2206.01091
RANDOM AND MEAN LYAPUNOV EXPONENTS FOR GLn(R)
Diego Armentano, Gautam Chinta, Siddhartha Sahi, Michael Shub
correctmedium confidenceCounterexample detected
- Category
- math.DS
- Journal tier
- Strong Field
- Processed
- Sep 28, 2025, 12:56 AM
- arXiv Links
- Abstract ↗PDF ↗
Audit review
The counterexample uses a measure µ supported on orthogonal conjugates of a fixed shear, i.e., µ is O(n)-conjugation-invariant. The paper’s Theorem 1 is proved for “orthogonally invariant” measures in the sense used in their argument: invariance under the left O(n)-action (indeed, they disintegrate along left O(n)-orbits and apply an inequality that integrates over U with A replaced by UA). This orbit-by-orbit step, and the fact that |det A|g| is constant on each left O(n)-orbit, are explicit in Theorem 2 and the proof of Theorem 3, and they do not hold for conjugation orbits. Hence the candidate’s measure does not satisfy the paper’s hypotheses, so no contradiction is obtained. See Theorem 1 and its reduction via Theorems 2–3 and the orbit argument .
Referee report (LaTeX)
\textbf{Recommendation:} minor revisions \textbf{Journal Tier:} strong field \textbf{Justification:} The paper successfully extends comparisons between mean log-eigenvalues and Lyapunov exponents to the real case with a universal constant, employing a refined geometric-analytic approach. The main contribution is both technically sound and conceptually interesting. A small clarification of the precise invariance hypothesis would preempt misreadings and strengthen accessibility.