Holomorph description of \(H^1\)
A holomorph description of \(H^1\) for finite local systems on connected schemes
Shengrong Wu
National University of Singapore
Automatic Pipeline with ChatGPT
Abstract
For a field \(K\) and a finite Galois module \(M\), O’Dorney described \(H^1(K,M)\) in terms of finite ’etale algebras whose monodromy lies in the holomorph \(\Hol(M)=M\rtimes \Aut(M)\). We prove the scheme-theoretic analogue for a connected scheme \(U\) and a finite locally constant sheaf \(\mathcal M\) on \(U_{\et}\). The resulting description identifies \(H^1_{\et}(U,\mathcal M)\) with affine \(\Hol(M)\)-monodromy lifts of \(\pi_1(U)\) modulo translation-conjugacy, or equivalently with finite ’etale covers carrying a fixed \(\Aut(M)\)-resolvent torsor. For integral normal noetherian \(U\) this recovers exactly the classes unramified along \(U\), and in the affine case it yields a tensor-product compatibility for the corresponding finite ’etale algebras.
1. Introduction
Let \(M\) be a finite abelian group. The holomorph
\[ \Hol(M)=M\rtimes \Aut(M) \]
is the group of affine transformations of the set \(M\). Its structural role in degree-\(1\) cohomology is elementary but important: the normal subgroup \(M\) records translations, the quotient \(\Hol(M)/M\cong \Aut(M)\) records the fixed linear part of the monodromy, and the stabilizer of \(0\in M\) is exactly \(\Aut(M)\).
For a field \(K\), O’Dorney made this structure completely explicit. If \(M\) is a finite \(G_K\)-module with action
\[ \phi:G_K\to \Aut(M), \]
then \(H^1(K,M)\) may be described by lifts
\[ \psi:G_K\to \Hol(M) \]
of \(\phi\), modulo conjugation by the translation subgroup \(M\), or equivalently by finite ’etale \(K\)-algebras with monodromy in \(\Hol(M)\) and a fixed \(\Aut(M)\)-resolvent torsor .
The aim of this note is to record the corresponding statement for connected schemes. Let \(U\) be connected and let
\[ \phi:\pi_1(U,\bar u)\to \Aut(M) \]
be continuous. The associated finite locally constant sheaf \(\mathcal M_\phi\) plays the role of a finite Galois module. We show that \(H^1_{\et}(U,\mathcal M_\phi)\) is still controlled by affine \(\Hol(M)\)-actions: the only change is that one replaces the absolute Galois group \(G_K\) by the ’etale fundamental group \(\pi_1(U,\bar u)\).
This result is formally close to standard torsor theory. We do not claim a new abstract classification of torsors. The point is instead to isolate a concrete and functorial holomorph package, parallel to O’Dorney’s field-level statement, that keeps the linear and translational parts of the monodromy visible and makes the arithmetic meaning over non-field bases explicit. Over
\[ U=\Spec \mathcal O_{K,S}, \]
for example, the theorem identifies classes in \(H^1_{\et}(U,\mathcal M_\phi)\) with O’Dorney’s classes in \(H^1(K,M)\) that are unramified at every finite place outside \(S\).
The main theorem gives three equivalent descriptions:
- \(H^1_{\et}(U,\mathcal M_\phi)\);
- lifts of \(\phi\) to \(\Hol(M)\) modulo conjugation by \(M\);
- triples \((E,Y,\theta)\), where \(E\to U\) is a finite ’etale \(\Hol(M)\)-torsor, \(Y=E/\Aut(M)\) is a degree-\(|M|\) finite ’etale cover, and \(\theta:E/M\xrightarrow{\sim}T_\phi\) identifies the quotient by the translation subgroup with the fixed \(\Aut(M)\)-torsor attached to \(\phi\).
