Baldwin’s Counter-Ontological Argument
June 7, 2004 — 8:23

Author: David Efird  Category: Existence of God  Comments: 9

In his paper, ‘There might be nothing’ Analysis 56: 231-38, Thomas Baldwin gives the following argument, which he terms ‘the counter-ontological argument’:
(1) It is a mark of concrete objects that they do not satisfy the Identity of Indiscernibles. So the identity of a concrete object is not determined by the intrinsic properties which determine what kind of thing it is.
(2) In the case of any being whose existence is necessary, the fact that its existence is necessary is determined by the kind of thing it is, and thus by its intrinsic properties.
(3) For any being whose existence is necessary, the intrinsic properties which determine its existence also determine its identity.
From these premises, it follows that there are no concrete, necessary existents. Now if one supposes also, that is, in addition to (1) – (3), that:
(4) If God exists, he is the sort of thing that can be causally efficacious.
(5) Anything that is the sort of thing that can be causally efficacious is concrete.
(6) If God exists, he is a necessary existent.
It follows that God does not exist. What to do?

Now many will be tempted to deny (1). But I quite like that characterisation of concreteness, and I would like to see how else one might resist Baldwin’s argument. I think (2), (4), and (6) are also true. So the one to deny, I think, is (3).
Denying (3) amounts to asserting that there could have been divine twins, that is, beings having all of the repeatable intrinsic properties that God has, which, let’s say, are the perfections. So, I’m claiming that there could have been two co-existing perfect beings.
Resistance to my claim will, I think, come on four fronts. Some will say that uniqueness is itself a perfection, so there couldn’t be divine twins. I’m just not moved by that assertion. Is a being that has the opposites of all of the usual perfections, like the omni-s, but is also unique perfect in any way? I’m just not moved to say that such an individual is. But intuitions, will of course, vary. And so so those who have intuitions other than mine will have to resist Baldwin’s argument in another way if they want to maintain God’s existence.
Another way of resisting my claim is conceding that uniqueness is not itself a perfection, but omnipotence is, and having that property entails being unique. Joshua Hoffman and Gary S. Rosenkrantz make this argument in their recent book The Divine Attributes. Suppose for reductio that there are two omnipotent agents, Dick and Jane.
“If this were possible, then it could happen that at some time, t, Dick, while retaining his omnipotence, attempts to move a feather, and at t, Jane, while retaining her omnipotence, attempts to keep that feather motionless. Intuitively, in this case, neither Dick nor Jane would affect the feather as to its motion or rest. Thus, in this case, at t, Dick would be powerless to move the feather, and at t, Jane would be powerless to keep the feather motionless! But it is absurd to suppose that an omnipotent agent could lack the power to move a feather or the power to keep it motionless. Therefore, neither Dick nor Jane is omnipotent. As a consequence, it is impossible that there be two co-existent omnipotent agents” (168).
The right response to make to this traditional argument, which appears in essentials also in work by Lacantius, John Duns Scotus, and William Wainwrignt, is to deny that moving the feather is a coherent task for Dick (given Janes’ will to keep it motionless) and that keeping the feather motionless is a coherent task for Jane (given Dick’s will to move the feather), and since inability to do incoherent tasks is no bar to omnipotence, such a thought experiment does not show that there couldn’t be two omnipotent agents. What this thought experiment shows is that the coherence of a task is dependent on contextual factors.
This resolution is rather similar to the resolution of the liar paradox on which the liar sentence does not express a proposition. But, as Kripke has shown, liar paradoxes arise not just in the case of a special sentence, but also in seemingly ordinary situations, when contextual factors render certain sentences paradoxical.
Consider the following case suggested by Kripke (‘Outline of a Theory of Truth’, pp. 691-92). Say that Jones makes the following claim:
(A) Most (i.e., a majority) of Nixon�s assertions about Watergate are false.
This assertion, in and of itself, has no paradoxical features that might be thought to prevent it from being true or false. Say that (A) is the only sentence that Jones utters about Watergate, or that all, except perhaps (A), of his assertions about Watergate are true. Now assume that Nixon�s assertions about Watergate are evenly balanced between true assertions and false assertions, and that Nixon then utters the following sentence:
(2) Everything that Jones says about Watergate is true.
Just as there is nothing intrinsically paradoxical about (A), there is nothing intrinsically paradoxical about (B), but together, they create a paradox. One might say that (A) and (B) are extrinsically paradoxical in Kripke�s thought experiment. With this example we see that, in Kripke�s words: �many, probably most, of our ordinary assertions about truth and falsity are liable, if the empirical facts are extremely unfavorable, to exhibit paradoxical features� (p. 691). Similarly, many � probably most � of the tasks we take to be coherent in ordinary cases are liable, if the case is extremely unfavourable, to exhibit paradoxical features. These paradoxical features render the sentences and the tasks that exhibit them incoherent. While, on the resolution of the liar paradox I have in mind, the liar sentence is incoherent in any context whatsoever, (1) and (2) are not coherent in the context given by Kripke, even though they are coherent in other, ordinary contexts. One might say that the liar sentence is intrinsically incoherent, while (1) and (2) are extrinsically incoherent in the context given by Kripke, and most, if not all, sentences are either intrinsically incoherent, or liable to be extrinsically incoherent depending on the contexts in which they are uttered. Thus, whether or not a sentence is coherent is dependent on contextual factors. Similarly, while the task of making two and two add up to five is not a coherent task in any context whatsoever, that is, it is intrinsically incoherent, the task of moving a feather and the task of keeping a feather motionless are incoherent in the one offered by Hoffman and Rosenkrantz, that is, they are extrinsically incoherent in this context, even though they are coherent tasks in ordinary contexts. Thus, whether or not a task is coherent is dependent on contextual factors, and an omnipotent being need not be able to perform an incoherent task, regardless of whether it is intrinsically incoherent or extrinsically incoherent. I conclude that there is no case from omnioptence against the possibility of there being divine twins.
A third objection to my claim that there might have been divine twins is that it’s impossible to individuate them. I suppose the only thing to do here is to appeal to haecceities, which is in effect a denial of a substantial need for individuation.
A final objection is that it’s just a matter of revealed faith that there couldn’t be divine twins. There’s not much I can do with that objection — only to ask, so what premise of Baldwin’s argument are you going to deny? I think (3) is the best.

