# Graphic Equalisation for LLMs > **Single-URL context entry point.** Hand this file's URL to a language model to > give it a compact, faithful operating model of Graphic Equalisation, then ask > your question. It is **not** the proof corpus and grants no result beyond what > its linked sources establish. Follow a source link before asserting a formal > theorem, empirical result, or domain-specific consequence. Each linked document is a plain-markdown rendering of a released draft > (draft-2026-09-06.1); the citable artifacts are the linked PDFs. ## 1. Theory in one screen **Base claim.** *To be is to stand in relation* (see [00 — *Is*](/papers/00_is/00_is.md)). To say that anything **happens** adds a second, independent admission: transformations occur. **Minimal running substrate.** A graph is G = (g, r, Γ): `g` = represented relational state; `r` = operative transformation structure; `Γ` = grammar deciding which graph-originated transformations cohere. `g` alone is sufficient for there to be structured distinction; `r` and `Γ` are the extra structure required for occurrence. **Dynamics (the Law).** ```text G_t ──present structure──> Poss(G_t) ──collapse/realisation──> G_(t+1) ``` `Poss(G_t)` is the set of immediate continuations expressible in the current vocabulary, meeting antecedent conditions, and admitted by `Γ`. Collapse/realisation selects one continuation; the realised successor is the next antecedent. The Law says this recurrence generates the dynamics. It does **not** identify collapse with deliberative choice. **Determinacy.** Given a genuinely identical, causally complete antecedent, the realised successor is fixed. Apparent indeterminacy may instead record a frame that omits causally active distinctions; this is not a license to postulate hidden variables in every case. **Residual.** A graph-originated deformation rejected by the relevant grammar is not thereby undone: it is **residual** — present but inadmissible in that frame. What a system does with residual (propagate, display, catch, cohere it to a value, terminate) is contingent handler/rule content, not substrate semantics. **Frames.** No operational "view from nowhere" exists. Every distinction, measurement, identity judgement, admissibility judgement, and quantity is relative to a frame. Formally, a frame is the boundary relation individuating a subgraph, not the subgraph itself. Frames can model other frames, never exhaustively close their own provenance or running from within. **Consequences to preserve while reasoning.** - **Identity is frame-relative:** continuity of a graph and a container's tracking of an instance are different relations; no absolute sameness is inferred from g, r, Γ alone. - **Time is not a substrate coordinate:** realised changes have enabling/succession order; a frame constructs time from closure cycles it holds. There is no required global tick or universal present. - **No implicit history:** the past affects a transition only via present graph/boundary structure that carries a causal trace. - **Reflection is ordinary graph dynamics:** if `Γ` represents and transforms itself, that is the loop; a new external homunculus is not required. - **Groundings are not the substrate:** a grounding fixes particular rule content, initial conditions, and a domain. Physics, minds, institutions, and software may instantiate different groundings without being reduced to one empirical model. ## 2. Economics and structure **Write economics.** In the substrate characterisation, a destructive delta/write is the characteristic paid operation; standing structure and the boundary are free **to that frame**. Payment may be inherited by a supplying frame, so "free here" never means universally free. This is a revisable detail claim, not an unconditional law of all possible substrates. **Well-formedness.** A graph is well-formed to the degree that it pays for nothing doing no work. Audit four costs: writes; retention; search over what is held; and silent writes to structure with no effect. An ill-formed graph still runs; it pays more. A useful compression is not "remove everything": retained structure can reduce future writes but increases search and inert-write risk. **Carvings are not privileged.** Boundaries/types/categories are frame-made. Under a shared binding constraint, independently developed, economical carvings may converge; convergence outside shared purpose would challenge this account. Do not convert an economical or widely shared classification into an ontological absolute. ## 3. What an LLM should do with the theory When analysing any claim/system, write the smallest explicit ledger: ```text Frame F: what boundary/aperture individuates the object and evidence? State g: what distinctions/relations are presently represented? Rules r: what can transform g? Grammar Γ: what is admissible; which gate/rule applies? Poss: which continuations are actually expressible and antecedent-authorised? Realisation: what one transition/event is claimed to occur? Residual: what was rejected, omitted, deferred, or made incoherent? Cost: who pays which write/search/retention cost, including suppliers? Grounding: which domain-specific laws/initial conditions have been added? Test: what observation would distinguish this account from its alternatives? ``` Use this to **clarify**, not to force-fit. If `g`, `r`, `Γ`, the frame, or the realisation rule cannot be independently specified, say the application is underspecified. Keep these separations intact: 1. `can represent/do` (**capability**) ≠ `can reach from current position/resources` (**accessibility**). 2. `not admitted at input` (**absent to the frame**) ≠ `entered then rejected` (**residual**). 3. `deterministic given a complete antecedent` ≠ `predictable by this observer`. 4. `a grounded/domain result` ≠ `a substrate theorem`. 