A.6.6 - Base Declaration Discipline: Say What Something Depends On
Status: Stable Type: Definitional relation-discipline pattern
Plain-name. Saying what something depends on.
Use this pattern when a sentence says that one thing is calibrated to, based on, attributable to, constrained by, or otherwise usable relative to another, and the relation is still hidden by words such as anchor, support, ground, or based on.
First useful move. Name the dependent and base, state the direct relation in an ordinary sentence, and apply that relation’s own predicate to the current facts. Stop when this readable assertion answers the receiving question.
What goes wrong if missed. An umbrella word hides the relation kind or reverses its participants. At the opposite extreme, a simple assertion is expanded into slots, witnesses, editions, and a new record even though no later use needs them.
What this buys. A direct, testable assertion first. Scope, time, evidence, a reusable RelationSignature, or a reviewable record is added only when the direct predicate or one named receiving use needs that distinction.
Not this pattern when. If the direct relation and its participants are already clear, use its direct pattern. If support means evidence use, assurance, ordinary help, work enablement, navigation, source description, or another non-basedness reading, use that reading’s direct pattern instead.
A.6.6:1 - Problem frame
FPF repeatedly needs to express a family of situations of the form:
The use, admissibility, or interpretation at issue depends on an explicit relation between the dependent and its base.
Examples occur across several disciplines:
- reference selection and identification (IDs, handles, pointers, registries),
- scale/datums/calibration (measurement traceability, baselines, normalisation),
- grounding of properties and abstractions to objects (attribution; “this property is about that thing”),
- admissibility/assurance (claims linked to evidence, checks, or proofs),
- publication discipline (what a statement is fit to be used for, where, and when).
An umbrella metaphor such as anchor can leave different relations and operations indistinguishable. When it hides the relation, recover that relation so its direction, invariants and applicable rules are clear.
This family can expose a typing mismatch: the named dependent or base may not satisfy the participant kind required by the direct relation, or its designation may not satisfy the receiving declaration’s reference mode. Resolve the mismatch under the applicable rule; renaming participants or reversing direction does not fix it. Add SlotSpecs only when a reusable declaration needs them.
Every ordinary basedness assertion first needs only:
- the dependent;
- the base; and
- the direct relation and its obtaining test.
Scope, time, evidence, continuity, or a reusable declaration is added only when the direct predicate or one named receiving use depends on it. Until the direct relation is named, umbrella words such as anchor, ground, attach, support, or based on usually mean only:
“There is an under-described relation here.”
The repair is therefore progressive: recover and test the direct relation, stop if the assertion is enough, and materialize declaration or assertion machinery only for a later use.
A.6.6:2 - Problem
Typical failure modes this pattern is designed to eliminate:
-
Relation-kind elision. One verb phrase is used to cover: ID-to-registry reference, claim-to-evidence admissibility, calibration-to-standard, property-to-object attribution, policy gating, etc. Rules and invariants cannot be stated because the relation kind is unspecified.
-
Perspective flip (dependent-view vs base-view). The same situation is described alternately as “X is anchored/grounded” and “Y is an anchor/ground”, with incompatible naming, hidden directionality, and silent re-typing of the ends.
-
Base–witness confusion. Evidence, pins, certificates, or proofs are treated as “the base”, even when they are only witnesses for a base relation (or conversely: a base is treated as a mere witness).
-
Scope/time collapse. A time-dependent assertion leaves its interval or temporal rule unclear, so readers cannot tell which case it covers.
-
Γ_timeused as a proxy for freshness. Authors treatΓ_timeas “freshness” or “evidence decay”, collapsing TimePolicy with witness-timespan/freshness predicates. -
Decision use without its required basis. Declarations that gate work, publication, or assurance omit the basis required by that decision. A witness, pin, or time selector is required only when that use needs it; omitting the required basis prevents its audit.
-
Grounding conflation. “Grounding” is used as if it were one relation, while FPF already distinguishes at least:
- constructive grounding of a model-edge by a trace (
tv:groundedBy), - situational/empirical grounding of an episteme via a grounding holon (C.2.1),
- source-local meaning recovery and, when needed, an F.17
SchemeSenseCellandLocalSenseBasisRelation(not a base declaration).
- constructive grounding of a model-edge by a trace (
-
Slot/basing conflation. A.6.5 separates a declaration-local SlotSpec, a designation carried in an episteme and the actual participant. Editing a designation does not itself change the obtaining relation.
-
Anchor relapse (source or meaning surrogate). “Anchor/anchoring” is used to mean “the source”, “the meaning”, “the global reference”, or “the thing that makes this true”. This hides the source, scheme, expression, local claim, and any obtaining basis relation behind a metaphor; the phrase alone does not let a reviewer recover them.
-
Support bucket relapse. “Support”, “support basis”, “support relation”, or “support record” is used as a generic container for unlike relations. Some cases are direct base-dependence; others are evidence use, assurance input, causal-use support basis, mathematical-lens use, work enablement, source description, publication companionship, or ordinary help. Treating them as one support relation recreates the under-described dependence that A.6.6 is meant to repair.
