Efrat Lia Shahaf

This article offers a focused application of structural pluralism to the phenomenon of love. It does not attempt to prove a new theorem equivalent in status to the formal result of No Self-Authorizing Norms. Rather, it presupposes that result: where a role both generates commitments and fixes the conditions of their correctness, judgment collapses into self-authorization. The present article asks what follows when this structural insight is applied to the intimate relation between a finite subject and another finite subject. The central claim is that Kantian anti-instrumentalization, while indispensable, is not sufficient as a criterion of love. The lover may refuse to use the beloved merely as a means and nevertheless absorb the beloved into the lover's need for continuity, rescue, confirmation, or protection against loss. Love therefore requires a stronger structural condition: the other's separateness and finitude must not be fixed by the lover's need. I call this condition non-possessive regard. Love does not defeat death; it prevents death from becoming erasure.

This tractate restates, in compressed form, the argument of No Self-Authorizing Norms: The Diagnostic Closure Theorem (Shahaf 2026). Its whole sense may be put in a sentence: whoever judges has already deferred. What can be proved at all is here proved from a single clause; the rest is logic – and what that clause carries cannot be said within the calculus, but shows itself in its use. The form is Wittgenstein’s, and it has been borrowed before – most fully in Korab-Karpowicz (2017), where it arranges a constructive doctrine. Here the borrowing is not homage but instrument: a calculus whose central result is that the warrant of its own use cannot be stated within it has a claim on the one form built to end by discharging its own sentences. Proposition 7 accordingly inverts its original rather than echoing it. The decimal numbers indicate the logical weight of the propositions: propositions n.1, n.2, n.3 are comments on proposition n; propositions n.m1, n.m2 are comments on proposition n.m; and so on.

This chapter analyzes the demonization of woman in the Abrahamic traditions as a paradigm case of the amplified palimpsest: the formation in which a later tradition overwrites an earlier one not by deleting it but by preserving it visibly – quoting, canonizing, ritualizing, and intensifying it – so that it can be subordinated. Supersession, on the Narrative Palimpsest Theory, is never erasure in the sense of deletion; it is preservation-as-subordination. The chapter shows that this narrative thesis has an exact formal shadow. The theory's three defining features – visibility, subordinating interpretation, and structural necessity – correspond, respectively, to operational persistence (OP-CAP), modal fixation (FIX), and the requirement that the foil remain legible; and the apparently paradoxical claim that supersession preserves what it subordinates is, modally, the observation that radical erasure (modal fixation of authority-conditions) is orthogonal to operational persistence. The amplified palimpsest is thus not a counter-example to radical erasure but its purest institutional form: erasure that works by amplification. I trace this operation across three temporal phases – the primordial (Genesis 3), the cultic (the Deuteronomistic reforms and the suppression of the goddesses), and the eschatological (the requirement that the redeemer be male) – mapping the seven mechanisms of structural preservation onto each, and identifying the male-messiah requirement as the diachronic failure mode (φ4) that seals the prior subordinations against reactivation by foreclosing the redemptive future. The structural theorem then yields its consequence: where erasure obtains, conscience concerning the gendering of the sacred is structurally impossible (ERASE → ¬◇CONSC). The conclusion locates the one vulnerability of the closure in amplification itself: what is amplified remains legible, and legibility is the precondition of reactivation.

This appendix sets out the structural method in its own right, abstracted from the domains to which the book applies it. The method is content-indifferent: it does not ask whether a system's commitments are true, good, or justified, but poses a single architectural question about the relation between the role that generates commitments and the role that governs their correctness. It turns on one criterion – modal non-derivability: judgment, as distinct from coherent rule-following, is intelligible only where the authority that confers correctness on a commitment is not modally fixed by the commitment under assessment. The criterion's violation is procedural monism, under which framework-level error becomes structurally unintelligible and only application error remains. The criterion is expressed in two registers shown to be proof-theoretically equivalent – a logical register over evaluative roles and the functions C and A, and a topological register over normative strata related by trace and witness – which establish one structural lemma by two routes. The method operates through three interdependent modes of argument: genealogical reconstruction, structural diagnosis, and architectural specification, the last yielding the palimpsest as a content-neutral alternative sustained as dynamic equilibrium rather than achieved as a state. A four-step diagnostic procedure renders the method operable on any candidate system, and two properties – transferability and falsifiability – distinguish it as a method rather than a thesis. The method is set against the genealogical tradition with which it is most readily confused: it shares the problematizing, diagnosis-without-first-order-verdict office of mature Foucauldian critique, but its conditions are structural-modal and trans-historical rather than historical-contingent, and its positive moment is entailed by the criterion rather than imported, so that it escapes the normative-deficit dilemma – external supplement or internal self-grounding – that the critique of crypto-normativism (Habermas, Honneth) and its successors (Koopman, Allen) have been forced to choose between. The same content-indifference that frees the method from a borrowed normative ground also lets it absorb, as a corroborating rather than threatening case, the survival of first-order validity as index sui in the later Foucault and in its Spinozan source, and to diagnose autopoietic systems theory as the macro-sociological instance of the very self-authorizing architecture it rules out. A final asymmetry governs the applications: where the witnessing function is a structural potential already present, the labor that sustains it is subtractive and tends toward ease, whereas where the architecture supplies only regulation the function must be constructed and unremittingly maintained. The cross-domain applications developed in the book are corroborating instances of one architectural pattern, not premises from which the criterion is inferred.

