On the contingency of perdurantism
Several people have claimed that perdurantism is only contingently true, or at least a posteriori: Mark Johnston expresses something like this at the end of "Is There a Problem about Persistence?"; Sally Haslanger in "Humean Supervenience and Enduring Things"; Frank Jackson in section 2 of "Metaphysics by Possible Cases"; and David Lewis in section 1 of "Humean Supervenience Debugged".
One of the arguments for this claim seems to be that both perdurantism and endurantism are to some degree intelligible, which is why philosophers still disagree about the issue. I find that strange. Philosophers also disagree about the existence of universals, arbitrary mereological fusions, possible worlds, and numbers. Are these also contingent matters?
Actually, I'm not sure what the claim amounts to. Here are two possibilities:
1) Neither perdurantism nor endurantism provide an analysis of intrinsic change. Rather, they are conflicting empirical hypotheses about how things change. In some possible worlds, things change their shapes by having temporal parts, in other worlds they do so by instantiating shapes-at-times, or by instantiating-at-times shapes.
2) The perdurantist analysis of intrinsic change is necessarily true. Whenever some thing in some world changes one of its intrinsic properties, then it must consist of temporal parts. However, some possible things don't ever change their intrinsic properties, or not all the time. It is a contingent matter whether these things (or these non-changing temporal segments of them) have temporal parts.
I think that (2) is a reasonable view, provided that it make sense to say that an extended thing might nevertheless lack parts. (1) seems to be very odd: How could shapes be intrinsic monadic properties in one world and relations to times at other worlds? What makes both of these rather different kinds of things shapes? On the other hand, saying that (2) makes perdurantism contingent would be an exaggeration: True, if (2) is true then some possible things endure identically through time. But endurantism is usually understood as more than just this, namely as an analysis of intrinsic change in enduring things. According to (2), there is no such thing. Moreover, the context of the papers mentioned above indicates that these philosophers really endorse (1).
If so, then at least for Lewis this would be a major change of view. For example, in "On the plurality of worlds" and, a bit more explicitly, in "Rearrangement of Particles", he takes perdurantism to be an analysis of intrinsic change. If he really gave up that view in "Humean Supervenience Debugged", I wonder why he doesn't even say so, or argue for the change, and its consequences.
One of the consequences concerns mental content: Lewis argued in several places that the content of propositional attitudes can be identified with (or modeled by) sets of temporal stages of possibilia. How does that work if some possibilia don't have temporal stages?