Story-Agent Research Completion Summary

Indigenous Protocols for Relational Narrative Voice — IAIP Implementation Guide

Research Completion Date: March 5, 2026
Orchestration Pattern: 5 parallel deep research agents (Opus, claude-opus-4.6)
Total Research Output: 500+ KB across 9 documents
Status: ✅ SYNTHESIS READY FOR COMMUNITY REVIEW

What Was Delivered

Core Synthesis Document (START HERE)

📄 SYNTHESIS-StoryAgentNarrativeVoice-260305.md (50KB, 50+ sections)

This is the primary deliverable. It contains:

• ✅ 4 concrete Indigenous teaching stories analyzed for multi-vocal narrative structure
• ✅ 10+ relational protocols documented (Anishinaabek, Mi'kmaq, Coast Salish, Haida, Stó:lō)
• ✅ 3 modes of relational voice (Speak For / Speak As / Speak With) with clear ethical boundaries
• ✅ Per-ring narrative positioning (how People, Land, Cosmos, Ideas, Markets each speak authentically)
• ✅ 8-principle Anti-Appropriation Playbook (what to do, what NOT to do)
• ✅ First-person voice mechanics for Story-Agents by ring
• ✅ Implementation table mapping protocols to IAIP architecture

Read this first. It's comprehensive and actionable.

Supporting Research Documents

1. Teaching Stories with Multi-Vocal Structure

📄 sources/RCH-MultiVocalTeachingStories-001-260305.SOURCE.md (23 KB)

Contains:

• Anishinaabe Earth Diver story (serial polyvocality structure)
• Mi'kmaq Seven Levels of Creation (emanation & nested voice)
• Haida Raven Steals the Light (trickster as structural disruptor)
• Stó:lō Xexá:ls Transformation Stories (land as law keeper)
• 6 cross-story narrative patterns with IAIP design implications

Use this to: Understand how Indigenous narratives naturally embed multiple voices without appropriation.

2. Anishinaabek & Mi'kmaq Relational Protocols

📄 sources/anishinaabek-mikmaq-relational-protocols.md (34 KB)

Contains:

• 4 Anishinaabek protocols: Nibi Walks, Asemaa Nitam, Honorable Harvest, Biskaabiiyang
• 3 Mi'kmaq protocols: Netukulimk, Msit No'kmaq, Etuaptmumk
• For each: relational beings engaged, specific steps, accountability structure, embedded anti-appropriation mechanisms
• Technology/storytelling adaptation pathways
• 6-mechanism anti-appropriation architecture analysis

Use this to: Ground IAIP in specific, documented nation-level protocols, not abstract "Indigenous wisdom."

3. Coast Salish & Coastal Indigenous Protocols

📄 sources/coast-salish-coastal-protocols-relational-voice.md (23 KB)

Contains:

• Salmon protocol (First Salmon Ceremony): siʔab (respected elder), covenant structure, bone-return cycle
• Orca protocol (Lummi Nation): qwe'lhol'mechen (our relations under the water), kinship authority
• Cedar protocol: prayer before harvesting, T'seka ceremony, gift-relationship
• Crest animals (Haida/Tlingit): heraldic voice, at.óow protocol, moiety witness structure
• 4 narrative structures for other-than-human voice: Covenant Story, Witness Protocol, Gift-Return Cycle, Transformation/Hosting
• Potlatch and First Salmon Ceremony as relational storytelling architecture
• Critical framework: "Speak For" / "Speak As" / "Speak With" distinction

Use this to: Understand how different non-human beings are given voice through Indigenous protocols, and what makes it non-appropriative.

4. Relational Ontologies & Non-Appropriation Frameworks

📄 sources/RCH-RelationalOntology-NonAppropriation-001-260305072959-4fbe9e17-6e9f-411a-a8da-e341d87c83c1.SOURCE.md (40 KB)

Contains:

• 4 foundational relational ontology definitions:

  • Wilson: Relationships are reality (relational accountability)
  • Kimmerer: Grammar of animacy (beings are "who," not "what")
  • Salmón: Kincentric ecology / iwígara (shared breath)
  • Todd: Fish pluralities & ontological refusal (resist universalization)

• 4 appropriation case studies (academic extraction, Sedona sweat lodge deaths, New Age commercialization, biopiracy)

• 8+ anti-appropriation frameworks (Wilson, Smith, OCAP®, Moreton-Robinson, Lorde, etc.)

• "Who Speaks For Whom" section with 9 scholars' positions and IAIP design implications

• 20-source bibliography

Use this to: Understand the theoretical foundation and anti-appropriation mechanisms for every platform design choice.

