering-justification-facing.

Stop condition. Stop at the mini-card when the coarsened rendering changes no next admissible work or reliance, disclosure, review, or planning-preparation move and blocks no concrete overclaim beyond its narrower admissible use.

Admissible-use examples.

Not this pattern when. Not this pattern when the primary question is ordinary same-entity wording, representation-medium change, explanation fidelity, comparison, bridge or substitution use, changed described entity, work authority, approval, adjudication, or gate authority. Use the neighboring governing FPF pattern or authoritySourceRef target for that primary question.

#Problem

FPF often needs a coarsened form of a source-bearing side: a manager summary, a redacted disclosure note, a dashboard tile, a lookup surrogate, a workshop simplification, or a didactic compression. The coarsened form can be valuable, but it becomes dangerous when readers forget that its admissible use is narrower than the source-bearing side.

The core failure is not ordinary omission by itself. The failure appears when the coarsened rendering stays honest only under an admissible-use card like this:

• the source-bearing side remains governing;
• the coarsened rendering has a declared source-loss mode or reduced recoverability;
• the target makes only the narrower use admissible;
• downstream use is non-admissible from the coarsened rendering alone;
• downstream use reopens the source-bearing side or moves to the exact governing FPF pattern or authoritySourceRef target that supports the requested use.

Without a named pattern for that relation, neighboring patterns repeat partial coarsening rules locally. The repetition hides the shared load and makes it too easy for coarsened renderings to travel as if they were the source-bearing side.

#Forces

#Solution

Controlled Semantic Coarsening governs one source-to-rendering relation.

• Source-bearing side means the governed U.Episteme, governed U.EpistemePublication, or declared source set that still carries the fuller claim, distinction, evidence relation, trace relation, or authority-reference relation. A declared source set must have a stable source-set identifier, bounded membership, and a reopen condition; an open corpus, folder, topic area, search-result cluster, or vague document neighborhood is not a source-bearing side.
• Coarsened rendering means the target readable form that carries a declared source-loss mode, reduced recoverability, reduced reliability transport, or narrower admissible use than the source-bearing side.
• Narrower admissible use means the practical use the coarsened rendering makes admissible, such as orientation, retrieval, bounded disclosure, workshop framing, or preliminary triage.
• Non-admissible downstream use means the use the coarsened rendering does not make admissible alone, such as approval, audit closure, release gate, work plan, equivalence, bridge or substitution use, accountability finding, or canonical technical claim.
• Reopen trigger means the condition that requires return to the source-bearing side, re-expansion in the current rendering or publication, or handoff to another governing FPF pattern or authoritySourceRef target.
• Load-bearing case means a coarsening case that will be cited, disputed, externally relied on, policy-bearing, bridge-adjacent, gate-adjacent, work-adjacent, privacy-sensitive, or assurance-facing.

#Ordinary mini-card

For ordinary use, publish only the smallest card that keeps the coarsened rendering honest.

A CSC card makes only the narrower admissible use named on the card admissible for the coarsened rendering. It never makes the non-admissible downstream use admissible; it only tells the reader when and where to reopen the source-bearing side or hand off to the governing pattern that carries that downstream use.

The card may live inline. Inherited source pins count when the surrounding publication already makes the source-bearing side visible.

If the coarsened rendering is used only for local orientation and the source-bearing side remains adjacent, the six-row card may be inline or implicit by immediate context; do not create a durable Controlled Semantic Coarsening object unless reuse, reliance, citation, or dispute appears.

#First check

Before using this pattern, ask five questions:

1. Is there exactly one source-bearing side: one governed source episteme, source publication, or declared source set with stable identifier, bounded membership, and reopen condition?
2. Does the target declare a source-loss mode against that source-bearing side, or has review shown that it can be retained only as a coarsened rendering?
3. Does the target make only narrower use admissible?
4. Is downstream use explicitly non-admissible from the target alone?
5. Is the source-bearing reopen or governing-pattern handoff trigger visible?

