Memo 04 — Religion as a Source of Norms

Series: The Landscape of Normative Systems (LNS) Memo number: 04 of 12 Primary JD questions: Q1 (sources of normative authority), Q2 (six-part analysis per source), Q3 (the norm-generated-by-revelation behavior class). Secondary: Q8 (what must stay above the kernel). Status of D0: Immutable input. This memo neither modifies, criticizes, nor defends D0. It maps structure above the kernel.

Scope note. This memo treats religion strictly as a norm-generation mechanism and analyzes it in the register of comparative religious studies and the sociology/anthropology of law — not theology. It does not adjudicate any truth claim, does not endorse or critique any faith, and does not rank traditions. Its object is structure: how a class of normative systems takes an authoritative, typically non-revisable source (a "revelation" — text or event) as an immutable root, layers a mutable interpretive jurisprudence on top of it, and produces obligations, permissions, and prohibitions binding on a community. Concrete traditions (halakha, fiqh and the sharia schools, canon law, dharmic and other traditions) are cited only as worked structural examples of a common mechanism. Wherever this memo appears to state a doctrine, it is describing how practitioners treat a norm, not asserting the doctrine's content is correct. Per §0.4 of the conventions, every substantive claim is tagged; when uncertain we tag down. Per M4, we treat the theological surface as presentation and look for the computational object beneath it.

4.0 Orientation: why religion is a distinct behavior class

Most sources surveyed in this series derive a norm's authority from a procedure that is, at least in principle, visible and repeatable: a vote (Memo 01), a contract (Memo 05), a precedent chain (Memo 02), a convergent convention (Memo 06). Religion is the paradigm case of a source whose root authority is asserted to originate outside any such procedure — in a revelation held to be non-derivable from and non-revisable by the community that receives it. INFERENCE This is what qualifies religion as its own computational-behavior class under JD Q3 ("norm generated by revelation") rather than a subspecies of authority or tradition. INFERENCE

The distinction matters for the kernel/above-kernel boundary because a revelation-rooted system exhibits a signature shape: an immutable base (the source) + a mutable interpretive layer (the jurisprudence) + an enforcement stack that migrates across conscience, community, and (historically) state. That shape — an immutable root of trust with a mutable interpretation layer over it — turns out to be computationally clean even though its content is maximally non-neutral. Keeping those two observations apart (clean shape, non-neutral content) is the central task of this memo.

Vocabulary reminder (from §0.6): we distinguish legitimacy (subjects treat the norm as binding independent of coercion), validity (formal membership of a norm in the system's rule-set), and efficacy (actual compliance). Religious systems are unusually instructive because these three routinely come apart, and the gaps are where the interesting engineering lives.

4.1 Established consensus

Claims here are broadly agreed within religious studies, sociology of religion, comparative law, and the anthropology of law. Agreement is about structure and history, not about any faith's truth.

4.1.1 Religion is, among other things, a normative system

FACT Every major religious tradition that has been studied at scale carries a body of prescriptions — obligations, permissions, prohibitions — governing conduct (ritual, diet, sexuality, property, contract, dispute resolution, inheritance, warfare, treatment of outsiders). Whatever else religion is, it demonstrably produces, maintains, and enforces ought-statements, which is the series' definition of a normative system (§0.6).

FACT Several traditions developed full-blown legal systems, not merely moral exhortation: rabbinic halakha, Islamic fiqh (the jurists' articulation of sharia), Roman Catholic and Eastern canon law, and the Hindu Dharmaśāstra literature are the standard scholarly examples of elaborated religious law with courts, procedures, professional interpreters, and remedies.

INFERENCE Because these are legal systems, the Hohfeldian primitives (§0.6) apply directly: religious law allocates duties and correlative rights, privileges and correlative no-rights, powers (e.g., to marry, divorce, vow, manumit, endow) with correlative liabilities, and immunities/disabilities (e.g., sacred property withdrawn from ordinary commerce). The eight jural relations describe religious norms without residue at least as well as they describe secular law. INFERENCE — provisional, as flagged in §0.6.

