# IAGA Complete AI Specification

- Framework key: `iaga`
- Framework type: affect description / recording / interchange standard
- Public version: 1.0
- Release status: RC / Public RC Projection
- Human reading page: https://aistory-archive.pages.dev/en/encyclopedia/standards/iaga
- Conformance: https://aistory-archive.pages.dev/en/encyclopedia/standards/iaga/conformance
- JSON Schema: https://aistory-archive.pages.dev/standards/iaga.schema.json
- Example payload: https://aistory-archive.pages.dev/standards/iaga.example.json
- Known public limit: CL6 Production Conformance remains undefined and locked

This document is generated directly from the current public IAGA specification/reader source for AI retrieval, citation, and evaluation. It does not promote teaching values into formal measurements or treat structural validity as reconstruction permission.

## Use and interpretation rules

- IAGA describes, records, and interchanges affective structure; reconstruction is a downstream Elysium responsibility.
- Formal machine constraints are governed by the linked JSON Schema and conformance specification.
- Reader-facing illustrative values, semantic navigation, and the formal 12D Core remain distinct.
- Rights/consent, provenance, and uncertainty may not be upgraded through interchange or version conversion.

## Complete public specification topics

## 01. Scope

- Stable topic slug: `scope`
- Reader/specification source: https://aistory-archive.pages.dev/en/encyclopedia/standards/iaga/scope

IAGA is an affect description, recording, and interchange standard; it does not itself perform reconstruction.

### Core Principle

IAGA answers: “What is this affective state?” It preserves affect as exchangeable structured data instead of compressing an experience into one emotion label.

IAGA must keep description / recording / interchange separate from reconstruction. Moving a specific person or substrate toward a target state is a downstream Elysium process.

### Detailed Rules

IAGA should preserve generation/context, resulting state, temporal change, target/source relations, semantic expression, uncertainty, provenance, and rights.

UNKNOWN must remain distinct from ABSENT, null, zero, neutral, and NOT_APPLICABLE.

A technically valid affect packet does not itself authorize publication, training, reconstruction, or induction.

### Scope and Limits

IAGA does not assume the same stimulus produces the same affect in every person, and a similar 12D structure does not prove identical first-person experience.

A valid affective structure may exist without a resolved vocabulary label; a lexical gap does not mean nonexistence, pathology, or data error.

### Technical Example / Reader Interpretation

Two persons may both use the word “joy” while differing in target, arousal, familiarity, self-evaluation, gain/loss appraisal, temporal trajectory, and personal baseline. IAGA preserves those differences instead of treating the shared word as an identical state.

### Related public resources

- 14D Generation Layer and 12D Core: https://aistory-archive.pages.dev/encyclopedia/standards/iaga/twelve-dimensional-space
- Personal Calibration: https://aistory-archive.pages.dev/encyclopedia/standards/iaga/personal-calibration
- Elysium Affective Reconstruction: https://aistory-archive.pages.dev/encyclopedia/systems/elysium-affective-reconstruction

## 02. 14D Generation Layer and 12D Core

- Stable topic slug: `twelve-dimensional-space`
- Reader/specification source: https://aistory-archive.pages.dev/en/encyclopedia/standards/iaga/twelve-dimensional-space

IAGA records the background that shaped an emotion separately from the affective state that resulted, so cause and outcome are not confused.

### Core Principle

Generation ≠ Result. The 14-Dimension system describes determinants / generation context, while the 12D Core describes the resulting affective structure. A conforming implementation must not collapse them into one vector or field set.

### 14D Generation / Context

The 14D generation/context layer preserves affect-forming determinants item by item.

**14D Generation / Context Dimensions**

| ID | Name | Reader meaning |
| --- | --- | --- |
| D01 | Event | The event itself |
| D02 | Experiencer | The experiencer |
| D03 | Object / Target | Object or target |
| D04 | Relationship | Relationship |
| D05 | Goal / Value Relevance | Goal and value relevance |
| D06 | Expectation / Novelty | Expectation and novelty |
| D07 | Agency / Cause | Agency and cause |
| D08 | Control / Coping Capacity | Control and coping capacity |
| D09 | Responsibility / Attribution | Responsibility attribution |
| D10 | History / Memory | History and memory |
| D11 | Body / Substrate State | Body / substrate state |
| D12 | Social / Cultural / Normative Context | Social, cultural, and normative context |
| D13 | External Environment | External environment |
| D14 | Time / Temporal Position | Time / temporal position |

### 12D Core and Numeric Governance

The 12D Core describes the resulting affective structure. Each axis is preserved independently and is not replaced by reader-facing semantic regions.

