Skip to content

Kirwan map

Restrict equivariant cohomology classes of a Hamiltonian group space to a regular moment-map level set and identify them with ordinary cohomology classes on the resulting symplectic quotient.

Version
v3 · 2026-09-19 · History
Domain-specific #
2132
Domain group
Formal Sciences
Origin domain
Mathematics
Subdomains
Symplectic Geometry, Hamiltonian Reduction and Equivariant Cohomology → Mathematics

Core Idea

For a Hamiltonian \(G\)-space with moment map \(\mu\) and regular value \(0\), the Kirwan map is the graded-ring homomorphism \(\kappa:H_G^*(M)\to H^*(\mu^{-1}(0)/G)\) obtained by restriction to \(\mu^{-1}(0)\) followed by the equivariant-to-quotient identification under the relevant freeness or orbifold hypotheses.[1] An equivariant class on the ambient Hamiltonian space restricts to the invariant level set, and when the group action there has the required regularity its equivariant cohomology represents the cohomology of the reduced space; Morse-theoretic analysis of the moment-map norm supplies the classical surjectivity theorem.

Its autonomous residual is the canonical ambient-to-reduced cohomology homomorphism attached to Hamiltonian reduction, not an arbitrary restriction map, moment map, quotient projection, or blanket claim that every reduction has the same cohomology. The identity fails when the target is not a justified reduced space, the level is singular without a revised framework, equivariant and ordinary cohomology are conflated, coefficients are suppressed where they matter, or Kirwan surjectivity is quoted outside its hypotheses.

Recognition requires an analyst to declare the group action, moment map and level, verify regularity and action hypotheses, specify coefficients, write the restriction and identification maps, and separate definition of the homomorphism from any surjectivity, kernel, or presentation theorem. Once established, it supports transporting characteristic and equivariant classes to symplectic quotients, computing quotient cohomology, describing kernels, and connecting Hamiltonian reduction with Morse and localization methods without turning those uses into the definition.

How would you explain it like I'm…

 

No faithful explanation at this level. At this level the Kirwan map becomes 'cutting out a piece of a shape and squashing it'. That is exactly the arbitrary restriction or quotient projection the concept is not. The picture also drops the cohomology, the equivariant-to-quotient identification and the regularity hypotheses.

Big Shape to Small Shape Recipe

Some mathematicians study shapes that are being moved around by a group of symmetries, along with a special measuring rule called a moment map. From such a shape you can build a smaller shape in two steps: keep only the points where the measurement reads exactly zero, then treat any points the symmetries move into each other as one single point. The Kirwan map is a fixed recipe that takes symmetry-respecting information about the big shape and turns it into information about the smaller shape. It only works when the zero level is 'nice' and the symmetries behave well there.

Reduction Cohomology Map

In symplectic geometry, a group G can act on a space M in a way described by a moment map μ, and the 'reduced space' is built by taking the level set where μ equals zero and dividing out by G. The Kirwan map is a homomorphism from the equivariant cohomology of M, which tracks topology while respecting the symmetry, to the ordinary cohomology of the reduced space. It works in two steps: restrict classes to the zero level set, then identify equivariant cohomology there with the cohomology of the quotient, which is valid when G acts freely or nearly so and 0 is a regular value. It is not just any restriction map or quotient projection. Whether the map is onto, the famous Kirwan surjectivity theorem, is a separate result proved with Morse theory, and it only holds under its stated hypotheses.

 

Let M be a Hamiltonian G-space with moment map μ, and suppose 0 is a regular value. The Kirwan map κ: H_G^*(M) → H^*(μ^{-1}(0)/G) is a homomorphism of graded rings. It is defined by first restricting equivariant classes to the invariant level set μ^{-1}(0). Then, if G acts freely there (or with finite stabilizers, in the orbifold setting with suitable coefficients), one uses the isomorphism H_G^*(μ^{-1}(0)) ≅ H^*(μ^{-1}(0)/G). The resulting map is canonical for Hamiltonian reduction; it is not just any restriction map or quotient projection. Kirwan's surjectivity theorem comes from Morse theory applied to the function |μ|², but it is a theorem about the map, not its definition. The construction breaks down at singular levels, when equivariant and ordinary cohomology are conflated, or when coefficients that matter are ignored. Its uses include pushing characteristic classes to symplectic quotients, computing quotient cohomology through kernels and presentations, and linking reduction to localization.

