Measure Theory: Lebesgue Integration, Sigma-Algebras, and Convergence Theorems
Introduction
Measure theory provides the mathematical foundation for integration and probability. It answers the question of what it means to measure the size of a set — its length, area, volume, or probability — and develops the tools needed to integrate functions over these sets in a rigorous and general way.
The subject was born from necessity. In the nineteenth century, mathematicians discovered that the Riemann integral, despite its power, could not handle certain functions and limits. Dirichlet’s function — the indicator of the rationals — is not Riemann integrable. The need for a more general theory of integration that could handle such functions and exchange limits with integrals led Lebesgue to develop his theory in 1902.
Measurable Spaces
A measurable space consists of a set X together with a sigma-algebra — a collection of subsets closed under complement, countable unions, and containing the empty set. Elements of the sigma-algebra are called measurable sets. They are the sets for which a measure can be defined.
Sigma-Algebras
The Borel sigma-algebra on R is the smallest sigma-algebra containing all open intervals. It includes open sets, closed sets, countable intersections of open sets (Gδ sets), countable unions of closed sets (Fσ sets), and much more. Every set likely to arise in practice is Borel measurable.
The Lebesgue measurable sets extend the Borel sets by adding all subsets of Borel sets of measure zero. This completion of the Borel sigma-algebra allows the Lebesgue measure to be defined on a larger collection of sets while preserving the measure of Borel sets.
Measures
A measure is a function μ defined on a sigma-algebra that assigns a nonnegative extended real number to each measurable set, satisfying countable additivity: the measure of a disjoint union of countably many sets equals the sum of their measures.
Lebesgue Measure
Lebesgue measure on R extends the notion of length. The Lebesgue measure of an interval is its length. Countable sets have measure zero. The Cantor set — an uncountable set with measure zero — shows that cardinality and measure are independent.
A key achievement of measure theory is the construction of a non-measurable set using the axiom of choice. Vitali’s construction partitions [0, 1] into uncountably many cosets of Q, then selects one point from each coset. The resulting set cannot be assigned a Lebesgue measure consistent with countable additivity and translation invariance. This result connects measure theory to set theory.
Lebesgue Integration
The Lebesgue integral builds functions from simple functions — finite linear combinations of characteristic functions of measurable sets. The integral of a nonnegative measurable function is the supremum of integrals of simple functions below it. General functions are integrated by separating into positive and negative parts.
Comparison with Riemann Integration
The Riemann integral partitions the domain into intervals and approximates the function by rectangles. The Lebesgue integral partitions the range into levels and measures the preimage of each level. This reversal of perspective is crucial: the Lebesgue integral works with a much wider class of functions and behaves much better under limits.
Every Riemann integrable function on a bounded interval is Lebesgue integrable, and the integrals agree. But many functions that are not Riemann integrable are Lebesgue integrable. The Lebesgue integral is the natural extension of the Riemann integral to a complete space of integrable functions.
Convergence Theorems
The Lebesgue integral excels where the Riemann integral fails: exchanging limits and integrals. The three fundamental convergence theorems are the cornerstones of measure theory.
Monotone Convergence Theorem
The Monotone Convergence Theorem states that if a sequence of nonnegative measurable functions increases pointwise to f, then the integrals converge to the integral of f. This theorem allows term-by-term integration of series of nonnegative functions.
Dominated Convergence Theorem
The Dominated Convergence Theorem is the most powerful tool: if a sequence of measurable functions converges pointwise almost everywhere to f, and if there is an integrable function g dominating all functions in absolute value, then the integrals converge to the integral of f.
These theorems are essential in functional analysis, where they justify the completeness of L^p spaces and the interchange of limits in many contexts.
Product Measures
Given two measure spaces, the product sigma-algebra is generated by rectangles A × B. The product measure satisfies μ × ν(A × B) = μ(A)ν(B). Fubini’s theorem states that for integrable functions, iterated integrals equal the double integral, regardless of the order of integration.
Applications
Fubini’s theorem allows the interchange of the order of integration in multiple integrals, provided the function is absolutely integrable. The theorem requires absolute integrability: the failure of this condition can lead to different iterated integrals having different values.
Product measures and Fubini’s theorem are essential for probability theory, where they describe independent random variables and expectations of products.
