If no joint distribution exists, then the value you assign to one observable cannot be fixed in advance independently of what else you measure. This is contextuality. In a classical hidden-variable model, each observable would carry a pre-assigned value, and measuring it would simply reveal that value. Quantum mechanics forbids this: the outcome statistics for one observable change depending on which other, non-commuting observables are measured alongside it. Contextuality has been confirmed in experiments and is the experimentally testable form of the failure of classical description.
Non-Commuting Observables: Why Quantum Mechanics Is More Fundamental
Measurement, Disturbance, and the Breakdown of Classical Description
Values Depend on Measurement Context
4 / 5
If no joint distribution exists, something has to give about the values themselves. In a classical hidden-variable picture, each observable carries a value that is already there, and measurement just reveals it. That picture requires the value to be fixed independently of what else you happen to measure.
0:00 / 0:00