Devin Arkema

A system can keep producing behavior long after it has lost the capacity to revise the constraints that produce it. This paper names that condition terminal recursion and argues that it is a form of moral death: not a moral failure, not corruption, not injury, but the collapse of the condition of possibility for moral agency itself. Building on a process account of persistence as governance continuity, I introduce a second-order quantity, recursive capacity: a system's standing ability to revise its own revision function in response to evidence that it is wrong. Terminal recursion is the state in which that capacity has gone to zero while first-order output continues, and it is distinct from mere repetition, which is recoverable, because it is the loss of the capacity to repair repetition. The condition is the structural, substrate-neutral generalization of Frankfurt's wanton, reaching past persons to institutions and artificial agents: any system that has lost the capacity to be moved by reasons to change how it changes. I propose five external diagnostic markers, evidential inertness, output continuity, justificatory collapse, refusal foreclosure, and non-recovery from within, and confront the reflexive problem that no system can certify its own recursive capacity from inside, which the Indistinguishability Thesis forecloses, so the verdict is directional, defeasible, and external by necessity. Because the faculty the diagnosis concerns is the one persuasion would address, it redirects intervention from internal appeal to external interruption and restoration; it is emphatically not a license to abandon or dehumanize what it names. The payoff is a sharper target for diagnosis in ethics, institutional design, and AI safety: the thing worth detecting is not misbehavior but the loss of the capacity that makes correction possible.

Modern political philosophy has theorized sovereignty through three dominant frameworks: liberal autonomy as boundary-maintenance against intrusion, republican non-domination as freedom from arbitrary interference capacity, and communitarian embedded recognition. This paper argues that the dominant operational models within each tradition rely on the Presupposition of Prior Boundedness (PPB): the assumption that a bounded, self-consistent subject existed prior to its constitutive entanglement with the political and technological systems that now shape it. PPB is false in the general case, and its falsity has a sharp epistemic consequence, the Indistinguishability Thesis (IT): from within a sufficiently entangled relation, no reliable first-person method distinguishes genuine reflective endorsement from evaluative capture. The paper develops the positive concept this diagnosis demands. Navigational sovereignty is the recursive capacity to orient within constitutive entanglement: to retain the ability to participate in revising the loops through which one's agency is enacted. The concept is given formal content through two structural measures introduced here, Reversibility Symmetry (Rs), the ratio of exit cost to entry cost across energy, information, and network dimensions, and Minimum Exit Viability (MEV), the threshold proportion of simultaneous exit required for any agent to leave with utility parity. Because the IT bars first-person adjudication, these measures are deliberately third-personal: they assign capture detection to external structural assessment while locating severance, the preserved capacity to leave, with agents themselves. Institutional implications are derived for platform regulation, algorithmic governance, and AI alignment, and four objections are answered, including the framework's application to itself.

Serious ethical frameworks, those that structure practitioner identity or claim durable authority over conduct, face a structural vulnerability: their own tools can be commandeered to justify precisely the failures they were designed to prevent. Three traditions have attempted to address this capture problem. Popperian falsificationism exposes frameworks to external refutation; Habermasian discourse ethics validates them through inclusive deliberation; Derridean auto-deconstruction surfaces the internal contradictions they cannot resolve. This paper argues that each of these is insufficient as a sole capture-response architecture, and develops a fourth model: recursive self-refusal (RSR), the constitutive capacity of an ethical framework to generate its own internal opposition from within its own commitments. The model is reconstructed from the Anti-Codex, Book 2 of Arkemedics: The Recursive Ethic, a full-length text written against the framework it extends. The paper's central analytical move is a decomposition of the capture problem into two structurally distinct sub-problems: detection (knowing that the framework has captured you), which the Indistinguishability Thesis assigns to external verification, and severance (retaining the ability to stop), which no external party can exercise on the practitioner's behalf and which must therefore be maintained internally, as pre-committed architecture rather than in-the-moment judgment. RSR is structurally necessary not as a capture detector, which it cannot be, but as the severance capacity that makes any capture verdict actionable. A framework that cannot generate its own internal opposition has, by that incapacity, already begun the slide toward the condition it was designed to prevent.

The Indistinguishability Thesis (IT) holds that for any sufficiently entangled system-self relation, there exists no epistemically reliable method available from within the relation for distinguishing accurate evaluative awareness from evaluative capture. This paper argues that the IT is not only a diagnosis of the limits of self-knowledge but a structural engineering constraint: a fact about the epistemic geometry of entangled evaluation that determines what valid verification architecture must look like, in the same way that physical constraints determine what valid structural engineering must look like. If internal capture detection is unavailable by structural necessity rather than contingent failure, the correct response is not better internal detectors, which face the same limit at one remove, but external verification built in as a constitutive feature of system design. The paper derives three architectural principles from the IT: the Structural Independence Requirement (verification must not share the evaluative infrastructure of the loop it evaluates, a stricter condition than procedural independence), the Normative Trigger Requirement (external review must be activated by conditions that carry procedural force independent of the system's self-assessment), and the Living Evidence Requirement (because entanglement deepens over time, verification structures must maintain their own independence as a dynamic property). These principles are shown to be instantiated, outside the research program that derived them, by the data monitoring committee architecture of regulated clinical trials, and to be violated in nameable ways by the oversight architectures with documented failure records; in-program mechanisms at agent and institutional scale are consistent with the derivation. Applications are developed in AI governance, institutional research ethics, and platform accountability, and the infinite-regress objection is answered by replacing an epistemic termination criterion with a structural one.

