Classical Ideological Audit — Cognitive Behavioural Therapy
Instrument: Classical Ideological Audit (CIA) v3.0. Instrument architecture: Dave Kelly. Theoretical foundations: Grant C. Sterling, Eastern Illinois University. Prose rendering: Claude (Anthropic), 2026.
Step 0 — Protocol Activation
Instrument invoked by name by the ratifying authority. Target named: cognitive behavioural therapy as a doctrinal system.
Corpus opened in this run: The Six Philosophical Commitments That Ground Stoic Practice (opened in full, view); CIA — Declaration Form v3.0 (opened in full); Classical Ideological Audit — Epicureanism (opened in full, as the ratified v3.0 architecture precedent); HSA v0.1 and HSA Run — Robertson (opened in full earlier in this session).
Narrowing printed. The remaining documents on the CIA required set — Core Stoicism, Nine Excerpts, SLE v4.4, Two and One-Half Ethical Systems, Stoicism and Self-Interest, the four Sterling commitment essays, the two Six Commitments integration documents, A Brief Reply Re: Dualism, Free Will and Causation, Substance Dualism in Sterling's Stoicism — were not reopened in this run. No prose is drawn from them. Each commitment is stated in the form the opened Six Commitments document gives it, and no governing passage is quoted from a document not opened here. Where a finding would turn on a formulation available only in an unopened document, that is declared at Step 6.
Quarantine. The CPA Run — Donald J. Robertson, the HSA Run — Robertson, and the AEI run on the decatastrophizing lesson are runs on an author and on documents, not on this target. Their verdicts are not carried in and no finding below is inherited from them. Where a finding here converges with one of them, that is noted at Step 6 as convergence, not as evidence.
Target eligibility. CBT is a system of ideas with no bearer, with declared variants and a statable intersection. It is a clinical-theoretical system rather than a political ideology; the CIA target constraint requires a doctrinal system, not a political one, and the target qualifies. The audit runs on presupposed content. Asserted clinical claims are recorded and not audited.
No prior conclusion about the findings is operative.
Self-audit — Step 0: Instrument invoked by name; target named and eligibility tested rather than assumed; corpus opened and the narrowing printed rather than absorbed; prior runs quarantined explicitly; no prior conclusion stated. Result: pass.
Step 1 — Ideology Statement and Variant Identification
Governing question: What is CBT, in propositional form, and what are its significant internal variants?
Core presuppositions
Shared across all declared variants. A practitioner or theorist who abandons any of these has left CBT as a system.
P1 — Cognitive mediation. Emotional and behavioural responses are not produced directly by events but by the person's cognitive processing of them — appraisals, beliefs, interpretations, or the relation he takes toward his own cognitions. Intervening at the cognitive level therefore alters affect and behaviour.
P2 — Naturalistic psychological ontology. The processes so intervened upon are natural processes belonging to the organism: lawfully related to prior conditions, empirically investigable, and modifiable by specifiable procedures. Both the system's mechanism claims and its outcome research require this.
P3 — Functional criterion of cognitive evaluation. A cognition is targeted because it is dysfunctional, maladaptive, unhelpful, or rigidly held — because of what it does to the person. Where truth-testing appears, it serves this criterion rather than replacing it.
P4 — Learning-history etiology. Present dispositions are products of prior learning, developmental history, conditioning, and reinforcement, and are modifiable by new learning delivered through structured practice.
P5 — Distress reduction and restored functioning as the criterion of success. The aim is measurable reduction of clinically significant distress and impairment. Efficacy is established by outcome measurement across populations.
P6 — Non-adjudication of value content. The system does not determine what the person ought to value. Goals are set collaboratively in the person's own terms; where values appear as a technical construct, they are chosen and clarified rather than discovered.
Intersection ruling. A position that abandons P2 or P5 — that treats the inner life as standing outside the natural causal order, or that measures success by something other than distress and functioning — is not a variant of CBT but a different kind of undertaking: philosophical direction, spiritual counsel, or moral formation. That ruling is made before the intersection is taken, and it is the point at which Favorable Variant Selection would enter if it entered at all.
Declared variants
Beckian cognitive therapy — Beck. The cognitive model in three levels: automatic thoughts, intermediate beliefs (rules, attitudes, conditional assumptions), core beliefs. Collaborative empiricism; reality-testing of predictions; identification of cognitive distortions.
Rational emotive behaviour therapy (REBT) — Ellis. The ABC model; irrational beliefs understood as absolutistic demands; vigorous disputing; unconditional self-acceptance. The variant that explicitly claims Epictetus as an ancestor.
