Alphabeta Math
LemmaStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedaudited 2026-09-27
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 eventual Hilbert function of a zero-dimensional projective quotient equals its total length

Statement

Assume the Axiom of Choice. Let k be a field, let I⊆k[x0,…,xn] be any homogeneous ideal, let S=k[x0,…,xn]/I carry its standard grading, and let X=Proj⁡S be zero-dimensional in the chartwise sense that every standard chart ring Ai=(Sxi)0 is either zero or of Krull dimension 0 (Projective scheme of a homogeneous quotient and its standard affine charts, Krull dimension of a nonzero ring); the empty case X=∅ is included. Then for all sufficiently large m

dim⁡kSm=len⁡k(X),

the total length of Total length of a zero-dimensional projective scheme. Saturation of I is not assumed, and the equality is between natural numbers.

Facts & Assumptions

Given: The Axiom of Choice, a field k, a homogeneous ideal I⊆k[x0,…,xn], the standard graded quotient S=k[x0,…,xn]/I, its standard chart rings Ai=(Sxi)0, each zero or of Krull dimension 0, and X=Proj⁡S.

[L1]

The points of X are the homogeneous primes p of S with S+⊈p; the standard charts D+(xi)=Spec⁡(Ai) cover X, chart points correspond to the primes of Ai, the local ring at a point is the localization of any chart ring containing it, and the chart identifications agree on overlaps (Projective scheme of a homogeneous quotient and its standard affine charts, Prime and local-ring correspondence on standard projective charts). A prime ideal is proper with multiplicative complement (Prime ideals and maximal ideals in a commutative ring).

[L2]

For a homogeneous element f of positive degree in S, the standard open D+(f)={x∈X:f∉px} is the affine chart Spec⁡((Sf)0), its ring of global sections is (Sf)0, and inside the chart D+(xi) the piece D+(fxi) corresponds to the degree-zero dehomogenisation f/xideg⁡f∈Ai (The standard open D+(f) of a projective quotient is the affine chart Spec⁡((Sf)0)).

[L3]

Assume AC. If k⊆K is a field extension and SK=S⊗kK, then SK is standard graded with (SK)m=Sm⊗kK, so dim⁡K(SK)m=dim⁡kSm; the projective scheme XK=Proj⁡(SK) is again zero-dimensional in the chartwise sense; and len⁡K(XK)=len⁡k(X) (Field extension preserves the graded pieces and the total length of a zero-dimensional projective quotient).

[L4]

Assume AC. For the zero-dimensional X: the point set is finite and discrete; each local ring OX,x is a finite-dimensional local k-algebra with nilpotent maximal ideal, finite length and finite residue degree; X is the finite disjoint union of the spectra Spec⁡(OX,x) of its local rings; and the total length is len⁡k(X)=∑xℓOX,x(OX,x)[κ(x):k], with the convention len⁡k(∅)=0. For an affine scheme Spec⁡B with B a finite-dimensional k-algebra one has len⁡k(Spec⁡B)=dim⁡kB (A zero-dimensional projective scheme has finitely many closed points with finite-dimensional local rings, Total length of a zero-dimensional projective scheme, Composition series and length of a module, The residue field at a point of an affine scheme, The degree [K:F]=dim⁡FK of a finite field extension, The underlying space of an affine spectrum, Schemes).

[L5]

Over an infinite field, a finite-dimensional vector space is not the union of finitely many proper linear subspaces (A finite-dimensional vector space over an infinite field is not a finite union of proper subspaces), and the fraction field k(t) of the polynomial ring k[t] is infinite, since the monomials tm have pairwise distinct images by the domain property (A polynomial ring over an integral domain is an integral domain, Frac⁡(D) is a field and d↦d/1 embeds the integral domain D).

[L6]

For a nonempty finite disjoint union of affine spectra, global sections multiply: Γ(∐j=1rSpec⁡Rj)=∏j=1rRj when r≥1, because the product-ring projections identify the spectrum with the disjoint union and restriction to the clopen pieces induces the product isomorphism (The spectrum of a finite product ring is the disjoint union of the factor spectra). For the empty union, the structure sheaf has exactly one section over its empty underlying space, so its ring of global sections is the zero ring (A set-valued sheaf has a unique section over the empty open set). The dimension of a finite direct sum of finite-dimensional spaces is the sum of the dimensions (If V=⨁i<nUi with every Ui finite-dimensional, then V is finite-dimensional and dim⁡FV=∑i<ndim⁡FUi; in particular dim⁡F(U⊕W)=dim⁡FU+dim⁡FW, Finite-dimensional vector space, and its dimension dim⁡FV; infinite-dimensional means having no finite basis).

[L7]

Localization is exact and commutes with itself: iterated localizations of S in any order agree up to canonical isomorphism, and kernels of localization maps are computed by the universal property (Localising twice is localising once at the multiplicative set generated by both denominator sets, Universal property of localisation: maps that invert S factor uniquely through S−1R, A localisation is unique up to a unique isomorphism compatible with the map from R); the localization SL of the graded ring S at a homogeneous element L of degree one is graded, the localisation map S→SL is degree-preserving, and its kernel is a graded ideal (Localisation at a homogeneous element is graded, with graded kernels and dehomogenised degree-zero parts).

Proof

technique · direct
1.1

If k is infinite, set K=k and SK=S; if k is finite, set K=k(t) with fraction field structure as in [L5], so that K is infinite, and set SK=S⊗kK; in the finite case [L3] gives dim⁡K(SK)m=dim⁡kSm for every m, len⁡K(XK)=len⁡k(X) for XK=Proj⁡(SK), and XK zero-dimensional in the chartwise sense.

