Theorem The Hilbert value of an actual saturated image
✓ theorem proved

The Hilbert value of an actual saturated image

Let \(k\) be a field and \(A\) a commutative \(k\)-algebra. Let \(M,N\) have compatible \(A\)- and \(k\)-actions and integer-indexed direct-sum decompositions. Suppose every \(M_n\) is finite-dimensional over \(k\).

Let \((A_r)_{r\in\mathbb Z}\) be homogeneous scalar subspaces acting compatibly on \(M\), so \(A_rM_s\subseteq M_{r+s}\). Choose \(d\in\mathbb Z\) and \(a\in A_d\). Let \(f:M\to N\) be \(A\)-linear of degree zero, with \[ \ker f=aM. \] Let \(S\subseteq M\) be a homogeneous \(A\)-submodule, saturated with respect to \(a\): \[ av\in S\quad\Longrightarrow\quad v\in S. \] Assume multiplication by \(a\) is injective on \(M\).

For \(q>0\), suppose the series \[ H_S(q)=\sum_{n\in\mathbb Z}\dim_k(S\cap M_n)q^n \] converges. Then the actual image has Hilbert value \[ H_{f(S)}(q)=(1-q^d)H_S(q). \]

Relation to the OpenAI notes

Actual image pieces and exact scalar-image values.

OpenAI notes: The comparison proof multiplies the upper-kernel Hilbert series by the linear-quotient factor and the correction degree shift. eq:step, edge:bound.

Lean (primary): The general image-shift theorem constructs an actual isomorphism of homogeneous intersection pieces and reindexes a summable integer-indexed series. The saturated-image theorem separately identifies the restricted map’s kernel with the scalar image, applies degreewise exactness and subtracts its Hilbert value. PositiveTransfer:273, PositiveTransfer:300, PositiveTransfer:316, Support:1227, Support:1256.

Role in the proof: The Blueprint explains both constructions and retains their distinct generic parameter hypotheses: nonzero q for the shift and positive q for saturated-image exactness, each with explicit summability.

Proof

Step 1. Restrict to the actual image

Define \[ f_S:S\longrightarrow f(S),\qquad v\longmapsto f(v). \] The map is surjective by the definition of the image. Write \(\pi_n^M,\pi_n^N\) for the component projections. Since \(S\) is homogeneous and \(f\) has degree zero, the component-transport theorem gives \[ \pi_n^N(f(v))=f(\pi_n^M(v))\in f(S). \] Thus the image is homogeneous, and \(f_S\) preserves degrees. These are the source's imageMap and map_homogeneous constructions.

Step 2. Identify the restricted kernel using saturation

If \(v\in S\) and \(f(v)=0\), the ambient kernel equality gives \(v=aw\) for some \(w\in M\). Since \(aw=v\in S\), saturation implies \(w\in S\). Therefore \(v\in aS\).

Conversely, for \(w\in S\), the vector \(aw\) lies in \(aM=\ker f\) and in \(S\), so it belongs to the kernel of \(f_S\). Hence \[ \ker f_S=aS. \]

Step 3. Apply graded exactness to the image presentation

The scalar image \(aS\) is homogeneous, because multiplication by \(a\in A_d\) is homogeneous and \(S\) is homogeneous. Steps 1–2 give the degree-zero exact sequence \[ 0\longrightarrow aS\longrightarrow S \xrightarrow{f_S}f(S)\longrightarrow0. \] Each \(S_n\) is finite-dimensional as a subspace of \(M_n\). Homogeneous preimages make the sequence exact in every degree, so the image pieces are finite-dimensional as well.

For \(q>0\), the Hilbert exactness theorem applies to this surjection and the assumed summable domain series. It gives \[ H_S(q)=H_{aS}(q)+H_{f(S)}(q). \]

Step 4. Compute the scalar image and subtract

Multiplication by \(a\) is injective on \(M\), hence on \(S\), and has degree \(d\). The scalar-image Hilbert theorem gives \[ H_{aS}(q)=q^dH_S(q). \] Its nonzero-parameter hypothesis follows from \(q>0\), and it uses the same summable series of \(S\).

Substitute this equality into Step 3 and rearrange: \[ H_{f(S)}(q)=H_S(q)-q^dH_S(q) =(1-q^d)H_S(q). \]

For a degree-one scalar, the factor is \(1-q\). For example, with \(A=k[X]\), \(M=S=A\), \(a=X\), and \(f:A\to A/(X)=k\), one obtains \(H_{f(S)}(q)=(1-q)/(1-q)=1\) for \(0<q<1\).

References

The local source declarations below give the construction and proof.

Lean formalization

OAI.KLInvariance.Hilbert.imageMap: Support.lean:1216.

OAI.KLInvariance.Hilbert.imageMap_surjective: Support.lean:1221.

OAI.KLInvariance.Hilbert.imageMap_ker: Support.lean:1227.

OAI.KLInvariance.Hilbert.map_homogeneous: Support.lean:1243.

OAI.KLInvariance.Hilbert.saturated_image_value: Support.lean:1256.

Used by