# RESEARCH_SYNTHESIS.md

## Activation Record

- Decision question:
- System recommendation: recommended / not-recommended
- Trigger reason codes:
- User decision: pending / confirmed / declined
- Confirmed by role:
- Confirmed at:
- Research mode: material-first hybrid
- External gap-fill: allowed / not-allowed
- Advisory boundary: this synthesis informs decisions; it is not acceptance, release, or effectiveness evidence.

### Trigger Reason Codes

| Code | Material signal | Present | Evidence / note |
|---|---|---|---|
| `explicit-multi-view` | User explicitly requests comparison, critique, or multiple perspectives | yes / no |  |
| `evidence-conflict` | Sources disagree on a decision-relevant claim | yes / no |  |
| `high-impact-decision` | The decision is costly, hard to reverse, or high consequence | yes / no |  |
| `credible-route-divergence` | Two or more credible routes depend on different assumptions | yes / no |  |
| `cross-domain-gap` | The decision has a material gap across disciplines or time periods | yes / no |  |

## Executive Layer

### One-sentence Synthesis

-

### Confidence-ranked Findings

| Finding ID | Finding | Confidence | Evidence strength | Source / claim IDs | Decision relevance |
|---|---|---|---|---|---|
| `FND-001` |  | high / medium / low | strong / moderate / weak / unverified |  |  |

### Recommended Actions

| Action | Supported by | Reversibility | Verification | Stop / re-evaluate trigger |
|---|---|---|---|---|
|  |  | reversible / costly / hard-to-reverse |  |  |

### Frontier Questions

| Question | Why it matters | Evidence needed | Owner / next step |
|---|---|---|---|
|  |  |  |  |

## Scope And Sources

- In scope:
- Out of scope:
- Material supplied by the user:
- Important gaps eligible for external fill:

| Source ID | Source class | Privacy-safe pointer | Coverage | Used by claims | Limitations |
|---|---|---|---|---|---|
| `SRC-001` | user-material / public-external / approved-private-external / synthetic |  |  |  |  |

### Source Rules

- Use supplied material first. External research may fill only a recorded important gap and must not silently expand scope.
- Prefer primary, official, or original research sources. Record access date for time-sensitive external material.
- Keep private content out of this file. Use the privacy-safe source attestation and pointer rules from `PROJECT_BRIEF.md`.
- Do not invent citations, evidence, quotations, missing values, or consensus.

## Claim And Evidence Ledger

| Claim ID | Claim | Claim type | Source IDs | Evidence strength | Confidence | What would change the conclusion |
|---|---|---|---|---|---|---|
| `CLM-001` |  | verified fact / inference / professional judgment / unknown |  | strong / moderate / weak / unverified | high / medium / low |  |

### Calibration Rules

- Evidence strength describes the support, not how persuasive the wording sounds.
- `strong`: direct project evidence or convergent high-quality primary sources with no material unresolved conflict.
- `moderate`: relevant support with a bounded limitation, dependency, or conflict.
- `weak`: indirect, single-source, analogy-heavy, or materially incomplete support.
- `unverified`: no reviewable support; keep it as a gap, not a finding.
- Confidence is claim-level. Do not use unsupported percentages.
- Every substantive finding must link to at least one claim ID. Inferences and professional judgments must be labeled.

## Five-lens Scan

Each lens is an analytical frame, not a claimed persona, credential, or separate agent. Complete all five rows unless a row is explicitly marked `not-applicable` with a reason.

| Lens | Core position | Strongest supporting evidence | Strongest counterevidence / objection | Hidden signal | Claim IDs |
|---|---|---|---|---|---|
| Practitioner |  |  |  |  |  |
| Scholar |  |  |  |  |  |
| Skeptic |  |  |  |  |  |
| Economist |  |  |  |  |  |
| Historian |  |  |  |  |  |

### Lens Boundaries

- Practitioner: execution constraints, operational feedback, failure modes, and what survives contact with reality.
- Scholar: definitions, theory, research quality, causality, and competing explanations.
- Skeptic: strongest falsification attempt, missing evidence, incentives to overclaim, and failure conditions.
- Economist: incentives, opportunity cost, distribution effects, market structure, and second-order effects.
- Historian: path dependence, comparable precedents, discontinuities, and false-analogy risk.

## Contradiction Map

| Conflict ID | Claim A | Claim B | Evidence for A | Evidence for B | Affected lenses | Status | Evidence needed |
|---|---|---|---|---|---|---|---|
| `CON-001` |  |  |  |  |  | unresolved / partially-resolved / resolved |  |

### Consensus

| Consensus ID | Shared conclusion | Participating lenses | Evidence boundary | Confidence |
|---|---|---|---|---|
| `CNS-001` |  |  |  |  |

### Unresolved Gaps

| Gap ID | Missing or disputed information | Affected findings | Can external research fill it? | Next evidence |
|---|---|---|---|---|
| `GAP-001` |  |  | yes / no |  |

## Cross-lens Connections

| Connection ID | Lenses connected | Hidden relationship | Why one lens alone would miss it | Decision impact |
|---|---|---|---|---|
| `LNK-001` |  |  |  |  |

## Self-review

| Finding ID | Weakest link | Strongest alternative explanation | Unresolved contradiction | Confidence after review |
|---|---|---|---|---|
| `FND-001` |  |  |  | high / medium / low |

### Academic-rigor Rubric

This transparent rubric approximates a demanding academic review. It does not claim Stanford University affiliation, endorsement, or an actual professor review.

| Dimension | Score 0-4 | Evidence for score | Required improvement |
|---|---:|---|---|
| Question and scope clarity |  |  |  |
| Source traceability and quality |  |  |  |
| Claim calibration |  |  |  |
| Contradiction and counterevidence handling |  |  |  |
| Decision usefulness and falsifiability |  |  |  |
| **Total / 20** |  |  |  |

Score meaning: `0` absent, `1` seriously deficient, `2` usable with material gaps, `3` strong with bounded gaps, `4` review-ready with explicit limitations.

## Completion Checklist

- [ ] The user-confirmed activation and trigger reasons are recorded.
- [ ] All five lenses contain a core position, strongest evidence, counterevidence, and hidden signal.
- [ ] Every substantive finding links to claim and source IDs or is marked unverified.
- [ ] Evidence strength and confidence are calibrated separately.
- [ ] Contradictions remain visible until evidence resolves them.
- [ ] External research, if used, fills only recorded important gaps.
- [ ] Recommendations include verification and stop / re-evaluate triggers.
- [ ] Self-review identifies weak links and alternative explanations.
- [ ] Canonical project decisions are updated only after user confirmation.