Why do we expect unitarity to be preserved in the black hole information paradox?

Question

Consider the following way of describing the black hole information paradox:

Suppose we start with a pure quantum state and a black hole of mass $M$. Now we throw the pure state into the black hole and wait until the black hole has radiated away sufficient energy to again have mass $M$. The argument (given eg. here) is that we started with a pure state and ended with a mixed state and thus the black hole has performed a non-unitary transformation on the system and information has been lost.

$\textbf{What I don't understand:}$ everything we know about quantum mechanics implies that the quantum nature of a system is extremely fragile - temperature changes, external forces, etc. can all destroy the "quantumness" of a system. Why do we expect that throwing a quantum system in a black hole is going to preserve its quantum nature? Wouldn't the expectation be that the system ceases to be described by quantum mechanics and thus one does not need to preserve unitarity and there is really no paradox at all in losing information? I feel I am misunderstanding something here...

Answer 1

quantum nature of a system is extremely fragile - temperature changes, external forces, etc. can all destroy the "quantumness" of a system.

This process is Quantum Decoherence, by which a system coupling with its environment slowly loses its quantum nature. And yes it can be modeled as a non-unitary process. However, the system plus the environment will still evolve in a unitary fashion.

In the case of black hole evaporation, currently, this isn't the case. That is, even the combined system is non-unitary.

Answer 2

It is not a consensus that black hole evaporation should be unitary. Some researchers, such as Wald and Unruh for famous examples and myself for an unfamous one, do not see any issue with the lack of unitarity. Instead, we see problems with the possible ways of "fixing" the "paradox", which might require failures of QFT or general relativity at fairly low curvatures, for example. More discussion is given in the link I provided.

To make a connection with S.G.'s answer, a way of thinking is that there is no need to preserve unitarity because the singularity inside the black hole turns the problem into something similar to an open system, with no notion of a closed composite system being available. Furthermore, as discussed in the review by Wald and Unruh, even in Minkowski spacetime you can get non-unitary evolution depending on how you define time. Since there is no straightforward definition of time in a black hole evaporation spacetime, non-unitarity is simply caused by this lack of definition, with nothing truly mysterious happening.