Birational Automorphism Groups

• Birational automorphism groups are defined as groups of invertible rational maps on algebraic varieties, capturing key birational invariants and classification properties.
• Reduction methods such as regularization, MRC fibration, and the Minimal Model Program reveal structural constraints by embedding finite subgroups into semilinear groups.
• Finite subgroup phenomena, including nilpotency of class two and Heisenberg-type structures, establish deep links between group properties and the birational type of varieties.

A birational automorphism group is the group of self-maps of an algebraic variety, defined as birational isomorphisms: these are rational maps that are invertible on a dense open subset, and their composition gives a group structure. Such groups encode deep birational and arithmetic properties of varieties, and their finite subgroups, structural constraints, and connections with cone decompositions have driven significant advances in algebraic geometry, birational classification, and group theory.

1. Definition and Foundational Properties

Given a variety $X$ (typically over an algebraically closed field of characteristic zero), the birational automorphism group $\operatorname{Bir}(X)$ consists of all $k$-birational self-maps of $X$. These are equivalence classes of isomorphisms between dense open subsets, and $\operatorname{Bir}(X)$ is a birational invariant of $X$ itself. The group may be equipped with topologies such as the Zariski or family-theoretic topology, which correspond to algebraic families of birational maps (Chen et al., 2 Dec 2025).

Nilpotently Jordan Property

A group $G$ is nilpotently Jordan of class at most $c$ if there exists a uniform integer $J>0$ such that every finite subgroup $H\subseteq G$ contains a nilpotent subgroup $K\leqslant H$ of nilpotency class at most $c$ and index at most $J$ (Guld, 2020). For $\operatorname{Bir}(X)$, the following sharp result is established:

Theorem: For any variety $X$ over characteristic zero, the birational automorphism group $\operatorname{Bir}(X)$ is nilpotently Jordan of class at most two: every finite subgroup contains a nilpotent-of-class-2 subgroup of uniformly bounded index (Guld, 2020).

This property unifies and strengthens previous Jordan-type theorems and serves as a structural control on finite subgroups.

2. Structure and Reduction Methods

Comprehensive reduction arguments underpin the study of the structure of $\operatorname{Bir}(X)$. The major methodological steps include:

• Regularization: For any finite $G\subseteq \operatorname{Bir}(X)$, one may construct a smooth projective model $X'$ and replace $G$ by its action on $X'$ via biregular automorphisms, possibly after passing to a subgroup of bounded index (Guld, 2020).
• MRC Fibration and Fibration Tower: Utilizing the maximally rationally connected (MRC) fibration $\varphi\colon X \dashrightarrow Z$, where $Z$ is non-uniruled, the fiber structure and the resulting base-fiber exact sequence allows for inductive arguments, central in bounding nilpotency classes (Guld, 2019).
• Minimal Model Program (MMP): Running a $G$-equivariant MMP produces a sequence of Mori fiber spaces and contractions, successively extracting structural constraints and functorial representations into semilinear groups.

These reduction techniques enable embeddings of finite subgroups into semilinear automorphism groups and application of group-theoretic results (Jordan's theorem for $\mathrm{GL}(n)$, Chermak–Delgado theorem).

3. Birational Automorphism Groups and Variety Type

The nature of $\operatorname{Bir}(X)$ is intimately connected to the birational type of $X$:

• Ruled Varieties: For $X$ birational to $Z\times \mathbb{A}^1$, $\operatorname{Bir}(X)$ determines $X$ up to birational equivalence and automorphism of the base field (Chen et al., 2 Dec 2025). The translations and normalizers of such factors organize the entire group, and field automorphisms represent a fundamental ambiguity.
• Non-uniruled Varieties: For non-uniruled $X$, $\operatorname{Bir}(X)$ can remain invariant under taking products with high-genus curves, demonstrating that $\operatorname{Bir}(X)$ often cannot distinguish between birational types in this regime (Chen et al., 2 Dec 2025).
• Uniruled Non-ruled: In this intermediate case (e.g., Fano threefolds, conic bundles), the structure of $\operatorname{Bir}(X)$ is not sufficient to determine the birational type; the detailed behavior here remains open.