If any answer is no, do not polish a coarsening story. Use the ordinary governing pattern or recover the exact project-side FPF kind and reference or authority-reference relation that actually supports the requested use. If the required support is missing, create only a prospective repair request, future decision request, prospective work-plan entry, or explicit source-gap note; do not treat that request or note as retroactive support for the coarsened rendering, earlier claim or effect, work occurrence, evidence, approval, gate passage, release permission, or engineering justification.

#Ordinary vs load-bearing

Ordinary cases should remain light. A short orientation summary, redacted partner note, workshop simplification, or lookup handle does not need the full assurance record if the six-row card is recoverable.

Load-bearing cases add only the fields that matter for the live use, dispute, reliance, citation, policy, bridge, work, gate, privacy, or assurance case. This list is not a daily gate for ordinary summaries, briefings, redactions, or lookup handles:

The fields below inherit the E.17:5.1e local-field rule. They are review aids for one coarsened-rendering case, not U.Kind, SurfaceKind, RelationKind, KindBridge, EvidenceKind, GateDecision, SpeechAct, Commitment, U.Work, authoritySourceRef target, or exact project-side FPF kind and reference unless another governing FPF pattern explicitly instantiates that object.

• sourceBearingSideRef and coarsenedRenderingRef when the source-bearing side, coarsened rendering, PublicationUnit, publication face, PublicationSurface, InteropSurface, or carrier could be confused;
• targetPublicationUnitIfAny when the coarsened rendering is carried by one PublicationUnit that is distinct from the publication, disclosure note, dashboard tile, or InteropSurface on which it appears;
• governingPatternRef, projectSourceRecordRef, or one privileged reopen path, so a coarsened rendering cannot reset its own provenance;
• coarseningBranch, sourceLossMode, and admissibleUsePosture as separate fields;
• recoverabilityAfterCoarsening when the source-loss mode affects claim support, accountability, admissible-use posture, or later citation;
• at least one kept claim bundle or distinction bundle, one coarsened or dropped bundle, and one reopen-only bundle when the case is disputed or later-cited;
• sourceSupportPosture when the E.17:5.1b postures could diverge: source pointer, source availability, source retrieval, source use, source faithfulness, claim support, contradiction, plausibility-only, omission, declared source-loss mode, added commitment, added linkage, independent verification, admissible use, non-admissible downstream use, or reopen trigger;
• uncertainty or abstention posture when branch reading, preserved distinctions, source pin, or admissible use cannot yet be stated stably;
• independent-verification question when downstream testing, assurance, gate, or external reliance appears;
• audienceOverReadRisk, plus a light reader-reliance or user-evidence check when readers may mistake the coarsened rendering for authority it does not carry;
• whether local re-expansion is enough to repair the current rendering or whether downstream use still needs return to the source-bearing side or named authoritySourceRef target.

#Branch and admissible-use discipline

coarseningBranch answers what sort of coarsening case this is. sourceLossMode names what was lost from the source-bearing side. admissibleUsePosture answers which use of the coarsened rendering remains admissible. Do not infer any one of the three from the others.

Ordinary admissible use covers aggregation, quotient-like orientation, didactic or report summaries, and briefings only for the named narrower use. Source-pinned-only use covers surrogate, index, retrieval-hint, lookup, and handle forms; these may help find or orient to the source but do not provide claim support themselves. authoritySourceRef-reopen-only covers the exceptional case where the coarsened rendering names the source whose named authority relation must be reopened; the coarsened rendering itself does not become the authoritySourceRef target, evidence source, gate source, or work source.

Privacy or redaction cases are admissible here only when the card names the sharing boundary, the source-loss mode, what was withheld or coarsened, the main re-identification or accountability risk being reduced, the source-bearing review path, and the accountability or gate uses that remain non-admissible.