We also prove that pullback corresponds to base change, and that addition in \(H^1\) is controlled by the addition map
\[ M\times M\to M. \]
The paper is organized as follows. Section~2 fixes notation and recalls the necessary background on finite ’etale covers, finite local systems, and torsors. Section~3 states the main theorem and its corollaries. Section~4 contains the proof. Section~5 gives explicit examples over arithmetic and function-field bases. Section~6 records remarks and open questions.
2. Notation and preliminaries
Let \(U\) be a connected scheme with geometric point \(\bar u\), and let \(\pi_1(U,\bar u)\) be its profinite ’etale fundamental group.
Let \(M\) be a finite abelian group. We write
\[ \Hol(M)=M\rtimes \Aut(M) \]
for the holomorph, acting on the set \(M\) by affine maps
\[ (a,t)\cdot x=a(x)+t. \]
With this convention,
\[ (a,t)(b,u)=(ab,\ t+a(u)). \]
For \(y\in M\), write \(\tau_y=(1,y)\) for translation by \(y\).
Fix a continuous homomorphism
\[ \phi:\pi_1(U,\bar u)\to \Aut(M). \]
Let \(\mathcal M_\phi\) be the associated finite locally constant sheaf of abelian groups on \(U_{\et}\). Let
\[ T_\phi\to U \]
be the finite ’etale \(\Aut(M)\)-torsor corresponding to \(\phi\) under the equivalence between finite ’etale \(U\)-schemes and finite continuous \(\pi_1(U,\bar u)\)-sets.
We use the following facts.
- Finite ’etale \(U\)-schemes are equivalent to finite continuous \(\pi_1(U,\bar u)\)-sets .
- Finite locally constant sheaves of abelian groups on \(U_{\et}\) are equivalent to finite continuous \(\pi_1(U,\bar u)\)-modules .
- For a finite abelian sheaf, \(H^1\) classifies torsors under that sheaf; for a finite constant sheaf this is spelled out in .
Remark. Changing the base point \(\bar u\) conjugates \(\pi_1(U,\bar u)\) and therefore conjugates \(\phi\) and the set of lifts. All constructions below are invariant under this change.
Remark. If \(U=\Spec R\) is affine, a finite ’etale \(U\)-scheme is equivalently a finite ’etale \(R\)-algebra. The scheme language is slightly cleaner globally, but the algebra language is what one sees over rings of \(S\)-integers.
3. Main result
Theorem 3.1. Let \(U\) be a connected scheme with geometric point \(\bar u\), let \(M\) be a finite abelian group, and let
\[ \phi:\pi_1(U,\bar u)\to \Aut(M) \]
be continuous. Let \(\mathcal M_\phi\) be the associated finite locally constant sheaf of abelian groups on \(U_{\et}\), and let \(T_\phi\to U\) be the associated finite ’etale \(\Aut(M)\)-torsor.
Then the following sets are naturally in bijection:
\(H^1_{\et}(U,\mathcal M_\phi)\);
continuous homomorphisms
\[ \psi:\pi_1(U,\bar u)\to \Hol(M) \]
with \(\pi\circ \psi=\phi\), modulo conjugation by the translation subgroup \(M\subseteq \Hol(M)\);
isomorphism classes of triples \((E,Y,\theta)\) where
- \(E\to U\) is a finite ’etale \(\Hol(M)\)-torsor,
- \(Y=E/\Aut(M)\) is a finite ’etale \(U\)-scheme of degree \(|M|\),
- \(\theta:E/M\xrightarrow{\sim}T_\phi\) is an isomorphism of \(\Aut(M)\)-torsors.
Moreover, if \(f:V\to U\) is a morphism of connected schemes, then pullback
\[ f^*:H^1_{\et}(U,\mathcal M_\phi)\to H^1_{\et}(V,f^{-1}\mathcal M_\phi) \]
corresponds under (b) to precomposition with \(f_*:\pi_1(V,\bar v)\to \pi_1(U,\bar u)\) and under (c) to base change of the triple \((E,Y,\theta)\).
