--- title: Adaptive Xypher Contact Result aliases: - CAL-ADAPT Result - Reversible Digital Memory Reservoir Result tags: - domain/physics - type/result - topic/thermodynamics - topic/xypher - topic/information-reservoir domain: Physics type: result status: exact-control-passed-native-law-open created: 2026-07-29 updated: 2026-07-29 related: - "[[Adaptive Xypher Contact Boundary]]" - "[[Q_H Thermodynamic Compatibility and Weak-Contact Frontier]]" - "[[Native Contact Identifiability for a Xypher Payment Substrate]]" --- # Adaptive Xypher contact result ## 0. Split verdict The exact CAL-ADAPT experiment matched every preregistered prediction. 1. **Append-only adaptive state is not a reversible equilibrium state.** A balance-reversing transfer does not reverse its receipt or monotone learning update. Hiding that state is valid only after an exact lumpability proof. 2. **A finite digital information reservoir can produce a nontrivial canonical subsystem law exactly.** The generated reversible causal-stack model passes microscopic reverse pairing, generator conservation, equilibrium detailed balance, strong lumpability, retained-rate no-quench, and exact current identities. 3. **This is not yet native Xypher thermodynamics.** The stack alphabet, reversible memory operations, and additive clocks are engineered protocol hypotheses. The current ledger and Praxor do not instantiate them. This is a positive constructibility result and a negative result for one invalid coarse-graining. It is not a derivation of temperature from $alpha$, heat from TAU, or an equation of state. ## 1. Frozen chronology The hypotheses and literal predictions were committed and pushed as `c4eb3af3042d03641bc70311c550c960f253276e` before the checker existed. Two independent reviews approved that preregistration. Implementation and execution then occurred in the independent crate `research/physics/xypher-calorimeter-analysis`. The final instrument was separately reviewed after adversarial hardening. It rejects malformed symbol endpoints, duplicate or non-unit channel instances, misclassified DTMC channels, and missing reverse IDs. It derives a quotient only after exact lumpability, freezes the full-support uniform reference instead of choosing an absorber-supported measure, and derives signed generator diagonals before checking exact zero row sums. ## 2. Exact DTMC boundary witnesses | Fixture | Exact outcome | Interpretation | |---|---|---| | Monotone two-state update | Rows normalize; one strict increase; reverse missing; only $z_1$ is recurrent; uniform reference is not stationary | A stationary delta on the absorber cannot be used as vacuous equilibrium evidence for the lost forward channel | | Apparent reciprocal support | Fiber-level support contains $A\to B$ and $B\to A$, but $A_0$ and $A_1$ give next-$B$ probabilities $1$ and $0$ | No autonomous balance kernel exists; projected detailed balance is `NotApplicable` | | Lumpable hidden-label control | Exact quotient $A\leftrightarrow B$; uniform law reversible; two closed full-state components | Hidden labels can be harmless when they do not change balance kinetics | | Reversible-memory square | Twelve labeled channels at probability $1/3$; one closed class; uniform law; zero currents; $Q_{\rm cross}=1/3$ and $Q_{\rm self}=2/3$ | Explicit write and erase channels restore a valid equilibrium model | | Separate `MA-` / `MB-` ablations | Row normalization and projected detailed balance still pass; full reverse support and full detailed balance fail | A reversible quotient does not make the full dynamics reversible and cannot substitute for a full-state check | The result does not prohibit learning. It requires learning, reset, erasure, clock, and receipt state to be modeled at the level where thermodynamic claims are made. ## 3. Exact causal-stack result For alphabet size $r\ge2$ and stack depth $D\ge1$, the generated microstates are $$ \mathcal Z_{r,D} = \{(n,s):0\le n\le D,\ s\in\Sigma_r^{D-n}\}. $$ The checker generated every unit-rate `REPLACE`, `POP`, and reverse `APPEND` channel. It derived the signed generator diagonal as $$ G_{zz}=-\sum_{z'\ne z}G_{zz'} $$ and verified every row sum exactly zero. The contacted graph is connected and symmetric, so the full stationary law is uniform. Counting the reservoir states gives $$ \Omega(n)=r^{D-n}, \qquad \pi(n)=\frac{r^{D-n}}{\sum_{j=0}^{D}r^{D-j}}, \qquad e^{-\beta\epsilon}=\frac1r. $$ The factor $1/r$ was not passed to the transition law. It followed from the enumerated density of states. The absolute energy quantum $\epsilon$ remains independently declared. ### Primary fixture: $r=2,D=2$ - seven microstates; - ten directed internal replacements; - six `POP` and six separately labeled reverse `APPEND` channels; - twenty-two contacted off-diagonal channel instances; - level exit rates $(3,4,2)$; - macrostate law $(4/7,2/7,1/7)$; - adjacent factor $1/2$ and two-gap consistency factor $1/4$; - macro rates $k_{n\to n+1}=1$ and $k_{n+1\to n}=2$; - low-subsystem-energy currents $(1,0)$; and - high-subsystem-energy currents $(0,-2)$. Every equilibrium boundary current is exactly zero. ### Preregistered extrapolation: $r=3,D=2$ - thirteen microstates; - forty-two replacements; - twelve `POP` and twelve `APPEND` channels; - sixty-six contacted off-diagonal instances; - level exit rates $(5,6,3)$; and - macrostate law $(9/13,3/13,1/13)$ with adjacent factor $1/3$. ### Metamorphic anti-hardcoding case The same constructor generated $r=2,D=3$ without a literal state table: fifteen states, thirty-four replacements, fourteen channels in each contact direction, sixty-two contacted instances, macro multiplicities $(8,4,2,1)$, and level exit rates $(4,5,4,2)$. A separate $r=4,D=1$ case also passed. ## 4. What this establishes Under the declared finite-state stochastic-thermodynamic model, a purely digital system can have all of the following simultaneously: - a conserved total energy lattice; - a counted reservoir density of states; - exact microscopic reverse channels; - a connected continuous-time generator; - exact equilibrium detailed balance; - a canonical subsystem marginal with a nontrivial fixed Boltzmann factor; - signed relaxation currents that vanish at the predicted law; and - additive contact clocks that do not quench retained internal hazards. That is a genuine exact thermodynamic construction in the standard statistical-mechanical sense. It is not new fundamental physics: exponential reservoir degeneracy and information reservoirs are established theory. The scientific contribution here is the exact application, falsification boundary, and executable admission gate for the Xypher architecture. ## 5. What remains open The current Xypher has not yet supplied the construction's load-bearing physics: - the Crystal does not yet force an $r$-ary reversible reservoir alphabet; - the live Praxor does not emit ordinary autonomous transfers; - the ledger appends receipts without a reverse receipt channel; - actor-local learning state lacks declared reset-reservoir accounting; - no protocol rule owns additive elementary transfer clocks; and - no independently derived energy scale identifies $alpha$ as temperature or TAU as heat/work. Therefore this experiment does not establish $\mathrm{TAU}=\alpha S_\tau$ as an equation of state or demonstrate a native blockchain thermodynamic phase. ## 6. Next native candidate A post-result audit found a mathematically clean graph-transfer family worth freezing next. For fixed inventory $K$, $$ \lambda_{i\to j}(n)=\kappa_{ij}u(n_i). $$ If $w_i\kappa_{ij}=w_j\kappa_{ji}$ and $g(m)=\prod_{q=1}^{m}u(q)$, its exact reversible occupancy law is $$ \pi_K(n) \propto \prod_i\frac{w_i^{n_i}}{g(n_i)}. $$ For $u(n)=n$, this is the multinomial law of $K$ independently clocked labeled units. It supplies reciprocal graph kinetics without a Gibbs sampler, but the current protocol does not yet justify one independent clock per `MicroTau`-edge or choose between additive edge clocks and normalized node clocks. Those two interpretations have different equilibria and different no-quench behavior. The next frontier is therefore not another long trajectory. It is an exact protocol-ownership test: derive the elementary clock and conductance from Crystal / Praxor semantics, or record that they are a newly engineered mechanism before executing ledger transfers. ## 7. Verification - focused public CAL-ADAPT integration tests: 4 passed; - adversarial CAL-ADAPT unit tests: 7 passed; - complete small-crate suite: 72 passed; - `cargo clippy --all-targets -- -D warnings`: passed; and - exact CAL-ADAPT execution time after build: approximately $0.01$ seconds. No trajectory dataset or large result directory is produced.