Divided differences and complex variations of multiple zeta-star values

arXiv:2606.16627 · math.NT · Submitted 2026-06-15 · Read on arXiv

math.NT

Submitted: 2026-06-15

Updated: 2026-11-02

Comments: 48 pages

License: http://creativecommons.org/licenses/by/4.0/

The gist: Motivated by the known derived set [1,+ infinity) of multiple zeta-star values, we study finite multiple star harmonic sums and their infinite-depth limits.

Terminology

Abstract

Motivated by the known derived set [1,+ infinity) of multiple zeta-star values, we study finite multiple star harmonic sums and their infinite-depth limits. Divided differences give a cube-integral representation of their holomorphic interpolation on Re(s)>0, which agrees with the Kawashima functions after reversal of the index. The finite-depth interpolations converge locally uniformly in s as the depth tends to infinity, uniformly over all positive integer index sequences on compact subsets. For real s>1, these limits define a bijection from the infinite positive integer indices onto (1,s]. For each admissible infinite index, they tend to the corresponding multiple zeta-star limit as Re(s) to+ infinity with bounded imaginary part. For Re(s)>1, we prove strict separation of every ordered pair of finite indices and of every ordered pair of infinite indices: either the difference or the difference divided by s has positive real part throughout this half-plane, with the choice fixed by the pair. In particular, both index maps are injective. The proof uses a summation principle for functions with positive real part, established through positive integral representations and explicit kernel estimates.

Sources

Related papers