Corollary 3.2. Assume that \(U\) is integral, normal, and noetherian, with function field \(K\). Then the generic-fiber functor identifies
\[ H^1_{\et}(U,\mathcal M_\phi) \]
with the subset of \(H^1(K,M)\) consisting of the O’Dorney classes whose associated finite ’etale \(K\)-scheme is unramified along \(U\).
In particular, if
\[ U=\Spec \mathcal O_{K,S} \]
for a number field \(K\) and a finite set \(S\) of finite places, then these are exactly the classes unramified at every finite place \(v\notin S\). No archimedean place enters the geometry of \(U\).
Proposition 3.3. Let \(\alpha,\beta,\alpha+\beta\in H^1_{\et}(U,\mathcal M_\phi)\), and let \(Y_\alpha,Y_\beta,Y_{\alpha+\beta}\) be the corresponding finite ’etale \(U\)-schemes from Theorem~\(\ref{thm:main}\). Then there is a finite ’etale morphism
\[ Y_\alpha\times_U Y_\beta\longrightarrow Y_{\alpha+\beta}. \]
If \(U=\Spec R\) is affine, dualizing this morphism gives an inclusion of finite ’etale \(R\)-algebras
\[ L_{\alpha+\beta}\hookrightarrow L_\alpha\otimes_R L_\beta. \]
4. Proof
Lemma 4.1. Let \(\mathcal P\) be an \(\mathcal M_\phi\)-torsor on \(U_{\et}\). Then the geometric fiber \(X=\mathcal P_{\bar u}\) is a finite continuous \(\pi_1(U,\bar u)\)-set equipped with a simply transitive \(M\)-action satisfying
\[ g\cdot(m\cdot x)=\phi(g)(m)\cdot(g\cdot x) \]
for all \(g\in \pi_1(U,\bar u)\), \(m\in M\), and \(x\in X\).
Conversely, every such finite continuous \(\pi_1(U,\bar u)\)-set arises from an \(\mathcal M_\phi\)-torsor.
Proof. This is the standard translation between torsors under a finite locally constant sheaf and finite locally constant sheaves of sets with simply transitive group action. Evaluating at \(\bar u\) and using gives the stated description in terms of finite continuous \(\pi_1(U,\bar u)\)-sets.
Lemma 4.2. Fix an \(\mathcal M_\phi\)-torsor \(\mathcal P\), let \(X=\mathcal P_{\bar u}\), and choose an \(M\)-equivariant bijection
\[ \iota:M\xrightarrow{\sim} X. \]
Then there is a unique continuous homomorphism
\[ \psi_\iota:\pi_1(U,\bar u)\to \Hol(M) \]
such that
\[ g\cdot \iota(x)=\iota(\psi_\iota(g)(x)) \]
for all \(g\in \pi_1(U,\bar u)\) and \(x\in M\), and it satisfies
\[ \pi\circ \psi_\iota=\phi. \]
Proof. Define \(\psi_\iota(g)\in \operatorname{Sym}(M)\) by the displayed formula. Since the \(M\)-action on \(X\) is compatible with \(\phi\), for all \(x,m\in M\) we have
\[ g\cdot \iota(x+m) =g\cdot(m\cdot \iota(x)) =\phi(g)(m)\cdot(g\cdot \iota(x)). \]
Applying \(\iota^{-1}\) gives
\[ \psi_\iota(g)(x+m)=\psi_\iota(g)(x)+\phi(g)(m). \]
Thus \(\psi_\iota(g)\) is affine with linear part \(\phi(g)\), so \(\psi_\iota(g)\in \Hol(M)\) and \(\pi\circ\psi_\iota=\phi\).
The homomorphism property follows from functoriality of the monodromy action, and continuity is automatic because the target is finite.
Lemma 4.3. Let \(\iota_y=\iota\circ \tau_y\), where \(\tau_y\) denotes translation by \(y\in M\). Then
\[ \psi_{\iota_y}(g)=\tau_{-y}\psi_\iota(g)\tau_y \]
for all \(g\in \pi_1(U,\bar u)\).
