Loading field map
Real analysis is the branch of mathematics that studies the real numbers, sequences and series of real numbers, and the concepts of continuity, differentiation, and integration built upon them. Its subject matter is the rigorous foundation for calculus, but it extends far beyond the computational rules taught in introductory courses. At its core, real analysis asks what the real number system actually is, what it means for a function to be continuous or differentiable, and under what conditions the limiting processes that underlie calculus are legitimate.
The discipline is organized around a cluster of fundamental questions. The first concerns the nature of the real numbers themselves. The rational numbers are dense—between any two rationals lies another—yet they have gaps, such as the absence of a number whose square is two. The real numbers are defined by filling these gaps, and the precise way this is done (via Dedekind cuts, Cauchy sequences, or axiomatic completeness) determines what can be proved about them. The central property is completeness: every nonempty set of real numbers that is bounded above has a least upper bound. This single axiom underpins the existence of limits, the intermediate value theorem, and the convergence of Cauchy sequences.
The second cluster of questions concerns limits and continuity. A sequence of real numbers converges if its terms eventually get arbitrarily close to a limit. A function is continuous at a point if small changes in the input produce arbitrarily small changes in the output. These definitions, formalized with epsilon-delta arguments, replace the intuitive but imprecise notions of "approaching" and "tending toward" that characterized early calculus. The central results here include the intermediate value theorem (a continuous function on an interval takes every value between its endpoints) and the extreme value theorem (a continuous function on a closed bounded interval attains a maximum and minimum).
The third cluster concerns differentiation. The derivative measures the best linear approximation to a function at a point. Real analysis asks when derivatives exist, how they behave under composition and inversion, and what the mean value theorem—which states that a differentiable function's average rate of change over an interval is attained at some interior point—implies about the function's global behavior. A key distinction emerges between functions that are merely differentiable and those that are continuously differentiable, and between derivatives that exist and derivatives that are integrable.
The fourth cluster concerns integration. The Riemann integral, the traditional integral of calculus, partitions the domain into intervals and sums the areas of rectangles. Real analysis reveals its limitations: some functions, such as the Dirichlet function (which is 1 on rationals and 0 on irrationals), are not Riemann integrable, and the interchange of limits and integrals requires stringent conditions. The Lebesgue integral, developed in the early twentieth century, partitions the range instead of the domain, allowing a much broader class of functions to be integrated and providing more flexible convergence theorems. The relationship between these two integrals, and the conditions under which they agree, is a central theme.
The precursors to real analysis lie in the calculus of Isaac Newton and Gottfried Wilhelm Leibniz in the seventeenth century. Their calculus operated with infinitesimals—quantities conceived as infinitely small but nonzero—and with intuitive notions of continuity and limit. The results were powerful, but the foundations were shaky. In the eighteenth century, Leonhard Euler and others manipulated infinite series with great freedom, often treating them as finite sums and obtaining correct results alongside paradoxes.
The nineteenth century brought a rigorous reconstruction. Augustin-Louis Cauchy redefined limits, continuity, and convergence in terms of inequalities, eliminating infinitesimals from the definitions. Bernard Bolzano and Karl Weierstrass sharpened these ideas further, with Weierstrass introducing the epsilon-delta formulation that remains standard. Weierstrass also constructed a function that is continuous everywhere but differentiable nowhere, demonstrating that geometric intuition could not be trusted as a guide to analytic truth. Richard Dedekind and Georg Cantor provided rigorous constructions of the real numbers from the rationals, settling the question of what the real numbers are.
The theory of integration developed more slowly. Cauchy defined the integral for continuous functions; Bernhard Riemann extended it to a broader class of functions in the 1850s. But the Riemann integral proved inadequate for the needs of analysis, particularly for the interchange of limits and integrals. At the turn of the twentieth century, Henri Lebesgue constructed a new integral based on measure theory, which assigns sizes to sets in a way that generalizes length. The Lebesgue integral, together with the measure theory that underlies it, became the standard framework for much of modern analysis.
The classical approach to real analysis, codified in the late nineteenth century, is organized around the epsilon-delta definition of limits. This approach treats the real numbers as a complete ordered field and develops all concepts—continuity, derivatives, integrals—as statements about inequalities. Its strength is its precision and its direct connection to the computational practice of calculus. Its limitation is that it becomes unwieldy for abstract or infinite-dimensional settings, and it does not naturally accommodate the measure-theoretic ideas needed for advanced integration theory.