A.6.6:3 - Forces
| Force | Tension |
|---|---|
| Universality vs precision | One discipline must cover calibration, evidence linking, reference selection, attribution, gating, etc., without collapsing them into one pseudo-relation. |
| Minimal kernel vs decision auditability | Few primitives are preferred, but decision-relevant declarations must expose the basis required by the decision, including witnesses, pins, or explicit time selectors when that use needs them. |
| Two perspectives, one reality | Dependent-view and base-view must both be expressible without renaming relation-end meanings or flipping meaning. |
| Compatibility with A.6.5 | When a reusable base-relation declaration needs SlotSpecs, keep SlotKind, ValueKind, and RefKind distinct and do not collapse designation edits with changes to a participant or relation. |
| Lexical guardrails | Umbrella metaphors can hide participants, direction, and the applicable relation rule; the wording must recover those values. |
| Cross-local integrity | When a declaration depends on a relation between different local kinds, local senses, scopes, or planes, that relation must remain explicit and reviewable; different sources alone do not create a Bridge. |
A.6.6:4 - Solution - State the direct relation, then add only what the receiving use needs
A.6.6:4.0 - Ordinary direct path
Start with a readable sentence:
Thermocouple channel TC-17 is calibrated to standard ITS-90 for rig R3.
Identify TC-17 and ITS-90, then apply the direct calibratedTo predicate and its applicability rule to the current facts. If the task only asks whether that calibration relation obtains for this rig, the sentence and predicate result are complete. Do not create a declaration record, witness set, edition, or assurance package merely because those fields could be written down.
Add a qualifier only when it changes the direct assertion or a named receiving use:
- name scope when the relation is limited to a range, population, rig, publication, or other extent;
- name time when the predicate or the use is time-dependent;
- cite an evidence-use or provenance relation when a claim about the relation is relied on;
- open occurrence identity only when another claim must refer to the same occurrence, compare it, qualify it, or record its history; and
- open a reusable declaration only when the reuse test in A.6.6:4.3 is met.
The assertion episteme, reusable declaration, world-side relation occurrence, evidence, and any Work remain different objects.
A.6.6:4.1 - Optional scoped assertion record
When replay, comparison, publication, or repeated review needs a stable representation, a project may show one C.2.1 assertion episteme in this local form:
scoped witnessed base declaration :=
< dependent,
base,
directRelationKind,
assertionPolarity,
scope?,
gammaTime?,
evidenceUseRefs? >
This is a representation of claim content, not a public kind, RelationSignature, or world-side occurrence. directRelationKind resolves to an already governed relation kind; an affirmative assertion requires that relation’s predicate to obtain for the actual participants, while a negative assertion requires nonobtaining to be established under the applicable criterion or closure basis. Failure to establish the affirmative does not establish the negative. scope and gammaTime are present only when the direct relation or named use needs them. evidenceUseRefs, when present, resolve to A.2.4 evidence-use relations for this assertion. The evidence epistemes, producing Work, operation result, carrier, provenance, currentness, and later reliance remain separately identified under A.2.4 and A.10.
A.6.6 admits neither U.BaseDeclarationDiscipline nor U.ScopedWitnessedBaseDeclaration. The latter is a retired spelling and must not be used as a kind or as a world-side relation occurrence.
The record’s C.2.1 identity follows its complete ClaimGraph, EntityOfConcern, and effective ReferenceScheme. Changing an identity-bearing value identifies another episteme; a representation-only edit need not. Neither change by itself begins, ends, or alters the world-side relation it describes.
A.6.6:4.2 - Direct relation and optional assertion are different objects
The picture is a direct arrow in ordinary reading:
dependent stands in the named direct relation to base.
The arrow is not a generic mathematical constructor. Its participant meanings, predicate, applicability, and occurrence identity come from the selected direct relation pattern. A scoped assertion episteme may affirm or deny that this predicate holds, and evidence may support reliance on that assertion. Filling the optional record or merely asserting the claim establishes no occurrence. An episteme or publication may have a constitutive role under the direct relation’s own rule; evidential support remains independently governed.
Calibration, attribution, policy dependence, constructive grounding, and other cases therefore remain different relation kinds. A.6.6 supplies a recovery discipline, not one universal BaseRelation kind.
A.6.6:4.3 - Reusable declaration only for a named reuse
Use the direct relation’s A.6.0 RelationSignature only after its kind is admitted and a named receiving use needs reusable participant typing or declaration content. An observed second consumer is not required; an anticipated use can justify the declaration. The signature restates the direct relation kind’s participant meanings, predicate, applicability and occurrence-identity rule for reuse. A.6.5 SlotSpecs belong inside it; an ordinary one-case assertion may need no signature or explicit occurrence individuation.
If no direct pattern supplies the relation kind, participants, or predicate, keep the local claim or return the A.6.RCD missing-governor result. Do not repair the gap by minting a generic BaseRelation kind or token, SlotSpecs, or a scoped-record type.
A.6.6:4.4 - What a reusable direct-relation declaration must say
For a named receiving use that needs a RelationSignature, the direct relation definition states:
- the dependent and base participant meanings and direction or symmetry;
- the obtaining predicate and applicability;
- the occurrence-identity rule, without requiring an occurrence to be explicitly individuated;
- admissible participant kinds and reference modes;
- any scope, time, evidence, or cross-local condition that changes this predicate or the named reuse; and
- the direct continuity or change rules, when that history is current.
