Monday, July 17, 2006

World Essentialism

Recall the distinction between the singular world vs. the various possible ways a world could be. (The latter are misleadingly called "possible worlds", but are better understood as world-states, or maximal properties.) We may think of the universe as a "lump", and the actualized possible world as the maximal "property" instantiated by the lump. But note that in general, it isn't the case that every object could instantiate any property. The entity that I am, for instance, is essentially human, and so couldn't instantiate the property of being a poached egg. (If an evil wizard transformed me into a poached egg, then the resulting object wouldn't be me any more.) This then raises the question of whether the world-lump has any essential properties, or whether it could actualize any possible world-state whatsoever whilst retaining its identity. Put in more intuitive terms, we can ask: could our world have turned out in any possible way? Or are some possibilities so extreme that they could only be realized by a different underlying universe?

I'm inclined towards anti-essentialism here. It seems tidiest to say that each possible world-state is a way our world could have been. But to explore the alternative, I think the most likely candidates for essential world-properties would be the laws of nature. We would then say that our world could have turned out in any nomologically possible way. It remains true that there could have been different laws of nature, but not in our universe. A change so radical would disrupt its very identity, and turn it into a different universe. (Just like turning me into a poached egg. The radically different properties can be realized, but only by a different entity.)

One might also want to build in the "initial conditions" of the universe as being part of its essential nature. Then, if the laws are deterministic, we would reach the conclusion that our world couldn't have turned out differently at all. (Again, things could have turned out differently. But that would be because a totally different universe existed in place of ours.) If indeterminism is true, at least a few more possibilities are opened up for us. But they will still be extremely limited compared to what philosophers usually take to be possible ways our world could have been.

This narrow account of local possibility might line up with pessimistic accounts of primitive modality. It doesn't settle those broader questions though. We are still left wondering whether, say, another universe with different laws really could have existed in place of ours. (This lack of progress is unsurprising, though, since I'm inclined towards a deflationary understanding of de re modality and debates about essentialism as merely terminological. There are no deep facts about my not possibly being a poached egg. All this really means is that of any entity that was a poached egg, we would refuse to hold it in the extension (or intension) of 'Richard Chappell'. Haeccaeities or identity facts are not fundamental aspects of modal reality. But that's a post for another day.)

But for those who take essentialism more seriously, do you think anything is essential to our world-lump?


  1. Hey Richard,

    It's not clear to me what sense can be made of a possible world's having certain properties essentially. Whether or not property P is essential to object x itself has to do with whether or not x has P in all possible worlds. But possible worlds aren't the sorts of things that can have properties in other possible worlds. So trivially, there is no property that is essential to some possible world. (That is, if you think that possible worlds aren't the sorts of things that can have properties in other possible worlds, including themselves. But if you deny the last clause, then it turns out that every property of a given possible world is essential to that world.)

    So maybe none of this applies to your "world-lumps". Maybe these are the sorts of things that can be destroyed and or occur in other possible worlds. Or maybe not; I'm not really sure what these "world-lumps" are supposed to be. (Are they spacetime manifolds?) In any case, whatever they are, world-lumps don't strike me as being possible worlds. Nor do I see why we need them, if we're inclined to think that possible worlds are maximal properties. (Or at least, Plantinga, Adams, and other philosophers who have defended this theory did not appear to think that "world-lumps" are required to bear these maximal properties.) Could you explain what you mean by a "world-lump"?

    An alternative to the above is to talk of natural kinds of possible worlds. John Bigelow, Brian Ellis, and Caroline Lierse explore this idea in "The World as One of a Kind: Natural Necessity and Laws of Nature", 1992, The British Journal for the Philosophy of Science 43(3): 371-388. This is a good essay if you're interested in the sorts of issues raised in your post. I'm skeptical that their particular strategy will work, since for them objects fall under natural kinds depending upon what their essential properties are. They suggest we define "essential property" as follows:

    [-]: Property P is essential to individual x iff necessarily, if any individual y lacks P, then y is not identical to x.

    But we have the same problem as before. How are we supposed to understand "necessarily" here if the indivduals in question are themselves possible worlds?

  2. Richard: Put in more intuitive terms, we can ask: could our world have turned out in any possible way?

    Sure. In fact, I'd wager that it's necessarily true that our world could have turned out in any possible way, provided that you're not equivocating on the sort of modality you're talking about in between the "could" and the "possible." After all, any way that our universe could not have turned out to be is (ipso facto) an impossible, not a possible, "way."

