Back to search
2507.17643

Arithmetic Degrees are Cohomological Lyapunov Multipliers

Jiarui Song, Junyi Xie, She Yang

correcthigh confidence
Category
Not specified
Journal tier
Top Field-Leading
Processed
Sep 28, 2025, 12:57 AM

Audit review

The paper proves precisely the statement the model labeled as likely open: for a surjective endomorphism f of a normal projective variety X and a point x with Zariski-dense orbit, the arithmetic degree satisfies α_f(x) ∈ {µ_i(f)} ∩ R_{≥1} (Theorem 1.2) . The proof works over finitely generated fields using Moriwaki heights, and it leverages: (i) functoriality through the Albanese to obtain α_f(x) ≥ λ_1(g) for the induced endomorphism g on Alb(X) and to dispatch the case α_f(x)=λ_1(g) via the inclusion of cohomological Lyapunov multipliers under semi-conjugacy ; (ii) a lifting lemma that splits off N^1(X)_R-eigen-directions into Pic(X)_R above the spectral radius on Pic^0(X) ; (iii) canonical height vectors for Jordan blocks to control growth from the big part of N^1(X)_R , ; (iv) an explicit polynomial–exponential growth bound for lower-rate components (Corollary 3.3) to contradict any α strictly between adjacent µ_i’s . The paper also recalls the monotonicity and functoriality of µ_i(f) and the standard upper bound α_f(x) ≤ λ_1(f) in this setting, along with the Moriwaki-height normalization α_f(x) ≥ 1, tying all ingredients together , , . In short, the paper establishes the general membership result; the model’s claim that no such proof is known is outdated relative to this paper.

Referee report (LaTeX)

\textbf{Recommendation:} minor revisions

\textbf{Journal Tier:} top field-leading

\textbf{Justification:}

This paper proves that arithmetic degrees along Zariski-dense orbits coincide with cohomological Lyapunov multipliers, resolving a central discreteness question and yielding a spectral-gap case of KSC. The approach is conceptually clear and technically robust: it combines Albanese functoriality, a carefully constructed lifting from N\^1 to Pic above the Pic\^0-spectral radius, canonical heights for Jordan blocks, and a linear-recurrence growth bound for heights. The result is of high significance for arithmetic dynamics. Minor clarifications (dependencies and an expository expansion on growth bounds) would further improve accessibility.