Glossary entry

Typed Decision Model

Glossary entry in System states & architectures. Definitions, checks, and evidence for Typed Decision Model.

Quick takeaways

A model that takes a block of state and a set of questions with fixed answers, and returns a chosen value with a probability distribution and a confidence margin instead of text.

Definition
Minimum evidence
Operational tests

Definition

A model that takes a block of state and a set of questions with fixed answers, and returns a chosen value with a probability distribution and a confidence margin instead of text. It gives no reasons. The first widely adopted product in the class, released in September 2026, is marketed as hundreds of times faster and cheaper than a language model for routing, triage, tool gating, and approval checks.

It matters because a cheap judgment gets placed at every branch, and authority follows it into places that do not look like decisions: the decision threshold that turns a score into an action, the shaping hop that decides what the decider sees, and the reasons nobody produced. The method governs it like any other delegate. The explainer names the six places its authority sits, and the decision record carries a typed_judgment block for what it decided.

StructuralVisibility

Scale: Systemic · Phases: Design, Deployment, Audit · Measurability: Fully measurable · Maturity: Core concept

Scope: D. System states & architectures. Operating postures, records, and architectures that decide whether an error is contained, recorded, or amplified.
Genealogy: A new substrate that tests the method's substrate independence: the unit of governance is the consequential decision, whatever component made it.

Adjacent terms

Decision Threshold Shaping Hop Reason Substitution Confidence-Gated Execution

References

Typed decision models and where their authority lives · Explainer
Decision-model gate on refunds · Worked example

Glossary territory

Operating postures, records, and architectures that decide whether an error is contained, recorded, or amplified.

Explore System states & architectures in the glossary index.

Entry details

Last updated: 2026-09-22
Term version: v1.0.0
Changelog: View updates

Tests and evidence

Operational tests (3)

Every consequential decision a typed decision model contributes to has a decision record with a typed_judgment block naming the pinned model version.
The threshold that acts on the model's output is a policy record, not a configuration value.
Confidence is recorded with its semantics, and a margin is not reported as a probability of being right.

Minimum evidence checklist

Artifact: Decision records with a typed_judgment block: pinned decider, question schema hash, distribution, confidence semantics, threshold policy, and reason provenance.
Behavior: A model upgrade suspending the thresholds tuned against the previous version until they are re-set.
Metric: Share of decisions involving a typed decision model whose record names a pinned model version and a threshold policy.

Copy citation (APA/BibTeX)

Cite this glossary entry Formats: APA, MLA, Chicago, BibTeX, RIS

Version

v1.13.0

Last updated

Sep 22, 2026

Permalink

https://ethotechnics.org/glossary/typed-decision-model

DOI

Pending Zenodo deposit

APA

Ethotechnics Institute Research Team. (2026). Typed Decision Model. Ethotechnics glossary. https://ethotechnics.org/glossary/typed-decision-model

MLA

Ethotechnics Institute Research Team. "Typed Decision Model." Ethotechnics glossary, 2026, https://ethotechnics.org/glossary/typed-decision-model.

Chicago

Ethotechnics Institute Research Team. "Typed Decision Model." Ethotechnics glossary. Sep 22, 2026. https://ethotechnics.org/glossary/typed-decision-model.

BibTeX

@misc{glossary_typed-decision-model,
  title={Typed Decision Model},
  author={Ethotechnics Institute Research Team},
  year={2026},
  howpublished={Ethotechnics glossary},
  url={https://ethotechnics.org/glossary/typed-decision-model},
  version={v1.13.0}
}

RIS

TY  - WEB
TI  - Typed Decision Model
AU  - Ethotechnics Institute Research Team
PY  - 2026
UR  - https://ethotechnics.org/glossary/typed-decision-model
ER  -

Scholarly metadata

Authorship

Ethotechnics Institute Research Team · Ethotechnics Institute · research@ethotechnics.org

Contact: research@ethotechnics.org

Publication details

Published: Dec 3, 2025
Last updated: Sep 22, 2026
Version: v1.13.0
DOI: Pending Zenodo deposit

License: CC BY-SA 4.0

Credit Ethotechnics Institute, include the term title + version, and link to the glossary permalink.

Archive: Wayback Machine history

Changelog

v1.13.0 · 2026-09-22 — Rewrote category descriptions and scope notes to the current framing, replaced vague or promotional definitions in 89 entries and 109 short definitions, removed unsourced adoption and effect claims, and cut filler words and redundant sentences from 38 entries and 20 short definitions. No term was added, removed, or renamed.
v1.12.0 · 2026-09-22 — Added Typed Decision Model and Confidence-Gated Execution, cross-linked to the worked example of a decision-model gate on refunds.
v1.11.0 · 2026-09-22 — Added Decision Threshold, Shaping Hop, and Reason Substitution: the places typed decision models locate authority, shape decisions without making them, and separate the reason from the decision.
v1.10.0 · 2026-09-19 — Promoted 64 orphaned short definitions to full canonical entries across core concepts, governance, decision states, and anti-patterns, and removed three redundant short definitions (procedural-realism, intake-cinema, legibility-illegibility).
v1.9.0 · 2026-09-19 — Added Feedback Proximity, Asymmetric-Capacity Restraint, Adaptive Capacity, Compensatory Adaptation, and Corrective Power, aligned the short definition of Ethotechnics with its canonical authority-and-correction framing, and refreshed the glossary's category scaffolding off the earlier moral-system vocabulary.
v1.8.0 · 2026-09-19 — Added Insulation, the unifying failure mode of the framework: whatever stops the consequences of institutional decisions from traveling back to those who can revise the generating rules, with its channels, its good and bad forms, and its self-ratifying loop.
v1.7.0 · 2026-09-18 — Added Non-Finality, Consequential Contradiction, and The Excluded Case, and generalized the falsified denominator into the falsified ledger — completing the five concerns and the fallibilist ground they stand on.
v1.6.0 · 2026-09-18 — Added Human Non-Substitutability, Safe Incompleteness, Anti-Conversion Rights, Abstention, Constitutional Debt, Discretion Migration, and Corrigibility Rent — the person-side principles, the three abstentions, and the rent concept.
v1.5.0 · 2026-09-18 — Added Exception Absorption, Exception Learning, Case Corrigibility, Structural Corrigibility, Institutional Learning, Corrective Debt, and the Workaround Presumption, formalizing the historical turn: reading institutions through their failure trajectories rather than their stated controls.
v1.4.0 · 2026-09-18 — Added Residual Complexity, Complexity Displacement, Falsified Denominator, Closure Without Remainder, Constitutional Governance, and Post-Optimization Rigor, embedding the foundational contention of burden redistribution.
v1.3.0 · 2026-09-11 — Added Attestation, Liability Record, Carve-Out, Delegation Chain, Composed Latency, and Silent-Harm Discovery, each bound to the standard clause that requires it.
v1.2.0 · 2026-09-08 — Added The Non-Conversion Principle, Exit Cost, and Jurisdictional Path Dependence, each linked to the theory essay that argues for it.
v1.1.0 · 2026-01-09 — Expanded scholarly metadata, operational tests, and provenance notes.
v1.0.0 · 2025-12-03 — Initial glossary release with stable permalinks.

Library search

Library search

Typed Decision Model