#Relation, Signature, Interface, Role, and Slot Precision Restoration

A.6.RSIR · Stable · FPF precision-restoration pattern · 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: FPF precision-restoration pattern Status: Stable Normativity: Normative unless marked informative

Plain name. Relation-signature-interface-role-slot recovery.

#Keywords

• relation-signature-interface-role-slot recovery
• direct relation participant
• participant meaning
• actual participant
• RelationSignature
• SlotSpec
• assertion or description designation
• operation argument or result declaration
• exact operation application and binding
• representation position and correspondence
• role
• role assignment
• interface
• port
• API
• reduced-use source label
• shadow ontology.

#Relations

#Content

#Use This When

Plain name. Relation-signature-interface-role-slot recovery.

Use this pattern when relation, signature, interface, role, role-holder grammar such as Holder#Role:Context, assignment, enactment, slot, field, parameter, argument, endpoint, port, API, protocol, capability, affordance, method, function, concern, interest, Markov-blanket, computational-boundary, or active-inference-boundary wording hides which FPF object or claim kind is current.

Primary EntityOfConcern. The EntityOfConcern is one encountered use of an ambiguous engineering phrase together with the claim that this use is intended to carry. RSIR recovers the direct governed object, direct relation and participant meaning, actual participant, declaration-local SlotSpec or operation declaration, exact operation application and binding, assertion- or description-side designation, representation position and correspondence, or claim before selecting its governing pattern. The phrase remains wording in an episteme or in speech; it is not the world-side object, occurrence, value, or relation named by the recovered claim.

Primary working reader. The first reader is an FPF pattern author, reviewer, or practitioner repairing a phrase before selecting the direct governing pattern. The downstream reader is the engineer, manager, analyst, or steward who needs the repaired phrase to preserve useful project language without minting a shadow ontology.

First useful move. Recover the project concern first, then recover the current governed EntityOfConcern or claim kind. Apply the direct governing pattern as soon as it is clear. Keep a reduced-use source label only when no governed value is being asserted.

What goes wrong if missed. The same word is used for differently governed objects without saying which claim is current. For example, interface may denote an API description, reusable signature, functional port, compatibility claim, or module-boundary relation; role may denote a work-facing U.Role or be misused for a direct relation-participant meaning, a declaration-local SlotKind, or a representation position. A later reader then cannot recover which relation obtains, which actual participant is meant, which SlotSpec or operation declaration is current, whether an exact application binds an actual value, or which representation correspondence is intended.

What this buys. The reader gets one small recovery move before the direct pattern is applied. The repair preserves useful engineering words while preventing a lexical cue from minting a new root kind or collapsing direct participation, reusable declaration, assertion or description, exact operation application and binding, and representation correspondence.

Not this pattern when. Do not use A.6.RSIR after the direct governing pattern is already clear. Do not use it for general relation repair after A.6.P is selected, for slot discipline after A.6.5 is selected, for function-like repair after A.6.F is selected, for module-interface repair after A.6.M is selected, for transformation wording after A.3.4.P is selected, or for publication and description repair after E.17, C.2.1, or C.2.P.DR is selected.

#Problem frame

The RSIR cluster sits at a common failure point in FPF texts. A project team sees one word and treats it as if it already selected the ontology:

• "role" in a work assignment, direct relation-participant meaning, declaration-local SlotKind, representation argument, RBAC-like status, or evidence use;
• "interface" in a module relation, functional port, API description, protocol, signature, or publication view;
• "slot", "field", "parameter", or "argument" in wording about an actual relation participant, a RelationSignature declaration, an A.6.1 argument or result declaration, one actual operation application and bound value, a data, formula, or method-call representation, or ordinary prose;
• "signature" in a law-governed declaration, API shape, interface specification, or plain sign-off phrase;
• "function" in architecture, capability, method, work, mathematical modeling, or quality wording.

A.6.RSIR is the first-level recovery pattern for this bounded cluster. It does not decide every neighboring subject ontology. It helps the practitioner recover which object or claim is current and then stop at the direct governing pattern.

