Nicolas Bourbaki:The Mathematician Who Never Existed
One of the most influential mathematicians of the twentieth century never lived. He had a daughter, a nationality, a publisher, a funeral and a seat at every important table in mathematics — and behind the name stood a secret society that is still meeting.
Founded
Dec 10, 1934
Born at
Café Capoulade
Life's work
Éléments de mathématique
Status
Active · members secret
The first Bourbaki congress, Besse-en-Chandesse, July 1935
Around 1949, the American Mathematical Society received a membership application from a mathematician named Nicolas Bourbaki. The Society’s secretary turned it down. Bourbaki, he explained, was not an individual; if the people behind the name wished to join, they could do so as an institution and pay institutional dues. Bourbaki declined to pay.1
Around the same time Ralph Boas, an editor at Mathematical Reviews, wrote in an encyclopedia that Bourbaki was a collective. Bourbaki answered with a letter of protest and a counter-rumor that spread happily through the profession: it was Boas who did not exist. “BOAS,” the story went, was an acronym for the editors of Mathematical Reviews.
It was a joke, and it was also a strategy. Nicolas Bourbaki is a collective pseudonym chosen in 1935 by a handful of young French mathematicians. None of them carried the name alone; all of them carried it together. Under it they launched the most ambitious publishing project in the history of mathematics: a series of books that rebuilds the subject from nothing, on axioms, with every definition stated and every step proved. Ninety years later, it is still being written.
§ 1Strasbourg, 1934
A Generation Missing
The story begins with a war. The First World War tore through France’s young scientists. Students of the École Normale Supérieure, the country’s nursery of mathematicians, went to the front as junior officers and died there in great numbers. When the survivors’ students came of age in the late 1920s, they found a missing generation between themselves and their teachers — and the textbooks those teachers still used.2
Henri Cartan and André Weil were both teaching at the University of Strasbourg. Their analysis course was built on Édouard Goursat’s Cours d’analyse, a respected book that had nothing to say about the mathematics being done in Göttingen in those years: Hilbert’s axioms, Emmy Noether’s abstract algebra, the new topology. Cartan came to Weil so often with questions about how to teach Stokes’s theorem that Weil finally proposed a cure.
“We are five or six friends. Let us all get together and sort out these problems once and for all.”
On Monday, December 10, 1934, six of them sat down at the Café Capoulade on the boulevard Saint-Michel in Paris. The plan was modest: write a modern treatise on analysis, finish it in about three years, and get back to research. Within months it had grown into something else. To do analysis properly, they decided, they would need algebra first; for algebra, sets; for sets, logic. The treatise had begun to eat its own foundations.
It was supposed to take three years. It has taken ninety, and it is not finished.
§ 2The table at the Capoulade
The Six Who Founded Bourbaki
AW
André Weil
1906–1998
Number theory and algebraic geometry. The instigator, and the group's intellectual engine.
HC
Henri Cartan
1904–2008
Complex analysis and topology. Weil's colleague at Strasbourg; the last founder to die, at 104.
JD
Jean Dieudonné
1906–1992
Wrote most of the drafts that became books. The loudest voice at every congress.
CC
Claude Chevalley
1909–1984
Algebra and Lie groups. The youngest founder, and the most radical about rigor.
JDe
Jean Delsarte
1903–1968
Analysis. Dean at Nancy, which became the group's second home.
RP
René de Possel
1905–1974
Analysis. Left in the early years — his wife later married Weil.
Others joined almost at once: Szolem Mandelbrojt, the geophysicist Jean Coulomb, and Charles Ehresmann among them. Nobody was elected. A newcomer was invited to a congress as a cobaye, a guinea pig, to sit through the shouting. If the guinea pig shouted back with something worth hearing, an invitation to return might follow.
