Methods.
The Cosmologies Project runs on a standardized four-week cycle so the discipline does not have to be reinvented each month. Three rules carry the weight: predictions are locked before evidence is read, sources are primary and from within the tradition, and every prediction is scored honestly at the end of the month.
The lock is the most important rule. A framework that predicts after reading is a framework that confirms itself. The lock makes the prediction memo a public artifact that exists before the contact, exposed and dated, with no edits permitted afterwards. Corrections enter the findings memo at month-end as data, not as retroactive edits to what was claimed.
Each month produces five artifacts: the locked prediction memo, a NotebookLM podcast trio (predictions, engagement, findings), an outreach engagement with one named scholar or practitioner, the locked findings memo, and the tradition page on the site updated to its complete state. Time investment runs about thirty hours per month, three hundred and sixty hours across the cohort.
Predict
~6 hours- ·Mon — name the tradition, the cohort question, and the calibration stake
- ·Tue — list framework predictions of convergence (specific to TFE terms, Ostrom principles, canon C-IDs)
- ·Wed — list what might falsify or revise the framework on contact
- ·Thu — list what might surface as new (concepts the framework lacks)
- ·Fri — lock primary sources, lecture targets, podcast targets, outreach target
- ·Weekend — final read-through, lock memo, generate Predictions Podcast
- →Locked prediction memo (no edits permitted after lock)
- →NotebookLM Predictions Podcast (12-18 min) generated and embedded
Read
~8 hours- ·Mon — primary text 1, opening 60-80 pages; note premises and untranslatable vocabulary
- ·Tue — continue text 1; mark contact points with the prediction memo
- ·Wed — text 1 final segments, or transition to text 2
- ·Thu — text 2, typically scholar-translator commentary or within-tradition critical voice
- ·Fri — text 3 where applicable, or revisit ambiguous passages
- ·Weekend — reading-log consolidation and prediction-evidence tagging
- →Daily reading log (engagement notes, not summary)
- →Each prediction tagged with current evidence: leaning-confirm, leaning-refute, leaning-surface, no-data
Watch & Listen
~8 hours- ·Mon — YouTube lecture #1: senior practitioner or scholar within the tradition's institutional setting
- ·Tue — YouTube lecture #2: critical or comparative voice
- ·Wed — podcast #1: long-form interview with a tradition scholar
- ·Thu — podcast #2: practitioner voice or second scholarly perspective
- ·Fri — outreach: prediction memo + calibration stake sent to one scholar/practitioner; ask is correction, not endorsement
- ·Weekend — Engagement Podcast generated; site mid-month status updated
- →Watch-log appended to reading log
- →Outreach email sent to one named scholar or practitioner
- →NotebookLM Engagement Podcast (15-20 min) generated and embedded
Synthesize
~8 hours- ·Mon — score every prediction: confirmed / partially confirmed / refuted / surfaced
- ·Tue — identify revisions the framework absorbs (named TFE terms, named canon concepts)
- ·Wed — identify what the framework refuses to absorb and why
- ·Thu — draft findings memo: scoring, revisions, refusals, openings, scholar replies
- ·Fri — update site tradition page; embed Findings Podcast; mark month complete
- ·Weekend — newsletter dispatch; decide whether findings warrant a graph Phase onboard
- →Locked findings memo with prediction scoring
- →Tradition page on site updated to complete state
- →NotebookLM Findings Podcast (20-25 min) embedded
- →Newsletter dispatch (~600-800 words) sent
- →Graph Phase onboard queued where revisions warrant
Three podcast artifacts per tradition, generated through NotebookLM from the project's own memos. Not marketing. The podcasts make the discipline portable — the prediction memo can be heard before the findings, and the findings can be heard without re-reading the source materials.
Predictions Podcast
12-18 minGenerated from the locked prediction memo. Names the tradition, the calibration stake, the framework's predictions of convergence, the falsification candidates, and the surfaces the framework expects to find missing. Listenable evidence the prediction was made before the reading began.
Engagement Podcast
15-20 minMid-month synthesis. Combines the reading log with the lecture and podcast notes. Honest about which predictions are still unresolved. Not a victory lap. Reports what surprised the engagement so far and what the outreach surfaced.
Findings Podcast
20-25 minThe score. Each prediction marked confirmed, partially confirmed, refuted, or surfaced-something-else. Names what the framework is revising and what it is refusing to absorb. The discipline exposed: what was wrong in the prediction memo, not just what was right.
What counts as a primary source
- ✓Scholar-translators within the tradition (e.g., Bhikkhu Bodhi for the Pali Canon, Mogobe Ramose for Ubuntu, Doi Takeo for Amae).
- ✓Within-tradition critical voices (scholars or practitioners critiquing the institutional substrate from inside it).
- ✓Recorded lectures and long-form interviews where the speaker is positioned within the tradition or has substantial sustained engagement with it.
- ✓Practitioner accounts when paired with at least one scholarly source that contextualizes the practitioner's position in the broader tradition.
What does not count
- ✕Wikipedia and encyclopedia entries treated as primary. They can orient; they do not engage.
- ✕Second-hand Western readings of the tradition where a within-tradition source is available in English.
- ✕Survey textbooks aimed at undergraduate religious-studies courses (not because they are wrong, but because they smooth over the metaphysical differences the project is built to engage).
- ✕AI-generated summaries of the tradition (the project uses NotebookLM to summarize its own work, not to substitute for primary engagement).
The discipline is not purist — it is calibrated to the project's purpose. The framework cannot test itself against substrates whose differences have already been smoothed over by the texts the framework reads. The first job of source selection is to preserve the difference.
I am an outsider. Please correct me.
The Cosmologies Project engages every tradition it studies in good faith, from the outside. It tests the framework against substrates whose metaphysical assumptions differ from its own. That is the work. The discipline is the prediction memo locked before reading begins, the primary sources from within the tradition, and the willingness to publish findings whether they confirm or refute.
If you read these pages and the framework gets something wrong about your tradition, please tell me. If you can point to sources from within the tradition that the project should engage with and has not, please send them. If you think a tradition is missing from the cohort that ought to be in it, please name it — a future cohort can absorb what this one cannot.
Reach me through the project's contact channels. Replies are read. Corrections enter the findings memo for the relevant month and, where warranted, revise the framework in the graph.