Hilbert additivity for the image of a homogeneous submodule
Let \(k\) be a field and \(A\) a commutative \(k\)-algebra. Let \(f:M\to N\) be degree zero between compatibly graded \(A\)-modules, and let \(S\subseteq M\) be a homogeneous \(A\)-submodule. Assume every \(M_n\) is finite-dimensional.
For \(q>0\), if the intersection series of \(S\) converges, then \[ H_S(q)=H_{S\cap\ker f}(q)+H_{f(S)}(q). \] Both series on the right converge.
Proof
Step 1. Define the map on the degree-n intersections
Restriction gives a \(k\)-linear map \[ f_n:S\cap M_n\longrightarrow f(S)\cap N_n,\qquad m\longmapsto f(m). \] The image lies in \(f(S)\) by definition and in \(N_n\) because \(f\) preserves degree.
Step 2. Lift a homogeneous image vector inside S
Take \(y=f(s)\in f(S)\cap N_n\), with \(s\in S\). Homogeneity of \(S\) gives \(\pi_n(s)\in S\cap M_n\), and component transport gives \[ f(\pi_n(s))=\pi_n(f(s))=\pi_n(y)=y. \] Thus \(f_n\) is surjective.
Its kernel consists exactly of the vectors in \(S\cap\ker f\cap M_n\). The correspondence is the identity on the underlying vector.
Step 3. Apply finite-dimensional rank–nullity
The domain \(S\cap M_n\) is finite-dimensional as a subspace of \(M_n\). Its image is therefore finite-dimensional, and rank–nullity gives \[ \dim_k(S\cap M_n) =\dim_k(S\cap\ker f\cap M_n) +\dim_k(f(S)\cap N_n). \] This supplies finiteness of the image pieces directly.
Step 4. Dominate both numerical series
Multiply by \(q^n>0\). Each right-hand term is nonnegative and at most the left-hand term. Since the series for \(S\) converges, comparison proves convergence of the kernel-intersection and image series.
Step 5. Sum the identities
Summing the degreewise equality is now valid and gives the stated Hilbert identity.
For an edge restriction \(\rho_e\), take \(S\) to be a partial outgoing kernel. Then \(S\cap\ker\rho_e\) is the kernel after adding that edge condition, and \(f(S)\) is its actual restriction image. The identity measures the change by that image's Hilbert value.
References
The local source declarations below give the construction and proof.
Lean formalization
OAI.KLInvariance.Hilbert.imagePieceMap: HilbertSeries.lean:451.
OAI.KLInvariance.Hilbert.imagePieceMap_surjective: HilbertSeries.lean:459.
OAI.KLInvariance.Hilbert.imageKernelEquiv: HilbertSeries.lean:476.
OAI.KLInvariance.Hilbert.finrank_image_exact: HilbertSeries.lean:492.
OAI.KLInvariance.Hilbert.value_image_exact: HilbertSeries.lean:513.