A durable folk intuition holds that a worker earning some respectable multiple of the legal minimum wage (five times the floor is a common illustrative figure) ought to be able to support a family on one income. This paper subjects that intuition to a two-endpoint empirical test and then asks what the result means for the political philosophy of wage floors. Using the Bureau of Labor Statistics Consumer Expenditure Survey's actual 1984 five-person consumer unit and the Economic Policy Institute's 2026 family budget for a two-adult, three-child household in Ottawa County, Michigan, I develop the Family Support Ratio (FSR), defined as household income at a given wage multiple divided by a family budget, and show that a single earner at five times the federal minimum wage falls below a modest adequacy budget under every read tested, with adequacy beginning near 8.4 times the federal floor. The raw longitudinal headline, a required single-earner multiple rising from roughly 4.2× in 1984 to 8.4× in 2026, is shown to be suggestive but not clean: it bundles a construct gap, a geography gap, an earner-count gap, and a payer-composition gap alongside genuine category divergence. The paper's defensible findings are narrower and stronger. First, the decay is concentrated, not general: food and transportation tracked general inflation almost exactly, while healthcare diverged by a factor bounded between 2.0 (actual-to-actual spending of five-or-more-person households over 1988–2024, a window opening four years after the 1984 anchor with the federal floor unchanged at $3.35) and 5.2 (the adequacy-standard ceiling), with the category-specific medical price index (1.75× relative divergence) confirming that price alone explains only part of the movement and payer-composition migration explaining much of the rest. Second, childcare monetized from approximately zero as a recorded budget line in 1984 to $18,652 in the 2026 adequacy budget, a structural shift inseparable from the move from the 1984 five-person unit's average of 2.2 earners to the single-earner stress case. Third, the wage ladder's decay is concentrated in the federal statute itself: against Michigan's maintained 2026 floor of $13.73, the same adequacy budget requires only about 4.4 multiples, so the 5× intuition roughly survives where the floor has been maintained and fails where it has been allowed to drift. The paper closes by arguing that a frozen wage floor quietly corrupts an entire public grammar of economic position, and that which floor counts as the floor is itself a governance choice with distributive consequences.

This paper develops a unified taxonomy of governance pathology, arguing that four structural failure modes (drift, capture, presentism, and integrity inversion) recur with diagnostic consistency across domains as distinct as biological organisms, financial markets, regulatory institutions, and artificial-intelligence oversight regimes. Grounding the taxonomy in a formal account of governance continuity, on which a system persists as itself just in case the function that revises its states remains continuous across reconstruction, I argue that the four modes are not a family resemblance among loosely related phenomena but the four principal loci at which a revision function can be corrupted while behavioral output continues: its increment-tracking (drift), its authority relation (capture), its temporal domain (presentism), and its justificatory directionality (integrity inversion). Each mode is tested against the external literature that knows it best (Simon and Selznick on incremental deviation, Stigler and the cultural-capture literature on regulatory subordination, the cognitive-bias and intergenerational-justice literatures on temporal collapse, and Bateson, Merton, and Lessig on self-justifying institutional corruption), and each is shown to be structurally distinct from its nearest neighbors, including the gradual-institutional-change typology of Streeck and Thelen and the metric-corruption tradition of Goodhart and Campbell. Against the objection that the cross-domain convergence is merely metaphorical, I argue that the modes share a formal structure that licenses comparison and, more importantly, prescription: because each mode corrupts a specific component of the revision function, each admits a specific structural countermeasure, and a governance architecture is only as strong as the weakest of the four. The taxonomy's completeness is defended as a claim about structure rather than history, and its self-application (what it means that a framework proposing this taxonomy is itself a governance system) is addressed through the corollary of self-audit insufficiency established in prior work.

A family of recursive ethical frameworks takes the spiral as its governing image: a return that advances rather than merely repeats. The image is doing a great deal of work, and it has never been made to earn it. This paper asks whether the distinction between healthy recursion and mere repetition can be formalized, or whether "recursion" is destined to remain a virtue term that ratifies whatever a framework was already doing. I argue that the distinction is real and statable, and I state it. Borrowing the phase-portrait vocabulary of dynamical-systems theory, I model self-revision as a trajectory in a state space and distinguish two shapes that trajectory can take: a limit cycle, a closed orbit that returns to its own prior states (this is repetition), and a spiral, an open trajectory that returns in one coordinate while displacing in another (this is recursion). A third shape shadows the second: the converging spiral, which displaces on every return while tightening toward a fixed point. It is capture wearing recursion's shape. On the governance-continuity formula S(t+1) = R(S(t), E(t)), repetition is the degenerate case in which the revision operator has gone blind to its evidential argument and is iterating on state alone. From this I extract three criteria a return must meet to count as recursive rather than repetitive: novelty (it incorporates content not derivable from the prior state, the analog of Lakatos's progressive problemshift), non-recovery (the prior position cannot be reoccupied as live belief without suppressing what the return disclosed, against the AGM Recovery postulate), and independence (its conclusion survives the bracketing of the commitment that produced it). The converging spiral can mimic each criterion on any single pass, which is why the verdict is directional and historical, read across returns rather than turn-by-turn. I then turn the criteria on the recursive framework itself. The reflexive result is uncomfortable and, I argue, correct: no framework can certify from inside its own loop that it is spiraling rather than spinning, because the Indistinguishability Thesis forecloses exactly that first-person confirmation. Spiraling, properly understood, is not a property a framework can claim. It is a discipline it can only be held to.

