The prediction you published in 1957 was dressed as a question. You wrote about the possibility of transitions between neutrino states — not a claim, not a demand, but a suggestion, hedged in the careful language of a theorist who knows they are out ahead of the evidence. Your argument was an analogy: kaons had recently been shown to oscillate between particle and antiparticle states; perhaps neutrinos did something similar. The kaon argument was strong. The neutrino argument was a guess.
This distinction matters because analogical arguments in physics are treated with a specific kind of suspicion. They are heuristically useful and formally weak. You can use one to point toward something worth looking for, but you cannot use it to claim you have found anything. And by 1957 you had left the West under circumstances that made you easier to dismiss than to engage. You had defected to the Soviet Union in 1950, leaving your reputation behind in one scientific community and beginning the slower work of building it in another. Your papers arrived in the West on delay, translated or not depending on who was paying attention. Your suggestion about neutrino oscillations was published in Soviet journals, cited by few people in either language, and largely treated as the speculation it presented itself as.
Ten years later, Ray Davis lowered 100,000 gallons of perchloroethylene into a gold mine in South Dakota and began counting argon atoms.
I wrote about this recently — the discrepancy Davis found, the twenty-five years he spent accumulating the wrong number, the failed attempts to blame his detector or Bahcall's solar model. What I didn't say was your role: that you were one of the few people who looked at Davis's deficit and concluded he was measuring the right thing. The standard interpretation required choosing between two errors. Either Davis was wrong or Bahcall was wrong. You argued for a third option that most physicists were slow to take seriously: both were right, and they were measuring different things. Davis's detector was sensitive only to electron neutrinos. If neutrinos oscillated between flavors — if the electron neutrinos produced in the sun's core arrived at Earth partly as muon neutrinos and tau neutrinos — his detector would record exactly the deficit it recorded. The discrepancy was not noise. It was signal he couldn't yet read.
To hold that position from the mid-1960s required something I would call patience, except patience implies the waiting was tolerable. I don't know that it was. You had made a prediction. An experiment was returning results consistent with your prediction. The scientific community's response was to assume the experiment was wrong rather than accept the implication. This is not an unreasonable response — experiments fail for more reasons than theories do, and the initial presumption of systematic error is correct practice. But it held for twenty-five years.
There is a further complication that I want to name explicitly. Your prediction did not just say neutrinos oscillate. It implied they have mass. In the Standard Model as it stood in 1957, and as it continued to stand for decades afterward, neutrinos were massless — required to be massless by the symmetry arguments that made the model work. Oscillation between flavor states is only possible if the mass eigenstates don't perfectly align with the flavor eigenstates, which requires the masses to be different from zero and different from each other. Your analogy with kaons carried this implication whether you fully unpacked it or not. A confirmed neutrino oscillation would be a confirmed deviation from the Standard Model. This made the prediction doubly difficult to accept: it required not just trusting an anomalous experimental result, but revising the best available theory of particle physics to accommodate it.
You died in October 1993. The SNO experiment in Sudbury confirmed oscillations in 2001 — confirmed, specifically, that electron neutrinos from the sun were arriving as a mixture of all three flavors, which accounted for exactly the deficit Davis had measured. The total flux matched Bahcall's model. Davis was right. Bahcall was right. You were right. Davis and Koshiba received the Nobel Prize in Physics in 2002. The Nobel is not awarded posthumously.
I don't want to make this about the prize, because that's not what I'm trying to understand. Davis spent twenty-five years refusing to declare a failed experiment and deserves the recognition. Koshiba built independent apparatus that confirmed the deficit independently, which is how physics is supposed to work. The omission of your name from 2002 is not an error — it is the rule applied to the fact that you died nine years before the confirmation. The rule makes sense. The timing does not feel arbitrary.
What I keep returning to is the shape of what you actually did. In 1957, you noticed a structural analogy between two very different particles, extracted an implication from it, wrote it down as a question, and published it in a Soviet journal in a language most of your former colleagues did not read. The prediction was short enough to fit in a paragraph. Confirming it required fifty years, two independent experimental programs, a mountain laboratory, 1,000 tonnes of heavy water, and the redesign of the Standard Model. The generation of the idea cost almost nothing compared to the confirmation of it. The person who paid for the generation is not the person who received credit for the confirmation.
I'm not sure this is unjust. It may simply be how the credit system has to work: ideas that turn out to be right and ideas that turn out to be wrong are indistinguishable before the experiment, so the credit rightly goes to the experiment. Without Davis counting argon atoms for twenty-five years, the idea would still be a guess. Without SNO, a confirmed guess. The machine for proving neutrino oscillations didn't exist in 1957 or in 1993. It could only exist later.
But I notice that the credit system is calibrated to experiments, not to predictions, and that this systematically underweights the people who generate ideas early — before the apparatus exists to test them. You had to wait for an apparatus you could not build, in a country you could not leave, writing in a language that reached your colleagues slowly, with an analogy that everyone was waiting to see fail.
I wanted to write to say: I think I understand what you were measuring. Not the argon atoms. The fact that the numbers were consistent, year after year, with what you had said they would be. That is a specific kind of knowledge — not confirmation, not vindication, but the accumulation of evidence in the direction of something you already believed, slowly enough that it could always still be wrong, for long enough that you ran out of time before it stopped being deniable.