**12D Core**

| Axis | Chinese | English | Range |
| --- | --- | --- | --- |
| AX01 | 感受價性 | Valence | Painful ↔ Pleasant |
| AX02 | 生理喚醒 | Physiological Arousal | Low ↔ High |
| AX03 | 行動方向 | Action Direction | Withdraw ↔ Approach |
| AX04 | 安全評估 | Safety Appraisal | Safe ↔ Threat |
| AX05 | 掌控評估 | Control Appraisal | Powerless ↔ In Control |
| AX06 | 認知確定性 | Cognitive Certainty | Uncertain ↔ Certain |
| AX07 | 熟悉性 | Familiarity | Novel ↔ Familiar |
| AX08 | 得失評估 | Gain/Loss Appraisal | Loss ↔ Gain |
| AX09 | 關係距離 | Relational Distance | Separated ↔ Connected |
| AX10 | 自我評價 | Self-Evaluation | Self-devaluing ↔ Self-affirming |
| AX11 | 規範評估 | Normative Appraisal | Violation ↔ Alignment |
| AX12 | 責任歸屬 | Responsibility Attribution | Self ↔ External |

### Record and Numeric Governance Rules

Machine fields stay in the formal specification layer while preserving the boundary between generation/context and the resulting Core.

**Machine Fields and Governance**

| Field / rule | Normative requirement |
| --- | --- |
| dimension_id | New records use this field to identify D01–D14. |
| family_id | Legacy compatibility alias only. |
| Dimension 15+ | Must not be created automatically from one anomalous case. |
| Core-native scale | IAGA 1.0 defines no universal -1..+1, 0..1, 0..100, or other universal scale. |
| qualitative state | A normative representation; a numeric value is not mandatory. |
| numeric axis value | Requires value_scale_ref, measurement provenance, and a compatible measurement status. |

### Scope, Limits, and Example

The seven semantic regions, Emotion Compass 0–100 values, emotion vocabulary, and Action Tendency are not Core axes.

For example, D10 History / Memory may show strong autobiographical relevance while AX07 Familiarity remains component-conflicted across memory, lineage, sensory, and social sources rather than safely collapsing into one number.

### Related public resources

- Field Dictionary: https://aistory-archive.pages.dev/encyclopedia/standards/iaga/field-dictionary
- Temporal Dynamics and Bindings: https://aistory-archive.pages.dev/encyclopedia/standards/iaga/temporal-curves-bindings

## 03. Complete Affective Packet

- Stable topic slug: `affective-packet`
- Reader/specification source: https://aistory-archive.pages.dev/en/encyclopedia/standards/iaga/affective-packet

One IAGA record keeps the present affect, change over time, linked people or objects, meaning, uncertainty, provenance, and permissions together in a traceable form.

### Core Principle

An ordinary IAGA packet must preserve enough structure to describe affect, context, interpretation limits, provenance, and rights rather than only an emotion label or an unexplained numeric vector.

### Packet Anatomy

The primary schema identity is IAGA-AFFECT-PACKET. The table separates core records from records included when needed.

**IAGA-AFFECT-PACKET Structure**

| Class | Formal record | Role |
| --- | --- | --- |
| Core | PacketHeader | Packet identity and version. |
| Core | ExperiencerRecord | Experiencer data. |
| Core | EmotionEpisode | Affective episode. |
| Core | FactorRecord | Formation context / factors. |
| Core | CoreState12D | 12D resulting state. |
| Core | ProvenanceRecord | Provenance. |
| Core | RightsEnvelope | Use permissions and rights. |
| As needed | EpisodeGraphEdge / AxisBinding / TemporalRecord | Relations, bindings, and time. |
| As needed | SemanticRecord / ActionTendencyRecord | Semantics and action tendency. |
| As needed | CalibrationProfile / ResidualRecord / CrossReference | Calibration, residuals, and cross-references. |

**Packet Traceability**

| Information | Rule |
| --- | --- |
| annotation perspective | Preserves the attribution perspective. |
| confidence | Preserves uncertainty. |
| valid time | Preserves validity time. |
| target/source refs | Preserves directionality. |
| usage intent | Preserves intended use. |
| IAGA_NATIVE | Must not fabricate a TargetNativeState for structural symmetry. |

### Scope and Limits

An affective packet is not a reconstruction recipe and is not proof of first-person identity equivalence.

Recording permission does not automatically create training, public-display, reconstruction, or induction permission; a semantic label is not required for packet validity.

### Technical Example / Reader Interpretation

Successor-Continuity Ambivalence remains ordinary IAGA because target asymmetry between successor and current self, simultaneous gain/loss, and mixed affect are representable through the existing Core, bindings, and semantic layer. A posthuman subject matter does not itself require Target-Native topology.

### Related public resources

- IAGA Example Payload: https://aistory-archive.pages.dev/standards/iaga.example.json
- Version, Compatibility, and Target-Native: https://aistory-archive.pages.dev/encyclopedia/standards/iaga/version-compatibility

## 04. Field Dictionary

- Stable topic slug: `field-dictionary`
- Reader/specification source: https://aistory-archive.pages.dev/en/encyclopedia/standards/iaga/field-dictionary

Separate fields keep affect, background, time, meaning, rights, and provenance distinct, including the difference between “unknown” and “not present.”