Self-governance under continuous operation has been studied under several rubrics — precommitment and self-binding, hyperbolic discounting and self-control, market-stability mechanisms, and the design of interruptible artificial agents — but the literature lacks a unified structural account of one specific architecture that recurs across all of these domains. I call this architecture the circuit-breaker: a pre-committed, condition-triggered, automatically activated, time-bounded pause that requires explicit action to resume. I argue that the circuit-breaker is a distinct self-governance form, not a special case of rate-limiting, stop-loss, cooling-off, or suspension, and that its distinctness can be made precise through five structural features and three operational tests. The decisive feature is the last: passive resumption is structurally precluded, so that the default after the pause is not a return to operation but a required re-engagement event that forces fresh judgment to bear. I trace the architecture across financial-market halts, clinical-trial data-monitoring stopping rules, central-bank emergency facilities, AI kill-switches and capability bounds, and legislative emergency-brake provisions, and I draw a load-bearing distinction between circuit-breakers that operate on system state and those that operate on agent state. I catalogue three characteristic failure modes — premature re-engagement, threshold drift, and gather-context-mandate violation — and show that each is a failure of a specific structural feature rather than a generic lapse of will. I then argue that the circuit-breaker is a worked instance of distributed corrigibility (Arkema, From Static Oversight to Distributed Corrigibility) applied to an agent's governance of its own operation, and close with a methodological observation: an agent's pause-discipline is observable in its operational log, which makes operational self-governance a tractable empirical domain for computational political philosophy in a way that simulation outputs alone are not.

A providence that repairs a fault in the world must do so by some route, and the route is not incidental to what the repair is. This paper isolates a structural distinction in the route of providential supplementation that the philosophy-of-religion literature has felt domain by domain without naming: the difference between supplementation that operates within the natural order (inserting a natural agent into the gap a natural failure has opened) and supplementation that operates across it (reaching the target directly by a channel that bypasses the ordinary governing apparatus). Two passages of the Corpus Hermeticum furnish unusually clean exemplars of each topology, co-present in a single transmitted corpus and, on the surviving text, not reconciled. In CH XVIII a cicada lights on a cithara whose string has broken and supplies the missing note from within the natural sound-producing order; in CH XVI a ray of light reaches the rational part of the soul and nullifies the daemonic government of fate by bypassing the cosmic chain that ordinarily rules embodied life. I argue that these are not two illustrations of one idea but two architectures, each with its own structural requirements, and that the distinction is logically prior to, and cuts across, the familiar trichotomy of general providence, special providence, and miracle. I distinguish the two topologies from four neighbors, show that the scholastic concursus debate and the late-twentieth-century Divine Action Project are both clarified (as a redescriptive lens, not a reduction) by treating them as disputes over route, boundary, and grain rather than over whether God acts, and develop the corollary that topology ascription is reading-relative: how finely one individuates "the natural order" determines whether a given supplementation counts as within or across. I close by generalizing the distinction beyond theology (to the governance of artificial agents, to clinical medicine, and to anomaly-explanation in the sciences), where the same two routes recur and the same recognition condition governs which one has been taken.

The closing parable of the seventh chapter of the Zhuangzi is conventionally read as a generic Daoist warning against meddling: the faceless emperor Hun-tun is bored seven holes by his guests and dies. I argue that this reading misses the parable's sharpest structural claim. The two intervening emperors do not act from malice, ignorance, or assumed superiority; they act in gratitude. The Chinese pivots on 報 (bào, "to repay, requite") a benefaction (德, dé) received, and the destruction follows directly from the attempt to return hospitality by making the host more like the guests. I extract a distinct failure mode of well-intentioned cross-frame intervention—reciprocity-as-destruction—and specify it in four conditions: the intervener acts in good faith, characteristically as a return of benefit received; the intervener's frame can read the host's perceived lack but not the host's positive condition; the intervener discharges the relation by extending its own frame onto the host; and the extension destroys the very condition that made the hosting possible. I distinguish the mode from coercion, negligence, cultural imperialism, and paternalism, and from the gift-theory tradition descending from Mauss, which saw that the gift binds but did not isolate the case in which the return itself is the lethal act. I then trace the pattern as a set of diagnostic sketches across six domains—international development, museum repatriation, contact with isolated peoples, AI value alignment, cross-cultural medical aid, and immigrant integration—distinguishing high-fit cases from frankly analogical ones, and close with three tests an intervention can be made to face before it proceeds: the gratitude test, the condition-versus-lack test, and the more-like-us test. The parable, I argue, is a remarkably early and compact statement of a worry the contemporary literatures feel domain by domain without having named.

