Lefschetz Homomorphism

• Lefschetz Homomorphism is a map that assigns an induced endomorphism on homology or cohomology groups from a self-map, with its alternating trace defining the Lefschetz number.
• It extends from classical fixed-point theory to combinatorial, dg-categorical, and dynamical systems approaches, providing tools for topological invariance and intersection pairings.
• In commutative algebra, the term refers to multiplication maps by powers of a linear form, crucial for understanding and verifying weak and strong Lefschetz properties in graded Artinian algebras.

The expression Lefschetz homomorphism does not denote a single universally fixed construction across mathematics. In the classical fixed-point-theoretic setting, it is the induced endomorphism on homology or cohomology associated with a self-map, and the corresponding Lefschetz number is the alternating trace

$$L(f)=\sum_q (-1)^q \operatorname{Tr}(f_*|_{H_q}) \quad\text{or}\quad L(f)=\sum_k (-1)^k \operatorname{tr}(f^*|_{H^k}).$$

In other settings, the same phrase or closely related language refers to multiplication maps by powers of a linear form, pairing-induced maps in Poincaré–Lefschetz duality, push-pull operators attached to correspondences, or induced actions on generalized homology theories (Staecker, 2013, Tu, 2022, Migliore et al., 2016, Laudenbach, 2020).

1. Classical fixed-point-theoretic meaning

For a smooth self-map $f:M\to M$ of a compact oriented manifold $M$, the classical Lefschetz number is defined by the alternating trace formula

$$L(f)=\sum_k (-1)^k\,\mathrm{tr}\!\left(f^*\mid H^k(M)\right).$$

When fixed points are isolated and nondegenerate, this equals the sum of local fixed-point indices,

$L(f)=\sum_{p=f(p)} \operatorname{ind}_p(f).$

In this setting, the operative homomorphism is the induced map $f^*$ on cohomology, or equivalently $f_*$ on homology (Tu, 2022).

An axiomatic characterization of the Lefschetz number for simplicial selfmaps of finite abstract simplicial complexes is given by a unique real-valued invariant $L(f,A)$ satisfying a valuation axiom

$$L(f,\emptyset)=0,\qquad L(f,A\cup B)=L(f,A)+L(f,B)-L(f,A\cap B),$$

and a simplex axiom

$L(f,x)=(-1)^{\dim x}c(f,x)+L(f,\partial x).$

Here $f:M\to M$0 records whether the simplex maps onto itself and with what orientation sign. From these axioms one recovers the chain-level trace formula

$f:M\to M$1

and, via the Hopf Trace Theorem, the homology-level formula

$f:M\to M$2

In this sense, the classical Lefschetz homomorphism is the passage from a self-map to its induced homology endomorphisms, whose alternating trace is forced by the axioms (Staecker, 2013).

The same axiomatic framework extends from simplicial maps to continuous selfmaps of compact polyhedra by simplicial approximation. The paper further shows that the usual homotopy invariance axiom can be weakened to continuity of the invariant on the mapping space $f:M\to M$3, so full homotopy invariance is stronger than necessary (Staecker, 2013).

2. Combinatorial and topological formulations

A combinatorial refinement is the combinatorial Lefschetz number for a homeomorphism $f:M\to M$4 of a simplicial complex and a definable $f:M\to M$5-invariant set $f:M\to M$6. It is defined by

$f:M\to M$7

where $f:M\to M$8 is the induced matrix on $f:M\to M$9-simplices. This construction is additive: $M$0 and in particular

$M$1

When $M$2 is open, this identifies $M$3 with the relative Lefschetz number $M$4 (López et al., 16 Jan 2026).

The central theorem in this setting is that the combinatorial Lefschetz number is a topological invariant. If $M$5 is a homeomorphism conjugating the restrictions of homeomorphisms $M$6 and $M$7, then

$M$8

For homeomorphisms, this topological invariance replaces the wedge-of-circles axiom and the cofibration axiom in the Arkowitz–Brown characterization. The same framework also yields a topological invariance theorem for the relative Lefschetz number and a generalization of O’Neill’s topological invariance of the fixed-point index (López et al., 16 Jan 2026).