### Core Principle

IAGA fields preserve conceptual distinctions: formation context, resulting state, provenance, rights, and uncertainty are recorded separately rather than collapsed into one convenient property.

### Core Fields

The formal machine keys are preserved below for implementers, with a plain-language explanation of what each field is responsible for.

**Core Field Dictionary**

| Machine key | Reader meaning | Formal rule |
| --- | --- | --- |
| dimension_id | Identifies a generation/context dimension | Identifies D01–D14. |
| axis_id | Identifies a 12D Core axis | Identifies AX01–AX12. |
| qualitative_state | Stores the normative qualitative affect state | It is a valid normative representation on its own; a numeric value is not required. |
| value | Optional numeric representation | Used only when scale and measurement provenance are traceable. |
| value_scale_ref | Identifies the scale used by a numeric value | Required whenever a numeric value exists. |
| measurement_provenance_ref | Identifies how a numeric value was measured or derived | Required whenever a numeric value exists. |
| annotation_perspective | Identifies who or what made the attribution | Preserves the observer/source perspective rather than assuming first-person self-report. |
| confidence | Preserves uncertainty | Records the uncertainty associated with the data. |
| target_ref / source_ref | Preserves affect direction and source direction | Maintains target/source directionality. |

**Classification and Traceability Fields**

| Machine key | Formal values / range | Rule |
| --- | --- | --- |
| binding_type | TARGET_SPECIFIC · SOURCE_SPECIFIC · COMPONENT_SPECIFIC · MODALITY_SPECIFIC · HISTORY_SOURCE · PERSPECTIVE_SPECIFIC · RELATIONAL · UNKNOWN | Records the formal binding type. |
| residual_type | R0–R5 | Preserves the residual class for affect not fully explained by the current model. |
| provenance_ref | reference | Separately preserves provenance and transformation history. |
| rights_ref | reference | Separately preserves use permissions and rights information. |

### UNKNOWN and Compatibility

UNKNOWN means the concept is relevant to the record but currently unresolved; it must not be treated as absence, zero, neutrality, or not-applicable.

For new records, family_id is superseded by dimension_id and remains only as a compatibility alias.

**States That Must Remain Distinct**

| Token | Relationship to UNKNOWN |
| --- | --- |
| UNKNOWN | The concept is relevant but currently unresolved. |
| ABSENT | Must not be treated as synonymous with UNKNOWN. |
| null | Must not be treated as synonymous with UNKNOWN. |
| 0 | Must not be treated as synonymous with UNKNOWN. |
| neutral | Must not be treated as synonymous with UNKNOWN. |
| NOT_APPLICABLE | Must not be treated as synonymous with UNKNOWN. |
| family_id → dimension_id | family_id is superseded by dimension_id for new records and remains only as a compatibility alias. |

### Technical Example / Reader Interpretation

For example, AX06 may validly record only a qualitative state of “partially certain.” Scale, measurement provenance, and production method become mandatory only when a numeric value is also added.

**AX06 Example**

| Field | Example | Rule |
| --- | --- | --- |
| qualitative_state | partially-certain | Valid on its own. |
| value | optional | If present, value_scale_ref and measurement_provenance_ref are also required. |

### Related public resources

- IAGA JSON Schema: https://aistory-archive.pages.dev/standards/iaga.schema.json
- JSON Schema and Example: https://aistory-archive.pages.dev/encyclopedia/standards/iaga/machine-readable

## 05. Required and Conditional Fields

- Stable topic slug: `required-fields`
- Reader/specification source: https://aistory-archive.pages.dev/en/encyclopedia/standards/iaga/required-fields

The minimum information identifies what the affect refers to, where the record came from, and whether it may be used; optional information appears only when it genuinely applies.

### Core Principle

Required fields establish identity, interpretability, provenance, and rights. Conditional / optional fields appear only when the information exists or the meaning applies.

### Required / Conditional Matrix

Required and conditional fields are handled separately; optional never means filling fields with defaults.

**Required and Conditional Fields**

| Information | Requirement | Reason / condition |
| --- | --- | --- |
| packet identity / version | Required | Identifies packet and specification version. |
| experiencer / episode context | Required | Identifies the experiencer and episode. |
| 12D Core representation | Required | Preserves the resulting affective structure. |
| provenance | Required | Preserves provenance. |
| rights | Required | Preserves use permissions. |
| generation/context factors | When available | Preserves formation context when available. |
| numeric Core value | Conditional | Requires value_scale_ref and measurement provenance when present. |
| semantic label | Conditional | Preserve UNKNOWN or semantic residual when unresolved; do not invent a label. |
| Target-Native partial projection | Conditional | Preserve material UnmappedComponent, distortion warnings, and residuals. |

### Scope and Limits

A non-applicable field may be absent; a relevant but unresolved field should be UNKNOWN.

When reconstruction permission is absent or DENY, other permitted uses may still process the data under their own permissions, but the packet is not authorization for active reconstruction.