#Problem

Without this pattern:

1. Lexical cues create shadow kinds. Interface, role, slot, endpoint, and function words become local root kinds because they sound technical.
2. Participant, declaration, and representation uses become roles. A direct relation-participant meaning, a declaration-local SlotKind, or an argument, field, or endpoint in a selected representation is called a role and then confused with U.Role; evidence-use, transformation, and interface claims lose their direct owners.
3. Role values become declaration or representation labels. A real U.Role is demoted into a declaration-local SlotKind or a source-schema field, so the role-taxonomy episteme, effective reference scheme, assignment occurrence, assignment window, role state, and work consequences can no longer be recovered.
4. Signatures absorb implementations. A law-governed U.Signature is used as if it were a mechanism, method, work-start gate decision, interface conformance proof, or publication.
5. Participant, declaration, application, and representation boundaries are skipped. A field or parameter is edited without deciding whether it denotes a direct relation-participant meaning or actual participant, a declaration-local SlotSpec, an A.6.1 argument or result declaration, one exact operation application and actual binding, or a position in a selected representation.
6. Evidence and status uses keep old role grammar. An episteme, standard, report, publication, or badge is said to have a role instead of being used in an evidence-use, source-use, status-use, publication-use, assurance-use, or gate relation.
7. Neighboring patterns are copied locally. A pattern repeats negative catalogues such as "not proof, not permission, not gate" instead of recovering the current object and applying the pattern that governs the claim.

#Forces

#Solution

Use A.6.RSIR as a first-level recovery move. RSIRRepairNote is optional working support, not a required record, schema, or publication layout. Omit every branch that is not current. The ordinary path may stop after projectConcern, recoveredEntityOfConcernOrClaimKind, selectedDirectGoverningPattern, and one result stated as retainedSourceLabelUse, blockedOverread, or nextAdmissibleUse.

RSIRRepairNote (optional working support; keep only current lines):
  projectConcern:
  recoveredEntityOfConcernOrClaimKind:
  selectedDirectGoverningPattern:
  retainedSourceLabelUse?:
  blockedOverread?:
  nextAdmissibleUse?:
  encounteredWording?:
  currentUse?:
  directParticipantMeaningAndActualParticipant?:
  relationDeclaration?:
  assertionOrDescriptionDesignation?:
  operationDeclaration?:
  exactOperationApplicationAndBinding?:
  representationUseAndCorrespondence?:
  neighboringCandidateValues?:
  stopCondition?:

When the optional note is used, it is complete when the current object or claim kind is clear enough to apply the direct governing pattern, keep ordinary prose, keep quote-only wording, or stop the stronger claim. No unused branch is filled for completeness.

#Recovery order

1. Recover the project concern. Say what the project is trying to do: assign work responsibility, declare a signature, check an interface, compare functions, name a port, use evidence, assert status, describe a method, or make another claim.
2. Recover the current governed object or claim kind. Decide whether the wording points to a direct relation or participant meaning, an actual participant, a reusable RelationSignature or SlotSpec, an assertion- or description-side participant designation, an A.6.1 argument or result declaration, one exact operation application and actual binding, a representation position and correspondence, a signature, interface claim, role value, role assignment, role description, port, boundary claim bundle, capability, affordance, method, function, concern, interest, publication, source label, or ordinary prose.
3. Name the direct governing pattern. Use the table in A.6.RSIR:4.2 only until the governing pattern is clear.
4. Separate direct participation, reusable declaration, and assertion or description. Use A.6.5 only when one complete SlotSpec in one exact RelationSignature is current. The direct relation pattern governs participant meaning, actual participants, obtaining, and occurrence identity. If an assertion or description episteme designates a participant, C.2.1 governs that episteme's identity and content, while the direct assertion, evaluation, evidence-use, or description family governs the exact predicate, polarity, or use relation. When a compatible SlotSpec is current, A.6.5 governs the designation's ValueKind and refMode discipline; an ordinary assertion may instead name actual participants directly without opening a reusable RelationSignature.
5. Separate operation declaration, actual application and binding, and representation. A.6.1 governs declaration-local ArgumentDeclaration and ResultDeclaration content. Open an actual operation-application binding only after one exact application has been independently identified and its actual bound value matters to a receiving claim. Keep a method-call, formula, tuple, edge, or schema place under C.29 or its exact representation owner and state correspondence separately.
6. Keep the source label reduced-use when no governed claim is current. A word can remain a cue, quotation, title, or local shorthand without being admitted as FPF-governed vocabulary.

Use Tech position only for a place in a selected representation, such as a tuple component, formula or method-call argument, graph-edge endpoint, or schema field. Until an explicit correspondence is stated, that position is neither a relation-participant meaning, actual participant, SlotKind, SlotSpec, nor evidence that the direct relation obtains.

