Andrey Markov

Andrey Andreyevich Markov (Russian: Андре́й Андре́евич Ма́рков; 14 June 1856 – 20 July 1922) was a Russian mathematician, widely regarded as one of the founders of modern probability theory and the originator of the theory of stochastic processes now known as Markov chains. A central figure of the St. Petersburg mathematical school and a student of Pafnuty Chebyshev, Markov made important contributions to number theory, continued fractions, and the theory of approximating functions, but he is chiefly remembered for his demonstration that the law of large numbers extends to sequences of dependent random variables. His work laid the theoretical foundation for a vast range of modern applications, from statistical mechanics and information theory to computational algorithms such as Markov chain Monte Carlo and PageRank. He was also known for his uncompromising character and his outspoken civic and political stances in late imperial Russia.

Early Life and Education

Markov was born in Ryazan, Russian Empire, the son of Andrey Grigorievich Markov, a minor civil servant who worked in forestry and railway administration. The family moved to St. Petersburg during his childhood. From an early age Markov suffered from a serious ailment of the knee that left him walking on crutches for much of his life, and as a schoolboy he was an indifferent student in most subjects apart from mathematics, in which he showed extraordinary precocity. A frequently related anecdote holds that, before receiving formal instruction in the subject, he independently devised a method for solving linear differential equations—which he called the "method of cascades"—and was delighted to learn that Newton had long since developed similar ideas.

In 1874 Markov entered the physico-mathematical faculty of St. Petersburg University, where he came under the decisive influence of Pafnuty Chebyshev, the leading Russian mathematician of the era. He graduated in 1878, receiving a gold medal for outstanding work, and began an academic career at the university that would span nearly half a century.

Academic Career

Markov defended his master's thesis, On Binary Quadratic Forms with Positive Determinant, in 1880 and his doctoral thesis, On Certain Applications of Algebraic Continued Fractions, in 1884, both at St. Petersburg University. He served as a privat-docent from 1880 and was appointed professor in 1893. In 1896 he was elected to the Imperial St. Petersburg Academy of Sciences, becoming a full academician in 1905. In that same year, having reached the retirement age then in force, he formally left the university but continued to teach there unofficially until near the end of his life. Among his students were mathematicians such as Jacob V. Uspensky, and his lectures and textbook shaped the training of a generation of Russian probabilists.

Early Scientific Work

Markov's earliest significant contributions lay in number theory and the theory of Diophantine approximation. His investigation of indefinite binary quadratic forms led to what are now called Markov numbers and the Markov spectrum: he determined precisely the values that the expression |x² + y² − 3xy| takes for coprime integers x and y, showing that the smallest values are √5, 2√2, √221/5, and so on, each realized by triples of integers (the Markov triples 1,1,1; 1,1,2; 1,2,5; …) generated through a branching structure later termed the Markov tree. This work proved influential in the subsequent theory of continued fractions and Diophantine approximation, and the associated unicity problem studied by Frobenius remains a topic of research to this day.

In analysis, Markov developed and extended Chebyshev's work on continued fractions and the moment problem, clarifying the connections between continued fractions, orthogonal polynomials, and convergent approximations. Together with his younger brother Vladimir (1871–1897), a talented mathematician who died young of tuberculosis, he is credited with the Markov brothers' inequality, which bounds the derivatives of polynomials in terms of the polynomials themselves—a foundational result of approximation theory. Markov also contributed to the method of least squares and to extremal problems concerning Chebyshev polynomials.

Probability Theory and the Birth of Markov Chains

Markov's most consequential scientific achievement was his creation of the theory of dependent trials, undertaken in the first two decades of the twentieth century. The immediate stimulus was a dispute with the Moscow mathematician Pavel Nekrasov, who maintained—drawing on religious and moralistic arguments—that independence of random trials was a necessary condition for the law of large numbers to hold. In his landmark 1906 paper, Markov proved that the law of large numbers extends to sequences of random variables that are not independent but form what he called a "simple chain," in which the probability of each outcome depends only on the preceding one. He subsequently developed the theory further, proving in later works versions of the central limit theorem for such chains and studying chains with finitely and countably many states.

