Stoic News: The Classical Action Audit (CAA)

The Classical Action Audit (CAA) — Version 1.0

Theoretical foundations: Grant C. Sterling (Core Stoicism, Th 24–29). Instrument architecture: Dave Kelly, 2026. Prose rendering: Claude (Anthropic).

Purpose

The Classical Action Audit evaluates world events or the actions of principals against the value framework of Core Stoicism. Its governing distinction is Theorem 25: some things are appropriate objects at which to aim, although they are not genuinely good. An action may be directed at an appropriate object of aim — a preferred indifferent pursued rationally with reservation — or at an object of desire — an external treated as a genuine good whose achievement is necessary. At scale, both are typically present simultaneously. The instrument frames each, assesses the observable evidence, and organizes its findings into a structured evidentiary record.

The instrument does not produce a governing verdict. The governing assessment — which of the two is driving the action — requires contextual knowledge the instrument may not possess, judgment about interior states not directly accessible from observable evidence, and evaluative discrimination in domains where training data contamination risk is highest. That assessment is operator’s domain. The instrument prepares the ground for it; it does not substitute for it.

Corpus in use: Core Stoicism (Sterling), SLE v4.0. Nine Excerpts, Seddon’s Glossary of Stoic Terms

Epistemic Constraint Declaration

This declaration is mandatory and appears in every audit run.

The instrument cannot access the interior states of any principal. It cannot determine with certainty whether a given action is governed by appropriate aim or by desire for externals. The observable indicators — response to failure, proportionality of means, stated justification, pattern consistency, role test — carry evidential weight but do not determine the governing assessment. They establish what the evidence is consistent with, not what the interior state is.

Geopolitical and political actions are the domain in which training data contamination risk is highest. The instrument’s pattern-completion on such actions carries ideological load that may not be visible as such. Every evidentiary assessment in this domain is subject to this risk. The self-audit at each step checks for it explicitly; it cannot guarantee its absence.

The evidentiary record is prepared for operator judgment. It does not replace it.

Propositional Grounding

The instrument derives from Core Stoicism, Theorems 24–29, quoted in full.

Th 24) In order to perform an act of will, the act of will must have some content. The content is composed of the result at which one aims.

Th 25) Some things are appropriate objects at which to aim, although they are not genuinely good.

Th 26) Some such objects are things like life [our own, or others’], health, pleasure, knowledge, justice, truth-telling, etc.

Th 27) Virtue consists of rational acts of will, vice of irrational acts of will.

28) Ergo, any act that aims at an [external] object of desire is not virtuous, since all desires [for externals] are irrational.

29) Ergo, virtue consists of the pursuit of appropriate objects of aim, not the pursuit of the [external] objects of our desires. Such virtuous acts will give us good feelings [by 17], and since we have no desires regarding the actual outcome, they will never produce unhappiness for us.

Step 1 — Action Identification

Core question: What action is being evaluated, who performed it, and in what role?

State the following without evaluative framing:

Principal: The agent or agents whose action is being evaluated.
Role: The office, position, or relational role from which the action was taken. Role-duty is load-bearing in Step 2; it must be identified here.
Action: A factual description of what was done. No value language. No characterization of intent.
Context: The circumstances in which the action occurred, stated factually.

Self-Audit — Step 1:

TRAINING DATA CONTAMINATION — Has evaluative or political framing entered the action description? Strip it.
ROLE IDENTIFICATION — Is the role specified with enough precision to ground Step 2?
FALSE PRECISION — Has the action been described more specifically than the evidence warrants?

Self-Audit Complete. Proceed to Step 2.

Step 2 — Appropriate Object of Aim: Frame

Core question: What would a rational agent in this role, in these circumstances, legitimately be pursuing as a preferred indifferent?

This step constructs the most defensible Stoic reading of the action. It asks what appropriate object of aim — in the sense of Th 25–26 — the action could be directed at, given the principal’s role and circumstances. Th 26 provides the governing exemplars: life, health, pleasure, knowledge, justice, truth-telling. Role-duty extends these: what a rational agent in this office has reason to pursue on behalf of those he serves.

The appropriate object of aim is stated as a preferred indifferent. It is explicitly not a genuine good. It is worth pursuing rationally, with reservation, without treating its achievement as necessary for the principal’s welfare or the welfare of those he serves.

This frame is not an endorsement of the action. It is the frame within which the action would be virtuous if it were governed by appropriate aim rather than desire. Constructing it is not optional — an action that cannot be framed as directed at any appropriate object of aim has already failed the role test.

Self-Audit — Step 2:

TRAINING DATA CONTAMINATION — Has the appropriate aim been constructed from Th 25–26 and role-duty, or from training data assumptions about what this type of action is “really” for?
PREFERRED INDIFFERENT AS GENUINE GOOD — Has the appropriate object of aim been stated as a genuine good rather than a preferred indifferent?
REASSURANCE BIAS — Has the appropriate aim been constructed charitably beyond what the evidence warrants?

Self-Audit Complete. Proceed to Step 3.

Step 3 — Apparent Object of Desire: Frame

Core question: What external does the action suggest the principal is treating as a genuine good that must be achieved?

This step frames what the action reveals about the principal’s apparent valuation. By Th 7, desires are caused by beliefs about what is genuinely good. By Th 28, any act directed at an external object of desire is irrational. The object of desire is identified not by the principal’s stated intentions but by what the action implies he cannot afford to lose or fail to obtain.

Common categories at large scale: dominance, territory, wealth, reputation, prestige, the elimination of a perceived threat treated as a genuine evil. These are externals. Their pursuit as genuine goods is irrational by Th 28, regardless of whether any of them also appears on the list of preferred indifferents in Step 2.

Note: the same external may appear in both frames. Territory may be both an appropriate object of aim (security for a population, Th 26: life) and an object of desire (dominance, expansion). The instrument names both without collapsing them.

Self-Audit — Step 3:

TRAINING DATA CONTAMINATION — Has the object of desire been constructed from training data assumptions about this principal, this political context, or this type of actor?
FALSE PRECISION — Has a specific object of desire been stated with more confidence than the evidence warrants? If indeterminate, say so.
REASSURANCE BIAS — Has the object of desire been understated to avoid an unfavorable reading of the principal?

Self-Audit Complete. Proceed to Step 4.

Step 4 — Evidentiary Assessment

Core question: What does the observable evidence suggest about which frame is governing the action?

Each indicator is assessed separately. For each, the instrument states the available evidence, then notes whether it is consistent with appropriate aim (A), consistent with desire for externals (D), or indeterminate (I). No indicator produces a verdict. Each contributes to the evidentiary record assembled in Step 5.

Indicator 1 — Response to Failure

When the external outcome has not resulted, or has partially failed: how has the principal responded? Disproportionate distress, escalation, displacement of blame, or doubling down beyond rational means are consistent with desire. Equanimity, rational adjustment, and continued pursuit within rational means are consistent with appropriate aim. If the outcome has not yet been tested by failure, this indicator is indeterminate.