5. `a model's self-description` ≠ evidence that it has closed its provenance, identity, or truth from within. ## 4. Limits, burden of proof, and falsification This is an information-level framework, not a replacement for field theories and not a blanket proof of the groundings. Its central empirical/compositional claim is that the same arrangement recurs across independently developed domains; that can fail if the vocabulary renders adversarially unrelated domains equally well merely by being generic. Important failure conditions include: (a) realised dynamics needing a primitive beyond g, r, Γ; (b) two causally complete identical antecedents yielding different successors with no operative distinction; (c) persistence/sameness derivable without frame individuation; (d) residual and non-admission coming apart; (e) a frame settling its own provenance/boundary from inside without a crossing/write; (f) an irreducible non-write cost where the economics claims none; or (g) a change paid nowhere in both the acting frame and sustaining container. Some failures revise a mechanism; others bound the theory's claimed range. Read the formal falsifiability section before treating any item as settled. **Do not infer:** a universal physical theory, a proof of consciousness, a specific quantum interpretation, a privileged ontology, an absolute ethics, or a successful implementation from this brief. The groundings are explicit, separately assessable programmes, not automatic consequences of the definitions. ## 5. Source map — follow links for authority Each link is a rendering for machine reading; the citable artifact is the PDF on each paper's page. ### Core formal record - [Is — Relation Without Time, Cause, or Rule](/papers/00_is/00_is.md) — the base relational claim and the removal method - [The Law of Graphic Equalisation — Theory — Minimal Recursive Dynamics, Collapse, Frame-Relative Identity, and Residual](/papers/01_law_of_graphic_equalisation_theory_referenced/01_law_of_graphic_equalisation_theory_referenced.md) — definitions, arguments, propositions, and falsifiers - [The Law of Graphic Equalisation — Statement](/papers/01s_graphic_equalisation_statement/01s_graphic_equalisation_statement.md) — the condensed statement of 01 - [Assessing an Account of This Kind — What a Universal Framework Can Risk, What It Earns Instead, and What the Record Shows](/papers/06_assessing_an_account_of_this_kind/06_assessing_an_account_of_this_kind.md) — what an account of this kind can claim and test ### Direct 01 groundings - [The Law of Graphic Equalisation — Physics Grounding — Change, c, the Reconciliation of Special and General Relativity, the Recovery of the Born Rule, the Dimensional Constants, and the Second Law](/papers/01a_graphic_equalisation_physics_grounding/01a_graphic_equalisation_physics_grounding.md) — change, c, relativity, Born recovery, constants, the second law - [The Law of Graphic Equalisation — Mind and Self Grounding — Frames, the Run and the Record, Free Will, Self-Models, Personal Identity, and Death](/papers/01b_graphic_equalisation_mind_and_self_grounding/01b_graphic_equalisation_mind_and_self_grounding.md) — frames, run and record, self-models, identity, death - [Graphic Equalisation — Cosmology Grounding — Horizons, Measures, and What Has No Outside](/papers/01c_graphic_equalisation_cosmology_grounding/01c_graphic_equalisation_cosmology_grounding.md) — horizons, measures, what has no outside - [The Law of Graphic Equalisation — Born Recovery — An Axiom-by-Axiom Note](/papers/01d_graphic_equalisation_born_recovery/01d_graphic_equalisation_born_recovery.md) — the Born rule, axiom by axiom - [The Law of Graphic Equalisation — Communication Grounding — Difference Across Frames, Kinetic Contact, Records, and Network Frames](/papers/01e_graphic_equalisation_communication_grounding/01e_graphic_equalisation_communication_grounding.md) — difference across frames, records, network frames - [The Law of Graphic Equalisation — Qualia and Consciousness Grounding — Qualia-Space, Aperture, and the Ill-Posed Hard Problem](/papers/01f_graphic_equalisation_qualia_and_consciousness_grounding/01f_graphic_equalisation_qualia_and_consciousness_grounding.md) — qualia-space, aperture, and the ill-posed hard problem ### Related machinery (use only when needed) - [Seldon Manifolds — Theory — The Geometry of Reachable Futures](/papers/02_seldon_manifolds_theory_referenced/02_seldon_manifolds_theory_referenced.md) — represented reachable futures - [Seldon Manifolds — Psychological Grounding — Forward Models, Acceptable Regions, and the Aperture of One Embodiment](/papers/02a_seldon_manifolds_psychological_grounding/02a_seldon_manifolds_psychological_grounding.md) — forward models and acceptable regions - [Seldon Manifolds — Social and Psychological Application — Four Classical Experiments, and What a Substrate Produces Without Being Fitted To It](/papers/02b_seldon_manifolds_social_and_psychological_application/02b_seldon_manifolds_social_and_psychological_application.md) — four classical experiments put to the substrate - [The Law of Narrative Momentum — Theory — First-Class History, Cost-Conditioned Persistence, and Revision](/papers/03_law_of_narrative_momentum_theory_referenced/03_law_of_narrative_momentum_theory_referenced.md) — first-class history, cost-conditioned persistence - [Narrativium — Theory — Frame-Relative Capacity for Consequential Deformation](/papers/04_narrativium_theory_referenced/04_narrativium_theory_referenced.md) — capacity for consequential deformation - [Frames, Relation, and the Empty Limit — A Metaphysical Companion to Graphic Equalisation](/papers/05_frames_relation_and_the_empty_limit/05_frames_relation_and_the_empty_limit.md) — two questions the Law raises and should not answer - [Causal Attitude Networks on the Substrate — What Moved, What Did Not, and What the Mean Field Costs](/papers/07_causal_attitude_networks_on_the_substrate/07_causal_attitude_networks_on_the_substrate.md) — an attitude model rebuilt as native graph rules - [Certified Answers as an Audit — Five Provably-False Keys on Humanity's Last Exam, and the Discipline That Makes Them Worth Reporting](/papers/08_certified_answers_as_an_audit/08_certified_answers_as_an_audit.md) — answer only when certifiable, otherwise miss honestly - [Geopolitical and Civilisational Future Application of Graphic Equalisation — The Account Applied to One Frame](/papers/10_applied_graphic_equalisation/10_applied_graphic_equalisation.md) — the account applied to the geopolitical and civilisational frame --- **Portable invocation:** give an LLM this file's URL (https://ozmalabs.com/papers/graphic-equalisation-for-llms.md), then state the target question and ask it to use §3's ledger, label inference vs. source-backed claim, and open only the relevant source links when verification matters. **Scope note.** This brief links only the papers released on this site. Some programmes referenced in the research corpus (adaptability, constructive implementation, and certain reconciliation programmes) are not publicly released and are intentionally omitted here.