  • jon kvanvig

    On the twin omnipotent beings case, I agree with your conclusion but not with the argument you give. You challenge the coherence of the task of Dick’s moving the feather. But notice you have to include a parenthetical qualifier. So it is not Dick’s moving the feather that is an incoherent task, but rather his moving the feather on condition that Jane wills something contrary.
    I think the right way to undermine the argument is to point out that you get the uniqueness claim only by assuming that it is possible for the two beings to will in opposition to each other. That’s an extra assumption, and there is nothing necessary about it. It is possible to have two beings whose wills are necessarily cooperative, and if so, the argument fails.

    June 7, 2004 — 8:53
  • jon kvanvig

    I’m wondering about how “the coherence of a task [can] be dependent on contextual factors”? On the face of it, coherence is a logical notion; for a task to be coherent is for a description of it to fail to imply a contradiction. If the description fails to imply a contradiction, it will fail to do so in every context.
    Maybe you want to build the contextual factors into the task. The task then becomes something of the form “doing X in context C.” But then we’ll have a hard time understanding omnipotence in terms of possible tasks (as an aside, Flint & Freddoso have other arguments against this way of thinking of omnipotence). You can’t say, in particular, that being omnipotent requires being able to perform any possible task. Think of “repenting when one has sinned,” “remembering what one has forgotten,” “carrying a twig when one is an ant.”
    So I prefer a different way of show that the Hoffman and Rosencrantz scenario fails to show that there can’t be two omnipotent beings.

    June 7, 2004 — 9:38
  • Judging by the subject matter, I assume this is David Efird and not David Talcott. Should we have a way to distinguish between the two Davids? Matthew?

    June 7, 2004 — 10:31
  • I have one suggestion, which I think may have been lurking behind Jonathan’s comments. One way of thinking about the Trinity involves erring on the side of tritheism rather than erring on the side of modalism, at least as van Inwagen describes it. If the Trinity is composed of three beings whose nature somehow involves some essential connection in their wills, then you get something like the case you’re discussing.
    Is it possible for the persons of the Trinity to act in opposition to each other? I think so, at least in one sense, if Lewis’ compatibilist sense of freedom is the right one, since each would be free with respect to factors not involving how the other persons will (but not free with respect to how the other persons will). I happen to like Lewis’ account of freedom, so this won’t work for me as easily, but if freedom isn’t contextually sensitive then Jonathan’s suggestion seems worth including.

