Entropy as Logarithmic Support Growth

Entropy comparisons depend on the reference distribution and the resolution of observation. In practical measurements, these choices determine which outcomes an instrument can distinguish and how much information its records provide about them. Making both choices explicit allows meaningful comparisons between instruments, datasets, or physical states measured at different resolutions. We compare the grid cells occupied by a smaller set with those occupied by a larger set containing it. Both sets are placed on a common grid, and each occupied cell is assigned equal probability within its set. Theorem 1 shows that the relative entropy of the smaller set’s cell distribution with respect to the larger set’s cell distribution equals the logarithm of the larger set’s occupied-cell count divided by the smaller set’s count. This measures the information gained by learning that an outcome lies in the smaller set. We call this quantity logarithmic support growth. An increase in area and an increase in dimension can produce the same cell-count ratio at one resolution. Under the theorem’s assumptions, refining the grid distinguishes them. For full-dimensional sets with positive volume and boundaries of measure zero, the logarithmic cell-count ratio converges to the logarithm of the volume ratio. For sets with different box dimensions, its leading growth is proportional to log(1/h), where h is the grid spacing, with a coefficient equal to the difference in dimension. We use two relative-entropy decompositions to describe the effects of restricting a distribution to a set and recording only which partition cell contains an outcome. We apply these decompositions to Shannon entropy, thermodynamic entropy differences, kinetic theory at finite resolution, and the scarcity of prime numbers among the integers. For prime numbers, we derive explicit bounds and identify a range of covering resolutions from which the twin-prime count can be recovered exactly. Across these applications, specifying the reference and resolution makes the entropy comparisons precise and computable.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-18
DOI
https://doi.org/10.5281/zenodo.24332038
Primary Topic
Statistical Mechanics and Entropy
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Entropy as Logarithmic Support Growth

Kevin Leon Merdy
1 links
Zenodo (CERN European Organization for Nuclear Research)
Statistical Mechanics and Entropy
preprint

Entropy as Logarithmic Support Growth

Kevin Leon Merdy
preprint en

Abstract

Entropy comparisons depend on the reference distribution and the resolution of observation. In practical measurements, these choices determine which outcomes an instrument can distinguish and how much information its records provide about them. Making both choices explicit allows meaningful comparisons between instruments, datasets, or physical states measured at different resolutions. We compare the grid cells occupied by a smaller set with those occupied by a larger set containing it. Both sets are placed on a common grid, and each occupied cell is assigned equal probability within its set. Theorem 1 shows that the relative entropy of the smaller set’s cell distribution with respect to the larger set’s cell distribution equals the logarithm of the larger set’s occupied-cell count divided by the smaller set’s count. This measures the information gained by learning that an outcome lies in the smaller set. We call this quantity logarithmic support growth. An increase in area and an increase in dimension can produce the same cell-count ratio at one resolution. Under the theorem’s assumptions, refining the grid distinguishes them. For full-dimensional sets with positive volume and boundaries of measure zero, the logarithmic cell-count ratio converges to the logarithm of the volume ratio. For sets with different box dimensions, its leading growth is proportional to log(1/h), where h is the grid spacing, with a coefficient equal to the difference in dimension. We use two relative-entropy decompositions to describe the effects of restricting a distribution to a set and recording only which partition cell contains an outcome. We apply these decompositions to Shannon entropy, thermodynamic entropy differences, kinetic theory at finite resolution, and the scarcity of prime numbers among the integers. For prime numbers, we derive explicit bounds and identify a range of covering resolutions from which the twin-prime count can be recovered exactly. Across these applications, specifying the reference and resolution makes the entropy comparisons precise and computable.

Zenodo (CERN European Organization for Nuclear Research)
Statistical Mechanics and Entropy
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon