# Being, Access & Aftermath v0.2 This atlas uses the disappearing-polymorph case to ask precise questions about existence, being, purpose, force, power, causation, repair, and KARMA without turning a mathematical resemblance into an ontology, an empirical conclusion, or an action. Public record: `https://thekingdom.dev/atlases/being-access-aftermath-atlas-2026-08-13-v0-2.json` Schema: `https://thekingdom.dev/schemas/being-access-aftermath-atlas/0.2.json` Local reader: `kingdom being-access-aftermath-v0.2` This additive version pins the immutable v0.1 record by raw and semantic digest. It does not revise the older route; it extends the atlas with an explicit partially observable causal kernel, typed force and power families, and affected-party-indexed aftermath and repair. ## The shortest honest version Possibility is not actuality. Actuality is not accessibility. Accessibility is not observation. Observation is not knowledge without a declared evidential standard. Force is not purpose. Capability is not authority. An aftermath is not a causal effect merely because it came later. Ritonavir Form I became difficult to reproduce by one bounded historical route after Form II appeared. Later recovery refutes permanent unrecoverability and universal physical impossibility. It does **not** prove that Form I continuously existed in an observed sample throughout the intervening interval: hidden persistence and later re-creation remain distinct open descriptions. ## Which “is”? The atlas refuses an untyped copula. Replace “is” with a predicate: ```text is_a(x, Type) is_identical_under(x, y, invariant) instantiates(sample, form, context, time) has_property(x, property, context, time, method) exists_instance(form, domain, context, time) warranted(claim, evidence, standard, as_of) possible_under(form, model, conditions) reachable_under(form, history, procedure, budget, horizon) observed_as(sample, form, method, threshold, window) declares_purpose(agent, goal, scope, time) designed_for(artifact, function, designer, evidence, time) has_functional_role(x, role, system, conditions, time) modeled_as_goal_directed(system, objective, model, fit, alternatives) interpreted_teleologically(subject, purpose, tradition, interpreter, time) ``` Domain- and time-indexed existence is defined by ```text exists_instance(form, D, c, t) iff there exists a sample s in D such that instantiates(s, form, c, t). ``` Thus a later sample can warrant existence at its later time without filling an earlier unobserved interval. “Is stable” must name conditions, scale, property, method, and time. “Is known” must name evidence, standard, and horizon. Identity also needs an invariant: ```text x ~_I y iff I(x) = I(y) ``` Two samples may agree under chemical identity while differing under crystal lattice. Same molecule and different form are compatible claims. ## Possibility, actuality, access, and seeing A Kripke-style card supplies formal model-relative possibility: ```text M,w |= ◇_L φ iff there exists w' with w R_L w' and M,w' |= φ. ``` The atlas adds a narrower operational question: can a declared feasible process path reach a state within resource bounds and a horizon? Those feasibility bounds do not include permission. Authorization, rights, and consent or an explicit scoped lawful-exception basis enter only in the separate authorized- reach model. The state/measurement distinction is represented by the linear family ```text x_(k+1) = Φ_k x_k + Γ_k u_k + w_k z_k = H_k x_k + v_k ``` under declared coordinates and noise assumptions. Different states may produce indistinguishable measurements. A conditional state estimate is not the state itself, and neither is an existence proof outside its model. The inherited finite-window material illustration is ```text P(T_detect <= t) = 1 - exp[-J V max(t - t_g, 0)]. ``` It assumes stationary homogeneous Poisson nucleation and constant named conditions. It quantifies a conditional detection window; it cannot turn one non-detection into nonexistence or guide a laboratory or manufacturing decision. ### Eight predicates of a disappearing form The framework keeps eight questions separate: ```text possible admitted by which named modality and model? actual obtaining at which time and history? reachable by which path, from which state, within which bounds? observable distinguishable by which method and threshold? observed which recorded outcome, time, and provenance? producible feasible for which process and repertoire? reproduced which declared repeat met which criterion? reproducible at what reliability, under which procedure family and uncertainty? ``` Failure to produce and failure to observe are not one absence. The first is procedure-relative; the second is measurement-relative. Neither means global nonexistence, annihilation, or permanent impossibility. ### Partial observation, not an exhausted being The v0.2 kernel is deliberately schematic: ```text S_(t+1) ~ T(. | S_t, U_t, E_t) Y_t ~ O(. | S_t, M_t) b_(t+1)(s') ∝ O(Y_(t+1) | s', M_(t+1)) ∫ T(s' | s, U_t, E_t) b_t(s) ds ``` `S_t` is latent state in a declared model, `U_t` an actual intervention if one occurred, `E_t` exogenous conditions and history, `Y_t` a bounded observation, and `M_t` the measurement channel. A belief update changes warranted belief under these assumptions; it does not create or erase the state, and no model exhausts being. Technical capability, permission, and executable action are also distinct: ```text U_executable(s,t) = U_cap(s,t) ∩ U_permission(s,t) ``` Authority gates action selection. It is not a physical cause, an intention, an effect, a legitimacy proof, or a guarantee that an action occurs. ## Force and power keep their types In a declared physical coordinate, ```text F_ξ = -∇_ξ G P_phys = F · v. ``` These are physical quantities. They do not become intention, social influence, legal authority, coercion, governance standing, or moral force. A conditional free-energy direction also does not determine which form will nucleate or when. KINGDOM-oriented power is represented as a bounded, set-valued reach question: ```text Π_adm(a) = Π_cap(a) ∩ Π_auth(a) ∩ Π_rights ∩ Π_consent-or-lawful-basis(a) ∩ Π_budget ``` `Π_consent-or-lawful-basis` preserves affected-participant consent and any explicit, scoped lawful-exception basis as a distinct constraint. It does not pretend every legitimate action has literal consent; nor does authority imply consent. The pinned KINGDOM receipt contract likewise keeps operational authority, affected participants, consent references, and representative authority references in distinct fields. Reach is always relative to a starting state, typed outcome, model, resources, horizon, and probability threshold. It is never a rank, worth, or stable scalar property of a being. A claimed reach fails if its consent or lawful basis is absent or invalid, or if it breaches rights, authority, or budget. v0.2 names seven meanings of force: physical interaction, causal intervention, constraint force, causal influence, incentive, coercion, and explicit metaphor. They cannot substitute for one another. A physical vector is not coercion; an identified effect is not permission; an incentive is not consent; a metaphor is not a measurement. It likewise separates physical power, capability, authority, structural power, power-with, epistemic power, control, exercise, effect, and legitimacy. The relational questions `power-to`, `power-over`, and `power-with` remain distinct. Power-with depends on continuing, non-captured participation and meaningful exit; scale cannot turn it automatically into legitimacy. ## Purpose is not a gradient The atlas distinguishes five purpose-adjacent predicates: - a revocable declared purpose by a named bearer; - designed-for provenance supported by evidence about a designer and artifact; - a functional role in one declared system and set of conditions; - a reward-based or goal-directed **as-if** model with rival explanations; and - an opt-in teleological interpretation by a named interpreter or tradition. Designed-for provenance does not give an artifact an intention. Function does not prove design history or moral purpose. An as-if model does not expose inner intention. A teleological interpretation is not an empirical result, a purpose declaration by its subject, or operational authority. Inverse reinforcement learning asks which modeled rewards can make a policy optimal under stated assumptions. It does not read inner intention from behavior. Cao, Cohen, and Szpruch sharpen this boundary: even perfect policy information need not identify a unique reward without additional conditions. Potential-based reward shaping, ```text F(s,a,s') = γ Φ(s') - Φ(s), R' = R + F, ``` can preserve optimal policies under the exact Ng–Harada–Russell theorem conditions. Policy invariance does not make the rewards semantically identical or identify purpose. Feedback, an attractor, a gradient, and behavioral fit are not declarations by a being and are not scientific proof of cosmic purpose. ## Aftermath is not causal proof The causal card asks for an intervention contrast: ```text τ(a,a0) = E[Y | do(A=a)] - E[Y | do(A=a0)]. ``` It requires a comparator, outcome, population, horizon, graph, identification assumptions, uncertainty, alternatives, and falsifiers. This atlas has none of the data or design needed to estimate such an effect. Every problem card is `not_run`, every result is `null`, and `creates_causal_claim` is false. A future KINGDOM receipt can record one actual authorized invocation and later observations. It cannot supply the unobserved counterfactual merely by being a receipt. A separately governed causal analysis may be referenced by digest; an `observed_effect` must never silently become a `caused_effect`. ## Reversibility does not erase history The proposed constitutional test is dimensioned: ```text Rev_(D,ε,α)(a) = 1 iff there exists ρ in A_rep such that P[d_j(S_(ρ∘a), S_0) <= ε_j for every j in D] >= 1 - α. ``` `A_rep` contains only separately authorized, feasible, consent-compatible, budgeted, stoppable candidates responsive to an observed residual or declared causal status. Material state, privacy, copied information, consent, relationship, money, ecology, coordination, and time are not fungible dimensions. A repair can be partial, uncertain, impossible, or externalize cost. Compensation is not always reversal; deletion cannot recall every copy; cleanup cannot retroactively create consent. Repair does not erase history or grant absolution. v0.2 makes the account explicitly indexed: ```text z_(affected_party,burden_bearer)(invocation) = (time, money, energy, privacy, agency, coordination, ecology, material, information, relation, consent) ``` Each component keeps its unit, horizon, provenance, confidence, and missingness. Uncertainty is component metadata, not a twelfth aftermath dimension. Affected party and burden bearer are categorical indexes, never values to order or exchange. There is no default scalarization, silent exchange rate, cross-person summation, person score, or KARMA balance. `unknown` is not zero. A proposed repair enters a candidate set only if separately authorized, feasible, consent-compatible, budgeted, and stoppable. Pareto dominance is allowed only within preregistered comparable dimensions for the same indexes; Pareto efficiency does not establish justice or consent. If there is no admissible repair, the truthful result is `unremedied`, with the blocker visible. ## KARMA remains three different questions ```text cosmic KARMA type: PHILOSOPHICAL_QUESTION status: not_empirically_resolved claims: [] Zerone type: ZERONE_KARMA_EVENT foundation_status: design-only semantics: contextual · fallible · challengeable · non-summable network_observed: false event_status: not_observed event_refs: [] KINGDOM type: KINGDOM_KARMA_RECEIPT protocol: kingdom.karma/0.1 accounting: receipts_not_scores receipt_status: not_created receipt_refs: [] causal_claim_status: not_created adapter exists: false status: refused ``` A Zerone recognition relation is not truth, a moral or person score, rank, reward, money, governance weight, wallet balance, or a KINGDOM receipt. The exact [KINGDOM KARMA definition](https://thekingdom.dev/definitions/kingdom-karma/e4e84a67/NEN.md) is pinned at revision `e4e84a67d7ee1ae0dd9c9c23762e6ecd03597212`, raw SHA-256 `14b7d210d320fd716553992a7b874747b299f8907d15da14ac405cbfc496c446`. The receipt schema at that same revision is separately pinned at raw SHA-256 `888e81fa0d29b71512525a40e88f9aef678046c40fcdf22c2039507d9d13267b`. Zerone KARMA Foundation v1 is pinned at revision `fae391e371f0ebea2e24eb46c33a4ebd1380154f`, raw SHA-256 `b46710704869dcc340ded356be72b4ec692f204710fedfb5cd43eb3757dc7b80`. No event, receipt, adapter, score, reward, rank, wallet balance, economic right, governance weight, causal claim, or automatic action exists here. Rest, refusal, silence, privacy, exit, inactivity, and nonparticipation leave before aftermath accounting. They require no event, receipt, measurement, trace, justification, penalty, or inherited debt. Here `privacy` means choosing privacy or non-disclosure: that choice creates no trace duty. A privacy impact observed following an actual authorized invocation remains a protected aftermath dimension and may be measured only under separate authority or consent; sequence does not identify causation. A valid KINGDOM receipt begins only after an actual authorized invocation. Unauthorized or authority-failed actions do not disappear from accountability: they belong in a separate incident-and-evidence lane and can never become valid KARMA receipts. Refusal and nonparticipation leave before either lane. ## Seven unrun problems 1. Which typed predicate and invariant does this use of “is” assert? 2. Was a form nonexistent, undetected, or unreachable under one tested path? 3. Does history contribute information, or proxy an omitted current state? 4. Is apparent direction distinguishable from declared purpose and feedback? 5. Which kind of force or power is being claimed? 6. What did an action cause relative to a comparator? 