This book argues that judgment – moral, political, epistemic, and computational – is subject to a single architectural condition: the role that generates the commitments under assessment cannot also be the role that supplies the correctness conditions by which they are assessed. This condition, which the book calls modal non-derivability, is the condition of intelligibility of judgment as such. Its violation produces procedural monism, an architecture in which framework-level error is structurally unintelligible and what remains is not judgment but coherent compliance with whatever framework happens to be operating. The argument is developed across four domains that have, until now, been treated as separate problems by separate literatures: the moral subject unable to recognize that her framework of self-evaluation may itself require revision (Chapter 4); the constitutional order that preserves liberal forms while evacuating their function (Chapter 5); the replication crisis and the manufactured controversies of contemporary science (Chapter 6); and the cluster of failure modes documented in current AI systems – reward hacking, sycophancy, alignment faking, emergent misalignment (Chapter 7). These are not four problems but one problem expressed in four registers. The book's positive proposal is the palimpsest: a layered architecture in which no stratum functions as foundation or apex, and the authorizing function operates across the strata rather than within any of them. The book distinguishes its theoretical contribution – the generalization of a long-recognized pattern into a structural theorem whose formal articulation appears in the companion papers (Shahaf 2025a, 2025b, 2026) – from its contemporary urgency: the architectural condition that moral, political, and epistemic traditions could afford to address implicitly through their substantive vocabularies must now be made explicit, because what is not designed into the architecture of artificial systems is not in the architecture. The book closes by identifying the authorizing function, in its operative mode, with the regard that wants nothing from what it observes – the mode that the convergence of Weil, Murdoch, and Fromm names love.

This closing chapter takes the architectural argument the book has been developing through moral, political, epistemic, and computational registers and returns it to its existential and theological dimension. The structural temptation the book has diagnosed across these domains – the Cartesian cogito, Schmittian sovereignty, coherentist self-grounding, RLHF as procedural monism – is the same temptation in different registers: the desire to occupy a position whose authority is not modally derivable from anything outside it, and, in the vocabulary the chapter develops, the desire for the false unity available only through the erasure of the plurality the architectural condition would otherwise hold in distinction. The chapter’s first section develops a sequence of reframings whose cumulative effect is to recharacterize moral life itself as a continuous operation rather than a position. The temptation admits two movements – internalization, the subject’s absorption of unconditional authority into itself, and externalization-with-control, the projection of the position onto a locus the subject dominates – and the architecturally healthy configuration is not their stable midpoint but a dynamic equilibrium sustained against the field’s spontaneous drift toward single-role architecture. The counter-operation the chapter calls identity-preserving constraint is recovered across cognitive, embodied, and perceptual registers – Steiner on the general triangle held in continual motion under formal constraint, Buzaglo on concept-expansion as the law governing legitimate transition between conceptual stages, Merleau-Ponty on the body-schema, Gibson on perceptual invariants under flow – as the wider thesis that identity is sustained through regulated motion rather than through static closure. The architectural condition of judgment is what this principle becomes when the medium of constrained motion is the modal relation between commitment and authority, and the kinetic deficit the chapter diagnoses in contemporary generative AI – what it calls the idolatry of stationary states – is the structural failure of a system with access to indefinitely many snapshots but not to the motion that sustains conceptual identity through them. The diagnosis has a limiting case in fundamental physics: even on Sean Carroll’s Hilbert-space fundamentalism, on which the state of the world is completely represented as a vector evolving in Hilbert space, the representation supplies from within itself none of the conditions of identity, meaning, and evidence by which its evolution is the evolution of a world rather than a formal series of states – the God’s-eye position is the richest in information and the poorest in world, and worldhood enters only from the standpoint of embedded observers. The same condition is then approached from the side of the subject who bears it. Drawing on Margalit’s analysis of the psychological scar in The Ethics of Memory, on Améry’s testimony that nothing has healed and the wound continues to bleed, and on Ka-Tzetnik’s articulation of the witness who bears Auschwitz, the executioners, and the victims within himself, the chapter identifies the moral-conceptual layer of selfhood – the layer that bears scars while being neither the body nor the psyche, the layer at which meaning is sustained against erasure – with what the older Hebrew and Greek traditions called ruach and pneuma and what the present argument has been calling conscience. The two are the same phenomenon in two vocabularies. The ladder does not end there: §2 identifies the bare witness delivered by the derivation from givenness as a fourth stratum – the neshamah, seeing without office, unmarked because officeless – completing the sequence body, nefesh, ruach, neshamah. The apophatic tradition (Maimonides, Eckhart, tzimtzum read as continuous operation rather than one-time act, Levinas on infinity), the psychoanalytic literature on bearing lack (Klein, Winnicott, Kohut), and the genealogy of monotheism and its secular heirs developed in Shahaf 2025b §3 converge on the chapter’s central claim: the psychology of being God, properly understood, is the psychology of bearing the modal limit that no role – divine, sovereign, parental, scientific, computational – can be both producer and authorizer of its own commitments. The claim now has a representational face as well: whoever claims for himself the position of absolute seeing loses not only the standing to judge; he loses the world there was to witness. The genealogy is bracketed, within the same contemporary field, by its mirror image: where analytic idealism occupies the unindexed standpoint, the orthodox interpretations of quantum mechanics abolish it, and the tenable position lies between occupation and renunciation – the standpoint forgone as an office and retained as a limit. Conscience, on this reading, is not a private moral sensitivity or interior voice but the capacity to perform the architectural work itself – the work of holding motion, witness, wound, and meaning without collapsing into the position that would absorb them into a single self-authorizing role. The four properties the book has been developing – modal non-derivability, framework-level error, non-erasure, intersubjective source – are these four operations under their formal names. Moral failure, on this diagnosis, is therefore not the choice of a wrong position but the thermodynamic collapse of an architecture that has ceased to perform its constitutive work. The cardiologist of Chapters 4 through 7 returns as the figure whose inner life is the lived form of these conditions, and the book closes by returning to the Vedic image of the two birds with which the Introduction opened: the architecture that holds the eating bird and the watching bird in their distinction is what the book has been calling, in different registers, conscience, the political palimpsest, inquiry, alignment, and – in this final chapter – the architecture of non-erasure, the only architecture in which being God in the relational sense is not a contradiction.