These dependent sequences were given his name only posthumously; today they are universally known as Markov chains, and the defining property—that the future depends on the past only through the present—is called the Markov property. The generalization of his ideas to continuous time and general state spaces, carried out notably by Andrei Kolmogorov, William Feller, and others, produced the modern theory of Markov processes, one of the central frameworks of probability theory. Markov also systematized his probabilistic teaching in the textbook Ischislenie veroyatnostey (The Calculus of Probabilities), first published in 1900 and subsequently translated into German, which served for decades as a standard reference in Russia and abroad.

The Eugene Onegin Analysis

In 1913, on the 76th anniversary of Pushkin's death, Markov presented to the Academy of Sciences a celebrated empirical study, "An Example of a Statistical Investigation of the Text of Eugene Onegin Illustrating the Linkage of Trials into Chains." Treating the succession of letters in the poem's first 20,000 characters as a chain of dependent trials, he classified each letter as a vowel or a consonant and computed the frequencies of transitions: he found that a vowel is followed by another vowel in roughly 12.8 percent of cases, while a consonant is followed by a vowel in about 66.3 percent of cases. From these data he derived the stationary distribution of vowels and consonants (about 43.2 percent and 56.8 percent, respectively), showing that the model converges to equilibrium in accordance with his theorems. Markov later extended the analysis to samples of 100,000 letters from Pushkin, Aksakov, and other authors. The study is now recognized as a pioneering application of probabilistic methods to linguistics and as an early instance of modeling text statistically—anticipating, in spirit, later work on information theory and the statistical analysis of language.

Political Engagement and Personality

Markov was famed for his blunt, principled character and his readiness for public controversy. An atheist and an opponent of autocracy, he protested in 1901 against the excommunication of Leo Tolstoy by writing to the Holy Synod requesting that he himself be excommunicated as well. In 1902, when the Academy of Sciences invalidated the election of the writer Maxim Gorky under government pressure, Markov publicly broke with the institution, sent a letter to the Tsar renouncing imperial decorations, and suspended his participation in Academy affairs for several years. His polemics with Nekrasov were likewise conducted in sharply public terms, and he took evident pleasure in demonstrating, mathematically, the falsity of claims he regarded as ideologically motivated. Contemporaries described him as austere, exacting, and incorruptible, though devoted to his family and to his students.

Personal Life and Family

In 1883 Markov married Maria Ivanovna Valvatieva, and the couple had one son, Andrei Andreevich Markov (1903–1979), who became a distinguished mathematician in his own right. The son, working in Moscow, made major contributions to mathematical logic, the theory of algorithms, and the foundations of constructive mathematics; notably, concepts such as Markov algorithms and Markov's principle are named after the son, whereas Markov chains and Markov processes commemorate the father—a distinction that requires care in the literature.

Death

In his final years, weakened by illness and by the hardship and famine that followed the revolution and civil war, Markov remained scientifically active. He died in Petrograd on 20 July 1922, of sepsis following an operation on his long-troubled knee.

Legacy and Influence

Markov's creation of the theory of dependent random variables transformed probability from a calculus of independent trials into a general theory of stochastic processes. The structures he introduced now permeate science and technology: statistical mechanics and chemical kinetics rely on Markov models of random evolution; queuing theory, genetics, demography, finance, and epidemiology use them as standard tools; hidden Markov models underpin speech recognition and bioinformatics; Markov chain Monte Carlo methods have become a cornerstone of Bayesian computation; and Google's PageRank algorithm is based on the stationary distribution of a chain over web pages. In theoretical mathematics, his name is attached to an extensive family of concepts, including Markov chains, Markov processes, the Markov property, Markov's inequality, the Chebyshev–Markov–Stieltjes inequalities, Markov numbers, the Markov spectrum and Markov constant, Markov chains in information theory, and the Gauss–Markov theorem in statistics. Historians of mathematics accordingly rank him, together with Chebyshev and Lyapunov, among the principal architects of the St. Petersburg school of probability and a founder of the modern probabilistic worldview.