Different local kinds, F.17 senses, scopes, or ReferencePlanes are handled by their applicable direct relations. Source difference alone creates no Bridge. A RelationSignature declares reusable content; the reusable form by itself neither asserts a current case nor establishes an occurrence.
A.6.6:4.4a - Claim-scoped non-kind predicate-base branch
When one identified derivation or criterion-selection claim uses claim content as its base, reuse A.6.6’s endpoint, scope, time, witness, Bridge/loss, change, and overread discipline without claiming that a new relation kind or special base-declaration occurrence has been admitted. Identify the dependent ClaimGraph and selected nonempty base subgraph as by-value NonemptyClaimContent values under C.2.1:4.1.2. State the derive or evaluate mode, derivation or evaluation-and-selection claim identity, bounded receiving use, and effective reference scheme. Add an A.2.6 ClaimScope, temporal policy/domain, source or witness qualification, or cross-scheme Bridge and loss account only when that independently varying fact changes the assertion.
The assertion is ordinary C.2.1 claim content under derivedUsingRuleContent or evaluatedAgainstRuleContent. The dependent and base are predicate parameters, not automatically A.6.5 SlotSpecs, participants of a reusable relation occurrence, or an intrinsic rule-bearing classification. Same-scheme use adds no Bridge. A source edition, designation, acceptance/currentness fact, trace, or witness qualifies the assertion but does not enter semantic-base identity. Equal graphs under the same scheme count as one semantic base with multiple qualifications; a changed graph is another base.
Change only the fact that changed: declare or withdraw a selected base, repoint the dependent, rescope, retime, refresh witnesses, or change the predicate relation. A changed subject, content, mode, bounded use, scope extension, temporal policy, or interpreted endpoint identifies another C.2.1 assertion. A claimed edition succession additionally requires the continuity conditions of C.2.1:4.5. Do not infer a new relation kind, occurrence, evidence result, Work, authority, or reliance from that change.
A basis-family analysis is a separate, optional C.2.1 episteme opened only for a named comparison, replay, material-conflict, or reliance receiver. Its candidate universe, evaluations, pairwise compatibility, temporal partition, established family, and disposition neither edit this reusable predicate declaration nor become fields of each derivation or criterion-selection assertion. A separate claim about actual use in Work follows its own rule; neither R7 predicate establishes that use.
A.6.6:4.4.1 - Perspective and voice
State the relation in the shortest ordinary sentence that keeps both participants and direction recoverable: TC-17 is calibrated to ITS-90 is valid. Functional or arrow notation may be added when it helps a formal receiver; it is not the default. Base-view wording is also valid when it preserves the same relation and direction. Do not turn B validates X into an inverse relation unless that inverse is independently defined.
A.6.6:4.5 - Lexical discipline
Normative lexical rule. In Tech or normative prose, do not use umbrella metaphors (anchor, attach, ground, or support) in place of the relation. Prefer an ordinary relation-specific sentence; add functional or arrow notation only when a named receiver benefits from it.
Red-flag rule (anchor* as dependence metaphor).
- In Tech or normative prose, repair
anchor*when it leaves the relation or participants unresolved. Keep an adequate ordinary or already governed use. - Source wording may remain for recognition or traceability alongside its recovered meaning. It cannot supply a missing rule merely by being renamed
validatedBy,verifiedBy,SupportRelation, or another formal-looking token.
Pattern-defined meanings. E.10 MG-DA token-to-EntityOfConcern anchoring and A.10 evidence anchors retain their own definitions. Apply the rule when that is the intended sense; do not open another basedness repair solely because the word appears.
Naming guard for relation vocabulary. Do not mint a new direct relation whose name merely preserves a metaphor such as Anchor*, Ground*, or Attach*. Name the relation kind and use the corresponding ordinary verb phrase. In an optional assertion record, the local directRelationKind field identifies that already admitted relation kind; the field is not another relation kind.
Lane guard for meaning. If the intent is “say what this expression means in this source”, do not introduce an Anchor… or Ground… relation. Recover the source-local claim under F.0.1; use F.17 only when a durable SchemeSenseCell or obtaining LocalSenseBasisRelation is needed. Semantic meaning assignment is not a base-declaration record.
Grounding disambiguation. When grounded leaves a consequential meaning unresolved, recover it. These branches illustrate distinct meanings, not an exhaustive classification:
- constructive grounding (
tv:groundedBy, base is a trace), - situational/empirical grounding (base is a grounding holon or experimental setup),
- source-local meaning lane (source, scheme, expression, local claim, and optional F.17 cell or basis relation; no special base-declaration object).
Bind deconfliction note. Do not use “bind/binding” as a synonym for declaring, refreshing, or changing an assertion or reusable relation declaration. This local edit vocabulary does not rename name binding or other already governed binding relations, including A.6.1 application bindings. Use the local declaration-change label only when a named receiver needs that history.
A.6.6:4.6 - Base-change operation lexicon
The following local labels classify changes to an optional assertion episteme or reusable declaration when a named receiver needs that history. They do not describe the beginning, ending, or change of the world-side relation itself, and an ordinary direct assertion needs none of them. In decision or publication use, preserve the prior episteme when changing its ClaimGraph, EntityOfConcern, or effective ReferenceScheme; such a change identifies another episteme. A representation-only edit need not do so. Claim edition continuity only when the C.2.1:4.5 rule and case facts establish it.
