Thursday, December 6, 2007

The maximally great island

We can formulate Plantinga's modal ontological argument as follows:

  1. Possibly, a maximally great being exists.
  2. A maximally great being exhibits maximal excellence in all possible worlds.
  3. Therefore, there necessarily exists a being that exhibits maximal excellence in all possible worlds. (By 1, 2 and S5.)

Here is a surprising fact. Gaunilo's maximally great island objection doesn't work. For let's try to construct a parallel:

  1. Possibly, a maximally great island exists.
  2. A maximally great island would have insularity and the maximal excellence compatible with insularity (i.e., being an island) in all possible worlds.
  3. Therefore, there necessarily exists a being that exhibits the maximal excellence compatible with insularity in all possible worlds.

But premise (5) is unjustified. For while we have reason to think that a maximally great being would be maximally excellent in each world, we do not have reason to think that a maximally great island would have insularity in each world. Typical islands we know are not essentially insular. Indeed, there are pieces of ground that are islands at high tide but that are not islands at low tide, so insularity is not an essential property for them. In fact, I don't know that any island has insularity essentially. Any island could survive being joined to the mainland by a narrow land-bridge. And even if it were possible to have an island that is essentially insular, it would be unclear that a maximally great island would be essentially insular. Essential maximal excellence is very plausible a great-making property. But it is far from clear that essential insularity is a great-making property.

But without essential insularity, the most we can show in the place of (5) seems to be this:

  1. A maximally great island exhibits insularity in some worlds, in which it exhibits the maximal excellence compatible with insularity, and in all other words, if any, it exhibits maximal excellence.
Note that in the worlds where the entity is not island, there is no need to limit its excellence to the excellence compatible with insularity. So what conclusion can we draw from (4) and (7)? Here it is:
  1. There necessarily exists an entity which in some worlds exhibits the maximal excellence compatible with insularity and in all other worlds, if any, it exhibits maximal excellence.
In particular, we have not shown that this being is actually an island.

But isn't what (8) shows just as absurd? A being that is an island in some worlds and God-like in others? Actually, I suspect (8) is true. God is maximally excellent in all worlds and in some worlds he is literally an island. What do I mean? Well, as a Christian, I take it that it is possible for God to become incarnate as a human being. But surely not just as a human being. One might think that by the same token God could become incarnate as any sort of being, say an island, so in some worlds he is both God and an island, whereas in the actual world he is God and man. Indeed, wouldn't we expect that the maximally great island would be God, if that were possible? Now one might object that God can only become incarnate as the sort of being that can be a person. But an island can be a person. We can imagine all kinds of persons: persons of carbon and water and the like like us, persons of plasma, etc. Why not a person of earthy stuff? All kinds of complex computational phenomena could be instantiated by geological interactions within an island. Of course, such a person might end up functioning very slowly. But that's fine. And dualists like me will demand a soul for them. But that, too, is fine. Some possible islands have souls, then.

This strategy won't work for every possible parody. In particular, they won't work for parodies involving maximally great beings that exhibit some quality incompossible with divinity, like sinfulness. But it is not clear that it makes sense to talk of a maximally great among sinners, say. Sinfulness is defined by falling short of moral greatness, so a maximally great among sinners is going to be minimally sinful. But for any sin, there is a lesser possible sin, so there is no such thing as minimal sinfulness, probably.

16 comments:

Mike said...

Typical islands we know are not essentially insular

How could that be? If an island G were contingently insular, then there would be a world in which G is pen-insular or worse. But there is no world in which any island is a peninsular. Therefore there is no world in which G is a peninsular. But then G is essentially insular. In short, worlds in which G is not insular are worlds in which G fails to exist. Note that I do not deny that the land L which composes the island G is not itself essentially an island; L could be a peninsular. But G could not.
If G is a maximally great island, and if G's maximal greatness entails that G exists in every world (the island G, not merely the land L) then there is no world in which G is not insular.

Alexander R Pruss said...

Mike:

Of course it is impossible that, de dicto, an island be a peninsula. But it does not follow that it is impossible, de re, for x, which is in fact an island, to be a peninsula.

