Home > Positioning > Persons > Eilenberg

Samuel Eilenberg (1913–1998)

Eilenberg, with Saunders Mac Lane, created category theory — the branch of mathematics that studies structure through relationships rather than through internal composition. Their 1942–45 papers introduced categories, functors, and natural transformations as a precise language for tracking how mathematical structures relate to one another across domains. The original motivation was algebraic topology: Eilenberg and Mac Lane needed to say exactly what it means for a construction in one mathematical setting to behave the same way as a construction in another. The tool they built for this purpose turned out to be far more general than the problem that prompted it. Category theory became a framework for studying mathematical structure itself — objects characterised entirely by their morphisms, with no reference to internal composition.

Life

Born 30 September 1913 in Warsaw, Russian Empire (now Poland). Studied mathematics at the University of Warsaw; PhD (1936) under Karol Borsuk, working in algebraic topology. Emigrated to the United States in 1939 — one of the last cohort of Polish mathematicians to leave before the war. Several of his teachers and colleagues in the Warsaw school were killed during the German occupation.

Positions at the University of Michigan (1940–45), then Indiana University (1946–47), then Columbia University (1947–82), where he spent the rest of his career. The Columbia years were prolific: he collaborated with Mac Lane on the category-theory foundations, with Norman Steenrod on the axiomatic foundations of homology theory, and with Henri Cartan on homological algebra. Later work extended into automata theory and formal languages, connecting algebra to theoretical computer science.

Eilenberg was also a serious collector of Southeast Asian art — bronzes, sculptures, and ceramics, primarily from India, Indonesia, and the Khmer civilisation. The collection, built over decades with a connoisseur’s eye, was donated to the Metropolitan Museum of Art in New York. He was known among mathematicians as “Professor” and among dealers as “the Professor.” Member of the National Academy of Sciences (1959). Wolf Prize in Mathematics (1986, shared with Atle Selberg). Died 30 January 1998 in New York.

Category theory

The founding papers — “General Theory of Natural Equivalences” (1945, with Mac Lane) and the preparatory work from 1942 onward — introduced the basic apparatus:

Categories. A category consists of objects, arrows (morphisms) between them, a rule for composing arrows, and an identity arrow for each object. The objects have no visible internal structure; everything the category knows about them lives in the arrows. The definition is minimal: three axioms (associativity of composition, existence of identities).

Functors. A structure-preserving map between categories: it sends objects to objects and arrows to arrows, respecting composition and identity. A functor carries the structure of one category into another without distortion.

Natural transformations. The concept that motivated the entire programme. A natural transformation is a systematic family of arrows relating one functor to another, coherent in the way they sit across the source category. Eilenberg and Mac Lane needed natural transformations to make precise a distinction that was felt but not formalised in algebraic topology: the difference between constructions that are “canonical” (natural — they arise from the structure) and constructions that depend on arbitrary choices. The naturality condition turned a vague intuition into a theorem-bearing concept.

The programme was initially regarded by many mathematicians as “abstract nonsense” — a label that was sometimes affectionate and sometimes not. The abstract apparatus proved indispensable: by the 1960s, Grothendieck’s refounding of algebraic geometry was built on categorical language, and the framework had spread into algebra, logic, topology, and eventually into computer science and theoretical physics.

The Eilenberg-Steenrod axioms

Foundations of Algebraic Topology (1952, with Norman Steenrod) axiomatised the homology and cohomology theories that algebraic topologists had been developing since Poincaré. Before Eilenberg and Steenrod, there were multiple competing definitions of homology (singular, simplicial, Čech, cellular), each with its own construction and its own proofs. Eilenberg and Steenrod identified the properties that all of these theories share — seven axioms — and proved that any theory satisfying the axioms is essentially unique. The effect was to separate what a homology theory does (its structural properties) from how it is constructed (the particular model). This is the categorical instinct applied to a specific mathematical discipline: characterise by relationships and properties, not by internal construction.

Where Eilenberg stops

Category theory’s power is its generality — the same language applies to algebra, topology, logic, and computer science. The generality is also the source of the standard criticism: the framework describes structure at a level of abstraction that can feel removed from the particular mathematical content it organises. The “abstract nonsense” charge, though often made in jest, reflects a genuine question: does category theory illuminate the mathematics it describes, or does it provide a language that mathematicians could do without? The answer has shifted over time — the Grothendieck revolution in algebraic geometry could not have been conducted without categorical language, and the same is true of modern homotopy theory and type theory. But the question recurs at each new application: whether the categorical framework adds insight or merely adds notation is a judgment made field by field.

Eilenberg’s later work in automata theory and formal languages — connecting algebraic structures (monoids, semigroups) to the classification of formal languages and computational processes — was technically substantial but less widely influential than the categorical work. The connection between algebra and computation that Eilenberg developed is now a specialised subfield (algebraic automata theory); whether it will prove as foundational as category theory is unclear.

The foundational question — whether category theory can replace set theory as the foundation of mathematics — was raised by Lawvere in the 1960s and remains unresolved. Eilenberg himself was not primarily interested in foundations; his work was motivated by the need for good tools, not by the desire for a new foundation. Whether category theory is a language, a framework, or a foundation is debated among mathematicians and philosophers of mathematics, and the question does not have a consensus answer.

Key works

See also: Mac Lane · Frege · Russell · Category theory