4.1.2 Origin: revelation and tradition as the two roots

FACT Religious traditions typically ground their norms in one or both of two roots:

1. Revelation — an authoritative disclosure held to originate from a transcendent source, fixed in a source-object: a text (Torah, Qur'an, Bible, Vedas as śruti — "that which is heard"), an event (Sinai, the Incarnation, the Prophet's career), or a person treated as a channel.
2. Tradition — an accumulated body of transmitted practice and teaching held to carry authority derived from, but distinct from, the revelation (rabbinic Oral Torah; Sunni Sunna transmitted via hadith; Catholic Sacred Tradition; Hindu smṛti — "that which is remembered").

FACT Traditions differ sharply in how they weight these two and whether they admit a third, derivative root produced by human reasoning (e.g., qiyas [analogy] and ijma [consensus] in classical Sunni usul al-fiqh; rabbinic middot / hermeneutic rules; scholastic reasoning in canon law). FACT

INFERENCE Computationally, the origin story has a recurring shape:

The base is asserted to be immutable; everything below it is where change is permitted to happen. This is the immutable-root-of-trust pattern developed in §4.5.1. INFERENCE

4.1.3 Legitimacy: sacred authority and faith (distinct from validity and efficacy)

FACT The characteristic legitimacy claim of a religious norm is sacred authority: the norm binds because its ultimate source is held sacred, and acceptance is mediated by faith — a disposition to treat the source as authoritative that is not itself grounded in ordinary empirical verification.

INFERENCE In the series' vocabulary this produces an unusual legitimacy profile:

• Legitimacy is anchored outside the community and outside time: the norm is binding because the sacred source commands it, not because a living authority enacted it. This is what Max Weber's typology labels traditional and charismatic authority as opposed to legal-rational authority. FACT that Weber drew this three-way distinction; the application here is INFERENCE.
• Validity (formal membership) is determined by the interpretive layer: a norm is "valid halakha" or "valid fiqh" if the jurisprudence certifies it as correctly derived from the roots. Validity is thus internal and procedural even where the root is not. INFERENCE
• Efficacy (compliance) depends on the enforcement stack (§4.1.4) and can diverge widely from validity — a norm can be unquestionably valid within the system yet widely unobserved. FACT , well documented across traditions.

HYPOTHESIS The reason religious legitimacy is robust to coercion-withdrawal (people keep complying when no one is watching and no state enforces) is that the enforcement locus is partly internalized as conscience (§4.1.4), which makes efficacy less dependent on external sanction than in most secular systems. Stated so as to be falsifiable: if this is true, measured compliance with purely religious norms should fall less than compliance with secular regulations of comparable cost when external monitoring is removed. HYPOTHESIS — M1-compliant, not asserted as established.

4.1.4 Enforcement: a three-layer stack (conscience, community, state)

FACT Religious norms are enforced through some combination of three mechanisms, which historically stack and shift:

FACT Layer 3 is historically common but not necessary: many traditions operated for long periods with only Layers 1–2 (diaspora communities; minority/persecuted periods; disestablished modern contexts). FACT

INFERENCE The engineering-relevant point: religion is the clearest natural example of a normative system whose primary enforcement can sit at Layer 1 (internal) and Layer 2 (community), independent of state monopoly on force. Most secular legal theory (Memo 02) assumes Layer 3 is central; religion shows Layers 1–2 can carry a functioning normative order for centuries. INFERENCE

OPEN Whether Layer-1 (conscience/transcendent-accounting) enforcement is representable at all in a governance computation — as opposed to being a purely internal state that the system can only observe indirectly through behavior — is unresolved and is flagged again in §4.3.

