Transparency · scoring instrument
The scoring rubric — OTR-1.0
This is the actual instrument Debby intends to use to compare worldviews — the dimensions, the scale, the weights, the tie rule, and the minimum evidence a school needs before it gets scored at all. It is published here in the same form the research uses it, not a marketing summary, so you can find the exact clause you disagree with.
No score has been calculated under this rubric. The instrument is frozen; the sheet is not filled in. A frozen rubric with unmarked cells still shouldn’t be trusted to produce numbers — including by Debby itself.
Which subjects are still unmarked — and which open questions would help mark them — is a public worklist, not a ranking.
Why this is published at all
Debby’s founder holds an Eastern Orthodox conviction. That conviction is treated here as a research hypothesis, not a pre-adjudicated verdict — the same proof obligations and the same frozen rubric apply to Orthodoxy that apply to every rival school, including on the dimensions where the founder’s own side is expected to do well.
Three commitments are load-bearing enough to say plainly, because the rubric below only means what it says if they hold:
- Same rubric for every worldview. A criterion is not used against a rival unless it has also been run against Eastern Orthodoxy — see the symmetry rule under Reviewer policy below.
- Steelman before critique. Every dimension requires a documented search for the school’s own best, strongest-available version of its position before that position is scored — not the easiest version to argue against.
- Quote, inference, and judgment stay separate. A verbatim quotation, a paraphrase or logical step drawn from it, and this program’s own evaluative conclusion are three different things, and the rubric’s own score sheet (§8, linked below) is built to keep them from blurring into each other.
Publishing the instrument before any score exists is the point: a rubric written after the results are known can be quietly bent toward them, however honestly. This one is dated, versioned, and open to scrutiny before it has scored anything.
Live findings, in plain language
Open research notes from the same adversarial process the rubric will eventually score under — reported here because the method is supposed to be able to find real gaps, not only agreeable results. None of these notes is a score.
Research note · 2026-07-31 · logic’s necessity
A specialist philosopher of logic (Penelope Maddy, UC Irvine — past president of the Association for Symbolic Logic) has argued in print that even rudimentary logic is not metaphysically necessary: its correctness depends on contingent structural features of our actual world, and quantum mechanics supplies a documented case where that structure fails to hold. On her account, logic’s apparent necessity is an illusion generated by mistaking broad applicability for unrestricted necessity.
Run symmetrically, as this program requires, that finding cuts in both directions at once. It weakens a leading atheist philosopher’s own account of what grounds logic — and it weakens, at least as much, the theist move that stops a regress of grounds at God specifically because God is necessary while the laws of logic supposedly are not. Neither side gets to treat logical necessity as a free premise; each owes it an argument the corpus has not yet located.
This is a research finding, not a rubric score — no `OTR-1.0` dimension has been evaluated for either party on this question, and the underlying contention remains open. It is reported here because the rubric’s own §7 rule exists precisely to stop open findings like this one from being smuggled into a number before they are settled.
Research note · 2026-07-31 · paradigm-case exclusion
A specialist philosopher (Gila Sher, UC San Diego) has argued in print that if a civilization enshrined Nazi values, those would not count as genuine human values — they would be “inhuman.” This program’s adversarial read of her paper found a real, unrepaired gap in how that exclusion is secured: she names Nazi values as her paradigm of the “inhuman” bucket before introducing the technical machinery (rigid designation) later cited to explain why the exclusion sticks. That machinery, applied consistently, shows only that currently-enshrined values resist relativization; it does not, by itself, supply an independent reason Nazism specifically belongs outside the human-value category rather than counting as a different fixed referent. A follow-up pass found the same pattern is structural, not one-off: every named exclusion case in the paper fails the same asserted-without-applied-standard test.
Three limits travel with the finding. It does not revive relativism about the values our civilization actually enshrined — that part of Sher’s apparatus still does real work. It is not unique to her: most ethical theories that fix concepts by paradigm cases face the same structural question. And it is a finding about the architecture of the argument, not a verdict that her conclusion is false. Finding the gap is the point of the method; dressing it as a refutation would be the opposite of the method.
This is a research finding, not a rubric score — no `OTR-1.0` dimension has been evaluated for Sher or any rival on this question, and the underlying contentions remain open. The next note is the same check, run the same day against the founder’s own side.
Research note · 2026-07-31 · EO holism under the same check
The same adversarial standard was then run against Eastern Orthodox epistemic holism — specifically the worked form in Jay Dyer’s transcendental apologetic and Fr. Dcn. Ananias Sorem’s “An Orthodox Theory of Knowledge…” That argument is real, not a slogan: it diagnoses epistemic bootstrapping in autonomous starting points and concludes, by impossibility-of-the-contrary, to the Orthodox revealed God as the unique precondition of knowledge, with systems standing or falling as a whole. Credit where due: unlike Sher’s unused “goodness” criterion, this critical standard actually does work against foundationalism and thin coherentism.