#Direct governing pattern selection

#Relation-defined wording dispatch

When wording derives a qualification, status, or category from participation in a relation, recover the object needed by the next use before naming it:

1. If the claim concerns an actual entity participating under one named relation-participant meaning, state the direct relation, that meaning, and the actual participant. The participant retains its direct kind.
2. If reusable typed declaration is current, use A.6.5 for the corresponding SlotSpec inside one exact RelationSignature. Its SlotKind is declaration-local and neither is the participant nor makes the relation obtain.
3. If an episteme asserts, evaluates, or describes the participation, C.2.1 governs the episteme's identity and content, while the direct assertion, evaluation, evidence-use, or description family governs the exact predicate, polarity, or use relation. When a compatible SlotSpec is current, A.6.5 governs the participant designation's ValueKind and refMode discipline; without reusable declaration, the assertion may designate the actual participants directly.
4. If repeated local quantification over such actual participants is current, use C.3 and C.3.1 for the local U.Kind, membership rule, and extent rule. Neither the participant-meaning label nor the declaration-local SlotKind admits that kind.
5. If the source exposes a tuple component, argument, edge endpoint, schema field, or other representation position, keep it under C.29 or the exact representation owner and state an explicit correspondence before an FPF claim consumes it. A value shown at that position establishes neither actual participation nor relation obtaining.

For parameter, argument, or result wording, separately recover the A.6.1 declaration content, one independently identified exact operation application and any obtaining declaration-local binding, and the selected representation position. Open the binding only when the actual bound value matters to a receiving claim. Neither the declaration nor representation syntax establishes the binding; a result binding is distinct from production, a produced entity, a result episteme, evidence, and work.

When a receiving use compares or constrains a whole organization of relation occurrences, A.22 may govern a selected U.Structure. One actual participant, corresponding SlotSpec or designation, operation binding, or representation position does not by itself establish such a structure.

#Replacement candidate rule

Do not replace one umbrella with another. The minimum admissible repair candidate names:

• the current object or claim kind;
• the governing pattern;
• one result current for the receiving use: a retained reduced-use source label, a blocked stronger reading, or the next admissible use.

Name a direct relation, claim-bearing episteme, declaration-local SlotSpec, A.6.1 operation declaration or actual application binding, or representation correspondence only when that exact object is current for the receiving use. Do not fill an unused branch or require both a retained source-label use and a blocked overread. If the minimum cannot be named, leave the phrase in quote-only or reduced-use form and record the blocker.

#Reduced-use source labels

Reduced-use labels are allowed. They are not failures. A source label remains reduced-use when it helps readers find or recognize the case but does not carry FPF-governed content.

Examples:

• "API role" can remain a quoted source phrase while the repair separately names the actual governed claim: a software API description, a provider role assignment, a promise-content episteme under A.2.3, a separately obtaining commitment under A.2.8, PromiseContentUse, a delivery or acceptance relation under its direct owner, another named direct relation, or an interface specification.
• "parameter" can remain ordinary prose while a complete SlotSpec is named only for a current reusable relation declaration, an operation ArgumentDeclaration or ResultDeclaration and any exact application binding stay under A.6.1, and a method-call, formula, or other representation position stays under C.29 or its exact representation owner.
• "function" can remain ordinary engineering language when no architecture, capability, method, work, mathematical, quality, or module claim depends on it.

#Shortcut Cost and Reopen Condition

A.6.RSIR is a deliberately weak first-level repair note. The baseline is full use of the direct governing pattern: A.6.P for relation repair, A.6.5 only for reusable RelationSignature SlotSpec discipline and compatible participant-designation typing, C.2.1 plus the direct claim family for assertion or description content, A.6.1 for operation declarations and any exact application binding, C.29 or the exact representation owner for positions and correspondence, A.2 and A.2.1 for role and role assignment, A.6.M for module-interface, A.6.F for function-like repair, or the evidence, status, publication, architecture, method, work, gate, or problem pattern named by value.

The saved effort is that a practitioner does not run several full patterns before knowing which one is current. The loss budget is narrow: RSIR may select a governing pattern, preserve a reduced-use source label, or record a blocker. It may not decide the role assignment, signature, operation application or binding, evidence-use relation, status assertion, exact service/access referent or direct relation, architecture description, or method relation that belongs to the selected pattern.

Reopen RSIR when the selected pattern shows that the source phrase carried more than one governed object, the object kind was selected too early, a needed slot distinction was missed, or evidence, status, publication, gate, method, work, architecture, capability, or concern claims were folded into one label. The reopened repair splits the phrase into multiple governed values or keeps the excess wording reduced-use.

