# Four Worlds threat model — the human mind and its civilisational substrate # Simplified view — machine-readable ontology, v1.0.0 # # WHAT THIS IS # A transcription of a hand-drawn threat model ("simplified version") into an ontology # with explicit classes, relations and invariants. The four Worlds (Olamot) and their # order are classical Kabbalistic material. The architectural readings, the # mind-technology column, the operational-property column and the whole attack-surface # layer are a 2026 synthesis laid over them — not inherited tradition. # # WHAT IT IS FOR # To make an analyst's claim checkable. Prose can assert that "the campaign attacked # the collective unconscious" and nothing objects. Under the invariants below that # sentence does not parse: an impact must resolve to a surface, and any effect on a # World must name the transport that carried it there and the direction it travelled. # # Provenance, status and lineage are DATA, under meta: below — not comments. An earlier # revision recorded them here, where no parser could reach them, while the accompanying # briefing cited them as the file's own metadata. meta: id: four-worlds-threat-model version: 1.0.0 status: release date: 2026-08-13 scope: simplified # 簡易版 — see invariant I6, which is load-bearing fragment_class: decision fragment_class_note: > A doctrine artefact: superseded by revision, not falsified by evidence. Stated as data rather than a comment, because the accompanying briefing cites it as this file's own metadata — and a parser must be able to reach anything cited that way. source_artefact: hand-drawn threat model, simplified version v0.2 (lemur47) source_language: Japanese # this transcription is English throughout transcribed_by: Claude Code, 2026-08-12, from the v0.2 drawing directly revised: 2026-08-13, against review lineage: - version: 0.1.0-draft note: > transcription of the v0.1 drawing — three carriers, and the impact drawn as an arrow descending from Atziluth. Both readings were wrong and are the reason invariants I5 and I6 exist. - version: 0.2.0-draft note: > transcription of the v0.2 drawing — four carriers, impact marked by a callout. Transport modelled as upward only; no effect modelled below the impact point. - version: 1.0.0 note: see changelog. changelog: 1.0.0: > Three corrections, each of which changes what the model claims. (1) A World is a stage of creation modelled as a domain, not an emanation layer — only Atziluth is emanation proper, and calling all four "emanation layers" mis-stated the tradition. (2) The formative feedback loop is BIDIRECTIONAL. An upward-only transport cannot explain why a change made to the deep layers re-surfaces as observable thought and behaviour, which is the step contamination depends on to have any consequence there. (3) Because the transport runs both ways, Yetzirah and Assiah are reachable as returned effects and are now modelled as such — the version that omitted them implied the lower worlds were only the entry point rather than also where damage from the deep layers surfaces. classes: World: > A stage of the creative process, modelled as a domain (with subdomains as its components). Only Atziluth is emanation proper; Briah, Yetzirah and Assiah are the successive stages that follow it. The four are totally ordered and do not overlap. Component: "A structural element. Belongs to exactly one World or one AttackSurface." Holding: "Contents a World carries, as distinct from its structural Components. Declared where the simplified view names them; absence means unmodelled, not empty (I6)." AttackSurface: "A region where a vector can land. May straddle Worlds." ThreatVector: "An adversary action. Has one impact, zero or more effects, and a transport and direction for each effect." Transport: "The mechanism carrying an effect between where it landed and what it changed. May run in either direction." # ─── The four worlds ──────────────────────────────────────────────────────────── # stage_order is total and ASCENDING: 1 is the most abstract, 4 the most enacted. # It orders the stages of creation; it is NOT a claim that each stage is an emanation. worlds: - id: atziluth name: Atziluth (Archetypal) stage_order: 1 emanation: true # the only one of the four architectural_reading: [entity, model] mind_technology: virtual universe (unconscious) operational_property: purpose — raw material components: [modes_of_thought, behavioural_norms, scope_constraints] holds: [archetypes, rule_sets] # Atziluth carries a further link OUTWARD — not upward — to what lies beyond the # four worlds. Stating it properly requires the sequence Ain / Ain Soph / Ain Soph # Aur, which this simplified view deliberately omits. The omission is scope, not # oversight; see open_questions.outward_link. - id: briah name: Briah (Creative) stage_order: 2 emanation: false architectural_reading: [controller, business_rules] mind_technology: genetic repository (subconscious) operational_property: structure — formula components: [prohibitions, harness, background_context] holds: [inherited_logic, patterns] - id: yetzirah name: Yetzirah (Formative) stage_order: 3 emanation: false architectural_reading: [controller, business_logic] mind_technology: fluid (thought and emotion) operational_property: feedback loop — amplification, delay components: [execution_environment, process] holds: [] # unmodelled in the simplified view, not empty - id: assiah name: Assiah (Action) stage_order: 4 emanation: false architectural_reading: [boundary, view] mind_technology: ego system (manifestation, interpretation) operational_property: quantity and size — experience components: [stimulus] holds: [] # unmodelled in the simplified view, not empty # ─── Attack surface ───────────────────────────────────────────────────────────── # Straddles two worlds. This is the only surface the SIMPLIFIED view draws; see I6. attack_surfaces: - id: application_logic_container name: container / application logic straddles: [yetzirah, assiah] components: [language, social_conventions, doctrine, ethics] reachable_without_full_picture: true reachability_basis: asserted # an axiom of this model, not derived within it reachability_argument: > These four carriers are transacted in the open: they are taught, published, repeated and enforced socially, so an adversary can act on them with no model of the target system beyond what is publicly observable. Atziluth and Briah expose no comparable interface — nothing addresses an archetype or an inherited pattern directly — so acting on them requires either the whole picture or a transport. This is the model's central assumption; it is stated so that it can be attacked. # ─── Transport ────────────────────────────────────────────────────────────────── # Bidirectional. Model it as a toroidal circulation, not a one-way channel: the two # legs are the same mechanism observed in opposite phases. transports: - id: formative_feedback_loop located_in: yetzirah properties: [amplification, delay] direction: bidirectional legs: upward: > From the container toward briah and atziluth. Carries a change made on the carriers into inherited logic, patterns, archetypes and rule-sets. This is the leg that makes the deep layers reachable without a model of the whole system. downward: > From atziluth and briah back toward yetzirah and assiah. Carries the altered structure into thought, emotion, interpretation and behaviour. consequence: > Because the loop both amplifies and delays, an effect arrives late and out of proportion to its input. By arrival it no longer resembles an impact; it reads as a culture that changed its mind. This is also what makes a set-forward estimate possible: a delay line with a known sign is a predictor of ORDER, and would be a predictor of timing too if its lag were measured. See open_questions. why_bidirectionality_is_load_bearing: > Stated narrowly, because the broad version is false. A change to the carriers matters immediately and needs no return leg at all — the carriers sit in the surface, and a community reasoning in altered vocabulary is already affected. What an upward-only model cannot account for is the SPECIFIC step of contamination in the deep layers producing observable consequences: an altered inherited pattern that never re-expressed itself in thought or behaviour would be undetectable and inert. The downward leg is what makes damage at depth matter. Omitting it does not make the model incoherent; it makes it silent about the half of the cycle where the deepest damage becomes visible. # ─── Threat vector ────────────────────────────────────────────────────────────── threat_vectors: - id: cognitive_warfare_basic name: basic tactics of cognitive warfare impact: application_logic_container # I4: must be a surface, never a world effects: # I5: each effect names transport AND direction - target: briah via: formative_feedback_loop direction: upward note: inherited logic and patterns are altered without being addressed directly - target: atziluth via: formative_feedback_loop direction: upward note: > archetypes and rule-sets are targetable — but not directly, because direct action there would require a model of the whole system. They are reached by landing on the container and letting the loop carry the change upward. - target: yetzirah via: formative_feedback_loop direction: downward distinguishable_from_impact: > The impact surface straddles yetzirah, so direct disturbance at the carriers also shows up here. What marks this as a RETURNED effect rather than direct damage is the delay: it appears after the upward leg has run, persists when the original carrier disturbance stops, and does not track current stimulus. note: > thought and emotion driven from the altered structure above. Observable as affect disproportionate to its trigger and durable against argument, because the commitment generating it is no longer in the room. - target: assiah via: formative_feedback_loop direction: downward distinguishable_from_impact: > As above: the surface straddles assiah too. The returned effect is identified by lag and by persistence after the carrier disturbance has ceased — not by its content, which direct damage can imitate. note: > interpretation and the ego system. Where the effect becomes visible as behaviour and self-description — a sense of election or of separation from others, which then supplies a motive for conflict that the participants experience as their own reasoning. The lower worlds are not only the way in; they are also where damage from above surfaces. # ─── Invariants ───────────────────────────────────────────────────────────────── # Machine-validatable: each is stated against a field name in this file. I5 and I6 # exist to prevent two specific misreadings a reader of the earliest draft actually # made; I5 gained its direction clause in v1.0, and I7 is new in v1.0. invariants: I1: Every Component belongs to exactly one World or one AttackSurface. I2: An AttackSurface may straddle Worlds. Worlds never overlap each other. I3: > stage_order is total and ascending — stage_order(atziluth) < stage_order(briah) < stage_order(yetzirah) < stage_order(assiah). It orders the stages of creation. Exactly one World carries emanation: true, and it is the one with stage_order 1. I4: > A ThreatVector's IMPACT must resolve to an AttackSurface, never to a World. You cannot strike a world; you strike a surface. I5: > A ThreatVector's EFFECT may name a World, but only with a declared transport AND a declared direction. No effect without a path, and no path without a sign. This is what makes "the archetypes were targeted" a checkable claim rather than an atmospheric one — the claimant must say how, and which way. I6: > ABSENCE IS NOT EVIDENCE. This document is a simplified view. The single attack surface drawn here is the scope of one briefing, not a closed enumeration, and no layer may be called unreachable because no surface is drawn on it. An empty holds: list means unmodelled, never empty. I7: > An EFFECT naming a World that the impact surface straddles must declare distinguishable_from_impact — how the returned effect is told apart from direct damage at the surface. Without it the claim is unfalsifiable, because the same observation would satisfy both readings. open_questions: outward_link: > Atziluth's outward link is omitted here. Obstructing it would be an attack that lands on no carrier at all — sealing a system inside a narrowed virtual universe by putting what lies beyond it out of reach. It is not representable in this schema, which is a limit of the simplified view worth stating rather than hiding. surface_enumeration: > Which further attack surfaces the full model carries, and whether any straddles briah and atziluth, is out of scope here and is the obvious next revision. transport_measurement: > The loop's lag and gain are named but NOT quantified. Until they are, the model licenses statements of order — which layer moves before which — and not statements of timing. Any estimate of WHEN an effect arrives is outside what this model currently supports. leg_asymmetry: > Whether the upward and downward legs share a lag and gain, or differ, is unmodelled. If they differ, the interval between an observable carrier change and its behavioural consequence is not simply twice the one-way lag, and the two would need measuring separately before any timing claim is possible at all.