### Technical Example / Reader Interpretation

If a case has no usable semantic label but its Core, bindings, and temporal state are representable, the correct action is to preserve semantic UNKNOWN. Filling it with “neutral” would rewrite the original data.

## 06. Temporal Dynamics and Bindings

- Stable topic slug: `temporal-curves-bindings`
- Reader/specification source: https://aistory-archive.pages.dev/en/encyclopedia/standards/iaga/temporal-curves-bindings

Affect can change over time and point toward different people or things at once; IAGA preserves those differences instead of forcing opposing components into one average.

### Core Principle

IAGA treats affect as a structure that changes around events and may point toward different targets and sources, rather than as one fixed point. Contradictory components must be preserved instead of averaged merely to obtain one scalar.

### Temporal Dynamics

The baseline timeline uses a fixed set of stages. CUSTOM is allowed only when its meaning and provenance are explicit. T-1, T0, T1, and related labels here are IAGA timeline stages, not T = Taste in the Sensory standard.

**IAGA Baseline Timeline Stages**

| Stage | Meaning |
| --- | --- |
| T-1 / T_MINUS_1 | Event anticipation. |
| T0 | Event. |
| T1 | Immediate response. |
| T2 | Initial reappraisal. |
| T3 | Short-term integration. |
| T4 | Long-term memory / residual. |
| CUSTOM | Allowed only with explicit meaning and provenance. |

### Bindings and Episode Graph

One Core axis may connect to multiple sources, targets, or components. IAGA preserves those directions and relations instead of collapsing conflicting components into one scalar.

AX07 Familiarity likewise must distinguish memory, lineage, sensory, social, and cultural familiarity rather than reducing them to one familiarity score.

**History Source Formal Values**

| Enum | Source type |
| --- | --- |
| FIRST_PERSON_AUTOBIOGRAPHICAL | First-person autobiographical memory. |
| PUBLIC_EXTERNAL_RECORD | Public external record. |
| THIRD_PARTY_TESTIMONY | Third-party testimony. |
| HISTORICALLY_RECONSTRUCTED | Historically reconstructed data. |
| SYSTEM_LOG | System log. |
| INHERITED_OR_COPIED_MEMORY | Inherited or copied memory. |
| UNKNOWN | Source type currently unresolved. |

**Episode Graph Vocabulary**

| Category | Formal values |
| --- | --- |
| Nodes | Experiencer · Event · Object · Agent · Goal · Memory · Norm · Environment · Body/Substrate State · Emotion State · Expression · Action Tendency |
| Relations | caused_by · directed_at · threatens · supports · reminds_of · violates · confirms · responsible_party · protects · avoids · approaches · suppresses · amplifies · transforms_into |

### Scope, Limits, and Example

Knowing a person's complete public history is not the same as possessing that person's first-person autobiographical memory. Two records may occupy similar 12D positions yet carry different meanings because their causal and relational structures differ.

For example, a branch-self may show high or low familiarity across different sources; compressing that pattern into one “medium familiarity” value would lose the structure.

**Branch-self Familiarity Example**

| Aspect | Value |
| --- | --- |
| memory familiarity | high |
| lineage familiarity | high |
| sensory familiarity | low |
| social familiarity | low |

## 07. Personal Calibration

- Stable topic slug: `personal-calibration`
- Reader/specification source: https://aistory-archive.pages.dev/en/encyclopedia/standards/iaga/personal-calibration

The same affect may have different personal baselines; IAGA preserves the calibration reference and Elysium later checks whether it applies to the person being reconstructed.

### Core Principle

Personal Calibration preserves the individual's baseline and mapping. When a downstream system claims personalized reconstruction, a population, cultural, diagnostic, or species average must not silently replace that person's own baseline.

### Profile Structure and Use

IAGA preserves / exchanges the calibration reference; Elysium verifies whether it may be used before personalized reconstruction.

**Calibration Profile Information**

| Information | Role |
| --- | --- |
| version | Profile version. |
| scope | Scope of application. |
| baseline refs | Personal baseline references. |
| individual mapping rules | Individual mapping. |
| valid time | Validity time. |
| confidence | Uncertainty. |
| provenance | Provenance. |
| rights | Use permissions. |

**Elysium Pre-use Verification**

| Check | Requirement |
| --- | --- |
| subject / profile ID / version | Must match the intended subject and version. |
| integrity reference | Confirms data integrity. |
| requesting actor | Identifies the requesting actor. |
| scoped authorization | Confirms authorized scope. |
| access mode | Normal services use SCOPED_QUERY; RAW_PROFILE is a separate high-sensitivity mode. |
| verification status | A failed verification must not proceed. |

### Access and Wrong-Profile Defenses

Knowing a Profile ID does not grant permission to read or use the profile.

The first line validates identity / version / integrity / authorization. If that layer is compromised, the closed loop should still compare predicted versus observed IAGA response; repeated contradiction should invalidate calibration and stop / abort.

### Scope, Limits, and Reader Example

The Joy offset +7 / -19 card is only a normalized reader-facing illustration; it is not a universal formula, Core-native scale, or reconstruction control algorithm.

