Alphabeta Math
LemmaStatement: Literature-sourcedProof: AI-generatedPipeline-generatedprecheck passverified 2026-11-05 (gpt-6.1-sol)
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

Absolute exponential-polynomial expansions with real exponents

Statement

Let D⊆R, N≥0, an,j∈C, and Pn(s)=∑j=0Nan,jsj. Suppose the family ∑n∈D∣ensPn(s)∣ is summable for every s in a nonempty complex open set U. Summability of an arbitrary family means that the supremum of its finite subsums is finite. Then, locally on U, the expanded family ∑n,jan,jenssj and all its termwise derivatives converge absolutely and uniformly on compact sets. If its sum is zero on a nonempty complex open subset of U, every coefficient an,j is zero. The exponents need not be discrete, bounded or contained in finitely many integer cosets.

Facts & Assumptions

Given: Exactly the data and pointwise absolute convergence in the Statement. Elementary finite-dimensional polynomial interpolation, Cauchy's integral formula, the holomorphic identity theorem, and the binomial formula are used with their explicit arguments below.

Proof

1.1givenalgebra

Fix s0∈U and choose ϵ>0 such that the closed rectangle ∣Re⁡(s−s0)∣≤2ϵ, ∣Im⁡(s−s0)∣≤2ϵ lies in U. On each of the two vertical lines with real parts c−=Re⁡s0−ϵ and c+=Re⁡s0+ϵ, choose N+1 distinct points s±,l in this rectangle. Polynomial interpolation has a fixed inverse Vandermonde matrix at these points, so for a finite constant C independent of n, ∣an,j∣enc±≤C∑l=0N∣ens±,lPn(s±,l)∣. The right sides have finite sums over n by hypothesis. Hence ∑n∣an,j∣enc±<∞ for every j. In particular all nonzero coefficients have countable support: for a summable nonnegative family, the sets of entries at least 1/k are finite, and their union contains its positive support. Interpolation shows the nonzero Pn have the same countability property.

2.1step 1.1algebra

On the smaller rectangle ∣Re⁡(s−s0)∣≤ϵ/2, ∣Im⁡(s−s0)∣≤ϵ/2, one has enRe⁡s≤enc+ for n≥0 and at most enc− for n<0. The finitely many powers sj are uniformly bounded there. Thus step 1.1 gives a summable majorant independent of s, proving absolute uniform convergence. For every integer r≥0, the bound ∣n∣re−∣n∣ϵ/2≤Cr,ϵ bounds the derivatives of enssj by the same endpoint coefficient sums on a slightly smaller rectangle. Every derivative is a finite combination of these terms. This proves normal convergence of all derivatives, holomorphy of the sum by integrating uniformly convergent sums over small circles, and legitimate termwise differentiation.

3.1givenstep 1.1step 2.1algebra

Suppose the sum is zero on a nonempty open subset, and fix an arbitrary n0∈D. Choose real c2<c1 through a rectangle contained in that zero set. Step 1.1 gives finite sums ∑n,j∣an,j∣encl for l=1,2. Define F+(s)=∑n≥n0e(n−n0)sPn(s),F−(s)=−∑n<n0e(n−n0)sPn(s). The first series is normally convergent on Re⁡s<c1, using its nonnegative exponents and the coefficient sum at c1; the second is normally convergent on Re⁡s>c2, using its negative exponents and the coefficient sum at c2. Their equality on the chosen rectangle follows from the zero sum. The identity theorem on the connected overlap strip gives equality everywhere on that strip. Hence they glue to one entire function F. On the first half-plane the finite constant coefficient sums at c1 bound ∣F(s)∣ by C1∑j=0N∣s∣j; on the second the analogous sums at c2 give the same bound with C2. These two half-planes cover C, so ∣F(s)∣≤C(1+∣s∣)N globally. Cauchy's formula on circles of radius R shows ∣F(k)(0)∣/k!≤C(1+R)N/Rk→0 for every k>N. Thus F is a polynomial of degree at most N.

4.1step 3.1algebra∎

Along the positive real axis use F−, divide by sN, and let s→+∞. Every summand tends to zero, since its exponent is strictly negative, and the summable majorant at any fixed s=M>max⁡(c2,1) permits passage of the limit through the sum. Thus the coefficient of sN in the polynomial F is zero. Along the negative real axis use F+ and let s→−∞. The analogous majorant at M′<min⁡(c1,−1) permits passage through the sum. Terms with n>n0 tend to zero and those with n=n0 tend to an0,N after division by sN. The same polynomial leading coefficient must therefore equal an0,N. It follows that an0,N=0. Since n0 was arbitrary, every degree-N coefficient vanishes. Repeat the argument with N−1, down to 0. All coefficients vanish, including when N=0.

Dependency tree · 0 levels

Nothing. This result depends on no other item in the library.

Sources