4.1.5 Mutation: how a system with an immutable base still changes

FACT Despite an immutable-base claim, religious normative systems change substantially over time. The observed mutation mechanisms are:

• Reinterpretation of a fixed source. The base text is unchanged; its applicable meaning is re-derived. This is the dominant, sanctioned mode. Examples treated structurally: sustained rabbinic exegesis; the operation of ijtihad (independent juristic reasoning) and the maqasid (purposes) literature in Islamic law; the development of Catholic doctrine articulated as organic development rather than change; case-by-case dharmic accommodation of local custom (ācāra, deśadharma). FACT that these mechanisms exist and are so described within the traditions.
• Authoritative deliberative bodies. Councils and synods (e.g., ecumenical councils in Christianity), responsa literature and rabbinic courts, schools of jurisprudence (madhahib) and their consensus (ijma). These institutionalize interpretation and can ratify change as authoritative. FACT
• Schism. When the interpretive layer cannot contain disagreement, the community forks: the same base source is retained, but a divergent interpretive lineage separates (major documented examples: the Sunni/Shia divergence; Chalcedonian and non-Chalcedonian churches; the Reformation; Theravada/Mahayana differentiation; numerous others). Schism is a fork of the interpretation layer, not of the root. INFERENCE , though the historical schisms themselves are FACT.
• Reform / revival movements. Deliberate attempts either to return to a purified reading of the base (restorationist/revivalist patterns) or to modernize application (reform/liberal movements). Both are re-derivations claiming superior fidelity to the same root. FACT such movements exist.
• Desuetude and de facto drift. Norms remain formally valid but fall out of efficacy; later interpretation may recognize this. FACT this occurs; attribution of intent is INFERENCE.
• "Closing/opening of the gate." Some traditions at times treated the interpretive frontier as effectively closed (the contested historiographic claim about a "closing of the gate of ijtihad") and later as reopened. The historicity of any hard closure is itself contested among scholars. OPEN / contested — do not assert.

INFERENCE The invariant across all of these: mutation is confined to the interpretive layer; the root is held constant, and change is legitimated as better access to an unchanged root, not as amendment of the root. This is the signature that distinguishes revelation-type mutation from legislative amendment (Memo 01) and from precedent evolution (Memo 02), where the authoritative text itself is edited or overruled. INFERENCE — this is the key structural claim of the memo; see §4.2 for a competing reading.

4.1.6 Hierarchy: scripture > tradition > interpretation

FACT Religious legal systems are explicitly hierarchical about source authority. The canonical ordering, present in structurally similar form across traditions, is:

FACT Conflict-resolution rules run upward: a lower-rank norm that conflicts with a higher-rank source is (in principle) invalid or must be reinterpreted to conform (a lexical/priority ordering). Traditions differ on the details (e.g., how authenticated tradition may specify or restrict apparent scripture), but the existence of a strict priority ordering with the root at the top is common. FACT for the existence of the ordering; the specific override rules are tradition-specific and INFERENCE when generalized.

INFERENCE This is a lexicographic priority structure: it resembles constitutional supremacy (Memo 01/02, "constitution > statute > regulation") with the crucial difference that Rank 0 is not amendable by any internal procedure. There is no religious analog to a constitutional amendment of scripture itself within orthodox forms; the amendment power that constitutions reserve is, for the root, absent by design. INFERENCE

4.1.7 Conflicts

FACT Religious normative systems generate several structurally distinct conflict types:

1. Intra-traditional (interpretive) conflict. Rival valid derivations from the same root. Managed by the interpretive institutions (schools, courts, councils); unresolved cases may precipitate schism (§4.1.5). FACT
2. Inter-faith conflict. Incompatible roots. Because each system's root is non-revisable and self-certifying, there is no shared higher norm to which both can appeal; resolution is therefore political/coexistential (treaties, toleration regimes, dhimma-type arrangements historically, modern pluralism) rather than jurisprudential. INFERENCE
3. Sacred-law vs secular-law conflict. Where a religious norm and a state norm both claim a subject's obedience. Modern states handle this through establishment/disestablishment settlements, freedom-of-religion regimes, conscientious-exemption doctrines, and confessional-court carve-outs (e.g., religious family-law tribunals operating within a secular state under defined limits). FACT that such mechanisms exist.
4. Establishment conflict. Contest over whether and which religion is backed by state power (Layer 3), and the correlative status of dissenters and non-believers. FACT this is a recurring historical and constitutional issue.

