Skip to content

Quasi-Invariant Measure

A measure whose class of null sets, though not necessarily its numerical values, is preserved by every transformation in a specified action, so each pushforward remains equivalent to the original and changes density through a Radon–Nikodym cocycle.

Version
v2 · 2026-08-30 · History
Domain-specific #
2600
Origin domain
measure theory
Subdomain
nonsingular transformations and group actions

Core Idea

Let a group \(G\) act by measurable automorphisms on a measurable space \((X,\mathcal B)\). A measure \(\mu\) is quasi-invariant under this action when every pushforward \(g_*\mu\), defined by

\[ (g_*\mu)(A)=\mu(g^{-1}A), \]

is equivalent to \(\mu\): each is absolutely continuous with respect to the other. Equivalently, transformations preserve the measure's null sets,

\[ \mu(A)=0 \quad\Longleftrightarrow\quad \mu(gA)=0, \]

even though they may change the positive numerical measure assigned to \(A\).

Quasi-invariance is therefore invariance of a measure class, not invariance of a measure's values. If \(g_*\mu=\mu\) for every \(g\), the measure is invariant and automatically quasi-invariant.

Scope of Application

The concept appears in measure theory, nonsingular ergodic theory, harmonic analysis, representation theory, probability on infinite-dimensional spaces, stochastic analysis, and geometry.

For group actions on homogeneous spaces \(G/H\), quasi-invariant measures allow integration and unitary representation construction even when no invariant measure exists. The Radon–Nikodym factor corrects the action so that operators preserve the \(L^2\) norm.

In nonsingular dynamics, a transformation preserves measure class rather than measure. Orbit structure and null-set statements remain meaningful, while the derivative cocycle captures compression and expansion. This supports ratio ergodic theory and type classifications beyond probability-preserving systems.

Clarity

The clearest diagnostic uses a null-set probe. Choose any measurable \(A\). If \(\mu(A)=0\), must its image or preimage under each transformation also have zero measure—and conversely? If yes for every set and action element, the measure is quasi-invariant.

A density-change calculation then explains the positive sets. Suppose on \(\mathbb R\), \(\mu(dx)=e^{-x^2/2}dx\), and translate by \(a\).

Manages Complexity

Exact invariance is often unavailable or unnecessarily strong. Quasi-invariance preserves the information needed for almost-everywhere analysis: a property true outside a null set remains true outside a transformed null set. This lets analysts transport function spaces, integrals, and probabilistic statements through transformations without demanding unchanged volume.

Abstract Reasoning

Measure equivalence is transitive. If \(g_*\mu\sim\mu\) and \(h_*\mu\sim\mu\), then \((gh)_*\mu\sim\mu\). The derivative satisfies a cocycle relation such as

\[ J_{gh}(x)=J_g(x)\,J_h(g^{-1}x) \]

under one common pushforward convention. This relation guarantees consistency under composition.

Knowledge Transfer

The definition transfers literally across homogeneous spaces, dynamical systems, Gaussian path spaces, and geometric transformations because all use the same pushforward–equivalence–density structure. What changes is the source of \(J_g\): a Jacobian determinant, a modular function, a Cameron–Martin exponential, or another cocycle.

The term should not be transferred metaphorically to approximate stability, robustness, or “almost unchanged” data. Quasi-invariance is exact at the null-set level. A small numerical change can fail it, and a large density change can satisfy it.

Relationships to Other Abstractions

Local relationship map for Quasi-Invariant MeasureParents 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.Quasi-InvariantMeasureDOMAINPrime abstraction: Invariance — is a kind ofInvariancePRIME

Current abstraction Quasi-Invariant Measure Domain-specific

Parents (1) — more general patterns this builds on

  • Quasi-Invariant Measure is a kind of Invariance Prime

    Quasi-Invariant Measure strictly instantiates Invariance at the level of null sets and measure class.

Hierarchy path (1) — routes to 1 parentless root

Neighborhood in Abstraction Space

Quasi-Invariant Measure sits in a sparse region of the domain-specific corpus (78th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Advanced Probability & Combinatorial Bounds (6 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-09-08