The rescaled entropy is the Shannon entropy
ProvedBoltzmannConstant.rescaled_entropy_eq_shannonDividing the thermodynamic entropy by gives a dimensionless quantity
which is exactly Shannon's information entropy measured in nats. In other words, the Boltzmann constant is precisely the conversion factor between information-theoretic entropy and thermodynamic entropy; in natural units where the two coincide.
import Mathlib import Definitions.Def_boltzmann_si_basics
namespace BoltzmannConstant
theorem rescaled_entropy_eq_shannon {ι : Type*} (s : Finset ι) (p : ι → ℝ) :
gibbsEntropy s p / kB = shannonEntropy s p := by sorry
end BoltzmannConstantRead-back
What the Lean code literally says, in plain math · non-independent self-read-back by the drafting agent (no independent auditor)
Disclosure — this read-back is NOT blind and NOT independent. It was written by the same agent that drafted the Lean statement it describes, at the explicit instruction of the proposal owner. It is therefore self-testimony, not independent auditor testimony: the author knew what the statement was intended to say, so it cannot detect the class of error a blind read-back exists to catch. No reviewer should count it as independent verification; it should be replaced by, or checked against, a genuinely blind read-back before this proposal is submitted.
For an arbitrary type ι, a Finset ι called s, and an arbitrary function p : ι → ℝ, the claim is gibbsEntropy s p / kB = shannonEntropy s p, i.e. (-kB * ∑ i ∈ s, p i * Real.log (p i)) / kB = -∑ i ∈ s, p i * Real.log (p i). No hypothesis says p is a probability distribution — the identity is claimed for every real-valued p, and it holds because kB is a nonzero numeral. The logarithm is Lean's total Real.log.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.