A puzzle about self-location and justification

The Puzzle

Imagine the following situation:

Physicists have settled on a grand theory (theory A) that unifies experimental results from a variety of different contexts. This theory makes certain predictions about the future of the universe: it will exist for an indefinitely long time, and in that long span of time, there will be many brains ("Boltzmann brains") that pop momentarily into existence as a result of quantum fluctuations. The vast majority of these brains will function improperly or not at all, -- but a few brains will be molecule-for-molecule identical to ours. The vast majority of these will then die in a matter of moments in hostile environment of empty space. Nevertheless, most of the brains that are like ours will be such Boltzmann brains. Theory B, on the other hand,  has been decisively refuted by experiments, but it also predicts that many Boltzmann brains would exist.

Here is the question: Are we justified in believing theory A over theory B?

Here is one reason to think not: Insofar as we believe in theory A, we should think that we're probably Boltzmann brains. After all, most brains like ours are Boltzmann brains. Insofar as we think we're probably Boltzmann brains, we should think that we've got no reason to favor theory A over theory B.  Therefore, we're not justified in both believing theory A and believing that we're justified in believing theory A. If we're not justified in believing that we're justified in believing theory A, we're not justified in believing theory A.

I am conflicted about this argument. On the one hand, it seems possible that a theory like theory A might be true. And if it is, then we could have evidence that points in its direction. Supposing we do, what should we believe?

On the other hand, it is hard to justify believing that we're one of the lucky brains, if we know that the majority of actual brains are Boltzmann brains.

Some Options:

1. Allow that we're justified in believing we're not Boltzmann brains. This requires denying the kind of restricted principle of indifference that Adam Elga defends in "How to defeat Dr. Evil with self-locating evidence. But it maybe something we can swallow.

2. Allow that we're justified in believing in theory A, even though we should think we're Boltzmann brains who have no real evidence for theory A. This *might* be justifiable on some grounds along the lines of the *new rational reflection principle* mentioned a few posts ago.

3. Admit that we could never be justified in believing in theory A. The most we could be justified in believing is that our evidence points to theory A. We should, perhaps, remain agnostic about which theory is true. Or else perhaps we should favor some other theory that also fits our evidence, even if it is otherwise more contrived or implausible.


Among these solutions, I am not sure which to favor. The second feels wrong to me, even though I am happy to admit that we might be justified in believing in things that we're not justified in believing that we're justified in believing in.