#Relational Precision Restoration — Recovering Direct Relations from Under-Specified Claims

A.6.P · 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 Reference 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

Plain name. Relation precision restoration.

Mint or reuse. This pattern reuses direct relation kinds, direct obtaining predicates, relation-participant meanings, RelationSignature, SlotSpec, U.Relation, U.Episteme, designators, references, descriptions, publications, and representations from their governing patterns. It introduces no U-kind, universal record-shaped relation object, qualification object, or generic relation-change object. A RelationKind token designates an already settled relation kind in a local or public vocabulary; the token is neither the kind nor an occurrence.

Plain object stack. A direct relation is what obtains among its actual participants under the participant meanings and obtaining condition stated by its direct pattern. Each participant keeps its independently governed kind. A compatible RelationSignature is a declaration episteme; one declaration-local SlotSpec can correspond to one participant meaning when reusable typed use is current. An assertion or occurrence-description episteme may designate the participants or an already recoverable occurrence. A table row, tuple, record, graph edge, functional expression, or arrow is a representation only through an explicit C.29 correspondence. None of those epistemic or representational objects makes the relation obtain or supplies occurrence identity by form.

Use this when. Use this pattern when a claim contains a relation-bearing phrase, but the phrase does not yet determine the direct relation, exact participants, direction, or detail needed by a later engineering claim or operation. Common recognition moments include a broad predicate such as "linked", "aligned", or "supports"; a participant named by metonymy; a qualifier that sounds precise while leaving the head kind unknown; service, server, provider, delivery, or access wording that leaves the promise, interface, system, role, method, work, or evidence object unclear; whole, part, complete, turnkey, or end-to-end wording that leaves a candidate whole, boundary, parthood, composition, coverage, or work claim unresolved; and integrity wording that still leaves open whether the sentence is about a structural whole, a characteristic or measurement, or evidence or assurance.

#Relations

#Content

#Problem frame

Use this when. Use this pattern when a claim contains a relation-bearing phrase, but the phrase does not yet determine the direct relation, exact participants, direction, or detail needed by a later engineering claim or operation. Common recognition moments include a broad predicate such as "linked", "aligned", or "supports"; a participant named by metonymy; a qualifier that sounds precise while leaving the head kind unknown; service, server, provider, delivery, or access wording that leaves the promise, interface, system, role, method, work, or evidence object unclear; whole, part, complete, turnkey, or end-to-end wording that leaves a candidate whole, boundary, parthood, composition, coverage, or work claim unresolved; and integrity wording that still leaves open whether the sentence is about a structural whole, a characteristic or measurement, or evidence or assurance.

Quoted, external, or ordinary source prose may remain as written. Open A.6.P only when an FPF statement will use the phrase to guide action, justify a decision or gate, support assurance or reliance, publish a claim, or reuse it across contexts. Repair that receiving FPF statement; preserve the source wording as a quotation or source expression instead of rewriting it as though the source had made the repaired claim.

Primary working reader, viewpoint, and concern. The working reader is an engineer viewing the sentence as input to a later claim or operation. The concern is that another person can find the same world-side or episteme-side objects, select the same direct governing pattern, and know which additional declaration, assertion, occurrence, designation, or representation detail that later use actually needs.

Primary EntityOfConcern. One relation-bearing claim in an episteme whose current expression leaves the direct relation kind or one or more actual participants unresolved, or leaves unclear whether a later claim or operation needs reusable declaration, explicit occurrence identity, designation, or representation.

First useful move. Replace the broad phrase with one readable sentence that names the exact participants and the direct relation believed to obtain. Name the governing pattern for that relation. If either the participants or the relation remain genuinely ambiguous, keep a small working candidate note and resolve that ambiguity before adding a reusable declaration, assigning a designator, or choosing a representation.

First-minute result. The draft Bearing_B is linked to Pump_P becomes Bearing_B isInstalledPartOf Pump_P during Interval_T after inspection identifies the physical part relation governed by A.14 and its current interval. If no later maintenance claim or operation distinguishes this installation episode from another, the repair stops there. A RelationSignature, explicit occurrence reference, or graph representation is added only when a named later claim or operation needs it.

What goes wrong if missed. A lexical replacement can make the sentence sound technical while preserving the same ambiguity. At the opposite extreme, an engineer can turn every relation phrase into a record-shaped episteme and then confuse that episteme, a declaration, or an identifier with the relation that obtains. Both failures obscure what is true, which object changes, and which pattern governs the needed claim or operation.

What this buys. The repaired claim remains readable. Load-bearing uses gain exact relation kinds, participant meanings, reusable typed declarations, occurrence identity, designations, and representations only where those distinctions change the later claim or operation.

Not this pattern when. Use the direct relation pattern when the relation and participants are already clear. Use A.6.5 when only reusable SlotSpecs are needed, A.6.REL when one obtaining occurrence needs explicit identity, C.2.1 when the issue is assertion or description identity, and F.18 when the object and relation are known and only designation remains unresolved.

FPF treats relation realism and epistemic access separately. A relation can obtain among its participants before anyone states, stores, diagrams, or names that fact. An assertion can affirm or deny the direct obtaining predicate. A declaration episteme carries reusable vocabulary and laws. A designator denotes an already recoverable object under an effective reference scheme. An episteme becomes available through a publication relation. A representation corresponds to an independently governed object or claim content. Relation precision restoration keeps these objects connected without collapsing them.

#Problem

An under-specified relation claim blocks a later claim or operation because several ontological questions remain hidden inside one phrase:

1. What kinds of objects are being related?
2. Which direct relation predicate is asserted to obtain?
3. Which relation-participant meanings are current, and are all actual participants named?
4. Does the current text state world-side participation, make an episteme claim about the direct relation, or define a local kind of entities participating under one meaning?
5. Does reusable typed use require a compatible RelationSignature and declaration-local SlotSpecs?
6. Does a later claim or operation need one obtaining occurrence to have explicit identity?
7. If something changes, is the changed object the occurrence, a declaration edition, assertion content or reliance posture, an evidence relation, a designation, a receiving-episteme reference, a description, a publication relation, a Bridge, or a representation edition?