The theory also extends beyond invertible maps. For open maps $M$9 with $$L(f)=\sum_k (-1)^k\,\mathrm{tr}\!\left(f^*\mid H^k(M)\right).$$0 and $$L(f)=\sum_k (-1)^k\,\mathrm{tr}\!\left(f^*\mid H^k(M)\right).$$1, the same trace-based definition is well-defined after triangulation and simplicial approximation, and it yields a fixed-point theorem: if

$$L(f)=\sum_k (-1)^k\,\mathrm{tr}\!\left(f^*\mid H^k(M)\right).$$2

then $$L(f)=\sum_k (-1)^k\,\mathrm{tr}\!\left(f^*\mid H^k(M)\right).$$3 has a fixed point in $$L(f)=\sum_k (-1)^k\,\mathrm{tr}\!\left(f^*\mid H^k(M)\right).$$4 (López et al., 16 Jan 2026).

3. Correspondences, dg-categories, and dynamical systems

For a smooth correspondence $$L(f)=\sum_k (-1)^k\,\mathrm{tr}\!\left(f^*\mid H^k(M)\right).$$5 on a compact oriented smooth manifold, with both projections $$L(f)=\sum_k (-1)^k\,\mathrm{tr}\!\left(f^*\mid H^k(M)\right).$$6 finite covering maps, the appropriate Lefschetz homomorphism is the push-pull operator

$$L(f)=\sum_k (-1)^k\,\mathrm{tr}\!\left(f^*\mid H^k(M)\right).$$7

Its Lefschetz number is

$$L(f)=\sum_k (-1)^k\,\mathrm{tr}\!\left(f^*\mid H^k(M)\right).$$8

If $$L(f)=\sum_k (-1)^k\,\mathrm{tr}\!\left(f^*\mid H^k(M)\right).$$9 meets the diagonal transversely, then the Lefschetz theorem for correspondences states

$L(f)=\sum_{p=f(p)} \operatorname{ind}_p(f).$0

This is a direct generalization of the fixed-point formula from maps to multivalued geometric data. In the same paper, the holomorphic correspondence case is formulated as conjectural, first for the structure sheaf and then for holomorphic vector bundles and Hecke correspondences (Tu, 2022).

In the dg-categorical setting, the Lefschetz construction is attached not to a map of spaces but to an endofunctor $L(f)=\sum_{p=f(p)} \operatorname{ind}_p(f).$1 of a smooth compact dg-category. Given morphisms

$L(f)=\sum_{p=f(p)} \operatorname{ind}_p(f).$2

there is an induced endomorphism

$L(f)=\sum_{p=f(p)} \operatorname{ind}_p(f).$3

and the categorical holomorphic Lefschetz formula states

$L(f)=\sum_{p=f(p)} \operatorname{ind}_p(f).$4

where $L(f)=\sum_{p=f(p)} \operatorname{ind}_p(f).$5 is the right adjoint of $L(f)=\sum_{p=f(p)} \operatorname{ind}_p(f).$6, $L(f)=\sum_{p=f(p)} \operatorname{ind}_p(f).$7 are boundary-bulk maps, and $L(f)=\sum_{p=f(p)} \operatorname{ind}_p(f).$8 is a canonical pairing on trace spaces. A second result is a reciprocity law for commuting endofunctors,

$L(f)=\sum_{p=f(p)} \operatorname{ind}_p(f).$9

which generalizes Lunts’ Lefschetz formula (Polishchuk, 2011).

A dynamical analogue appears for signed Smale spaces $f^*$0. The paper constructs signed homology groups $f^*$1 and an induced action of $f^*$2. The corresponding Lefschetz theorem identifies the alternating trace on rationalized signed homology with the signed count of periodic points: $f^*$3 Here the Lefschetz homomorphism is the induced action on the signed Putnam-type homology theory rather than on singular homology (Deeley, 2016).

