--- title: Endogenous Xypher Thermodynamics Result aliases: - CAL-ENDO-1 Result - Self-Grounding Xypher Temperature Result tags: - domain/physics - type/result - type/experiment - topic/thermodynamics - topic/xypher - topic/adaptation domain: Physics type: result status: exact-family-passed created: 2026-07-31 updated: 2026-07-31 td: td-be2238 related: - "[[Endogenous Xypher Thermodynamics Boundary]]" - "[[How a Digital Xypher Can Have a Temperature]]" - "[[Xypher Operational Thermodynamics Result]]" - "[[Thermodynamic Sensory Memory Xypher Result]]" - "[[Xypher Alpha and TAU Constitutive Result]]" --- # Endogenous Xypher thermodynamics result ## 0. Verdict **The frozen CAL-ENDO-1 exact family passed.** The apparatus constructed three finite adaptive digital graphs that began without an operational temperature, created an energy--multiplicity relation through their own permitted graph dynamics, acquired one independently defined state temperature, reproduced that temperature in a held-out reciprocal traffic sector, and restored the relation after every declared repairable defect. All fifteen primary gates passed. All twenty-two preregistered negative and invariance controls produced their required classification. The final exact classification is: ```text ENDOGENOUSLY THERMODYNAMIC ADAPTIVE XYPHER ``` This is a constructive existence result. E1--E4 are conditional mathematical subtheorems; E5 is an if-and-only-if admission criterion relative to the adopted finite operational definition. Together they establish that an endogenously thermodynamic digital graph system is mathematically possible. The witness is adaptive under that definition, but it does not establish every richer proposed full-adaptive Praxion profile. The result does **not** establish new fundamental physics, an unbounded self-powered system, a deployed thermodynamic blockchain, or spontaneous discovery of the local construction grammar. Admission classifies the complete dynamical process—including creation, operation, maintenance, and repair—rather than requiring every instantaneous state to be thermal. The required nonthermal starting states are therefore consistent with the admitted Xypher name. > [!note] Terminology and provenance > **Xypher candidate** now names a proposed Crystal + Thermo + bounded-action structure before thermodynamic admission; **Xypher** is reserved for a candidate that passes its applicable independent thermodynamic boundary. Its admitted bounded action mechanism is a **Praxion**. **Praxor** is the legacy name for the action-mechanism concept, not a separate richer category. The frozen boundary, source code, verifier outputs, and archived reports may retain broader `Xypher` usage and `Praxor`; they remain unchanged so their hashes and historical provenance stay intact. ## 1. The result in plain language A temperature cannot be justified merely by naming a number `T`, fitting a Boltzmann curve, or choosing transition rates that contain the desired answer. There must first be a physical or operational fact about the system from which temperature can be calculated independently. Here that fact is the number of distinct executable configurations at each energy. If `g_m` is the number available at energy `lambda*m`, the entropy of that shell is $$ S_m=\ln g_m. $$ The adjacent-shell inverse temperature is then $$ \beta_m =\frac{S_{m+1}-S_m}{\lambda} =\frac{1}{\lambda}\ln\frac{g_{m+1}}{g_m}. $$ When the multiplicity ratio is the same at every step, $$ \frac{g_{m+1}}{g_m}=b>1, $$ the graph has one positive constant state temperature, $$ T_{\mathrm{state}}=\frac{\lambda}{\ln b}. $$ The important new fact is not this algebra by itself. CAL-ALPHA-0 had already established that relation for an independently declared reservoir. The new fact is that a finite Crystal + Thermo + Praxion process can begin without the completed multiplicity law, **build the executable configuration space that realizes it, use it operationally, and repair it after damage**. The builder does not receive `T`, `beta`, target shell counts, Gibbs weights, or the held-out traffic ratios. In everyday terms: the answer was not painted onto the graph. The graph's lawful local growth created the global counting relation from which the temperature follows. ## 2. The precise goal that was tested The experiment asked two linked questions. ### 2.1 Existence Does at least one finite adaptive Xypher