The measure-theoretic approach, initiated by Lebesgue, reframes integration in terms of the size of sets. A measure assigns a nonnegative number to certain subsets of the real line, generalizing the notion of length. The Lebesgue integral is defined by partitioning the range of a function and measuring the sets where the function takes values in each interval. This approach yields powerful convergence theorems: the monotone convergence theorem and the dominated convergence theorem give conditions under which limits can be passed through integrals. It also reveals the structure of measurable functions and sets, and it provides the foundation for probability theory, functional analysis, and partial differential equations.
The measure-theoretic approach is not a replacement for the epsilon-delta approach but a complement. The epsilon-delta definitions remain the working tools for continuity and differentiability; measure theory extends the integral and provides a richer framework for limits. Modern real analysis courses typically present both, using the Riemann integral to motivate the Lebesgue integral and then developing the latter in full.
A third approach, which emerged in the early twentieth century, views real analysis through the lens of function spaces. Instead of studying individual functions, this perspective studies spaces of functions—such as the space of continuous functions on an interval, or the Lebesgue spaces $L^p$ of functions whose p-th power is integrable. These spaces are equipped with norms or metrics, and the tools of real analysis become statements about convergence, completeness, and boundedness in these spaces. This perspective connects real analysis to functional analysis and provides the language for modern partial differential equations and harmonic analysis.
The functional-analytic perspective is not a rival to the earlier approaches but a later layer built upon them. It presupposes the measure-theoretic construction of the Lebesgue integral and the epsilon-delta theory of continuity, and it reorganizes these results into a more abstract and powerful framework.
The concepts of real analysis form a tightly interconnected web. Completeness of the real numbers implies the Bolzano-Weierstrass theorem (every bounded sequence has a convergent subsequence) and the Cauchy criterion for convergence. Continuity is defined in terms of limits, and differentiability in terms of continuity of a certain quotient. The fundamental theorem of calculus links differentiation and integration: under suitable conditions, differentiation and integration are inverse operations. The Lebesgue integral extends the Riemann integral, and the two agree for functions that are Riemann integrable, but the Lebesgue integral handles a wider class and has better convergence properties.
A central theme is the distinction between pointwise and uniform notions. A sequence of functions may converge pointwise—at each point individually—without converging uniformly, and uniform convergence is needed to interchange limits with integrals or derivatives. Similarly, a function may be continuous at each point without being uniformly continuous on an interval, and the distinction matters for the validity of certain theorems. The Arzelà-Ascoli theorem, which characterizes when a family of functions has a uniformly convergent subsequence, is a deep result that ties together compactness, equicontinuity, and pointwise convergence.
Another central theme is the role of sets of measure zero. A property that holds except on a set of measure zero is said to hold almost everywhere. Many theorems in measure-theoretic analysis hold only almost everywhere, and functions that agree almost everywhere are often identified in function spaces. The Cantor set—a set of measure zero that is uncountable and has no isolated points—illustrates the subtlety of measure and the limitations of geometric intuition.
Contemporary real analysis is a mature field whose core results are settled, but it remains an active area of research at its boundaries. The classical theory of the real numbers, continuity, and integration is taught as a fixed body of knowledge, but research continues in areas that build on real analysis: harmonic analysis, which studies functions through their decompositions into basic waves; geometric measure theory, which extends measure and integration to curved spaces; and the theory of partial differential equations, which relies heavily on the tools of real analysis.
The field also maintains a close relationship with other branches of mathematics. The Lebesgue integral is the foundation of probability theory, where expectation is an integral and independence is a statement about product measures. The function spaces of real analysis are the setting for functional analysis, and the completeness of these spaces is essential for the existence theorems of differential equations. The distinction between the Riemann and Lebesgue integrals, once a topic of foundational debate, is now a standard part of the curriculum, with the Lebesgue integral serving as the default framework for advanced work.
Real analysis is sometimes described as the "rigorous version of calculus," but this understates its scope. It is the study of the real number system in its full depth, and its methods and results permeate modern mathematics. The epsilon-delta definitions, the measure-theoretic integral, and the function-space perspective are not competing doctrines but successive and complementary layers of a single edifice, each built to address limitations in the previous one and each retaining its own domain of applicability.