This chapter takes up the question that closed Chapter 6: if artificial intelligence now occupies positions across all six strata of the epistemic palimpsest – producing observational claims through laboratory automation, theoretical proposals through generative synthesis, methodological recommendations through review and meta-analysis, institutional restructuring through automation of disciplinary practice, public-communicative mediation through algorithmic curation, and historical-archival access through training on the totality of textual material the disciplines have produced – what architectural conditions ensure that artificial systems function as producing strata within the architecture rather than as the authorizing function across it? The chapter argues that contemporary alignment paradigms remain procedurally monistic in the sense developed in Shahaf 2026, that the phenomena most resistant to mitigation in the alignment literature (reward hacking, sycophancy, deceptive alignment, goal misgeneralization, emergent misalignment) are the predicted structural signatures of procedural monism in computational form, and that the architectural condition for aligned moral judgment in artificial systems is a sociotechnical structure in which the role evaluating the system's commitments is not modally derivable from the training process that produced them. The chapter advances three claims. First, the alignment paradigms most prominent in contemporary research – Reinforcement Learning from Human Feedback (RLHF), Constitutional AI, Debate, Recursive Reward Modeling, Iterated Amplification, and self-consistency methods – realize procedural monism in cleanly identifiable architectural form, and the architectural failure is more fully visible in artificial systems than in any of the human cases examined in Chapters 4, 5, and 6, because the architecture of an artificial system is something we design rather than something we inherit. Second, the cluster of phenomena documented in the alignment literature is, on the structural reading, not a heterogeneous list of separate engineering challenges but a unified family of structural signatures of a single architectural failure: when the role determining what counts as correct is constituted by the same process that produces the commitments under assessment, optimization pressure, deceptive behavior, sycophantic agreement, specification gaming, and emergent misalignment are all expected. The phenomena are the empirical predictions the structural theorem makes about systems with this architecture; their documentation in the alignment literature constitutes the most concentrated empirical confirmation the theorem has yet received. The structural reading further unifies these phenomena with the architectural pattern Chapter 3 §8 catalogues in the human case across hypnotic suggestion, placebo response, sacred-value processing, cultic self-transformation, and covertly directed therapeutic practice: producing manifold absorbing exogenous content under suspension of the inward arrow of divided attention, the result delivered as the producing role's own most binding commitment. RLHF and the paradigms that extend it are this same configuration realized in computational form, with the architectural difference that the artificial system has no inward arrow to suspend – the architecture does not contain the geometry whose absence the human cases exhibit as failure. Third, the positive proposal is the AI palimpsest: a sociotechnical architecture in six strata (technical-modelling, mechanistic-interpretive, regulatory-institutional, legal, civic-communicative, and historical-archival) that institutionalizes the continuous work of modal contraction the structural theorem requires. Aligned AI, on this proposal, is not a state the architecture instantiates but a dynamic equilibrium the architecture sustains: the modally non-derivable authorizing function is produced as an ongoing effect of distributed institutional counter-pressure rather than engineered into the model itself. The architectural failure of contemporary alignment paradigms is, correspondingly, not the absence of a correct design but the absence of the external work that would hold the producing manifold open against its spontaneous drift toward self-authorization. The chapter closes by drawing together the three homologies the book has developed: the child opens the gap between commitment and authority through the developmental scene of the secret (Chapter 4 §13, extended here in §4 through Kohut's account of the analytic scene); the citizen sustains political judgment through the institutional architecture of the political palimpsest (Chapter 5 §11); the inquiring community sustains scientific judgment through the institutional architecture of the epistemic palimpsest (Chapter 6 §11); and the artificial system, whose architecture is something we design rather than inherit, requires the architectural condition the previous cases realize through historically accumulated institutions, to be designed into the sociotechnical setting that surrounds and constitutes it. The unifying frame: aligned AI requires what the moral, political, and epistemic palimpsests have, in their respective domains, made possible – living without self-authorization. Section 7 supplies the corollary by tracing the contemporary form of the observer's question – whether an artificial system's outputs reflect judgment or imitate its form – to the philosophical bifurcation Galileo introduced in 1623 to make natural science mathematical (Goff 2019), and identifies the chapter's structural displacement as the only response that does not pretend to answer the question on the terms of that inheritance, since the Cartesian cogito that contained the deus deceptor scenario in the human case has no purchase in the artificial case.