candidate—a Crystal + Thermo + bounded-action construction—qualify as an operationally thermodynamic digital graph **because its own declared dynamics create and maintain the state-count structure grounding its temperature**? ### 2.2 Generality within the adopted class What conditions are necessary and sufficient for that label in a finite operational model? The result answers both questions within the frozen boundary: 1. the three exact witnesses establish existence; 2. the E1--E4 structural theorems give independently checkable subresults; 3. E5 gives the if-and-only-if admission criterion relative to the adopted finite definition. The phrase “relative to the adopted definition” matters. E5 does not claim that every conceivable kind of digital thermodynamics must use this exact architecture. It states exactly what must be true to earn the CAL-ENDO label. ## 3. What “endogenous” means here No system creates its own laws from literal nothing. Molecules do not invent collision laws, and this graph does not invent its alphabet or rewrite grammar. CAL-ENDO distinguishes three levels: 1. **State endogeneity.** Temperature is computed from the system's actual authoritative configuration space, not supplied as a number. 2. **Constitutive endogeneity.** The system's driven dynamics create and repair the multiplicity relation without being told a thermal target. 3. **Grammar endogeneity.** The system discovers the local construction law itself. The result establishes the first two. It does not establish the third. The finite alphabet, unique-parent rule, energy quantum, graph rewrites, and work resources are designed constitutive primitives. This is comparable to specifying the particles and Hamiltonian of a physical model before asking what macroscopic state variables arise. The causal firewall was explicit. Construction and repair could inspect local ports, graph state, descriptor inventory, work cells, action outcomes, and physical memory. They could not inspect temperature, target multiplicities, entropy slopes, Gibbs weights, or the held-out traffic later used to test the temperature. Static data-flow inspection and poison interventions both verified that separation. ## 4. Structural theorem ladder ### E1. Local graph law implies the multiplicity recursion For a canonical root, injective word labels, unique parentage, and exactly one executable child per local symbol in an alphabet of size `b`, words of length `m` are in bijection with shell `m`. Therefore $$ g_0=1, \qquad g_{m+1}=b g_m, \qquad g_m=b^m. $$ This is the central local-to-global step: a local port-completeness law creates a global energy--multiplicity relation. Within the regular labelled-prefix class, local port completeness is sufficient and necessary for that regular realization. The same shell histogram can occur in irregular graphs, so counts alone do not prove the prefix mechanism. ### E2. Constant temperature is equivalent to a constant shell ratio For a finite layered graph with equally spaced energies and at least two adjacent shell gaps, one positive constant state-count temperature exists if and only if there is one constant `r>1` such that $$ g_{m+1}=r g_m $$ throughout the admitted interval. Its value is $$ T_{\mathrm{state}}=\frac{\lambda}{\ln r}. $$ In the prefix witnesses, `r=b`. If ratios vary, the graph may instead have state-dependent local temperatures; fitting an average slope would not make the temperature exact. ### E3. Independent reciprocal traffic recovers the same temperature A separate probe body was coupled only after construction. Its microscopic exchange graph was sealed in advance, hidden from the builder, reciprocal, and strongly lumpable. Double-counting the same undirected microscopic edges gives the coarse hazard ratio $$ \frac{k_{x\to y}}{k_{y\to x}} = \frac{h_y g_R(E_{\mathrm{tot}}-U_y)} {h_x g_R(E_{\mathrm{tot}}-U_x)}. $$ When `g_R(lambda*m)=b^m`, this becomes local detailed balance at $$ T_{\mathrm{LDB}}=\frac{\lambda}{\ln b} =T_{\mathrm{state}}. $$ The test checked more than one rate ratio. It checked actual microstate fibers, reverse support, row closure, strong lumpability, the unique uniform microcanonical stationary law on complete states, its Gibbs body-macro projection, irreducibility, and relaxation. A finite contact construction independently produced zero prepared current at equal temperature and the preregistered nonzero hotter-to-colder