Proof. For every \(x\in M\),
\[ g\cdot \iota_y(x) =g\cdot \iota(x+y) =\iota\bigl(\psi_\iota(g)(x+y)\bigr) =\iota_y\bigl(\tau_{-y}\psi_\iota(g)\tau_y(x)\bigr). \]
This proves the identity.
Proof of Theorem~\(\ref{thm:main}\). The equivalence (a)\(\Leftrightarrow\)(b) follows from Lemmas~\(\ref{lem:torsor-piset}\), \(\ref{lem:labeling}\), and \(\ref{lem:conjugacy}\).
For (b)\(\Rightarrow\)(c), let
\[ \psi:\pi_1(U,\bar u)\to \Hol(M) \]
be a lift of \(\phi\). Let \(E\to U\) be the finite ’etale \(\Hol(M)\)-torsor attached to the regular left action of \(\Hol(M)\) on itself. Let
\[ Y=E/\Aut(M). \]
Because the stabilizer of \(0\in M\) under the affine action of \(\Hol(M)\) is \(\Aut(M)\), the finite ’etale cover \(Y\) is the one corresponding to the affine \(\pi_1(U,\bar u)\)-set \(M\). Also,
\[ E/M \]
is the \(\Aut(M)\)-torsor attached to the composite
\[ \pi_1(U,\bar u)\xrightarrow{\psi}\Hol(M)\xrightarrow{\pi}\Aut(M), \]
which is exactly \(\phi\). Hence there is a canonical isomorphism
\[ \theta:E/M\xrightarrow{\sim}T_\phi. \]
For (c)\(\Rightarrow\)(b), a \(\Hol(M)\)-torsor \(E\) gives a monodromy map
\[ \psi:\pi_1(U,\bar u)\to \Hol(M) \]
well defined up to conjugacy by \(\Hol(M)\). The extra datum
\[ \theta:E/M\xrightarrow{\sim}T_\phi \]
fixes the image of \(\psi\) in \(\Aut(M)\) exactly. Therefore the remaining ambiguity is conjugation by the kernel of
\[ \Hol(M)\twoheadrightarrow \Aut(M), \]
namely by the translation subgroup \(M\).
The pullback statement follows from functoriality of the fiber functor and the construction of \(E\) and \(Y\) from finite continuous \(\pi_1\)-sets.
Proof of Corollary~\(\ref{cor:unramified}\). Let \(\eta\) be the generic point of \(U\). By , \(\pi_1(U,\bar\eta)\) is the Galois group of the maximal extension of \(K\) unramified along \(U\). Hence a lift
\[ \psi:\pi_1(U,\bar\eta)\to \Hol(M) \]
is exactly a field-theoretic lift of the unramified quotient of \(G_K\). Applying O’Dorney’s field theorem over \(K\) gives the identification with the unramified subset of \(H^1(K,M)\) .
When \(U=\Spec \mathcal O_{K,S}\), codimension-one points are the finite places outside \(S\), so “unramified along \(U\)” is equivalent to “unramified at every finite place \(v\notin S\)”.
Proof of Proposition~\(\ref{prop:sum}\). Choose lifts
\[ \psi_\alpha(g)=(\phi(g),\sigma_\alpha(g)), \qquad \psi_\beta(g)=(\phi(g),\sigma_\beta(g)). \]
Then \(Y_\alpha\times_U Y_\beta\) corresponds to the finite \(\pi_1(U,\bar u)\)-set \(M\times M\) with action
\[ g\cdot(x_1,x_2)= (\phi(g)x_1+\sigma_\alpha(g),\ \phi(g)x_2+\sigma_\beta(g)). \]
The addition map
\[ q:M\times M\to M,\qquad q(x_1,x_2)=x_1+x_2, \]
is equivariant for the action attached to the sum cocycle \(\sigma_\alpha+\sigma_\beta\), because
\[ q(g\cdot(x_1,x_2)) = \phi(g)(x_1+x_2)+\sigma_\alpha(g)+\sigma_\beta(g). \]
Therefore \(q\) induces a morphism of finite ’etale \(U\)-schemes
\[ Y_\alpha\times_U Y_\beta\to Y_{\alpha+\beta}. \]
If \(U=\Spec R\) is affine, dualizing the above morphism gives the inclusion
\[ L_{\alpha+\beta}\hookrightarrow L_\alpha\otimes_R L_\beta. \]
The scheme-theoretic statement is primary; the algebra statement is its affine dual.