4. Subgroup Structure: Finite Groups, $p$-Groups, and Heisenberg Phenomena

Finite subgroups of $\operatorname{Bir}(X)$ admit strong constraints:

• Index and Nilpotency Class: By the nilpotently Jordan property, every finite $H \leqslant \operatorname{Bir}(X)$ contains a nilpotent-of-class-2 subgroup $K$ of index at most $J$.
• Optimality and Heisenberg Groups: Examples such as surfaces birational to $E\times \mathbb{P}^1$ (with $E$ elliptic) demonstrate the sharpness: Heisenberg-type $p$-groups can embed into $\operatorname{Bir}(X)$, being nilpotent of class 2 but not abelian (Guld, 2020).
• $p$-Group Criterion for Rationality: For rationally connected $X$ of dimension $n$, the presence of a large prime elementary abelian $p$-subgroup (rank $n$) in $\operatorname{Bir}(X)$ implies $X$ is rational (Xu, 2018), highlighting a group-theoretic characterization of rational varieties.

Case	Subgroup Property	Index/Bound Type
Ruled surface $E \times \mathbb{P}^1$	Heisenberg $p$-group, class 2	Nilpotent, sharp, not abelian
Rationally connected, dim $n$	$(\mathbb{Z}/p\mathbb{Z})^n$ for $p \gg 0$	Forces rationality
Non-uniruled	Finite, abelian	Jordan property holds

5. Movable Cones, Coxeter Groups, and Cone Conjecture

The geometry of the movable cone relates $\operatorname{Bir}(X)$ to Coxeter group actions, fundamental domains, and cone decompositions:

• Wehler-Type Calabi–Yau Manifolds: $\operatorname{Bir}(X)$ is isomorphic to the universal Coxeter group $UC(n+1)$, acting via Picard–Lefschetz reflections on divisor classes. Movable and nef cones are identified as chambers and Tits cones in Coxeter theory (Cantat et al., 2011, Yáñez, 2021).
• Fractal Boundaries: For $n + 1 \ge 4$, the boundary of the movable cone is fractal, matching limit sets of Kleinian groups and, in specific dimensions, Apollonian gaskets (Cantat et al., 2011). This reveals intricate accumulation loci for divisor classes under birational automorphisms.
• Cone Conjecture: The Morrison–Kawamata movable cone conjecture is verified in Coxeter–movable cone settings. The birational automorphism group organizes the movable cone as a union of rational polyhedral chambers, each corresponding to the action of a distinct birational automorphism (Yáñez, 2021).

6. Special Cases and Applications: Surfaces, Picard Number Two

For projective varieties of Picard number two, $\operatorname{Bir}_2(X)$ (group of pseudo-automorphisms) admits a precise dichotomy:

• Discrete or "Almost Infinite Cyclic": The group is finite if any extremal ray is rational, and is almost infinite cyclic otherwise, i.e., it contains a cyclic subgroup of finite index (Zhang, 2013).
• Cone Structure Verification: In the klt Calabi–Yau setting with Picard number two, the cone conjecture is achieved with explicit rational polyhedral fundamental domains organized by $\operatorname{Bir}_2(X)$ actions.

7. Open Problems and Directions

Key unresolved problems and conjectures include:

• Structural classification of $\operatorname{Bir}(X)$ for varieties that are uniruled but not ruled.
• Uniform bounds for the Jordan constant $J$ in terms of dimension alone, regardless of birational class (Guld, 2019).
• Birational characterization problems: for which $X$ does $\operatorname{Bir}(X)$ determine birational type up to field automorphism (Chen et al., 2 Dec 2025).
• Extension to positive characteristic and to infinite subgroups beyond the algebraic regime.

These lines of enquiry continue to shape modern birational geometry and group theory.

---

In summary, birational automorphism groups encode tight structural constraints, admit deep connections to the birational type and rationality of the underlying variety, enforce strong nilpotency and Jordan-type bounds, and organize the geometry of cones via Coxeter group and reflection structures. Their finite subgroups, especially in the context of reduction via MMP and fibration theory, offer a unifying algebraic view bridging group theory and modern algebraic geometry (Guld, 2020, Guld, 2019, Chen et al., 2 Dec 2025, Cantat et al., 2011, Yáñez, 2021, Zhang, 2013, Xu, 2018).