What Happened
Physics · Particle Physics · Standard Model · History of Science · Explainerpapers and prize citations read in full · the anecdotes left where they were found
A category proposed for two particles, outgrown twice, and still the name of the family
Six Leptons, a Word Made for Two
Six leptons are usually told as six discoveries with dates beside them. Read against the papers that announced them — and the positron’s, which belongs to the same story — four of those announcements went out under a title that hedged: apparent existence, note on the nature of, a confirmation, anomalous production. And the word that holds them all together was coined, in 1948, for a class of two that did not include the muon.
What a lepton is The modern definition says nothing about weight. The word says nothing else.
A lepton is a spin-½ fermion that takes no part in the strong interaction. The Standard Model arranges the six of them into three pairs, each pair a charged lepton with its own neutrino: the Particle Data Group's review of neutrino masses states it as one active neutrino for each charged lepton, and adds that neutrinos have neither strong nor electromagnetic interactions. Nothing in that definition is about weight. Rosenfeld built the word in 1948 from the Greek leptos, thin or slight, and meant it for particles lighter than a proton — a criterion the tau broke in 1975 by weighing nearly twice as much as one. The name survived the criterion because by then it named a place in the theory rather than a mass range.
- ½The spin every lepton carries, and the reason all six are fermions. The Particle Data Group's 2025 summary table prints J = ½ against the electron, the muon and the tau alike.
- 3Generations, each a charged lepton paired with its neutrino. The count is measured rather than assumed: the invisible width of the Z boson closes the family at three light neutrino species, and did so eleven years before the sixth member was seen.
- 1.89×The tau's mass as a multiple of the proton's — 1776.93 MeV against 938.27 MeV. The heaviest of the “light particles” outweighs the object the word was coined to exclude.
The four numbers Each one read out of a record named in the sources line.
- 2particles the word “lepton” was coined to cover, in Rosenfeld’s Nuclear Forces, 1948
- 6leptons in the family now, in three generations
- 11years between the published count of generations (October 1989) and the first observed tau-neutrino interaction (December 2000)
- 4tau-neutrino candidate events in DONUT’s 203 interactions, against an estimated background of 0.34
What is established, and how well Sorted by standing. The four bands are not interleaved, because a title checked in a database and a footnote read in the book are not the same kind of fact.
| Standing | What |
|---|---|
| Read in full | Rosenfeld’s Nuclear Forces (1948) adopts “lepton” for a particle of small mass “irrespective of its charge”, having “two kinds of states, in which it appears as an electron and a neutrino”. |
| Read in full | The same book discusses the muon throughout as a meson. Across its full text, “meson” appears 248 times and “mesotron” not once. |
| Read in full | Stoney had the name in print by October 1894: an estimate of the unit “for which I have since ventured to suggest the name electron”. He dates the underlying idea to August 1874 and February 1881. |
| Read in full | Thomson’s 1897 paper calls the objects corpuscles, not electrons, and reports that “the value of m/e is independent of the nature of the gas”. |
| Read in full | DONUT analysed 203 neutrino interactions and found four tau-neutrino candidates against an estimated background of 0.34 events. |
| Read in full | The LEP and SLC combination gives the number of light neutrino species as 2.9840 ± 0.0082, from 17 million Z decays plus 600 thousand more. |
| Citation checked | The announcing papers and their dates: Yukawa (February 1935), Anderson (September 1932), Neddermeyer and Anderson (1937), Conversi–Pancini–Piccioni (1947), Lattes–Muirhead–Occhialini–Powell (1947), Reines and Cowan (November 1953), Danby and colleagues (July 1962), Perl and colleagues (August 1975), ALEPH (October 1989). |
| Citation checked | The pion paper has four authors. Muirhead is the one usually dropped from the retelling, and the 1950 Nobel Prize named only Powell. |
| Not established here | That Stoney first printed “electron” in 1891, and that he called the unit “electrolion” in 1881. The 1891 citation came back from a search; the paper was not opened. |
| Not established here | That the tau took its name from Greek triton; that DONUT’s four events amounted to about 3.5 standard deviations; that Kaufmann had better data than Thomson and withheld the claim; that Zhao Zhongyao saw positron tracks in 1929 and 1930. |
| Anecdote, unsourced | Rabi’s “Who ordered that?” and Pauli’s letter of 4 December 1930. Both are quoted everywhere. Three attempts to open a scan of the letter failed, and no contemporaneous record of the remark was found. |
Timeline
How it actually went Every dated entry in the record. Prize citations sit where the prize was given, not where the work was done.