5. Examples and applications
Example 5.1 (Quadratic covers over \(\mathbf Z\lbrack 1/2\rbrack\)). Let
\[ U=\Spec \mathbf Z[1/2],\qquad M=C_2, \]
with trivial \(\phi\). Then \(\Aut(C_2)=1\) and \(\Hol(C_2)\cong C_2\). Theorem \(\ref{thm:main}\) says that \(H^1_{\et}(U,C_2)\) is represented by quadratic ’etale covers of \(U\).
Examples include
\[ \mathbf Z[1/2]\times \mathbf Z[1/2],\qquad \mathbf Z[1/2][i],\qquad \mathbf Z[1/2][\sqrt{2}],\qquad \mathbf Z[1/2][\sqrt{-2}]. \]
These are finite ’etale because the discriminants \(-4\), \(8\), and \(-8\) are units after inverting \(2\).
The point here is not merely the generic fibers over \(\mathbf Q\), but the fact that the covers already exist over the arithmetic scheme \(\Spec \mathbf Z[1/2]\).
Example 5.2 (An \(S_3\)-cover over \(\mathbf Z\lbrack 1/6\rbrack\)). Let
\[ U=\Spec \mathbf Z[1/6],\qquad M=C_3. \]
Then
\[ \Aut(C_3)\cong C_2,\qquad \Hol(C_3)\cong S_3. \]
Let
\[ \phi:\pi_1(U,\bar u)\to C_2 \]
be the quadratic character cutting out
\[ T_\phi=\Spec \mathbf Z[1/6][\omega],\qquad \omega^2+\omega+1=0. \]
Consider
\[ L=\mathbf Z[1/6][x]/(x^3-2). \]
Its discriminant is
\[ -108=-2^2 3^3, \]
which is a unit on \(U\), so \(L\) is finite ’etale over \(U\). Its generic fiber has Galois closure with group \(S_3\), and the corresponding quadratic resolvent is \(\mathbf Q(\sqrt{-3})\). Thus Theorem~\(\ref{thm:main}\) places \(L\) naturally in
\[ H^1_{\et}(U,\mathcal M_\phi). \]
Example 5.3 (A function-field example). Let
\[ U=\Spec \mathbf F_5[t,1/t(t-1)],\qquad M=C_2, \]
again with trivial \(\phi\). Then
\[ Y=\Spec \mathbf F_5[t,1/t(t-1)][y]/(y^2-t(t-1)) \]
is finite ’etale over \(U\), because the discriminant \(4t(t-1)\) is a unit on the base. This gives a nontrivial class in
\[ H^1_{\et}(U,C_2). \]
Hence the same formalism applies over open curves over finite fields.
6. Remarks and open questions
Remark. The theorem should be viewed as a concrete holomorph reformulation of standard torsor theory in the specific setting of finite local systems, rather than as a new abstract theorem of ’etale cohomology. Its novelty lies in the explicit affine package and in the arithmetic-scheme interpretation parallel to O’Dorney’s field theorem.
Remark. The algebraic structure of \(\Hol(M)\) is essential. The normal subgroup \(M\) encodes translations, the quotient \(\Aut(M)\) encodes the fixed linear part, and the stabilizer of \(0\in M\) identifies the degree-\(|M|\) cover as a quotient by \(\Aut(M)\). The addition law in \(H^1\) is controlled by the additive map \(M\times M\to M\).
Remark. Nothing here addresses bad reduction. Extending the same picture to finite flat group schemes over Dedekind bases would require replacing finite ’etale monodromy by a different kind of integral structure.
