hellsteiger said:
Please explain to me further if there is a difference between a superposition and mixed state (without and observer, or interaction - that interaction typically being how an observation is carried out).
The big difference between a superposition and a mixed state is the presence of "interference terms".
Suppose you have a system that has 3 states of interest: [itex]A, B, C[/itex]. You start the system in either state [itex]A[/itex] or [itex]B[/itex], and later check to see whether it is now in state [itex]C[/itex]. (Assume that states [itex]A[/itex] and [itex]B[/itex] are orthogonal)
Define:
- [itex]P_A[/itex] is the probability of the system starting in state [itex]A[/itex].
- [itex]P_B[/itex] is the probability of the system starting in state [itex]B[/itex].
- [itex]P_{AC}[/itex] is the probability of the system winding up in state [itex]C[/itex], given that it started in state [itex]A[/itex].
- [itex]P_{BC}[/itex] is the probability of the system winding up in state [itex]C[/itex], given that it started in state [itex]B[/itex].
Then if the system starts in a mixture of states [itex]A[/itex] and [itex]B[/itex], then the probability that the system will end up in state [itex]C[/itex] is simply:
[itex]P_C = P_A P_{AC} + P_B P_{BC}[/itex]
On the other hand, if the system starts in a pure state that is a superposition of states [itex]A[/itex] and [itex]B[/itex], then the probability will be different:
[itex]P_C = P_A P_{AC} + P_B P_{BC} + 2 \sqrt{P_A P_{AC} P_B P_{BC}} cos(\phi)[/itex]
where [itex]\phi[/itex] is a phase depending on the exact superposition you started with, and the nature of the interaction leading to state [itex]C[/itex]
The term involving the phase [itex]\phi[/itex] is an "interference" term that only appears when you have a superposition, rather than a mixture.
The significance of the difference is that a mixture can always be explained using the "ignorance" interpretation: The system is really in state [itex]A[/itex], or state [itex]B[/itex], we just don't know which. The interference term cannot be explained (at least not straight-forwardly) as ignorance about the true state of the system.
Decoherence makes the interference terms negligible, so after decoherence, we can pretend that we have a mixture, and pretend that the system is in one state or the other, and we just don't know which. Prior to decoherence, we can't consistently pretend that's true.