Ten faces behind one name — Cartan, Weil, de Possel, Ehresmann, Schwartz (top); Dieudonné, Chevalley, Samuel, Serre, Douady (bottom)
§ 3A general, a prank, a country
Where the Name Bourbaki Came From
General Charles-Denis Bourbaki was one of the more colorful soldiers of nineteenth-century France. He won early fame, lost an army in the snow during the Franco-Prussian War of 1870–71, and tried to shoot himself; the bullet missed. Generations of French students knew the name. At the École Normale Supérieure, it also belonged to a hazing ritual: an older student in disguise would deliver a solemn, deliberately absurd lecture to first-year students, ending with a “Bourbaki’s theorem” that made no sense at all.4
The founders remembered the joke. At the first official congress, in July 1935 at Besse-en-Chandesse in the Auvergne, the name was adopted. The first name, Nicolas, was added when the group needed to sign a note to the Paris Academy of Sciences. The note had to be presented by a real academician; Élie Cartan, Henri’s father, agreed to put it forward without asking too many questions.
Then they gave him a life. Bourbaki became a citizen of Poldavia, a fictitious country that ENS students had invented for an earlier hoax. His books came from the “University of Nancago” — Nancy, where Dieudonné and Delsarte taught, fused with Chicago, where Weil had settled after the war. He had relatives, a home address and, eventually, enemies.
§ 4 · The paperwork of a man who never lived
A Wedding, an Application and a Funeral
The joke needed props. Over three decades, Bourbaki’s friends produced the documents a real life leaves behind. The wording below is paraphrased.
Wedding announcement1939
Betti Bourbaki weds H. Pétard
A printed card announced the marriage of Bourbaki's daughter Betti to Hector Pétard — himself a pseudonym, invented in 1938 by two Anglo-American mathematicians for a joke paper on the mathematical theory of hunting lions. The signatory: Nicolas Bourbaki, of the Royal Academy of Poldavia.
Membership applicationc. 1949
Bourbaki applies to the AMS
The American Mathematical Society was asked to admit Nicolas Bourbaki as a member. Its secretary refused: Bourbaki was not an individual. He could join as an institution — and pay institutional dues. Bourbaki declined.
Funeral notice1968
Nicolas Bourbaki, deceased
A mock death notice circulated in Paris announcing that Bourbaki had died on November 11, 1968. Mourners were directed to the Cemetery of Random Functions. The group, of course, kept meeting.
Greatly exaggerated
Hector Pétard, the bridegroom, was the creation of Ralph Boas and Frank Smithies — the same Boas who would later tell the world that Bourbaki was a committee. The feud had started as a wedding.
§ 5How the work was done
Inside a Bourbaki Congress
Bourbaki worked like nothing else in science. Two or three times a year the members gathered for a congrès of a week or two, usually in a quiet hotel in the French countryside. A member would arrive with a draft chapter. It was then read aloud, line by line, and torn apart by everyone else. Specialists had no protection: the algebraist was expected to attack the topology, and the topologist the algebra. At the end the draft went to someone else to rewrite, and the cycle began again at the next congress.
Shorts, sandals and set theory — a Bourbaki congress at the blackboard
“
Certainly, at first, the impression of a gathering of madmen. Two or three monologues shouted at top voice, seemingly independently of one another.
A single chapter could go through six, eight, ten complete drafts across a decade. Output was slow by design. The rule that held it all together was unanimity: nothing went to the printer until every active member accepted it. In return, the standard was absolute. There was nothing careless in a Bourbaki text, because everyone at the table had tried to find it.
The society’s unwritten rules
01Anonymity
No text carries an individual name. Everything is published as Nicolas Bourbaki.
02Unanimity
No chapter goes to print until every active member accepts it. One objection sends it back.
03Fifty
Members retire from the group by the age of fifty, so the next generation has room.
04Secrecy
The current roster is never published. Former members talk; active ones do not.
05No chair
No president, no leader, no casting vote. Everyone may attack anyone's draft.