Exceptional interop-facing simplification is not ordinary coarsening. It is admissible here only when it stays source-tethered and names the operative relation kind, such as bounded contrast, broader or narrower, partial overlap, proxy, lossy normalization, or context-bounded match. If the coarsened rendering makes bounded contrast across contexts or source epistemes or source publications is the primary question, use E.17.ID.CR. If it implies equivalence, substitution, projection, or bridge or substitution use, use F.9 or F.9.1.

#Source-loss mode, recoverability, and anti-overread

The card must name the live sourceLossMode before a coarsened rendering is treated as admissible for its stated use. A source-loss mode is not a strength scale. It names which source-bearing distinction failed to travel into the coarsened rendering.

Recoverability and admissible use are separate. A recoverable coarsened rendering is not automatically admissible for downstream use, and a non-admissible use is not repaired merely by saying the source could be found.

A coarsening chain may not silently reset provenance. If one coarsened rendering is reused to make another, the same source-bearing side must stay explicit, the earlier source-loss mode and uncertainty posture must remain visible, and the new rendering must declare only the added source-loss delta. If that cannot be stated cleanly, reopen the source-bearing side rather than extending the chain.

Aggregation or quotient-like coarsening remains inside this pattern only while the coarsened rendering keeps one bounded described set, slice, case bundle, or alternative bundle explicit as the described entity or described set. If several entities, alternatives, or slices become one new class target or proxy target, apply A.6.4.

#Neighbor exits

Neighboring governing patterns may point here when a coarsened rendering relation becomes primary. They do not govern the shared coarsening relation by local repetition.

#Well-formedness constraints

Well-formedness constraint CSC-WF-1 (source-to-rendering relation). A controlled-coarsening case is well formed only when it contains exactly one source-bearing side, at least one coarsened-rendering side, one declared narrower admissible use, one non-admissible downstream use, and one visible source-bearing reopen or governing-pattern handoff condition. The source-bearing side may be one governed source episteme, source publication, or declared source set with stable source-set identifier and bounded membership; it must not be an open, vague corpus.

Well-formedness constraint CSC-WF-2 (no authority upgrade). A coarsened rendering does not gain evidence, bridge, work, approval, gate, or adjudication authority by repetition, fluency, audience convenience, citation, or publication on a more visible publication face or channel.

Well-formedness constraint CSC-WF-3 (source path continuity). A coarsening chain remains well formed only while the same source-bearing side, prior source-loss mode, uncertainty posture, and added source-loss delta remain recoverable.

#Archetypal Grounding

Tell. Controlled semantic coarsening is the disciplined act of making a coarsened rendering useful while keeping the source-bearing side and the non-admissible downstream uses visible. It is not simplification as style. It is simplification under a source, use, loss, and reopen card.

Show (System). A service team has an incident review with trace details, confidence bands, and alternative branches. A manager dashboard tile says: Cache failover evidence is the leading concern; details remain in IR-42. The tile may orient planning, but it may not approve release, close audit, prove causality, or trigger work without reopening IR-42.

Show (Episteme). A research review bundle is given the lookup handle cache-failover risk. The handle is admissible for retrieval and orientation only. Any claim-bearing use reopens the review bundle because the handle does not carry the evidence, alternatives, or source support.

#Worked slices

Manager orientation summary. The source-bearing side is incident review IR-42 with trace details, confidence bands, and alternative branches. The coarsened rendering is Cache failover evidence is the leading concern; details remain in IR-42. Its narrower admissible use is orientation for planning conversation. Its non-admissible downstream uses are approval, audit closure, release gate, causal proof, and work order.

Redacted partner note. The source-bearing side is a full incident record with actor identity, trace path, and recovery evidence. The coarsened rendering is a partner-facing redacted note that withholds actor identity and trace path. Its narrower admissible use is bounded disclosure and coordination. Accountability, legal, audit, readiness, and gate uses reopen the full incident record or name the relevant authoritySourceRef target.