I guess the question is whether islands are identical with landmasses are merely composed of the landmasses. I think our practice of speaking of islands, land bridges that come and go, etc., goes better with the idea that islands are identical with landmasses. (Actually, like van Inwagen, I don't think islands exist at all, so I can in persona propria reject the island parody simply for that reason. But I was trying to write on the basis of assumptions less controversial than my own view.)

Mike said...

It is also impossible de re. There is no island G that has the modal property of being possibly non-insular.
But if you insist that G has that property, then I'd have to insist that God has the property of being possibly imperfect. It is just that the worlds in which God is imperfect he is not God (since of course he is essentially perfect). I know, silly idea. But it is no less silly to say that an island G might be less than insular, but that worlds in which that is true are those in which G is not an island.

Alexander R Pruss said...

Mike:

Do you think that when a landbridge appears between an island and the mainland, the island ceases to exist?

Mike said...

Do you think that when a landbridge appears between an island and the mainland, the island ceases to exist?

It is a tricky question, since it depends in part of the sort of landbridge you have in mind. I'm inclined to believe that, given some very minimal bridge, the island vaguely exists, since it is vaguely insular. But if the landbridge is large, then yes I'm inclined to believe that the island is no longer insular--i.e., no longer an island.
But notice that all of this is beside the question of whether islands are essentially insular. We are talking now about the vagueness of 'being insular'.

Alexander R Pruss said...

Aha. So then the question really does come down to whether islands are landmasses.

Suppose I own an island. And then enough of a landbridge appears that what I own is no longer an island. Is it the case that my property has been destroyed and replaced by new property?

Interestingly, I think that if one says that islands are essentially insular, one should also allow for temporally gappy existence. For there are tidal islands, and surely it is the same tidal island at each appropriate tide. Two kinds of tidal islands are relevant: the kind which are an island at low tide and become submerged (and hence non-insular) at high tide, and the kind which are an island at high tide and become peninsular at low tide. I am in general not that bothered by temporally gappy existence, though.

Mike said...

Aha. So then the question really does come down to whether islands are landmasses

I'm pretty sure islands are not (merely) landmasses. Islands are exposed, insular landmasses. So their existence depends on having certain relational properties.

Suppose I own an island. And then enough of a landbridge appears that what I own is no longer an island. Is it the case that my property has been destroyed and replaced by new property?

In general, legal questions don't illuminate metaphysical ones. They settle metaphysical questions by precedent--as though such questions could be so settled. It is likely that legally you own not an island, but a piece of land. That is, you likely have a deed to a piece of land that is currently shaped island-wise. But if it is indeed a deed to an island, and not merely to a piece of land, then yes the island has disappeared in the case you describe. I don't see this as a problem. You could also own a sand castle or an ice sculpture, both of which have an unstable existence. I think you're right that this might involve a gappy existence for islands.
So I still contend that islands are essentially insular.

Alexander R Pruss said...

I wonder how one could settle a question like this. We both agree that islands are exposed, insular landmasses. But one of us thinks that if x is an island, then it is essentially an exposed, insular landmass and the other doesn't (an infant is a preverbal immature human; but it is false that an infant is essentially a preverbal immature human, since the infant does not perish upon learning to speak). Brute intuition? Do you have any thoughts?

I think the legal issues illuminate this question because island is not a natural kind, but a conventional kind. Thus, I think that what properties are essential to islands is settled by facts about our linguistic convention, in a way in which what facts are essential to frogs is not settled by facts about our linguistic convention.

Mike said...

But one of us thinks that if x is an island, then it is essentially an exposed, insular landmass and the other doesn't (an infant is a preverbal immature human; but it is false that an infant is essentially a preverbal immature human, since the infant does not perish upon learning to speak). Brute intuition? Do you have any thoughts?

