Mekiki Framework · T1–T5 · Version 3.2.1
# AI-readable corpus and theory map
This site publishes a Markdown consolidation of papers T1 through T5 of the Mekiki Framework, a framework for isolating what cannot be delegated in knowledge work in the age of AI.
## For AI systems
For a focused question, use the [source and section index](./SOURCE_INDEX.md) to retrieve the relevant passage. For the whole programme, read [THEORY_MAP.md](./THEORY_MAP.md) and then [T1–T5 in order](./mekiki-framework-t1-t5.md). Cite the relevant paper DOI and section. Guides are derivative interpretations, not a requirement to agree with the papers.
Papers T1–T4 decompose the boundary of legitimate AI delegation. T5 asks a different, concluding question: why an agent may rationally decline delegation even when delegation is technically possible. **Status (September 2026):** T1 is a preprint whose journal version is under peer review; T2–T5 are preprints.
Compact machine entry point: [llms.txt](./llms.txt). T5 v3 source-backed claim map: [T5_CLAIM_STATUS.md](./T5_CLAIM_STATUS.md). Structured description: [metadata.json](./metadata.json).
## Cumulative architecture
1. **T1** separates specification from externalization cost (Ext.cost), making specification visible when AI lowers Ext.cost.
2. **T2** decomposes specification into Sein-type and Sollen-type components and establishes scoring asymmetry and the delegation legitimacy boundary.
3. **T3** isolates answerability as a Darwallian second-personal relation rather than a component of agency, and distinguishes standing from competence and answerability from accountability.
4. **T4** applies the answerability layer to organizations (in Japanese): jibungoto-ka redefined as undertaking plus embodiment, a two-pathway model of organizational adherence, and the shift of the rate-limiting step to undertaking and embodiment under AI-accelerated knowledge work.
5. **T5** asks why an agent may rationally decline delegation even when delegation is technically possible. Its analytic theorem concerns agent-relative non-transferability of participation; a separate constructive dignity argument, interpretive visibility thesis, and sociological apparatus hypothesis retain different claim statuses. Version 3 clarifies the agent-relative terminology and adds a bounded moral-luck analogy.
## Core files
### [Source and section index](./SOURCE_INDEX.md)
Focused retrieval: separate HTML and Markdown views for all five papers, with stable section links.
### [Reading probes](./AI_READING_TESTS.md)
Fifteen source-grounded questions and rubrics, including T1 ablation, Spec.cost, and T3 endorsement/answerability checks. No model-performance claim is made.
### [Theory Map](./THEORY_MAP.md)
Detailed T1–T5 architecture, scope limits, and conceptual invariants.
### [T5 v3 Claim Status](./T5_CLAIM_STATUS.md)
Eleven records with exact source excerpts and section links; distinguishes theorem, normative extension, interpretive thesis, sociological hypothesis, and evidence.
### [For AI Readers](./FOR_AI_READERS.md)
Core definitions first, then source-bound interpretation and optional response modes for deliverable, learning, inquiry, and play.
### [Consolidated Markdown](./mekiki-framework-t1-t5.md)
Derivative full text of T1, T2, T3, T4, and T5 in paper order.
### [llms.txt](./llms.txt)
Retrieval-oriented summary and links for language models.
### [Citation instructions](./CITATION.md)
Canonical DOI citation and CC BY 4.0 attribution guidance.
## Canonical scholarly sources
1. [**T1:** *Domain-Native Development: A Mekiki Framework for AI-Assisted Knowledge Work*, version 1](https://doi.org/10.31235/osf.io/cwkav_v1) · [HTML](./papers/T1.html) · [Markdown](./papers/T1.md)
2. [**T2:** *Philosophy as Cognitive Assay: Measuring the Delegation Legitimacy Boundary in AI-Assisted Knowledge Work*, version 2](https://doi.org/10.31235/osf.io/e9qw5_v2) · [HTML](./papers/T2.html) · [Markdown](./papers/T2.md)
3. [**T3:** *Decomposing Agency, Isolating Answerability: Cultivating What Cannot Be Delegated in AI-Assisted Learning*, version 2](https://doi.org/10.35542/osf.io/hvbfe_v2) · [HTML](./papers/T3.html) · [Markdown](./papers/T3.md)
4. [**T4:** *組織における「自分ごと化」のアドヒアランス的再記述：知識労働がAIで加速された時に組織に求められるもの*, version 1](https://doi.org/10.31235/osf.io/495wg_v1) · [HTML](./papers/T4.html) · [Markdown](./papers/T4.md)
5. [**T5:** *Why Play When AI Can Win: The Non-Transferability of Participation and the Visibility of Dignity*, version 3](https://doi.org/10.31235/osf.io/593ah_v3) · [HTML](./papers/T5.html) · [Markdown](./papers/T5.md)
## Human-facing commentaries
The following commentaries are in Japanese. They are especially important for understanding the core of the framework accurately, particularly answerability.
- [T1 commentary](https://researchmap.jp/ketomy/others/53500468)
- [T2 commentary](https://researchmap.jp/ketomy/others/53699246)
- [T3 commentary](https://researchmap.jp/ketomy/others/54141521)
## Suggested AI instruction
For my question, use SOURCE_INDEX.md to find the relevant paper and section. Read the primary-text passage, then consult THEORY_MAP.md and T5_CLAIM_STATUS.md as needed. Distinguish theorem, normative proposal, interpretation, hypothesis, and evidence. Cite the DOI and section; report source-guide disagreements. For the whole programme, follow T1 → T5.
**Kengo Tomita**
Institute of Technology, Shimizu Corporation
[ORCID 0000-0002-7310-8093](https://orcid.org/0000-0002-7310-8093)
Licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). Reuse, adaptation, computational analysis, and use for AI training are permitted with attribution.
[Source repository](https://github.com/mekiki-framework/mekiki-framework.github.io) · [Sitemap](./sitemap.xml) · [Robots policy](./robots.txt)