The Instrument That Cannot Register the Doctrine v1.0
The Instrument That Cannot Register the Doctrine v1.0
Theoretical foundations: Grant C. Sterling (Eastern Illinois University). Analysis and synthesis: Dave Kelly. Prose rendering: Claude (Anthropic). 2026.
I. Where the Question Started and Where It Arrived
The question that opened this line of work was whether cognitive behavioural therapy undermines Stoicism. It is a natural question to ask, and it invites a natural kind of answer: an examination of what proponents of the reconciliation believe, and a demonstration that their beliefs diverge from the classical commitments. That is the shape the corpus’s existing instruments are built to produce. The Classical Presupposition Audit takes a figure’s argumentative record and recovers what he must hold in order to argue as he does. The Handbook Standard Audit takes a document’s practical instructions and measures them against the Enchiridion standard. The Classical Ideological Audit takes a doctrinal system and tests its core presuppositions against the six commitments.
All three of those instruments were run. The CPA on Robertson returned Full Dissolution. The HSA on a single lesson returned Severance. The CIA on cognitive behavioural therapy returned Full Dissolution with Structural Imitation at foundationalism and correspondence. The findings converged, and each was arrived at independently.
But following the thread further produced something the three instruments were not built to catch, and it is a different kind of object. The Stoic Attitudes and Behaviours Scale — the measuring instrument of the Modern Stoicism research programme, developed in its first version by Donald Robertson in 2013 and validated in 2026 with more than eight thousand respondents across a hundred and sixteen countries — contradicts three of the six commitments in its architecture, before any item is written and before any respondent answers. No author’s belief is required for the contradiction. It is not a claim anyone makes. It is a property of a response format, an arithmetic operation, and a validation design.
That is the finding this essay develops. It is worth developing because a displacement built into an apparatus is more durable than one built into a book. A book can be answered. An apparatus is used.
II. The Three Contradictions
The response format
Every item on the scale is answered on a seven-point continuum running from strong disagreement to strong agreement. Scores above the midpoint are recorded as agreement with Stoic principles and scores below it as disagreement, with higher scores indicating stronger Stoic attitudes.
Whatever proposition on the instrument states the doctrine of the good — and the Virtue dimension retains six items, so something in that region is present — enters as a quantity a respondent possesses to a degree. He can hold it at five. He can move from four to six over the course of a training programme, and the movement is the finding the programme reports.
Under the corpus this is not a description of a partial state. Only virtue is genuinely good is a proposition about how things are. It corresponds to an objective fact about value or it does not. A man who holds it at five holds a false value judgment about the remaining externals; he does not hold five-sevenths of a truth. The correspondence commitment and the moral realism commitment together make the proposition binary in a way a Likert scale has no representation for.
The point is not that the scale’s designers deny this. It is that the response format has already decided the question in a way no item wording could reverse.
The arithmetic
The scale carries seven dimensions: beliefs about control, beliefs about happiness, Stoic mindfulness, virtue, benevolence and compassion, ethical development, and Stoic worldview. Six items compose the Virtue dimension, eight the Benevolence and Compassion dimension, eight Beliefs about Happiness. The total score is produced by summing all forty items after reverse-scoring and dividing by forty.
That operation makes the dimensions commensurable. A low Virtue score is offset arithmetically by a high Benevolence score, and the resulting single number is the respondent’s Stoicism.
Robertson himself states the doctrine this violates, and states it correctly. In the 2019 paper with Codd he distinguishes the intrinsic value belonging to character from the selective value belonging to externals, and holds the two completely incommensurate. Incommensurability means precisely that no quantity of the second converts into the first. A scale that averages across dimensions has performed the conversion as a matter of arithmetic.
This finding requires no access to the item text. It follows from the scoring instructions alone.
The validation criterion
The scale earns its standing by correlating positively with flourishing, resilience, life satisfaction and positive affect, and negatively with anger and negative emotion. The convergent measures are the Satisfaction with Life Scale, the Flourishing Scale, the Scale of Positive and Negative Experience, the WHO-5 wellbeing index, the Anger Disorder Scale, and the Brief Resilience Scale.
Consider what this design can and cannot represent. A respondent who scored high on the scale and reported persistent distress would count as evidence against the instrument — as noise, or as a validity problem to be addressed in the next revision. Under the corpus he may simply be a virtuous man having a hard year, which the doctrine not only permits but requires as a live possibility. A respondent who scored low and reported contentment would likewise register as a failure of the scale rather than as what the corpus would call a man comfortable in false judgments.
The validation design cannot distinguish philosophical Stoicism from whatever correlates with feeling well. That is not a defect by its own lights. Its stated purpose is to separate philosophical Stoicism from the colloquial trait of emotional suppression, and the wellbeing correlation is exactly the discriminant that does the separating, since the colloquial trait correlates with worse outcomes. The purpose is legitimate and the execution is careful. The consequence remains: the instrument is anchored to a hedonic and functional standard, and everything measured with it inherits that anchor.