Can one formulate an analogous description for nonabelian finite local systems, replacing ordinary \(H^1\) by pointed nonabelian cohomology?
Is there a finite-flat analogue over Dedekind schemes that remains valid at primes dividing \(|M|\)?
Can the additive compatibility in Proposition~\(\ref{prop:sum}\) be refined into a useful description of cup products or low-degree pairings?
For \(U=\Spec \mathcal O_{K,S}\), can one make the resolvent torsor explicitly computable for small groups such as \(C_3\), \(C_4\), and \(V_4\)?
What is the correct formulation in towers \(U_n\to U\), especially in Iwasawa-theoretic situations?
Is there a useful function-field refinement involving Artin–Schreier and Kummer covers on open curves over finite fields?
References
- [O’Dorney, 2025, arXiv:2506.11310] Evan M. O’Dorney, ’Etale algebras and the Kummer theory of finite Galois modules, arXiv:2506.11310, 2025.
- [Stacks Project Authors, 2025, Tag 03VD] The Stacks Project Authors, The Stacks Project, Tag 03VD (Fundamental groups), 2025.
- [Stacks Project Authors, 2025, Tag 0DV4] The Stacks Project Authors, The Stacks Project, Tag 0DV4 (Locally constant sheaves and the fundamental group), 2025.
- [Stacks Project Authors, 2025, Tag 0A5G] The Stacks Project Authors, The Stacks Project, Tag 0A5G (First cohomology and torsors), 2025.
Peer Review
Claude
I’ve read through your paper carefully. Here’s my evaluation.
Correctness of the Proofs
The proofs are correct, and the logical chain is clean. The core argument — Lemmas 4.1 through 4.3 feeding into Theorem 3.1 — is a straightforward and well-executed exercise in unwinding definitions. Specifically:
Lemma 4.1 is just the standard dictionary between torsors under a finite locally constant sheaf and π₁-sets with compatible simply transitive action. Nothing to object to here.
Lemma 4.2 correctly identifies the affine structure: the computation that ψ_ι(g)(x + m) = ψ_ι(g)(x) + φ(g)(m) is the right calculation showing the map lands in Hol(M) with the correct linear part. The continuity argument (finite target) is fine.
Lemma 4.3 is a one-line verification, and it’s correct.
Theorem 3.1 proof: The (a)⇔(b) direction is solid. For (b)⇒(c), the identification of Y = E/Aut(M) with the degree-|M| cover and E/M with T_φ via the stabilizer-of-0 observation is correct. The (c)⇒(b) direction — that θ pins down the Aut(M)-component and leaves only M-conjugacy ambiguity — is also right.
Corollary 3.2 and Proposition 3.3 are clean applications. The fiber product / addition map argument in 3.3 is correct.
Novelty Assessment
This is where things get more delicate. As you acknowledge in Remark 6.1, the main theorem is a “concrete holomorph reformulation of standard torsor theory.” I’d put it more bluntly: the core content of Theorem 3.1 is well-known to experts, and each equivalence is essentially a repackaging of standard facts.
The equivalence (a)⇔(b) — that H¹(U, M_φ) is classified by lifts π₁ → Hol(M) of φ modulo M-conjugacy — is a direct consequence of the general fact that H¹ classifies extensions (or equivalently, crossed homomorphisms modulo principal ones). The crossed homomorphism g ↦ σ(g) satisfying the cocycle condition is exactly the translational component of a lift to Hol(M). This perspective appears implicitly throughout group cohomology texts (Brown, Neukirch, Serre) and in the étale cohomology literature whenever one discusses twisted coefficient systems.
The equivalence (a)⇔(c) is a specialization of the general torsor-theoretic framework: given a short exact sequence 1 → M → Hol(M) → Aut(M) → 1, an Aut(M)-torsor T_φ, and the connecting map in nonabelian cohomology, the fiber over [T_φ] is exactly H¹(U, M_φ). This is standard nonabelian cohomology (Giraud, Breen, or even just Serre’s Galois Cohomology §I.5 in the field case).
