1 Basic concepts

Measurability is the property that allows a set, function, or other object to be assigned a size in a way that is compatible with a chosen measure. In practice, it identifies the objects on which measure theory can operate reliably. The notion appears in the study of length, area, volume, probability, and integration, where one needs a consistent framework for describing “size” and “average behavior.”

1.1 Measure spaces

A measure space consists of a set, a sigma-algebra of subsets of that set, and a measure defined on those subsets. The sigma-algebra specifies which sets are measurable, while the measure assigns a nonnegative size to each measurable set. Common examples include the real line with Lebesgue measure and a probability space with total measure one.

1.2 Sigma-algebras

A sigma-algebra is a collection of subsets closed under complements and countable unions. This closure ensures that the usual operations on sets remain within the same family of measurable sets. Sigma-algebras are the basic domains on which measures are defined, and they determine which questions about size can be answered.

1.3 Measurable sets

A measurable set is a set that belongs to the sigma-algebra of a measure space. Such sets can be assigned a measure without ambiguity in the given framework. The class of measurable sets is built to be stable under standard set operations, making it suitable for analysis and probability.

1.4 Measurable functions

A measurable function is one whose preimages of measurable sets are measurable. This property ensures that the function interacts well with the measure structure on its domain and codomain. Measurable functions include many of the standard functions encountered in analysis, and they form the proper class for integration and probability.

1.4.1 Preimages of measurable sets

The defining condition for a measurable function is usually stated using preimages: if a set in the codomain is measurable, then its inverse image under the function must be measurable in the domain. This formulation is natural because it transfers structure backward from the target to the source. It is especially convenient when checking measurability for real-valued functions, where intervals often suffice.

1.4.2 Equivalent characterizations

For functions between certain measurable spaces, measurability can be verified using alternative criteria, such as checking preimages of a generating family of sets. For real-valued functions, it is enough to test preimages of open intervals, or even rays of the form \((-\infty, a)\). These equivalent descriptions simplify many proofs and calculations.

1.5 Measurable spaces

A measurable space is a set equipped with a sigma-algebra, independent of any specific measure. It provides the abstract setting in which measurability is defined. Many constructions in analysis begin with a measurable space and later add a measure when size or integration is needed.

2 Types of measurability

Different measures and sigma-algebras lead to different notions of measurability. A set or function may be measurable in one sense and not in another, depending on the underlying structure. The main types arise from standard choices in topology, analysis, and probability.

2.1 Borel measurability

Borel measurability refers to measurability with respect to the Borel sigma-algebra, which is generated by open sets in a topological space. Borel measurable sets include open and closed sets, along with many sets formed from them by countable operations. Borel measurable functions are important in topology and analysis because they connect measurable structure with open-set structure.

2.2 Lebesgue measurability

Lebesgue measurability is tied to Lebesgue measure on the real line and related spaces. It includes all Borel sets and also many additional sets obtained by completing the measure. Lebesgue measurable functions are the standard objects in real analysis because they are broad enough for integration while remaining well behaved.

2.3 Universal measurability

A universally measurable set is measurable with respect to every complete probability measure on a given space, or in related formulations, with respect to all measures in a suitable class. This notion is stronger than Borel measurability in some settings and is useful in advanced analysis and descriptive set theory. Universal measurability captures sets that remain manageable across many measure-theoretic contexts.

2.4 Completeness and completion of measures

A measure is complete if every subset of a null set is measurable. If a measure is not complete, it can be enlarged to its completion by adding all subsets of measure zero sets. Completion often turns a more limited measurable structure into one better suited for analysis, especially when exceptional sets of measure zero must be included.

3 Measurable sets

Measurable sets are the building blocks of measure theory. They are the sets to which a measure can be consistently assigned, and they are organized so that standard set operations remain available. Their behavior under unions, intersections, and complements is one of the central advantages of sigma-algebras.

3.1 Definition in a sigma-algebra

Within a given measurable space, a measurable set is simply an element of the sigma-algebra. This definition makes measurability a structural property rather than a geometric one. The choice of sigma-algebra determines which subsets are considered measurable.