INFERENCE Types 2–4 are the reason religion cannot be resolved inside a content-neutral kernel: they are conflicts between incommensurable roots, and any kernel that picked a winner would be importing non-neutral content. This directly motivates the Q8 exclusion argument (§4.6).

4.2 The six-part analysis (JD Q2), consolidated

The conventions require, for each source, an explicit six/seven-part analysis: origin, legitimacy, enforcement, mutation, hierarchy, conflicts, computational implications. §4.1 developed these discursively; this section states them in compact, auditable form for the source "religion (revelation-rooted normative system)."

4.3 Structural worked examples (comparative, non-doctrinal)

These are included to show the same mechanism instantiated differently. The descriptions are of legal-structural organization, deliberately silent on doctrinal content. Any error here should be read as a structural simplification, not a doctrinal claim.

A. Halakha (rabbinic Judaism). FACT , structural - Root: Written Torah (Rank 0). Authoritative tradition: Oral Torah, codified through Mishnah and Talmud (Rank 1). Derived reasoning: hermeneutic rules (middot), later codes and responsa (Rank 2–3). - Interpreters: rabbinic scholars and courts (bet din); authority via scholarly ordination and reception, not a single central magisterium. - Enforcement: strongly Layers 1–2 (conscience + community, including herem); Layer 3 only where a polity adopts it. - Mutation: continuous responsa and codification; movement differentiation (e.g., Orthodox/Conservative/Reform) as interpretive forks over the authority and method of change. FACT such movements exist and differ on method.

B. Fiqh / sharia (Islam). FACT , structural - Root: Qur'an (Rank 0) + Sunna via authenticated hadith (Rank 1). Derived sources in classical usul al-fiqh: ijma (consensus) and qiyas (analogy), with school-specific additional principles (Rank 2). Particular fatawa and court rulings (Rank 3). - Interpreters: trained jurists (fuqaha) organized into schools (madhahib); no single universal magisterium in Sunni Islam; distinct authority structures in Shia branches (e.g., living jurists' authority). FACT - Mutation: ijtihad (independent reasoning) vs taqlid (following established rulings); maqasid (higher purposes) reasoning; the contested historiography of a "closed gate." FACT the concepts; OPEN the closure historiography. - Enforcement: Layers 1–2 pervasively; Layer 3 where a state adopts it.

C. Canon law (Roman Catholic, with Eastern analogues). FACT , structural - Root: Scripture (Rank 0) + Sacred Tradition (Rank 1), with an authoritative living magisterium empowered to interpret both (a strong, centralized interpretive authority — a notable structural contrast with A and B). - Codification: a promulgated, revisable code of canon law (the interpretive layer is explicitly written and periodically re-promulgated). FACT - Mutation: ecumenical councils and magisterial teaching; doctrine framed as development, not change. FACT the framing. - Enforcement: Layers 1–2 (incl. excommunicatio); Layer 3 historically where established. - Note: canon law is the standard scholarly bridge case between religious law and Western secular legal technique (procedure, courts, codification). FACT

D. Dharmic traditions (Hindu Dharmaśāstra; and, differently, Buddhist Vinaya). FACT , structural - Hindu: śruti (Vedas, Rank 0) > smṛti (remembered texts incl. Dharmaśāstra, Rank 1) > ācāra/custom of the good (Rank 2), with pronounced accommodation of local and community-specific norm-sets rather than a single universal code. Interpretation is decentralized; no single magisterium. FACT /INFERENCE - Buddhist Vinaya: a monastic rule-set with explicit procedural rules for its own application and for communal decision-making — a case where the religious norm-system is heavily procedural and community-adjudicated, and where the "root" includes rules about how disputes over the rules are handled. FACT , structural.