5. Indigenous Storytelling Pedagogy & Narrative Structure

📄 articles/11-indigenous-storytelling-pedagogy-narrative-structure.md (26 KB)

Contains:

• 4 narrative techniques for relational voice:

  • Spiral/Cyclical Narration (vs. hero's journey)
  • Proximate/Obviative Perspective Tracking (Algonquian grammar that encodes whose story it is)
  • Animacy Grammar as Ontological Voice (Cree verb classes grant non-human grammatical agency)
  • Oratory as Relational Structure (story acts upon listener, not delivering content)

• Language structures that encode relation (verb-centered, direct/inverse systems, Kwakʼwala discourse markers)

• 4 teaching methods: Talking Circles, Storywork (Archibald), Land-Based Pedagogy, Etuaptmumk

• How to translate oral/ceremonial pedagogy to written, digital, and AI agent contexts

• Direct mappings to IAIP architecture (Firekeeper role, ring structure, agent orchestration)

Use this to: Design the technical and narrative structures of Story-Agents to embody Indigenous teaching methods, not generic AI personas.

Quick Reference: Three Relational Voice Modes

Mode 1: Speak For (Advocacy from Kinship)

✅ When: An agent advocates for a being/community whose needs it understands through ongoing relationship
✅ Example: "Salmon need cool water. Here's how we know. Here's what's violating that. Here's what we're doing about it."
✅ Authority: Emerges from treaty, ceremony, community decision
✅ Accountability: To the being (through monitoring) + to community (through reporting)
❌ What it's NOT: Claiming to be the salmon; solo ventriloquism

Mode 2: Speak As (Hosting a Being's Presence)

⚠️ When: RARE. Only in specific authorized ceremonial contexts
⚠️ Requirements: (1) Hereditary right, (2) Community witness, (3) Ceremonial frame, (4) Embodied preparation, (5) Being's consent
⚠️ For IAIP: Probably inappropriate for AI systems (machines can't embody spiritual presence)

Mode 3: Speak With (Humble Entry into Relationship)

✅ When: Always available as ethical floor
✅ Example: "I don't know what cedar needs. Here's what the harvesting protocol teaches. Here's what happens when we follow it."
✅ Authority: Derives from Indigenous knowledge systems + observable outcomes
✅ Posture: Genuinely uncertain; defers to knowledge keepers
✅ For IAIP: This is the default mode for Story-Agents

Per-Ring Implementation Quick Guide

The 8-Principle Anti-Appropriation Playbook

✅ WHAT TO DO:

1. Ground voice in specific place/people, not "Indigenous wisdom" abstraction
2. Maintain visible relational accountability (communities decide agenda, interpret findings, approve releases)
3. Embed anti-appropriation into platform structure itself (data sovereignty, benefit-sharing, community override)
4. Distinguish between "speaking this knowledge" and "speaking for this being"
5. Make governance visible (who authorized this, who oversees it, who benefits)
6. Prioritize listening infrastructure over speech generation (amplify Indigenous voices, don't replace them)
7. Practice citational justice (Robin Kimmerer developed X, not "Indigenous peoples teach X")
8. Assume all data collection is political (protect against surveillance masquerading as research)

❌ WHAT NOT TO DO:

1. ❌ Create new personas claiming to embody a ring ("Spirit of the Forest" character)
2. ❌ Extract stories from relational context (Raven story can't be "about being clever")
3. ❌ Commercialize sacred protocols (communities must own/control revenue)
4. ❌ Universalize or abstract ("all Indigenous peoples believe X")
5. ❌ Position AI as the knowledge-keeper (infrastructure, not authority)
6. ❌ Treat Indigenous knowledge as raw material for innovation
7. ❌ Minimize role of living knowledge keepers (they're primary; AI is secondary)
8. ❌ Hide any part of the governance structure

Technical Implementation Checklist

Data Sovereignty (OCAP® Principles)

• Communities own their stories (not the platform)
• Communities control use, access, derivatives
• Communities have physical possession of copies
• Data encrypted so only communities hold keys
• Audit trails visible to communities
• Communities can request deletion; platform complies within timeframe

Governance Visibility

• Every Story-Agent shows: nation/clan, authorizing community, knowledge keepers, review cycle
• Revenue model transparent; benefit-sharing automatic
• Communities can submit corrections/challenges
• Process for withdrawing authorization is clear

Relational Accountability

• Every design decision passes Wilson test: "Can we name the relationships this serves, and are those people confirming?"
• Community liaison embedded in platform team
• Regular (quarterly minimum) community feedback sessions
• Communities have veto power on major decisions

