An arXiv preprint dated 26 Aug 2026 presents a mathematical framework that ties together three ways of tracking arithmetic volume and the measures built from spaces of sections. Its main comparisons say that the L2, supremum and theta-based formulations of arithmetic volume coincide, while normalized theta distortion and a Bergman-type distortion have the same limsup when the latter is applied to a specified truncated section module. The analysis concerns abstract arithmetic-geometric objects and asymptotic behavior.
A shared language for abstract arithmetic objects
The basic setting is a smooth projective arithmetic variety over Spec(Z), of dimension n + 1, equipped with a continuous Hermitian line bundle and a smooth volume form. The volume form gives the global sections a Euclidean lattice structure, providing the framework for comparing different norms and distortion quantities.
Two distortions, one asymptotic picture
Two central quantities are an arithmetic distortion called theta and an L2-versus-supremum distortion called rho. The paper states a pointwise inequality in which theta is no larger than rho. To control sections with large Euclidean norm in the volume estimates, the analysis also uses a second-moment Gaussian tail estimate for Euclidean lattices.
On the theorem's truncated section submodule, the normalized theta and rho measures have the same limsup. In ordinary terms, after the prescribed normalization and restriction to controlled sections, the two distortion constructions become indistinguishable at the level of their limiting upper behavior. The qualification matters: this is not a general pointwise limit in every setting.
From section growth to positivity
That comparison feeds into the paper's treatment of arithmetic volume. Under the stated smooth volume-form and continuous-metric assumptions, its L2, supremum and theta-based asymptotic formulations all give the same volume. The paper also gives a weaker integral representation, expressing arithmetic volume through normalized theta distortion while the metric is shifted by weights up to the asymptotic maximal slope. Because the representation uses a limsup, it is weaker than a fully established limit formula.
The framework gives the distortion measure a geometric interpretation under specific positivity assumptions. For a smooth weakly nef Hermitian line bundle with an ample underlying line bundle, the asymptotic theta measure converges to the classical equilibrium measure, with the upper and lower forms identified with that same measure. The smoothness, weak nefness and ampleness conditions are part of the result.
The equilibrium measure is also used as a test for the paper's notion of arithmetic bigness. In the continuous Hermitian setting and under the specified perturbations, bigness is equivalent to a uniform positive lower bound on the lower arithmetic equilibrium measure when sufficiently small constant weights are added. A related variational statement says that, under weak nefness and ampleness, the right directional derivative of arithmetic volume in a nonnegative smooth direction is the integral of that perturbing weight against the equilibrium measure.
Sharper results in special settings
Several sharper statements are restricted to toric geometry. In the smooth toric setting, the full-lattice arithmetic equilibrium measure agrees asymptotically with the measure generated by sections that are small in the supremum norm. For smooth toric nef Hermitian line bundles, normalized arithmetic distortion converges weakly to the equilibrium measure, and the total mass of that measure gives the volume derivative.
In the positive toric case, when the first Chern form is positive, the normalized distortion measure converges weakly to a measure restricted to a theorem-defined subset U, described through the paper's Legendre-Fenchel construction. The concentration result locates where the asymptotic mass sits, but only within that positive toric setting.
The framework is finally connected to an equidistribution statement in arithmetic dynamics. Under the paper's stated algebraic dynamical assumptions, integrals over Galois-orbit measures of a generic small sequence converge, for every continuous test function, to the corresponding normalized integral against the equilibrium measure. This is an application in a specified dynamical system, rather than a general equidistribution theorem for unrelated settings.
The general case is still open
The broadest claim remains open. The document conjectures that the limsup used to define the arithmetic equilibrium invariant is actually a true limit and does not depend on the auxiliary volume form. Stronger volume representations and variational extensions remain tied to the assumptions used in the theorems, including smoothness, positivity, nefness, ampleness or toric structure. The preprint therefore presents a coherent mathematical program with several established links, while its full generality is still to be determined.
Paper data and sources
Original title: Arithmetic theta invariants and arithmetic equilibrium measures
Authors: Mounir Hajli
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-26
DOI: Not available
Original paper · Full text