Without answers, readers cannot tell whether two statements disagree, whether one participant may replace another, whether an inverse sentence preserves meaning, or whether a later claim refers to an obtaining occurrence rather than to an assertion or representation of it.

#Forces

#Solution

Begin with the objects named by the claim. Recover exact actual participants and one direct relation first. Then add only the declaration, assertion detail, occurrence identity, designation, reference, or representation demanded by the exact later claim or operation.

Local RPR mantra — five moves. Name the referents. State the direct relation or comparison with its actual participants, then follow its governing pattern. For the next named reader or task, add a declaration only to reuse typed rules, occurrence identity only to distinguish occurrences, a designation only to refer back to one object, or a representation only to show it in another form; otherwise add none. If a later sentence says something changed, name which object changed—the relation occurrence, claim-bearing episteme, designation relation, or representation—and follow that object's pattern. Then shorten without hiding the relation or its participants.

Referents means the objects recovered in 4.1; it is not a shared kind. Comparison means the direct comparison relation governed by A.19.CPM. Actual participants means the independently governed entities that participate under the relation's participant meanings. The mantra does not ask the reader to fill slots or positions or to create a record.

The mantra keeps the repair order and stop in attention. Sections 4.1-4.12 remain the governing Solution for hidden arity, world-side and epistemic separation, demand-driven declaration and individuation, relation-dependent wording, polarity, unresolved candidates, exact changed objects, Plain relaxation, and direct-pattern exit. The mantra is Plain didactic wording, not a second method, work plan, or performed work.

#Recover the objects before choosing relation notation

Start from the claim as written and ground each load-bearing head:

1. Identify the exact referent intended by each participant expression.
2. State the independently admitted kind of each referent under its direct governing pattern.
3. Separate a world-side object from an episteme about it and from a publication or representation of that episteme.
4. Recover metonymy explicitly. The phrase at the table may state physical location or participation in a negotiation meeting. Evidence from the current case selects the direct relation; neither reading by itself establishes a role assignment.
5. Leave the claim unresolved when the current evidence does not select one referent. A more technical synonym is not a repair.

The result of this step is an ordinary sentence containing identifiable objects. It is not a newly minted object kind. When several candidates remain live, use the small working note in A.6.P:4.9.

If the material is still a cue and no relation-bearing claim can yet be stated, stay with A.16.1 or B.4.1 instead of forcing relation publication. If the cue has stabilized into an open explanatory question but still has no selected relation answer, continue through B.5.2.0.

If counter-evidence or a failed use shows that a published relation statement overstates its articulation, closure, or framing, use A.16.2 to reopen, back off, or respecify that publication. A.16.2 records the retreat; A.6.P repairs the relation again only after the engineer can name a grounded candidate relation, its participants, and a discriminating check. Use A.16.0 only when readers must see lineage, branching, loss, or responsibility-transfer history; a local return needs no trajectory account.

A.6.P begins when the available observations or claims let the engineer name at least one grounded candidate relation, its participants, and a discriminating check.

#State the direct relation, participant meanings, and obtaining condition

Write the smallest readable direct-relation sentence that answers the current question:

<actual participant 1> <direct relation predicate> <actual participant 2> ...

Then name the direct governing pattern and recover from it:

• the admitted direct relation kind and its explicit governed RelationKind token;
• the relation-participant meanings and actual participants, each retaining its independently governed kind;
• the condition under which the relation obtains and its semantic predicate is satisfied by those participants considered under the participant meanings;
• applicability, direction, symmetry, inverse law, polarity, and temporal qualification when they change the predicate;
• the occurrence-identity rule, whether or not the current use needs explicit individuation;
• when the direct ontology says that a new occurrence is constructed or constituted, the constructor, inputs, construction work or process, and their contribution to occurrence identity.

Every in-scope direct subject-relation claim that exits A.6.P as positive or governed negative names an explicit admitted RelationKind token. When no suitable token exists, first settle the governed relation value and any required relation-kind admission under its direct pattern, [A.6.RCD](/generated/patterns/A.6.RCD), and [E.24](/generated/patterns/E.24); then apply [F.8](/generated/patterns/F.8) and, for durable naming, [F.18](/generated/patterns/F.18) and [F.17](/generated/patterns/F.17). Naming does not admit a kind or occurrence. An exact [A.6.1](/generated/patterns/A.6.1) operation-application binding, local [A.15.PROD](/generated/patterns/A.15.PROD) or [A.6.RCD](/generated/patterns/A.6.RCD) claim, or non-assertability result keeps its direct owner's semantics and is not coerced into this relation-kind family.

An ordinary assertion may name the actual participants directly. When reusable typed use is current, a compatible RelationSignature declaration can restate the participant meanings, obtaining predicate, applicability, and identity rule and contain only the declaration-local SlotSpecs needed by the receiving typed uses. The declaration remains an episteme; it neither makes the relation obtain nor supplies occurrence identity.

Assertion polarity remains claim-side. An affirmative assertion claims that the direct predicate is satisfied; a negative assertion denies it. Refutation or unresolved reliance belongs to [A.10](/generated/patterns/A.10) or the receiving evaluation. Denial, refutation, or unresolved reliance creates no negative world-side occurrence.

Do not select ontology from grammar. A verb-shaped phrase supplies neither constructive identity nor agency. Use the exact direct pattern for the relation, object, work, method, change, role, or acting system named by the current claim.

#Recover actual participants, hidden arity, qualifiers, and typed declaration only when needed

Ask which actual participation belongs to the direct relation's obtaining condition. Add a participant or qualifier only when it changes one of these:

• predicate satisfaction or relation obtaining;
• applicability or admissible use;
• occurrence identity;
• whether one participant can replace another without changing the claim;
• interpretation under an effective reference scheme;
• scope, Γ_time, viewpoint, view, or another exact qualification owned by the direct relation or receiving claim;
• witness or evidence expectations for a named decision or publication use;
• the exact later claim or operation.

