Rolling proof infrastructure

Instruments

Operational evaluators for claims, cognition, transfer, and rolling proof.

"Do not score the aura. Score the claim."

Ψ.world instruments convert artifacts into scored, gated, revisable outputs. Each instrument is a distinct evaluative machine with a defined input, scoring grammar, output artifact, and gate condition.

View instruments Read the doctrine
Instrument registry

Evaluators

Instrument Function Input Output Gate State
ARCHON Ψ Scores claims and artifact trails for rigor, transfer, and evidentiary sufficiency Claims, manuscript packet, evaluator notes Structured scorecard + gate decision Pass / Hold / Revise / Deny Live
IVY Writ Issues a rolling credential entry tied to demonstrated cognition Scored packet + revision history Writ artifact + renewable proof ledger Issue / Defer / Renew Pilot
Quill Forces clean claim articulation before scoring begins Raw assertions, draft packet, questions Claim-normalized submission Admit / Reject / Rewrite Live
Crown Decides publication legitimacy and release state Scored claim set Release, hold, or conditional publication Release / Hold / Conditional Internal
Factorizer Evaluates trending claims for product intelligence and market signal Product artifacts, traces, prompts Signal scorecard + registry entry Signal / Noise / Monitor Forthcoming

Each instrument has a burden, a trail, and a gate. Registry updated on instrument state changes.

Instrument specifications

Machine specifications

Problem it solves
Claims circulate without a shared standard for rigor. ARCHON Ψ makes the evaluative burden explicit and auditable.
Input accepted
Claims, manuscript packets, evaluator notes, prior revision records, supporting artifact links
Artifact produced
Structured scorecard with axis scores, composite QTAC₇ rating, gate decision, and revision path
Example decision
HOLD — Evidence trail incomplete. Evidentiary sufficiency score below threshold (5.5/10). Revision required on claims 3 and 7 before resubmission.
Scoring axes
Claim specificity Evidence trail Transfer potential Evidentiary sufficiency Revision discipline Cognitive rigor Decomposability
Hold / Deny conditions: Claims not decomposable into falsifiable assertions; evidence trail gaps without explanation; missing revision acknowledgment; composite QTAC₇ below floor threshold; evaluator notes contradicting submitted claim set.
Problem it solves
Static credentials decay. IVY Writ issues a rolling credential entry that tracks demonstrated cognition over time, not a single test moment.
Input accepted
Scored ARCHON packet + revision history + evaluator signature
Artifact produced
Writ artifact (timestamped) + renewable proof ledger entry + transfer-ready credential summary
Example decision
ISSUE — Three claim cycles complete. Revision discipline demonstrated. Writ issued, valid 180 days. Renewal trigger: one new scored packet.
Scoring axes
Revision consistency Claim volume over time Transfer breadth Gate passage rate Ledger integrity
Defer condition: Insufficient revision history (minimum two scored cycles required); ARCHON Ψ gate not yet passed; claim set does not demonstrate cross-domain transfer. Writ deferred until cycle minimum met.
Problem it solves
Vague claims cannot be scored. Quill forces clean articulation before any evaluative work begins — a submission does not reach ARCHON Ψ without passing Quill.
Input accepted
Raw assertions, draft packets, unstructured questions, verbal notes converted to text
Artifact produced
Claim-normalized submission ready for ARCHON Ψ intake — structured claims, labeled assumptions, evidence slots identified
Example decision
REWRITE — Four of seven claims contain compound assertions. Decompose each into single-predicate form. Return for re-admission.
Evaluation criteria
Claim atomicity Predicate clarity Assumption labeling Evidence slot presence Ambiguity index
Reject / Rewrite conditions: Compound claims (more than one predicate per claim); undefined terms without lexical anchor; missing assumption declaration; ambiguity score above rejection floor. What cannot survive decomposition should not demand belief.
Problem it solves
Publication without a gate produces noise. Crown determines whether a scored claim set has earned public release, conditional visibility, or must remain in the ledger only.
Input accepted
Fully scored ARCHON Ψ claim set + IVY Writ status + submitter's transfer record
Artifact produced
Release status (Public / Conditional / Held) + publication timestamp + visibility scope definition
Example decision
CONDITIONAL — Released to verified reviewers only. Three claims require independent corroboration before full public release. Re-evaluated in 60 days.
Release dimensions
Claim integrity score Corroboration level Transfer readiness Submitter ledger standing Public harm potential
Hold condition: ARCHON Ψ composite below release floor; IVY Writ not yet issued; submitter ledger shows no completed revision cycles; claim set includes unresolved contested assertions.
Problem it solves
Product claims trend without verification. Factorizer applies the same evaluative grammar to product intelligence — separating signal from noise across market categories.
Input accepted
Product artifacts, behavioral traces, market prompts, claim sets from product submissions
Artifact produced
Signal scorecard + trending registry entry + category position map
Example decision
MONITOR — Category signal detected. Insufficient artifact trail for scoring. Flag for re-evaluation when evidence threshold met.
Signal dimensions
Signal strength Category specificity Evidence density Adoption velocity Competitive differentiation Claim-to-artifact ratio
Noise condition: Claim-to-artifact ratio below threshold; no traceable adoption evidence; category signal indistinguishable from adjacent noise. Instrument in active development — acceptance criteria subject to change.
Universal flow