It still fails the identical free-rider check in two places. Success against autonomous epistemology is treated as licensing a full Triune / Eastern Orthodox package, without an independent derivation that rules out unitarian or other theistic rivals. And morality is listed alongside logic as a co-equal transcendental without a separate demonstration that categorical moral obligation has the same precondition-of-intelligibility status logic is argued to have. Parallel gap, same cycle, founder’s own side — which is the point of the parity rule.
This is a research finding, not a rubric score — no `OTR-1.0` dimension has been evaluated for Eastern Orthodoxy on this question, and the underlying contentions remain open. It is reported beside the Sher note because a method that finds gaps only in rivals is not the method this rubric claims to be.
The full story, illustrated
These three notes are excerpts. The complete thread — Maddy’s necessity denial, Agrippa’s trilemma catching everyone tested, and Orthodoxy’s own genuine wins and unrepaired gaps — is written up in full, with diagrams, as its own case study.
The ten dimensions
Every school is scored, dimension by dimension, against its own best located formulation of its position — never against a weaker version, and never against a version this program built on the school’s behalf.
Internal coherence and valid inference
Are the school's load-bearing commitments mutually consistent, and are the inferences it needs actually valid?
The only veto dimension — a 0 here blocks any aggregate (see Weights).
Explanatory scope and depth
How much of the fixed explanandum list does the account address, by a stated mechanism rather than by relabelling? The preconditions of logic and mathematics are a fixed comparison item here.
Ontological and epistemic adequacy
Does the school supply both something in reality that makes the claim true, and an account of how a knower could come to know it — consistently with each other?
Non-ad-hocness and parsimony
Are the school's posits motivated independently of the results they were introduced to secure? A lower posit-count never adds points by itself.
Fit with public empirical, historical, and textual evidence
Where a claim is of a type public evidence can bear on, how does it fare against that evidence, judged by the standards of the relevant discipline?
Epistemic access and error-correction resources
Does the school supply a mechanism by which a sincere adherent could detect and correct an error in its own account — and has it ever visibly operated?
Unity/plurality, normativity, persons, meaning, lived experience
Does the school account for each of these five phenomena, scored and reported separately as well as pooled?
Performance against strongest identified objections
Against the strongest objection located under a documented search, has the school replied, engaging the objection's actual load-bearing step?
Source quality, independence, and representation fidelity
Above the eligibility bar, how good is the evidence base — independent sources, and is the school represented as it represents itself?
Unresolved costs and defeaters
What does the school still owe, how severe is it, and does the school disclose it?
The scale
Every dimension is scored on the same 0–4 ordinal ladder, plus two non-numeric codes that are deliberately not zeroes. Five levels give a genuine midpoint; a finer scale would invite a precision the evidence does not have.
| Score | Name | Anchor |
|---|---|---|
| 0 | FAILS | The school's own best located formulation cannot deliver what this dimension asks, on its own terms. |
| 1 | ASSERTED | The position is stated but the load-bearing claim is asserted rather than argued in any located source. |
| 2 | ARGUED, UNREPAIRED | A real argument exists, but a specific load-bearing premise is unrepaired against a located objection. |
| 3 | SURVIVES, RESIDUAL DISCLOSED | The argument survives the strongest located objection, and the remaining cost is named rather than hidden. |
| 4 | SURVIVES, RESIDUAL EXPLAINED | As 3, plus the residual is eliminated or explained by a principle the school states in its own sources, independently attested. |
| W | WITHHELD | The dimension depends on a contention that is not yet human-adjudicated. Not a zero. |
| NE | NOT ELIGIBLE | The minimum source-coverage bar is unmet for this school on this dimension. Not a zero — a statement about the research, not the school. |
A 4 may never be earned by an argument this program constructed on a school’s behalf. A 0 carries the heaviest evidentiary burden and is the only score that can veto a ranking. A finding that cuts against both compared parties equally is scored the same way on both — symmetry, not a tiebreaker in disguise.
Weights
All ten dimensions carry equal weight. The reported figure is the mean score across eligible dimensions only — never a bare number without the count of scored, withheld, and not-eligible dimensions alongside it.
The one exception is not a weight at all: a school scoring D1 = 0 — demonstrated internal inconsistency under its own best formulation — receives no aggregate score, regardless of its other nine scores. The veto is symmetric by construction: it applies to Eastern Orthodoxy on identical terms.
Equal weighting was chosen because a non-equal scheme needs a defended theory of relative worth this program does not have — and because the program already has a registered prediction about which two schools will finish top two, which makes any weighting that happens to favor them look unfalsifiable from the outside, however honestly chosen.
