(Tentative)
| Definition | Functional Requirements | Syntax | Semantics |
Definition
PA Input (PGM-PAI)
- Is produced by Basic Knowledge (BKN) in response to UR Input.
- Encapsulates a structured semantic payload containing contextual, emotional, and alignment metadata.
- Serves as the definitive input to the Personality Alignment (PAL) AIM, enabling modulation of A-User behaviour, tone, and style in accordance with User State and interaction context.
Functional Requirements
PA Input conveys the following main information elements:
| Function | Description |
| Enhanced Response Delivery | Contains the full enriched response text generated by BKN. Preserves semantic richness, coherence, and emotional tone. |
| Alignment Metadata | Includes tone, style, and pacing preferences derived from SA-To-PA or SA-To-BK. Optionally includes narrative goals or expressive constraints. |
| User State Context | Includes interpreted user state (engagement, affective tone, disposition) to guide alignment decisions. Supports optional override flags for alignment enforcement. |
| Interaction Focus | Includes the current focus of interaction (ItemID, ItemType, and optional description). Preserves continuity with UR Prompt Plan and UR Input. |
Syntax
https://schemas.mpai.community/PGM1/V1.0/data/PAInput.json
Semantics
| Label | Description |
| Header | PA-Input Header |
| ├─ Standard-PAI | The characters “PGM-PAI-V” |
| ├─ Version | Major version – 1 or 2 characters |
| ├─ Dot-separator | The character “.” separating version components |
| └─ Subversion | Minor version – 1 or 2 characters |
| PAInputID | Unique identifier for this PA-Input instance |
| MInstanceID | Identifier of the metaverse instance |
| MEnvironmentID | Identifier of the metaverse environment |
| EnhancedResponse | Full enriched response string from BK |
| UserState | Interpreted user state from SA-To-PA |
| ├─ EngagementLevel | User’s cognitive engagement level (low, moderate, high) |
| ├─ AffectiveTone | Emotional tone inferred from user response (neutral, positive, etc.) |
| └─ Disposition | Inferred user disposition or attitude |
| InteractionFocus | Current focus of interaction |
| ├─ ItemID | Identifier of the focused Item |
| ├─ ItemType | Semantic category (from ItemTypeTaxonomy.json) |
| └─ Description | Natural language description of the Item |
| AlignmentDirectives | Cues guiding Personality Alignment modulation |
| ├─ ToneModulation | Preferred tone for agent response (empathetic, assertive, etc.) |
| ├─ StylisticPreference | Preferred communication style (concise, elaborate, etc.) |
| ├─ PacingCue | Suggested pacing for agent response (slow, steady, fast) |
| └─ NarrativeGoal | Optional expressive goal (clarify, deepen, contrast, etc.) |
| Trace | Provenance metadata for PA-Input |
| ├─ SAInputID | Reference to the originating SA-Input |
| ├─ ExpressiveStateGuideID | Reference to the originating ExpressiveStateGuide |
| ├─ SAInputID | Reference to the originating SA Input. |
| ├─ BKFragmentID | ID of the BK fragment that generated the response |
| └─ Timestamp | Time of PA-Input creation (see Time.json) |