#Archetypal Grounding

System case: module interface claim. A team says "the cooling module exposes the heat-exchanger interface." RSIR first asks what claim is current. If the claim is substitutability or separate change, use A.6.M. If a reusable relation declaration for exchanged-medium and boundary-condition participant meanings is current, use A.6.0 plus A.6.5 for the RelationSignature and complete SlotSpecs. If the current use is a diagram, API schema, or other representation, keep its positions under C.29 or the exact representation owner and state explicit correspondence. If the claim is a functional port in a transformation-flow structure, use A.6.F, A.3.4, and E.18. RSIR does not create U.Interface.

Role case: API provider role. A source says "the API role is provider." RSIR first recovers what participates in work. If provider is a work-facing role, use A.2.1 to name the holder system, ProviderRole, role-taxonomy episteme, effective reference scheme, and assignment window. Add a model-use structure only when an independently selected DDD-style organization changes interpretation. If the API is a publication or protocol description, use E.17 for publication; use A.6.P:4.11a when service or service-access wording hides the concrete subject or relation. Send promise content to A.2.3 and an accountable provider or consumer commitment to A.2.8. For protocol, contract, SLA, or agreement-like boundary wording that bundles several claims, use A.6.C to unpack them before sending each recovered object to its direct owner; if module-interface semantics are current, use A.6.M; if boundary-package statement classification is current, use A.6.B. Do not assign a work role to the API description.

Evidence case: reviewer evidence role. A report says "reviewer evidence role approved the gate." RSIR blocks the composite. ReviewerRole may be assigned to an admitted U.System under A.2 and A.2.1. A report episteme may be used in an evidence-use relation under A.10, B.3, F.10, or E.17. A gate approval may be a gate decision under A.21 or a speech-act case under A.2.9. No episteme gets a work role by being evidence.

Slot case: method parameter. A method description says "parameter target controls the model." That sentence has no exact governor in this case, so it is not retained as the repaired claim; keep target only as a reduced-use source label and write one positive A.6.1 use instead. In the current recognizeAdmittedHolonCandidate declaration, candidate is an ArgumentDeclaration meaning one exact entity being evaluated, with ValueKind = U.Entity; recognitionJudgment is the declared result meaning. Under A.6.1, the project independently identifies the bounded recognition-evaluation invocation P-37 by that declaration's application predicate, identity rule, and extent rule. During P-37, Pump #37 is actually bound under candidate, and the returned value unknown is bound under recognitionJudgment. In a call representation such as recognizeAdmittedHolonCandidate(target = Pump-37, ...), the named-argument position target corresponds to the declared candidate meaning but is neither the declaration nor either binding. The practitioner writes: "target is the call label; A.6.1 declares candidate : U.Entity; during exact application P-37, Pump #37 is bound as candidate." Stop there unless the receiving claim needs the result binding or another direct owner.

#Near-Miss Checks

#Bias-Annotation

This pattern has a relation-cluster bias because it sits in A.6. It mitigates that bias by stopping as soon as the direct governing pattern is clear.

It has an interface and software-language stress case because API, endpoint, protocol, and interface wording often enters from software. The pattern deliberately keeps the recovery general: architecture interfaces, physical ports, functional ports, service-access descriptions, and publication forms are all possible, and none is selected by word choice alone.

It resists semio-bias by keeping descriptions, publications, records, reports, standards, and source labels under the patterns that govern those objects and uses: C.2.1, E.17, C.2.P.DR, A.10, B.3, F.10, C.28, E.10, or E.10.ARCH when those objects or uses are current. A source label may help recognition; its presence is not evidence that the denoted object is the current EntityOfConcern or that a proposed action is admissible.

#Conformance Checklist

1. The repair starts with project concern, not with a replacement word.
2. The current EntityOfConcern or claim kind is named before a direct governing pattern is applied.
3. The repair stops at the direct governing pattern once it is clear.
4. When reusable relation declaration is current, slot discipline uses A.6.5 and states one complete SlotSpec = <SlotKind, ValueKind, refMode> inside one exact RelationSignature; actual participants and representation positions remain outside it.
5. Role claims preserve the four participants of generic U.RoleAssignment and derive occurrence extent from uninterrupted obtaining; claims about role description, role state, selected role relation structure, capability, method, planned work, and performed work exit to their direct patterns.
6. Evidence-use and status-use cases are not represented through U.RoleAssignment for epistemes.
7. Interface wording is kept as a recognition cue but is not admitted as generic U.Interface.
8. Every neighboring object family selected in the dispatch table exits to its direct governing pattern rather than being redescribed inside RSIR.
9. Relation-defined wording dispatches separately to the direct participant meaning and actual participant; a declaration-local SlotSpec when reusable typing is current; an assertion- or description-side designation whose episteme identity and content stay with C.2.1, whose predicate, polarity, and use stay with the direct claim family, and whose typing stays with A.6.5 only when a compatible SlotSpec is current; a C.3 local kind when repeated quantification is current; or a representation position plus explicit correspondence. It does not create one umbrella qualification object.
10. Operation wording keeps A.6.1 ArgumentDeclaration or ResultDeclaration content, one independently identified exact application and obtaining argument or result binding, and any call or formula representation position distinct; it infers neither a public application kind nor production, a produced entity, a result episteme, evidence, or work from the binding.
11. Quote-only or reduced-use labels carry no action-facing claim beyond the claim admitted by the selected governing pattern.