Structural Signature

  • Carrier: a Hamiltonian action of a compact Lie group on a symplectic manifold, a moment map, a suitably regular level set, and the associated symplectic quotient
  • Inputs or antecedent state: group action, moment map, chosen level, regularity and compactness hypotheses, equivariant coefficient ring, restriction to the level set, quotient identification, and cohomological grading
  • Constitutive operation: An equivariant class on the ambient Hamiltonian space restricts to the invariant level set, and when the group action there has the required regularity its equivariant cohomology represents the cohomology of the reduced space; Morse-theoretic analysis of the moment-map norm supplies the classical surjectivity theorem
  • Invariant: the map has the ambient equivariant cohomology as source, the reduced-space cohomology as target, and is induced by level-set restriction plus the justified quotient identification
  • Recognition test: declare the group action, moment map and level, verify regularity and action hypotheses, specify coefficients, write the restriction and identification maps, and separate definition of the homomorphism from any surjectivity, kernel, or presentation theorem
  • Output or consequence: transporting characteristic and equivariant classes to symplectic quotients, computing quotient cohomology, describing kernels, and connecting Hamiltonian reduction with Morse and localization methods
  • Failure boundary: the target is not a justified reduced space, the level is singular without a revised framework, equivariant and ordinary cohomology are conflated, coefficients are suppressed where they matter, or Kirwan surjectivity is quoted outside its hypotheses

What It Is Not

  • It is not the whole field of symplectic geometry; many objects in that field do not satisfy its constitutive rule.
  • It is not its canonical example. For a compact Hamiltonian \(G\)-manifold with \(0\) a regular value and a free action on \(\mu^{-1}(0)\), restriction followed by \(H_G^*(\mu^{-1}(0))\cong H^*(M/\!/G)\) gives the Kirwan map to the smooth quotient. That is an instance, not a definition.
  • It is not Symplectic Reduction. Symplectic reduction constructs the quotient space from a moment-map level set. The Kirwan map is the induced cohomological map from the unreduced equivariant theory to the quotient's ordinary cohomology.
  • It is not an unrestricted metaphor. If the group action on the regular level set is only locally free, the quotient can be an orbifold and coefficient choices affect the clean identification; singular values require intersection-cohomology or other modified constructions rather than automatic reuse

Scope of Application

Kirwan map applies when the analyst can specify a Hamiltonian action of a compact Lie group on a symplectic manifold, a moment map, a suitably regular level set, and the associated symplectic quotient and establish that the map has the ambient equivariant cohomology as source, the reduced-space cohomology as target, and is induced by level-set restriction plus the justified quotient identification. The entry centers the classical cohomological construction; singular, noncompact, K-theoretic, stacky, and derived variants require separately stated hypotheses and targets.[2]

  • Recognition. declare the group action, moment map and level, verify regularity and action hypotheses, specify coefficients, write the restriction and identification maps, and separate definition of the homomorphism from any surjectivity, kernel, or presentation theorem
  • Comparison. Compare legitimate instances through compactness, group, coefficient ring, moment-map level, regularity, free versus locally free action, source grading, target singularity, surjectivity, kernel, and localization data.
  • Boundary. If the group action on the regular level set is only locally free, the quotient can be an orbifold and coefficient choices affect the clean identification; singular values require intersection-cohomology or other modified constructions rather than automatic reuse
  • Use. Preserve every assumption when using the identity for transporting characteristic and equivariant classes to symplectic quotients, computing quotient cohomology, describing kernels, and connecting Hamiltonian reduction with Morse and localization methods.

Clarity

A clear claim names the carrier, governing rule, assumptions, and recognition test. This matters because Kirwan map can be used for related K-theory or generalized-cohomology maps, and prose often merges the map's definition with Kirwan's surjectivity theorem. The disciplined statement is that the object counts as Kirwan map exactly when the map has the ambient equivariant cohomology as source, the reduced-space cohomology as target, and is induced by level-set restriction plus the justified quotient identification

Identity and measurement remain separate. The map and its properties are established algebraically and topologically from the action and level-set hypotheses; a dimension count or computational presentation alone does not prove the required identification. Approximation or noisy evidence may weaken a classification without changing its definition.