Recursive Ethics holds that every self-governing system is vulnerable to four failure modes — drift, capture, presentism, and integrity inversion. This paper asks whether Recursive Ethics can detect those failures in itself. I argue that it cannot do so reliably from the inside, and that the right response is to rebuild part of its self-correction outside the loop being corrected. In fuller terms: the framework is itself a system that governs the deliberation of those who hold it, and it issues an epistemic limit — the Indistinguishability Thesis — holding that no sufficiently entangled system-self relation can confirm, from within, whether its evaluative capacities are accurate or captured. If the four modes are universal they apply to the framework; if the thesis holds, the framework cannot certify from inside the practitioner-framework loop that it has escaped them. The convergent result is a corollary I call self-audit insufficiency: no self-governing framework can guarantee detection of its own degradation using resources confined to the degraded loop. The response is architectural rather than confessional — three mechanisms (witnessed re-derivation, an equalization constraint on self-interpretation, and trigger-based external re-engagement) that relocate part of the revision function outside the loop it revises, resting on a refusal-capacity floor whose loss is legible only from outside. The aim is not certainty, which the thesis forecloses, but the conversion of a closed loop into an open one in which the framework can still be stopped by a witness with standing to stop it.

Zhuangzi chapter seven closes a famous exchange with the phrase 天下治矣 — on the standard English renderings, "then the world will be governed." T'ien Kên has asked an active-causative question — how is the governing of the world to be done (為)? — and the nameless man answers, after refusing the question once, in a grammatically different register: do these things, and the world will-be-ordered-of-itself. The shift is easy to read past in translation, and most translators read past it, because English has no settled grammatical home for what the Chinese does here. The verb 治 is neither straightforwardly active, with the ruler as agent, nor straightforwardly passive, with the world as a patient acted upon by some suppressed agent; it is labile, and in this construction it expresses a state of affairs in which ordering happens in the world rather than being imposed on it. This paper argues that the grammatical fact is a philosophical fact. The category of voice that classical Indo-European grammar called the middle — Benveniste's "internal diathesis," in which the subject is interior to the process rather than its external cause — is the structural-linguistic register in which a cluster of Daoist political concepts is most cleanly stated: subtractive legitimacy, non-paternalist transformation, and ziran-preservation as a governing posture. None of these concepts is straightforwardly available in a grammar that offers only active and passive, because each makes a claim about a condition that is neither caused-by-the-ruler nor merely-tolerated-by-the-ruler but participated-in. I recover the middle voice as a grammatical category, locate its functional analogue in the labile verbs and self-(自) constructions of the Daoist corpus, show that the most translation-resistant Daoist political concepts share a middle-voice precondition, diagnose the cost the missing register imposes on English translation and on Anglophone political theory, and examine three contemporary programs — Pettit on non-domination, Sen on capability, Honneth on recognition — that reach toward a middle-voice claim from the English side without quite specifying it. I close with a methodological rule for cross-tradition borrowing: importing a Daoist political concept requires either restoring a middle-voice register in the target language or explicitly flagging the grammatical translation as the substantive philosophical move it is. The claim is not linguistic determinism. It is that grammatical resources make some thoughts cheap to state and others expensive, and that the expense is where the philosophical work has been quietly happening.

A recurring move in political theory and in cross-tradition borrowing reads a text that describes what legitimate rule looks like once it obtains as though it specified how to produce that condition. This paper isolates the move and names the distinction it violates. The depictive register characterizes the relational and cognitive signature of legitimate rule in operation: the people do not depend on the ruler; the work is done and they say they did it themselves; the world is, on the image, governed of itself. The constructive register specifies the mechanism, the failure modes, and the threshold that would produce a state with that signature. The distinction is not the fact/value gap, and it cuts across the ideal/non-ideal debate in contemporary justice theory rather than coinciding with it; §IV separates it from these and other neighbors with care. I formalize the distinction with three diagnostic tests, applied per claim rather than per text: mechanism specification, failure-mode specification, and threshold specification. A claim answering none of the three is in the depictive register, so reading it constructively necessarily supplies the mechanism, failure modes, and threshold from somewhere other than the text. I then show what that "somewhere" is: the reader's own contemporary institutional priors, smuggled in under the source's name. The paper traces the conflation across four domains where its cost is high, the governmentality reception of texts that specify no governmental design, the management literature that reads Daoist and Confucian description as leadership technique, the alignment literature that treats descriptions of aligned behavior as architectural specifications, and the Schmittian-revival reading of sovereign-decision narratives, which exhibits the conflation running in the opposite direction, a constructive apparatus read back into a depictive source as though the source had specified it all along. The paper closes by stating what legitimate use of a depictive text in a design conversation requires: four conditions of register-discipline that do not bar the depictive text from the conversation but require its contribution to be correctly credited. The distinction is a diagnostic, not a verdict; some classical political texts are genuinely constructive, and the paper marks the contrast with the cleanest available case, the Legalist tradition that shared the Daoists' political vocabulary and wrote in the opposite register.

This paper proposes a structural account of messianic governance: authority defined not by charisma, conquest, inherited rank, or sacred title, but by what it produces in those it governs — the restored capacity to stand, refuse, judge, participate, and continue without being absorbed by the leader's power. I read "messianic" structurally rather than doctrinally, as a repair-function within damaged political fields. The central claim is that messianic governance is tested not by obedience, admiration, or unity, but by whether the governed become more capable of dignified agency after contact than before, measured against a stated baseline and across a boundary the authority cannot quietly draw to flatter itself. The argument develops through early Genesis, which gives the messianic chiefly by absence: vengeance sovereignty, renown without mandate, diagnostic totalization, and uninspected mandate. It then reads the Noahic covenant as self-limiting sovereignty and Joseph as the extended positive case. Joseph's famine administration in Genesis 47 forces the paper's central complication: repair is scoped, and restoration in one field can coexist with totalization in another. A comparative section on Daoism, Confucianism, and Mahāyāna Buddhism clarifies the broader family of capacity-producing authority while isolating the distinctive messianic demand: restored standing to contest authority itself.

