Chevalley groups, commutative rings, class-preserving endomorphisms, locally inner endomorphisms, Sha-rigidity

1. Problem Statement and Context

The Core Question:
The paper investigates the X-rigidity of adjoint Chevalley groups $G_{ad}(\Phi, R)$ and their elementary subgroups $E_{ad}(\Phi, R)$ over a commutative ring $R$. Specifically, it addresses the converse of a standard group theory fact: while every inner automorphism is class-preserving (maps every element to a conjugate of itself), is every class-preserving endomorphism necessarily inner?

Definitions:

• Locally Inner (Class-Preserving) Endomorphism: An endomorphism $\phi: G \to G$ such that for every $g \in G$, $\phi(g)$ is conjugate to $g$ in $G$.
• X-Rigidity: A group $G$ is X-rigid if its Shafarevich–Tate set $X(G)$ is trivial. In terms of endomorphisms, this means $\text{End}_c(G) = \text{Inn}(G)$ (Property E).
• The Gap in Previous Work: The authors note that a previous claim regarding X-rigidity for local rings of rank $>1$ (in reference [3]) contained a logical gap. That proof relied on a structural description of injective endomorphisms, which had not been established, whereas the available literature only covered automorphisms.

Objective:
To provide a direct, self-contained proof of X-rigidity for the low-rank types $A_1, A_2, B_2,$ and $G_2$ over arbitrary commutative rings, avoiding the unverified assumptions of prior work.

2. Methodology

The authors employ a direct, constructive approach analyzing the structure of locally inner endomorphisms $\phi$. The methodology relies on the following key techniques:

1. Localization and Reduction:

   • Using the fact that a commutative ring $R$ embeds into the product of its localizations at maximal ideals ($R \hookrightarrow \prod R_{\mathfrak{m}}$), the problem is reduced to the local case.
   • Further reduction is made to residue fields $k(\mathfrak{m}) = R_{\mathfrak{m}}/\text{Rad}(R_{\mathfrak{m}})$.
   • The authors utilize Gauss decompositions (for local rings) and Bruhat decompositions (for fields) to analyze the structure of elements.

2. Normalization Strategy:

   • Since $\phi$ is locally inner, it is conjugate to the identity on specific generators. The authors compose $\phi$ with an inner automorphism to normalize it such that $\phi$ fixes specific "test" elements (e.g., $X_0 = x_{\alpha}(1)x_{\beta}(1)$).
   • The goal is to show that after this normalization, $\phi$ must be the identity map on the elementary subgroup.

3. Centralizer Analysis:

   • For each type, the authors explicitly compute the centralizer $C_G(X_0)$ of a carefully chosen element $X_0$ (usually a product of simple root unipotents).
   • They determine the form of elements in the centralizer within the unipotent subgroup $U^+$, often deriving polynomial constraints on the parameters.

4. Trace Identities and Polynomial Constraints:

   • A crucial tool is the trace invariance in the adjoint representation. Since $\phi(g)$ is conjugate to $g$, $\text{tr}(\phi(g)) = \text{tr}(g)$.
   • For types $A_2$ and $B_2$ (and $G_2$), the authors use explicit trace formulas for products of root elements, such as $\text{tr}(x_{\alpha}(t)x_{-\alpha}(s))$, which are polynomials in $s$ and $t$.
   • By equating the trace of $\phi(x(t))$ with the trace of $x(t)$, they derive functional equations for the ring endomorphism $\rho: R \to R$ induced by $\phi$.

5. Rank-One Calculus and Commutator Relations:

   • The proofs heavily utilize standard Chevalley commutator relations and rank-one identities (e.g., $x_{-\alpha}(u)x_{\alpha}(v) = \dots$) to propagate constraints from simple roots to the entire root system.

3. Key Contributions and Results

The paper proves Theorem 1.1, establishing X-rigidity for the specified types under specific unit assumptions in the ring $R$.

Assumptions:

• Types $A_1, A_2, B_2$: $2 \in R^\times$ (2 is invertible).
• Type $G_2$: $2, 3 \in R^\times$ (2 and 3 are invertible).