3.2 Examples of measurable sets

Typical examples include intervals on the real line, open sets in a topological space, and finite or countable sets in many common measure spaces. In probability, events are measurable sets in the sample space. More complex examples arise from countable unions and intersections of simpler measurable sets.

3.3 Non-measurable sets

Non-measurable sets are subsets excluded from the sigma-algebra. They may exist in spaces where a measure cannot be consistently extended to all subsets while preserving natural properties. Their existence shows that measurability is a genuine restriction, not a trivial feature of every set.

3.4 Closure properties

Measurable sets are designed to be stable under the main operations used in analysis. This stability allows one to build complicated measurable sets from simpler ones without leaving the measurable class. The closure properties are essential for defining measures and proving basic theorems.

3.4.1 Complements

If a set is measurable, then its complement is measurable as well. This property ensures that the measurable structure handles both a set and the region outside it. It is one of the defining features of sigma-algebras.

3.4.2 Countable unions

A countable union of measurable sets is measurable. This is crucial because many constructions in analysis use sequences of sets rather than single sets. Countable closure is one of the reasons sigma-algebras are preferred over simpler set systems.

3.4.3 Countable intersections

A countable intersection of measurable sets is measurable, as follows from closure under complements and countable unions. Intersections are often used to define limiting objects, such as sets described by infinitely many conditions. This property supports limit arguments throughout measure theory.

4 Measurable functions

Measurable functions are the functions that preserve measurable structure through preimages. They are the natural class of functions in integration, probability, and many parts of analysis. Their importance comes from the fact that many operations on functions are easiest to control when measurability is available.

4.1 Real-valued functions

For real-valued functions, measurability is commonly checked using preimages of open intervals or half-lines. Such functions include many familiar examples from calculus and analysis, such as continuous functions and monotone functions. Real-valued measurable functions are the main input for the Lebesgue integral.

4.2 Extended real-valued functions

Functions taking values in the extended real line may also be measurable. This broader setting allows one to handle unbounded functions and limits that may diverge to infinity. It is especially useful in integration theory, where nonnegative functions are often treated in this extended framework.

4.3 Functions between measurable spaces

A function between measurable spaces is measurable if the preimage of every measurable set in the codomain is measurable in the domain. This abstract definition generalizes the familiar real-valued case. It provides a uniform language for functions in algebra, topology, probability, and analysis.

4.4 Composition of measurable functions

The composition of measurable functions is measurable when the intermediate spaces are appropriately chosen. This closure property makes measurable functions into a robust category for analysis. It ensures that building more complicated functions from simpler measurable pieces does not break measurability.

4.5 Measurable mappings in topology

In topological settings, continuous maps are measurable with respect to the Borel sigma-algebras. This link connects topology and measure theory in a simple and powerful way. However, not every measurable map is continuous, so measurability is generally weaker and more flexible than continuity.

5 Measurability in integration

Integration theory depends on measurability, since only measurable functions can be integrated in the standard Lebesgue framework. Measurable sets and functions provide the domain on which approximation and limiting processes work correctly. The structure is designed so that integrals are stable under passage to limits under suitable hypotheses.

5.1 Simple functions

Simple functions take only finitely many values and are measurable when their level sets are measurable. They serve as the basic approximating functions in Lebesgue integration. Because they are easy to integrate directly, they form the foundation for defining the integral of more general functions.

5.2 Measurable functions and the Lebesgue integral

The Lebesgue integral is first defined for nonnegative measurable functions and then extended to broader classes. Measurability ensures that the integral can be constructed from simple approximations and behaves well under limits. This framework is one of the principal achievements of modern analysis.

5.3 Approximation by measurable simple functions

Many measurable functions can be approximated pointwise by sequences of measurable simple functions. This approximation is central to the construction of the integral and to proofs of convergence results. It shows that complicated measurable behavior can often be built from finitely valued components.

5.4 Convergence theorems

Key results such as the monotone convergence theorem and dominated convergence theorem rely on measurability. These theorems justify exchanging limits and integrals under suitable conditions. They are among the main reasons measurability is indispensable in analysis.