For Sample_S wasMeasuredBy Instrument_I, a later evidence claim may separately refer to the measurement work occurrence, its interval, the applied measurement method, a measurement-result episteme, and a calibration episteme. The measured-by relation includes only the actual participants selected by its direct obtaining condition; the other objects remain participants or content of their own work, evidence, temporal, method-use, measurement, assertion, or description relations.

When reusable typed use is current, declare each participant meaning needed by that use through A.6.5:

SlotSpec := <SlotKind, ValueKind, refMode>

One SlotKind names one participant meaning locally inside one exact RelationSignature. ValueKind states the independently governed kind of the corresponding actual participant. refMode states how a receiving assertion or occurrence-description episteme designates that participant. The SlotSpec is declaration content; the participant does not become or occupy that declaration component. If one proposed ValueKind hides objects for which the predicate has different meaning, recover a real common kind or split the direct relation kind instead of preserving a hidden union as a prose list.

#Keep world-side, declaration, assertion, designation, and representation objects distinct

A representation can correspond to a direct relation, assertion content, declaration, participant designation, or already recoverable occurrence. State the exact source element, represented FPF object or claim content, and explicit C.29 correspondence. Representation form neither makes the relation obtain nor supplies participant or occurrence identity.

Functional and arrow forms are therefore assertion or representation notation, not world-side relation objects:

installedPartOf(Bearing_B, Pump_P, during=Interval_T)
Bearing_B --installedPartOf{during=Interval_T}--> Pump_P

The first can represent the content of a relational assertion; the second is a binary projection in a selected representation. A use that relies on either notation declares how its argument or endpoint elements correspond to the actual participants, direct predicate, qualifications, and any designated occurrence. The ordinary readable sentence remains sufficient when no representation-dependent use is current.

#Increase explicitness only for a named receiving use

Here receiving use is Plain shorthand for the exact later claim or operation that needs an additional object. It is not a shared FPF kind. Name that claim or operation and its direct governing pattern before using it to justify more apparatus.

Use progressive elaboration from one recovered direct relation:

readable direct-relation sentence with actual participants
  +-- compatible RelationSignature and SlotSpecs, when reusable typed declaration is current
  +-- explicit occurrence individuation, when a named receiver needs occurrence identity
      +-- occurrence-description episteme or stable designation, only when that receiver needs it
  +-- relational assertion detail, when polarity, modality, or reliance is current
  +-- C.29 representation and correspondence, when a representation-dependent use is current

This diagram is itself a [C.29](/generated/patterns/C.29) representation of independent elaboration branches, not a world-side structure or mandatory process. A RelationSignature is not a prerequisite for explicit occurrence identity. An assertion may name actual participants without a reusable declaration. A relation can obtain under its direct rule even when no local episteme exposes an occurrence designator. Conversely, a stored row, graph edge, tuple, or identifier does not establish obtaining.

Apply the [A.6.REL](/generated/patterns/A.6.REL) receiving-use test before explicit individuation. Comparison, occurrence history, nesting, and participation of an occurrence in another direct relation normally need identity. A direct relation assertion can stop without explicit occurrence identity when no later claim or operation distinguishes that occurrence. Repeated occurrences may have the same participants; the direct identity rule, not participant equality or row identity, supplies the discriminator.

#Resolve relation-dependent wording by the actual object

These readings leave no fourth qualification object. A readable word such as result, input, problem bearer, or next continuation can remain in Plain prose when the direct relation or claim is recoverable. Naming that reading creates neither a kind nor an occurrence. The world-side participant never becomes a declaration-local SlotSpec; the receiving episteme's designation denotes the participant without replacing it.

#Name change by the object that actually changes

There is no universal relation-edit operation. First point to the object that the sentence says changed. If the sentence says that the same object continued, use that object's identity rule to test that claim. If identity-bearing episteme content changed, name the result as another episteme rather than an in-place edit:

The object in the first column controls the operation and continuity test. Revision may name an activity family; one actual revision is one exact dated W : U.Work occurrence under A.15.1. If S : U.System performed it, recover the exact obtaining RA : U.RoleAssignment, check that S = RA.HolderSystemSlot, and state S performed W under RA or performedUnderAssignment(W, RA) under F.6. The revised episteme output is a separate object; performing the revision does not let an identity-bearing episteme change in place. A shared title, sequence, identifier, or authoring intention does not establish an edition, refinement, or supersession relation. If none of the identifying facts for the selected row changed, do not invent a change claim.

#Preserve polarity and inverse meaning

Participant order is part of many relation predicates. Bearing_B isPartOf Pump_P and Pump_P hasPart Bearing_B can be paired as inverse readings only when that inverse law is declared under the direct parthood pattern. A symmetric relation is symmetric under its direct law, not because a sentence sounds reciprocal.

When two viewpoints use different readable directions:

1. keep the same participant referents and their exact kinds;
2. name the forward predicate selected by the direct pattern;
3. use an explicit inverse predicate or inverse reading when one is available under that pattern;
4. keep scope, time, viewpoint, and reference scheme fixed while checking equivalence;
5. treat a change of participant kind or predicate as a semantic change rather than a stylistic rewrite.

#Use an actionable guide and keep a small candidate note when grounding is unresolved

For each ambiguity cluster, guide the reader through this order:

trigger expression -> candidate grounded objects and direct relations -> discriminating observations or tests -> readable direct-relation rewrite -> only the additional declaration, assertion, occurrence, designation, or representation needed by the named receiver -> exact governing exit

Do not organize the guide as a synonym list or make a table field the ontology. A qualifier such as comparative, safe, interactive, or reliable narrows wording but does not restore the head kind by itself.

When grounding remains unresolved, use this informative temporary episteme. The prompts are not a reusable schema or tuple kind.

For Alice is at the table, the physically present place and participation in a meeting are both plausible only while local evidence leaves them open. A location observation selects a located-at relation. A meeting roster and role assignment may instead select participation in meeting work. The note does not combine those relations and does not infer a role from a place expression.

When alternatives remain unresolved, the note may support explanation only. It cannot justify a decision, mechanism gate, publication claim, assurance, reliance, or cross-context reuse. Before stopping, name the reader, decision, or work that is blocked and name the observation, test, or direct-pattern condition that would separate the alternatives. Continue only after that discriminator selects one grounded reading; otherwise keep the alternatives explicit and keep the named use blocked.