Redacted functional-description publication. The source-bearing side is a functional architecture note that names flow relations, method-selection limits, work-plan prerequisites, result-measurement requirements, and two exception cases. The coarsened rendering is a partner-facing table that keeps the main flow relation and removes the exception cases and result-measurement details. Its narrower admissible use is bounded orientation for coordination. Work planning, gate passage, evidence, engineering justification, control-architecture use, and release permission reopen the source-bearing side or apply A.15, A.10, B.3, A.20, A.21, or B.2.5 as the live claim requires.

Exceptional interop-facing simplification. The source-bearing side is two pinned context notes plus their bridge or comparison basis. The coarsened rendering is: For this exchange only, Field A is treated as broader than Field B; see source notes for exceptions. The rendering may orient the exchange, but any equivalence, substitution, projection, bridge-row, or approval use applies F.9 or F.9.1 or reopens the source-bearing side basis.

Bad fit: hidden work authority. Deployment may proceed; see summary S-3. This is not an admissible controlled coarsening card. The sentence tries to convert a coarsened summary into execution or gate authority. Use A.15, A.20, or A.21, and reopen the source-bearing side before any work or approval claim proceeds.

#Bias-Annotation

Lenses tested: Gov, Arch, Onto and Epist, Prag, Did. Scope: Universal for source-to-rendering relations that claim controlled semantic coarsening inside FPF.

This pattern favors Prag and Did by allowing useful coarsened renderings to remain cheap and readable. It also favors Gov and Arch by requiring non-admissible downstream use, source reopen, and neighboring-pattern application when release, policy, assurance, adjudication, bridge, work, evidence, or gate use is attempted. The mitigation for over-governance is the ordinary mini-card: ordinary cases stay light, and only dispute, citation, external reliance, policy, bridge, work, gate, privacy, assurance, release, or adjudication use adds load-bearing fields.

#Conformance Checklist

A conformance check is retained only if it changes the next admissible use of the coarsened rendering, blocks a concrete overclaim, or preserves the source-bearing reopen path needed for the declared admissible use.

#CSC-Core

#CSC-Conditional

#Common Anti-Patterns and How to Avoid Them

#Consequences

#Rationale

Controlled coarsening is useful because FPF work often needs cheap readable forms. It is risky because cheap readable forms often travel farther than their admissible use. The pattern therefore does not ban coarsened renderings; it makes the source-to-rendering relation explicit enough that later users know when to stop, reopen, or hand off to another governing FPF pattern or authoritySourceRef target.

This pattern is narrower than a general simplification pattern. It applies only when the coarsened rendering remains tied to a source-bearing side and carries a narrower-use card.

The core memory aid is simple: a coarsened rendering may help reading, but it must not become the source-bearing side it was derived from. It may expose or cite the source-bearing side or the exact project-side FPF kind and reference that carries the requested support; that exposed source or value remains the support, not the coarsened rendering's readable face. If support is missing, a repair request, source-gap note, or reopen note may guide only future repair or return to source; it does not backdate the coarsened rendering into source support.

#SoTA Alignment: Adopted Or Adapted Invariants And Rejected Shortcuts

SoTA alignment rule. Read each row here as source idea -> local FPF invariant -> practical local test -> popular shortcut rejected. A source citation governs nothing by reputation; it counts only when the cited idea is translated into the Solution, conformance checks, boundary rules, worked slices, and Relations of this pattern.

Purpose. This section justifies the pattern's safeguards. It is not an additional operational checklist. The Solution, Conformance Checklist, worked slices, and Relations above carry the live pattern discipline.

Positive SoTA role. Use CSC when a coarsened readable rendering is still worth using in project work, but only for a narrower admissible use and without pretending that the rendering carries the source-bearing side's support.

The practical implication is the same across these traditions: coarsened readable publication faces or renderings are valuable, but their admissible use depends on source support, relation kind, validation evidence, audience, and reopen path. The worked slices in A.6.3.CSC:5.1 are the nearest recovery anchors for those SoTA rows.