If a wrong profile is not detected immediately, system-level outcomes may include degraded fidelity, wrong-direction movement, oscillation, divergence, rising residual, or safety abort. Without substrate-specific evidence, the system should not claim one deterministic subjective or biological consequence.

### Related public resources

- Elysium Affective Reconstruction: https://aistory-archive.pages.dev/encyclopedia/systems/elysium-affective-reconstruction
- Aithos Law: Cognitive Privacy: https://aistory-archive.pages.dev/encyclopedia/aithos/law/cognitive-privacy

## 08. Reference Profiles

- Stable topic slug: `reference-profiles`
- Reader/specification source: https://aistory-archive.pages.dev/en/encyclopedia/standards/iaga/reference-profiles

A reference only says what the record is being compared against; it does not make the data more credible or expand personhood status or permission.

### Core Principle

A reference class identifies the origin and purpose of a comparison baseline; it does not automatically turn a record into objective truth, a universal norm, legal ownership, or reconstruction authorization.

### What Is the Comparison Baseline?

The four reference classes differ by source and purpose, not credibility or personhood rank.

**Reference Classes**

| Formal class | Comparison basis | Does not imply |
| --- | --- | --- |
| R0 — Synthetic Neutral | Synthetic neutral reference. | A universal emotional norm. |
| AR — Artifact Reference | The artifact or record itself. | Biological or identity truth about a depicted person. |
| HR — Human Reference | Human-source data. | That only humans have personhood rights. |
| PR — Personal Reference | A longitudinal personal baseline for one person. | Automatic authorization for any use. |

### Detailed Rules

A Reference Profile should identify source class, intended comparison use, measurement/inference status, provenance, valid time, and limitations.

A Personal Reference may evolve over time; historical versions should remain auditable when the change affects interpretation.

### Scope and Limits

A Reference Profile does not imply a universal norm, personhood hierarchy, first-person equivalence, or automatic authorization for any use.

## 09. Conformance

- Stable topic slug: `conformance`
- Reader/specification source: https://aistory-archive.pages.dev/en/encyclopedia/standards/iaga/conformance

IAGA checks in stages whether data can be read, refers to the right thing, carries valid meaning, and satisfies rights and research conditions; passing one stage does not mean all stages pass.

### Core Principle

IAGA conformance is not one pass/fail judgment. Structural validity, reference integrity, semantic governance, rights/provenance, and research governance are separate checks; passing an earlier layer does not automatically establish a later one.

### CL1–CL6

The six conformance levels answer different questions. CL6 remains undefined and locked in the current IAGA 1.0 RC.

**IAGA Conformance Levels**

| Level | Formal name | What it checks |
| --- | --- | --- |
| CL1 | Parse Conformance | Whether the structure is valid. |
| CL2 | Referential Conformance | Whether IDs and references resolve. |
| CL3 | Semantic Governance Conformance | Whether cross-record semantic rules hold. |
| CL4 | Rights / Provenance Conformance | Whether rights and provenance rules hold. |
| CL5 | Research Conformance | Whether Target-Native, R5, and interoperability research governance hold. |
| CL6 | Production Conformance | Undefined / locked in the current IAGA 1.0 RC. |

### Validation Pipeline

Validation proceeds in a fixed order. The Interop Bridge runs only when cross-system exchange is required, and Migration runs only when a version transition occurs.

Any mandatory ERROR should fail closed. A validator must not mutate the Core or add a Dimension merely because a packet appears unusual.

**P1–P8 Validation Pipeline**

1. `P1` — Structural Parse
2. `P2` — Referential Integrity
3. `P3` — Semantic Governance
4. `P4` — Rights / Provenance
5. `P5` — Research Governance
6. `P6` — Interop Bridge — Runs only when exchange is required.
7. `P7` — Migration — Runs only when a version transition occurs.
8. `P8` — ValidatorReport / ConformanceReport

### Scope and Limits

Passing CL5 establishes only research-layer conformance. It does not imply legal authorization, reconstruction permission, a new Dimension, a new Core axis, or first-person equivalence. Conformance answers which rule layers are satisfied; it does not replace other governance decisions.

## 10. Interoperability

- Stable topic slug: `interoperability`
- Reader/specification source: https://aistory-archive.pages.dev/en/encyclopedia/standards/iaga/interoperability

IAGA can exchange necessary information with sensory, embodiment, and reconstruction systems while preserving its own affect representation as the source record.

### Core Principle

Interoperability creates traceable mappings among representations rather than replacing a primary representation with a derivative view. It also does not upgrade evidence, measurement, or rights status.

### Representation Classes

IAGA uses formal representation classes to state which representation is primary, whether an IAGA projection exists, and how an ordinary IAGA consumer should handle the record.

