Cohomologically Trivial Automorphisms

• Cohomologically trivial automorphisms are maps that leave the (co)homology rings unchanged, linking geometry, arithmetic, and representation theory.
• They are characterized by strict bounds and classifications in surfaces, threefolds, and Kähler manifolds using tools like Lefschetz’s formula and canonical inequalities.
• Their study employs methodologies including deformation theory, lattice analysis, and the Chevalley–Eilenberg complex to reveal deep topological and algebraic properties.

A cohomologically trivial automorphism is an automorphism of an algebraic or geometric object that acts identically on its (co)homology or cohomology ring, typically with integral, rational, or torsion coefficients. The study of such automorphisms provides deep insight into the interplay between geometric structures, group actions, and their topological invariants, and serves as a technical bridge between geometry, arithmetic, and representation theory.

1. General Definition and Foundational Properties

Let $X$ be a (smooth) complex projective variety, compact Kähler manifold, or, in an algebraic context, a scheme of finite type over a field. The automorphism group $\Aut(X)$ acts via pullback on the graded (co)homology rings $H^*(X, R)$ for suitable coefficient rings $R$ (typically $\mathbb{Z}$, $\mathbb{Q}$, or their adèles). The subgroup of cohomologically trivial automorphisms is defined as

$\Aut_{R}(X) := \left\{ g \in \Aut(X) : g^*|_{H^*(X, R)} = \mathrm{id} \right\}.$

For integral coefficients, one refers to $\Aut_{\mathbb{Z}}(X)$; for rational (or numerical) coefficients, $\Aut_{\mathbb{Q}}(X)$. The group $\Aut_{\mathbb{Z}}(X)$ is always a closed subgroup of $\Aut(X)$0, and the inclusions

$\Aut(X)$1

hold. In the absence of cohomological torsion, these subgroups coincide.

Cohomologically trivial automorphisms are subject to stringent geometric and arithmetic constraints. In particular, objects of general type or high Kodaira dimension typically admit only trivial (identity) cohomologically trivial automorphisms, while low Kodaira dimension or certain group actions can allow finite or even infinite groups of such automorphisms.

2. Surfaces of General Type: Bounds and Classification

On minimal surfaces of general type, the structure of cohomologically trivial automorphisms exhibits exceptional rigidity, depending acutely on topological and arithmetic invariants. For a smooth projective surface $\Aut(X)$2 of general type, one can generally assert:

• When $\Aut(X)$3 (fake quadrics, including surfaces isogenous to a product), the largest possible group order is $\Aut(X)$4, realized for $\Aut(X)$5 isogenous to a product with $\Aut(X)$6; typically, $\Aut(X)$7 is trivial except possibly in the case $\Aut(X)$8, where $\Aut(X)$9 (Catanese et al., 26 Jan 2026).
• For $H^*(X, R)$0, $H^*(X, R)$1; this bound is sharp and achieved precisely for surfaces isogenous to a product of unmixed type, with both $H^*(X, R)$2 and $H^*(X, R)$3 possible (Cai et al., 2013).
• For $H^*(X, R)$4, the only candidates for nontrivial $H^*(X, R)$5 arise when $H^*(X, R)$6 and $H^*(X, R)$7, in which case exactly a single involution ($H^*(X, R)$8) occurs; for $H^*(X, R)$9 one is always rationally cohomologically rigidified (Cai et al., 2012).

In threefolds of general type, cohomologically trivial automorphisms are tightly bounded in the maximal Albanese dimension case: for minimal projective threefolds with only Gorenstein quotient singularities and generically finite Albanese map, $R$0, dropping to $R$1 if $R$2 is ample and $R$3 is nonsingular. For threefolds isogenous to a product of curves, possible groups are $R$4, $R$5, or $R$6 (Zhao, 2019).