Operation classes (conceptual):
- declareBase - create a new optional assertion with explicit
dependent,base,directRelationKind, andassertionPolarity, or a new reusable declaration for that same already governed direct relation kind; add only the scope, time, evidence-use, or other qualifications that its direct predicate or named receiver needs. - withdrawBaseDecl — retire an assertion or declaration (or render it inapplicable by scope narrowing or time restriction, depending on the direct relation’s declaration).
- rebase — change
basewhile keeping the samedependentanddirectRelationKind(legality depends on the direct relation’s declaration; often requires witness refresh). - repointDependent — change
dependentwhile keeping the samebaseanddirectRelationKind. - rescope — change
scope(widen/narrow/translate) under the direct relation’s scope rule; widening often triggers witness refresh. - retime — change
Γ_timeselector/policy when time matters; not a substitute for witness-timespan/freshness predicates. - refreshWitnesses — add/refresh witnesses/pins when decision use continues across time advances, scope widening, or evidence refresh.
- changeDirectRelationKind — not an edit-in-place. Changing
directRelationKindchanges claim meaning; mint a new assertion or declaration rather than silently rewriting the kind. When edition history is needed, relate it to the prior episteme only if C.2.1:4.5’s continuity rule and case facts establish that relation. Use F.13 for a separately current lexical-continuity claim.
Relation to A.6.5 designation operations (informative). Apply A.6.5 when the changed assertion designation is governed by a compatible RelationSignature SlotSpec. Other assertion or declaration fields retain their own C.2.1 or direct definition. State which content changed. A separately claimed world-side change uses the direct relation’s rule; ordinary editing requires no invented Work occurrence.
Relation to E.18 assurance ops (informative). E.18:5.4 applies ConstrainTo, CalibrateTo, CiteEvidence and AttributeTo to the annotation of a resolved connection(q), retaining the relation r and any participant subject the operation requires. Each operation keeps its own meaning and basis. A project may use the local declaration-change labels for changes in a represented assertion; those labels neither perform an assurance operation nor establish a subject relation.
A.6.6:4.7 - Disambiguation guide for selecting the direct relation
When umbrella wording leaves the relation unclear, recover the direct meaning that fits the claim:
| Colloquial intent | Direct relation or reading (illustrative) | Participants and conditions | Typical supporting material, when needed |
|---|---|---|---|
| “This ID refers to that thing” | Identification (identifies) | For “ID I identifies entity E”, name I and E. If the source means an entity-ref or slot-content value instead, name that referent. | issuance record, registry pin |
| “This thing is indexed by that ID” | Indexing (indexedBy) | Entity E is indexed by ID I; apply the indexing rule. | issuance record, registry pin |
| “This thing is registered” | Registration (registeredIn) | Recover the participant registered, its registry or registry entry, and the domain’s registration predicate and direction. | issuance record, registry pin |
| “Make measurements comparable by calibration” | Calibration (calibratedTo) | Name the instrument, model, or output said to be calibrated and the standard or datum; the domain’s calibration rule must determine their roles. | calibration Work plus certificate pin |
| “This is the datum of that” | Datum relation (datumOf) | Recover what is the datum of what, and the domain rule; datumOf does not supply the calibration or normalisation predicate. | the domain-required basis; calibration Work or certificate pin only if that rule needs it |
| “Make measurements comparable by normalisation” | Normalisation (normalisedTo) | Name what is normalised—an instrument, model, or output—and the standard or datum used; recover the normalisation predicate and direction from its domain rule. | the domain-required basis; calibration Work or certificate pin only if that rule needs it |
| “This result bears on that claim” | Evidence use under A.2.4, with A.10 only when replayable provenance or reliance is needed | result or other evidence episteme and target claim, with the evidence-use direction defined by A.2.4 | evidence-use relation; producing Work, result binding, carrier, provenance, currentness, and reliance remain separate |
| “This edge is grounded in construction” | Constructive grounding (tv:groundedBy) | dependent: WM edge; base: constructor trace (Γ_m) | trace pins, edition pins |
| “This description is about X” | Ordinary aboutness under A.7/C.2.1 | description episteme and its EntityOfConcern X; aboutness alone does not establish a source-to-receiving construction | the source or describing relation required by the use |
| “Construct a view of the same entity” | Viewing arrow under A.6.3 | separately identified source and receiving epistemes, the same EntityOfConcern, and the construction rule | viewing pins |
| “Retarget a description to another entity” | Retargeting arrow under A.6.4 | separately identified source and receiving epistemes with different EntitiesOfConcern, and the retargeting arrow r | r; a separate assertion q for a named receiving use, and a current-case judgement when that case is evaluated |
| “Allowed only under policy P” | Constraint / policy (constrainedBy, permittedUnder) | dependent: work-step / publication item; base: policy/rule | policy pin, waiver/work ref |
| “Property belongs to object” | Attribution / aboutness (attributedTo, aboutEntity, characterises) | dependent: property/abstraction; base: object | observation/derivation witnesses |
| “This expression means … in this source” | Source-local meaning lane (F.0.1; F.17 only when a durable address or basis relation is needed) | local expression and local-sense claim | source passage and, when current, an obtaining basis relation |
This table is illustrative. Each row keeps its own direct relation and governor; it is not a list of species of one universal base relation or record. Grammatical subject and object expose a sentence’s direction but do not by themselves determine dependent/base allocation. Where the domain predicate or participant allocation is not supplied, recover it from the source or leave that local choice open; this guide does not define it. Ordinary aboutness and source-local meaning do not by themselves select a basedness or construction branch.