INFERENCE Cross-example generalization: the traditions differ most in the centralization of the interpretive layer (from a single living magisterium [C] to decentralized scholarly reception [A, B, D]) and in who may perform authoritative mutation. They differ least in the shape: immutable root, ranked derived layers, mutation confined above the root, enforcement anchored first in conscience and community. This near-invariance of shape across divergent content is the memo's main structural finding. INFERENCE / candidate HYPOTHESIS for Memo 10's taxonomy.

4.4 Competing theories (presented without adjudication)

Per §0.5, live disagreements are presented, not resolved.

C1. What is religion, for norm-generation purposes? - Substantive definitions (religion = orientation to a transcendent/sacred referent) vs functional definitions (religion = whatever plays a certain social-integrative or meaning-providing role, per the tradition associated with Durkheim). The functional view would fold some secular ideologies into the same behavior class; the substantive view would not. FACT these two definitional families exist; the choice is contested. The choice matters for the taxonomy (Memo 10): does "revelation-rooted" pick out religions specifically, or any system with a treated-as-sacred immutable root (including secular founding myths, "civil religion", constitutional veneration)? OPEN

C2. Is the immutable root really immutable, or is that a legitimating story? - One reading (broadly the "law is interpretation" family, and the position associated with legal realism transplanted to religious law): the root does effectively change because interpretation can reach almost any result, so "immutability of the root" is a legitimation device masking ordinary normative change. FACT this critique is made. - The competing reading (internal/participant and much of comparative law): the root genuinely constrains — not every interpretation is available, the interpretive layer has its own rules of the game, and communities treat certain readings as out of bounds. FACT this position is held. - This memo does not adjudicate. It notes only that for engineering purposes the two readings converge on the same data structure (immutable base + constrained-but-mutable interpretation function); they disagree about how tight the constraint is, i.e., about a parameter, not the shape. INFERENCE

C3. Secularization vs persistence. - The secularization thesis (modernization erodes religious norm-authority) vs the persistence/desecularization readings (religion remains or resurges as a norm-source; the position associated with Weber is often invoked on the disenchantment side, and various sociologists of religion on the persistence side). FACT both theses are actively defended. OPEN as to which holds where; irrelevant to the kernel design but relevant to efficacy estimates D2 might rely on.

C4. Charismatic vs traditional vs legal-rational authority — is religion reducible to one? The Weberian three-type schema is widely used FACT, but whether religious authority is primarily charismatic (founder), traditional (inherited), or has become legal-rational (codified canon law, bureaucratic religious courts) is contested and varies by tradition and era. OPEN

C5. Natural law as a bridge or a category error. Some theories (the natural -law family, associated with Aquinas historically and with later secular natural-law theorists) treat religiously articulated norms as accessible to reason and hence continuous with secular ethics (Memo 03) and jurisprudence (Memo 02). Others treat revealed norms as categorically discontinuous with reason-derived norms. Whether "revelation" is a distinct behavior class or a special case of "authority + tradition" depends partly on this dispute. OPEN

4.5 Computational implications

This is the section the compiler architects care about. Everything here is INFERENCE or HYPOTHESIS; none of it is asserted as established, per M1.

4.5.1 The immutable-root-of-trust analogy

HYPOTHESIS A revelation-rooted normative system maps cleanly onto the security-engineering pattern of a root of trust:

The analogy's payoff: security engineering already has vocabulary and machinery for exactly this shape — an immutable root that signs a mutable, versioned policy layer; trust that flows in one direction; the ability to fork the policy layer while pinning the same root; and the impossibility of the policy layer rewriting its own root. INFERENCE If D1 needs to represent revelation-type sources, it likely does not need a new mechanism; it needs a generic immutable-root + versioned-interpretation-layer construct, of which religion is one instantiation and (per C1/C3 in §4.4) constitutional foundationalism may be another. HYPOTHESIS

4.5.2 Interpretation as the mutable layer over an immutable base

INFERENCE The correct decomposition for D1 is:

• The source-object (root) is data: an opaque, versioned, content-hashed blob the kernel treats as fixed. The kernel never needs to understand it.
• The interpretation function is where all normative computation happens: it takes (root, tradition, context) and emits applied norms (Hohfeldian relations / deontic O,P,F statements). It is versioned, forkable, and carries its own procedural metarules.
• Validity = "certified by the interpretation layer as correctly derived." This is internal and checkable even though the root's authority is not. INFERENCE Consequently, a kernel can mechanically check validity-relative -to-an-interpretation without ever ruling on the root's legitimacy. This is the key separability result of the memo. HYPOTHESIS — offered for destructive testing per M5 and Memo 11.

INFERENCE Enforcement maps to the three-layer stack (§4.1.4). Only Layer 3 (state coercion) resembles what most secular models assume. Layers 1–2 imply a governance computation might need to represent self-enforcement and community sanction as first-class, not as degenerate cases of state enforcement. Whether Layer 1 is representable at all is OPEN (see §4.3 note and §4.6.3 below).

4.5.3 Non-falsifiable axioms and the M1 boundary

INFERENCE The root of a revelation-rooted system is, by construction, a set of non-falsifiable axioms: statements the community treats as authoritative and that are not offered for empirical refutation. This is not a criticism — it is a structural fact about the source class and precisely what makes it a distinct behavior (Q3).

The connection to M1 (falsifiability) is sharp and must be stated carefully:

• M1 governs this series' own scientific claims. It does not require that the content handed to the kernel be falsifiable. A governance system routinely operates over non-falsifiable value-axioms; that is normal, not a defect. INFERENCE
• Therefore the scientific claim we may make under M1 is not "the revelation is true/false" (unfalsifiable, hence out of bounds) but rather the structural claim: "systems in this class exhibit the immutable-root + mutable-interpretation shape, with enforcement on the three-layer stack." That structural claim is falsifiable — it predicts we will not find an orthodox tradition that amends its root by ordinary internal procedure, and it can be broken by a counterexample. HYPOTHESIS , deliberately falsifiable per M1.
• OPEN Candidate counterexamples to stress in Memo 11: traditions with ongoing revelation (a root that keeps growing by authoritative addition), which blur the "immutable root" claim by allowing the root itself to be extended (not edited, but appended). Does "append-only root" still count as immutable? This is an important edge case for D1's data model.

INFERENCE Design consequence: the kernel must be able to carry non-falsifiable axioms as opaque roots without evaluating them, and must keep the falsifiable structural machinery (interpretation, validity-checking, conflict priority) strictly separate from the non-falsifiable content. This separation is itself a testable design commitment (M7: use the separation only if it emerges — here it appears to emerge cleanly).

4.5.4 Text-diagram: religion in the normative dependency graph (JD Q4/Q7 preview)

INFERENCE Note the feedback loop religion -> ethics -> culture -> politics -> law -> (re)establishment -> religion, previewed for Memo 09. The sacred-vs-secular edge (bottom) is bidirectional and is the site of the Q8 exclusion argument.

4.6 Why specific religions must stay ABOVE the kernel (JD Q8)

JD Q8 asks which concepts must never enter the Governance Kernel and demands a justification for each exclusion. This memo owns the exclusion of specific religions (Christianity, Islam, Buddhism, Judaism, Hinduism, and any named tradition or its doctrines). The argument has three independent legs; any one suffices, and they are stated so they can be attacked (M3/M5, Memo 11).

4.6.1 Content non-neutrality (the primary argument). INFERENCE A specific religion's norms are indexed to a particular non-falsifiable root that other traditions reject. Embedding any such root in the kernel would make the kernel non-neutral among sources — it would be taking a side on an incommensurable inter-faith conflict (§4.1.7, type 2). The kernel is by hypothesis content-neutral (§0.6); a specific religion is maximal content. Therefore specific religions belong above the kernel by definition of the kernel. This is the same argument that excludes "democracy", "justice", "capitalism", etc. (Memo 03, Memo 11) — religion is simply the cleanest case because its non-neutrality is explicit and self-declared.