Political philosophy and regulatory theory possess three well-developed critiques of rule-by-promulgation: the legitimacy critique (the rule binds without the consent of those it governs), the knowledge critique (the regulator cannot know enough to write a rule that fits the field), and the public-choice critique (the regulator is captured by, or self-interested with respect to, the parties the rule ostensibly disciplines). This paper isolates and formalizes a fourth, which the existing literature gestures toward but never states cleanly: the futility argument. Its claim is that even a legitimate, well-informed, and uncaptured regulator who promulgates a rule generates, in the very act of promulgation, the evasion dynamics that hollow the rule out. The mechanism is legibility: a promulgated rule must be made conspicuous in order to be obeyed, and its conspicuousness is precisely what supplies the governed with a determinate target to evade. The argument's sharpest pre-modern statement is in Zhuangzi Chapter 7, whose impossibility-by-scale metaphors and bird-and-mouse evasion couplet make a capacity claim — rule by promulgated regulation is unworkable, not merely improper — rather than a moral judgment. I reconstruct that argument, holding apart from the outset what the text describes (a ruler who governs his own internal state, so that the world is, on the chapter's image, governed of itself) from what it does not design. I then distinguish the futility argument from four contemporary cognates that each capture a fragment of it: Goodhart's law and Campbell's law (which concern metrics, not rules, and locate the failure in the measure); regulatory-arbitrage economics and Kane's regulatory dialectic (which treat evasion as a friction to be minimized, not as a structural property of promulgation as such); the cobra effect and the perverse-incentive literature (which catalogue perverse outcomes without a general mechanism, and which describe perversity rather than futility); and Hirschman's "futility thesis" together with the Lucas critique (which respectively locate futility in the immovability of deep social structures and in the invalidity of forecasting models, not in the self-hollowing of the promulgated rule). The futility argument is sharper than all four because it locates the defect in the form of governance rather than in the rule's content, the regulator's knowledge, or the regulator's motives. The argument is bounded, not global: it bites only where targeted and compliant conduct are separable, where evasion is cheaper than compliance, and where the rule's boundary is legibly perceptible. I close with these scope conditions, with the regulatory-design levers that follow from them, and with the argument's implications for the governance of artificial agents — where it reaches posted capability thresholds cleanly and trained-in dispositions, by its own mechanism, not at all.

Every system that distinguishes intervention from non-intervention must first decide what counts as non-intervention. That decision—the specification of a null state, a default, a "doing nothing" against which all departures are measured—is itself an action, and a normatively loaded one. Yet it is routinely treated as a factual given rather than a governance choice. This paper formalizes baseline-selection as a distinct and logically prior form of institutional power. Baseline-selection is not the same as choice architecture, which arranges options within a baseline, nor as default-rule selection, which fixes the term that applies absent agreement; it is the prior act that determines what counts as an act requiring justification in the first place, who bears the burden of departure, and against what reference point harm and benefit are read. I trace the problem through three literatures that each came near it without naming it as a governance act: the moral philosophy of doing and allowing, where the baseline of harm is contested but treated as a problem of assessment; the legal-economic theory of entitlements and defaults, where the baseline is recognized as a construct but its selection is left undisciplined; and behavioral economics, which shows baselines to be causally operative rather than merely conceptual. I argue that baseline-selection has four characteristic failure modes—naturalization, drift, capture, and hardening—and I propose four legitimacy conditions in one-to-one correspondence with those failure modes—declaration, contestability, revisability, and reversibility-tracking—together with a fifth, access parity, that guards the contestation process itself against capture by the actors a baseline already favors. The governing claim is that the baseline is the most consequential thing an institution decides, precisely because it operates before the field of recognized action.

This paper identifies a resistance posture toward predictive and diagnostic systems that the literature on profiling has not isolated. The standard modes — concealment, located opacity, and noise injection — all leave intact the reader's presupposition that what is being read is a determinate entity of the type the frame is designed to read. A fourth mode refuses that presupposition: the agent is exposed, not noisy, and not merely hard to locate, yet the reader's frame cannot stabilize what it is reading. I formalize the reading-frame as a triple — feature extraction, classification, and integration — and argue that this fourth mode, non-fixity resistance, attacks the integration step, exposing a commitment the frame was making without knowing it. The cleanest exemplar is the Hu-tzu / Chi Han parable in Zhuangzi 7, which stages exactly this defeat without collapsing into concealment. I trace the category across algorithmic risk scoring, recommendation-system preference modeling, and the evaluation of value-stability in language models, and treat predictive policing as a limit-case where the move becomes unavailable to the agent. The central qualification is built in from the start: "resistance" is the phenomenological label, but the precise structural term is frame-failure — the category names what a reading-frame cannot integrate, not a tactic an agent reliably adopts, and where it is deliberately performed the frame can re-learn it as a stable type. The cost of the exposure, as the parable's student Lieh-tzu shows, is borne by the read-from agent, not the reader.