Contextual or third-wave CBT — ACT, DBT, MBCT. Grounded in functional contextualism, a philosophy of science that takes successful working as its truth criterion. Substitutes acceptance and cognitive defusion — changing one's relation to a thought — for disputation of the thought's content.
Protocol CBT as delivered — manualized, disorder-specific treatment protocols. Philosophically the thinnest variant and the form in which most people actually encounter the system.
Self-audit — Step 1: Presuppositions stated propositionally rather than as slogans; core identified as the intersection across all four declared variants rather than selected from the most favourable; the intersection ruling stated before the intersection was taken; four variants declared for Stage Two; no finding anticipated. Result: pass.
Step 2 — Stage One: Core Presupposition Audit
Governing question: What does each core presupposition entail for each of the six commitments?
C1 — Substance Dualism
The commitment holds that mind and body are ontologically distinct substances, and that the rational faculty therefore has causal powers not reducible to physical causation. Without it, the dichotomy of Enchiridion 1 has no ontological ground.
Bearing of P2 and P4. CBT locates the cognitions it treats inside the natural causal order. Its mechanism claims — that this procedure produces that change through an identified pathway — and its outcome research — that the change occurs reliably enough to be measured across populations — both require the inner life to be lawfully related to prior conditions and to intervention. P4 makes the point historically as well as prospectively: what the person now believes is the product of a learning history.
A limit, recorded. CBT does not assert materialism as doctrine. It is methodologically naturalistic rather than metaphysically declared, and an individual dualist can and does practise it. The finding is at the level of presupposition, which is what this instrument audits, not at the level of assertion.
Load-bearing test. Could CBT abandon P2 and argue as it does? No. Without the lawful modifiability of the processes it targets, the system has no mechanism, no outcome literature, and no basis for manualization.
Finding: Divergent. Load-bearing. The rational faculty as a substance standing apart from the physical causal order is not available within the framework, and the framework cannot do without what excludes it.
C2 — Libertarian Free Will
The commitment holds that assent is genuine origination — that the Pause is a real interruption of an automatic process and that both paths remain open until the will moves.
Bearing of P1, P2, and P4. The central explanatory unit of Beckian CT is the automatic thought, and the word carries its full weight: the cognition arrives without deliberate origination. The therapeutic move interposes a trained procedure between that arrival and the response. The reliability of the interposition is then a function of practice, rehearsal, and homework — that is, of learning history. Contextual CBT is more explicit still: functional contextualism descends from behaviour-analytic selectionism, in which behaviour is selected by its consequences.
Load-bearing test. Could CBT abandon the account of change as produced by practice and contingency? No. Its entire delivery model — sessions, homework, skills acquisition, relapse prevention — presupposes it.
Finding: Divergent. Load-bearing. What the framework supplies is an interposed and trainable procedure, not origination by a distinct faculty. The architectural resemblance between that interposition and the Pause is registered at Step 4 under this finding, not converted into a separate verdict.
C3 — Ethical Intuitionism
The commitment holds that some moral truths are apprehended directly by the rational faculty, without inference — that the agent does not conclude that only virtue is genuinely good but sees it.
Domain engagement: engaged. CBT evaluates beliefs constantly and adjudicates which goals are worth pursuing. The domain is not absent, and Orthogonal is unavailable.
Bearing of P3 and P6. The route by which CBT decides what a person should believe or pursue runs through two channels: what the person himself endorses, and what the outcome evidence shows reduces distress. Neither is direct rational apprehension of a moral fact. ACT's values construct makes the position explicit — values are freely chosen life directions, clarified rather than discovered. Ellis defines rationality as goal-conduciveness. Beck's reality-testing tests empirical predictions, not moral propositions.
Finding: Divergent. Load-bearing. P6 is not incidental to the system; non-adjudication of value content is written into the professional ethics of the discipline. A faculty that directly apprehends what is genuinely good would make P6 untenable.
C4 — Foundationalism
The commitment holds that some propositions are self-evident stopping points — necessarily true, not matters of arbitrary definition, known directly rather than by inference — and that others are justified by derivation from them.
Bearing of P1. Beckian CT has a genuinely hierarchical architecture terminating in basic items. Automatic thoughts are traced to intermediate beliefs, which are traced to core beliefs, and the core belief is where the tracing stops. The formal structure is that of foundationalism, and the vocabulary is literally foundational.