06Guinea pigs
Promising young mathematicians are invited to a congress as cobayes. If they speak up, they may be asked back.
A congress of the late 1930s. The woman on the left is usually identified as Simone Weil, André's sister, who sat in
§ 6The life's work
Éléments de mathématique: Rebuilding Mathematics From Scratch
The title is a manifesto in two words. Éléments points straight back to Euclid, whose Elements was the first attempt to derive a whole science from a short list of assumptions — you can see how that looked on the page in the first English edition of 1570. And mathématique is singular, where French normally says mathématiques. For Bourbaki there were not many mathematics, only one.5
The order of the books is the argument. Book I is the theory of sets. Everything else is built on top of it, and nothing may be used before it has been defined. The plan drew on Hilbert’s axiomatic program and on the logic of Russell and Whitehead, but its ambition was practical rather than philosophical: to give working mathematicians one shared, reliable language.
In 1950 an essay signed by Bourbaki appeared in English, The Architecture of Mathematics. Dieudonné wrote it. It explained the plan in one image. At the bottom lie sets. On top of them stand three “mother structures”: algebraic structures, where you combine elements; order structures, where you compare them; and topological structures, where you say what it means for elements to be close. Every field of mathematics, from number theory to integration, is some combination of the three.8
Bourbaki’s architecture, simplified: one foundation, three mothers, many children
1939first volume printed
2016most recent new book
0pictures in the core volumes, by design
That last number is not quite literal — a handful of diagrams slipped in over the years — but the spirit is exact. Compare it with Oliver Byrne’s 1847 Euclid, which replaced letters with colored shapes so a reader could see every proof. Byrne and Bourbaki stand at opposite ends of the same question: should mathematics be shown, or only stated?
§ 7The words you already use
The Language Bourbaki Left Behind
Many mathematicians use Bourbaki every day without opening one of the books. The group spent decades arguing about the right word and the right symbol, and the winners spread through lecture halls everywhere. A few of them:
∅
The empty-set sign
André Weil borrowed the Norwegian letter Ø for the set with nothing in it. It entered print in the 1939 volume on sets.
Three words for three ways a function can pair two sets. Bourbaki coined them; every first-year course now uses them.
The dangerous bend
A road sign in the margin warned readers of a tricky passage. Donald Knuth later borrowed it for The TeXbook.
𝓕
Filters
Henri Cartan's 1937 device for handling limits in any topological space, made standard by the General Topology volume.
The three words in the middle of that list are worth a minute of play. A function sends each element of one set to an element of another. It is injective if no two inputs land on the same output, surjective if every output gets hit, and bijective if both are true — a perfect pairing. Build a few below and watch the labels change.
Bourbaki’s vocabulary, live
Size of A
3
Size of B
4
Send each number to a letter
1 →
2 →
3 →
InjectiveSurjectiveBijective
No two numbers share a letter, but some letter is left alone. Injective, not surjective.
Try it: a bijection needs A and B to be the same size. Make A bigger than B and no injection is possible; make it smaller and no surjection is.
That small observation — two finite sets have the same size exactly when a bijection exists between them — is how Cantor measured infinity, and it is the first idea in Halmos’s Naive Set Theory, the friendliest way into Bourbaki’s Book I. Halmos, fittingly, wrote the first popular portrait of Bourbaki, for Scientific American in 1957.9
§ 8 · The public face
The Séminaire Bourbaki
The congresses are closed. The seminar is not. Since 1948 the Séminaire Bourbaki has met in Paris, now at the Institut Henri Poincaré, three times a year. Speakers — who need not be members — explain a recent breakthrough by someone else. A Bourbaki talk is an honor, and a written exposé that newcomers to a field still read first.
