Possible Worlds and the Objective World

It could be that, at the moment you finish reading this essay, the entire universe is wiped out in a great cosmic reset. After the reset, events exactly like those of the past 14 billion years are replayed. You and I and everyone we know have doppelgänger echoes in the very distant future. Call this scenario First. Here’s a different way things could be: this could have already happened. That is, it could be that our Big Bang was a cosmic reset, and before it there were 14 billion years of history just like those that have happened since the reset up to when you finish reading. In this case, you and I and everyone we know are doppelgänger echoes of people in the very distant past. Call this scenario Second. In each of these scenarios, two copies of a long stretch of history take place, one after the other. In First, you and I live in the first of two copies; in Second, we live in the second. But this is a bit odd. First and Second tell precisely the same qualitative story: they only differ concerning the identities of particular individuals—which of the two people exactly like you is really you, and such questions. You might think it is implausible that there are really two different possible worlds that differ merely concerning which things are which. The pull of a thought like this creates a tension: First and Second both seem possible, but it doesn’t seem like there should be such subtle differences between possible worlds. David Lewis suggests a way of splitting the difference between these two seemings. His crucial claim—one that will occupy a good deal of attention in this essay—is that “Possibilities are not always possible worlds” (1986, 230). In the case at hand, there really