Do neutrino oscillations generate back-reaction in matter? Typically all interactions generate some back-reaction. Neutrino oscillations are a subtle long-range interaction that induce weak-force interacting neutrinos to flavor-oscillate, and the properties of these oscillations are influenced by the presence of matter according to the Mikheyev–Smirnov–Wolfenstein effect.
But is not clear to me if there is backreaction in normal matter from all these oscillations that are incurred, even under the assumption of absolute zero neutrino absorptions. If there is no backreaction, it would be (I think) the only known case of an interaction that does not have some form of reaction or counterpart in one of the physical subsystems
If there is some some reaction expected from these oscillations (even if tiny or unobservable) what shape and form would such reaction take?
 A: I'll try to give an intuitive explanation. Imagine a neutrino which is produced at point A and detected at point B. 
Two things:


*

*Neutrino oscillation in vacuum is no interaction but a result of the fact that the interacting flavor eigenstates are linear combinations of the propagating mass eigenstates. We detect different neutrino flavors in charged current weak interactions typically with protons or neutrons where they produce either a electron, muon or tau (or the corresponding anti particles). Neutrino production is typically the inverse process. That is neutrinos are detected and produced as flavor eigenstates. However, these are no eigenstates of the Hamiltonian. Therefore if we want to propagate a neutrino from A to B we have to translate the flavor eigenstates into the Hamiltonian or mass eigenstate basis. The mass eigenstates are the "real" neutrinos in the sence that we understand other particles as electrons and quarks i.e. they have a real physical mass.
Now think of your flavor neutrino as a wave which is the sum of 3 wave packets which represent the mass eigenstates. At point A where the flavor state is produced all 3 mass wave packets are located at A. However when propagating to B the 3 mass eigenstates propagate at different speeds such that when arriving at B the total wave which represents the flavor neutrino has changed. This induces the total wave to now no longer be a flavor eigenstate but a mixture of different flavors such that there is a certain ptopability that we detect any flavor at B although we produced a fixed flavor at point A. This is in a intuitive way what causes oscillation. You see oscillation itself is no interaction but of course to detect it you need some interaction. And of course in the detecting interaction there is a backreaction like momentum transfer. 

*Neutrino oscillations in matter are influenced by forward scattering of neutrinos typically on electrons inside the matter. This is the MSW effect and this forward scattering is without momentum transfer so there is again no backreaction. However, the presence of interactions changes the Hamiltonian of the system. Changing the Hamiltonian changes its eigenstates which in response changes the way oscillations appear.
Note that to have oscillations it is crucial that we do not know which physical mass eigenstate has propagated from A to B. That is we need some uncertainty on the kinematics of the system.  (one would still see changes in flavor though, just no oscillating pattern). A case where this might happen would be the detection of (extra galactic) supernova neutrinos and it could be used to find the exact neutrino mass which is yet unknown.
I notice that although this was supposed to be intuitive its rather long and maybe graphs or formulas would've helped... 
A: Neutrinos oscillate whether they are travelling through matter or free space. Each flavor has a different mass, with an associated energy due to $e=mc^2$. But the neutrino's energy must be conserved. So the difference when it changes flavor is counterbalanced by a corresponding change in its kinetic energy $e=1/2 mv^2$. Since $e$ and $m$ are determined, the change is theoretically accompanied by a small (but indetectable) change in velocity $v$.