1948first session
3×a year: spring, summer, autumn
1,000+exposés published
Openanyone may attend
§ 9Out at fifty
Three Generations Under One Name
The age limit was the founders’ cleverest rule. It stopped Bourbaki from aging with its members and forced each generation to hand the name to the next. Weil, true to it, left in 1956. Laurent Schwartz and Jean-Pierre Serre, both Fields medalists, belonged to the second wave; so did the American Samuel Eilenberg, one of the inventors of category theory. Pierre Cartier, Adrien Douady and later Alain Connes came with the third.
Birth to fiftieth birthday — the latest the rule would let them stay
■ Founders■ Second generation■ Third generation
The chart shows a relay. The founders’ bars end in the 1950s, just as the second generation’s are running at full speed; the second runs out in the late 1970s as the third takes over. Nobody knows exactly who holds the name now. That is the point.
A later congress — papers on laps, a hotel terrace, a week of arguments ahead
§ 10The brightest member
Grothendieck and the Break Over Foundations
Alexander Grothendieck arrived in Paris in 1948, a stateless twenty-year-old with an unusual past and almost no formal training. He sat in on the Séminaire, moved to Nancy to work with Dieudonné and Schwartz, and within a few years had rewritten the theory of topological vector spaces. By the mid-1950s he was attending congresses, and he quickly became the most productive member the group had ever had.
He was also impatient with it. Grothendieck had come to think that the right foundation was not sets and structures but categories and functors — a language about how structures relate rather than what they are made of. He pressed the group to adopt it, rewriting drafts in the new style. Bourbaki, whose Book I was already in print and whose method depended on not changing direction mid-stream, would not follow. Around 1960 Grothendieck left.6
He took the Bourbaki ambition with him and made it larger. With Dieudonné he wrote the Éléments de géométrie algébrique — the title is a deliberate echo — and the seminars that rebuilt algebraic geometry on the idea of a scheme. Those tools proved the last of Weil’s own conjectures, the analogue of the Riemann Hypothesis for curves and varieties over finite fields. Grothendieck won the Fields Medal in 1966, left his institute in 1970, and spent the end of his life alone in a village in the Pyrenees. His influence is still visible in papers like the one Thomason co-signed with a dead friend, written for Grothendieck’s sixtieth-birthday Festschrift.
“Without the language Bourbaki invented, there would be no words to describe contemporary mathematics.”
§ 11The case against
Abstraction or Sterilization?
Bourbaki was never universally admired, and the sharpest blows landed on something the group never did. In the 1950s and 60s, school reformers in France, the United States and elsewhere carried Bourbaki’s language — sets, structures, axioms — into classrooms of ten-year-olds under the banner of the “New Math.” In most places it went badly. Children learned to write unions and intersections without learning why anyone would want to. Richard Feynman sat on California’s textbook committee and wrote about it with characteristic exasperation.
Bourbaki had proposed no school reform; its name was borrowed as a seal of authority. But the critics also had a point about the books themselves. They leave out history, motivation and pictures. The Russian mathematician Vladimir Arnold, who thought geometry and physics were the lifeblood of the subject, attacked the “Bourbakization” of teaching for decades. His 1997 essay on teaching opens with a line that reads as a direct reply: “Mathematics is a part of physics.”10
In the age of Bourbaki, classical geometry fell out of fashion almost entirely; the story of the man who kept it alive is one of the best entries in our list of geometry books.
Hardy draws a similar line in A Mathematician’s Apology: pure mathematics is not useful the way a hammer is useful, and was never meant to be. And one more honest caveat belongs here. Bourbaki did not prove that mathematics rests on a complete foundation; after Gödel, nobody can. What it did was choose one coherent foundation and build on it, in public, for ninety years.
§ 12Still meeting
The Legacy of Nicolas Bourbaki
Open a graduate textbook in algebra, topology or functional analysis and you are reading Bourbaki’s choices, made in shouted debates in rural French hotels between 1935 and the 1970s. Structure, morphism, filter, injective, compact in its modern sense: much of the standard grammar of the subject was fixed in those rooms. The group’s place in the larger story is part of our history of mathematics in 50 milestones.
