Alphabeta Math
TheoremStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedprecheck passjudge pass (gpt-6.1-sol)audited 2026-10-08
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.

The Bergman metric is positive definite on bounded domains and biholomorphically invariant

Statement

Assume the Axiom of Countable Choice ACω (The Axiom of Countable Choice (ACω)) and let m≥1. Let Ω⊆Cm be a bounded domain with Bergman kernel KΩ, Bergman metric form gΩ, and quadratic form BΩ2(z;X)=gΩ(z)(X,X) as in The Bergman metric form on a bounded domain, and use the one-based coordinate aliases of The Levi form and strict plurisubharmonicity.

  1. For every z∈Ω and every X∈Cm∖{0}, BΩ2(z;X)=1KΩ(z,z)sup⁡{∣∂Xf(z)∣2: f∈A2(Ω), ∥f∥L2≤1, f(z)=0}, where ∂Xf=∑j=1mXj∂f∂zj. The supremum is finite, positive, and attained. In particular BΩ2(z;X)>0, so the Hermitian form gΩ(z) is positive definite.

  2. If F:Ω→Ω′ is a biholomorphism of bounded domains, then for all z∈Ω and all X,Y∈Cm, gΩ(z)(X,Y)=gΩ′(F(z))(DF(z)X, DF(z)Y). Equivalently, BΩ2(z;X)=BΩ′2(F(z);DF(z)X) for every X.

Facts & Assumptions

[A1]

The only choice principle assumed is ACω (The Axiom of Countable Choice (ACω)): it enters through the Bergman Hilbert/Riesz structure and the orthogonal projection. No full Axiom of Choice is used.

[F1]

For a nonempty open Ω⊆Cm, A2(Ω) is a closed complex Hilbert subspace of L2(Ω) for the first-variable-linear pairing, each class has a unique holomorphic representative, point evaluation has a unique Riesz section kw=KΩ(⋅,w) with f(w)=⟨f,kw⟩, and KΩ(z,w)=⟨kw,kz⟩=KΩ(w,z)‾ (The Bergman space A2(Ω) and the Bergman kernel, A2(Ω) is closed, and the Bergman kernel is the sum over any complete orthonormal system).

[F2]

For a bounded domain Ω, KΩ∈C∞(Ω×Ω), the diagonal is smooth with KΩ(z,z)≥1/λΩ(Ω)>0, and KΩ is holomorphic in its first and antiholomorphic in its second variable (Smoothness of the Bergman kernel and positivity of its diagonal on bounded domains, A2(Ω) is closed, and the Bergman kernel is the sum over any complete orthonormal system).

[F3]

For every nonempty compact K⊆Ω there are finite constants with sup⁡K∣f∣≤CK∥f∥2 and sup⁡K∣∂zαf∣≤CK,α∥f∥2 for every multi-index ∣α∣≤1 (Sup-norm and first-derivative bounds by the L2 norm on compact subsets).

[F4]

Every bounded linear functional on a real or complex Hilbert space has a unique representing vector, isometrically; in the first-variable-linear convention f(x)=⟨x,y⟩ (Riesz representation for Hilbert spaces).

[F5]

A closed linear subspace M of a Hilbert space satisfies H=M⊕M⊥ uniquely, and the orthogonal projection onto M is defined by that decomposition (Orthogonal decomposition by a closed subspace, The Hilbert orthogonal projection onto a closed subspace).

[F6]

Cauchy–Schwarz: ∣⟨x,y⟩∣≤∥x∥∥y∥, with equality exactly for linearly dependent pairs (Cauchy–Schwarz: ∣⟨x,y⟩∣≤∥x∥ ∥y∥, with equality exactly for dependent pairs).

[F7]

If real C2 functions satisfy u≥v near a with u(a)=v(a), then Lu(a;X)≥Lv(a;X) for every X (The complex Hessian of a C2 function dominates that of a minorant at a common minimum).

[F8]

The Bergman metric form is the Levi form of uΩ=log⁡KΩ(z,z), and BΩ2(z;X)=LuΩ(z;X)=∑j,k=1m∂2uΩ∂zj∂z‾k(z)XjXk‾ for the one-based coordinate aliases (The Bergman metric form on a bounded domain, The Levi form and strict plurisubharmonicity).

[F9]

Wirtinger operators are ∂zj=12(∂xj−i∂yj), ∂z‾j=12(∂xj+i∂yj), for holomorphic functions the real-coordinate derivative identities in [F12] give ∂z‾jf=0, real mixed partials commute by Clairaut--Schwarz theorem for continuous second partial derivatives, log⁡′(x)=1/x for x>0 by The natural logarithm has derivative 1/x and equals the integral from 1 to x of 1/t, and the product and chain rules hold for these first-order operators (Wirtinger operators in Cm, Complex m-space and its real coordinate dictionary, Ck maps and multi-index derivative notation in Euclidean space, The chain rule for total derivatives: D(g∘f)(a)=Dg(f(a))∘Df(a)).