Manages Complexity

The abstraction compresses abelian and nonabelian actions, smooth and orbifold quotients, different regular levels, generalized cohomology theories, K-theoretic analogues, and singular-reduction extensions into a stable carrier, rule, invariant, and failure boundary. It makes comparison tractable while retaining the variables that control validity.

Compression can hide assumptions. A responsible use therefore declares compactness, group, coefficient ring, moment-map level, regularity, free versus locally free action, source grading, target singularity, surjectivity, kernel, and localization data and returns to the full diagnostic whenever a convention or boundary case changes.

Abstract Reasoning

  1. Type the carrier. Establish a Hamiltonian action of a compact Lie group on a symplectic manifold, a moment map, a suitably regular level set, and the associated symplectic quotient and reject examples from a different problem.
  2. Lock the rule. Express that the map has the ambient equivariant cohomology as source, the reduced-space cohomology as target, and is induced by level-set restriction plus the justified quotient identification independently of one notation or implementation.
  3. Derive carefully. Infer transporting characteristic and equivariant classes to symplectic quotients, computing quotient cohomology, describing kernels, and connecting Hamiltonian reduction with Morse and localization methods only under the stated assumptions.
  4. Stress-test. Contrast the legitimate boundary case—If the group action on the regular level set is only locally free, the quotient can be an orbifold and coefficient choices affect the clean identification; singular values require intersection-cohomology or other modified constructions rather than automatic reuse—with this counterexample: the moment map itself is a geometric map from the manifold to the dual Lie algebra, not the Kirwan map between cohomology rings.

Knowledge Transfer

Transfer within symplectic geometry is strong when new cases preserve the same carrier, mechanism, and diagnostic. The move from For a compact Hamiltonian \(G\)-manifold with \(0\) a regular value and a free action on \(\mu^{-1}(0)\), restriction followed by \(H_G^*(\mu^{-1}(0))\cong H^*(M/\!/G)\) gives the Kirwan map to the smooth quotient. to In a torus reduction, one can use fixed-point and moment-polytope data to identify classes that vanish after reduction and thereby present the quotient's cohomology ring. demonstrates that continuity.[3]

Outside the domain, only the skeleton—restrict invariant information from an ambient structured space to a constraint locus and then descend it through a quotient—travels automatically. The terms Hamiltonian action, moment map, equivariant cohomology, regular value, level set, symplectic quotient, restriction, surjectivity, kernel, and localization retain domain-specific meanings, so every role and inference must be revalidated.

Examples

Canonical

For a compact Hamiltonian \(G\)-manifold with \(0\) a regular value and a free action on \(\mu^{-1}(0)\), restriction followed by \(H_G^*(\mu^{-1}(0))\cong H^*(M/\!/G)\) gives the Kirwan map to the smooth quotient. The map is defined before one asks whether it is onto; compactness and the Morse theory of the squared moment map are among the hypotheses behind the classical surjectivity result. It is canonical because the carrier, rule, invariant, and consequence are all inspectable.[1]

Mapped back: a Hamiltonian action of a compact Lie group on a symplectic manifold, a moment map, a suitably regular level set, and the associated symplectic quotient → An equivariant class on the ambient Hamiltonian space restricts to the invariant level set, and when the group action there has the required regularity its equivariant cohomology represents the cohomology of the reduced space; Morse-theoretic analysis of the moment-map norm supplies the classical surjectivity theorem → the map has the ambient equivariant cohomology as source, the reduced-space cohomology as target, and is induced by level-set restriction plus the justified quotient identification → transporting characteristic and equivariant classes to symplectic quotients, computing quotient cohomology, describing kernels, and connecting Hamiltonian reduction with Morse and localization methods

Applied / In Practice

In a torus reduction, one can use fixed-point and moment-polytope data to identify classes that vanish after reduction and thereby present the quotient's cohomology ring. Localization data assist the computation, but the destination remains the chosen reduced space and the kernel statement must retain regular-value, coefficient, and compactness assumptions. It qualifies only after the same diagnostic and failure boundary are checked.[2]

Mapped back: declared instance → recognition test → boundary check → qualified use