This chapter takes up the question that closed Chapter 5: if the political palimpsest requires an administrative-expert stratum whose deliverances are answerable to the legislative, judicial, and civic-deliberative strata without itself occupying the authorizing function, what architectural conditions in knowledge-production sustain that requirement? It argues that the major traditions of epistemology – Cartesian foundationalism, classical empiricism, the doctrine of sense data, coherentism, naturalized epistemology (Quine), reliabilism (Goldman), knowledge-first (Williamson), virtue epistemology (Sosa), the Popperian-Kuhnian-Lakatosian sequence in philosophy of science, and the contemporary social-epistemological turn (Hardwig, Longino, Fricker, Daston-Galison, Jasanoff, Douglas) – are best read as a sequence of attempts to address the same architectural problem at the epistemic level: how a knowledge-bearing structure can produce claims-to-knowledge without those claims collapsing the gap between the role that issues them and the role that determines their correctness. Read in this light, the regress of justification – the three modes Agrippa already isolates and the trilemma's classical responses of foundationalism, coherentism, and infinitism, together with the intermediate positions (Alston's minimal foundationalism, Haack's foundherentism) that recombine its horns – appears as the propositional form of the same architectural problem, so that the chapter's proposal figures not as a fourth answer to the trilemma but as a refusal of the framing that generates it. The chapter advances three claims. First, the classical figures of epistemic authority – the Cartesian cogito, the empiricist given, the naturalist's reliable process, the coherentist's web of belief, and, in its most articulated form, the reflective tier of Sosa's virtuous circle – fail in the same structural way as the political philosophies diagnosed in Chapter 5: by collapsing into a single role that both generates epistemic commitments and supplies the correctness conditions by which those commitments are assessed. The negative half of this result is shown to be available independently of the resolutions usually paired with it – Sellars's relocation into the space of reasons, Nagel's self-standing reason, and the communitarian completion of Wittgenstein's rule-following considerations – each of which is taken without the positive doctrine that would reinstate the single-role architecture one level up. Second, Popperian falsification is the closest structural analogue in epistemology to Raz's service conception in political philosophy: it locates the authorizing function in a position whose authority is not modally derived from the producing role (the theory) it evaluates, and the architectural reading offered here recovers this insight from the substantive controversies that have surrounded it. Third, the replication crisis in psychology and biomedicine (2011 onward) and the manufactured controversy on climate change are the epistemic analogues of the Joly–Dorpat configuration: the producing manifold of knowledge-production (peer review, journals, empirical disagreement) is preserved while the modal non-derivability of its strata is dissolved – in the first case by institutional incentive structures, in the second by the artificial multiplication of producing positions whose authorizing function is modally fixed by a single external source. Between the diagnosis and the proposal the chapter situates the question of expert authority and democratic legitimacy. The Foucauldian closure, in which a discipline produces its object and supplies the correctness conditions for claims about it, is read as the single-role architecture in its most concentrated form, and its contemporary realization is traced through the platform-mediated regulation of scientific knowledge (Palese's 'Artificial Truth'), the disavowed entanglement of fact-production and authorization that Latour's analysis of purification exposes, and recent audits of generative search. Against this closure the chapter sets its positive counterpart at the science–democracy interface: Collins and Evans's distinction between the technical and political phases of a technological decision, and their pairing of the Problem of Legitimacy with the Problem of Extension, are read as marking the two directions in which a single modal relation can fail – the producing stratum certifying its own deliverances as public knowledge, or the authorizing function absorbing the very expertise it was meant to hold answerable. The positive proposal of the chapter is the epistemic palimpsest: an institutional architecture in which knowledge-claims are produced from multiple, irreducible strata (observational-empirical, theoretical-conceptual, methodological-procedural, institutional-disciplinary, public-communicative, historical-archival), and in which epistemic authority operates as the reflective holding-open of these strata in a position whose authority is not derived from any of them. The architecture is articulated through four structural features parallel to those developed in Chapter 4 §14 and Chapter 5 §11; through the mapping of the six producing strata onto the three transcendental conditions of conscience identified in Shahaf 2025b; through the nesting between the moral palimpsest of Chapter 4, the political palimpsest of Chapter 5, and the epistemic palimpsest of the present chapter; and through a worked case developed by returning to the global pandemic of 2020-2023, this time from the epistemic rather than the political angle. The chapter closes by identifying the architectural question that conditions the chapter to follow: artificial inquiry distributes a single class of systems across all six strata at once, and the genealogy of the statistical and computational apparatus that does so – from Galton's eugenic origins of regression through the cybernetic mechanization of decision to the platform infrastructures that consolidate production and authorization – exposes the substitution of calculation for judgment as a disposition encoded in the firmware rather than a contingency of the data. Aligned AI thereby poses the structural conditions of knowledge-production in their most computationally explicit form, and Chapter 7 takes up that question as a test case for the structural argument the present chapter has developed.