    To put it another way, isn't your question here just equivalent to asking whether the world could have turned out in any way it could have turned out, or whether it's possible for the world to have turned out in any possible way?

    Richard: Or are some possibilities so extreme that they could only be realized by a different underlying universe?

    I don't think I know what this even means. What would it mean to have a numerically (not just qualitatively) different universe from the one we've got? I can conceive of things going quite differently in this universe, but I really have no idea what I'm supposed to imagine about this numerically distinct, alien world that would make it something other than the world we're in, under different circumstances. (If they are numerically distinct, could they both be actual at once? If not, why not? But if so, doesn't that conflict with the condition of totality?)

    Alex Skiles: Whether or not property P is essential to object x itself has to do with whether or not x has P in all possible worlds. But possible worlds aren't the sorts of things that can have properties in other possible worlds.

    I don't see why they wouldn't be. If there are statements that are true of possible worlds, (e.g. that in at least one of them I went to bed instead of writing these remarks), then that seems like a good prima facie reason for saying that those statements are actual, i.e., are among the states of affairs that constitute the actual world. But if you accept that, then you accept applying predicates to worlds within other possible worlds (e.g. that the world in which I go to bed has, in this the actual world, the property of having been accessible for me).

    Of course, you could deny the prima facie case, and argue either (a) that modal facts aren't in any possible world, but float outside (over? under?) any possible world; or (b) that there aren't really any facts about possible worlds at all. But the sort of transworld realm of facts required for (a) seems to undermine the idea of possible worlds as total; and to assert (b) just seems to be to give up on possible worlds as anything other than a useful fiction.

  3. Hi Alex, I assume that any given possible world-state could not itself have been different, and so has all its properties essentially. I was instead considering essentialism about the world-lump (which is not the same thing as a possible world-state). If we think of possible worlds as maximal properties, then we need something to instantiate such properties. That "something" is what I call a "world-lump". (Maybe it's a spacetime manifold, or universe, if those are the sorts of entities that instantiate maximal properties.) We need this to account for the metaphysical status of actuality. After all, only one possible world is actualized. It's the way that our world-lump is. The other possible world-states are merely possible, unrealized, uninstantiated. Their non-actuality consists in the fact that there are no world-lumps in such a state. They are ways that no world is.

    I think we have an independent grasp of modal notions like 'possibility' and 'necessity' (and, indeed, 'essentially'). Possible worlds talk can help illuminate some aspects of these bedrock notions, but I wouldn't take it as more fundamental. It seems merely confused to talk of our world lump as existing "in" various possible worlds. That is to treat an object as part of the properties it instantiates, which seems backwards. If you want to retain the claim that "P is necessary iff P is true in all possible worlds", you can talk about what's true in -- better, "according to" -- a possible world-story. That is, what would be true if the possible world-state were instantiated. It's this counterfactual fact that's fundamental, however. (N.B. I don't think a Lewisian analysis of this counterfactual in terms of possible worlds would be illuminating. Better to take it as primitive.) In other words, I use modal facts to specify possible worlds, rather than vice versa.

    Rad Geek: any way that our universe could not have turned out to be is (ipso facto) an impossible, not a possible, "way."

    Isn't that question-begging? I was proposing that we distinguish ways a world could be from ways this world could be. If these are distinct, then possible worlds, as maximal world-properties, should be identified with the former. My question, then, is whether there are ways that other universes, but not ours, could be. Your response merely asserts the denial of this claim.

    If multiple world-lumps could coexist, then multiple possible world-states could be co-actualized. If we accept the totality condition, then modus tollens will lead us to deny that multiple world-lumps can coexist. But perhaps it still makes sense to think that a numerically distinct lump could exist in place of ours. They just couldn't both exist at once. (This might be expected if we think of a world-lump as being a maximal object, such that all existing things are mereological parts of it.)

    "I really have no idea what I'm supposed to imagine about this numerically distinct, alien world that would make it something other than the world we're in, under different circumstances."

    Yeah, as a deflationist about essential identities, I'm sympathetic. But many philosophers often seem to take numerical distinctness to be a deep (brute?) metaphysical fact. If the wizard turns me into a poached egg, that's supposed to be "something other than [me] under different circumstances." So if you can get an intuitive grasp on the kind of numerical distinctness essentialists propose elsewhere, just apply that very same notion to the universe as a whole.

