Dante Research Requests — Design Spec

Date: 2026-06-28
Source: work/TODO.md TODO-16 ("Dante research requests")
For: Dante Diotallevi (Synergetics University; "Synergetics Conversations"), a philosophy-trained Fuller scholar.

Purpose

Run three research passes over the Buckyverse corpus and produce shareable, cited findings answering the questions Dante posed:

• (a) Algebra / "cipher": every statement Fuller made regarding algebra, including every instance of the word "cipher."
• (b) Philosophers: all philosophers mentioned, with their distribution.
• (c) Scientists & theories: the distribution of scientists / scientific theories mentioned.

Deliverables

Three separate output reports in .wiki/output/ (type: report):

1. research-fuller-algebra-and-cipher-2026-06-28.md
2. research-fuller-philosophers-mentioned-2026-06-28.md
3. research-fuller-scientists-and-theories-2026-06-28.md

Each report is self-contained and follows the wiki output frontmatter (title, type: report, sources, generated).

Corpus

Everything in the repo: all book content (books by and about Fuller, in the root book directories), the Synergetics Dictionary (~20K cards, synergetics_dictionary/), and .wiki/ articles. Search the actual source markdown, not just the wiki layer.

Method (per report)

This is interpretive research, not a keyword concordance. Grep is the entry point to find passages; the work is reading those passages in context to surface Fuller's actual answers and positions — and it is fine (often better) when other people's words (biographers, interviewers, co-authors, scholars in the about-books) point to or frame Bucky's answers. A mere name-drop is noted but weighted far below a passage where Fuller engages the idea.

A locate → read-in-context → interpret → evidence pass:

1. Locate. Define a generous term/name list (e.g. algebra, algebraic, cipher, ciphering for (a); a broad philosopher list for (b); scientists + theory terms for (c)) and grep across the whole corpus to find candidate passages.
2. Read in context. Open the surrounding text of the strongest hits — including others' framing (an interviewer's question, a biographer's gloss) that reveals Fuller's stance. Distinguish genuine engagement (Fuller takes a position, argues, redefines) from incidental mention (a name in a list).
3. Interpret. Synthesize Fuller's actual answer/position: e.g. what he says algebra/ciphering is and why he distrusts or reframes it; which philosophers he wrestles with vs name-checks and how; which scientists/theories anchor his thinking and to what end.
4. Evidence. Support each claim with cited quotes (path + passage), drawing on both Fuller's own books and the about-books where their framing illuminates his answer.

Report structure

---
title, type: report, sources, generated
---
# <Title>
> abstract — Fuller's answer/position in one paragraph
## Method (corpus, how located, date; interpretive not concordance)
## Fuller's position / answer  (the synthesis — the heart of the report)
## Evidence (cited quotes; mark whether from a book BY or ABOUT Fuller)
## Distribution / inventory (table: name/term → weight → where engaged)
## Notes (honesty: OCR noise; whole-repo corpus mixes voices; curated lists; grep-located not exhaustively disambiguated)
## Sources (searched corpus; Synergetics Dictionary dataset)

The synthesis ("Fuller's position / answer") leads; the distribution table supports it rather than being the point.

Honesty / limitations

• Searching the whole repo conflates Fuller's words with biographers'/scholars'. Reports will flag this and, where it matters, separate hits in books by Fuller from books about him.
• OCR'd scans carry noise; counts are approximate. Name lists are curated, so the distribution reflects the chosen list (reports will state the list).
• This is a grep-and-read survey, not a disambiguated NLP entity-extraction; reports say so.

Integration & verification

• Files land in .wiki/output/; update output/_index.md and master stats; keep frontmatter lint-clean and links resolving.
• Verification = the project's content lint (wiki structure/link/frontmatter checks) plus confirming every cited file path resolves. No build/test suite applies (markdown-only deliverable).

Out of scope (YAGNI)

• No new tooling, NLP pipeline, or app. No changes to the corpus itself.
• No automated recurring job — this is a one-time research pass; the recipes are reproducible from each report's Method section.

Pluck lifecycle

Implement on the feature branch → verify → squash-merge to main → queue an ACC acceptance (a human/Dante should sanity-check the findings) → yank TODO-16.