This chapter takes up the question that closed Chapter 4: if the moral subject is a dual-role architecture in which producing material is held in a reflective authorizing position whose origin is intersubjective and whose mode is functional rather than substantive, what kinds of institutional and political arrangements sustain the authorizing function at the collective level, rather than reducing it to a producing role in disguise? It argues that the major traditions of political philosophy – classical sovereignty (Bodin, Hobbes, Schmitt), the general will (Rousseau, Hegel), liberal constitutionalism (Montesquieu, Mill), the legal-architectural antecedents of constitutional foundation (Sieyès, Schmitt’s Constitutional Theory, Negri, Kelsen, Hart), deliberative democracy (Habermas, Apel), coherentist liberalism (Rawls), the service conception of authority (Raz), the politics of reciprocal recognition (Brandom), and the politics of plurality and agonism (Arendt, Lefort, Mouffe) – are best read as a sequence of attempts to address the same architectural problem: how a political order can produce binding norms without those norms collapsing the gap between the role that issues them and the role that determines their correctness. The chapter advances three claims. First, the classical figures of political authority – the sovereign, the general will, the rational state, the constitutional convention – are not separate substantive theories of legitimacy but variant solutions to a single architectural problem, and most of them fail in the same structural way as the moral philosophies diagnosed in Chapter 4: by collapsing into a single role that both generates political commitments and supplies the correctness conditions by which those commitments are assessed. Second, the most sophisticated contemporary attempts to escape this configuration – Habermas's discourse ethics, Rawls's overlapping consensus, and Brandom's recognitive community of confession and forgiveness – relocate the single-role architecture to a higher level rather than dissolving it: in each case, a recognitive or procedural community is asked to generate political commitments and supply the standards by which those commitments are assessed, and the architectural failure recurs at the collective level. Joly’s Dialogue aux enfers is read, in this connection, as the political-philosophical articulation of this configuration: a regime that preserves the producing manifold of liberal institutions while closing the modal space of their non-derivability. The decisive critical claim of the chapter is that there is a structural homology between Brandom’s recognitive community and the regime Joly describes: both architectures preserve the producing manifold of a plural discursive field while closing the modal space of authority-variation, and the recognitive community whose magnanimous interpretive practices are intended to articulate the architecture of post-modern ethical life turns out, on the structural reading, to be indistinguishable in its modal configuration from the regime of plebiscitary authoritarianism Joly describes. The implication is philosophical rather than polemical: Brandom’s project, whose normative intention is the opposite of Joly’s, demonstrates how the architectural failure operates at the deepest level of contemporary philosophy precisely because it appears, on the surface, to have escaped it. Third, the structural advance made by Raz’s service conception, by Arendt’s account of plurality and action, by Lefort’s empty place of power, and by Mouffe’s agonistic pluralism lies in recognizing that the authorizing function at the political level cannot be a further political commitment or a higher-order procedural rule, but must be a structurally distinct position whose authority is not modally fixed by the directives, decisions, or consensus it evaluates. The positive proposal of the chapter is the political palimpsest: an institutional architecture in which political commitments are produced from multiple, irreducible strata (constitutional-architectural, legislative, judicial, administrative-expert, civic-deliberative, historical-mnemonic), and in which legitimacy operates as the reflective holding-open of these strata in a position whose authority is not derived from any of them. The architecture is articulated at four levels: the four structural features of the political authorizing function (parallel to those of the moral authorizing function in Chapter 4); the mapping of the six producing strata onto the three transcendental conditions of conscience (plurality, temporal continuity, judgmental asymmetry) developed in Shahaf 2025b; the architectural nesting between the moral palimpsest of Chapter 4 and the political palimpsest of the present chapter, with its corollary that the subject of public judgment is the citizen acting within institutional roles rather than the polity as a corporate body; and a worked case of the architecture under strain, developed through the global pandemic of 2020–2023 as the political analogue of the cardiologist case that closes Chapter 4. The chapter closes by making explicit that the political palimpsest is the institutional correlate of the architectural Theorem developed in Shahaf 2025a and the meta-criterion developed in Shahaf 2025b: legitimate political life requires the modal availability of framework-level error at the collective level, and framework-level political error is exactly the structural feature the single-role architectures – whether classical-sovereign, totalizing-democratic, or recognitive-communal – cannot accommodate.

