We know that neutrino eigenstates are not mass eigenstate and this therefore produces neutrino oscillations. This is, however, deduced from the fact that the neutrino of one flavor produces the corresponding partner lepton of the same flavor (electron neutrino only produces electrons, etc.).

My question is : how do we know that this is indeed the case experimentally?

I know that the standard model predicts that it is the case, but how can we verify that, if the only way to tell that there was a neutrino of a given flavor is to detect the corresponding partner ?

EDIT : To give some precision about the kind of answer I'd like :

Saying : we know the flavor of a neutrino by the lepton it produces is not enough. In that case, we could imagine that a lepton does not always produces the corresponding neutrino (say, an "electronic neutrino" is by definition a neutrino that creates electrons when it interacts with a nucleon, but we could imagine that an electron could produce half of the time a "muonic neutrino" (always producing muons) and half of the time an "electronic neutrino"). I know that's not what predicts the standard model, but has it been verified experimentally ? If so, how can we do that, as we don't know what is produced before detecting it...

I ask that because we could imagine that it could be another way to "explain" at least part of the neutrino oscillations physics.


1 Answer 1


This is what we mean by the flavor state of a neutrino.

A neutrino which is involved in a weak interaction with a charged lepton has the same flavor as the charged lepton. By definition.

The observation of a charged partner is the operator for neutrino flavor measurement.

  • $\begingroup$ Instead of assuming neutrino oscillations, we could imagine that the creation of neutrino is not the same than their annihilation. How do we prove experimentally that this is always the case. Or I state it the other way around : if I know the flavor of a neutrino by the partner it produces, how do I know that the partner always produces the corresponding flavor. Is there an experimental fact to say that ? I'll edit my question. $\endgroup$
    – Adam
    Commented Nov 1, 2013 at 4:12
  • $\begingroup$ I know what you are trying to get at, but it doesn't lead anywhere: the observation of a charge partner is an operator and the eigenstates of that operator are called "flavor states". That is a definition. You are imagining that there is some other more basic meaning of flavor, but if there is you are going to have to call it something else at least until you can convince the establishment that your version is more useful. $\endgroup$ Commented Nov 1, 2013 at 4:16
  • $\begingroup$ Saying : "it's how it is in the model" is not an experimental fact. Maybe the answer is "we haven't/can't verify it", but I don't think your answer address my question. $\endgroup$
    – Adam
    Commented Nov 1, 2013 at 4:20
  • $\begingroup$ Observation is a operator; that is how quantum mechanics works. And operators have eigenstates (also how QM works). The eigenstates that go with seeing a charged partners of a given flavor are given the same flavor label. They have to be labeled somehow, and this is the choice we've made. That there are eigenstates is built into QM. Labeling them is a human convention and therefore arbitrary. There might be another interesting basis, but you haven't defined it yet, and you'll need to distinguish it from the one everyone currently uses. $\endgroup$ Commented Nov 1, 2013 at 4:23
  • 1
    $\begingroup$ At a minimum Daya Bay is recording data that show the oscillation probability appears to approach zero at zero distance now (especially when taken in combination with previous data sets such as KamLAND). Double Chooz should have the near detector running by next spring and will add to that data. $\endgroup$ Commented Nov 1, 2013 at 5:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.