Real analysis, the rigorous study of functions of real numbers, evolved from the practical need to understand change, continuity, and area. Its central questions concern limits, convergence, continuity, differentiation, and integration. The field's history is defined by a long struggle to ground the powerful but intuitively perilous calculus of Newton and Leibniz in a secure, arithmetical foundation, free from geometric intuition and the metaphysical vagueness of infinitesimals.
The initial phase, Infinitesimal Calculus, developed independently by Isaac Newton (fluxions) and Gottfried Wilhelm Leibniz (differentials) in the late 17th century. It was a brilliantly effective algorithmic framework but rested on the logically problematic concept of an "infinitely small" quantity. For over a century, mathematicians like the Bernoulli family, Leonhard Euler, and Joseph-Louis Lagrange operated within and expanded this framework, producing monumental results while attempting alternative justifications, such as Lagrange's theory of analytic functions and power series. This period is often characterized as Formal Manipulation Calculus, where algorithmic prowess often outpaced logical scrutiny.
A decisive turn toward rigor began in the early 19th century. Bernard Bolzano and, most influentially, Augustin-Louis Cauchy initiated the Epsilon-Delta Framework. In his Cours d'analyse (1821), Cauchy systematically replaced infinitesimals with limits defined by inequalities, giving the first coherent definitions of continuity, convergence, and the definite integral as a limit of sums. However, Cauchy's work still occasionally relied on geometric intuition and the vague concept of "infinitesimal."
The complete arithmetization of analysis was achieved in the second half of the 19th century, primarily by Karl Weierstrass and his students in Berlin. The Weierstrassian Arithmetization Program eliminated all geometric and kinematic language, basing analysis entirely on the properties of real numbers. Weierstrass introduced the modern ε-δ definitions, constructed pathological counterexamples (like a continuous, nowhere differentiable function), and emphasized uniform convergence. This program demanded a clear understanding of the real number system itself, leading to the Construction of the Real Numbers via theories by Weierstrass, Richard Dedekind (Dedekind cuts), Georg Cantor, and Charles Méray.
The late 19th and early 20th centuries saw the formalization and generalization of the Weierstrassian framework into Classical Real Analysis, characterized by point-set topology and measure theory. Cantor's set theory provided a new language. The inadequacy of the Riemann integral for advanced analysis led to Henri Lebesgue's revolutionary Lebesgue Integration Theory, which, by measuring the size of sets more subtly, created a powerful integral with superior convergence properties. This gave rise to the central modern framework of Measure-Theoretic Analysis, encompassing not only integration but also the study of function spaces like L^p spaces, foundational for functional analysis and probability.
Parallel to this mainstream, alternative foundational philosophies emerged. In the early 20th century, L.E.J. Brouwer's Intuitionism rejected the law of the excluded middle for infinite sets, proposing a constructivist revision of analysis. While not becoming the mainstream, it spurred deep investigations into the logical foundations of analysis and persists in Constructive Analysis. Similarly, Abraham Robinson's Nonstandard Analysis (1960s) used mathematical logic to rigorously resurrect infinitesimals, providing an alternative but equivalent framework to the epsilon-delta approach.
The 20th century also saw the rise of specialized, powerful methodologies. The Hardy-Littlewood Circle Method, initially developed for additive number theory, became a major technique in analytic number theory and Fourier analysis. The theory of Singular Integrals, pioneered by Alberto Calderón and Antoni Zygmund, became a cornerstone of modern harmonic analysis and partial differential equations.
Today, real analysis exists as a mature, layered discipline. The Weierstrassian Arithmetization Program provides the universal foundational language. Measure-Theoretic Analysis is the dominant advanced paradigm for pure analysis, probability, and ergodic theory. Specialized methodologies like the Calderón-Zygmund Theory of singular integrals drive cutting-edge research. Meanwhile, alternative frameworks like Nonstandard Analysis and Constructive Analysis continue as active, philosophically distinct research programs.
**