But what stands at the foundational node. C4 requires that the foundations of ethical knowledge be self-evident propositions, necessarily true, apprehended directly. Beck's core beliefs are contingent products of developmental learning history — I am unlovable, I am incompetent, I am defective — and they are foundational in the causal sense, not the justificatory one. They are basic in a person's psychology; nothing about them is basic in a structure of knowledge. Their expected fate is revision. Everywhere C4 requires a necessarily true self-evident proposition, CBT has a contingent learned schema.
Finding: Structural Imitation. The foundationalist architecture is intact and explicit. The content at every foundational node is of a different kind. This matters practically: it lends the system the appearance of a rigorous epistemic structure while the items in the foundational positions cannot do the work C4 assigns them.
C5 — Correspondence Theory of Truth
The commitment holds that a judgement is true when it corresponds to how things are, and that the corpus requires this specifically in the moral domain: only virtue is genuinely good is true because it corresponds to an objective fact about value.
Bearing of P3. CBT has a correspondence structure and deploys it visibly. Beck's cognitive distortions are named as distortions — mismatches with the facts. Collaborative empiricism tests predictions against outcomes. You believe you will be fired; let us see what the evidence shows.
Where it is deployed and where it is not. The correspondence architecture governs empirical predictions about the world, where it is philosophically uncontroversial. It does not govern the value premise. The appraisal this would be terrible is handled as a rigidity to be held more flexibly, an overestimate to be right-sized, or a thought to be defused from — not as a proposition that is false about value. Contextual CBT states the substitution outright: functional contextualism takes successful working as its truth criterion and declines the correspondence criterion by name. Ellis takes rationality to be what serves the person's goals.
This is the same distribution the Epicureanism run found: correspondence supplied precisely where it is safe and absent precisely where the corpus requires it.
Finding: Structural Imitation. The correspondence architecture is present and is not a pretence; it is simply pointed away from the domain in which C5 is load-bearing for the corpus.
C6 — Moral Realism
The commitment holds that moral facts obtain independently of anyone's beliefs about them, and specifically that only virtue is genuinely good and only vice genuinely evil.
Bearing of P5 and P6. CBT's operative predicates of value are adaptive, functional, healthy, workable, and valued by the client. The first three are biological and anthropological; the fourth is pragmatic; the fifth is preferential. None is a mind-independent moral fact. P5 installs a substantive account of what is worth having — reduced distress and restored functioning — and that account is precisely what Theorem 10 denies is the genuine good. P6 then withholds the discipline's authority to say otherwise. Ellis, the variant with the most explicit philosophy, denies that persons have intrinsic worth at all and treats self-rating as itself the pathology.
Finding: Divergent. Load-bearing on two counts. The good the framework operates with is a hedonic and functional state, and its warrant is client preference together with population outcome data.
Self-audit — Step 2: Each finding grounded in a named core presupposition rather than in a variant. C1's methodological-not-doctrinal character recorded rather than suppressed, and the finding held at the presupposition level where the instrument audits. Orthogonal considered and refused on C3 with the positive engagement shown. Structural Imitation used only where architecture and content were separately identified. No Divergent finding softened by a Structural Imitation verdict. Result: pass.
Step 3 — Dissolution Finding
C1: Divergent. C2: Divergent.
Full Dissolution.
The framework requires those who adopt it as a governing account of their condition to understand the inner life as a set of natural processes shaped by learning history and modifiable by trained procedure. The prohairesis — the faculty of choice, the rational faculty in the corpus sense, the thing Epictetus identifies as what the agent actually is and as the only locus of genuine good — is not present in that account. What occupies its position is a repertoire of skills of varying strength.
Self-audit — Step 3: Dissolution computed from the two Divergent findings by the stated rule; not inflated, since no Partial Convergence was read as Divergent; stated as a framework implication rather than as a claim about any person. Result: pass.
Step 4 — Structural/Content Distinction Layer
Governing question: Where does the system replicate the formal architecture of a commitment while substituting incompatible content?
C4 — what is imitated, what is substituted
Imitated: a dependency hierarchy of beliefs terminating in basic items that are not themselves derived from anything further, with the therapeutic work proceeding by tracing the derived back to the basic.
Substituted: causal basicness for justificatory basicness, and contingent learned schema for necessarily true self-evident proposition. A core belief is where a person's psychology bottoms out. A foundational proposition is where justification bottoms out. The two are architecturally alike and epistemically unrelated.
C5 — what is imitated, what is substituted
Imitated: the assessment of a belief against how things actually are, with a named failure category for beliefs that fail the test.
Substituted: the domain. The test is applied to empirical predictions and withheld from value appraisals, and in the contextual variant the criterion itself is replaced with workability. The result is a system that looks like it corrects false beliefs, and does correct false predictions, while leaving the value judgement untested.