Semantic-web boundary. In the W3C row, Data on the Web, SHACL, and DCAT support publication metadata, provenance, validation, cataloging, and interoperability. They do not by themselves support work occurrence, gate passage, bridge or substitution use, equivalence, release permission, or project claim support; those uses require the governing pattern or exact project-side FPF kind and reference that carries that claim.

#Relations

• Specializes: A.6.3 U.EpistemicViewing for declared source-loss mode in a same-lineage source-to-rendering relation.
• Coordinates with: A.6.3.CR, A.6.3.RT, E.17.EFP, E.17.ID.CR, F.9, F.9.1, A.15, A.6.4, A.20, and A.21.
• Does not replace: conservative retextualization, representation transduction, explanation profiling, bounded comparative reading, bridge-card discipline, stance overlay, changed-object discipline, work authority, gate authority, or adjudication authority.
• Entry relation: neighboring patterns may hand off here when a coarsened rendering's narrower-use, non-admissible-use, and reopen card becomes the primary question.
• Governing-pattern relation wording: this pattern is a specialization under A.6.3, not a bundle, suite, profile, overlay, or review pack. Its governing role is limited to the controlled-coarsening relation itself.

#Boundary with quantum-like state-representation coarsening

Use CSC first when one source-bearing side, model, state representation, or evidence set is made less detailed for a narrower use, or when review discovers that a target already in circulation can be retained only under a narrower-use card: summary, dashboard row, orientation note, partner-safe version, simplified diagram, or coarse working description. Ordinary controlled simplification remains CSC even when it is lossy.

Action path:

1. Name the source-bearing side and the coarsened version.
2. State the use scope of the coarsened version before stating what it means.
3. State the lost distinctions, evidence paths, comparability, uncertainty, state dimensions, or alternatives.
4. State admissible use and non-admissible use in practical terms.
5. State when to reopen the source-bearing side.
6. If the coarsened rendering claims to preserve action, intervention, manipulation, explanation, or cross-abstraction structure, state the causal-abstraction or approximate-causal-abstraction mapping before treating the shortcut as QL coarsening.
7. Ask whether the shortcut depends on a QL cue such as incompatible probes, contextual probability, instrument-like update, open-information-system update whose update rule, probe frame, or export admissibility is part of the modeling requirement, or no faithful-enough export of the represented state for the admissible use. If not, stay in CSC.
8. If yes, coordinate with the C.26 state-representation coarsening support section while leaving CSC as the controlled-use boundary for the coarsened version.

For ordinary use, start with the standard shortcut mini-form:

Use a fuller CSC and C.26 coarsening boundary record only when the coarsened state representation will be reused, formalized, empirically compared, used in a high-stakes decision, or tied to a comparative performance claim:

Useful outputs:

• a CSC mini-form when the issue is controlled simplification;
• a fuller C.26 coarsening support record only when a QL cue remains and the claim is reusable, formal, empirical, high-stakes, or comparative-performance-bearing;
• no QL wording when the case is only summary, anonymization, diagramming, audience adaptation, or ordinary coarsening.

#C.29 MLA relation

When controlled semantic coarsening depends on mathematical abstraction, quotienting, coarse-graining, or a learned coarse representation, A.6.3.CSC still governs the source-bearing return condition, narrowed admissible use, non-admissible downstream claim, and coarsened rendering. The applicable C.29 output for the stated use (MLA.LensCandidateNote, MLA.OneLine, MLA.MiniCard, or MLA.FullCard when required) may be cited only for adequacy of the mathematical abstraction or coarse-graining lens. It does not make the coarsened rendering a bridge, replacement source, evidence record, or causal-use support.

#A.6.3.CSC:End

---

Last Updated: 2026-05-15 — this section last modified in upstream FPF commit 37a19061 (github.com/ailev/FPF)