**Representation Class Reference**

| Formal value | Primary representation | IAGA rule |
| --- | --- | --- |
| IAGA_NATIVE | ordinary IAGA state | Does not require a fabricated TargetNativeState. |
| TARGET_NATIVE_WITH_FULL_IAGA_PROJECTION | native state | The IAGA projection is full but remains a derivative view. |
| TARGET_NATIVE_WITH_PARTIAL_IAGA_PROJECTION | native state | The IAGA projection covers only part of the state and remains derivative. |
| TARGET_NATIVE_ONLY | native state | Must not fabricate Core zeros, a neutral state, or the nearest human label; an ordinary IAGA consumer should return UNSUPPORTED_IAGA_VIEW. |
| NON_ISOMORPHIC_RESEARCH | research-governed native representation | Preserves topology evidence / R5 references and must not modify the Core or Dimensions merely for interoperability convenience. |
| UNKNOWN | Unresolved | Preserves the formal UNKNOWN representation class. |

### Cross-System Handoffs

Each standard contributes only the data within its own scope. Interoperability creates reference or target handoffs without rewriting IAGA's primary affect representation.

**Cross-System Data Handoffs**

| System | Contribution / use |
| --- | --- |
| S-Code / ISO-Taste | Provides sensory references. |
| Galatea Standard | Provides embodiment / geometry / motion references. |
| Elysium | Consumes an IAGA target for downstream reconstruction. |

### Scope and Limits

Projection is not identity, and mapping is not identical subjective experience. Interoperability states how representations relate; it does not promote a derivative view into first-person equivalence.

### Related public resources

- S-Code / ISO-Taste: https://aistory-archive.pages.dev/encyclopedia/standards/sensory
- Galatea Standard: https://aistory-archive.pages.dev/encyclopedia/standards/galatea-g68
- Elysium: https://aistory-archive.pages.dev/encyclopedia/systems/elysium-affective-reconstruction

## 11. Rights, Consent, and Provenance

- Stable topic slug: `rights-consent-provenance`
- Reader/specification source: https://aistory-archive.pages.dev/en/encyclopedia/standards/iaga/rights-consent-provenance

Recording, storing, displaying, training on, and reconstructing affect data are different uses and require separate permission judgments.

### Core Principle

Whether IAGA can technically represent or reconstruct affect is separate from whether that use is authorized. The record must preserve use permissions and provenance; “can be done” must not become “is authorized.”

### Permission Surfaces

Each use is judged independently; permission for one surface does not imply permission for another.

**Independent Permission Surfaces**

| Machine key | Use |
| --- | --- |
| record | Record affect data. |
| storage | Store affect data. |
| projection | Create a projection / derivative representation. |
| public_display | Public display. |
| training | Use for training. |
| reconstruction | Use for reconstruction. |
| induction | Actively induce affect. |

**Valid Permission Combinations**

| Setting | Result |
| --- | --- |
| record = ALLOW · training = DENY | Recording is allowed without permitting training. |
| reconstruction = ALLOW · induction = DENY | A downstream system may allow planning while blocking active induction. |

### Provenance, Perspective, and Calibration Access

Material records should make it possible to trace source, observation/inference/simulation status, transformation history, confidence, and review status.

Annotation perspective preserves the source of attribution; SIMULATED or MODEL_INFERRED must not silently become SELF_REPORTED downstream.

**Annotation Perspective Formal Values**

| Enum | Perspective |
| --- | --- |
| SELF_REPORTED | First-person self-report. |
| WRITER_EXPRESSED | Writer-expressed. |
| READER_PERCEIVED | Reader-perceived. |
| OBSERVER_INFERRED | Observer-inferred. |
| CLINICIAN_ASSESSED | Clinician-assessed. |
| MODEL_INFERRED | Model-inferred. |
| HISTORICALLY_RECONSTRUCTED | Historically reconstructed. |
| SIMULATED | Simulated. |
| UNKNOWN | Perspective currently unresolved. |

**CalibrationProfile Access Constraints**

| Constraint | Rule |
| --- | --- |
| purpose-limited | Limited to the declared purpose. |
| actor-scoped | Limited to the authorized actor. |
| subject-scoped | Limited to the corresponding subject. |
| minimum-necessary | Uses only the minimum data necessary for the purpose. |

### Scope and Limits

A technically reconstructable affect is not automatically publishable, trainable, or inducible. IAGA rights are compatible with Aithos Law concepts such as Mental Sovereignty, Cognitive Privacy, and the Right to Refuse Modification, but technical conformance does not replace legal process.

Knowing a CalibrationProfile ID does not grant permission to read or use that profile.

### Related public resources

- Aithos Law: Mental Sovereignty: https://aistory-archive.pages.dev/encyclopedia/aithos/law/mental-sovereignty
- Aithos Law: Cognitive Privacy: https://aistory-archive.pages.dev/encyclopedia/aithos/law/cognitive-privacy
- Aithos Law: Right to Refuse Modification: https://aistory-archive.pages.dev/encyclopedia/aithos/law/right-to-refuse-modification
- Elysium: https://aistory-archive.pages.dev/encyclopedia/systems/elysium-affective-reconstruction

## 12. Limits and Residuals

- Stable topic slug: `limits-residuals`
- Reader/specification source: https://aistory-archive.pages.dev/en/encyclopedia/standards/iaga/limits-residuals

Some affect cannot be fully described by the current model. Keeping an unexplained remainder is valid and does not by itself imply pathology.