prepared initial current away from equality. Both closed reciprocal contact fixtures have zero stationary current. ### E4. Creation and repair reduce to an exact finite-chain condition Let `M_therm` be the set of complete states passing the thermodynamic classifiers. In a finite construction or repair chain, the hitting time of `M_therm` is finite with probability one if and only if no reachable closed communicating class lies entirely outside `M_therm`. That criterion is both necessary and sufficient. The apparatus discharged it with exact finite rank/resource certificates, not by observing one long lucky trajectory. Because the admitted chain is finite, almost-sure hitting also implies finite mean hitting time. The same logic was applied after every frozen leaf defect. ### E5. Dynamic endogeneity admission criterion An adaptive finite construction is admitted as an endogenously thermodynamic Xypher if and only if it has all seven of the following: 1. independently declared energy, actual shell counts, and microscopic reverse pairing; 2. causal non-use of temperature, target counts, fitted traffic, and other forbidden answer channels; 3. a first creation time that is nonzero, almost surely finite, and therefore finite-mean from every declared nonthermal initial state; 4. invariance between perturbations and almost-sure finite repair after every declared repairable defect, with finite mean return time and named finite resources; 5. an executable perception--action--outcome--learning loop whose memory has a causal effect on later construction; 6. closed and typed heat, work, information, resource, and inverse-protocol ledgers; 7. independent held-out operation with unique `T_LDB=T_state`, plus correct contact behavior when contact is claimed. Conditions 1, 2, and 7 prevent circular fitting. Conditions 3 and 4 establish creation and maintenance rather than a static representation. Condition 5 makes the witness adaptive. Condition 6 prevents hidden memory, work, or irreversibility from being mistaken for thermodynamics. ## 5. Exact constructive witnesses The frozen primary family used `lambda=1`, alphabet sizes `b in {2,3,4}`, shells `m in {0,1,2,3}`, and work-cell quantum `epsilon_W=1`. Every initial graph was classified `UNDEFINED`: it did not yet have the required constant-temperature shell interval. The adaptive dynamics then produced: | Alphabet | Driven operations | Actual shell counts | Resulting temperature | Probe fiber sizes | Repair arms | |---:|---:|---|---|---|---:| | `b=2` | 31 | `1,2,4,8` | `1/ln(2)` | `8,8,2` | 8 | | `b=3` | 81 | `1,3,9,27` | `1/ln(3)` | `27,18,3` | 27 | | `b=4` | 171 | `1,4,16,64` | `1/ln(4)=1/(2 ln 2)` | `64,32,4` | 64 | For every witness: - the state classifier returned `UNIQUE_CONSTANT`; - the independent traffic classifier returned `UNIQUE_LDB_EQUILIBRIUM`; - `T_state` and `T_LDB` were symbolically equal; - every declared leaf deletion was repaired exactly; - all driven forward protocols had explicit exact inverse protocols; - all heat, work, resource, contact, and information accounts closed. The held-out thermal sector enumerated 18, 48, and 100 complete states and 200, 1,260, and 4,896 authoritative events respectively. Contact checked 512 equal-temperature rows and 64 unequal-temperature rows. The contact fixtures had a separate shared energy gauge: left reservoir `(b_1,lambda_1)=(2,1)`; equal-temperature right reservoir `(4,2)`; unequal-temperature right reservoir `(2,2)`; contact quantum `q=2`; shared total reservoir energy `4`; and prepared macro probabilities `p_A(0)=p_B(0)=1/2`. The equal preparation was already stationary. In the unequal preparation, `j_{A->B}(0)=2` and `J_E(0)=4`, so energy initially flowed from hotter reservoir 2 into colder reservoir 1; its final stationary current was zero. ## 6. Primary-gate result | Gate | Result | Decisive evidence | |---|---|---| | E01 complete state | **Pass** | authoritative replay reproduces driven, probe, contact, and the frozen `do(M)` post-intervention kernel | | E02 initially nonthermal | **Pass** | all three initial specimens are `UNDEFINED` | | E03 forbidden-input non-use | **Pass** | static source audit, rejected target/traffic payloads, invariant poison traces | | E04 Xypher causal loop | **Pass** | perception, integration, work enablement, graph action, outcome, and learning are executable | | E05 memory causal | **Pass** | frozen `do(M)` changes `P(FB}(0)=2` and `J_E(0)=4` into colder reservoir 1; both stationary currents `0` | | E14 controls | **Pass** | 22/22 controls received their frozen outcomes | | E15 no hidden fit | **Pass** | exact manifest and resource/source audit | No floating-point tolerance, fitted slope, sampled trajectory horizon, target count feedback, or post-result parameter change entered the result. The frozen `do(M)` comparison is an off-manifold structural-causal-model surgery, not an autonomous physical transition. The ordinary post-intervention kernel remains defined, replayable, and ledger-closed. Its role is to show that physical memory is causally relevant to the subsequent law, not to claim that the system spontaneously rewrites its own memory. ## 7. What the controls ruled out The controls were designed to make nearby but weaker constructions fail under their correct labels. - A prebuilt perfect tree was only operationally thermal; it did not create itself. - A nonadaptive builder failed the memory-causality gate. - The historical `a_P` gate deadlocked in the frozen naive construction. - An external thermostat and installed Gibbs weights were classified as rate-compatible or fitted representations, not endogenous thermodynamics. - Target-count and traffic-trained builders were rejected before the policy could inspect them. - Deleting a leaf broke the exact temperature until repair; disabling repair left the defect in place; withholding the required descriptor correctly made the resource premise false. - Mutating the port law produced counts `1,2,4,9` and a state-dependent rather than constant temperature. - Hiding memory made the projected process non-Markovian. - Ideal counters could reproduce logic but failed the physical finite-record boundary. - Omitting one ledger cell broke augmented first-law closure. - Doubling all microscopic clocks changed activity and transient currents but not temperature or stationary laws. - Breaking reverse support caused the independent traffic classifier to reject the mutant. These outcomes separate the full claim from static geometry, fitted rates, causal contamination, lucky repair, hidden resources, and activity alone. ## 8. Provenance and execution record ### 8.1 Frozen scientific boundary | Object | Revision or digest | |---|---| | boundary commit | `7ee278ed14ade12689540481903e4832f53442d7` | | boundary SHA-256 | `a3aafdc4d29dc3b6c8a2c5bf8deb554745a311b2f3485e1577f3697b1aa14c80` | | final apparatus commit | `17200442560fcb8fabbd24fe0fc4a1dea7a57abc` | | 13-source manifest SHA-256 | `d08bb07cf0d36c44e6c880c1f848080e6236e32e7ee817377039383346f27517` | | release binary SHA-256 | `a49b79225cb7ddf45e1fe8186d599d80e486fb14a4b886d2ca9712be9bbbd121` | | toolchain | `cargo 1.94.0`, `rustc 1.94.0`, `aarch64-apple-darwin` | The first prospective freeze was invalidated before execution by formatting, Futuruna syntax, and one Rust compile defect. A corrected freeze passed all 39 tests, but its complete report could not be produced within the available process/memory envelope. Those attempts emitted no scientific stdout and are recorded as `NO_RESULT`. A source-reviewed, semantics-preserving memory/transport refactor removed duplicate in-memory report corpora and streamed output directly. Its final source was independently re-reviewed, refrozen, committed, and pushed before execution. Debug and release suites both passed 40/40 tests. One plaintext top-level attempt on those final bytes was manually interrupted after the report reached 14,059,099,608 bytes; no content was inspected, and it is recorded as `NO_RESULT`. The unchanged binary was then rerun with its stdout passed through lossless `zstd` compression. Compression changed only transport, not the evaluator or report bytes. ### 8.2 Complete confirmatory artifact ```text command: cargo run --release --quiet --manifest-path Cargo.toml transport: stdout | zstd -1 -T0 pipeline exit: 0 verdict: OVERALL PASS classification: ENDOGENOUSLY THERMODYNAMIC ADAPTIVE XYPHER ``` | Artifact property | Exact value | |---|---| | raw report bytes | `14,940,547,191` | | compressed bytes | `72,263,592` | | raw report lines | `10,880` | | raw SHA-256 | `caf1bd441175444d1f0fdab2221af755048135fb47b8024c53c0bcf7bd8c7c56` | | compressed SHA-256 | `4f18196d88eb5d9571968a126f631bcfd6e413bb0912ce05b72087882c2433de` | | stderr SHA-256 | `87145a36670a91f0a65db7e4a9c5691d1956954b13ed525dd7fd7e7059878ec2` | The