Note. A.6.3 defines the viewing arrow. A.6.4 distinguishes the retargeting arrow r from a bounded-use assertion q and a current-case judgement. When a named receiving use is claimed, q’s ClaimGraph contains that use, its invariant, visible loss, conditions, and affirmative or negative polarity. When a case is evaluated, the judgement compares its facts with q and returns exactly satisfies, fails, or cannot decide. A cannot decide judgement names the missing fact and reopen condition. This table directs those construction cases to their own patterns; it defines no second operator, arrow, assertion, judgement, application, or Work.
A.6.6:4.7a - Support wording selection test
When a draft uses support, supported by, supporting, support basis, support relation, or a support-headed compound, do not first choose a more formal synonym. Ask what assertion the next reader needs.
If the sentence is about basedness, write the smallest direct form:
dependent stands in <direct relation> to base
Identify the participants and apply that direct predicate. Stop there when it answers the use. Add scope, time, an assertion record, a reusable RelationSignature, occurrence identity, or evidence only when the predicate or one named receiver needs it.
If the sentence is not basedness, use the matching ontology:
| Support wording means… | Use… |
|---|---|
| an episteme bears on a claim | the A.2.4 evidence-use relation; use A.10 when provenance, currentness, rival explanations, or bounded reliance must be replayed |
| a claim is acceptable for material reliance | A.10 for the bounded-reliance basis; B.3 only for a separately identified assurance claim, with the evidence-use relations kept separate |
| a causal, intervention, counterfactual, or simulation-only use is admissible | C.28 |
| a mathematical lens exposes preserved or lost structure | C.29 for that lens; C.26 only for its quantum-like/contextual-model case; F.9 only for a separately obtaining correspondence between local sense cells; the direct mathematical pattern for the mathematical object or rule |
| one thing helps or enables work | the work, resource, capability, or action relation, or ordinary Plain help |
| a file, section, packet, or companion helps a reader | E.17, E.11, or ordinary orientation |
| a source, model, diagram, or view describes something | A.7, C.2.1, E.17, and the direct describing or source-use relation |
Do not create SupportRelation, SupportBasis, SupportRecord, validatedBy, or verifiedBy as a fallback. Work, a result episteme, its carrier, provenance, evidence use, and later reliance remain separate.
A.6.6:5 - Archetypal Grounding
A.6.6:5.1 - System archetype: calibration to a standard
Tell. A lab instrument channel TC‑17 is described as “anchored to ITS‑90”. Later, the reference standard is swapped, the phrase “still anchored” is kept, and the applicability window silently expands. Downstream work disagrees and nobody can reconstruct what changed.
Show. First state the direct assertion: TC-17 is calibrated to ITS-90 for rig R3 over 0–200 °C during the stated calibration interval. Apply the calibration predicate and stop there if this answers the use. When a later publication or comparison needs the assertion edition, show the same claim in an optional scoped record:
BD#Calib_TC17_v5 :=
〈 dependent = ThermocoupleChannelRef(TC-17),
base = StandardRef(ITS-90 / CalStd-2025-09),
directRelationKind = calibratedTo,
assertionPolarity = affirmative,
scope = WorkScope{rig=R3, range=[0..200]°C},
gammaTime = interval[2025-09-01, 2026-03-01] 〉
When a later decision relies on this assertion, identify the calibration-certificate episteme and its A.2.4 evidence-use relation to the assertion. Use A.10 when the decision needs a bounded-reliance judgment or the producing Work, operation result, carrier, provenance or currentness path. Distinguish changes by what changed:
- New standard ⇒ rebase; establish the evidence-use relation required for the new assertion, reusing suitable witnesses or refreshing them when needed.
- Wider applicability window ⇒ retime and likely refreshWitnesses.
- Relation-kind change (“not calibration, just normalisation”) ⇒ changeDirectRelationKind is not an edit; mint a new assertion or declaration. Claim continuity with the prior episteme only when that history is needed and C.2.1:4.5’s rule and case facts establish it.
A.6.6:5.2 - Episteme archetype: an evaluation result used as evidence
Tell. A report says that model M improved accuracy by 4%. The team points to EvalRun-2025-10-12, but that Work occurrence is neither the claim nor an evidence relation, and its log carrier does not become evidence merely by being attached.
Show. First recover the accuracy measure, reference condition or population, and percentage convention from the reported comparison. If those operands cannot be recovered, the 4% comparison claim remains unidentified; stop before relying on that numerical target. Once they are recovered, identify the result episteme that states the measured comparison and the target claim about the 4% improvement. State the A.2.4 evidence-use relation between that episteme and claim, including the ClaimScope, polarity, window, and receiving use. If the decision also needs replayable source, carrier, provenance, currentness, or bounded-reliance information, use A.10 to cite the evaluation Work, its operation-result binding, the result episteme, the log carrier, and their independently obtaining direct relations.