4.6.2 Non-amendability would freeze the kernel. INFERENCE A defining structural property of a religious root is that it is not amendable by internal procedure (§4.1.6). If such a root were placed in the kernel, the kernel would inherit an unamendable normative commitment — contradicting the requirement that the kernel be content-neutral machinery reusable across all sources. The immutability that is a feature at the source layer would be a fatal rigidity at the kernel layer. Hence the root must sit above the kernel, as source-data the kernel carries but does not adopt.

4.6.3 Layer-1 enforcement is not mechanizable by the kernel. HYPOTHESIS /OPEN Religious norms characteristically rely on conscience and transcendent accounting (Layer 1), which are internal states the kernel cannot observe, verify, or actuate. A kernel that tried to own religious enforcement would either (a) reduce it to Layer 3 coercion — distorting the source — or (b) claim to compute something it cannot access. Either way the kernel oversteps. Therefore religious enforcement's Layer-1 component must remain above/outside the kernel; the kernel at most represents the structure of Layers 2–3. OPEN whether Layer 1 is representable even in principle.

Positive statement of what the kernel MAY carry. INFERENCE The exclusion is of content, not shape. The kernel (or D1) may legitimately carry, in fully content-neutral form: - the immutable-root-of-trust construct (§4.5.1), instantiated by some root supplied from above; - the versioned/forkable interpretation-layer abstraction (§4.5.2); - lexicographic priority ordering for conflict resolution (§4.1.6); - a multi-layer enforcement representation covering Layers 2–3. None of these names a religion; each is a generic mechanism that religion happens to exhibit with unusual clarity. This is the memo's constructive contribution to Memo 12's interface.

4.7 Open questions

Per §0.5. These are unresolved and must be treated as unresolved by D1/D2.

• OPEN O-1. Append-only roots. Do traditions with ongoing/continuing revelation violate the immutable-root claim, or is an append-only root still "immutable" for engineering purposes? Directly affects D1's root data model (immutable blob vs. append-only log). (§4.5.3.)
• OPEN O-2. Boundary of the behavior class. Is "norm generated by revelation" a distinct primitive, or a composite of "authority + tradition + non-falsifiable root"? If composite, it should be decomposed in Memo 10, not kept as a primitive (M2 elimination). (§4.4 C1, C5.)
• OPEN O-3. Representability of Layer-1 enforcement. Can conscience / transcendent accounting be represented at all, or only its behavioral shadow? If only the shadow, D2 must not claim to model religious compliance directly. (§4.1.4, §4.6.3.)
• OPEN O-4. How tight is the interpretive constraint? The C2 dispute is a parameter (how much the root actually constrains interpretation). D1 needs some value or range for this parameter to model validity; the memo cannot supply it. (§4.4 C2.)
• OPEN O-5. Inter-root conflict has no jurisprudential solution. Confirmed structural feature (§4.1.7): incommensurable roots share no higher norm. Any D2 that must arbitrate between religious systems is doing politics (Memo 01/08), not computation, and should be labeled as such.
• OPEN O-6. Self-certification and Gödel-adjacent worries. A system whose root certifies its own authority is self-referential. Whether this raises a genuine formal obstacle (cf. Memo 11's treatment of Gödel/Turing) or is harmless because the kernel never evaluates the root's truth is unresolved and is handed to Memo 11.

4.8 Research opportunities

Per §0.5. Where D1/D2 work could contribute or must decide.