#Classify boundary claims and keep engineered rewriting epistemic

Use A.6.B only when a sentence at the boundary does at least one of four things: defines a truth-conditional relation or signature rule (L); decides whether one identified mechanism application may start or continue (A); assigns a duty to an accountable actor (D); or states which execution effect or evidence can be observed and under which conditions (E). A sentence about claim scope or use, when to start or stop A.6.P, how to correct an endpoint kind, or whether a Bridge is needed does not qualify merely because it limits the repair.

Before giving a sentence an A label, answer two questions: Which mechanism application is about to start or continue? What predicate is checked at that point to admit or reject it? If either answer is missing, do not label the sentence A. Keep its scope or use, A.6.P start or stop decision, endpoint-kind correction, and Bridge need with the patterns that govern those questions. Split any mixed sentence before classifying its claims:

• L states the direct relation semantics, declaration invariants, polarity, participant meanings, and any reusable SlotSpec typing;
• A states one predicate checked when an identified mechanism application starts or runs. Its result says whether that application is admitted, may continue, or is rejected. A condition does not become A merely because it limits a claim, tells an author when to enter or stop this pattern, asks for an endpoint-kind correction, or requires a Bridge;
• D states duties of accountable systems or role assignments and does not turn a declaration into an actor;
• E states work and evidence expectations, witness carriers, observation conditions, and freshness under their direct owners.

Scope, Γ_time, viewpoint, reference scheme, witnesses, admissible use, and non-admissible overread stay with the direct relation or claim that actually needs them. They are not a universal qualifier kit, and an admissible use sentence is not an A claim unless the reader can point to both the mechanism application and its runtime entry predicate.

If a later task must describe an engineered episteme operation, first identify input X and output Y independently under C.2.1. A difference in claim content, EntityOfConcern, or effective reference scheme identifies another episteme; no component is rewritten in place. Keep the operation separate from X and Y. Use A.6.3 only for an exact compatible viewing or construction case: its entry must accept independently identified X and Y about the same exact EntityOfConcern, and its result must state the construction, preservation and loss, and applicability without substituting a slot/ref record for the operation. This edition does not route to A.6.2 or A.6.4: their current bodies still describe component-slot rewriting and retargeting by replacing EntityOfConcernSlot. When the needed operation is morphing or retargeting, stop with X, Y, the changed EntityOfConcern if any, and the sentence the next task needs; name A.6.2 or A.6.4 as the future owner that still requires a compatible repair. Call X and Y editions, refinements, or successors only when that direct continuity relation independently obtains; a shared title, sequence, or authoring intention is not enough.

If a system authors, materialises, checks, or publishes the output, that dated activity is U.Work under A.15.1; it is not the effect-free operation. Neither the operation nor that Work by itself makes the repaired world-side relation begin or cease, changes its actual participants, or supplies occurrence identity. Ordinary A.6.P repair stops before this step unless the reader's later task actually needs the operation described.

#Relax wording, then exit to the exact governor

After the relation has been recovered, Plain wording may be shorter than the Tech explanation. The shorter wording remains usable when a reader can still recover the exact participants, direct relation and governing pattern, every qualification that changes the declared use, and the point at which reusable declaration, occurrence identity, assertion detail, or representation becomes necessary.

A.6.P ends when the direct relation and participants are selected. The selected direct pattern governs that relation. Separate assertion, occurrence-identity, evidence, work, Bridge, description, publication, designation, and representation questions leave through their own patterns.

When generic relation recovery identifies one current claim at a method, intended-work, actual-work, production, evaluation, delivery, acceptance, transfer, or receiving-use boundary, apply A.6.P.WMR. It returns exactly one of four families:

1. an exact direct subject-relation claim, positive or governed negative;
2. an exact A.6.1 operation-application binding;
3. a local A.15.PROD claim or another local relation-bearing claim selected under A.6.RCD disposition 2;
4. an exact non-assertability result independently reasoned as factually unsupported, missing-information, or missing-governor.

Only missing-governor is an ontology blocker, and it names the affected receiving use and future owner. When participant referents and the named receiving claim are exact but no current direct relation closes that claim outside A.6.P.WMR, exit to A.6.RCD rather than improvising a relation or kind.

#Recover service/access claims through direct owners

Start with the decision, not a facet list. Ask what the reader must choose, do, accept, explain, restart, or stop. Then write one plain sentence naming the concrete subject or relation. If the source sentence carries several claims, write several sentences and send each to its own owner. The first useful result names the referent or relation, its current governing pattern, and the next action.

Source-domain guard. Bare service has no default system reading. In ordinary business and physical-world talk it may name a dated occurrence of service provision, a reusable way of providing it, offered outcome or eligibility content, provider participation, or another direct claim. In software talk it may be metonymic wording for an exact process, deployed component, endpoint, application, host, or cluster. Name the referent before choosing Work, Method, U.PromiseContent, U.RoleAssignment, U.System, or a relation. Never rewrite service automatically as server or as a system.

The table below is one recovery aid inside A.6.P. It is not a second router, a record schema, a common kind, or evidence that several rows obtain together.

Four language probes. Use these to select the question; do not treat them as templates whose rows all obtain.

• “My service stopped.” Ask what stopped. Service-provision Work may have ceased; an exact deployed software or physical bearer may have stopped or become unavailable; or promised availability or fulfilment may have failed. The sentence alone selects none. State, Work, promise, evidence, and fulfilment remain separate. Enter A.1.SCR only if the repaired bearer claim itself depends on systemhood.
• “Which services do we provide?” Name the offerings or promise contents being compared and any separately current provider assignments. Do not infer performed Work, a provider system, or fulfilment from the catalogue wording.
• “How is this service provided?” A reusable way of providing it selects a Method; a procedure or API text selects an episteme and perhaps MethodDescription; a dated provision selects Work. The wording alone selects none.
• “Restart the service.” Name the exact process, deployed component, endpoint, host, application, cluster, or other bearer to restart and the action's governor. Do not silently rewrite service as server, U.System, or service-provision Work.