4. Duality, pairings, and algebraic correspondences

In Morse theory for a compact oriented manifold $f^*$4 with nonempty boundary, the Lefschetz construction appears as an intersection pairing between the relative and absolute Morse complexes. The paper defines

$f^*$5

with

$f^*$6

and proves that it induces the classical homological pairing

$f^*$7

This is described as the Lefschetz homomorphism or Poincaré–Lefschetz pairing in Morse-complex form (Laudenbach, 2020).

An intersection-homological version is established for oriented subanalytic stratified $f^*$8-pseudomanifolds, without requiring a collared neighborhood of the boundary. For complementary perversities, the pairing-induced maps

$f^*$9

are isomorphisms. In this framework, the Lefschetz homomorphism is the map sending a relative intersection-homology class to the functional defined by intersection with Borel–Moore classes (Valette, 2010).

In the algebraic-geometric literature of the Lefschetz standard conjecture, the relevant homomorphism is an algebraic correspondence inverse to hard Lefschetz. For a smooth projective variety $f_*$0 of dimension $f_*$1, the conjecture $f_*$2 asserts that for each $f_*$3 there is an algebraic correspondence

$f_*$4

inverse to cup product by a hyperplane class,

$f_*$5

For projective irreducible holomorphic symplectic manifolds of generalized Kummer deformation type, this conjecture is proved in degrees

$f_*$6

where $f_*$7 is the smallest prime dividing $f_*$8; in particular, when $f_*$9 is prime, the Lefschetz standard conjectures hold for $L(f,A)$0 (Foster, 2023).

5. Multiplication maps and Lefschetz modules

In commutative algebra, the phrase Lefschetz homomorphism commonly refers to multiplication by powers of a general linear form in a graded Artinian algebra. For

$L(f,A)$1

with the $L(f,A)$2 general and all exponents equal, the basic maps are

$L(f,A)$3

The weak Lefschetz property is maximal rank for $L(f,A)$4 in every degree, while the strong Lefschetz property պահանջs maximal rank for all $L(f,A)$5, $L(f,A)$6. The paper explicitly refers to $L(f,A)$7 as the Lefschetz homomorphism in degree $L(f,A)$8 and exponent $L(f,A)$9 (Migliore et al., 2016).

For uniform powers of general linear forms in three variables, the paper proves that

$$L(f,\emptyset)=0,\qquad L(f,A\cup B)=L(f,A)+L(f,B)-L(f,A\cap B),$$0

has maximal rank for every degree $$L(f,\emptyset)=0,\qquad L(f,A\cup B)=L(f,A)+L(f,B)-L(f,A\cap B),$$1 and for any number $$L(f,\emptyset)=0,\qquad L(f,A\cup B)=L(f,A)+L(f,B)-L(f,A\cap B),$$2 of generators. In the almost complete intersection case

$$L(f,\emptyset)=0,\qquad L(f,A\cup B)=L(f,A)+L(f,B)-L(f,A\cap B),$$3

the maps $$L(f,\emptyset)=0,\qquad L(f,A\cup B)=L(f,A)+L(f,B)-L(f,A\cap B),$$4, $$L(f,\emptyset)=0,\qquad L(f,A\cup B)=L(f,A)+L(f,B)-L(f,A\cap B),$$5, and $$L(f,\emptyset)=0,\qquad L(f,A\cup B)=L(f,A)+L(f,B)-L(f,A\cap B),$$6 are classified by congruence classes of $$L(f,\emptyset)=0,\qquad L(f,A\cup B)=L(f,A)+L(f,B)-L(f,A\cap B),$$7, with failure of maximal rank restricted to at most one degree and often by dimension $$L(f,\emptyset)=0,\qquad L(f,A\cup B)=L(f,A)+L(f,B)-L(f,A\cap B),$$8 (Migliore et al., 2016).