6 Measurability in probability theory

Probability theory uses measurability to define random variables, events, and expectations. A probability space is a measure space with total measure one, so measurability becomes the language for describing random outcomes in a mathematically precise way. Many probabilistic constructions depend on the same structural features that appear in integration.

6.1 Random variables

A random variable is a measurable function from a probability space to the real numbers or another measurable space. This definition ensures that events described in terms of the random variable are themselves measurable. Random variables are the basic objects of probabilistic modeling.

6.2 Events and sigma-algebras

Events are measurable subsets of the sample space, and the sigma-algebra specifies which questions about outcomes are meaningful. Different sigma-algebras correspond to different levels of available information. This organization allows probability to formalize uncertainty in a structured way.

6.3 Conditional expectation

Conditional expectation is defined using measurability relative to a sub-sigma-algebra. It represents the best measurable approximation to a random quantity based on limited information. This concept is central to modern probability and underlies martingales and many forms of statistical reasoning.

6.4 Measurable stochastic processes

A stochastic process is measurable when its dependence on time and outcome satisfies suitable measurability conditions. Such processes are studied in areas including stochastic analysis and Markov processes. Measurability ensures that one can define pathwise and probabilistic properties rigorously.

7 Advanced topics

Advanced measure theory studies how measurability behaves under products, extensions, and more elaborate constructions. These topics are important in higher-dimensional integration, probability, and abstract analysis. They show how the basic idea of measurability scales to more complex settings.

7.1 Product sigma-algebras

A product sigma-algebra is generated by measurable rectangles in a product space. It provides the natural measurable structure on Cartesian products of measurable spaces. This construction is essential for working with multiple variables and joint distributions.

7.2 Product measures

A product measure assigns measure to sets in a product space in a way compatible with the measures on each factor. It is the measure-theoretic counterpart of combining independent dimensions. Product measures are fundamental for repeated integration and multivariable probability.

7.3 Fubini and Tonelli theorems

Fubini’s theorem and Tonelli’s theorem describe when iterated integrals can be exchanged with each other and with the integral over the product space. These results depend on measurability and integrability assumptions. They are among the most useful tools in analysis and probability.

7.4 Carathéodory’s extension theorem

Carathéodory’s extension theorem provides a method for extending a premeasure on a smaller collection of sets to a complete measure on a generated sigma-algebra. It is a cornerstone of measure construction. This theorem explains how measures such as Lebesgue measure can be built from simpler starting data.

7.5 Measurability of set functions

A set function assigns values to sets, such as a measure, content, or probability. Measurability in this context is tied to how the set function behaves with respect to the measurable structure and limiting processes. Such considerations are important in advanced measure theory and in the study of capacities and related notions.

Several concepts are closely related to measurability but arise from different viewpoints or weaker structures. These notions often appear in topology, descriptive set theory, and real analysis. They help clarify the boundaries between measurable, topological, and approximate notions of size and regularity.

8.1 Baire measurability

Baire measurability is defined using the Baire sigma-algebra, which is generated by continuous functions or by compactly controlled topological structure in suitable spaces. It is closely related to Borel measurability but can differ in some settings. This notion is useful in topology and functional analysis.

8.2 Analytic sets

Analytic sets are projections of Borel sets in product spaces and often enjoy strong regularity properties. They are not necessarily Borel, but they are typically measurable in many standard contexts. Analytic sets occupy an important place in descriptive set theory and advanced measure theory.

8.3 Outer measure

An outer measure assigns size to all subsets, not just measurable ones, and is used as an intermediate step in constructing measures. Measurable sets are then identified through a criterion that makes the outer measure behave like a true measure on them. This approach is central to the development of Lebesgue measure.

8.4 Almost everywhere concepts

An almost everywhere statement holds except on a set of measure zero. These notions are closely tied to measurability because they allow one to ignore exceptional null sets while preserving meaningful properties. Almost everywhere equivalence is pervasive in integration, probability, and functional analysis.