Stop with the short evidence-use statement when it answers the question. No validatedBy(claim, Work) edge or scoped base-declaration record is required. If a project later needs a reusable evidence-relation declaration, that direct relation must first have its own participant meanings, predicate, applicability, and occurrence-identity rule.
A.6.6:5.3 - Structural archetype: constructive grounding of a model edge
Tell. Suppose a B.3.5 assurance requirement for this publication requires a recoverable constructor trace. A structural edge is published (“A componentOf B”) without that trace. It becomes treated as “obvious”, while the construction chain is not recoverable. Ordinary parthood does not require this trace; the failure here concerns the selected hypothetical assurance requirement.
Show. Locate the required constructor trace and test the direct tv:groundedBy predicate for the model edge and trace. If the trace remains unavailable, retain that gap and stop the dependent assurance claim. For illustration, suppose ComposeCAL#c17 is recovered and the case facts satisfy the grounding predicate for Edge:componentOf#17. The short affirmative assertion may be enough; a publication needing a stable assertion edition with its qualifiers can represent it as:
BD#EdgeGrounding_ComponentOf_17 :=
〈 dependent = WMEdgeRef(Edge:componentOf#17),
base = TraceRef(Γ_m:ComposeCAL#c17),
directRelationKind = tv:groundedBy,
assertionPolarity = affirmative,
scope = PublicationScope{view=WMCardLite, system=S, line=L3},
gammaTime = snapshot(2025-11-02) 〉
The printed trace reference names the intended constructor trace; its presence does not establish the trace or the grounding relation. If another use relies on the assertion that the grounding relation obtains, cite its evidence-use and provenance relations separately. This example shows why “grounding” must be disambiguated: here it is a declared constructive relation with an explicit base (trace).
A.6.6:6 - Bias-Annotation
| Lens | Bias introduced by this pattern |
|---|---|
| Governance / assurance | Prefers explicit witnesses and time selectors when the decision’s basis requires them; increases auditability but adds authoring overhead. |
| Architecture | Prefers the direct assertion and predicate first. It permits a reusable declaration or scoped assertion record only for a named receiver, reducing both hidden relations and record-first over-formalization. |
| Onto-epistemic | Makes the relation kind and direct predicate explicit; resists both metaphor-only wording and a universal base-relation kind. |
| Didactic | Teaches the short dependent–base–direct-relation question first; the optional record vocabulary appears only for a named later use. |
A.6.6:7 - Conformance Checklist
An A.6.6 use conforms when the checks for its selected branch pass:
- CC-BD-1 - Direct assertion first. The dependent, base, direct relation, and readable assertion or unresolved factual question are recoverable. The direct pattern supplies the predicate. Test affirmative obtaining or establish nonobtaining under its applicable criterion or closure basis; a record, label, or failure to establish the affirmative supplies no case result. Missing case facts leave the question unknown; only a missing governing rule requires A.6.RCD.
- CC-BD-2 - Ordinary stop. If that assertion answers the receiving question, no SlotSpecs, declaration record, witnesses, edition, occurrence identity, or assurance package is required.
- CC-BD-3 - Reusable declaration is demand-driven. A
RelationSignaturesatisfies the reuse test in A.6.6:4.3 and applies only to an already admitted relation kind. - CC-BD-4 - Assertion and occurrence stay separate. A scoped witnessed record, when used, is a C.2.1 assertion or description episteme. It is not the world-side relation occurrence, and its form alone does not establish that occurrence; any constitutive contribution follows the direct relation’s rule.
- CC-BD-5 - Qualifiers are local. Scope and time are explicit when the selected predicate or named receiving use depends on them; they are not a universal field kit.
Gamma_timeis not used as a proxy for evidence freshness. - CC-BD-6 - Evidence ontology is direct. Evidence use follows A.2.4 and A.10. Work, operation result, result episteme, carrier, provenance, evidence-use relation, and reliance remain separate; no generic
verifiedByorvalidatedByedge is minted. - CC-BD-7 - Crossings are conditional. A relation between two F.17 cells uses F.9 only when its predicate obtains and keeps the bounded-use claim separate. A ReferencePlane crossing uses its applicable plane relation. One creates neither the other.
- CC-BD-8 - No silent retyping or direction flip. Participant kinds and direction follow the direct relation. A mismatch is repaired by the applicable narrowing, Bridge, retargeting, or direct relation rule, not by renaming an endpoint.
- CC-BD-9 - Plain language remains sufficient. Ordinary relation-specific prose is preferred. Functional or arrow notation is optional and may not replace the readable assertion.
- CC-BD-10 - Wording does not create ontology. Recover the meaning of anchor, ground, attach or support when it is unresolved. Keep an adequate ordinary or governed meaning; a metaphor or new label cannot supply an absent relation rule.
- CC-BD-11 - Meaning lane stays separate. Source-local meaning starts with F.0.1 and uses F.17 only when a durable sense address or basis relation is needed; it is not a base-declaration record.
- CC-BD-12 - Change claims name the changed object. Changing an assertion’s or reusable declaration’s ClaimGraph, EntityOfConcern, or effective ReferenceScheme identifies another episteme; representation-only editing need not. An actual relation change requires the direct relation’s own change predicate and any separately current Work.