C2 — the resemblance registered, the finding unchanged
The most consequential architectural resemblance in this target sits under C2. CBT interposes a step between the arriving cognition and the response: notice the automatic thought, examine it, respond deliberately rather than reflexively. That is formally the position the Pause occupies in the Five-Step Method, and the resemblance is close enough that the sequences can be laid side by side almost line for line.
The content at that position is a trained skill whose reliability varies with practice, embedded in a framework in which the practice itself is a shaping influence on a lawful process. The content the corpus requires is origination by a faculty that is the genuine first cause of its own assent. Following the ruling the Epicureanism run made on the clinamen, this is registered here rather than converted into a Structural Imitation verdict: a skill and an origination are not two versions of one thing, and the divergence at the level of content is total.
Self-audit — Step 4: Both Structural Imitation findings characterized with the imitated architecture and the substituted content named separately. The C2 resemblance registered under the Divergent finding on the Epicureanism precedent rather than converted, with the precedent cited. Structural Imitation not used to soften any Divergent finding. Result: pass.
Step 5 — Stage Two: Variant Differential Analysis
Governing question: Do the presuppositions distinguishing the declared variants shift any Stage One finding?
Beckian cognitive therapy
This variant carries the strongest case for shifting C5 toward Partial Convergence. Reality-testing is genuine correspondence work, cognitive distortions are named as departures from the facts, and the therapy's characteristic move is to confront a belief with evidence.
The shift is refused. The correspondence apparatus operates on predictions — will this actually happen, how likely is it, what happened the last four times — and on the estimate of one's own capacity. The evaluative premise embedded in the appraisal is handled by proportion: the catastrophe is right-sized rather than denied. Right-sizing presupposes a correct non-zero magnitude, which presupposes that the object carries some genuine weight. That is what C6 denies. The variant is more correspondentist than the others in its methods and identical to them in the domain it leaves untouched.
Differential finding: confirm. No shift. C5 remains Structural Imitation.
REBT
The variant that names Epictetus, and therefore the one where a shift would matter most.
No shift, in the direction that matters. Ellis defines a rational belief as one that serves the person's goals — a pragmatic criterion, which confirms C5 rather than unsettling it. His target is the form of the belief, the absolutistic demand, rather than its truth about value: the correction of I must not fail is I would strongly prefer not to fail, which retains the preference and removes the demand. Preference for an external is exactly what the corpus permits under reservation, so the surface result can coincide — but the route is the moderation of demandingness, not the recognition that the external carries no genuine weight. And unconditional self-acceptance is grounded in the denial that persons have intrinsic worth, not in an objective account of where worth lies.
Differential finding: confirm. The variant closest to the corpus in vocabulary confirms C3, C5, and C6 rather than shifting them. This is the most significant result of Stage Two.
Contextual and third-wave CBT
Functional contextualism declares the pragmatic truth criterion openly rather than leaving it implicit, and defusion abandons content disputation altogether. Where Beckian CT leaves the value judgement untested, this variant declines on principle to test it — the thought is to be held differently, not assessed.
Differential finding: confirm, with intensification. C5 Structural Imitation is made explicit by this variant rather than modified. C4's imitation weakens here, since the belief hierarchy is de-emphasized, but no node acquires content C4 could accept.
Protocol CBT as delivered
The philosophically thinnest variant. P6 operates at maximum: the protocol adjudicates nothing about what is worth wanting.
Differential finding: confirm. No shift on any commitment.
Self-audit — Step 5: Variants selected by philosophical significance. The strongest available case for a shift — Beckian correspondence — examined on its merits and the refusal grounded rather than asserted. No core presupposition reassigned to variant status to soften a finding. No separate verdict issued per variant. Result: pass.
Step 6 — Summary Finding
Part A — Commitment pattern
Commitment Finding
──────────────────────────────────────────────
C1 Substance Dualism Divergent
C2 Libertarian Free Will Divergent
C3 Ethical Intuitionism Divergent
C4 Foundationalism Structural Imitation
C5 Correspondence Theory Structural Imitation
C6 Moral Realism Divergent
──────────────────────────────────────────────
Dissolution Full Dissolution
Four Divergent, two Structural Imitation. No Convergent, no Partial Convergence, no Orthogonal.