And the work goes on. A volume on algebraic topology appeared in 2016, and new chapters of the spectral theory book followed. The congresses still happen, at undisclosed places. The members are still unnamed. For more than ninety years an organization with no funding of its own, no president and no public roster has kept a single signature alive. Nicolas Bourbaki never lived. But the language mathematics speaks today was largely written in his sentences.
§ 13Ninety years at a glance
Bourbaki Timeline
1934
December 10. Six mathematicians meet in the Café Capoulade, Paris, to plan a new analysis textbook.
1935
First congress at Besse-en-Chandesse. The name Nicolas Bourbaki is adopted. A first note appears in the Comptes rendus.
1939
The first fascicle of the Éléments, on set theory, is published. The wedding card for Betti Bourbaki circulates.
1948
The Séminaire Bourbaki opens in Paris. It still runs three times a year.
1949
Weil states the Weil conjectures. Armand Borel attends his first, very loud, congress.
1950
“The Architecture of Mathematics” appears in English under Bourbaki's name, setting out the three mother structures.
c. 1960
Grothendieck, the group's most productive member, breaks with it over foundations.
1968
A mock funeral notice announces Bourbaki's death. The group keeps working.
2008
Henri Cartan, the last founding member, dies at 104.
2016
A new volume, Algebraic Topology, appears — 82 years after the first meeting.
Today
The group still meets and still publishes. Its members are still not named.
§ 14Quick answers
Questions About Bourbaki
Was Nicolas Bourbaki a real person?+
No. Nicolas Bourbaki is a collective pseudonym adopted in 1935 by a group of young French mathematicians, among them André Weil, Henri Cartan, Jean Dieudonné and Claude Chevalley. Everything the group writes is published under that one invented name.
Where does the name Bourbaki come from?+
From General Charles-Denis Bourbaki, a French officer of the Franco-Prussian War. At the École Normale Supérieure, students knew his name from an old prank in which a fake lecturer delivered a deliberately absurd mathematics talk; the founders took the name from that joke.
Does the Bourbaki group still exist?+
Yes. The group still meets in private congresses and still publishes volumes of the Éléments de mathématique; a book on algebraic topology appeared in 2016. Its current members are not made public, and members leave by the age of fifty.
What did Bourbaki contribute to mathematics?+
A unified, axiomatic presentation of modern mathematics built on set theory and three 'mother structures': algebraic, order and topological. The group also standardized vocabulary and notation still in daily use, including injective, surjective, bijective and the empty-set symbol ∅.
Why did Grothendieck leave Bourbaki?+
Around 1960 Alexander Grothendieck broke with the group over the direction of its foundations. He wanted mathematics rebuilt in the more flexible language of categories and functors, and the group did not adopt it. He went on to reshape algebraic geometry on his own terms.
Further reading
Six Books for Going Deeper Into Bourbaki
Two histories, a memoir, two volumes by Bourbaki himself and one novel. Links go to Amazon; Abakcus may earn a small commission on purchases.
1.The AMS application and the Boas episode: Mashaal, M. Bourbaki: A Secret Society of Mathematicians. AMS, 2006. Also Science News, “Founder of the secret society of mathematicians,” 2019.
2.War losses and French mathematics: Quanta Magazine, “Inside the Secret Math Society Known Simply as Nicolas Bourbaki,” November 2020.
4.The General Bourbaki prank at the ENS and the naming of the group: Mashaal (2006), chapters 2–3.
5.The singular “mathématique” and the reference to Euclid: CNRS News, “Bourbaki and the Foundations of Modern Mathematics,” 2017.
6.Grothendieck and the group: Jackson, A. “Comme Appelé du Néant — As If Summoned from the Void: The Life of Alexandre Grothendieck.” Notices of the AMS 51 (2004).
7.Borel, A. “Twenty-Five Years with Nicolas Bourbaki, 1949–1973.” Notices of the AMS 45 (1998).