These results are summarized in the table below:

Invariants	Group Type(s)	Max Order	Reference
Gen. type, $R$7	$R$8 (for $R$9), $\mathbb{Z}$0, $\mathbb{Z}$1, exceptional cases	192	(Catanese et al., 26 Jan 2026)
Gen. type, $\mathbb{Z}$2	$\mathbb{Z}$3, $\mathbb{Z}$4, cyclic	4	(Cai et al., 2013)
Gen. type, $\mathbb{Z}$5	$\mathbb{Z}$6 (exceptional)	2	(Cai et al., 2012)
Threefolds, max. Alb. dim.	$\mathbb{Z}$7, up to $\mathbb{Z}$8, $\mathbb{Z}$9 in unmixed prod.	6	(Zhao, 2019)

Finite bounds for these groups rely on the interplay of canonical volume inequalities (Bogomolov–Miyaoka–Yau, Clifford-Severi), Lefschetz fixed-point formulas, and comparison of quotient maps with controlled singularities.

3. Elliptic and Enriques Surfaces: Finiteness and Group Structures

Elliptic Surfaces

For minimal properly elliptic surfaces, the behavior of cohomologically trivial automorphisms depends acutely on the Euler characteristic $\mathbb{Q}$0:

• If $\mathbb{Q}$1, $\mathbb{Q}$2 is finite, with only $\mathbb{Q}$3, $\mathbb{Q}$4, or $\mathbb{Q}$5 possible ($\mathbb{Q}$6), and explicit examples realizing the first two occur. In the infinite-component case (pseudo-elliptic), there are at most two components in the automorphism group (Catanese et al., 2024).
• For $\mathbb{Q}$7 and $\mathbb{Q}$8, the uniform bound $\mathbb{Q}$9 can be attained; if the elliptic fibration is isotrivial, then $\Aut_{R}(X) := \left\{ g \in \Aut(X) : g^*|_{H^*(X, R)} = \mathrm{id} \right\}.$0, again sharp. If $\Aut_{R}(X) := \left\{ g \in \Aut(X) : g^*|_{H^*(X, R)} = \mathrm{id} \right\}.$1 and the fibration admits an additive fibre, then $\Aut_{R}(X) := \left\{ g \in \Aut(X) : g^*|_{H^*(X, R)} = \mathrm{id} \right\}.$2 must be trivial—further, while it might grow with $\Aut_{R}(X) := \left\{ g \in \Aut(X) : g^*|_{H^*(X, R)} = \mathrm{id} \right\}.$3 or the bigenus, it is always less than $\Aut_{R}(X) := \left\{ g \in \Aut(X) : g^*|_{H^*(X, R)} = \mathrm{id} \right\}.$4 (Catanese et al., 2024).

Enriques Surfaces

Let $\Aut_{R}(X) := \left\{ g \in \Aut(X) : g^*|_{H^*(X, R)} = \mathrm{id} \right\}.$5 be an Enriques surface over an algebraically closed field of arbitrary characteristic. The groups of numerically and cohomologically trivial automorphisms satisfy:

• For classical $\Aut_{R}(X) := \left\{ g \in \Aut(X) : g^*|_{H^*(X, R)} = \mathrm{id} \right\}.$6 ($\Aut_{R}(X) := \left\{ g \in \Aut(X) : g^*|_{H^*(X, R)} = \mathrm{id} \right\}.$7), $\Aut_{R}(X) := \left\{ g \in \Aut(X) : g^*|_{H^*(X, R)} = \mathrm{id} \right\}.$8 is cyclic, order at most $\Aut_{R}(X) := \left\{ g \in \Aut(X) : g^*|_{H^*(X, R)} = \mathrm{id} \right\}.$9. The numerically trivial group is cyclic of order $\Aut_{\mathbb{Z}}(X)$0 or $\Aut_{\mathbb{Z}}(X)$1, and the quotient is at most $\Aut_{\mathbb{Z}}(X)$2.
• In characteristic $\Aut_{\mathbb{Z}}(X)$3, for supersingular $\Aut_{\mathbb{Z}}(X)$4, exceptional cases allow larger groups: odd cyclic groups of orders in $\Aut_{\mathbb{Z}}(X)$5 or the quaternion group $\Aut_{\mathbb{Z}}(X)$6 (Dolgachev et al., 2017, Dolgachev, 2012).