Deepest divergence: C6, by way of P6. This differs from the Epicureanism run, where C1 was load-bearing for everything else. Here the ontology is comparatively lightly held — CBT is methodologically naturalistic rather than doctrinally materialist — while the refusal to adjudicate the content of value is constitutive of the discipline, written into its professional ethics, and is what makes the system structurally unable to do what Enchiridion 1 does. C1 supplies the enabling ontology; C6 is where the system cannot move.
Strongest point of convergence: none at the level of finding, and the reason is the important result. The point of maximal resemblance lies outside the commitment set entirely. P1 — that disturbance is produced by the person's judgement of the event rather than by the event — is formally the same causal proposition the corpus draws from Enchiridion 5. It is a genuine agreement, and it is not agreement on any of the six commitments. That is why the resemblance produces no convergent finding, and it is also why the resemblance is so effective at concealing the six divergences behind it.
Part B — Dissolution
Full Dissolution. C1 and C2 both Divergent, and no declared variant shifts either.
Part C — Agent-level implication
An agent who takes up CBT as a governing account of his condition has accepted the following self-description: he is an organism whose beliefs were installed by a learning history, whose responses are mediated by processes open to lawful modification, whose task is the reduction of distress and the restoration of functioning, and whose values are his own to choose because no faculty available to him apprehends which values are correct.
Compared with the Epicurean case, the displacement here is harder to see, and the difference is structural rather than a matter of degree. Epicureanism announces itself as a rival system with a rival account of the soul; an agent choosing it knows he is choosing. CBT presents to a reader of Stoicism as corroboration — the vocabulary matches, the mediational proposition matches, the sequence of noticing and examining and responding matches. What has changed is the content at each node: contingent schema where the corpus has self-evident proposition, workability where the corpus has correspondence, functioning where the corpus has virtue, trained skill where the corpus has origination, and client-chosen value where the corpus has moral fact. An agent can traverse the whole architecture, recognize every landmark, and arrive somewhere else.
That is the precise sense in which the claim that CBT undermines Stoicism is supported by this run. Not that it contradicts Stoic doctrine — largely it does not, at the level of stated content. That it supplies a complete substitute architecture under the same names.
Scope declarations
The CIA has issued no finding on whether CBT is clinically effective, whether it should be used, whether a person in distress should seek or decline it, or whether any practitioner is a sound clinician. Those questions lie outside the instrument. Nothing here is counsel about treatment. The finding concerns the philosophical presuppositions the system requires and what those presuppositions entail for an agent who adopts the framework as a governing account of himself.
No finding here is about Robertson or any other named person. The prior runs on him remained quarantined and contributed no evidence.
Declared items
Convergence with the HSA run, declared and not counted. The Step 5 refusal to shift C5 for Beckian CT rests on the same structure the HSA found at the document level — value appraisal handled by proportion rather than by truth. The two runs were on different targets under quarantine, and the convergence is recorded as convergence only. It is not evidence within this run.
Failure Mode 6 — Symmetry Bias, disclosed. This run returns the identical distribution to the Epicureanism run: four Divergent, two Structural Imitation on C4 and C5, Full Dissolution. There is a substantive reading available — both are naturalistic systems with a hedonic or eudaimonic-functional telos, an empiricist epistemology, and a foundationalist and correspondence architecture deployed off the moral domain — and the load-bearing commitment differs between them, which a mere pattern transfer would not produce. But the identity of the distribution cannot be distinguished from inside the run from transfer of the earlier run's shape. This is the D2 problem in its exact form and it is yours to adjudicate.
Latent gap in the dissolution rule. Not triggered here. The rule computes Full from two Divergent, Partial from one Divergent and one Partial Convergence, and None from neither Divergent. A target returning Divergent on one of C1/C2 and Structural Imitation on the other falls outside all three clauses. C2 came close to that verdict in this run before the Epicureanism precedent settled it. The rule is incomplete and would need a clause before a target reaches that combination.
Corpus narrowing carried from Step 0. Twelve required documents were not reopened. The findings most exposed to that narrowing are C3 and C4, where Sterling's own formulations of intuition and of the three conditions on a self-evident proposition sit in documents not opened here.
Self-audit — Step 6: Summary follows from Steps 2 through 5 with no new material introduced at synthesis. Agent-level implication stated as framework implication, without conversion into a clinical verdict — the analogue here of Failure Mode 3. The absence of any convergent finding examined rather than left standing, and the genuine agreement located and named where it actually sits. Symmetry Bias disclosed rather than defended. Result: pass. CIA v3.0 run complete.
Theoretical foundations: Grant C. Sterling (Eastern Illinois University). Analysis and synthesis: Dave Kelly. Prose rendering: Claude (Anthropic), 2026.