- Aug 1874Stoney presents a fundamental unit of electricity at the British Association in Belfast — known here only from his own account of it, twenty years later.
- Feb 1881He reads the same paper before the Royal Dublin Society.
- Oct 1894“Of the ‘Electron,’ or Atom of Electricity”: the name is proposed in print.
- 1897Thomson publishes “Cathode Rays” and calls the objects corpuscles.
- Sep 1932Anderson: “The Apparent Existence of Easily Deflectable Positives”.
- Feb 1935Yukawa predicts a carrier of the nuclear force with a mass between the electron’s and the proton’s.
- 1936Nobel Prize to Carl David Anderson “for his discovery of the positron”.
- 1937Neddermeyer and Anderson: “Note on the Nature of Cosmic Ray Particles”.
- 1947Conversi, Pancini and Piccioni show that the cosmic-ray particle does not interact strongly — so it is not Yukawa’s.
- 1947Lattes, Muirhead, Occhialini and Powell find the pion — the particle Yukawa predicted — in the same year.
- 1948Rosenfeld’s Nuclear Forces coins “lepton” for electrons and neutrinos. The muon is a meson in the same book.
- 1950Nobel Prize to Cecil Frank Powell alone, for the photographic method and “his discoveries regarding mesons”.
- Jul 1953Konopinski and Mahmoud propose a conserved lepton number.
- Nov 1953Reines and Cowan: “Detection of the Free Neutrino”.
- Jul 1956The result again, retitled “Detection of the Free Neutrino: a Confirmation”, in Science.
- Jul 1962Danby and colleagues establish that there are two kinds of neutrino, not one.
- Aug 1975Perl and colleagues: “Evidence for Anomalous Lepton Production”. No particle is named in the title.
- 1988Nobel Prize to Lederman, Schwartz and Steinberger for “the doublet structure of the leptons”.
- Oct 1989ALEPH publishes the number of light neutrino species; OPAL publishes the Z mass and width the same month.
- 1995Nobel Prize to Perl for the tau and Reines for the neutrino, in one citation forty years wide.
- Dec 2000DONUT submits the first observation of tau-neutrino interactions.
The two orders disagree By date, and by how well evidenced — and the gap between them is the point.
The chronology and the evidence run in opposite directions. The oldest entry above is the weakest: Stoney’s 1874 presentation survives here only in his own recollection of it, twenty years later. The best-evidenced entry sits in the middle of the list — the 1948 coinage, read from the book’s own preface and footnote — and it is the one the standard account leaves out. Meanwhile the two most confidently dated moments in every retelling, Pauli’s letter and Rabi’s question, are the two this page could not place at all. The anecdotes carry sharper dates than the papers do, which is the wrong way round.
- 1948read in full, and the entry nobody tells
- 1874the oldest date here, on the thinnest evidence
- 2famous documents this page could not open
The Argument
A word that outlived its own definition The 1948 footnote has three clauses. All three have since been overtaken.
Rosenfeld adopts “lepton” “as a pendant to ‘nucleon’” for a particle of small mass “irrespective of its charge”; he says such a particle has “two kinds of states, in which it appears as an electron and a neutrino”; and he keeps “electron” for “a particle of small mass with an elementary charge of either sign”. Today the class has six members rather than two, an electron and a neutrino are different particles rather than two states of one, and an electron has one sign of charge. The name survived and every definition under it was replaced. The preface adds the part that usually gets dropped: the word was not Rosenfeld’s. It was, he writes, “coined by Prof. Møller in consultation with a hellenist”.
- 2 → 6members of the class, 1948 to now
- 248times “meson” appears in the book that coined “lepton”
The hedge is in the title What each announcing paper actually claimed, in its own words. Four of these six hold something back in the title itself; 1962 and 2000 say plainly what they found.
| Lepton | The announcing title | What it claimed |
|---|---|---|
| Positron | “The Apparent Existence of Easily Deflectable Positives” (1932) | An appearance, hedged twice inside six words. |
| Muon | “Note on the Nature of Cosmic Ray Particles” (1937) | A note, on a nature. No particle is named. |
| Electron neutrino | “Detection of the Free Neutrino” (1953), then “…: a Confirmation” (1956) | A detection, and then, three years later, its own confirmation. |
| Muon neutrino | “… and the Existence of Two Kinds of Neutrinos” (1962) | A distinction between two things, not the finding of one. |
| Tau | “Evidence for Anomalous Lepton Production” (1975) | An anomaly in a rate. The particle is not in the title at all. |
| Tau neutrino | “Observation of Tau Neutrino Interactions” (2000) | The only flat claim in the set. It rests on four events. |
What Others Add
The six members, and what is on record for each Masses and mean lives from the Particle Data Group's 2025 lepton summary table. The year is the announcing paper's, whose wording section three examines.