    June 7, 2004 — 12:17
  • jon kvanvig

    Jeremy’s response here ties in nicely with his post on Leibniz, in the following way. It is certainly logically possible for the three persons of the Trinity to will differently; the crucial question is whether it is metaphysically possible. I claimed that it was (metaphysically) possible that they all will alike essentially. But of course that implies that it is not metaphysically possible that they will differently.
    I have no argument here… The only argument I have shows that it is not logically impossible that they will alike essentially, and it’s the weak kind: basically, I challenge those who think otherwise to derive a contradiction from the description, and I claim it can’t be done. Too bad for me… I need to go find an argument…

    June 7, 2004 — 16:22
  • The issues here overlap the issues that underlie Scotus’ argument for unicity and Kvanvig’s first point seems quite right.
    Scotus thought he had an argument against the possibility of two omnicompetent beings which I don’t think works. Whereas he tried to argue that there couldn’t be two omnicompetent beings from the possibility of their conflicting and the absurdity of such a state of affairs, I argued:
    (1) It seems that an omniscient being would know how to attain its ends.
    (2) Such a being would know that for any impossible act, that it was impossible and wouldn’t form an intention to perform it.
    (3) Contraverting the will of an omnipotent being would be such an act (along with squaring circles, striking dead the shortest spy without harming Ortcutt).
    (4) It seems that two omniscient beings, knowing how to avoid trying the impossible, would know how to coordinate its intentions with all other omnipotent beings.
    (5) Thus, it seems describing two subjects as omnicompetent but in conflict is latent nonsense.
    We can reject the possibility of such a scenario I think using the above reasoning.

    June 8, 2004 — 18:53
  • I find (2) and (3) dubious in the Baldwin argument. So, here’s (2)
    (2) In the case of any being whose existence is necessary, the fact that its existence is necessary is determined by the kind of thing it is, and thus by its intrinsic properties.
    How are we to understand this? I’m assuming that the fact that a sortal property p is “determining” that existence is necessary entails that p is had essentially (or here, necessarily, as well). Then is the claim that the proposition *God exists* supervenes on the proposition *God has p* Well, OK, but they entail each other, and thus the supervenience will go the other way, too. Are we to understand the claim counterfactually? Then we’re into the quagmire that is counterpossibles.
    (3) For any being whose existence is necessary, the intrinsic properties which determine its existence also determine its identity.
    I’m not sure how to understand this. So, the claim is something like x’s having p1, …pn underlies the truth of the claim that x=y? I would have thought we had enough trouble reducing identity for mere contingent entities. I’ve no idea what to think of this claim when it’s applied to necessary beings, especially non-spatial (and perhaps atemporal) ones. I guess it might be true in a very uninteresting way: if p1=*being omnipotent*, and this is an individual essence of God…
    Also, a case can be made that (1) is false too: For coincident objects, one may take the sortal properties to be explanatorily prior to other intrinsic properties.

    June 8, 2004 — 19:07
  • Matt, why is it a problem if supervenience goes the other way? Its necessary existence is determined by the kind of thing it is, and the kind of thing it is is determined by its necessary existence. It’s not as if the conjunction is impossible, and it’s not as if you leave something unexplained (since both are necessary and not in need of explanation in terms of something else).
    All he needs for (1) is that the identity isn’t wholly determined by the intrinsic properties. If there are coincident entities, that’s still so. Consider the case of two duplicates of coincident entities. The existence of the other set of coincident entities shows that it’s not wholly the intrinsic properties at work in distinguishing them. I think the argument will be unaffected by this modification.
    In (3), it has to be something like “x has p1, … pn” and “y has q1, … qn” underlies “x=y”. I’m not sure if that affects your point, but you need to have the intrinsic properties listed for y also, or you couldn’t ground the proposition that x=y.

    June 9, 2004 — 8:52
  • Kevin Vallier

    What is wrong with denying premise (5)?
    (5) Anything that is the sort of thing that can be causally efficacious is concrete.
    If you take the platonist position on universals, then universals are both abstract and yet, must be, from what I can tell at least, causally efficacious.
    Why not just resurrect the old doctrine of formal causation?

    October 12, 2004 — 22:41