MetaScience, Scientific Inquiry, and Agents
Scientific advance through plural inquiry
A thirteen-meeting reading club studying scientific advance through plural epistemic modes, and designing agents evaluated against documented scientific work.
About this Group Meeting
Drafted v0.5 · public book narrowed to what the group has covered · October 20, 2026
MetaScience, Scientific Inquiry, and Agents · Denolle Lab · meetings taken one at a time
What this book is for
- Define the modes of inquiry as the literature invokes and uses them: observation, exploration, hypothesis assessment, theory, simulation, instruments, methods, synthesis, robustness, estimation, regularities.
- Define the outcomes we want to maximize: scientific advance, impact, novelty, and breadth or polymathy, with how each has been measured and where the measures fail.
- Build a general graph of inquiry progressively, one reading at a time, and test it on successful and ordinary papers: where each paper enters, which steps it takes, where it ends, and how it scores.
- Each mode will become a specialist agent. The definitions and tests here are what those agents’ rubrics start from.
Read in this order
| Step | Page | What it gives |
|---|---|---|
| 1 | Definitions | The outcome terms: advance, novelty, impact, uptake, breadth; the thirteen mode profiles in one table |
| 2 | Modes of inquiry | Where each mode was defined, the evidence on modes and outcomes, what each mode writes into Cleland’s schema |
| 3 | The unit graph | Graph v0.4: the nodes, their sources, the change log |
| 4 | Walking the graph | Eight papers, step by step through the graph, with a success profile for each |
| 5 | From graph to decision tree | Graph v0.5, proposed: Cleland’s schema as the record, the modes as units with inner loops, eight questions that call them |
| 6 | What has been measured | Novelty, impact, breadth, and advance metrics from the literature |
| 7 | Success set | The eight test papers, matched controls, and the metrics against them |
- The thirteen mode specifications are the rubric each mode agent will be held to.
Where things stand
| Material | Status |
|---|---|
| Mode definitions | Drafted from the literature, October 19, 2026; to be discussed at the modes meeting |
| Graph | v0.4.2 is the vocabulary; the v0.5.1 decision tree is the controller (adopted by the convenor, October 21, 2026; room discussion pending) |
| Test papers | Eight coded from the full text; 24 matched controls scored; no unsuccessful papers yet |
| Success score | v0 defined on the walks page; two of four dimensions measured, two pending the group’s ratings |
| Meetings held | Meeting 1, September 15, 2026; Meeting 10, September 22, 2026; notes |
| Next meeting | Modes of inquiry, with slides |
- Drafted: proposed text. Discussed: taken up in a recorded meeting. Adopted: an explicitly recorded decision.
- Pages for meetings not yet held are kept in the repository and return to the book when the group takes them up.
How we meet
- One anchor reading for everyone, one companion presented by a rotating reader; optional readings are resources.
- Each meeting leaves one ledger row and three sentences: what changed, what evidence warrants it, what follows for the agents or their evaluation.
- Distinguish a primary result, a philosophical argument, a historical reconstruction, an empirical association, and a benchmark.
Access
- Bibliography: every work cited, with open-access links where they exist. No journal PDFs are redistributed.