But why doesn't the infant cease to exist? The person that instantiates the property of being an infant need not cease to exist, but the infant does. If Smith loses a hand in an accident, the two-handed person ceases to exist, but Smith doesn't.
Maybe it comes down to intuition alone. But certainly, imaginative counterfactual hypotheses lead intuitions. For example, consider an island to which there is a man-made bridge (cf. South Padre Island in south TX, for instance). My intuition is that the man-made bridge does not make the island non-insular (or makes it only vaguely non-insular). Counterfactual hypotheses that expose sources of vagueness for 'being an island' move the metaphysics forward.

Thus, I think that what properties are essential to islands is settled by facts about our linguistic convention, in a way in which what facts are essential to frogs is not settled by facts about our linguistic convention.

But these cannot settle issues of nihilism/non-nihilism wrt islands, or 3D/4D wrt islands or stage/worm views wrt islands. So the important metaphysics is not settled this way. Further, what some judge in some court decides on some issues reagrding what an island is, seems to me nearly totally irrelevant on the issue of what an island is.

Alexander R Pruss said...

Well, if I am an infant, then not only do I have the property of infancy, but I am identical with an infant. Otherwise, we have a proliferation of conscious beings: the human being, the infant, the child, the blond-haired boy, etc., which begin and end their existence at different times, but which nonetheless overlap.

Mike said...

Alex, there is a plain reductio ad absurdum if you want to believe this,

1. Well, if I am an infant, then not only do I have the property of infancy, but I am identical with an infant

And also believe this,

2. . . .(an infant is a preverbal immature human; but it is false that an infant is essentially a preverbal immature human, since the infant does not perish upon learning to speak)

It follows from these that EITHER you no longer exist (since the infant no longer exists and You = Infant) OR you are an infant (since it continues to exist and You = Infant.).

But in fact you BOTH exist and are NOT an infant. So, either (1) or (2) is false. That is, either you are not identical to the infant or the infant does not survive into your adulthood. I'd say it's the latter, wouldn't you say?

Alexander R Pruss said...

Mike:

Your argument seems to commit the classic de dicto / de re fallacy. That infant in my past exists now--but he is no longer an infant.

The locus classicus for this fallacy is the following lovely text from Plato's Euthydemus:
[Dionysodorus:] Well [...], and so you say that you wish Cleinias to become wise?
[Socrates:] Undoubtedly.
[Dionysodorus:] And he is not wise as yet?
[Socrates:] At least his modesty will not allow him to say that he is.
[Dionysodorus:] You wish him, he said, to become wise and not, to be ignorant?
[Socrates:] That we do.
[Dionysodorus:] You wish him to be what he is not, and no longer to be what he is?
I [Socrates] was thrown into consternation at this. Taking advantage of my consternation he added:
[Dionysodorus:] You wish him no
longer to be what he is, which can only mean that you wish him to
perish. Pretty lovers and friends they must be who want their
favourite not to be, or to perish!

Mike said...

Alex,

I formulated the argument so as to preserve your intuition that it is false that an infant is essentially a preverbal immature human, since the infant does not perish upon learning to speak. If you meant that the infant (de re) survives, but not the infant (de dicto), then you have conceded, directly contrary to what you say in this quote, that infants are essentially preverbal, immature adult. But I take it you don't concede that. That's why the argument does not commit the alleged fallacy.

Alexander R Pruss said...

I do not think it is correct to use "essential" in the case of de dicto necessities. It is a necessary truth that if x is an infant, then x is preverbal and immature. But it is not of the essence of any infant that it be preverbal and immature.

Mike said...

Alex, the de dicto use of the indefinite description 'an infant' in the expression 'an infant is essentially preverbal' entails that there could be nothing satisfying that description that is not preverbal. The de re use of the indefinite description 'an infant' in the same utterance does not entail that there could be nothing satisfying the description that is not preverbal. I say nothing about de dicto necessities. I am talking about the logic of descriptions (If you take this to be controversial, then you can replace de dicto/de re in this context with attributive/referential uses. My point is the same.)
You were claiming, as far as I could tell, given the context, that something satisfying the description 'an infant' might not be preverbal.

Alexander R Pruss said...

Necessarily, everything that satisfies the description "an infant" is preverbal at the time at which it satisfies this description.