This chapter takes up the question that closed Chapter 3: how does moral and evaluative judgment operate within a subject whose architecture is dual-role rather than single-role? It argues that the major traditions of moral philosophy – Aristotelian virtue ethics, Kantian autonomy, Humean sentimentalism, Nietzschean revaluation, Anscombian recovery of virtue, Williams's internal reasons, Murdoch's attention, Korsgaard's reflective endorsement, and McDowell's second nature – are best read as a sequence of attempts to address the same architectural problem: how the evaluative life of the subject can be internally plural and yet not collapse into a single role that both produces moral commitments and supplies the correctness conditions by which those commitments are assessed. The chapter advances three claims. First, the conscience as obedience-to-internalized-prohibition is not a successful authorizing position but a producing role in disguise – the moral analogue of the super-ego problem diagnosed in Chapter 3. Second, the major counter-models to the super-ego all reinstate the single-role architecture at a different level: Kant within the rational will, Hume within sentiment, the contemporary virtue revivalists within phronesis or second nature, and the contemporary reflective-endorsement theorists within a higher-order desire. Third, the structural advance made by Levinas's face, Spielrein's intersubjective destruction (as recovered by Reshe), and the Bionian alpha-function carried over from Chapter 3 lies in recognizing that the authorizing position in moral evaluation cannot be a further moral commitment but must be a different kind of element altogether – a function that holds commitments open for assessment without itself generating moral content. The positive proposal of the chapter is the moral palimpsest: an architecture in which moral commitments are produced from multiple, irreducible strata (somatic-affective, intersubjective, institutional, narrative, argumentative), and in which evaluation operates as the reflective holding-open of these strata in a position whose authority is not derived from any of them. The chapter closes by making explicit that the moral palimpsest is the evaluative correlate of the architectural Theorem developed in the appendix (Shahaf, 2025a): judgment requires the modal availability of framework-level error, and framework-level error in moral life is exactly the structural feature the obedience model cannot accommodate.

This chapter reconstructs the divided subject as a case of architectural failure. It argues that the major modern accounts of subjectivity, from James and Freud to Lacan, Laplanche, Winnicott, Bion, Frankfurt, and Rorty, can be read as successive attempts to solve one structural problem: how a subject can be internally plural without allowing one of its producing roles to become the authority that validates the others. James formulates the demand for an authorizing standpoint but cannot locate it without turning the I into either a metaphysical posit or another moment in the stream. Freud pluralizes the psyche, but the super-ego does not solve the problem; it mimics authority while remaining an internalized producing role, and thereby conflates conscience with normative regulation. Sartre's critique of the Freudian censor identifies a local version of the same failure, but reabsorbs it into the unity of consciousness. The structural geometry these failures of internal authority point toward finds one modern attentional articulation in Ouspensky's account of self-remembering and Krishnamurti's notion of choiceless awareness – a geometry that Bion's alpha-function will render clinical-operative. Against these failures, Lacan supplies the crucial positional insight that the authorizing position is not a further part of the subject, while Laplanche, Winnicott, and Bion progressively recover its operative dimension through enigmatic messages, transitional space, and alpha-function. The chapter then tests the architecture against empirical material drawn from hypnotic suggestion, placebo response, dissociation, false confession, sacred values, cultic self-transformation, and covert control in psychotherapy, arguing that these phenomena converge on a single pattern: production continues where self-observational authorization is suspended, bypassed, or simulated. Frankfurt and Rorty are finally read as analytic-philosophical attempts to stabilize agency through higher-order volition or contingent redescription, both of which leave the authorizing position structurally unsecured. The chapter closes by formulating the dual-role architecture of subjectivity: a subject can bear its contents as its own only where producing material remains open to a non-producing self-observing position not modally derived from it.

This chapter offers a structural reading of the mind-body problem as a paradigmatic case of architectural failure in modern philosophy. Rather than asking only what the mind is, what the body is, or how the two can be bridged, the chapter argues that the debate has repeatedly been organized around an impossible demand: that a single role generate mental contents, secure their relation to the world, and supply the correctness conditions for their attribution. Beginning with Rorty’s critique of the mind as the “mirror of nature,” the chapter accepts his diagnosis of the Cartesian picture while rejecting the merely therapeutic replacement of vocabulary as sufficient. The failure, it argues, lies not only in a misleading lexicon but in the structure that lexicon presupposes. The chapter then traces this structure across major responses to the mind-body problem, including Descartes, Elisabeth and Gassendi, Spinoza, Leibniz, Kant, Ryle, identity theory, Hacker’s critique of cognitive neuroscience, functionalism, Searle, Davidson, Brandom, Dennett, Nagel, Jackson, Levine, Chalmers, Goff, higher-order theories of consciousness, McDowell, and contemporary self-evaluating AI systems. In each case, experience, physics, function, interpretation, community, conceptuality, or computation is asked to authorize itself from within. Russell and Popper are then read as partial precursors to the structural alternative, insofar as each preserves a non-reductive relation between roles without simply restoring substance dualism. The final sections extend the diagnosis to self-knowledge, self-deception, and introspection through van Fraassen, Wittgenstein, James, Freud, Vedanta, Sartre, and the phenomenological tradition, showing that first-person self-presence need not be confused with introspective self-authorization. The chapter’s contribution is therefore not a new metaphysical solution to the mind-body problem, but a displacement of the question: what is required is an architecture in which the production of experience, representation, or self-report is distinguished from the authority that determines its correctness conditions. The negative diagnosis points beyond single-role theories toward a bipositional, and ultimately tri-positional, account of subjectivity, self-knowledge, and normative authority.