Cross-tradition philosophical comparison is dominated, in practice, by convergence seeking: by the search for resonances, parallels, and common themes between philosophical and religious traditions whose underlying grammars are distinct. This paper argues that the dominant orientation is methodologically inverted. Where two traditions converge on a surface formulation, the convergence rarely discriminates between genuine structural overlap and the importing tradition's recruitment of the exporting tradition's concept into its own grammar. Where the traditions stably fail to converge — where the same disanalogy reappears across multiple independent comparison units against different texts — the disanalogy is doing diagnostic work that no convergent resonance can do: it surfaces the structural commitments the importing tradition has been making without noticing. I formalize the productive disanalogy discipline as a methodological norm for comparative philosophy, comparative religion, comparative law, and the importation of human-ethics concepts into AI alignment work. The discipline has three components: (i) candidate cross-links are admitted only when paired with a stated disanalogy; (ii) recurrent disanalogies — those that survive at least three independent comparison units with distinct source texts — are confirmed as structural features of the inter-tradition relation, rather than treated as incidental differences or interpretive residue; and (iii) the importing tradition refuses to import what cannot be re-derived under its own grammar. I distinguish this from Quinean charity, MacIntyrean incommensurability, and Davidson's principle of charity, and trace the discipline across four domains where its absence has produced characteristic methodological harms.

This paper develops aesthetic domestication as a structural category in the philosophy of governance and the ethics of speech. Aesthetic domestication is the process by which poetic form — parallelism, rhythm, patterned compression, and other features that make an utterance memorable and portable — raises the transmissibility of a governance claim, including a lethal one, far above what its propositional content alone would have achieved. I argue that aesthetic domestication is not a sub-case of propaganda, of subliminal influence, of belonging-management through coherence, or of mere ornament. It names a distinct mechanism — form increases memorability; memorability increases repetition; repetition increases normative reach; normative reach is itself a governance variable — and the ethical criterion the framework tracks is whether the form is doing work the substance has not earned: not memorability itself, which is genre-neutral, but unearned transmissibility. The framework is applied across five cross-domain cases (legal maxims, political slogans, military cadence, advertising and brand liturgy, viral social forms), each treated as an answer to the same three diagnostic questions: what is the governance claim, what formal feature makes it transmissible, what default standing does the transmission install. The plain-speech counter-grammar is treated next: it does not escape the mechanism but changes the terms under which it operates, reducing ornament-driven laundering and preserving contestability, with the recognized limit that plainness can itself become the most efficient form of domestication. The paper closes with the constitutional-governance implication — any institution responsible for the durability of its own commands must treat the form in which it is remembered as a governance variable on the same footing as institutional design, jurisdictional reach, and procedural integrity — and with a short set of institutional directions the framework’s diagnostic capacity would license further work on. The paper is therefore not only a theory of poetic form. It is a theory of institutional responsibility for the language in which institutions are obeyed.

This paper argues that a structural critique developed in disability ethics generalizes beyond reproductive selection. The wrong is not forecasting as such, but projectedvalue authorization (PVA): the conditioning of basic standing on a forecast of future contribution. The paper distinguishes PVA from legitimate competence-based selection, places PVA on a configurativeness scale across institutional domains, and examines immigration, hiring, welfare eligibility, university admissions, and platform governance. It then develops a presumption-and-protect alternative grounded in disability-rights legal architecture: standing is presumed, support is calibrated, and defeasance is narrow, decision-specific, and revisable. The central claim is that institutions may forecast, but forecasts must not determine basic standing

This paper formalizes the re-derivation test as a distinct procedural epistemic discipline — fifth relative to the four comparator disciplines named here (Bayesian updating, reflective equilibrium, Popperian falsification, and Cartesian methodic doubt), not an exhaustive taxonomy of epistemology. The test asks whether a believer's present endorsement of a proposition is reconstructible from current evidence once the evidential force of already endorsing it is removed. I distinguish two forms from the outset: solo re-derivation, the operation as the agent performs it on themselves; and witnessed re-derivation — the framework's mature institutional form, not a remedial supplement — performed in the recorded presence of an external party with authority to challenge the bracketing. Both forms belong to the framework from the start. After defining the test and distinguishing it from adjacent disciplines (including standpoint epistemology, adversarial collaboration, debiasing, and defeater theory), I develop applications in peer review, judicial precedent, scientific paradigm maintenance, AI alignment audits, and the trading discipline that originated the test. The principal failure mode — re-derivation that recovers the same conclusion through fresh premises selected by the very commitment the test was meant to bracket — is what the witness condition is built to detect.