- CC-BD-13 - Optional history is proportional.
declareBase,rebase,rescope,retime, orrefreshWitnessesis used only when a named receiver needs that declaration history. The label establishes no world-side fact.
A.6.6:8 - Common Anti-Patterns and How to Avoid Them
| Anti-pattern | Why it fails | Repair |
|---|---|---|
| Generic support bucket | Hides whether support means basedness, evidence use, assurance, work enablement, navigation, source description, or ordinary help | Apply the support wording selection test; state the direct relation or keep ordinary help instead of minting a support-headed relation or record |
| Umbrella “anchored/attached/grounded” with no direct relation | Hides relation kind and predicate | Name the participants, use the relation-specific verb, and apply its direct predicate |
| Perspective flip without recoverable participants | Direction and typing become ambiguous | Keep the same participants and direction in both active and passive wording; add formal endpoint names only when reused |
| Work or carrier treated as evidence relation | Collapses producing Work, result episteme, carrier, provenance, evidence use, and reliance | State the A.2.4 evidence-use relation; open A.10 only for the replayable provenance or reliance path |
| Unclear temporal applicability | Current or latest leaves the intended time-dependent claim ambiguous | State the interval or temporal rule, using explicit Γ_time when the governing use needs it. Apply the evidence-use or reliance rule’s freshness check when that use requires it. |
| Decision use without its basis | A relied-on assertion cannot be checked | Cite the evidence-use, provenance, currentness, or assurance relations required by that decision; do not add a generic witness field or new document |
| Semantic meaning expressed as basedness | Confuses source-local meaning with another relation | Recover the source-local claim under F.0.1 and add an F.17 cell or basis relation only when needed |
| Relation-kind change presented as an edit | A semantic shift masquerades as continuity | State the new direct relation and use the continuity rule when that history matters |
Using *Slot for a participant or designation | Conflates declaration content, participant and reference | Reserve ...Slot for a declaration-local SlotKind. Name the dependent and base directly, distinguishing their designations and any reference values under A.6.5. |
| Optional record field treated as a carrier or free-text kind | Lets a record label stand in for the direct relation | Make the field identify the already admitted relation vocabulary entry; keep the assertion, carrier, and relation occurrence separate |
A.6.6:9 - Consequences
Benefits
- A readable direct assertion can close an ordinary question without record-first work.
- Reusable declarations are available when a receiving use needs their participant typing and laws.
- Scope, time, evidence, and continuity are explicit exactly where they change a predicate or receiving use.
- Evidence use, source-local meaning, semantic Bridges, plane crossings, and ordinary help keep their own ontologies.
Trade-offs and mitigation
- The author must identify the direct relation instead of hiding it behind support or anchor. The mitigation is one ordinary sentence and the direct predicate, not a universal form.
- A later replay or publication use may require more detail. Add only the missing declaration, qualifier, assertion, evidence, or occurrence identity at that point.
Adoption test (informative). A team has adopted A.6.6 when it can answer three questions in order: What are the participants and direct relation? Does its predicate hold for this case? Does a named later use require a reusable declaration, occurrence identity, scope, time, evidence, or history? A negative answer to the third question is a valid stop.
A.6.6:10 - Rationale
Why focus on base declaration rather than a metaphor. The recurring ambiguity is not “how to attach”, but which direct relation is being asserted between which participants. A readable relation-specific sentence exposes that answer; an optional declaration can then preserve it for a named reuse.
Why keep the direct relation, assertion, and evidence separate. The relation’s predicate determines whether the world-side fact obtains. A C.2.1 episteme may affirm or deny it under the polarity test, and A.2.4/A.10 may support reliance on that assertion. Conflating these objects lets a record or carrier stand in for truth. A base is a participant in the selected direct relation. Evidence or other supporting material justifies an assertion only through its own direct relations. Conflating the two makes both reasoning and audit unreliable.
Why add scope and Gamma_time conditionally. They are required when the direct predicate or receiving use changes across extent or time. Adding them everywhere hides the ordinary relation behind a universal qualifier form.
Recover extent and time from the direct relation and the receiving use. An already clear interval or temporal rule needs no duplicate field; an implicit current or latest that leaves the intended case ambiguous does.
Why distinguish semantic revision from representation editing. Changing the relation kind changes the assertion’s meaning. Preserve the former episteme when history matters and establish any claimed continuity under C.2.1; a new wording or representation alone need not change identity.
Why retain a local declaration-change lexicon. When a named receiver tracks assertion or declaration history, the labels distinguish which episteme field changed. They are optional and do not describe actual relation change without the direct relation’s own predicate. When that history is needed, explicit change classes distinguish rebase, retime, rescope, and witness refresh; a generic edit label hides those differences and recreates the ambiguity A.6.5 removed at the slot layer.
