Quantitative norm paving over finite-index subfactors

Let N ⊂ M be an inclusion of II₁ factors of finite Jones index, and let E be the trace-preserving conditional expectation onto N′ ∩ M. To ε-pave a finite family, we seek a single projection partition (pi) of 1 in N satisfying ‖∑i pi(x − E(x))pi‖ ≤ ε‖x − E(x)‖ for each member x. For 0 < ε < 1 we prove the bound ⌈9ε−2⌉ + ⌈3ε−1⌉ on the number of pieces, uniformly in the index and the family size. At each fixed inclusion, the supremum n∞(ε) of the least paving counts over finite families satisfies 2 ≤ lim infε↓0 ε²n∞(ε) ≤ lim supε↓0 ε²n∞(ε) ≤ 4. Finite families of trace-zero self-adjoint unitaries in N give the lower bound. This determines the optimal finite-family exponent in Sorin Popa’s paving-size question from his preprint of 10 October 2022. For a single self-adjoint operator the worst count has order ε−1 for N = M and at index two, settling those cases of Popa’s single-operator question. Across all finite-index inclusions the worst count has order ε−2 even for individual self-adjoint unitaries with E(x) = 0. For trace-zero self-adjoint contractions in a single factor, the worst infimum error with at most r ≥ 2 pieces is 1/(r − 1). At index two the same value holds for trace-zero self-adjoint contractions whose square lies in N, and for those with at most two spectral values. Each operator in the latter class admits an r-piece paving with strictly smaller error, giving the exact worst count ⌈1 + ε−1⌉ for 0 < ε < 1. At index two the worst count is asymptotic to ε−1 for trace-zero self-adjoint contractions commuting with their image under the involution of M with fixed algebra N. GitHub repository: eulogica/quantitative-norm-paving-finite-index-subfactors.

Authors

Institutions

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-19
DOI
https://doi.org/10.5281/zenodo.24364040
Primary Topic
Advanced Operator Algebra Research
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Quantitative norm paving over finite-index subfactors

Martino Lupini
Zenodo (CERN European Organization for Nuclear Research)
Advanced Operator Algebra Research
preprint

Quantitative norm paving over finite-index subfactors

Martino Lupini
preprint en

Abstract

Let N ⊂ M be an inclusion of II₁ factors of finite Jones index, and let E be the trace-preserving conditional expectation onto N′ ∩ M. To ε-pave a finite family, we seek a single projection partition (pi) of 1 in N satisfying ‖∑i pi(x − E(x))pi‖ ≤ ε‖x − E(x)‖ for each member x. For 0 < ε < 1 we prove the bound ⌈9ε−2⌉ + ⌈3ε−1⌉ on the number of pieces, uniformly in the index and the family size. At each fixed inclusion, the supremum n∞(ε) of the least paving counts over finite families satisfies 2 ≤ lim infε↓0 ε²n∞(ε) ≤ lim supε↓0 ε²n∞(ε) ≤ 4. Finite families of trace-zero self-adjoint unitaries in N give the lower bound. This determines the optimal finite-family exponent in Sorin Popa’s paving-size question from his preprint of 10 October 2022. For a single self-adjoint operator the worst count has order ε−1 for N = M and at index two, settling those cases of Popa’s single-operator question. Across all finite-index inclusions the worst count has order ε−2 even for individual self-adjoint unitaries with E(x) = 0. For trace-zero self-adjoint contractions in a single factor, the worst infimum error with at most r ≥ 2 pieces is 1/(r − 1). At index two the same value holds for trace-zero self-adjoint contractions whose square lies in N, and for those with at most two spectral values. Each operator in the latter class admits an r-piece paving with strictly smaller error, giving the exact worst count ⌈1 + ε−1⌉ for 0 < ε < 1. At index two the worst count is asymptotic to ε−1 for trace-zero self-adjoint contractions commuting with their image under the involution of M with fixed algebra N. GitHub repository: eulogica/quantitative-norm-paving-finite-index-subfactors.

Zenodo (CERN European Organization for Nuclear Research)
University of Bologna (IT)
Advanced Operator Algebra Research
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