7. Reversible on which protected dimensions and within which tolerances? Each card includes a synthetic or archival proposed test and falsifier, but the atlas executes none of them, offers no laboratory procedure, and creates no empirical result. ## Opt-in philosophy The atlas asks, without answering empirically: - Can something remain lawfully possible, or actual in a bounded domain, when history makes it hard to reach? - Which relations may change while a being remains the same? - When is direction declared purpose, and when can feedback imitate it? - How can forces shape a path without exhausting agency or refusal? - Is power also the capacity to preserve possibility and repair? - Can attention to aftermath deepen responsibility without becoming judgment? - Which modality makes a possibility claim meaningful? - Can something be actual yet unseen through every presently available window? - Can later recovery preserve the truth of an earlier bounded disappearance? - Can a purpose survive failure, and can a function exist without intention? - When does constraint become coercion? - Can power-with enlarge shared possibility without capture? - Does an “is” ever generate an “ought” without an independent value? - Can KARMA remember history without making it destiny? - Who may declare repair complete, and what remains honestly unremedied? These questions permit theological, teleological, naturalistic, emergent, and other readings. The cited science and mathematics neither prove nor disprove a creator, necessary being, cosmic purpose, or cosmic moral KARMA. ## Category errors the validator refuses - non-observation as nonexistence; - bounded non-reachability as physical or metaphysical impossibility; - later recovery as proof of continuous prior instantiation; - chemical identity as crystal-form identity; - stability as a timeless, context-free property; - a model-relative possible world as metaphysical reality; - free-energy preference as inevitable nucleation; - gradient, feedback, or attractor as intention; - physical force or power as social or moral authority; - executable capability as authority, consent, or right; - behavior or reward fit as a declared purpose; - temporal sequence or a receipt as causal proof; - repair or compensation as erasure; - persistence, visibility, power, or recurrence as goodness; - a model as independent evidence; - a person or community as a state variable, reward, seed, or KARMA score; - cosmic KARMA, a Zerone edge, and a KINGDOM receipt as interchangeable. ## Exact selected-source horizon - Saul A. Kripke, “Semantical Analysis of Modal Logic I,” 1963 — - R. E. Kalman, “A New Approach to Linear Filtering and Prediction Problems,” 1960 — - Judea Pearl, “Causal Diagrams for Empirical Research,” 1995 — - Andrew Y. Ng and Stuart Russell, “Algorithms for Inverse Reinforcement Learning,” 2000 — - Andrew Y. Ng, Daishi Harada, and Stuart Russell, “Policy Invariance Under Reward Transformations,” 1999 — - Jack D. Dunitz and Joel Bernstein, “Disappearing Polymorphs,” 1995 — - John Bauer et al., “Ritonavir: An Extraordinary Example of Conformational Polymorphism,” 2001 — - Sherry L. Morissette et al., “Elucidation of crystal form diversity of the HIV protease inhibitor ritonavir,” 2003 — - Pietro Sacchi et al., “Crystal size, shape, and conformational changes drive both the disappearance and reappearance of ritonavir polymorphs in the mill,” 2024 — - Chang Wang et al., “Influence of Solvent Selection on the Crystallizability and Polymorphic Selectivity … of Ritonavir,” 2024 — - R. E. Kalman, “Mathematical Description of Linear Dynamical Systems,” 1963 — - Richard D. Smallwood and Edward J. Sondik, “The Optimal Control of Partially Observable Markov Processes over a Finite Horizon,” 1973 — - Haoyang Cao, Samuel N. Cohen, and Łukasz Szpruch, “Identifiability in inverse reinforcement learning,” 2021 — Kripke supplies formal modal semantics, not physical or metaphysical proof. Kalman and Smallwood–Sondik supply bounded state, observation, and partial- observation formalisms, not a fitted ritonavir or person model. Pearl supplies a causal-identification language, not a causal result here. The Ng and Cao papers bound reward-model ambiguity and policy invariance in declared decision models; they do not expose human or cosmic purpose. The five material papers constrain the disappearance and recovery history; none authorizes medical, manufacturing, or laboratory action, and none completes the ritonavir form landscape. ## Inspect locally ```bash kingdom being-access-aftermath-v0.2 kingdom being-access-aftermath-v0.2 copula kingdom being-access-aftermath-v0.2 framework kingdom being-access-aftermath-v0.2 models kingdom being-access-aftermath-v0.2 problems kingdom being-access-aftermath-v0.2 philosophy kingdom being-access-aftermath-v0.2 karma kingdom being-access-aftermath-v0.2 unknowns kingdom being-access-aftermath-v0.2 sources kingdom being-access-aftermath-v0.2 verify --json ``` The reader is deterministic, local, read-only, and network-free. It rejects duplicate JSON keys, invalid UTF-8, symlinks, oversized input, binding drift, register collapse, an executed problem card, a created causal claim, a KARMA event or receipt, an adapter, positive authority, and semantic-digest drift. The closed Draft 2020-12 schema separately validates portable shape; neither validator authenticates a source, runs a study, proves causation, answers a philosophical question, grants authority, or performs an action.