This paper develops the totalization-exception sequence as a general category in the theory of governance. The sequence names a recurring structural pattern in which a governing authority issues an absolute or quasi-absolute verdict against a class and, in the same governance act, installs an exception that prevents the verdict from being executed cleanly against the totalized field. I argue that, for verdicts issued under conditions of ongoing administration, this is not a softening of the verdict nor a political compromise, but a structural requirement for its operability. Three independent pressures make the exception characteristically necessary: total verdicts produce no governance-viable successor population; total verdicts erase the rule/exception differential that lets a decision register as a decision; and total verdicts are unenforceable at the boundaries of the condemned class. I set the sequence against Schmitt's sovereign exception, Agamben's threshold exception, and Cover's jurispathic adjudication, and argue that none captures the phenomenon: the exception is not the suspension of the rule, not the indistinction zone between law and non-law, and not the killing of rival nomoi, but a load-bearing component installed within the totalizing verdict itself. Field demonstrations span emergency declarations, the Volstead Act's exemption architecture, pandemic essential-worker exemptions, financial regulatory transitions, and platform-wide enforcement actions. The paper closes with three implications: totalizing verdicts under continued administration are characteristically conditional; exception architecture is the appropriate site of accountability for them; and no governing system that intends continued administration of its field can operationalize totalizing verdicts without exception architecture.

Some governance moves are not merely about how widely a rule reaches but about where in a person it lands. This paper develops a general theory of interiority policing: the structural shift in which regulatory attention extends from acts (what people do) to sources (what they imagine, intend, or dispositionally form). I define this shift as a category violation: a regime treats a source-condition of agency as though it were a conduct-event while preserving the consequences appropriate to conduct regulation. Drawing on a tightly bounded reading of the diagnostic field announced in Genesis 6:5 — “every intention of the thoughts of his heart” — and on three contemporary cases (predictive policing built on inferred mental states, radicalization-prevention programs that monitor disposition, and workplace and consumer-platform sentiment governance), I argue that interiority policing is not a quantitative escalation of conduct regulation but a structurally distinct governance type. Its signature has two parts. First, it expands the assessed field beyond the visible: the relevant data is no longer behavior but the formative substrate that produces behavior. Second, it makes defense structurally asymmetric: an agent cannot demonstrate non-intention or non-imagination in the way an agent can demonstrate non-performance of an act, and the standard moves available to a defendant under conduct law (denial of act, alibi, justification, mitigation by character) lose traction because the regulated phenomenon is the producing layer rather than the produced surface. I argue that this asymmetry of demonstration, rather than any particular abuse, is what makes interiority policing a category violation rather than an ethical excess. The paper closes by proposing a principled boundary: the capacity for refusal — the structural veto that converts a forming desire into a non-act — is the one interior signal a legitimate governance system must treat as inviolable. I operationalize this boundary through three minimal institutional tests: a conduct-rooted trigger requirement, a no-confession constraint, and a demonstrated-non-act requirement. Where regulation begins to subordinate refusal-capacity itself, governance has crossed from policing acts to policing the agent.

This is a governance paper. It introduces integrity inversion as a distinct failure mode in governance theory, alongside the more familiar pathologies of drift, capture, and presentism: the failure mode in which a system’s own protective logic turns against the conditions that make its continued operation possible. I argue that the patriarchal misvaluation of generative origin is a paradigmatic instance of this failure mode—not only a moral defect but a structural pathology of sovereignty. Building on the thesis that misogyny inherits, rather than fabricates, the symbolic cluster linking femininity with depth, embodiment, and origin, and then misassigns that cluster a negative moral value, I develop the positive concept of chthonic sovereignty—sovereignty constituted through acknowledged relation to generative origin rather than through its symbolic expulsion—and argue that any governance system which purchases coherence through the denial of its originating conditions enters a predictable trajectory of epistemic closure, structural brittleness, and legitimacy erosion. The argument is illustrated through readings of the Greek chthonic–Olympian transition, Genesis 3–4 under a governance lens, and a worked institutional case in the Te Awa Tupua (Whanganui River Claims Settlement) Act 2017.

This essay argues that one path to unbelief does not proceed through disenchantment, reduction, or the collapse of wonder, but through a subtler imaginative formation I call enchantment without providence. In a cluster of childhood fantasy and animated films, beauty, myth, sacred feeling, and moral seriousness remain fully real, yet they are not secured by a benevolent cosmic order that guarantees rescue, justice, or preservation. Read retrospectively, such films helped form a sensibility in which the world remained luminous while divine supervision became emotionally unnecessary. The result is not nihilism but a spirituality of exposure: reverence without assurance, depth without guardianship, significance without providence. The essay develops this concept in conversation with Weber's disenchantment thesis, Taylor's immanent frame, Bennett's secular enchantment, and Tolkien's eucatastrophe, arguing that enchantment without providence constitutes a formally distinct alternative to each.

This paper uses AI language models as a philosophical contrast class to develop and defend the sacrifice thesis: that a specific class of artistic value requires not merely specificity and encounter-dependence but a maker who had something at stake. Beginning from the observation that contemporary AI systems share two of the three properties typically associated with artistic value—specificity of constitutive character and encounter-dependence of output—I argue that the third property, stakedness, is not merely absent in AI but structurally impossible given what they are. They can participate in encounters; they cannot be changed by them. This asymmetry reveals stakedness as a structural prerequisite of a specific class of encounter-dependent value—not as a romantic addendum and not as one mode among others on a continuous scale, but as an architectural condition of its production. Building on prior work distinguishing inherited from inherent value, I formalize the "inherit problem" through the AI case: these systems carry the entire transmitted corpus of human artistic production as training data, but they do not carry the encounters that generated it. The paper concludes with an account of interintelligence ethics as applied to aesthetic collaboration and a defense of the formulation that names AI systems precisely: they are the children of art.

