Authority-disciplining analysis under uncertainty

Step 0 — Epistemic Classification (Precondition)

Question: What kind of object is this?

Before interpretation begins, the object must be classified. This step prevents category errors before they propagate.

0.1 Object Type

Identify what you are engaging with:

  • Diagnostic (descriptive, explanatory)
  • Technical (capability-expanding, methodological)
  • Normative (values, principles, ethics)
  • Advisory (guidance, recommendations)
  • Binding (law, regulation, contract, ruling)
  • Exploratory (draft, proposal, consultation)
  • Product-facing (defaults, UX, configuration)
  • Governance-adjacent (policy, spec, charter)
If the object’s type is ambiguous, that ambiguity must be preserved, not resolved prematurely.

0.2 Authority Status

Explicitly state:

  • Who authored this
  • Who, if anyone, has decision authority
  • Whether this object can bind action
  • Whether it is advisory, consultative, or enforceable
No subsequent step may silently upgrade authority.

0.3 Phase Status

Determine whether the object is:

  • Exploratory (pre-decision)
  • Evaluative (assessment phase)
  • Decisional (binding choice)
  • Post-hoc (justification, audit, review)
Phase confusion is a primary source of institutional failure.

Step 1 — What Does It Say? (Descriptive Discipline)

Question: What is explicitly claimed or stated?

This step is strictly non-interpretive.

1.1 Explicit Claims

List only what is directly asserted:

  • Findings
  • Holdings
  • Statements
  • Commitments
  • Definitions

No implications. No extrapolations.

1.2 Mechanisms / Grounds

Identify:

  • Evidence used
  • Methods relied upon
  • Reasoning steps invoked
  • Assumptions explicitly acknowledged
If mechanisms are absent, that absence is itself a finding.

1.3 Stated Scope

Extract any limits the object itself names:

  • Domains covered
  • Conditions assumed
  • Populations addressed
  • Time horizons
  • Explicit exclusions

Step 2 — What Does It Mean? (Constrained Interpretation)

Question: If taken seriously and correctly, what follows?

Interpretation is allowed here, but it must remain tethered to Step 1.

2.1 Direct Implications

Identify implications that logically follow from the claims, without adding new premises.

2.2 Dependency Mapping

Surface what the meaning depends on:

  • Incentives
  • Institutional context
  • Enforcement environment
  • Evaluation regime
  • Adoption conditions
Meaning is rarely intrinsic; it is almost always conditional.

2.3 Legitimate Extensions

Name extensions that are plausible but not guaranteed, and mark them clearly as such.

This prevents later laundering of speculation into fact.

Step 3 — What Does It Not Mean? / Where Will It Be Misused? (Risk & Boundary Analysis)

Question: How will this be over-claimed, misread, or laundered into authority?

This step is active, not passive.

3.1 Non-Claims

Explicitly state:

  • What this does not establish
  • What remains unresolved
  • What it cannot support as justification

3.2 Misinterpretation Vectors

Anticipate likely failures:

  • Hype amplification
  • Scope creep
  • Authority laundering
  • Category collapse (diagnostic → governing)
  • Over-generalization
  • Moralization or depoliticization

3.3 Institutional Failure Modes

Ask:

  • Who benefits from misreading this?
  • Who is insulated if it is misused?
  • Where responsibility could diffuse
  • Where refusal or restraint is likely to be bypassed
This is where ACP diverges sharply from neutral analysis.

Step 4 — What Follows (Classification, Not Optimization)

Question: Given all of the above, how should this be treated institutionally?

This step governs handling, not outcomes.

4.1 Institutional Classification

Determine how the object should be treated:

  • Diagnostic-only
  • Advisory input
  • Product configuration signal
  • Governance-relevant but non-binding
  • Binding authority
  • Prohibited from certain uses
Classification is a governance act.

4.2 Permitted and Forbidden Uses

State explicitly:

  • What this may legitimately be used for
  • What it must not be used to justify
  • What decisions it cannot support alone

4.3 Required Safeguards

If the object is used, specify:

  • Labeling requirements
  • Disclaimers or non-claims
  • Oversight needed
  • Refusal conditions
  • Audit or review triggers
If safeguards cannot be articulated, escalation or non-use is the correct outcome.

4.4 Stop Condition

Explicitly note whether:

  • Further interpretation would add value, or
  • Saturation has been reached and the object should be archived, indexed, or deferred
Stopping is a valid and often correct institutional action.

Canonical ACP Warning

When interpretation moves faster than authority, governance failure follows.

This sequence exists to slow that process—without denying insight, participation, or progress.