74 From Classical Mechanics to Quantum Field Theory. A Tutorial
Remark 2.1.4. Obviously this formula does not make sense ifμ(ψA)(a)=0as
expected. Moreover the arbitrary phase affectingψdoes not lead to troubles, due
to the linearity ofPa(A).
(3) Compatible and Incompatible Observables. Two observables are com-
patible – i.e. they can be simultaneously measured – if and only if the associated
operatorscommute,thatis
AB−BA=0.
Using the fact thatHhas finite dimension, one easily proves that the observables
AandBare compatible if and only if the associated spectral projectors commute
as well
Pa(A)Pb(B)=Pb(B)Pa(A) a∈σ(A),b∈σ(B).
In this case,
||Pa(A)Pb(B)ψ||^2 =||Pb(B)Pa(A)ψ||^2
has the natural interpretation of the probability to obtain the outcomesaandb
for a simultaneous measurement ofAandB.IfinsteadAandBare incompatible,
it may happen that
||Pa(A)Pb(B)ψ||^2 =||Pb(B)Pa(A)ψ||^2.
Sticking to the case ofAandBincompatible, exploiting (2.9),
||Pa(A)Pb(B)ψ||^2 =
∣∣
∣∣
∣
∣∣
∣∣
∣P
(A)
a
Pb(B)ψ
||Pb(B)ψ||
∣∣
∣∣
∣
∣∣
∣∣
∣
2
||Pb(B)ψ||^2 (2.10)
has the natural meaning ofthe probability of obtaining firstband nextain a
subsequent measurement ofBandA.
Remark 2.1.5.
(a) Notice that, in general, we cannot interchange the rˆole ofA andBin
(2.10) because, in general, Pa(A)Pb(B)=Pb(B)Pa(A)ifAandBare incompatible.
The measurement procedures “disturb each other” as already said.
(b)The interpretation of (2.10) as a probability of subsequent measurements
can be given also ifAandBare compatible. In this case, the probability of ob-
taining firstband nextain a subsequent measurement ofBandAis identical
to the probability of measuringaandbsimultaneously and, in turn, it coincides
with the probability of obtaining firstaand nextbin a subsequent measurement of
AandB.