The classification depends on the interplay between the Picard group, the lattice structure of $\Aut_{\mathbb{Z}}(X)$7, and the detailed geometry of half-fibers and genus-one pencils.

4. Kähler, Hyperkähler, and Complete Intersection Manifolds

Compact Kähler and Hyperkähler Manifolds

The faithfulness of the cohomological representation is a central question. For Kähler manifolds, the chain of normal subgroups

$\Aut_{\mathbb{Z}}(X)$8

encodes increasingly coarse topological invariants, culminating in cohomological triviality. For low Kodaira dimension, cohomologically trivial automorphism groups can be arbitrarily large, even realizing cyclic groups of arbitrary order on blow-ups of $\Aut_{\mathbb{Z}}(X)$9 (Catanese et al., 2020).

For compact hyperkähler manifolds:

• In dimension $\Aut_{\mathbb{Q}}(X)$0, $\Aut_{\mathbb{Q}}(X)$1 acts faithfully on $\Aut_{\mathbb{Q}}(X)$2; i.e., every numerically trivial automorphism is the identity (Jiang et al., 2023).
• For deformation types associated to Hilbert schemes of $\Aut_{\mathbb{Q}}(X)$3 surfaces and generalized Kummer manifolds, the only automorphisms acting trivially on $\Aut_{\mathbb{Q}}(X)$4 are the identity map, even though the action on $\Aut_{\mathbb{Q}}(X)$5 can have large kernel (translations/tensorizations) (Oguiso, 2012).
• For O'Grady's 10-dimensional example, the kernel is trivial; for the 6-dimensional case it is $\Aut_{\mathbb{Q}}(X)$6, generated by $\Aut_{\mathbb{Q}}(X)$7-torsion points and line bundles on the underlying abelian surface, all acting trivially on $\Aut_{\mathbb{Q}}(X)$8 (Mongardi et al., 2014).

Complete Intersections

For smooth complete intersections in projective space, the cohomological representation is generically faithful except for odd-dimensional intersections of two quadrics. There, the kernel of the action is $\Aut_{\mathbb{Q}}(X)$9 for dimension $\Aut_{\mathbb{Z}}(X)$0 odd (Chen et al., 2015).

5. Algebraic and Noncommutative Settings: Lie Algebras and Frobenius Algebras

Lie Algebras

For an extension of Lie algebras

$\Aut_{\mathbb{Z}}(X)$1

with $\Aut_{\mathbb{Z}}(X)$2 abelian, a pair $\Aut_{\mathbb{Z}}(X)$3 of automorphisms of $\Aut_{\mathbb{Z}}(X)$4 and $\Aut_{\mathbb{Z}}(X)$5 is called cohomologically trivial (or inducible) if a certain attached extension cocycle in $\Aut_{\mathbb{Z}}(X)$6, constructed using the Chevalley–Eilenberg complex, vanishes. These automorphisms precisely lift to automorphisms of $\Aut_{\mathbb{Z}}(X)$7 compatible with the original short exact sequence, and fit into an exact sequence

$\Aut_{\mathbb{Z}}(X)$8

(Bardakov et al., 2015).

Frobenius Algebras and Hochschild Cohomology

For a Frobenius algebra $\Aut_{\mathbb{Z}}(X)$9 with Nakayama automorphism $\Aut(X)$00, $\Aut(X)$01 acts trivially on Hochschild cohomology $\Aut(X)$02: $\Aut(X)$03 Any automorphism preserving the Frobenius structure is cohomologically trivial on Hochschild cohomology, and the homotopical triviality can be explicitly exhibited at the level of cochains. This is foundational for the construction of invariants such as the Nakayama Jacobian, Nakayama divergence, and potentially for extensions to Calabi–Yau and derived settings (Suárez-Álvarez, 6 Feb 2025).

6. Methodologies and Proof Techniques

Cohomological triviality is analyzed through the following principal tools and techniques:

• Lefschetz Fixed-Point Formula: Connects trace computations on cohomology to the geometry of fixed loci of automorphisms, which is essential in ruling out or bounding the order of cohomologically trivial automorphisms.
• Volume and Canonical Class Inequalities: Inequalities bounding $\Aut(X)$04, such as the Miyaoka–Yau and Clifford–Severi inequalities, are combined with quotient maps to constrain possible group orders.
• Lattice Theory and Monodromy: Understanding the action of automorphisms on the intersection lattice (Picard or Néron–Severi groups) and especially on integral 2-cohomology is fundamental in the hyperkähler and Enriques cases.
• Specialization and Deformation Theory: Deformation invariance of the kernel of the cohomological action, as well as density arguments in moduli, prove faithfulness or determine explicit group structures.
• Chevalley–Eilenberg Complex: In noncommutative and Lie algebra contexts, cohomological triviality is encoded transparently by explicit cocycle computations.

7. Open Problems and Future Directions

• Boundedness in Elliptic and General Type Cases: For elliptic surfaces with $\Aut(X)$05, it remains open whether an absolute uniform bound on $\Aut(X)$06 exists, independent of $\Aut(X)$07. The maximum order, attained for $\Aut(X)$08, is $\Aut(X)$09 (Catanese et al., 2024).
• Rigidity and Faithfulness in Higher Dimension: In higher-dimensional compact hyperkähler varieties (beyond fourfolds), the full scope of faithfulness for the cohomological representation remains unresolved (Jiang et al., 2023).
• Group Structure and Torsion: The precise classification in terms of the interplay between geometric structure, arithmetic invariants, and the presence of torsion in cohomology is only fully understood in select regimes.
• Stack-Theoretic and Moduli Implications: Cohomologically trivial automorphism groups influence the stratification of moduli spaces, jumping loci, and the topological structure of diffeomorphism groups (Catanese et al., 2020).
• Extensions to Noncommutative Geometry: Generalizing the analysis of cohomologically trivial automorphisms to Gorenstein, Calabi–Yau, and derived settings is a frontier area with connections to invariants and deformation theory (Suárez-Álvarez, 6 Feb 2025).
• Differentiable Rigidity vs. Cohomological Triviality: The distinction between cohomological, homotopical, and diffeomorphic triviality raises subtle questions, particularly in the context of bounded domain quotients (Cai et al., 2012).

• (Zhao, 2019) Automorphisms of 3-folds of general type acting trivially on cohomology
• (Dolgachev, 2012) Numerical trivial automorphisms of Enriques surfaces in arbitrary characteristic
• (Cai et al., 2013) Automorphisms of surfaces of general type with q=1 acting trivially in cohomology
• (Cai et al., 2012) Automorphisms of surfaces of general type with q≥2 acting trivially in cohomology
• (Mongardi et al., 2014) Automorphisms of O'Grady's Manifolds Acting Trivially on Cohomology
• (Suárez-Álvarez, 6 Feb 2025) The action of the Nakayama automorphism of a Frobenius algebra on Hochschild cohomology
• (Chen et al., 2015) Automorphism and Cohomology II: Complete intersections
• (Jiang et al., 2023) On numerically trivial automorphisms of compact hyperkähler manifolds of dimension 4
• (Dolgachev et al., 2017) Numerically trivial automorphisms of Enriques surfaces in characteristic $\Aut(X)$10
• (Oguiso, 2012) No cohomologically trivial non-trivial automorphism of generalized Kummer manifolds
• (Catanese et al., 2020) On topologically trivial automorphisms of compact Kähler manifolds and algebraic surfaces
• (Catanese et al., 26 Jan 2026) Cohomologically or numerically trivial automorphisms of surfaces of general type
• (Catanese et al., 2024) On the cohomologically trivial automorphisms of elliptic surfaces I: $\Aut(X)$11
• (Catanese et al., 2024) On the numerically and cohomologically trivial automorphisms of elliptic surfaces II: $\Aut(X)$12
• (Bardakov et al., 2015) Extensions and automorphisms of Lie algebras