Does one have to know the state of the whole universe?
According to current thinking, one can analyse experiments independently of the rest of the universe; in contrast, the state of the universe plays an important role in the thermal interpretation – everything depends on it, as only its dynamics are deterministic.
However the microscopic state of the measured system together with the measurement device should suffice to record the dominant effects. Their temporal development is dependent on their surroundings, etc. so that after enough layers of the onion one has the whole universe. The external layers naturally contribute very little, but they prevent a completely deterministic picture.
Presumably a sufficiently careful definition of the boundary of an experiment can restrict the role of the rest of the universe to specification of the temporal development of the micro-state of the boundary.
One must then grapple with a initial-boundary value problem instead of the initial value problem, as the micro-state of the boundary must be known for all times relevant to the experiment.
Arnold Neumaier (Arnold.Neumaier@univie.ac.at) A theoretical physics FAQ