A.6.6:11 - SoTA-Echoing
-
RDF-star and statement qualification. Adopt/Adapt. RDF-star/SPARQL-star explored attaching qualifiers/provenance to statements and edges. RDF 1.2’s Candidate Recommendation Snapshot of 7 April 2026 distinguishes representing a proposition from asserting that it holds. We adopt the “qualified statement” intuition, but adapt it by requiring an explicit relation kind and by making
Γ_timeand USM scopes explicit when the direct relation or receiving use needs them. Primary sources: Hartig and Thompson, Foundations of an Alternative Approach to Reification in RDF (first submitted 2014), retained as history and marked obsolete by its authors; RDF 1.2 Concepts and Abstract Data Model, Candidate Recommendation Snapshot, 7 April 2026. -
Wikidata-style statements with qualifiers and references. Adopt/Adapt. Wikidata statements separate a core statement from optional qualifiers and references. We adopt that separation and adapt it by making decision-relevant basis requirements explicit through evidence-use relations, with slots only for a reused declaration, and explicit scope/time where the assertion or time-dependent use needs them. Primary source: Wikidata, Help:Statements.
-
Metrology traceability and calibration competence. Adopt/Adapt. Calibration is an operation relating a standard’s quantity values and uncertainties to indications, then using that information to obtain measurement results. Metrological traceability is a property of a measurement result related to a reference through a documented calibration chain. We retain the need for documented calibration evidence and adapt time-dependent applicability through explicit
Γ_timeand the witnesses or pinned calibration records required by the assertion or receiving use. Primary sources: JCGM VIM, calibration (2.39) and metrological traceability (2.41); ISO/IEC 17025:2017, edition confirmed in 2023. -
Assurance case metamodels for claim–evidence structure. Adopt/Adapt. SACM formalises claim/evidence structures and emphasises structured support relations. We adopt the idea that decision-relevant admissibility links should be explicit, and adapt it by using FPF’s scope/time discipline and by treating relation-kind elision as a first-order defect. Primary source: OMG Structured Assurance Case Metamodel (SACM), version 2.3, October 2023.
-
Objects over a base as a mathematical lens. Adopt/Adapt. Category-theory treatments of objects over a base, including slice categories, offer a mathematical way to compare an X-relative-to-B construction. Use that lens only when the selected mathematical construction and its correspondence to the direct domain relation are established under C.29. It supplies no universal obtaining predicate, scope, time or witness rule for calibration, evidence, attribution and the other distinct domain relations. Primary source: Riehl, Category Theory in Context (2016).
A.6.6:12 - Relations
Specialises A.6.P Relational Precision Restoration. A.6.6 handles basedness wording by recovering the dependent, base, and direct relation, then stopping or opening only the additional object required by a named use.
Builds on A.6.REL and A.6.0. The direct pattern supplies relation obtaining and occurrence identity. A reusable RelationSignature is justified only for an already admitted relation kind and shared declaration content; the declaration form alone establishes no occurrence.
Builds on A.6.5 only when reusable declaration content is current. A SlotSpec declares a participant meaning, ValueKind and designation mode inside one RelationSignature. An ordinary assertion needs no SlotSpec record unless its receiving use requires that declaration.
Coordinates with A.2.4 and A.10. A.2.4 states the evidence-use relation. A.10 represents the independently established sources, Work, result epistemes, carriers, provenance, currentness, and later-use relations needed for bounded reliance. Neither pattern admits generic verifiedBy or validatedBy edges, and Work is not an evidence carrier.
Coordinates with A.14 and C.2.1. Constructive grounding and empirical grounding retain their direct predicates and participants. Their assertion epistemes and evidence remain separate from the world-side relations.
Coordinates with A.6.3 and A.6.4. A.6.3’s viewing construction separately identifies the source and receiving endpoints with the same EntityOfConcern; the endpoints may identify the same episteme in identity viewing. A.6.4 retargets between different EntitiesOfConcern; its retargeting arrow r, bounded-use assertion q, current-case judgement, operation application, and Work remain distinct. Ordinary aboutness alone selects neither construction. A.6.6 adds no second arrow or universal relative-to object.
Coordinates with F.9 and ReferencePlane rules conditionally. F.9 applies only to an obtaining Bridge between two F.17 cells and keeps its bounded-use claim separate. A ReferencePlane crossing uses its plane relation. If both are current, state both; if only one is current, introduce no object from the other branch.
Coordinates with A.2.6, A.7, and E.24.UK: A.2.6 governs scope and Γ_time when the assertion or receiving use depends on them. A.7 distinguishes EntityOfConcern, description, specification and publication uses. E.24.UK supplies the U-kind admission boundary; an optional local record form admits no new kind.
Coordinates with C.3.3 and E.18. C.3.3 supplies U.KindBridge only for an independently established cross-local kind correspondence, with its kinds, direction, declared CL^k, and use/loss conditions; different participant kinds alone establish no KindBridge or silent re-typing. E.18 retains the meanings and subject resolution of its assurance operations on connection(q), including CalibrateTo, CiteEvidence, AttributeTo, and ConstrainTo; A.6.6’s declaration-change labels do not replace them.
Coordinates with E.8, F.0.1, F.15, and F.17: E.8 governs pattern-authoring order and SoTA discipline. F.0.1 recovers source-local meaning; F.17 supplies an optional durable sense address or basis relation. F.15 supplies the carrier/source-currentness, provenance, and refresh validation harness when that validation is current.
Feeds E.10, E.10.D1, and F.18 lexical governance. Umbrella words trigger recovery of the direct relation. Ordinary relation-specific prose remains valid; notation and durable public names are added only for a named use.