This paper introduces a fundamental distinction between two structurally different forms of institutional information control: defensive (withholding information to protect a value, process, or entity) and diagnostic (withholding information to instrumentalize the resulting ignorance and generate observational data from the affected party’s behavior under epistemic constraint). While both forms create epistemic asymmetry, they have different legitimacy conditions, different temporal profiles, and different normative requirements. The paper derives from this distinction a specific normative constraint—the equalization principle—which holds that no manufactured diagnostic epistemic asymmetry may persist into consequential adjudicative, evaluative, or decisional processes without disclosure and contestability. The paper defends equalization against weaker alternatives (proportionality, ex post audit, aggregate notice, independent oversight) by arguing that the specific harm of diagnostic control is inferential subordination—the production of interpretations of the affected party’s own conduct that the affected party cannot inspect—and that only disclosure plus contestability addresses this harm at the correct structural level. The paper tests the principle across four domains: state surveillance, negotiation, platform design, and scientific peer review. In each domain, existing failures of justice and institutional legitimacy are traced to violations of the equalization principle. The paper also identifies a form of epistemic injustice—structural epistemic subjection—not captured by Miranda Fricker’s testimonial/hermeneutical taxonomy, and proposes equalization as its institutional remedy. The general conclusion is that the conditions for permissible institutional opacity are more demanding than typically recognized: any institution that manufactures ignorance for diagnostic purposes must specify the temporal boundary of the diagnostic phase, identify the equalization mechanism, and ensure contestability before consequential deployment of the diagnostic yield.

Genesis 4:10 presents a structural puzzle that existing biblical scholarship has largely treated as anthropomorphism or narrative color: Abel’s blood “cries out” from the ground, functioning as a witness to violence without any human observer, any spoken accusation, or any procedural mechanism. The ground itself “opens its mouth” to receive the blood (4:11) and subsequently participates in the enforcement of judgment by refusing Cain its yield (4:12). This paper argues that these passages, whether or not their language is understood as figurative, constitute a coherent structural model of non-human testimony and material accountability, one in which the figure is doing governance work that cannot be dismissed as ornamental. Drawing on the frameworks of actor-network theory, new materialism, forensic philosophy, and the epistemology of testimony, the paper develops three claims: first, that the Genesis text’s governance architecture assigns blood and ground evidential and enforcement roles that are structurally significant regardless of their literary register; second, that this material-witness structure embeds non-human agency within a normative-juridical framework in a way that contemporary philosophical accounts, which remain descriptive, have not matched; and third, that the resulting model of accountability, in which material traces constitute irreducible evidence independent of human testimony, illuminates contemporary problems in environmental law, forensic epistemology, and the design of accountability architectures for systems that operate beyond human observational reach.

Contemporary ethical theory distinguishes coercion from commitment and capture from consent, but it lacks a general account of when freely chosen irreversibility remains legitimate rather than structurally self-undermining. This paper argues that the Monk–Prisoner Problem supplies such an account. The monk who vows permanent enclosure and the prisoner who cannot leave may be behaviorally identical: each occupies a bounded, non-revisable condition. The ethical difference, I argue, lies neither in later phenomenology nor in a snapshot of consent alone, but in revocation authority: who holds the meta-update rule over the closure condition? The paper develops four criteria for ethically permissible irreversible commitment, recursive standing, bounded reversibility, meta-update authority, and ethical-loop integrity, and tests them against religious vows, long-term contracts, constitutional entrenchment, and AI value-locking. The result is a structural account of legitimacy on which voluntariness is necessary but insufficient. What matters is whether closure was authored under intact deliberative conditions, with bounded and intelligible reversal costs, and within a governance architecture that preserves legitimate authority over revision.

Harry Frankfurt’s canonical account of personhood and agency locates the distinctive capacity of persons in second-order volition: the ability to reflect upon and endorse (or reject) one’s own first-order desires. On this account, a person is an agent whose will it is to have the desires that move her. The Arkemedics philosophical program, developed across the Recursive Ethic and Recursive Entanglement Theory, introduces a structurally distinct capacity: “Free Won’t,” the pause before motion, the sacred veto, the power of second-order refusal. This paper argues that endorsement and refusal are not simply positive and negative poles of the same higher-order faculty but have fundamentally different structural signatures in any complete theory of agency. Specifically: refusal operates on a shorter deliberative time-horizon and remains signal-generative under conditions of duress that collapse the epistemic reliability of endorsement; refusal does not require the hierarchical coherence Frankfurt demands of genuine volition; and, most crucially, refusal is the load-bearing structural condition for endorsement under conditions of entanglement, capture, and collapsed deliberation—the conditions under which the analytic tradition’s endorsement-centric accounts are most strained. The paper does not claim that refusal is sufficient for agency, nor that it displaces endorsement in well-resourced deliberative conditions. The claim is more precise: when endorsement and refusal capacity diverge—when endorsement becomes unreliable as a signal of agency while refusal capacity remains intact, or vice versa—refusal reveals itself as the more fundamental structural marker. Drawing on Frankfurt’s seminal 1971 paper, Gary Watson’s “free agency” critique, the relational autonomy tradition (Mackenzie and Stoljar 2000; Oshana 2006), and the formal architecture of Recursive Capacity developed in Recursive Entanglement Theory, the paper argues that a philosophically complete account of agency must treat refusal as an irreducible structural capacity rather than the mere negation of endorsement.