| Lepton | Gen. | Charge | Mass | Mean life | Announced |
|---|---|---|---|---|---|
| Electron e− | 1 | −1 | 0.510 998 950 00(15) MeV | > 6.6 × 10²⁸ yr | 1897 |
| Electron neutrino νe | 1 | 0 | not listed by flavour | no decay seen | 1956 |
| Muon μ− | 2 | −1 | 105.658 3755(23) MeV | 2.196 9811(22) × 10⁻⁶ s | 1937 |
| Muon neutrino νμ | 2 | 0 | not listed by flavour | no decay seen | 1962 |
| Tau τ− | 3 | −1 | 1776.93(9) MeV | 290.3(5) × 10⁻¹⁵ s | 1975 |
| Tau neutrino ντ | 3 | 0 | not listed by flavour | no decay seen | 2000 |
Why three of the six rows have no mass in them Not a gap in the measurement. A neutrino of a given flavour does not have one.
The Particle Data Group's lepton summary table gives a spin and a mass for the electron, the muon and the tau, and then stops. Under Neutrino Properties it lists a single mass limit for all three flavours together. That is not an omission. A neutrino produced as an electron neutrino is a superposition of the three states that do have definite masses, so the question “what does the muon neutrino weigh” has no single answer to give. What is known instead is the difference between the squared masses, and those are measured well: they fix the heaviest of the three at no less than about 50 meV, while the tightest direct limit on the mass itself, from tritium decay at KATRIN in 2025, stands at 0.45 eV — still some eight times above the floor.
- 0Neutrino masses listed by flavour in the Particle Data Group's lepton summary table. One combined limit covers all three.
- 50 meVThe floor under the heaviest neutrino, from the measured mass-squared splitting of 2.451 × 10⁻³ eV². Not a measurement of a mass, but a mass no smaller than this.
- 0.45 eVThe tightest direct upper limit, at 90% confidence, from 259 measurement days of tritium decay at KATRIN, published in 2025.
Three things the papers settle that the story does not Each drawn from a record named in the sources line, not from the retelling.
1935 → 1947 · reclassification
The muon was found as somebody else’s particle
- Yukawa’s paper is dated February 1935 and the Neddermeyer–Anderson note 1937. Something of roughly the right mass turned up two years later, in the right place, and was the wrong particle.
- Both corrections landed in 1947: Conversi, Pancini and Piccioni showed the cosmic-ray particle does not interact strongly, and Lattes, Muirhead, Occhialini and Powell found the one that does.
- A monograph printed in 1948 still files it under mesons. Reclassification runs slower than discovery.
October 1989 · counting
The family was closed before it was complete
- ALEPH published the number of light neutrino species in October 1989, and OPAL the Z mass and width the same month. DONUT submitted the first observed tau-neutrino interactions in December 2000.
- The final combination, from 17 million Z decays, gives 2.9840 ± 0.0082 — a value that excludes two and excludes four at the same time.
- It constrains only neutrinos light enough and coupled strongly enough to appear in Z decay. It is not a census of everything that could exist.
1936–1995 · attribution
A prize citation is evidence of credit, not of authorship
- The 1950 Prize names Powell for “his discoveries regarding mesons”. The paper behind it has four authors.
- The 1988 citation frames the muon neutrino as “the demonstration of the doublet structure of the leptons” — a structure, not a particle.
- The 1995 Prize covers the 1953–56 neutrino detection and the 1975 tau in a single sentence, forty years apart.
Conclusion
What to take from it And what this page is deliberately not carrying.
What holds
Six leptons, three generations, and a published count of generations that runs eleven years ahead of the sixth member’s observation. Every date and title above was checked against a record this page names, and the two that were read in full — the 1948 book and the 1894 paper — are the two that carry the argument.
What is deliberately missing
Pauli’s letter, Rabi’s question, Kaufmann’s reticence, Zhao Zhongyao’s near miss. These are the best stories in the subject and not one of them is sourced here. A page that repeated them on the strength of everyone else repeating them would be a better read and a worse record, and the choice between those two is not close.
What to hold loosely
That the announcing titles hedged is a fact about the titles. Reading it as timidity, or as rigour, is an inference and is marked as one — the contents of most of those papers were not opened here, only their citations. And a category rebuilt twice may well be rebuilt again: the Z-width count says three light neutrinos couple to the Z, which is not the same statement as three being all there is.