L^p Spaces
The L^p spaces are Banach spaces of functions whose p-th power is integrable. For 1 ≤ p < ∞, the L^p norm is ‖f‖_p = (∫ |f|^p dμ)^(1/p). For p = ∞, the L^∞ norm is the essential supremum. Hölder’s inequality relates L^p and L^q norms for conjugate exponents 1/p + 1/q = 1.
The completeness of L^p spaces — the Riesz-Fischer theorem — is one of the fundamental results of measure theory. It establishes that L^p spaces are Banach spaces, providing the foundation for functional analysis and the study of partial differential equations.
Interpolation of L^p Spaces
The Riesz-Thorin interpolation theorem shows that if an operator is bounded on L^p and on L^q, then it is bounded on L^r for all r between p and q. This theorem is essential in harmonic analysis, where it is used to prove boundedness of Fourier transform and singular integral operators.
Marcinkiewicz interpolation extends these ideas to weak-type estimates. The connection between L^p spaces and functional analysis is central to modern analysis.
Signed Measures and Radon-Nikodym
A signed measure extends measures by allowing negative values. The Hahn decomposition theorem states that any signed measure can be decomposed into the difference of two positive measures supported on disjoint sets.
Radon-Nikodym Theorem
The Radon-Nikodym theorem states that if ν is absolutely continuous with respect to μ (meaning μ(A) = 0 implies ν(A) = 0), then there exists a function f such that ν(A) = ∫_A f dμ. The function f is the Radon-Nikodym derivative.
This theorem is fundamental in probability theory, where it guarantees the existence of conditional expectations. It also underlies the definition of the derivative of measures in economics, physics, and real analysis.
Haar Measure
Haar measure is a translation-invariant measure on locally compact topological groups. The existence and uniqueness of Haar measure was proved by Haar in 1933. On the real line, Haar measure is Lebesgue measure. On the circle group, Haar measure is arc length.
The theory of Haar measure connects measure theory to topology and group theory. The convolution of functions on a group, defined using Haar measure, turns the space of integrable functions into a Banach algebra. The representation theory of compact groups uses Haar measure to construct invariant inner products.
Applications in Harmonic Analysis
Fourier analysis on groups uses Haar measure to define the Fourier transform. The Pontryagin duality theorem describes the dual group of characters and shows that the Fourier transform maps functions on a group to functions on its dual.
The theory of almost periodic functions uses Haar measure on the Bohr compactification. The connection between measure theory and harmonic analysis is essential for understanding the structure of L^p spaces and convolution operators.
Integration on Manifolds
The theory of integration on manifolds uses measure theory to integrate differential forms. The Riemannian volume form provides a canonical measure on a Riemannian manifold. The divergence theorem and Stokes theorem are expressed in terms of integrals of differential forms over manifolds with boundary.
The theory of currents, developed by Federer and Fleming, generalizes integration on manifolds to objects with less regularity. Currents are dual to differential forms and satisfy compactness properties that make them useful in geometric measure theory and the calculus of variations.
Geometric Measure Theory
Geometric measure theory studies sets with finite perimeter — Caccioppoli sets — and rectifiable sets that are built from Lipschitz images of Euclidean space. The area formula generalizes the change of variables formula to Lipschitz maps, and the coarea formula relates integrals over level sets.
The Plateau problem — finding the surface of minimal area spanning a given boundary — is solved using geometric measure theory. The existence of minimizers in the class of integral currents was proved by Federer and Fleming. The regularity of minimizers was studied by De Giorgi, Almgren, and others, connecting measure theory to differential geometry.
What is a measurable set? A measurable set is a member of a sigma-algebra — a collection of sets closed under complement and countable unions that forms the domain of a measure.
What is the difference between Riemann and Lebesgue integration? Riemann integration partitions the domain; Lebesgue integration partitions the range. Lebesgue integration handles a wider class of functions and has better convergence properties.
Why is measure zero important? Properties that hold everywhere except on a set of measure zero are said to hold almost everywhere. Many theorems in analysis hold almost everywhere rather than everywhere.
How is measure theory used in probability? Probability is a measure space where the total measure is 1. Random variables are measurable functions. Expected values are integrals. Convergence theorems justify limit theorems.
Real Analysis Guide — Functional Analysis — Set Theory Guide