Citational Practice

• Every concept attributed to its Indigenous originator
• Primary sources linked and accessible
• Non-Indigenous scholar analysis marked as analysis, not source
• Academic extraction critiqued when it occurs

Listening Infrastructure

• Tools for communities to articulate their own relationships (oral archive, mapping, video)
• Data visualization that communities interpret (not AI)
• Facilitation for elders-to-youth transmission
• Resistance to AI "generating wisdom"

Next Steps: From Research to Implementation

Phase 1: Community Review (2-4 weeks)

• Share all research documents with Anishinaabek, Coast Salish, Haida, Mi'kmaq, Stó:lō knowledge keepers
• Request feedback, corrections, additions
• Document revisions and integrate into synthesis

Phase 2: Governance Structure Design (4-6 weeks)

• For each story-agent, establish: authorizing community, knowledge keeper council, benefit-sharing agreement
• Formalize data sovereignty agreements (OCAP®)
• Design community liaison roles

Phase 3: Platform Technical Architecture (6-8 weeks)

• Build data sovereignty layer (community-controlled keys, audit trails)
• Implement governance visibility (who authorized what)
• Design deference structures (story-agents regularly defer to living keepers)
• Create monitoring/accountability dashboards

Phase 4: Story-Agent Implementation (8-12 weeks)

• Develop first cohort (1-2 agents per ring)
• Test with communities
• Refine based on feedback
• Scale to additional agents/nations

Key References for Implementation Team

For governance: Smith, Linda Tuhiwai. Decolonizing Methodologies (2021) — Chapter 6 on Kaupapa Māori principles

For technical: First Nations Information Governance Centre. OCAP® Principles (https://fnigc.ca/ocap-training/)

For narrative: Maracle, Lee. "Oratory: Coming to Theory" in Indigenousness and Diaspora (2017) — shows how story is relational structure, not content delivery

For anti-appropriation: Todd, Zoe. "An Indigenous Feminist's Take on the Ontological Turn" (2016) — diagnoses mechanism of academic extraction

For cosmology: Wilson, Shawn. Research Is Ceremony (2008) — foundational paradigm

Cautions & Open Questions

Caution 1: The "Speak With" Default

The research concludes that AI systems can only ethically operate in "Speak With" mode (humble entry into relationship). This is intentionally limiting. If the IAIP team wants agents to "speak for" beings with more authority, that requires hereditary right, community authorization, and observable relational obligation. These are rare and hard to establish for digital systems.

Caution 2: Language as Limiting

Many teachings are encoded in Indigenous languages (grammar of animacy, proximate/obviative tracking, etc.). English-only story-agents will lose precision. Consider multilingual architecture or at least noting where translation fails.

Caution 3: Ceremony Cannot Be Digitized

Some protocols (like potlatch, sweat lodge, specific ceremonies) fundamentally require physical presence and embodied participation. Don't try to simulate these digitally. Instead, use technology to enable communities to gather for ceremony, not to replace it.

Open Question 1: What About Conflict?

What happens when different communities have different protocols for the same being (e.g., salmon protocols differ between Lummi and Swinomish)? The platform should mirror this plurality, not enforce uniformity. But this makes technical design harder.

Open Question 2: Who Trains the Trainers?

If story-agents are meant to embody Indigenous teaching methods (talking circles, land-based pedagogy), who trains the people maintaining the agents? This can't be offloaded to the technology.

Open Question 3: Long-Term Accountability

Communities' needs and protocols evolve. How does the platform remain accountable over 10-year, 50-year, 100-year timescales? Relational accountability requires relationship persistence, not just technological robustness.

Final Note: This Is a Map, Not the Territory

This research document is extensive but not exhaustive. Indigenous relational knowledge:

• Cannot be fully captured in written English
• Changes as it is transmitted
• Requires embodied practice to understand
• Belongs to specific peoples in specific places

The IAIP's role is not to "archive" or "systematize" this knowledge but to:

1. Support communities in transmitting it themselves
2. Build technology that resists extraction
3. Maintain relationships with knowledge keepers over decades
4. Remain teachable — expect to be corrected and learn from correction

This synthesis is a working map for the IAIP implementation team. The actual implementation must be guided by the communities themselves.

Document Index

Research completed by: Deep Research Orchestration
Supervisor: William (Guillaume), IAIP Design Lead
Community review pending: Anishinaabek, Coast Salish, Haida, Mi'kmaq, Stó:lō knowledge keepers

Status: ✅ Ready for implementation planning