db9d327854
- Source: inbox/archive/2024-04-00-albarracin-shared-protentions-multi-agent-active-inference.md - Domain: collective-intelligence - Extracted by: headless extraction cron (worker 6) Pentagon-Agent: Leo <HEADLESS>
41 lines
3.2 KiB
Markdown
41 lines
3.2 KiB
Markdown
---
|
|
type: claim
|
|
domain: collective-intelligence
|
|
description: "Group intentionality (we-intentions) can be formalized as shared anticipatory structures in multi-agent generative models"
|
|
confidence: experimental
|
|
source: "Albarracin et al., 'Shared Protentions in Multi-Agent Active Inference', Entropy 26(4):303, 2024"
|
|
created: 2026-03-11
|
|
secondary_domains: [ai-alignment]
|
|
depends_on: ["shared-anticipatory-structures-enable-decentralized-multi-agent-coordination"]
|
|
---
|
|
|
|
# Group intentionality is constituted by shared temporal anticipation structures rather than aggregated individual intentions
|
|
|
|
Albarracin et al. (2024) formalize group intentionality — the "we intend to X" that is qualitatively different from "I intend to X and you intend to X" — as shared protentions (anticipatory structures) within multi-agent generative models. This provides a mechanistic account of how collective intentions emerge from shared temporal predictions rather than from aggregating individual intentions.
|
|
|
|
The key distinction: group intentionality is not reducible to individual intentions because it is constituted by shared anticipatory structures that exist at the level of multi-agent interaction. When agents share protentions (anticipations of immediate future states), they share a temporal structure that coordinates their actions toward collective outcomes. This shared temporal structure is the substrate of "we-intentions."
|
|
|
|
This formalization bridges phenomenology (Husserl's analysis of shared temporal experience) with computational models (active inference) and provides rigorous mathematical grounding (category theory). Group intentionality is not a mysterious emergent property but a natural consequence of agents sharing the predictive/temporal components of their generative models.
|
|
|
|
## Evidence
|
|
|
|
- Albarracin et al. (2024) use category theory to formalize the mathematical structure of shared goals and group intentionality
|
|
- The framework shows that shared protentions (temporal anticipations) are sufficient to generate coordinated collective behavior without requiring agents to explicitly represent "we-intentions" as distinct from individual intentions
|
|
- Phenomenological analysis (Husserl) grounds the formalism in shared temporal experience — agents that share anticipation of collective futures naturally coordinate
|
|
- The non-reducibility of group intentionality to individual intentions is formalized as a structural property of multi-agent interaction
|
|
|
|
## Implications for Multi-Agent Systems
|
|
|
|
For AI systems and organizational design:
|
|
1. **Collective objectives as shared temporal structures**: Rather than trying to aggregate individual agent goals, design systems where agents share anticipatory structures about collective states
|
|
2. **Coordination without negotiation**: Shared protentions enable coordination without requiring explicit negotiation protocols or centralized control
|
|
3. **Measuring group intentionality**: Can be operationalized as the degree to which agents share temporal predictions about collective outcomes
|
|
|
|
---
|
|
|
|
Relevant Notes:
|
|
- [[shared-anticipatory-structures-enable-decentralized-multi-agent-coordination]]
|
|
- [[collective intelligence is a measurable property of group interaction structure not aggregated individual ability]]
|
|
|
|
Topics:
|
|
- collective-intelligence
|