Addressability is an aid, not a classification rule. If the sentence says call, visit, connect to, route to, restart, deploy, or scale, use it to ask which exact access point, delivery bearer, or other entity the claim concerns. Apply A.1 only when the repaired claim depends on systemhood. Actual Work still passes A.15.1 and every relation passes its direct governor. An endpoint may be an access point without being the whole delivery system.

Internet-access case. “We sell internet access” first becomes the commercial claim the reader needs: promise content, permission, provider or consumer participation, status, fulfilment, or another direct relation. For the promise reading, state the concrete claim—for example, Customer-18 may rely on PromiseContent-IA-18 for the named connectivity outcome and acceptance content—under its owner. If actual participation matters, state provider and consumer role assignments separately. If the source instead proposes the physical or operational whole InternetAccessArrangement-CA17, preserve that exact entity beside ProviderGateway-2, HomeRouter-18, and the status claims; enter A.1.SCR only when the decision depends on whether the arrangement itself is a system. Do not substitute one gateway or endpoint before that evaluation. Keep the connection Method and API or procedure description separate. ProvisionConnectionPlan-18 is intended Work; ConnectionEstablishmentWork-42 is one dated occurrence. A real grant uses A.2.8.PER and is not inferred from a credential. Connectivity status, measurement, evidence, evaluation, delivery, fulfilment, and acceptance each need their own claim. Nothing creates U.Access, AccessRelation, or one service bundle. After these direct subjects are recovered, use A.1.STM only when the live question is how one result contributes to use of a named project system-of-interest; otherwise stop at the direct owner.

Physical repair-shop case. “The repair service is delayed” is not yet one subject. If the blocked decision is where to leave the machine, name the front desk or intake point and test systemhood only if that decision needs it. If the decision is what physically performs the repair, name the workshop, equipment, or other exact bearer. If it is who is responsible, name an admitted holder and obtaining provider assignment only when current. What the customer was promised is promise content; an accountable deadline or duty is a commitment; how repair is done is a Method; the procedure card is a separate description; the repair that happened is dated Work. No phrase establishes all of these, and fulfilment or acceptance still requires its own relation.

No duplication boundary. A.6.P owns only this recovery move. It owns none of the routed objects or relations and creates no serviceSituation, QRR, facet bundle, universal situation, access kind, status kind, or compatibility record. A.1.SCR receives only an exact bearer or arrangement claim whose decision depends on systemhood. A.1.STM receives an already recovered result only when the next question asks how that result contributes to project use. The direct owners decide everything else.

The word that triggered the repair does not govern the result. The exact direct pattern does.

#Lexical guardrails

Overloaded words are diagnostic entry points, not relation kinds. In Tech or normative prose, same, synced, linked, connected, anchored, grounded, supported, and similar words cannot substitute for an unnamed direct relation or claim family. Bind and rebind remain name-binding or direct-owner vocabulary and are not generic relation-change verbs. A Plain gloss is admissible when its direct reading and governing exit remain recoverable.

E.10 owns the trigger scan and E.10.ARCH the shared wording-use recovery architecture. F.18 owns durable name selection after the governed object is known. A.6.P does not maintain a second trigger registry or mint one specialization for every repeated word.

#Archetypal Grounding

#Physical assembly and repeated occurrences

Tell. A maintenance note says replacement bearing is linked to Pump_P.

Show. Inspection finds that Bearing_B participates as the installed part and Pump_P as the assembly whole in the direct installed-part relation during Interval_T. Both remain physical holons of their independently governed kinds. A.14 governs the parthood predicate and obtaining condition. The maintenance assertion names them directly. If no later claim distinguishes installation episodes, the repair stops at the readable sentence.

Show the identity-dependent use. A reliability analysis compares the installed-part relation before removal with the relation after reinstallation. The same bearing and pump can participate in two occurrences. The analysis applies the direct identity rule and A.6.REL, exposes the two already obtaining occurrences, and designates them separately in its assertion. Maintenance database rows and diagram edges may represent the assertion or occurrence descriptions through explicit C.29 correspondences; row or edge identity is not physical relation identity.

When repeated maintenance assertions need one typed declaration, an InstalledPartRelationSignature may contain declaration-local InstalledPartSlot and AssemblyWholeSlot SlotSpecs corresponding to the two participant meanings. Those declaration components type the assertions' participant designations. They are not world-side places occupied by the bearing and pump.

Installation work constitutes the beginning of another installed-part occurrence only when the direct parthood identity rule says that it does. Creating or updating the maintenance row is representation work and is not an ontological constructor. The material character of the installed-part relation also does not by itself introduce a separate relator; the direct parthood ontology would have to identify and justify any constitutive truth-maker.

#Clinical evidence use and negative reliance

Tell. A care note says the measurement supports the dose change.

Show. The repair distinguishes the measurement work occurrence, its measurement-result episteme, the work-plan claim, the measured characteristic, the applicable range, and the exact evidence relation. C.16 governs measurement construction and applicability; A.10 governs whether that episteme bears on the named dose-change claim for this use. The broad predicate is not replaced by a universal support relation.

Show the boundary. If current evidence refutes the dose-change assertion or leaves reliance unresolved, the receiving evaluation records that posture. It does not create a negative world-side measurement, treatment, or evidence occurrence. A fresh publication of the same measurement-result episteme changes availability, not the measured condition or evidence relation by itself.

#Episteme correspondence and representation

Tell. Two teams say the models are aligned, and one tool draws an edge between their model nodes.

Show — models. Model epistemes DesignModel_E and MaintenanceModel_E have EntityOfConcern PumpAssembly_P204 and effective reference schemes CAD-Part-Scheme-v7 and Asset-Register-Scheme-2026Q2, respectively. AlignmentEdge_17 joins nodes labelled BRG-6204 and bearing-4471.

Structure decision. No established model-use structure changes either label's reading or whether replacement approval may use the pairing. Add no BoundedModelUseStructure; a diagram boundary is not evidence.

