Pith. sign in
module module moderate

IndisputableMonolith.Verification.ProbabilityNormalizationCert

show as:
view Lean formalization →

Verification module that turns recognition costs into a discrete probability distribution and certifies the elementary axioms. It normalizes exponential weights, proves non-negativity, the unit upper bound, and sum-to-one, then bundles those facts. Anyone auditing RS measurement statistics or paper exports of the lightweight Born rule would cite it. The argument is a short positivity-and-summation stack over real exponentials, not a deep derivation.

claimFrom recognition costs $C_i$, form $p_i=e^{-C_i}/\sum_j e^{-C_j}$. The module proves $p_i\ge 0$, $p_i\le 1$, and $\sum_i p_i=1$, and packages these facts as a probability-normalization certificate over the lightweight Born-rule cost-to-weight map.

background

Recognition Science maps measurement alternatives to recognition costs; the induced statistics are obtained by exponential reweighting. The upstream lightweight Born-rule module supplies a minimal algebraic cost-to-weight map intended for paper exports, deliberately avoiding heavy analysis dependencies.

This verification module sits on that map. It introduces the normalized probabilities $p_i=e^{-C_i}/Z$ with partition sum $Z=\sum_j e^{-C_j}$, records that $Z>0$ (hence nonzero), and states the standard distribution inequalities. Sibling names in the module mark the pieces: the cost-to-probability map, positivity and nonvanishing of the exponential sum, non-negativity and unit bound for each $p_i$, the sum-to-one identity, and the bundled certificate.

proof idea

Short verification stack, not a deep proof. Real exponentials are positive, so each weight and the partition sum are positive; the sum is therefore a valid nonzero denominator. Normalized coordinates are consequently well-defined, nonnegative, and at most one. Sum-to-one is immediate from the definition of $Z$. Those lemmas are collected into the certificate structure. No forcing-chain or phi-ladder reasoning appears.

why it matters in Recognition Science

The import graph shows no downstream consumers inside the monolith, so the module functions as a leaf audit artifact for external checks and paper exports. It closes the elementary side of the story begun in the lightweight Born-rule module: once costs are given, the induced weights form a genuine discrete probability distribution. That underwrites any claim that RS measurement statistics are probabilistically well-formed without reopening real-analysis obligations in the core development. It does not itself advance the T0–T8 forcing chain, RCL, or the mass ladder.

scope and limits

depends on (1)

Lean names referenced from this declaration's body.

declarations in this module (7)