External research runs

Bring-back date: 2026-05-02 to 2026-05-03 (per-file mtime preserved in filename). Source folder (off-repo): /Users/mikewest/Desktop/Deep Research Emotions. Prompts: docs/research/emotion-corpus-expansion/external-deep-research-prompts.md. Bring-back protocol followed: filename <prompt-id>--<llm>--<YYYYMMDD>.md per the prompts file's Bring-back protocol section.

This is the durable archive. Do not edit these files. They are the verbatim LLM responses; the bring-back protocol synthesizes ACROSS them downstream.

What's here

36 runs total = 30 per-emotion (Prompt 1, OpenAI) + 6 cross-cutting (Prompts 2–5, OpenAI + Claude).

Prompt 1 — per-emotion literature reviews (30 runs)

All OpenAI Deep Research, run 2026-05-02 / 2026-05-03.

Emotion	File	Already in EMOTION_PRIMARY_OPTIONS?
anger	p1-anger--openai-deep-research--20260502.md	✅
anxiety	p1-anxiety--openai-deep-research--20260502.md	✅
awe	p1-awe--openai-deep-research--20260502.md	❌ candidate addition
bewilderment	p1-bewilderment--openai-deep-research--20260503.md	❌ candidate addition
boredom	p1-boredom--openai-deep-research--20260503.md	❌ candidate addition
contempt	p1-contempt--openai-deep-research--20260502.md	❌ candidate addition
contentment	p1-contentment--openai-deep-research--20260503.md	❌ candidate addition
desire	p1-desire--openai-deep-research--20260502.md	✅
despair	p1-despair--openai-deep-research--20260503.md	❌ candidate addition
disappointment	p1-disappointment--openai-deep-research--20260502.md	✅
disgust	p1-disgust--openai-deep-research--20260502.md	✅
embarrassment	p1-embarrassment--openai-deep-research--20260503.md	✅ (shame family)
fear	p1-fear--openai-deep-research--20260502.md	✅
gratitude	p1-gratitude--openai-deep-research--20260503.md	❌ candidate addition
grief	p1-grief--openai-deep-research--20260502.md	✅
guilt	p1-guilt--openai-deep-research--20260503.md	✅ (shame family)
hope	p1-hope--openai-deep-research--20260503.md	❌ candidate addition
jealousy	p1-jealousy--openai-deep-research--20260503.md	❌ candidate addition
joy	p1-joy--openai-deep-research--20260502.md	✅
loneliness	p1-loneliness--openai-deep-research--20260502.md	❌ candidate addition
longing	p1-longing--openai-deep-research--20260503.md	⚠ overlaps with yearning — research-driven decision needed
love	p1-love--openai-deep-research--20260503.md	✅
nostalgia	p1-nostalgia--openai-deep-research--20260503.md	✅
pride	p1-pride--openai-deep-research--20260502.md	✅ (shame family)
relief	p1-relief--openai-deep-research--20260503.md	✅
resentment	p1-resentment--openai-deep-research--20260503.md	❌ candidate addition
sadness	p1-sadness--openai-deep-research--20260502.md	✅
shame	p1-shame--openai-deep-research--20260502.md	✅ (shame family)
tenderness	p1-tenderness--openai-deep-research--20260502.md	✅
trust	p1-trust--openai-deep-research--20260503.md	❌ candidate addition

Candidate additions to EMOTION_PRIMARY_OPTIONS (Phase 4 work): awe, bewilderment, boredom, contempt, contentment, despair, gratitude, hope, jealousy, loneliness, resentment, trust. Plus a research-driven decision on longing↔yearning.

Prompt 2 — Emotion Corpus Finder / Source Material Discovery (cross-cutting, 2 runs)

Memoir traditions · oral-history archives · therapy transcripts · open-access journals · diaries / letters · ethnographic catalogs · documentary film · honest gaps.

• p2-source-discovery--openai-deep-research--20260503.md
• p2-source-discovery--claude-deep-research--20260503.md

Convergence reading needed: these are designed for cross-LLM synthesis — Phase 2's first pass.

Prompt 3 — Historiography of named emotions (cross-cutting, 2 runs)

How Western emotion-naming has changed over time · emotions that arrived late · emotions that disappeared · therapeutic culture's reshaping · cross-tradition contrast · contemporary live debate.

• p3-historiography--openai-deep-research--20260503.md
• p3-historiography--claude-deep-research--20260503.md

Editorially load-bearing: the historiography directly informs Vela's editorial register and the developmental-theology arc.

Prompt 4 — Cross-cultural emotion ethnography (cross-cutting, 1 run)

Foundational ethnographies · linguistic-anthropology · indigenous + decolonial framings · religious + contemplative-tradition emotion mapping · affective-anthropology methodology · honest gaps.

• p4-cross-cultural--claude-deep-research--20260503.md

OpenAI counterpart not yet run. Future bring-back can fill the convergence pair.

Prompt 5 — Therapy-transcript + clinical case-study source discovery (cross-cutting, 1 run)

Public-domain primary sources · trade-published case-study collections · master-clinician demonstration video series · open-access journals with verbatim transcripts · family/group-therapy transcripts · recovery + self-help confessional · historical psychoanalytic case literature · selection-bias note · ethical considerations.

• p5-therapy-transcripts--claude-deep-research--20260503.md

OpenAI counterpart not yet run. Future bring-back can fill the convergence pair.

What downstream surfaces these feed

Per the prompts file's Bring-back protocol §3–5:

1. Convergent sources (where multiple LLMs agree on canonical works) → highest-priority candidates for ASN-1023 (docs/research/corpus-acquisition/EMOTION-CORPUS.md) and ASN-1024 (docs/research/corpus-acquisition/THERAPY-TRANSCRIPT-CORPUS.md) acquisition trackers.
2. Divergent points (where LLMs disagree on emphasis or contested readings) → editorial use for the magazine; signal contested terrain Vela's editorial work can engage with directly.
3. New source categories surfaced → potential new acquisition trackers (e.g., a sibling CROSS-CULTURAL-EMOTION-ETHNOGRAPHY.md for Prompt 4 material).
4. Taxonomy-relevant findings → feed into lib/loom/taxonomy/emotion.ts (EMOTION_PRIMARY_OPTIONS, SHAME_FAMILY_PRIMARIES, complexity markers) — Phase 4 work.
5. Research-program scaffolding → seed a new top-level research thread (sibling to Christianity-sex-shame and Text-aesthetic) — Phase 3 work, plan filed at docs/research/emotion-corpus-expansion/2026-05-03-deep-research-arrival-plan.md.

What is NOT in this index

• Synthesis content. The runs above are raw LLM output. Cross-LLM synthesis happens in the per-tracker docs (Phase 2) and in 2026-05-03-deep-research-arrival-plan.md.
• Acquisition decisions. What to actually purchase / acquire / scan lives in the canonical trackers under docs/research/corpus-acquisition/. This INDEX only points to the source material.
• Open gaps. Each run includes its own Where you are uncertain / Honest gaps section. Future bring-backs (additional LLMs on Prompts 4 and 5; per-emotion runs from Claude or Gemini) can close those gaps.