L3L5
1.2

Assume now that k is infinite. If X≠∅, let p1,…,pr be its finitely many points, written as homogeneous primes of S by [L1], and for each j let Vj={a=(a0,…,an)∈kn+1:a0x0+⋯+anxn∈pj}; each Vj is a proper k-subspace, because it is the kernel of the linear map kn+1→(S/pj)1, which is nonzero as xi∉pj for some i; if X=∅ let L=x0, and otherwise [L5] provides a∉⋃jVj and we set L=a0x0+⋯+anxn∈S1. In both cases L∉px for every point x∈X, so X⊆D+(L).

L1L5
2.1

Consequently it suffices to prove the displayed equality for the pair (SK,K): if dim⁡K(SK)m=len⁡K(XK) holds for all m≥m0, then in the finite case dim⁡kSm=dim⁡K(SK)m=len⁡K(XK)=len⁡k(X) for all m≥m0, and in the infinite case the equality is the claim itself; from here on we therefore assume that k is infinite.

L3step 1.1
2.2

Since D+(L)⊆X on the other hand, we have D+(L)=X as open subschemes; by [L2] the open subscheme D+(L) is the affine scheme Spec⁡(A) with A=(SL)0=Γ(X,OX), so X=Spec⁡A.

L2step 1.2
3.1

By [L4] the space X is the finite disjoint union ∐x∈XSpec⁡(OX,x). If X=∅, [L6] gives A=Γ(X,OX)=0, so dim⁡kA=0=len⁡k(X) by the empty-sum convention in [L4]. If X≠∅, [L6] applies with the positive number of factors and gives A=∏x∈XOX,x, a finite-dimensional k-algebra with dim⁡kA=∑x∈Xdim⁡kOX,x; applying [L4] to the affine scheme Spec⁡A and to the local rings OX,x gives len⁡k(X)=dim⁡kA=∑xℓOX,x(OX,x)[κ(x):k]. Thus A is finite-dimensional and dim⁡kA=len⁡k(X) in either case.

L4L6step 2.2
3.2

For every index i there is Ni≥1 with xiNi∈LS: the open subschemes D+(xi) and D+(Lxi)=D+(L)∩D+(xi) of X coincide by step 2.2, so inside the chart D+(xi)=Spec⁡(Ai) the localization Ai→(Ai)vi at the degree-zero dehomogenisation vi=L/xi of L on that chart is an isomorphism, vi is a unit of Ai with inverse w, and writing w=a/xim with a∈Sm gives (L/xi)(a/xim)=1 in (Sxi)0, hence La−xim+1 is killed by a power of xi and xiNi=L⋅(xiMa)∈LS for Ni=m+1+M.

L1L2L7step 2.2
4.1

For m≥0 define the k-linear map φm:Sm→A, s↦s/Lm, using A=(SL)0⊆SL; every element of A is a fraction s/Lm with s∈Sm, so A=⋃m≥0φm(Sm), and since A is finite-dimensional by step 3.1 the ascending chain of images φm(Sm) stabilizes: there is m0 with φm surjective for every m≥m0.

L7step 3.1
4.2

Let K0=∑i=0nNi; every monomial in x0,…,xn of degree K0 is divisible by some xiNi by the pigeonhole principle, and S+K0 is generated by those monomials, so S+K0⊆LS.

step 3.2
5.1

Let T=ker⁡(S→SL)={s∈S:LMs=0 for some M}, a graded ideal of S by [L7], and put Tm=T∩Sm; then ker⁡φm=Tm for every m, because s/Lm=0 in (SL)0 exactly when s is killed by a power of L.

L7step 4.1
5.2

Consequently Sm=LSm−1 for every m≥K0: S+K0⊆LS gives (S+K0)m⊆(LS)m=LSm−1, and (S+K0)m=Sm because every degree-m monomial with m≥K0 is divisible by some degree-K0 monomial.

step 4.2
6.1

For every m≥K0 one has Tm=LTm−1: if t∈Tm, then t∈Sm=LSm−1 by step 5.2, say t=La with a∈Sm−1, and LMt=0 implies LM+1a=0, so a∈Tm−1; conversely L Tm−1⊆Tm. Hence dim⁡kTm≤dim⁡kTm−1 for m≥K0, and the sequence of dimensions is eventually constant, say equal to d for all m≥m1.

step 5.1step 5.2
7.1

One has d=0: for m≥m1 the map L:Tm→Tm+1 is surjective by step 6.1, so LM:Tm1→Tm1+M is surjective for every M; but Tm1⊆Sm1 is finite-dimensional and every element of T is killed by a power of L, so some LM kills all of Tm1 and Tm1+M=0, forcing d=0 and Tm=0 for all m≥m1.

step 6.1
8.1

For m≥max⁡(m0,m1) the map φm:Sm→A is surjective by step 4.1 and has kernel Tm=0 by steps 5.1 and 7.1, hence is an isomorphism of k-vector spaces and dim⁡kSm=dim⁡kA=len⁡k(X) by step 3.1.

step 3.1step 4.1step 5.1step 7.1
9.1

If k is infinite, step 8.1 proves the claim; if k is finite, step 8.1 applied over the infinite field K=k(t) to SK and XK gives dim⁡K(SK)m=len⁡K(XK) for all large m, and step 2.1 converts this into dim⁡kSm=len⁡k(X) for all large m; the empty case is included, since then L=x0 still gives A=(Sx0)0=0, and all steps above remain valid.

step 2.1step 8.1
10.1

The proof is complete: the equality dim⁡kSm=len⁡k(X) holds for all sufficiently large m, no saturation of I was used, the case X=∅ is covered by len⁡k(∅)=0, and the Axiom of Choice is inherited from the finite-chart, base-change and finiteness suppliers of [L3], [L4] and [L5].

L3L4L5step 9.1∎

Depends on

Used by

Dependency tree · two levels

131 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