#Common Anti-Patterns and How to Avoid Them

#Consequences

A.6.RSIR adds a small first-level decision before heavy repair. That extra step prevents E.10 from carrying substantive recovery content and prevents each neighboring pattern from repeating the whole RSIR diagnosis.

The pattern also keeps useful source vocabulary alive. Engineers can still say interface, API, role, parameter, function, and endpoint. FPF simply refuses to let those words select ontology by themselves.

The cost is one explicit stop: after the direct pattern is clear, RSIR must stop. Otherwise it becomes the giant repair pattern it was created to avoid.

#Rationale

The RSIR cluster needs a first-level pattern because E.10 should remain a trigger and lexical-governance pattern, while A.6.P, A.6.5, A.6.M, A.6.F, A.2, A.15, and publication, evidence, and status patterns each govern only their respective objects.

The main ontological principle is participant, declaration, application and binding, assertion and designation, and representation separation. An actual direct-relation participant retains its direct kind under one participant meaning. A corresponding SlotSpec, when reusable typed relation declaration is current, states a declaration-local SlotKind, exact ValueKind, and refMode. In an assertion or description, C.2.1 governs the episteme's identity and content, the direct claim family governs predicate, polarity, or use, and A.6.5 governs participant-designation typing only against a compatible current SlotSpec; an ordinary assertion can name actual participants without one. An A.6.1 ArgumentDeclaration or ResultDeclaration states reusable operation meaning, while one exact application and obtaining binding relate that independently identified occurrence to an actual bound value. A C.29 representation position may correspond to any of those meanings without becoming the participant, declaration, application, or binding.

The second principle is direct governance. Once the current object is recovered, the pattern that governs that object governs the repair. RSIR only identifies the direct governing pattern.

#SoTA-Echoing

This pattern does not introduce new external SoTA sources beyond the source uses already admitted by E.24 for ontic introduction. It applies those source uses to the narrower RSIR recovery problem.

#Relations

E.10 detects trigger wording. E.10.ARCH states that RSIR is the first-level restoration pattern for this bounded cluster when the direct governing pattern is not already clear.

A.6.5 governs complete declaration-local SlotSpec = <SlotKind, ValueKind, refMode> content inside one exact RelationSignature and, only when a compatible SlotSpec is current, participant-designation typing. C.2.1 still governs the assertion or description episteme's identity and content, and the direct claim family governs predicate, polarity, and use. An ordinary assertion may designate actual participants directly without reusable declaration.

A.6.P governs relation precision restoration after the recovered object is a relation or relation-bearing claim.

A.6.0 governs U.Signature; A.6.1 governs operation argument and result declaration content plus any independently identified exact application and declaration-local binding; E.20 governs mechanism introduction. A.6.1 admits no public OperationApplication U-kind or universal input/output/result relation, and its result binding alone establishes none of production, a produced entity, a result episteme, evidence, or work.

A.2, A.2.1, A.2.2, A.2.5, A.2.7, A.15, and Part F role-description and naming patterns govern role, role assignment, capability, role state, role relation structure, role-method-work, and durable role-name claims.

A.6.M, A.6.F, A.6.A, A.3.4.P, E.18, C.30, C.30.ASV, C.30.AD, and C.30.TFS-REL govern module-interface, functional, affordance, transformation, transformation-flow, architecture-of, structural-view, and architecture-description cases.

C.2.1, E.17, C.2.P.DR, A.10, B.3, G.6, F.10, and C.28 govern episteme identity and content, publication, declarative representation, evidence, assurance, provenance, status, and causal-use cases; the exact direct claim family still governs the predicate, polarity, or use asserted through that content.

#A.6.RSIR:End

Last Updated: 2026-08-05 — upstream FPF commit 3dbce514 (github.com/ailev/FPF)