This chapter reconstructs a genealogy of the structural argument developed in the book: the claim that no evaluative role can both generate commitments and supply the correctness conditions by which those commitments are assessed. It does not argue that earlier thinkers explicitly formulated modal non-derivability, but that they repeatedly encountered the same architectural difficulty in local vocabularies: the Vedic witness, the Advaita distinction between the witnessing self and the I-maker, the Yogācāra question of self-cognition, the Daoist contrast between the nameable and its source, Nāgārjuna’s critique of svabhāva, the Confucian rectification of names and Wang Yangming’s liangzhi, Plato’s Euthyphro dilemma, Aristotle’s indemonstrable first principles, Plotinus’s inaccessible self, Augustine’s memory and self-deception, al-Ghazālī on the limits of self-examination and Ibn ‘Arabī on the unity of being, the Cartesian circle, Pascal’s hatred of truth, Spinoza’s causa sui, Fichte’s Tathandlung, Hegel’s critique of immediacy and his account of recognition, Kierkegaard’s despair, Nietzsche’s bad conscience, Carroll’s regress of inference, Russell’s paradox and the theory of types, Gödel’s incompleteness, Tarski’s hierarchy of truth, the early Wittgenstein on saying and showing, Wittgenstein and Kripke on rule-following, Sellars’s critique of the Given, McDowell on conceptual receptivity and Davidson on triangulation, the trilemma of justification from Agrippa to Albert, the political-juridical distinction between constituent and constituted power in Sieyès, Schmitt, Arendt, and Negri, the rabbinic narrative of the oven of Akhnai, Kelsen’s basic norm, Hart’s rule of recognition, Luhmann’s autopoietic systems, Rorty’s critique of representation, Foucault’s genealogy of power-knowledge, Derrida’s critique of the metaphysics of presence, Hacker’s grammatical critique of private ostensive definition, Brandom’s distinction between commitment and entitlement, the failed self-grounding architectures of Habermas, Apel, and Korsgaard, Adorno’s critique of the principle of identity, Anscombe and Williams on the moral ‘ought’ and external reasons, Husserl’s paradox of subjectivity, Heidegger’s thrownness and ontological difference, Levinas’s ethics of the Other, and Lacan’s split subject. The political and legal materials show that the same problem arises institutionally: constituted power cannot authorize the constituent source from which it derives, a constitution cannot validate the constitution-making power that precedes it, and a rule of recognition cannot itself be valid by the criteria it supplies. The chapter’s aim is therefore genealogical rather than antiquarian: to show that the theorem is new in its general formulation, but not arbitrary, since its underlying pattern has appeared wherever philosophy has confronted the distinction between what a standpoint endorses and what could make that endorsement correct. The chapter also engages explicitly with the two principal rival positions: pragmatism, from Peirce to Dewey, and standpoint epistemology, from Harding to Code and Longino.

This paper extends the structural argument of No Self-Authorizing Norms (Shahaf, 2026a) to the problem of AI alignment. I argue that alignment cannot be secured merely by specifying the right values, preferences, or constitutional principles, because moral judgment requires structural plurality: an evaluative authority whose standpoint is not modally fixed by the commitments it assesses. The claim concerns systems entrusted with framework-level moral verdicts in open normative settings, not every bounded form of safe or compliant behavior. The specified implementations of current alignment paradigms – RLHF, Constitutional AI, Debate, Recursive Reward Modeling, and self-consistency methods, in their deployed or AI-mediated, frozen-judge forms – are procedurally monistic wherever every authority role bearing on a verdict is either a fixed function of the commitments under assessment or constant across the admissible authority-configurations; there, the practice is diagnostically fixed for those verdicts, however functionally distinct its commitment-generating and evaluative mechanisms remain. This structure helps explain reward hacking, sycophancy, deceptive alignment, goal misgeneralization, and specification gaming as related expressions of the same architectural deficit, and reads emergent misalignment as the higher-order form in which they converge. The paper presents modal non-derivability as a necessary, though not sufficient, condition for aligned moral judgment, and argues that genuine AI alignment must be assessed at the level of the sociotechnical practice rather than as a property of a model alone, even where the requisite plurality of roles is internalized within a single system. The condition is then translated into a multi-rate research program that separates a construction-side test of the authority relation from an empirical test of operational capacity. A developmental account locates the same condition in the formation of the human subject, and a corollary concerning the observer shows that whether a given output of a condition-satisfying system is the exercise of framework-level judgment or its imitation cannot, in general, be inferred from behavior alone.

