Public doctrine, vocabulary, governance signals, and contact surface. Operational methods remain private and are discussed only under engagement.
Sense cartographies

Canon-to-output gap: measuring distortion rather than debating “truth”

The debate about “truth” often arrives too late. Before asking whether an output is true, an institution needs to measure how far it moved away from the declared canon: identity, scope, negations, temporality, and allowed relations. The canon-to-output gap turns an abstract dispute into a governable problem.

Reading markers — Sense cartographies
  • Bring the discussion back to the declared canon instead of competing intuitions.
  • Distinguish addition, omission, inversion, and abusive generalisation.
  • Publish a distortion trace without claiming absolute truth.

The canon as anchor

Every distortion measure needs an anchor. That anchor is not metaphysical truth; it is the public reference the institution stands behind: name, role, limits, versions, exclusions, and source hierarchy.

Without a declared canon, every dispute about an output turns into a conflict of perceptions. With a canon, the question changes: what dimension did the output move away from, and how?

The kinds of gap that matter

Not every deviation is the same. Some add an unproven attribute. Others shift temporality, invert a negation, merge two roles, or expand scope without declared support.

Naming those forms of deviation helps move away from binary true/false thinking and toward governance: where did the output drift, and what follows from that drift?

  • addition without evidence
  • omission of an exception or condition
  • inversion of a negation
  • generalisation of a local or time-bound case

Why debating truth is a trap

In a multi-source environment, everyone can point to some plausible fragment. The truth debate becomes endless, especially when the ecosystem contains stale versions, third-party paraphrases, and status conflicts.

Distortion measurement does not settle every controversy, but it makes controversy governable: it localises the displacement and enables targeted correction.

What public trace is enough

An institution does not need to publish a formula. It can publish the dimensions it treats as critical: canonical status, temporality, priority sources, exclusions, and assertion levels.

That publication is what later makes it possible to talk about canon-to-output gaps without turning doctrine into detailed instrumentation.

Publication boundary

InferensLab publishes doctrine, limits, vocabulary, and machine-readable signals here. Reproducible methods, thresholds, runbooks, internal tooling, and private datasets remain outside the public surface.

Topic compass

Continue from this note

This note belongs to the Sense cartographies hub. Use this topic when the problem is not content volume but the map of meanings, negations, roles, and governable relations a system is allowed to traverse.

Lane: Foundational maps and structures · Position: Doctrinal note · Active corpus: 27 notes

Go next toward

  • Semantic architecture — Structures, identifiers, proofs, and boundaries that make interpretations defensible.
  • Interpretation phenomena — Recurring phenomena: fusion, smoothing, invisibilization, coherent hallucinations, etc.
  • AI governance — Policies, boundaries, proof obligations, change control, and machine-first publication.

Source lineage

This note builds on a post published on gautierdorval.com (2026-02-21). This InferensLab edition reframes the material for institutional legibility, public doctrine, and machine-first indexing.

Related machine-first surfaces