[F11]

A biholomorphism F:Ω→Ω′ satisfies KΩ(z,w)=JF(z)KΩ′(F(z),F(w))JF(w)‾ with JF=det⁡CDF≠0, and UFg=(g∘F)JF is a unitary isomorphism A2(Ω′)→A2(Ω) (Transformation law of the Bergman kernel under a biholomorphism).

[F12]

A holomorphic function on an open set is of class Cn in the real coordinates for every natural n, hence smooth (Holomorphic functions of several variables are smooth and their complex derivatives are holomorphic).

[F13]

The complex plane is complete, hence Banach for its modulus norm. A continuous differentiable complex-valued curve whose derivative is bounded by C changes by at most C times the parameter distance (The complex plane is complete, and convergence is equivalent to convergence of real and imaginary parts, Mean value inequality for a differentiable Banach-valued curve).

Proof

technique · direct, using the reduced Riesz kernel of the subspace $\{f(z)=0\}$ and a second-order comparison with $\log K_\Omega$

Given: ACω, the bounded domain Ω, a point z∈Ω, and a direction X∈Cm.

1.1A1F1F2F3F4F5

Put M:=KΩ(z,z). By [F2], M>0, so kz≠0 and the subspace H′:={f∈A2(Ω):f(z)=0}={f:⟨f,kz⟩=0}=(Ckz)⊥ is closed; by [F5], A2(Ω)=Ckz⊕H′. By [F3] with the singleton K={ζ} the evaluation εζ(f):=f(ζ) is bounded on A2(Ω) and hence on H′ for every ζ∈Ω; [F4] gives a unique hζ∈H′ with f(ζ)=⟨f,hζ⟩ for all f∈H′. Define E(ζ):=∥hζ∥22.

1.2F1F2F5F6algebra

By [F5] the projection of kζ onto H′ is hζ=kζ−⟨kζ,kz⟩Mkz, and ⟨kζ,kz⟩=KΩ(z,ζ) by [F1]. Hence, pointwise in η, hζ(η)=KΩ(η,ζ)−KΩ(z,ζ)MKΩ(η,z), and reproducing on hζ gives E(ζ)=∥hζ∥22=KΩ(ζ,ζ)−∣KΩ(z,ζ)∣2M. By [F2] the function E is C∞ on Ω; by [F6] it satisfies E(ζ)≥0 with E(z)=0, and E(ζ)>0 for ζ≠z, since then some coordinate satisfies ζj≠zj and the bounded function η↦ηj−zj lies in H′ with nonzero value at ζ.

1.3F2F8F9algebra

Compute the Levi form of E. Write D(ζ):=KΩ(ζ,ζ) and A(ζ):=KΩ(ζ,z), so that E=D−∣A∣2M, D(z)=A(z)=M and ∣A(z)∣2=M2, and D is C∞ near z. Set aj:=∂jA(z) and cjk:=∂j∂k‾D(z). Since KΩ is holomorphic in the first and antiholomorphic in the second variable [F2, F9], the Wirtinger derivatives ∂ζ‾KΩ and ∂ηKΩ vanish identically, and the product and chain rules give ∂jD(z)=aj, ∂k‾D(z)=ak‾, ∂j∣A∣2(z)=ajM, ∂k‾∣A∣2(z)=Mak‾, and ∂j∂k‾∣A∣2(z)=ajak‾; consequently ∂jE(z)=∂j‾E(z)=0, ∂j∂kE(z)=∂j‾∂k‾E(z)=0, and ∂j∂k‾E(z)=cjk−ajak‾M. Since D(z)=M>0, the same rules applied to log⁡D give ∂j∂k‾log⁡D(z)=cjkM−ajak‾M2, so M ∂j∂k‾log⁡D(z)=∂j∂k‾E(z); summing against XjXk‾ and using [F8] yields LE(z;X)=M Llog⁡D(z;X)=M BΩ2(z;X).

2.1A1F5F6F7F8F12step 1.2

Let f∈H′ with ∥f∥2≤1. By [F6] and the reproducing identity of step 1.1, ∣f(ζ)∣2=∣⟨f,hζ⟩∣2≤∥f∥22E(ζ)≤E(ζ) for every ζ∈Ω, with equality at ζ=z. Both E and ∣f∣2 are C2: E was shown C∞ in step 1.2, and f is holomorphic, hence smooth in the real coordinates by [F12]. So [F7] gives LE(z;X)≥L∣f∣2(z;X)=∣∂Xf(z)∣2; hence sup⁡{∣∂Xf(z)∣2:f∈H′,∥f∥2≤1}≤LE(z;X).

2.2F2F9F13step 1.2step 1.3

Fix X≠0 and δ>0 such that z+τX∈Ω for real ∣τ∣<δ. Set H(τ,σ):=hz+τX(z+σX). The explicit formula in step 1.2 makes H smooth, with H(τ,0)=H(0,σ)=0. Differentiating that formula gives ∂τ∂σH(0,0)=∑j,kXjXk‾(cjk−ajak‾/M)=b, where b:=LE(z;X) by step 1.3. Put R(τ,σ):=H(τ,σ)−bτσ. For every ε>0, continuity makes ∣∂τ∂σR∣<ε on a small rectangle about (0,0). Since ∂σR(0,σ)=0 and R(τ,0)=0, applying [F13] first to τ↦∂σR(τ,σ) and then to σ↦R(τ,σ) yields ∣R(τ,σ)∣≤ε∣τσ∣. Thus H(τ,σ)/(τσ)→b as both nonzero real parameters tend to 0, and in particular E(z+τX)/τ2=H(τ,τ)/τ2→b.

