slug: wisniewska-2025-gurland-remainder
bibkey: wisniewska2025gurland
doi: null
url: https://arxiv.org/html/2512.07028v1
triage: theorem
motivation_gids:
  - D5/S3/Analytic/SeriesInequalities/GurlandLogRemainder.result

Gurland’s original logarithmic remainder

Problem

Halina Wiśniewska, Some inequalities for Gurland’s ratio of the gamma functions, arXiv:2512.07028v1, section 2, Open Problem 1:

In Corollary 1 we showed that R_n^m(x,y) tends to zero under a sufficient condition. Does R_n^m(x,y) tend to zero hold for every x,y > 0?

For positive real x,y put a=(x+y)/2 and d=(x-y)/2. In equations (2)–(4),

$$G_{\ast }=\frac{\Gamma (1+x)\Gamma (1+y)}{\Gamma (1+a{)}^2},H(s,b)=\sum _{j=0}^{\infty }\frac{1}{(j+b{)}^s},$$ $$S_m=\sum _{k=1}^{m-1}\frac{d^{2k}H(2k,1+a)}{k},R_m=\log G_{\ast }-S_m.$$

The expansion order m tends to infinity through integers, with m >= 2 in the source. Its n is already summed out. Every positive real pair, including x=y and positive integer Gamma arguments, is in scope.

Motivation

D5/S3/Analytic/SeriesInequalities/GurlandLogRemainder.result has type claim. The closed proposition claim is precisely the all-positive limit above. The source’s sufficient condition Q(x,y)<1 is unnecessary.

Gap

Preregistration #18508 binds the source, tier-1 question, full quantifiers and bounded literature scope. The source year is 2025. The inspected versioned primary poses the question. The bounded bibliographic input covers related Crossref titles and Merkle 2005, DOI 10.1016/j.camwa.2004.01.016, metadata. No exact prior settlement was established in that bounded scope. Merkle’s full mathematical text remains unverified. This is no priority claim or certification that earlier work lacks the identity.

Route

For n >= 0 put c_n=d^2/(n+1+a)^2. The identity

$$1-c_n=\frac{(n+1+x)(n+1+y)}{(n+1+a{)}^2}$$

and positivity give 0 <= c_n < 1. In the real Euler logarithmic sequence at 1+x, 1+y and 1+a, the factorial and log(N) contributions cancel. Consequently the sum of -log(1-c_n) is exactly log(G_*). This uses the all-positive real Gamma limit and retains integer arguments.

The logarithmic power series has nonnegative terms c_n^(k+1)/(k+1). Its joint summability follows from the summable logarithmic row sums. The source Hurwitz columns are separately summable by p-series comparison. Exchanging the sums identifies the coefficient series with log(G_*). Finite interchange and the index k+1 identify its partial sums with the source S_m. Their convergence proves R_m tends to zero.

Falsifier

A mismatch in the Gamma shift, the real Hurwitz shift, the exponent 2k, the finite interval 1 <= k < m, or the limit parameter would invalidate the source correspondence. These quantities occur explicitly in the formal definitions and in the finite Euler and interchange identities. No arbitrary Taylor-point family is substituted for the source remainder.

Evidence

The formal module contains sourceHurwitz, gammaRatio, partialExpansion, remainder, claim and result : claim, with only helpers used in their proof. The finite Euler cancellation, positive-shift summability, finite interchange, sum exchange and diagonal calculation are kernel proof terms. Classical summation and Gamma theorems are reused from the pinned library. The mathematical result introduces no axiom, sorry, admit, native_decide or certificate-trust assumption. Its axiom closure and current build evidence are supplied by the normal Lean report; this prose is not a second machine status.

Triage

What the settlement shows

The mechanism is the exact nonnegative logarithmic double series, whose row sums are fixed by the all-positive Gamma Euler limit. The theorem is pointwise on the entire positive quadrant; no global uniform rate or stronger quantitative remainder estimate is asserted. At x=y, G_*=1 and every S_m and R_m equals zero. This branch is included directly.

The equivalent identity in Open Problem 2 has zero difference and belongs to the same family, rather than a second problem settlement. Open Problem 3, concerning the location of t(x,y), is unresolved. The source’s proved Q<1 special case remains a special case of the whole-domain result. The diagonal uses equation (4) and never asserts existence in the empty interval (0,c_n). Source-selected Taylor points off the diagonal must satisfy equation (14); no such points are exposed by the native claim.

ASSUMED-UNVERIFIED

Exact overlap with Merkle’s full text and wider literature priority remain unverified. Formal correctness of this source statement does not settle priority, independent acceptance, integration or a solved-problem count.