Structural Tensions

  • T1: Exact identity vs. practical recognition. The constitutive condition may be exact while evidence is indirect. Diagnostic: Can the reviewer state both the condition and the warrant?
  • T2: Canonical form vs. variants. abelian and nonabelian actions, smooth and orbifold quotients, different regular levels, generalized cohomology theories, K-theoretic analogues, and singular-reduction extensions can preserve or change the identity. Diagnostic: Which named role is invariant across the variants?
  • T3: Compression vs. hidden assumptions. The label is useful only while prerequisites remain visible. Diagnostic: Can each downstream inference be traced to a declared assumption?
  • T4: Autonomy vs. reduction. The candidate uses broader structures but claims the canonical ambient-to-reduced cohomology homomorphism attached to Hamiltonian reduction, not an arbitrary restriction map, moment map, quotient projection, or blanket claim that every reduction has the same cohomology. Diagnostic: Does that residual still support independent recognition after the parent and neighbors are subtracted?

Structural–Framed Character

The entry is structurally mixed but domain-framed. Its portable skeleton is restrict invariant information from an ambient structured space to a constraint locus and then descend it through a quotient; its identity-bearing terms are Hamiltonian action, moment map, equivariant cohomology, regular value, level set, symplectic quotient, restriction, surjectivity, kernel, and localization. Those terms determine admissible objects, evidence, and consequences inside symplectic geometry.

Structural Core vs. Domain Accent

The structural core is a carrier governed by An equivariant class on the ambient Hamiltonian space restricts to the invariant level set, and when the group action there has the required regularity its equivariant cohomology represents the cohomology of the reduced space; Morse-theoretic analysis of the moment-map norm supplies the classical surjectivity theorem and tested by declare the group action, moment map and level, verify regularity and action hypotheses, specify coefficients, write the restriction and identification maps, and separate definition of the homomorphism from any surjectivity, kernel, or presentation theorem. The domain accent is constitutive rather than decorative, so an analogy that preserves only the skeleton is not another instance of Kirwan map.

This entry is a kind of Function (Mapping).

The Kirwan map literally assigns each ambient equivariant cohomology class a class on the reduced space through a specified composite homomorphism; Hamiltonian reduction and its hypotheses supply the specialization. What makes it its own entry is the canonical ambient-to-reduced cohomology homomorphism attached to Hamiltonian reduction, not an arbitrary restriction map, moment map, or quotient projection, and not the blanket claim that every reduction has the same cohomology.

Relationships to Other Abstractions

Local relationship map for Kirwan mapParents appear above the current abstraction, mutual partners to the right, and children below. Node labels state whether each abstraction is prime or domain-specific; colors identify relation types.Kirwan mapDOMAINPrime abstraction: Function (Mapping) — is a kind ofFunction(Mapping)PRIME

Current abstraction Kirwan map Domain-specific

Parents (1) — outgoing typed relations

  • Kirwan map is a kind of Function (Mapping) Prime

    The Kirwan map literally assigns each ambient equivariant cohomology class a class on the reduced space through a specified composite homomorphism; Hamiltonian reduction and its hypotheses supply the specialization.

Hierarchy path (1) — routes to 1 parentless root

Neighborhood in Abstraction Space

Kirwan map sits in a crowded region of the domain-specific corpus (31st percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.

Family — Manifolds, Stacks & Classifying Spaces (18 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-25

Not to Be Confused With

  • Moment map. Encodes infinitesimal Hamiltonians and selects the reduction level; it is not a cohomology homomorphism.
  • Restriction in ordinary cohomology. Lacks the equivariant source and quotient identification constitutive here.
  • Quotient map. Maps points of a level set to orbits, whereas the Kirwan map maps cohomology classes.
  • Kirwan surjectivity. A theorem about the map under hypotheses, not part of the bare definition in every generalized setting.

References

[1] Frances Kirwan, Cohomology of Quotients in Symplectic and Algebraic Geometry, Princeton University Press, 1984, ISBN 978-0-691-21456-6. registry ↩a ↩b

[2] M. F. Atiyah and R. Bott, 'The Moment Map and Equivariant Cohomology,' Topology 23(1), 1–28 (1984), DOI 10.1016/0040-9383(84)90021-1. registry ↩a ↩b

[3] Lisa C. Jeffrey and Frances C. Kirwan, 'Localization for Nonabelian Group Actions,' Topology 34(2), 291–327 (1995), DOI 10.1016/0040-9383(94)00028-J. registry ↩