Needed sentence. For bearing-replacement planning on PumpAssembly_P204, CAD-Part-Scheme-v7 label BRG-6204 and Asset-Register-Scheme-2026Q2 label bearing-4471 denote the same installed bearing. This is a candidate claim, not yet a fact.

Edge boundary. Under C.29, AlignmentEdge_17 represents that sentence. It does not make the sentence true, prove one referent, or identify a Bridge occurrence.

Adjacent reading. Record ninety-seven percent of endpoint pairs passed the mapping test with C.16. If approval relies on that measurement, A.10 governs the evidence relation. The percentage does not establish the needed sentence.

Result. A.6.RCD missing-governor: bearing-replacement approval is blocked; participants are BRG-6204 and bearing-4471; the needed sentence is above; the edge remains a representation. No current direct correspondence or Bridge pattern states when the cross-scheme claim holds. A future direct pattern must state its predicate, conditions, and, if occurrences must be distinguished, identity rule. Until then, do not assert the models are aligned or mint a Bridge from the edge or a Card.

Show the boundary. The graph edge and its endpoint positions remain representation elements. An explicit C.29 correspondence states which assertion content, participants, and direct relation the edge represents. The edge does not make the correspondence obtain, prove same EntityOfConcern, or individuate a relation occurrence. Shared labels likewise establish neither same world-side referent nor substitutability.

When the claim is that both model epistemes concern the same world-side holon, recover each episteme's exact EntityOfConcern or grounding designation plus the observations, trajectory, or identity evidence governed for that holon. A Bridge preserves stated correspondences and losses; it does not prove the common world-side referent.

Show the viewpoint-conformance use. A review must decide whether model episteme E conforms to maintenance viewpoint episteme P. Identify E from its claim content, EntityOfConcern, and effective reference scheme, and resolve P as one claim-bearing viewpoint edition before reading a diagram, field, or reference as ontology. Then apply the direct predicate governed by E.17.0: EpistemeViewpointConformanceRelation(E,P). Plainly, E conforms to this exact viewpoint.

For unchanged E and P, the pair <E,P> determines one positive occurrence. Selecting P for this review is a separate use qualification: selection neither makes conformance obtain nor enters occurrence identity. If another review selects P2, test <E,P2> instead of retagging E. A viewpointRef or diagram label does not identify either participant by itself. A query, transformation, or projection may supply construction history for E, but neither that history nor an assertion, evaluation, representation, or publication makes conformance obtain.

Add an occurrence designator only if the next decision must distinguish two conformance occurrences. Add an evaluation result or evidence path only if the reader must rely on the conformance verdict. Add construction history only if the task asks how E was produced; add a representation or publication only if another reader must inspect or receive it; add associated Work only if the task asks who performed which activity. Otherwise add none of these objects. A.6.P stops after recovering E, P, and the direct relation; E.17.0 governs the predicate and, if the use separately asks whether an episteme is a U.View, that recognition.

#Method, work, role, and agency

The sentence the inspection method checks Pump_P uses active grammar. The repaired ontology says: one exact RA : U.RoleAssignment obtains under A.2.1 with four actual participants—admitted System_S as holder, InspectorRole, InspectorRoles_2026 : U.Episteme as the role-taxonomy episteme, and InspectionReferenceScheme : U.ReferenceScheme as the effective scheme. F.6 then states that System_S performed InspectionWork_W under RA (performedUnderAssignment(InspectionWork_W, RA)); InspectionWork_W applies InspectionMethod_M; and the direct examination relation connects the work occurrence to Pump_P. The example names the assignment participants but does not duplicate its interval or full assignment card. Each object keeps the identity and relations of its direct pattern. Only the holder system acts; the assignment does not.

#Formal reduced case

3 < 5 is assertion content in mathematical notation. The numeral occurrences, comparison sign, and operand places are representation elements under C.29. An explicit correspondence can relate them to the values, direct less-than predicate, and any compatible declaration used for typed reuse. No receiving use here distinguishes one obtaining occurrence from another, so the engineer stops at the assertion. A graph edge, tuple, or statement reifier introduced by a tool represents the proposition or assertion; it does not constitute the direct relation.

#Bias-Annotation

This pattern favors ontology-first recovery, direct relation sentences, and the lightest explicit form that serves the named receiving use. That bias counters record-first and vocabulary-first repair.

The counter-risk is under-specification: an engineer may stop before a load-bearing participant, applicability condition, occurrence identity, or reference is exposed. The receiving-use test, hidden-arity check, three-way relation-dependent wording distinction, and neighboring-pattern exits provide the counterweight.

The pattern also favors neutral domain language. Examples span physical assembly, clinical work, epistemes, roles, methods, and formal relations so that one publication technology or professional tradition does not become the default ontology.