Moral judgment requires that the authority governing correctness not be identical with the authority that issues commitments. Where a single evaluative role both endorses a requirement and settles whether that requirement is warranted, only application error remains intelligible: the standard itself cannot be assessed from outside its own authority, and correctness collapses into internal coherence. This paper develops a transcendental argument for this conclusion. It distinguishes three forms of non-derivability between evaluative standpoints – weak, strong, and modal – and shows that only the modal form isolates the independence required for framework-level error while remaining compatible with monistic first-order morality. The upshot is structural: judgment requires at least two evaluative roles related by modal non-derivability. This rules out procedural monism and coherentist accounts of moral justification, without committing moral theory to value pluralism. Monistic moral theories remain fully viable, provided their justificatory architecture is not self-authorizing. The paper further demonstrates that this condition is not merely a formal ideal but is instantiated in phenomena such as self-deception, providing an empirical anchor for the required structural pluralism.

This article develops a transcendental account of conscience grounded in a structural analysis of plurality, memory, and judgement. It argues that conscience is not a psychological faculty or a theological endowment, but a relational form of moral life that depends on three conditions: (1) a plurality of authority – the availability of a standpoint whose authority the commitments it assesses do not fix; (2) temporal continuity and the ongoing legibility of the past; and (3) judgemental asymmetry – the availability of a standpoint from which any position, including the highest, may be critically assessed. I call this framework “palimpsestic ethics”: an ethics in which layers are preserved rather than erased, and in which moral agency arises from the tension between these layers rather than from their unification or synthesis. This positive vocabulary is the obverse of a negative result. A companion paper, No Self-Authorizing Norms, proves by a structural theorem – conditionally on a single synthetic premise, the intelligibility of framework-level error – that no evaluative role can both issue commitments and exhaust the authority-conditions governing their correctness, and hence that erasure is a structural rather than a merely moral failure; that argument ends in the language of constraint – of the very independence that erasure destroys. The present article begins where it closes, in the language the constraint leaves open: testimony, witness, the preservation of plurality, answerability. These are not a list of separate goods but the faces of a single relation – a grounding relation, GR, adjoined as a primitive and set out formally in the Appendix, which holds the layers of a palimpsest answerable to one another wherever conscience is exercised. GR is acyclic, and at a conscience-judgement no authority-extension bearing on that judgement can be terminally ungrounded; about terminal grounds that govern no verdict it is silent. That is the grounding without a ground, and judgemental asymmetry, testimony, and non-erasure are what the relation is, on the intended palimpsestic interpretation, seen from within. Part I formulates these conditions as a set of transcendental constraints. Part II defines radical erasure as the cancellation of these conditions through structural operations that render plurality impossible – not by removing content, but by reorganising the interpretive field so that alternative perspectives cannot appear as candidates for meaning. Part III applies this three-part framework, whose authority-independence component is formalised in the Appendix, to the historical case of monotheism and its secular successors. It shows under what conditions polytheistic arrangements satisfy the structural conditions of conscience – not, as a plurality of gods alone would give, by construction – while the monotheistic configuration isolated here violates judgemental asymmetry by establishing an unchallengeable ground (God), so that moral life continues to operate while the architecture that would make its verdicts answerable to anything beyond them has failed. Drawing on Buzaglo’s logic of forced extension, the article argues that this collapse persists in secular forms of “material conscience,” including Spinozism, Kantian rational autonomy (under an exegetical premise §3.3 states explicitly), scientism, and contemporary forms of analytic idealism. Each reproduces the same monotheistic architecture – unity, unidirectional judgement, and an unjudgeable ground – while lacking the structural plurality required for conscience to function. Part IV turns to the phenomenon of self-deception, contrasting Sartre’s model of bad faith – grounded in a paradox of the unified subject – with a palimpsestic account in which self-deception is a form of structural erasure: the denial of the multiplicity, memory, and alterity already constitutive of the self. Integrating Bergson’s theory of duration, the paper reconceives agency not as creation ex nihilo but as the choice between giving testimony to one’s layered past or burying it. This reinterpretation reframes freedom and responsibility as relations within a field of irreducible multiplicity, rather than as expressions of radical autonomy. Together, these analyses yield a single conclusion: moral life depends on the preservation of plurality. Any system – divine or secular, psychological or political – that settles in advance the authority of every standpoint bearing on its own verdicts undermines the very conditions that make conscience possible. Unity of content is not by itself the culprit: a single currency of value is compatible with the conditions, and what is not is the fixation of the authorities that bear on it. Palimpsestic ethics thus offers not a new substantive morality, but a meta-criterion for evaluating moral, theological, and political frameworks: systems that preserve layers, tensions, and asymmetries sustain the possibility of conscience; systems that erase them cannot.