Loading field map
Mathematical analysis is the branch of mathematics that studies limits, continuity, differentiation, integration, and infinite series, together with the rigorous logical foundations on which these concepts rest. Its subject matter is the precise behavior of functions and sequences, especially as they approach limiting values. Where algebra studies structure through equations and discrete operations, analysis studies processes of approximation and passage to the limit. The field's central achievement is to have turned intuitive notions of motion, accumulation, and infinitesimal change into exact statements about real and complex numbers, and then to have extended those statements to far more general settings.
At its heart, analysis asks: What does it mean for a process to converge? When can we legitimately interchange a limit with another operation, such as summation or differentiation? How do we measure the size of a set or the length of a curve when ordinary geometric intuition fails? These questions are not idle. The entire edifice of calculus—the mathematics of change and accumulation—depends on them. Without a rigorous account of limits, the derivative is merely a ratio of vanishing quantities and the integral a sum of infinitely many zero-width slices. Analysis supplies the logical scaffolding that turns these heuristics into theorems with precise conditions and exceptions.
The stakes are visible in the phenomena that analysis exists to explain. A Fourier series may converge at every point yet fail to represent the function that generated it. A function can be continuous everywhere and differentiable nowhere. An infinite series can be rearranged to sum to any value you please, or to no value at all. These pathologies are not curiosities; they mark the boundaries where naive manipulation of the infinite leads to contradiction. Analysis maps those boundaries, and in doing so reveals why some intuitive procedures are safe and others are not.
The origins of analysis lie in the calculus of the seventeenth century, developed by Isaac Newton and Gottfried Wilhelm Leibniz. Their calculus was a collection of powerful algorithms for tangents, areas, and rates of change, justified by appeals to infinitesimals—quantities "smaller than any assignable magnitude" but not zero. This was not a rigorous theory but a working practice, enormously successful in physics and geometry. The eighteenth century, dominated by Leonhard Euler and his contemporaries, extended the calculus into a general "analysis of the infinite," treating functions as expressions to be manipulated formally, often with breathtaking disregard for convergence. Euler's work on infinite series and products, for example, produced correct results through methods that later generations would deem illegitimate.
The nineteenth century brought the demand for rigor. Augustin-Louis Cauchy redefined the derivative and integral in terms of limits of finite differences and sums, eliminating infinitesimals from the foundations. Karl Weierstrass then gave the modern epsilon-delta definition of limit and continuity, reducing all of calculus to inequalities involving ordinary real numbers. This "arithmetization of analysis" reached its culmination in the work of Richard Dedekind and Georg Cantor, who constructed the real numbers themselves from the rationals, and in Cantor's theory of infinite sets, which gave precise meaning to the sizes of infinite collections. By the end of the century, analysis had become a fully rigorous discipline, and its methods had revealed a menagerie of counterexamples—continuous nowhere-differentiable functions, space-filling curves, and non-measurable sets—that forever separated the subject from naive geometric intuition.
Modern mathematical analysis is not a single monolithic enterprise but a family of approaches that share a common core while addressing different problems. The most useful way to organize the field is around the distinct foundational frameworks that have emerged, each with its own questions and tools.
The oldest and most direct approach works with functions of real or complex variables, using limits, continuity, and differentiability as the primary tools. Real analysis in this tradition studies the behavior of functions on intervals of the real line: their continuity, differentiability, integrability, and representation by power series or trigonometric series. Its central theorems—the intermediate value theorem, the mean value theorem, the fundamental theorem of calculus—are the rigorous backbone of calculus. Complex analysis, developed in the nineteenth century by Cauchy, Bernhard Riemann, and Weierstrass, studies functions of a complex variable. Its central discovery is that differentiability in the complex sense is an extraordinarily strong condition: a function differentiable once in a neighborhood is automatically differentiable infinitely often, and its values everywhere are determined by its values on any tiny curve. This rigidity produces powerful results—Cauchy's integral formula, the residue theorem, the maximum modulus principle—that have no analogue in real analysis. Complex analysis is not a separate subject from real analysis but a specialization whose extra structure yields far stronger conclusions.