#Conformance Checklist

1. Recognition. The use begins with one actual relation-bearing claim and states which later claim or operation is blocked by its ambiguity.
2. Grounded heads. Every load-bearing head refers to an exact object or remains explicitly unresolved; a qualifier does not substitute for the head kind.
3. Direct relation. Every positive or governed-negative direct subject-relation exit names exact actual participants, an explicit admitted RelationKind token, and the direct governing pattern. The A.6.P.WMR non-relation exits remain under their exact owners.
4. Participant meanings. The direct pattern states the participant meanings, actual participation, obtaining predicate, applicability, and occurrence-identity rule; every participant retains its independently governed kind.
5. No negative occurrence. Negative assertion, refutation, or unresolved reliance remains claim- or evaluation-side and creates no negative world-side occurrence.
6. Demand-driven declaration. A compatible RelationSignature and declaration-local SlotSpecs appear only when reusable typed use is current; an ordinary assertion may name participants directly.
7. Designation separation. A participant designation or occurrence reference remains content of a receiving episteme and neither replaces the referent nor makes the relation obtain.
8. Hidden arity and qualifiers. Every participant or qualifier included in the repair changes predicate satisfaction, applicability, identity, substitution, interpretation, admissible use, witness needs, or the named later claim or operation.
9. Occurrence threshold. Explicit U.Relation identity appears only when a named receiver needs one occurrence distinguished from another; repeated occurrences with the same participants use the direct identity discriminator.
10. Construction choice. When construction is identity-bearing, the direct owner names the constructor, inputs, construction work or process, and identity contribution; otherwise the repair introduces no constructor.
11. Object separation. Direct relation kind, participant meaning, actual participant, declaration, assertion, occurrence description, occurrence, designator, reference, publication, Bridge, and representation remain distinct where present.
12. Relation-dependent wording. The reading resolves to world-side actual participation, an assertion- or description-side designation, or a justified C.3 local kind; no fourth qualification object is introduced.
13. Polarity. Participant order, inverse wording, symmetry, assertion polarity, and temporal qualification follow the direct relation law.
14. Changed object. Every change statement selects one exact changed object and its governing row in A.6.P:4.7; no generic relation-edit operation remains.
15. Boundary classification. Apply L, A, D, and E only to an actual A.6.B boundary statement. Before marking A, point to the particular mechanism application and the predicate checked at entry; if either is absent, leave claim use or scope, A.6.P entry or stop, endpoint-kind correction, and Bridge need with their own patterns.
16. Candidate guide and stop. Unresolved alternatives are grounded objects, kinds, or relations with a discriminating check, not a synonym list or representation-first ontology. Until the check selects one reading, the note remains Plain or informative, names the blocked reader, decision, or work and the needed discriminator, and carries no decision, gate, publication, assurance, reliance, or cross-context reuse.
17. Representation boundary. A table, row, field set, tuple, graph edge, functional expression, arrow, formula, or reifier has explicit C.29 correspondence for any relied-on FPF use and does not constitute the represented relation by form.
18. Optional episteme operation. Use this path only when a later task must describe an operation on an episteme or representation. Identify input and output independently under C.2.1; use only an exact compatible A.6.3 viewing or construction case, keep any continuity relation separate, and use A.15.1 for actual authoring, materialisation, checking, or publication Work. Do not route from this edition to the current slot/write profiles in A.6.2 or A.6.4; for morphing or retargeting, preserve the exact input, output, changed EntityOfConcern if any, and needed sentence as the explicit future-owner stop. Neither an operation nor that Work by itself changes the repaired world-side relation or supplies occurrence identity.
19. Plain relaxation. Short final wording retains a recoverable direct relation, actual participants, and visible escalation points.
20. Neighbor exit. The repaired claim leaves A.6.P through one exact governing exit in A.6.P:4.11, including exactly one of the four A.6.P.WMR families when that specialization is current.

#Common Anti-Patterns and How to Avoid Them

#Consequences

Benefits

• Engineers can retain short practical relation language without surrendering exact ontology.
• Hidden participants and kind collapses become repairable through one repeatable sequence.
• Reusable declarations, occurrence identity, designations, and representations appear only for named receiving uses.
• Assertion, declaration, occurrence, evidence, designation, publication, Bridge, and representation changes remain independently reviewable.
• Neighboring patterns receive the object they actually govern rather than a generic container.

Costs and mitigations

• Grounding an ambiguous phrase takes more work than lexical replacement. The first useful move and small candidate note keep that work bounded.
• Some claims become longer when their truth depends on scope, time, viewpoint, witness, or another participant. Plain relaxation restores readability after the precise reading is stable.
• Domain-specific direct relation patterns still need their own obtaining and identity rules. A.6.P supplies recovery and governing-pattern selection, not a universal relation ontology.

#Rationale

The problem is ontological before it is lexical. A broad expression is dangerous because it can hide different objects, predicates, participant meanings, actual participants, or later claims and operations. Replacing the expression without recovering those objects merely moves the ambiguity.

FPF therefore begins with relation realism: a relation may obtain independently of an assertion about it. Actual participants retain their independently governed kinds while participating under relation-local meanings. An optional RelationSignature makes those meanings reusable as typed declaration content; an assertion or description can designate participants or an already recoverable occurrence; a representation corresponds to the selected object or claim content. None of those adjacent objects substitutes for world-side obtaining or identity.

Construction matters only when the direct ontology makes it identity-bearing. In that case, constructor, inputs, construction work or process, and output identity explain how another occurrence begins. Other relations may obtain without such a constructor. A.6.P therefore uses constructional analysis as a discriminating test rather than as a universal ontology.

Progressive elaboration preserves affordability. An engineer can stop at a readable direct fact. Typed declaration, assertion detail, occurrence identity, stable references, descriptions, publications, and representations appear only when a named later claim or operation depends on them. This order also explains why naming comes late: a good name helps designate a recovered distinction but cannot create it.

The changed-object rule prevents a second ontology collapse. Revising an assertion, declaration, evidence relation, designation, reference, publication, Bridge, or representation is not automatically a change in the direct relation being discussed. Naming the changed object makes continuity, evidence, and return conditions intelligible.

Neutral terminology matters because A.6.P is cross-domain. Every object named by the method remains under its direct governing pattern; A.6.P does not collect unlike objects under one local umbrella. Domain-specialized vocabulary enters only when a direct local pattern governs the current object.

#SoTA-Echoing

#Ontological SoTA and constructive grounding

These sources constrain relation occurrence, identity, and construction. They are not interchangeable: FPF takes constructor-and-input discipline from the constructional-ontology line, uses BORO as a distinct 4D comparison, and treats implementation ontologies as stress comparators rather than proof.

#Representation and implementation stress tests

These sources do not decide what exists. They test whether a representation can preserve the ontological distinctions selected above without turning a statement, row, graph term, tuple, or reifier into the world-side occurrence.

#Service and access separation pressure

These sources constrain the recovery of service or access wording; they do not define a service ontology for FPF.

Earlier public service lineage also cited ITIL 4, ISO 24617-2 speech-act practice, and SRE literature. They remain bounded examples rather than ontological governors: ITIL offer and service-level wording can cue A.2.3 or A.6.C; a communicative act is separated from its content and any enduring binding by A.2.9, A.2.3, and A.2.8; SRE interface, SLO, deployment, telemetry, and incident distinctions can help name separate claims. None licenses an always-unpack word rule, a mandatory facet family, every deontic phrase becoming a commitment, every performative phrase becoming a speech act, or actuals becoming Work and evidence automatically.

