Algebraic groups generated by semisimple elements

Technical Summary: Algebraic Groups Generated by Semisimple Elements

Problem Statement
The paper addresses the structural characterization of connected linear algebraic groups $G$ over an algebraically closed field $K$ of characteristic zero, specifically focusing on the subgroup generated by semisimple elements. While it is well-established that reductive groups are generated by semisimple elements, the behavior of non-reductive groups (those containing a non-trivial unipotent radical) remains less clear. The central problem is to determine the precise structure of the subgroup $G_{\text{sem}}$ generated by all semisimple elements in an arbitrary connected linear algebraic group $G$ and to characterize the class of groups that are "multiplicatively generated" (i.e., generated entirely by their semisimple elements).

Methodology
The author employs the structural theory of linear algebraic groups and Lie algebra techniques specific to characteristic zero. Key methodological tools include:

• Decompositions: Utilizing the standard decompositions $G = G_{\text{ss}} \cdot R(G)$ and $G = G_{\text{red}} \ltimes R_u(G)$, where $G_{\text{red}}$ is a maximal connected reductive subgroup, $G_{\text{ss}}$ is a maximal connected semisimple subgroup, $R(G)$ is the radical, and $R_u(G)$ is the unipotent radical.
• Lie Algebra Correspondence: Analyzing the tangent algebra $\mathfrak{g} = \text{Lie}(G)$. The unipotent radical's Lie algebra $\mathfrak{n} = \text{Lie}(R_u(G))$ is treated as a module under the adjoint action of $G_{\text{red}}$.
• Module Theory: Decomposing $\mathfrak{n}$ into simple $G_{\text{red}}$-submodules. Specifically, the author defines $\mathfrak{n}_1$ as the sum of all non-trivial simple $G_{\text{red}}$-submodules in $\mathfrak{n}$.
• Exponential Map: Constructing the unipotent subgroup $U = \exp(\mathfrak{s})$, where $\mathfrak{s}$ is the subalgebra generated by $\mathfrak{n}_1$.
• Density Arguments: Using the fact that the set of semisimple elements in a reductive group contains a non-empty open subset, and that the product of this set with itself covers the group.

Key Contributions and Results

1. Characterization of Multiplicatively Generated Groups (Proposition 1):
   The paper establishes five equivalent conditions for a connected linear algebraic group $G$ to be multiplicatively generated (generated by its $G_m$-subgroups):

   • $G$ is generated by semisimple elements.
   • Any homomorphism $G \to G_a$ is trivial.
   • $G$ has no proper normal subgroup containing $G_{\text{red}}$.
   • The derived subgroup $[G, G]$ equals $G_{\text{ss}} \ltimes R_u(G)$.

2. Explicit Description of $G_{\text{sem}}$ (Theorem 1):
   The primary result provides an effective description of the subgroup $G_{\text{sem}}$ generated by all semisimple elements in any connected linear algebraic group $G$.

   • Let $\mathfrak{n}_1$ be the sum of non-trivial simple $G_{\text{red}}$-submodules in $\text{Lie}(R_u(G))$.
   • Let $\mathfrak{s}$ be the subalgebra generated by $\mathfrak{n}_1$, and $U = \exp(\mathfrak{s})$.
   • Result: $G_{\text{sem}} = G_{\text{red}} \ltimes U$.
     This implies that $G_{\text{sem}}$ is the minimal normal subgroup of $G$ containing $G_{\text{red}}$.

3. Corollary on Regular Subgroups:
   The paper proves that any regular subgroup of a reductive group (a subgroup containing a maximal torus of the ambient group) is generated by semisimple elements. Consequently, any parabolic subgroup of a reductive group is multiplicatively generated.

4. Counter-Example and Structural Nuance:
   An example of a solvable group is provided where $G_{\text{sem}}$ is a proper subgroup of $G$. In this case, the center of $G_{\text{sem}}$ is shown to be one-dimensional and unipotent, illustrating that $G_{\text{sem}}$ can possess unipotent elements even though it is generated by semisimple ones.

Significance and Scope
The paper claims to provide a complete structural description of the characteristic subgroup $G_{\text{sem}}$ within the context of characteristic zero. The significance lies in extending the known property of reductive groups (being generated by semisimple elements) to the broader class of non-reductive groups by identifying exactly which part of the unipotent radical is "captured" by the semisimple generation process.

The author explicitly notes the limitations of the current approach: the results rely heavily on techniques specific to algebraically closed fields of characteristic zero, such as the exponential map, full reducibility of representations, and specific Lie algebra decompositions. The paper does not claim these results transfer directly to positive characteristic, stating that new ideas would be required to address arbitrary ground fields.

Open Problems
The paper concludes by posing four specific problems related to the generation of algebraic groups:

• Problem 1 & 2: Determining the minimal number of $G_a$ or $G_m$ subgroups required to generate a given additively or multiplicatively generated group.
• Problem 3 & 4: Determining the minimal number $l$ (or $m$) such that any element in the group can be written as a product of at most $l$ unipotent (or $m$ semisimple) elements.
  • The author notes that Alexey Galt provided a solution to Problem 3: if the semisimple part $G_{\text{ss}}$ is adjoint, $l=2$; otherwise, $l=3$.

The paper also touches upon the concept of "A-images" and "M-images" (varieties admitting surjective morphisms from affine spaces or punctured affine spaces), linking the generation of groups by one-parameter subgroups to the unirationality of homogeneous spaces.