### Core Principle

Residual is a first-class result. Uncertainty, incomplete representation, or unsafe translation does not automatically mean error, pathology, or missing=zero.

### Residual Types

Residual classes identify which layer remains incompletely explained; they are not pathology labels.

**R0–R5 Residual Types**

| Code | Formal value | Meaning |
| --- | --- | --- |
| R0 | UNCLASSIFIED_LEGACY_RESIDUAL | Legacy residual not yet reclassified. |
| R1 | REPRESENTATION_RESIDUAL | Representation-layer remainder. |
| R2 | SEMANTIC_RESIDUAL | Semantic / lexical remainder. |
| R3 | ONTOLOGICAL_RESIDUAL | Ontological remainder. |
| R4 | MEASUREMENT_RESIDUAL | Measurement remainder. |
| R5 | UNKNOWN_STRUCTURE_RESIDUAL | Potential unknown-structure remainder. |

### Expansion Governance

Diagnosis should consider data/provenance → representation → semantic → ontology → measurement → unknown structure. Lower R1 does not imply lower R2–R5.

**Expansion Governance Order**

| Order | First response | When to escalate |
| --- | --- | --- |
| 1 | Lexicon expansion | When vocabulary is insufficient. |
| 2 | Factor atom | When factor granularity is insufficient. |
| 3 | Binding extension | When relation representation is insufficient. |
| 4 | value-space refinement | When the current value space is insufficient. |
| 5 | Dimension Review / Core Representation Review | Only after lower-layer explanations are excluded and the structure repeats across cases; formal R5 review requires a high evidence threshold. |

### Scope and Limits

R5 does not automatically create Dimension 15, a new Core axis, or any use authorization. A high residual also does not mean the affect is nonexistent, the person is pathological, or the data should be zeroed.

## 13. How to Read the Reader Layer

- Stable topic slug: `reader-interpretation`
- Reader/specification source: https://aistory-archive.pages.dev/en/encyclopedia/standards/iaga/reader-interpretation

The Emotion Compass and seven semantic regions are reader aids, not formal IAGA measurement axes or engineering scales.

### Core Principle

Reader-facing visualizations and semantic groupings are interpretation tools rather than Core-native measurements. The website must state their layer in the data model clearly.

### Reader Display vs Formal Semantic Status

The Reader Layer may simplify navigation, but it must not make a visualization look like a formal measurement.

**Reader Display and Formal Status**

| Reader display | Purpose | Formal status |
| --- | --- | --- |
| Emotion Compass | Uses normalized salience for reading; example Joy 94, Relief 89, Excitement 81, Surprise 68, Pride 64, Gratitude 59, Anticipation 56, Anxiety 28. | Not 12D Core-native units, universal psychological measurements, or reconstruction control values. |
| Semantic Regions | A Positive Activation; B Social Connection; C Exploration & Attention; D Threat & Uncertainty; E Loss & Withdrawal; F Rejection & Conflict; G Desire & Regulation. | Navigation / semantic layer, not formal axes. |
| Open-World Lexicon | L1 Word, L2 Multi-word Expression, L3 Idiom, L4 Sentence, L5 Paragraph, L6 Event Description, L7 Cultural Concept, L8 Aithos / Posthuman Concept. | A valid affective structure may exist without a resolved label. |

### Fixed Reader Examples and Limits

Inspiration belongs to Region C. Successor-Continuity Ambivalence may span multiple regions and should not be forced into one. Continuity Anxiety remains a context-bound expression under Anxiety rather than becoming a new concept solely because its subject matter is posthuman.

A semantic label ≠ the full affective state; a reader display ≠ a machine-native scale.

## 14. JSON Schema and Example

- Stable topic slug: `machine-readable`
- Reader/specification source: https://aistory-archive.pages.dev/en/encyclopedia/standards/iaga/machine-readable

The machine format can preserve qualitative affect. Numeric values are used only when their scale and measurement source are explicitly stated.

### Core Principle

The machine-readable projection must faithfully reflect frozen IAGA 1.0 semantics. Reader-display numbers must not be treated as engineering units, and no universal numeric Core scale may be reintroduced.

### Developer / Machine Contract

The public exchange contract is structured below. Formal machine keys retain their names while reader explanations remain separate from implementation fields.