  4. Hey Rad Geek,

    My point was simply that possible worlds, in virtue of being maximal, aren't the sorts of things which overlap. But in order for a something to have a property in a world, it would have to overlap that world. Thus, a possible world is not something that could have a property in a world. (I originally thought that Richard's talk of the essential properties of "world-lumps" committed him to overlap, but now I am not so sure.)

    Note that this is a completely different issue from whether there are actual truths about non-actual possible worlds. Nowhere did I deny that there are such modal facts. But I think you could agree with me that a very bad way to analyze such facts is by saying that we truly ascribe predicates to possible worlds just in case there is located somewhere within the actual world something which bears the corresponding properties, namely that possible world! If you agree with me that this analysis is absurd, then I doubt that we're in disagreement elsewhere.

    Hey Richard,

    As I wrote before, none of the defenders of possible worlds as maximal properties thought that there is literally something in which these maximal properties are instantiated. For good reasons: What the hell could such a thing be? (Note that it could not be a spacetime manifold or the universe, since presumably it is possible that there could have been neither.)
    How many "world-lumps" are there? Ten? Just one? As many as there are way things might have been? I'm sure we could think of still more potentially embarrassing questions, but you probably get the point.

    Fortunately for those interested in the maximal property view, not every theory of property instantiation involves bare particulars or other pincushions for properties, so in any case it does not follow that the maximal property view is committed to anything like your "world-lumps". My impression is that defenders of this view took it that exactly one maximal property has an instance, took the instance of that maximal property to just be be the actual world, and left "instantiation" here unanalyzed. But it's been a while since I read those who defend this theory, so I'm interested now to see what they had to say about this issue, if anything. I'll post again if I come across anything radically different than what I've said above.

  5. "took the instance of that maximal property to just be be the actual world"

    In what sense of 'world'? Presumably you do not mean the possible world (maximal property) itself. A property is not its own instance. (Or if it is, then presumably all the other maximal properties are their own instances too, so they're all actualized.)

    So then you must mean 'world' here not to mean a property, but rather an object. That is, a world-lump, not a world-state. And that's exactly what my view is.

    (I thought Adams' view was something like this too. There are many possible world-stories, but only one of them accurately describes how the world[-lump] actually is.)

    You suggest instead that properties might be instantiated, though not instantiated by anything. That just sounds incoherent to me.* But maybe I'm missing something. (I'll be interested to hear what you "come across".)

    * = [Even on the Bundle Theory, the resulting bundle is the instantiating object, right? Maybe this is terminological.]

  6. Hey Richard,

    I may have wrongly attributed the maximal property view to Adams and Plantinga. In fact, I did: Their view is that possible worlds are something like maximal states-of-affairs, not properties. (I think that both of them use maximal properties in their theories of transworld identity, but that is neither here nor there. But I guess it explains the confusion.)

    You are correct: No theory of properties identifies them with their instances. I did not say as much. What I did say was that some theories of properties, and their attendent theories of property instantiation, are flexible enough to allow for the instances of a particular property to be identical to the objects which instantiate it. Here are the two examples which spring first to mind:

    For the bundle theorist, a property P is instantiated by an object x just in case P is compresent with the bundle of properties {P1,P2,...} which "make up" x. So the degenerate case will be if P is the only property instantiated by x. In that case, to say that P is instantiated by x is to just say that P is compresent with itself. (This sounds very odd, but this is how bundle theorists handle counterexamples involving "lonely" properties.)

    Likewise for the trope theorist, for whom properties just are classes of primitively resembling tropes, and for whom instantiation is understood in terms of something like compresence, as above. (But here there is a big problem: For the trope theorist, it turns out that each non-actual maximal property is the null set. So every non-actual possible world corresponds to the same maximal property. Something like this problem crops up with inconsistent properties, so perhaps the clever trope theorist could solve this problem, as well.)

    Hopefully this has shown that the maximal property view of modality is not committed to "world-lumps" to instantiate these maximal properties. But I'm still curious: If there are "world-lumps" numerically distinct from the one we actually inhabit, then how many are there? Would this be an item of a posteriori knowledge, or something about which we could consult our modal intuitions? It would help me to understand your view if I had some idea.

  7. I think there's only one world lump, and it exists necessarily. But on the alternative views I have in mind, there might be as many possible world-lumps as there are (say) possible laws of nature. If you change the laws, you've changed the identity of the lump. Something like that. (We can still insist that necessarily, only one world-lump exists. It's just that which possible world-lump it is might be a contingent matter.)


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)