Code, data, figures and manuscript for
R. Chen, How Much Infinity Does Physics Need? Reverse Mathematics of Physical Limits and the Empirical Boundary (2026). DOI: 10.5281/zenodo.23212848
- Four equivalences over RCA₀. Each is equivalent to arithmetical comprehension ACA₀:
- Fekete's lemma for subadditive sequences (existence of thermodynamic limits);
- the existence of the infrared limit of every bounded monotone c-function;
- the existence of the scaling exponent of every Cantor construction with nondecreasing dimension parameters;
- the existence of the set of indices whose computable fractal limit sets are connected.
- Individual quantities. A limit with a computable modulus that RCA₀ verifies is provably existent in RCA₀; a limit that is not a computable real is not, because the ω-model REC contains no real equal to it.
- Moduli in the series. The Ising free energy on an L×L torus has error 0.639913/L² at K_c, matching the Ising CFT torus amplitude ln(θ₂+θ₃+θ₄)/2η = 0.639912; the diamond-lattice critical-cluster mass converges geometrically with ratio Λ₂/Λ₁ = 0.3553; the embezzlement asymptotics has error O(1/(n ln n)). The Ω approximations have no computable modulus.
- The empirical boundary. The limits that entered explicit certification budgets are provable in RCA₀; those shown to be empirically inaccessible (Ω-dimensions, limit connectivity, finite-dimensional game values) require ACA₀. Since ACA₀ is conservative over Peano arithmetic, the infinity physical limits require is real but logically mild.
| Path | Contents |
|---|---|
paper/ |
LaTeX source and compiled PDF of the manuscript |
code/reverse_math_moduli.py |
Script producing every number and the figure in the paper |
figures/ |
Figure as vector PDF (used by the paper) and PNG |
data/ |
Numerical data as CSV (metadata in # header lines) |
results/ |
Console output of the script (the numbers quoted in the paper) |
Requirements: Python ≥ 3.10 and the packages in requirements.txt
(tested with Python 3.12.4, NumPy 1.26.4, SciPy 1.13.1, Matplotlib 3.8.4). Runtime is about 15 seconds.
pip install -r requirements.txt
python code/reverse_math_moduli.py # add --show to display the figureRun from the repository root; the figure goes to figures/ and data to data/ (override with RM_FIG_DIR,
RM_DATA_DIR).
| Data file | Content | Paper |
|---|---|---|
ising_free_energy_error.csv |
free-energy error on the L×L torus at K_c and 0.8 K_c | §4, Fig. 1 |
dhl_mass_convergence.csv |
convergence of the normalised critical-cluster mass | §4, Fig. 1 |
embezzlement_error.csv |
error of the embezzlement asymptotics | §4, Fig. 1 |
omega_certified_interval.csv |
certified lower and upper bounds for Ω of binary lambda calculus | §4, Fig. 1 |
To rebuild the paper (pdfLaTeX, two passes):
cd paper
pdflatex Chen_2026_Reverse_Math_Physics.tex
pdflatex Chen_2026_Reverse_Math_Physics.tex@misc{Chen2026ReverseMathPhysics,
author = {Chen, Ruqing},
title = {How Much Infinity Does Physics Need? Reverse Mathematics of Physical Limits and the Empirical Boundary},
year = {2026},
doi = {10.5281/zenodo.23212848},
url = {https://doi.org/10.5281/zenodo.23212848}
}- Code (
code/): MIT License - Manuscript, figures, data and results (
paper/,figures/,data/,results/): CC BY 4.0
Ruqing Chen — GUT Geoservice Inc., Montreal — ruqing@hotmail.com