A second major approach, originating in the work of Henri Lebesgue at the turn of the twentieth century, rethinks the integral from the ground up. The Riemann integral, which partitions the domain into small intervals and sums the function's values times interval lengths, fails to handle functions with infinitely many discontinuities and behaves badly under limits. Lebesgue's insight was to partition the range instead of the domain: group together all points where the function takes similar values, measure the size of each such set, and sum. This requires a theory of measure—a way to assign sizes to sets far more general than intervals or unions of intervals. The resulting Lebesgue integral is defined for a vastly larger class of functions, and it satisfies far better limit theorems: under mild conditions, the integral of a limit equals the limit of the integrals. Measure theory also resolves the ancient problem of "length" for arbitrary sets, revealing that some sets are so pathological that no consistent measure can be assigned to them. This approach, sometimes called the Lebesgue theory, became the standard foundation for probability theory, functional analysis, and much of modern analysis.
A third approach, emerging in the early twentieth century, shifts attention from individual functions to spaces of functions. Instead of asking what a single function does, functional analysis asks what properties hold for all functions in a given class, and how those functions relate to one another as points in an infinite-dimensional space. The key move is to treat a collection of functions—say, all continuous functions on an interval, or all square-integrable functions—as a vector space with a norm or inner product, and then to apply the methods of geometry and topology to this space. This perspective, developed by David Hilbert, Stefan Banach, and others, unifies disparate problems: differential equations become statements about operators on function spaces, and solving an equation becomes finding a fixed point of an operator. The central results—the Hahn–Banach theorem, the open mapping theorem, the uniform boundedness principle—are theorems about infinite-dimensional spaces that have no finite-dimensional analogues. Functional analysis is not a rival to classical analysis but a reorganization of it, one that reveals the common structure underlying many apparently different problems.
Two further approaches deserve mention for the light they shed on the foundations. Nonstandard analysis, developed by Abraham Robinson in the 1960s, resurrects the infinitesimals of Leibniz and Newton, but now rigorously, by constructing an extension of the real numbers that contains genuine infinitesimal quantities. In this framework, the derivative can be defined as an ordinary ratio of infinitesimal differences, and the integral as a genuine sum over infinitesimally thin slices. The approach is logically equivalent to standard analysis—every theorem provable in one is provable in the other—but it often makes proofs more intuitive and closer to the original calculus. Constructive analysis, by contrast, refuses to accept proofs that merely show the existence of a mathematical object without providing a method to construct it. This approach, rooted in the work of L. E. J. Brouwer and later formalized by Errett Bishop, rejects the law of excluded middle in certain contexts and requires that every existence theorem come with an explicit algorithm. It is not a separate theory of analysis but a stricter standard of proof, one that yields theorems valid in any computational setting.
The contemporary landscape of analysis is shaped by the interplay of these approaches. Measure theory and functional analysis have become the default language for large parts of the subject, particularly in the theory of partial differential equations, harmonic analysis, and probability. Classical real and complex analysis remain essential, both as the source of the field's core intuitions and as active research areas in their own right. The distinctions among approaches are not sharp boundaries but matters of emphasis: a working analyst may use measure-theoretic tools for one problem, complex-analytic methods for another, and functional-analytic frameworks to organize both.
The field's enduring questions remain those it has always asked, though in ever more general settings. How do we make sense of the infinite—infinite sums, infinite processes, infinite-dimensional spaces? What are the precise conditions under which approximation is legitimate? How do we measure the size of objects that resist ordinary geometric description? These questions have driven analysis from the calculus of Newton to the abstract spaces of Banach and Hilbert, and they continue to organize the subject today. The pathologies that nineteenth-century mathematicians discovered—the continuous but nowhere-differentiable functions, the non-measurable sets, the conditionally convergent series—are not anomalies to be eliminated but permanent features of the mathematical universe, and analysis is the discipline that has learned to live with them, to characterize them precisely, and to build theories that accommodate them.