A more structural abstraction is the Lefschetz module. Here one starts from a finite-dimensional commutative graded $$L(f,\emptyset)=0,\qquad L(f,A\cup B)=L(f,A)+L(f,B)-L(f,A\cap B),$$9-algebra

$L(f,x)=(-1)^{\dim x}c(f,x)+L(f,\partial x).$0

with an open convex cone $L(f,x)=(-1)^{\dim x}c(f,x)+L(f,\partial x).$1, together with a graded $L(f,x)=(-1)^{\dim x}c(f,x)+L(f,\partial x).$2-module $L(f,x)=(-1)^{\dim x}c(f,x)+L(f,\partial x).$3 and an $L(f,x)=(-1)^{\dim x}c(f,x)+L(f,\partial x).$4-invariant symmetric bilinear form $L(f,x)=(-1)^{\dim x}c(f,x)+L(f,\partial x).$5. The module is Lefschetz of degree $L(f,x)=(-1)^{\dim x}c(f,x)+L(f,\partial x).$6 if it satisfies analogues of Poincaré duality, Hard Lefschetz, and the Hodge–Riemann relations. The paper proves that when $L(f,x)=(-1)^{\dim x}c(f,x)+L(f,\partial x).$7 is decomposed over a subalgebra $L(f,x)=(-1)^{\dim x}c(f,x)+L(f,\partial x).$8 generated by elements in $L(f,x)=(-1)^{\dim x}c(f,x)+L(f,\partial x).$9, each indecomposable summand is itself a Lefschetz module over $f:M\to M$00. This supplies an algebraic decomposition theorem parallel to the decomposition theorem for morphisms of complex projective varieties (Amini et al., 3 Nov 2025).

6. Terminological scope, related notions, and common confusions

A persistent source of confusion is that several papers with “L” or “Lefschetz” terminology do not define a classical Lefschetz homomorphism. The paper “A Homeomorphism Invariant of Polyhedra” does not define a Lefschetz homomorphism, a Lefschetz number, or a fixed-point theorem. It defines instead a bigraded $f:M\to M$01-homology $f:M\to M$02 for finite simplicial complexes via a double complex built from chains and links, proves invariance under stellar subdivision, and concludes that this $f:M\to M$03-homology is a homeomorphism invariant of polyhedra. Any identification with a Lefschetz homomorphism is therefore only indirect and metaphorical (Zheng, 2011).

In the theory of finite topological spaces, a Lefschetz complex $f:M\to M$04 is a finite graded set with incidence coefficients satisfying

$f:M\to M$05

Its associated chain complex has homology $f:M\to M$06, called Lefschetz homology groups. The central comparison map is a chain homomorphism

$f:M\to M$07

from Lefschetz chains to singular chains of the associated finite $f:M\to M$08-space, inducing

$f:M\to M$09

Under an augmentability hypothesis and point-closure acyclicity, $f:M\to M$10 is an isomorphism (Kubica et al., 2019).

In equivariant combinatorics, the Lefschetz invariant of a $f:M\to M$11-monomial $f:M\to M$12-poset is an element of the $f:M\to M$13-monomial Burnside ring $f:M\to M$14,

$f:M\to M$15

and it is used to define a multiplicative generalized tensor induction map

$f:M\to M$16

Here the term “Lefschetz” refers to an alternating chain sum in a Burnside-ring setting, not to an induced endomorphism on homology (Bouc et al., 2019).

A different use appears in hyperelliptic broken Lefschetz fibrations, where a rational-valued homomorphism

$f:M\to M$17

is defined on a subgroup of the hyperelliptic mapping class group preserving a simple closed curve $f:M\to M$18. For non-separating $f:M\to M$19,

$f:M\to M$20

and for separating $f:M\to M$21,

$f:M\to M$22

This homomorphism enters the signature formula

$f:M\to M$23

so it is a signature-localization homomorphism rather than a homology endomorphism (Hayano et al., 2011).

Taken together, these usages show that Lefschetz homomorphism is best understood contextually. In fixed-point theory it is the induced action on homology or cohomology whose alternating trace gives the Lefschetz number; in duality theory it may be an intersection-pairing map to a dual group; in commutative algebra it is multiplication by powers of a linear form; and in several adjacent literatures “Lefschetz” names alternating-sum invariants or correction homomorphisms that are related to, but not identical with, the classical construction.