III. A Fourth Point, About Which Propositions Survive
The scale’s development ran from Robertson’s nineteen items in 2013, through dialogue with Christopher Gill and Tim LeBon, to a thirty-one item version in 2015, a thirty-seven item version tested at Stoic Week, a seventy-seven item version circulated for comment, and a sixty-item version validated by Gill and DiGiuseppe. The final forty were selected by exploratory graph analysis and confirmatory factor analysis.
Retention was decided by statistical coherence. An item survives if it loads cleanly on a factor and contributes to internal consistency. Ethical Development retained three items. Stoic Worldview retained three. Benevolence and Compassion retained eight.
Nothing in that procedure attends to whether a proposition is load-bearing in the philosophy. A proposition can be the foundation on which the rest of the system depends and be dropped because respondents answer it inconsistently — and a proposition most people find counterintuitive will tend to answer inconsistently. Counterintuitiveness is precisely the mark of the Stoic doctrine of the good, which asks the agent to deny what pain, death, and defeat appear to be. The selection procedure is not neutral with respect to the corpus’s content. It systematically disfavours exactly the propositions the corpus holds most central.
IV. Why This Is a Different Order of Finding
The three prior runs found displacement in things people wrote. The CPA found what Robertson must hold to argue as he does. The HSA found what a lesson instructs a reader to do. The CIA found what a system presupposes.
Each of those is answerable. An author can be shown the finding and can respond, defend, revise, or refuse. The displacement lives in claims, and claims have authors.
An instrument has no author in that sense. The response format was chosen because seven-point Likert scales are what psychometrics uses. The averaging was chosen because sum scores are how dimensional instruments report. The validation targets were chosen because those are the measures with established psychometric properties. Every one of those decisions is correct by the standards of the discipline making it, and no one of them was a decision about Stoic doctrine. The contradiction is assembled out of methodological defaults, and no participant had to hold a counter-commitment for it to arrive.
This is the Corrective Layer problem in a new location. The corpus has held that counter-commitments radiate into practice without the philosophical argument that generated them ever being examined, and often without being recognised as philosophical at all. A measurement instrument is that process in its most concentrated form: the commitments are not stated, not argued, not attributed to anyone, and not visible to the people using them, because they present as method rather than as doctrine.
And the apparatus governs a research programme. Stoic Week, the resilience training, and the published correlations between Stoicism and wellbeing all take their readings from this instrument. If it cannot represent incommensurability, no study using it can find the virtuous man who is unwell or the untroubled man who is wrong. Those two cases — the ones on which the Stoic doctrine of the good actually turns — fall through the instrument and cannot appear in any result it produces.
V. What Is Not Claimed
No finding about any person. Nothing here concerns LeBon, Robertson, Gill, or any of the authors. The prior runs on Robertson were quarantined from this analysis and contributed nothing to it. What is under examination is a scoring procedure.
No finding about psychometric quality. The validation appears careful by its own standards: five iterations, expert content review, a large and international sample, appropriate factor-analytic methods, reported reliabilities. Nothing above is a criticism of the statistics.
No finding about the research programme’s usefulness. Whether Stoic Week helps its participants is a question this analysis does not reach and could not settle.
And a real limit on the finding itself. The forty items sit in tables behind the publisher’s paywall and have not been read. The three contradictions above follow from the scoring instructions, the dimensional structure, and the validation design, all of which are open — but whether the doctrine of the good appears in the item set at all, and in what wording, is unsettled. If it proves absent, the finding changes character: it becomes an omission rather than a misrepresentation, and a different argument would be required. The study data is deposited at OSF under an open licence and LeBon’s research guide is on Zenodo. Either should close the gap.
VI. The Instrument Gap
The corpus cannot presently issue this finding as a verdict, and the reason is structural.
The CPA takes an argumentative record and returns findings about what a figure must presuppose. The HSA takes practical instructions and measures them against the five-element standard. The CIA takes a doctrinal system with declared variants and audits its core presuppositions. Each of the three takes assertions as its object.
A measurement scale asserts nothing. Its commitments live in four places for which none of the three instruments has a locus: the response format, the aggregation rule, the item retention procedure, and the validation criterion. Those four are where the whole of the analysis above sits, and a scale audit would need to take them as its loci by construction.
Building it would extend the corpus into a domain it has not yet entered — the domain in which philosophical commitments are carried not by claims but by procedures. That domain is large. Diagnostic criteria, outcome measures, actuarial instruments, assessment rubrics, and rating scales of every kind operate in it, and all of them make commitments about value, about persons, and about what can be true of a person, without stating any of them.
Whether the corpus should go there is a decision for the ratifying authority. What this essay establishes is that there is somewhere to go, and that the existing instruments stop at its edge.
Theoretical foundations: Grant C. Sterling (Eastern Illinois University). Analysis and synthesis: Dave Kelly. Prose rendering: Claude (Anthropic). 2026.
0 Comments:
Post a Comment
<< Home