How scoring works

01
Claim
A specific, falsifiable assertion enters the pipeline via Quill
02
Evidence
Artifact trail attached and labeled with source and timestamp
03
Score
ARCHON Ψ evaluates across seven axes, produces QTAC₇ composite
04
Gate
Pass, Hold, Revise, or Deny — decision recorded in ledger
05
Revision
Held claims return with documented changes — trail preserved
06
Reissue
Passed claims enter the registry; IVY Writ issued if thresholds met

The pipeline is linear but revisable. Every entry in the ledger preserves the full revision trail — not just the final state.

Legitimacy by refusal

Example: a held claim

A scoring system earns trust through refusal, not praise. This is what a held claim looks like in the ledger — what was submitted, what failed, and what must change.

ARCHON Ψ — Gate decision: HOLD
SCD-2024-0031 · Submitted 14 Nov
Submitted claim
"Our training methodology produces measurably superior outcomes across all knowledge worker domains and is more effective than any existing credentialing system."
Failed axis
Claim specificity: 3.1/10
Compound assertion — "all domains" and "any existing system" are untestable as stated
Failed axis
Evidence trail: 2.4/10
No artifact linked. "Measurably superior" references no measurement instrument or baseline
Passed axis
Transfer potential: 7.8/10
Core concept has high transfer value if claim is decomposed and supported
Gate verdict
HOLD — Resubmit with: scoped domain, named baseline, linked measurement artifacts, and decomposed claim structure
Evaluator note
The underlying claim may be valid. The submitted form is not scoreable. Decompose into three claims: one per domain, one per comparison baseline, one per measurement artifact. Each is then evaluable on its own merits. The instrument is not hostile — it is precise.
Enter the pipeline

Submit for scoring

Bring 1–3 artifacts, 5–10 claims, and one transfer question. Quill normalizes your submission before it reaches ARCHON Ψ. The ledger preserves every revision.

  • 1–3 artifacts — manuscripts, project records, documented outputs, or research notes with provenance
  • 5–10 claims — specific, single-predicate assertions you are prepared to defend through revision
  • One transfer question — where does this knowledge apply outside its origin domain?
  • Revision acknowledgment — consent to receive a hold decision and resubmit with documented changes
Submit a packet Read the doctrine
Instruments in your pipeline
Live
Quill
Normalizes your claims before scoring begins
Live
ARCHON Ψ
Primary evaluator — scores and issues the gate decision
Pilot
IVY Writ
Issued after two successful scored cycles
Internal
Crown
Publication gate — not part of standard submission flow

"A ledger of demonstrated cognition."