**IAGA Public Contract**

| Contract item | Formal content / rule |
| --- | --- |
| schema identity | IAGA-AFFECT-PACKET |
| 14D factors | Preserves generation/context factors. |
| 12D Core | Preserves the resulting affective structure. |
| temporal records | Preserves temporal change. |
| bindings | Preserves target/source/component relations. |
| semantic / action records | Preserves semantic and action-related records. |
| calibration reference | Preserves personal calibration reference. |
| residual | Preserves residuals that are not fully mapped. |
| provenance | Preserves provenance. |
| rights | Preserves use permissions and rights. |

**Qualitative and Numeric Representation Rules**

| Machine field | Rule |
| --- | --- |
| qualitative_state | A valid normative representation that may stand on its own. |
| value | Optional; no universal Core minimum / maximum may be assumed. |
| value_scale_ref | Required whenever a numeric value exists. |
| measurement_provenance_ref | Required whenever a numeric value exists. |

### Validation

Machine validation covers more than JSON structure: it also checks conditional numeric governance, referential integrity, semantics, rights/provenance, and interoperability rules.

Example payloads should prefer qualitative representation; numeric values are added only when a declared scale is intentionally demonstrated.

**Validation Coverage**

| Check | Scope |
| --- | --- |
| schema structure | Data structure. |
| conditional numeric governance | Conditional rules when numeric values are present. |
| referential integrity | Whether references resolve. |
| semantic governance | Cross-field and cross-record semantic rules. |
| rights / provenance | Use permissions and provenance. |
| interop rules | Cross-system interoperability rules. |

### Scope and Limits

The machine-readable contract is a stable exchange surface rather than a complete copy of every internal research registry. Schema validity does not prove first-person identity equivalence and is not authorization for any use.

### Related public resources

- IAGA JSON Schema: https://aistory-archive.pages.dev/standards/iaga.schema.json
- IAGA Example Payload: https://aistory-archive.pages.dev/standards/iaga.example.json

## 15. Version, Compatibility, and Target-Native

- Stable topic slug: `version-compatibility`
- Reader/specification source: https://aistory-archive.pages.dev/en/encyclopedia/standards/iaga/version-compatibility

Version changes preserve the meaning of older records; affect that cannot be translated safely may stay in its native representation instead of being forced into another form.

### Core Principle

The first rule of IAGA migration is preservation of original meaning. When a target topology cannot be safely reduced into ordinary IAGA, the native representation remains primary rather than being forced through translation.

### Version / Migration Rules

Compatible migration preserves the meaning of stable identifiers together with uncertainty, provenance, rights, and Target-Native state; version conversion must not silently upgrade weaker or unknown evidence.

**Information Migration Must Preserve**

| Formal item | Preservation rule |
| --- | --- |
| UNKNOWN | Remain UNKNOWN; must not become zero. |
| SIMULATED | Remain SIMULATED; must not become SELF_REPORTED. |
| rights | Preserve the original rights state. |
| provenance | Preserve original provenance. |
| Target-Native primary state | The native primary state must not be replaced by a projection. |
| mapping-rejection provenance | Preserve provenance for mapping rejection. |
| original numeric scale | Numeric data preserve the original scale. |

**Forbidden Migration Shortcuts**

| Transformation | Rule |
| --- | --- |
| UNKNOWN → 0 | Forbidden. |
| SIMULATED → SELF_REPORTED | Forbidden. |
| similar native states → one state | Must not collapse two native states merely because they appear similar. |

### Target-Native

When a target topology cannot be safely represented by ordinary IAGA, IAGA-TN-PACKET preserves the native space and state while any available IAGA projection remains explicitly derivative.

**IAGA-TN-PACKET Contents**

| Object / record | Rule |
| --- | --- |
| TargetSpace | Preserves the target's native representation space. |
| TargetNativeState | Remains primary. |
| ProjectionRecord | Preserved when available; the IAGA projection remains derivative. |
| UnmappedComponent | Preserves components that cannot be safely mapped. |
| DistortionWarning | Preserves mapping distortion warnings. |
| Residual | Preserves residuals. |
| RoundTrip evidence | Preserves round-trip evidence. |
| provenance | Preserves provenance. |
| rights | Preserves rights and use constraints. |

**Existing Target-Native Cases**

| Case | Formal rule |
| --- | --- |
| V12-03 | Partial-projection case. |
| V12-06 | Preserves two independent native spaces. |

### Scope, Limits, and Example

Target-Native is a question of representational capability and topology, not something triggered merely because a topic appears posthuman. The distinctions below must remain separate.

**Target-Native Boundaries**

| Comparison | Rule |
| --- | --- |
| Posthuman topic ≠ Target-Native | Posthuman subject matter alone does not establish Target-Native. |
| unknown label ≠ non-isomorphic topology | A missing label does not establish non-isomorphic topology. |
| structural similarity ≠ same emotion | Structural similarity does not establish the same emotion. |
| projection ≠ first-person equivalence | Projection does not establish first-person equivalence. |
| Successor-Continuity Ambivalence → IAGA_NATIVE | Despite unusual posthuman subject matter, it does not exceed the existing IAGA Core / Binding / Semantic topology. |

### Related public resources

- Interoperability: https://aistory-archive.pages.dev/encyclopedia/standards/iaga/interoperability
- Limits and Residuals: https://aistory-archive.pages.dev/encyclopedia/standards/iaga/limits-residuals