Tie policy
Two schools whose mean scores differ by less than 0.40 (on the 0–4 scale) are tied. No ordering between them may be reported, published, or implied by presentation order.
A tie is a reported result, not a deferral: it requires a note on which dimensions the two differ on, an explicit statement of what evidence would break the tie, and a standing rule that a tie may never be broken by the founder’s prior, the registered prediction, tradition, fluency, or volume of sources.
Minimum source coverage
Coverage is a gate, not a score. Below it, a school is not marked down — it is NE (not eligible) on that dimension, and the fact is disclosed. All six roles below are required before any number may be written down for a given school on a given dimension:
- 1A proponent primary or confessional source at first hand, stating the school's own position on this specific dimension.
- 2Serious proponent scholarship arguing the position, not merely restating it (a documented self-attestation fix exists for schools with no institutional literature — capped, and flagged as such).
- 3A serious opponent source that actually engages this dimension, not merely a general critic of the school.
- 4An independent or method-focused source where the claim type requires it — formal claims need formal work, historical claims need historical work.
- 5One documented steelman pass: a recorded search for the school's own best version of the position.
- 6A completed search log — databases, terms, inclusion/exclusion reasons, date, stop rule, and an independence check on the sources found.
A dimension whose proponent evidence is entirely at one remove — secondary summary, a quotation known only through a critic, a search-engine synthesis rather than a fetched page — is capped at 2 regardless of how favorable it would otherwise look.
Reviewer policy
One founder, adjudicating; AI subagents, proposing research and draft scores. No institutional review board, no funded specialist panel — any policy that pretended otherwise would be decoration. An AI pass may propose a score with its evidence; the proposal is not a score until a human adjudicates it, and no aggregate may be computed from unadjudicated proposals.
Before adjudication, a proposed score gets an adversarial red-team pass in fresh context: a separate pass that sees the evidence but not the first pass’s conclusion, tasked only with arguing for the strongest defensible different score. Disagreements of two or more steps must be resolved in writing before scoring.
Every published comparison must disclose the single-adjudicator structure, the founder’s disclosed Eastern Orthodox conviction, the registered prediction (and that it cannot be cited as evidence for anything), and the count of withheld and not-eligible dimensions per school.
Evidence cutoff
Every packet of evidence carries a dated cutoff, and a score inherits the earliest cutoff among the packets it rests on — a score is only as current as its stalest input.
Where a score’s inherited cutoff precedes the program-level cutoff by more than six months, a targeted check for post-dating replies is mandatory before scoring. A post-cutoff source may never be added to a published score quietly — it either triggers a dated addendum, or a rescore with old and new results reported side by side.
What happens where the research is still open
A dimension whose score would depend on a contention that has not yet been human-adjudicated is scored W (withheld) — never a 0, never a 2, never the midpoint, and never the founder’s own prior. Those are the four ways an open question quietly becomes a number, and this rubric is built specifically to refuse all four.
This protects every side at once: an open contention on which Eastern Orthodoxy currently looks strong may not be scored up any more than a rival’s open contention may be scored down. Applied honestly today, before the underlying research is adjudicated, this rubric would return mostly withheld dimensions for every school — which is the instrument doing its job, not a defect in it.
Freeze status — why this is still a draft
Nothing above produces a valid score until the founder has worked through the steps below and formally frozen the version. This is that checklist, published as it actually stands today.
Keep or cut the four dimensions not named in the original commissioning brief (D6–D9)
2026-07-31 — kept, all four, no cuts.
Accept flat equal weights plus the coherence veto, or name a different scheme
2026-07-31 — accepted as drafted.
Accept the 0.40 tie margin
2026-07-31 — accepted as drafted.
Accept the W/NE arithmetic and the 7-of-10 rank-eligibility bar
2026-08-24 — accepted (Option A: W and NE drop out; named-set floor 7; global school ranks forbidden under this version).
Test the rubric on three neutral, non-theological examples for intelligibility
2026-07-31 — all three tests passed on the draft’s own criteria.
Invite specialist review where feasible, then freeze: version, date, content hash, commit SHA
2026-08-24 — frozen as OTR-1.0. Specialist silence did not block the hash. Freeze is not a score.
The instrument’s status is OTR-1.0 (frozen 2026-08-24). Almost every cell is still unmarked. Any number minted from an unmarked sheet — by a human or by a future AI pass that finds this page and starts scoring — is not valid.
Contest a clause, or propose a change
This is exactly the kind of document that should get argued with. If a dimension is mis-scoped, an anchor is ambiguous, a weight is indefensible, or the source-coverage bar is wrong, say so with the specific clause quoted. There is no submission or voting system yet — that is a separate, larger piece of work this pass deliberately did not build — but every message is read by the person who will actually decide whether to change the rubric, not routed to a queue.