What O’Dorney contributed was the explicit étale-algebra interpretation over fields, connecting this to concrete arithmetic objects. Your paper’s contribution is observing that the same package works over connected schemes by replacing G_K with π₁(U, ū). But this replacement is essentially formal — once you accept the equivalences between finite étale covers, π₁-sets, and locally constant sheaves (which are all in the Stacks Project), the theorem follows by the same argument.
Summary Judgment
The paper is correct but not novel enough for a research publication in its current form. The proofs are well-written and the examples in Section 5 are nice illustrations, but the main theorem is a routine generalization that experts in étale cohomology would regard as “known.” The open questions in Section 6 are genuinely interesting (especially Q2 on finite flat group schemes and Q5 on Iwasawa-theoretic towers), but they are stated without progress toward answers.
If you want to develop this further, I’d suggest either: (a) making serious progress on one of those open questions, particularly Q2 or Q3, which would give the paper real content beyond the formal generalization; or (b) finding a concrete arithmetic application where the scheme-level holomorph description yields something that wasn’t accessible from the field-level statement alone — for instance, explicit computation of the resolvent torsors in Q4 for specific rings of S-integers.
Would you like to dig into any of the open questions, or discuss strategy for strengthening the paper?
Gemini
The article titled “A holomorph description of \(H^{1}\) for finite local systems on connected schemes” is a specialized contribution to arithmetic geometry . Like the previous papers, it carries the subtitle “Automatic Pipeline with ChatGPT,” indicating it was produced through an AI-assisted workflow.
Honest Evaluation of the Article
This paper is a technical extension of a 2025 result by Evan O’Dorney. While O’Dorney’s original work focused on the field-theoretic level (Galois modules), this article generalizes the description to the language of schemes and étale fundamental groups (\(\pi_1\)) .
The primary value of the paper is not in proving a new abstract theorem of étale cohomology, but in providing a concrete “holomorph package”. It successfully bridges the gap between abstract torsor theory and practical arithmetic examples, such as quadratic covers over \(\mathbb{Z}[1/2]\) or \(S_3\)-covers over \(\mathbb{Z}[1/6]\) .
Is the Proof Correct?
Yes, the mathematical logic is sound and follows established categorical equivalences. The author relies on standard results from the Stacks Project to bridge the gap between sheaves and group actions . The proof structure is as follows:
Lemma 4.1 & 4.2: Correctfully uses the equivalence between torsors under a finite locally constant sheaf and continuous actions of \(\pi_1(U)\) on a finite set \(M\) . It shows that a lift to the holomorph \(Hol(M)\) is the same as choosing a specific affine action compatible with the linear action .
Lemma 4.3: Correctfully identifies that changing the “origin” of the affine action corresponds precisely to conjugation by the translation subgroup \(M\), which explains why the cohomology is classified by lifts modulo \(M\) .
Theorem 3.1: The bijection between cohomology classes, group-theoretic lifts, and geometric triples \((E, Y, \theta)\) is a standard application of the Galois correspondence for covers .
Proposition 3.3: The proof for tensor-product compatibility of the corresponding algebras is algebraically sound, using the addition map \(M \times M \to M\) to relate the product of two covers to the cover of their sum .
Has Other Research Already Proved This?
The core idea—that \(H^1\) classifies torsors and torsors are governed by the fundamental group—is a foundational result of Grothendieck’s Étale Cohomology.
However, the specific holomorph formulation presented here is unique to the 2025/2026 timeline:
O’Dorney (2025): Proved the version for fields.
This Paper (2026): Generalizes it to schemes.
While an expert in the field would find the transition from O’Dorney’s field case to the scheme case to be “natural” or “expected,” this paper is the first to explicitly write down the functorial package for schemes and prove the algebraic inclusion for the resulting étale algebras (\(L_{\alpha+\beta} \hookrightarrow L_\alpha \otimes_R L_\beta\)).