Across these sources, FPF adopts separation pressure and adapts it to the direct-owner routes in 4.11a. It explicitly rejects U.Access, AccessRelation, a service bundle, word-induced systemhood, and blanket actuals-to-Work.

The first table governs the general ontological moves. The second checks representability only after those moves have been selected. The service-and-access table constrains one recurring recovery branch without importing a service ontology. The physical, clinical, episteme, work, and formal cases test that the resulting method is not specialized to information systems.

Reopen the smallest affected passage. Start with the one claim, case, exit, or source row that uses the changed fact. Reopen it when its governing pattern changes who participates, when the relation obtains, or how an occurrence is reidentified; when newer source evidence overturns or narrows the construction or reification distinction used there; or when an actual use can no longer reach the practical result or stopping boundary promised by that passage. Do not reopen the whole pattern unless the same change reaches several passages. If an exit no longer matches its owner's entry and result, stop using that exit until it is repaired.

#Relations

• The direct relation pattern governs obtaining, predicate satisfaction, and the occurrence-identity rule. After A.6.P recovers that relation, A.6.REL governs demand-driven explicit individuation, application of that identity rule, and occurrence-as-participant use.
• A.6.0 governs U.Signature and compatible RelationSignature declarations; A.6.5 governs declaration-local SlotSpec, SlotKind, participant ValueKind, and receiving-episteme designation mode.
• A.6.RSIR selects among direct participation, declaration, operation, assertion or description, and representation when interface, role, slot, field, parameter, or endpoint wording is the entry cue.
• A.6.B separates L, A, D, and E claims after the direct relation is recovered. When the next task explicitly needs an operation on an episteme or representation, identify its input and output independently under C.2.1; use A.6.3 only for an exact compatible viewing or construction result, and A.15.1 for actual authoring, materialisation, checking, or publication Work. Current A.6.2 and A.6.4 are not exits from this edition because their slot/write profiles fail that compatibility test; keep them as named future owners for the explicit morphing or retargeting stop. Assert an edition or successor relation only after its own predicate is satisfied. Ordinary A.6.P repair stops before these objects.
• A.6.6 provides specialized recovery for basedness. Service/access wording stays in A.6.P:4.11a and exits to its exact direct owner. It reaches A.1.SCR only after one exact bearer or arrangement claim has been recovered and the decision depends on systemhood; it reaches A.1.STM only after recovery when the separate question is contribution to use of a named project system-of-interest. Cross-context and whole-part wording may use A.6.9 or A.6.H only when that pattern's entry accepts the objects named in 4.11 and its result returns the direct predicate and participants or an explicit blocker. If either check fails, stay with the 4.11 exit and its direct owner. A situation record, Card, or bundle does not replace the direct relation, claim-bearing episteme, or representation.
• A.6.P.WMR governs its method, work, result, production, delivery, acceptance, transfer, and receiving-use boundary and returns one of the four results listed in 4.11.
• Use A.6.RCD after the reader can name the participants and the sentence the next task needs, but no current pattern supplies its predicate. Broad wording alone is not a missing-governor result.
• C.2.1 governs assertions and descriptions. A.6.P identifies one candidate episteme E, one viewpoint episteme P, and the conformance question; E.17.0 tests EpistemeViewpointConformanceRelation(E,P). If the next task also asks whether E is a U.View, E.17.0 handles that recognition separately. Viewpoint selection, evaluation, construction, representation, and publication do not establish conformance; A.10 governs reliance, while E.17 and E.24.PUB govern publication.
• Use C.3 and C.3.1 only if the next claim must quantify over a locally defined participant set, test membership or substitution, or order kinds. Otherwise do not mint a local kind.
• A.1, A.2, A.2.1, A.3.1, A.3.4, and A.15.1 govern the direct criteria for systems, roles, role assignments, methods, actual bounded change, and work.
• A.10 governs evidence relations and B.3 assurance. F.9 governs Bridge-description or Bridge Card content and its admitted cross-context use; it does not supply the separate direct Bridge predicate, obtaining condition, or occurrence identity. Use only a pattern that states that predicate and identity rule. If none exists after both endpoints and the needed sentence are explicit, return missing-governor through A.6.RCD. C.30.P governs architecture wording, C.16.P characteristic wording, G.2 palette/front/archive distinctions, and A.6.F function-like wording. Each direct domain pattern governs its recovered relation.
• A.16.1 and B.4.1 retain cue material that has not reached a grounded relation-bearing claim; B.5.2.0 carries a stabilized open explanatory question that has no selected relation answer; A.16.2 records reopen, backoff, or respecification when a published relation overstates articulation, closure, or framing. Use A.16.0 only when lineage, branching, loss, or responsibility-transfer history itself must be published.
• E.10 governs wording triggers and E.10.ARCH shared recovery architecture. F.18 governs designation after objects and relations are recovered.

#C.29 mathematical-lens account and representation boundary

C.29 governs a declared mathematical-lens-use account or representation only when a mathematical object, tuple, graph, function, arrow, table, or other formalism is used for a stated subject and purpose. It owns the candidate mathematical object designation, mapping mode, explicit correspondence, preserved structure, lost structure, declared use, blocked overread, and stop condition. It does not assert world-side participation, make a direct relation obtain, admit its kind, or supply occurrence identity.

When a sentence claims cross-context meaning, export, correspondence, or substitution, first name what each endpoint means locally and write the Bridge predicate being asserted. The direct Bridge owner decides whether that predicate obtains and whether one occurrence continues; F.9 and C.2.1 govern the separate Bridge description or Card episteme and its direction, CL, loss, and admitted-use claims. A changed Card, evidence item, or publication is not a changed Bridge occurrence. Use C.29 only when the reader relies on a representation or mathematical-lens account; a representation does not establish the Bridge or the represented world-side relation. If the sentence needs no cross-context correspondence or substitution claim, add no Bridge. If it does need one but no current pattern supplies the predicate or identity rule after the endpoints are named, return missing-governor through A.6.RCD.

#A.6.P:End

Last Updated: 2026-08-04 — upstream FPF commit 8b727cba (github.com/ailev/FPF)