3.1A1F4F8step 2.1step 1.3

Assume X≠0. The function f0(η):=⟨η−z,X⟩=∑j=1m(ηj−zj)Xj‾ is holomorphic on Ω and bounded there because Ω is bounded, so f0∈H′; also ∂Xf0(z)=∑j=1mXjXj‾=∣X∣2. Applying step 2.1 to f0/∥f0∥2 gives LE(z;X)≥∣X∣4/∥f0∥22>0 for X≠0. Hence by step 1.3, BΩ2(z;X)=LE(z;X)/M>0: the quadratic form of gΩ(z) is positive on every nonzero vector, so the Hermitian form is positive definite.

4.1A1F1F4F5F6F10step 1.1step 3.1step 2.2

For real τ≠0 put gτ:=hz+τX/τ∈H′. Reproduction gives ⟨gτ,gσ⟩=H(τ,σ)/(τσ), so step 2.2 implies ∥gτ−gσ∥22→0 as τ,σ→0. Choose, for example, τn=δ/(n+2). The closed subspace H′ is complete by [F1, F5], so gτn→g∈H′; the same Cauchy estimate gives gτ→g for all real τ→0. By step 2.2, ∥g∥22=b. For every f∈H′, continuity of the pairing and holomorphic differentiability give ⟨f,g⟩=lim⁡τ→0f(z+τX)/τ=∂Xf(z). Hence g represents the derivative functional, and Cauchy–Schwarz gives sup⁡f∈H′,∥f∥2≤1∣∂Xf(z)∣2=∥g∥22=b, attained at g/∥g∥2 because b>0 by step 3.1.

5.1step 2.1step 1.3step 3.1step 4.1

Combining steps 1.3, 2.1 and 4.1 yields sup⁡{∣∂Xf(z)∣2:f∈A2(Ω),∥f∥2≤1,f(z)=0}=LE(z;X)=KΩ(z,z)BΩ2(z;X), which is the displayed formula, with the supremum attained; step 3.1 gives its strict positivity for X≠0. This proves the two assertions of part 1.

6.1A1F8F10F11step 5.1

For part 2, let F:Ω→Ω′ be a biholomorphism, J:=JF=det⁡CDF, and let UFg=(g∘F)J be the unitary isomorphism of [F11], so that UF maps the closed unit ball of A2(Ω′) onto that of A2(Ω) and {g:g(F(z))=0} onto H′={f:f(z)=0} (because J(z)≠0). If f=UFg, then the product rule and [F10] give ∂Xf(z)=J(z) ∂DF(z)Xg(F(z))+g(F(z)) ∂XJ(z), and for f∈H′ the second term vanishes; hence the supremum identity of part 1 applied on both domains, together with the diagonal kernel law KΩ(z,z)=∣J(z)∣2KΩ′(F(z),F(z)) from [F11], gives KΩ(z,z)BΩ2(z;X)=∣J(z)∣2KΩ′(F(z),F(z))BΩ′2(F(z);DF(z)X). Since KΩ(z,z)>0 and ∣J(z)∣2KΩ′(F(z),F(z))=KΩ(z,z), dividing by the positive number KΩ(z,z) yields BΩ2(z;X)=BΩ′2(F(z);DF(z)X) for every nonzero X; when X=0, both quadratic forms are zero by definition.

7.1F10step 6.1∎

The quadratic forms of the Hermitian forms agree under the complex-linear map DF(z). For all X,Y, applying step 6.1 to the four directions X+Y,X−Y,X+iY,X−iY and using complex linearity of DF(z) plus the polarization identity g(X,Y)=14(q(X+Y)−q(X−Y)+iq(X+iY)−iq(X−iY)) for the quadratic form q(Z)=g(Z,Z) gives gΩ(z)(X,Y)=gΩ′(F(z))(DF(z)X,DF(z)Y).

Remarks

The supremum in part 1 is taken over all of A2(Ω) with the single constraint f(z)=0; since f(z)=⟨f,kz⟩, this is exactly the closed subspace H′ of the proof. In the orthonormal-system proof of the source, the reduced kernel of H′ plays the role of the element g of step 4.1. The constant hidden in the formula is carried by KΩ(z,z); the Bergman metric is normalized so that on the disc at the origin one obtains BD2(0;X)=2∣X∣2.

Depends on

Used by

Dependency tree · two levels

144 results within two dependency steps of this one, each drawn at its shortest distance from it. An arrow runs from a result to what uses it, so the chart reads left to right and ends at this result, which carries a heavier outline. Every node is a link to that result. Click elsewhere on the chart to enlarge it.

Sources