Main Results:

1. Elementary Subgroups: Every locally inner endomorphism of $E_{ad}(\Phi, R)$ is inner for $\Phi \in \{A_1, A_2, B_2, G_2\}$.
2. Adjoint Groups:
   • For $\Phi \in \{A_2, B_2, G_2\}$, every locally inner endomorphism of $G_{ad}(\Phi, R)$ is inner.
   • For $\Phi = A_1$, the result holds for $G_{ad}(A_1, R)$ provided $E_{ad}(A_1, R)$ is normal in $G_{ad}(A_1, R)$.
3. Conclusion: All listed groups are X-rigid.

Specific Technical Breakthroughs by Type:

• Type $A_1$:

  • The proof involves a detailed matrix analysis in the $3 \times 3$ realization.
  • The authors use trace constraints to determine the form of $\phi(x_{-\alpha}(s))$.
  • A critical step involves showing that a specific matrix form derived from the conjugacy condition forces the off-diagonal parameters to vanish, leading to the identity map.

• Type $A_2$:

  • The authors analyze the centralizer of $X_0 = x_{\alpha_1}(1)x_{\alpha_2}(1)$.
  • They prove that the image of the highest root element $x_{\alpha_1+\alpha_2}(1)$ must be of the form $x_{\alpha_1}(a)x_{\alpha_2}(a)x_{\alpha_1+\alpha_2}(b)$ with $a^2=0$ and $b \in R^\times$.
  • Graph Automorphism Obstruction: A significant contribution is the proof that a potential "graph automorphism" (swapping simple roots) is not locally inner. They construct an element $Y$ such that $Y$ and $\phi(Y)$ are not conjugate, using trace and determinant arguments over fields of characteristic $\neq 7$ and a specific matrix construction for characteristic 7.
  • Trace identities are used to prove the induced ring endomorphism $\rho$ is the identity ($4(\rho(t)-t)=0 \implies \rho(t)=t$).

• Type $B_2$:

  • Similar to $A_2$, but utilizing the specific commutator relations of $B_2$ (e.g., $[x_\alpha, x_\beta] = x_{\alpha+\beta}x_{\alpha+2\beta}$).
  • The proof shows that the image of the long root involution $h_\alpha(-1)$ and the Weyl group elements $w_\beta(1)$ forces the parameter $c$ (representing a deviation from the identity) to be zero.

• Type $G_2$:

  • This is the most complex case due to the higher number of roots and more intricate commutator relations involving coefficients like 3.
  • The authors perform a "direct computer calculation" (symbolic) in the adjoint representation to determine the centralizer of $X_0$ and the constraints on the image of the highest root $x_{2\alpha+3\beta}(1)$.
  • They derive a system of polynomial equations for the parameters of the conjugating element, showing that $b=0$ (the deviation from the identity) is forced by the invertibility of 2 and 3.
  • The proof concludes by showing that fixing the images of simple root elements and Weyl group elements forces the entire endomorphism to be the identity.

4. Significance

1. Resolution of a Logical Gap: The paper explicitly addresses and bypasses the gap in the previous literature (reference [3]) regarding injective endomorphisms. By providing a self-contained proof for low-rank types that does not rely on the classification of injective endomorphisms, the authors validate the X-rigidity property for these groups independently.
2. Generality: The results hold for arbitrary commutative rings (subject to the unit conditions), not just fields or local rings. This is a significant generalization of rigidity results in algebraic K-theory and group theory.
3. Foundation for Higher Ranks: The authors suggest that the direct analysis approach developed here can be extended to all reduced irreducible root systems, although the complexity of the commutator relations increases with rank.
4. Methodological Rigor: The paper demonstrates the power of combining localization techniques, explicit matrix calculations, and trace identities to solve structural problems in infinite groups over rings.

In summary, this paper establishes that for low-rank adjoint Chevalley groups over commutative rings, the only way to map every element to a conjugate of itself is via an inner automorphism, thereby confirming their X-rigidity and solidifying the theoretical foundation for the study of class-preserving maps in this context.