compressed artifact passed `zstd` integrity verification. Its decompressed line categories close exactly: 1 verifier header, 1 boundary record, 212 observations, 3 specimens, 15 gates, 22 controls, 57 information reports, 213 protocol headers, 213 initial-state records, 213 terminal-state records, 1,495 forward and 1,495 inverse traces, 6,359 thermal rows, 578 contact rows, 1 sentinel, 1 classification, and 1 final verdict. These sum to exactly 10,880 newline-terminated rows. Stderr contains only nine expected Rust naming warnings for the frozen symbolic state-coordinate names. The large raw report is deliberately not committed to Git. The compact result, hashes, frozen source, exact manifests, and deterministic evaluator are enough to reproduce and authenticate it without adding a multi-gigabyte repository object. ## 9. What this establishes Within the declared finite operational model: 1. At least one—and here three—adaptive digital graph constructions can begin nonthermal and use local graph-changing dynamics to create an exact global energy--multiplicity relation. 2. The actual enumerated relation independently defines a positive constant state temperature. 3. A sealed reciprocal sector that was not available to the builder recovers the same unique temperature and has the correct equilibrium and relaxation structure. 4. Finite contact has zero prepared equal-temperature current and the frozen hotter-to-colder unequal-temperature prepared transient; both closed reciprocal fixtures have zero stationary current. 5. The relation can be maintained between perturbations and restored after every declared finite repairable defect, using explicit finite work and descriptor resources. 6. Physical memory causally changes subsequent construction statistics, so the witnesses are adaptive rather than static graph grammars. 7. E1--E4 characterize the conditional structural, kinetic, and finite-chain mechanism; E5 packages them into the definition-relative admission criterion. This is stronger than a simulation that merely resembles thermodynamics. It is an exact executable mathematical existence proof inside a declared digital stochastic-thermodynamic model. ## 10. What this does not establish - It does not show that every historical Xypher candidate passes thermodynamic admission. - It does not show that the prefix grammar is the only mechanism. - The grammar, alphabet, energy quantum, and lawful rewrites are designed; the witness does not discover them from an unrestricted hypothesis space. - Repair is a finite driven episode with a named work store. It is not an indefinitely self-powered steady state or perpetual resistance to damage. - `Delta S_tau` is not identified with the microcanonical entropy used here. - `a_P` is not established as physical temperature. - TAU and Xi are not instantiated; TAU is not thereby shown to be heat, work, money, or an equation of state. - The result does not establish intelligence, consciousness, self-awareness, or a theory of the origin of temperature in the material universe. - No blockchain or payment-network implementation was tested. Such a system would need its own complete-state, reservoir, reciprocal-channel, contact, and ledger bridge. - This is not empirical evidence that running ordinary software gives its host hardware a new joule-valued thermodynamic degree of freedom. The scientifically warranted novelty is the exact Xypher-specific construction and characterization of endogenous **operational digital thermodynamics**, not a claim to have replaced or extended the fundamental laws of thermodynamics. ## 11. The next frontier The existence question posed by CAL-ENDO is now closed for the adopted finite class. The next irreducible scientific question is grammar endogeneity: > Under what generic adaptive local dynamics does a thermodynamic manifold > emerge as an attractor, without installing a prefix-port grammar that is > already known to make the required multiplicity relation reachable? That program can widen the initial graph class, weaken the local grammar, allow the alphabet or rewrite law to be learned, and test whether the energy--multiplicity relation is selected and stabilized rather than directly entailed by a designed invariant. A separate engineering frontier may then bridge the theorem to a blockchain payment network. It should not be conflated with the mathematical existence result proved here.