• R-1. Generic immutable-root primitive. Specify a content-neutral "immutable root + versioned interpretation layer" object in D1, with religion, constitutional foundationalism, and canonical-text systems as test instantiations. Payoff: one mechanism covers several sources (economy of primitives, M2). (§4.5.1–4.5.2.)
• R-2. Fork semantics for interpretation layers. Model schism as a fork of the interpretation layer pinning the same root. Define merge/no- merge semantics and how two forks' rulings are compared. Directly reusable for legal-precedent forks (Memo 02) and standards forks (Memo 07). (§4.1.5.)
• R-3. Multi-layer enforcement type. Design an enforcement representation with explicit Layer-2 (community) and Layer-3 (coercive) slots and a marked, possibly-null Layer-1 (internal). Forces every source to declare where its enforcement actually sits, exposing the common secular assumption that enforcement = state coercion. (§4.1.4, §4.6.3.)
• R-4. Lexicographic conflict resolver. Reuse the scripture>tradition> interpretation>ruling ordering as a general priority resolver, parameterized by which layer is non-amendable. Test against constitutional supremacy (Memo 01/02). (§4.1.6.)
• R-5. Falsifiable structural predictions. Turn §4.5.3's structural claim into an explicit test suite for Memo 11: search for (a) orthodox internal amendment of a root, (b) a tradition with no interpretive layer, (c) a purely Layer-3 religion with no conscience component. Each would falsify part of the model. (M1, M5.)
• R-6. Non-falsifiable-axiom carrier. Prototype in D1 a mechanism to carry opaque, non-evaluable value-axioms as content-hashed roots, with a hard guarantee the kernel never branches on their truth, only on interpretation-certified validity. This is the concrete separability test for §4.5.3. (M7.)

4.9 Handoff to D1/D2 (concrete computational implications)

For the compiler architects. Each item is an actionable design input; none requires accepting any religious content.

1. Provide a generic ImmutableRoot object. Content-hashed, versioned or append-only (resolve O-1), never edited by in-system procedure, never evaluated for truth by the kernel. Religion is one supplier; constitutions may be another. (§4.5.1, §4.6.2.)
2. Make the interpretation layer the locus of all normative computation. A versioned, forkable InterpretationFunction: (root, tradition, context) -> {Hohfeldian relations, deontic O/P/F} with its own procedural metarules. Schism = fork; reform = new version pinning the same root. (§4.5.2, R-2.)
3. Define validity as interpretation-certified, independent of root legitimacy. The kernel checks "correctly derived under interpretation version V" and must be able to do so without ruling on the root. This is the separability commitment; test it destructively (R-6, Memo 11). (§4.5.2.)
4. Represent enforcement as a three-layer stack with Layer-1 (internal) marked and possibly null; do not silently collapse it into Layer-3 (state coercion). Treat Layer-1 as at best observable via behavioral shadow (O-3). (§4.1.4, R-3.)
5. Adopt a lexicographic priority resolver (root > tradition > derived > ruling), parameterized by which layer is non-amendable. Reuse for constitutional supremacy. (§4.1.6, R-4.)
6. Do NOT admit any specific religion, doctrine, or sacred text as kernel content. Carry them only as opaque roots supplied from above. Justification: content non-neutrality, non-amendability-induced rigidity, and non-mechanizable Layer-1 enforcement (§4.6). This is a hard boundary, not a default.
7. Flag inter-root conflict as out of computational scope. Any arbitration between incommensurable religious roots is politics (Memo 01/08), not kernel computation; D2 must label such decisions as imported political inputs, not derived results. (§4.1.7, O-5.)
8. Carry non-falsifiable axioms without evaluating them. The kernel branches only on interpretation-certified validity, never on a root's truth-value. This keeps the whole construction M1-clean at the level of the series' claims while still operating over unfalsifiable content. (§4.5.3, R-6.)

One-line summary for D1/D2: Religion contributes not content but a reusable shape — an immutable root of trust beneath a versioned, forkable interpretation layer, enforced across conscience/community/state — and the firm rule that the root's content, however clean the shape, stays above the kernel.

End of Memo 04. Cross-references: Memo 02 (validity/priority, legal interpretation), Memo 03 (ethics/natural-law bridge), Memo 06 (culture, civil religion, functional definitions), Memo 09 (mutation, feedback loops), Memo 10 (taxonomy: is "revelation" a primitive?), Memo 11 (destructive tests, Gödel/ self-reference), Memo 12 (interface synthesis).