Monday, July 23, 2007

Contingency and the principle of sufficient reason

This is a redated post, once again. The dialogue between exapologist and Steve in the comments may help to support, or not, my responses to Parsons on Secular Outpost two posts back.

In reading Steve Lovell's comments on the gap between the claim that the physical universe is contingent on the one hand, and the claim that it is dependent on the other, I think some version of the principle of sufficient reason is supposed to do th is kind of work. So I am redating this post on Wainwright's treatment of the Thomistic argument.

In trying to make sense of the Thomistic cosmological argument, I have found William Wainwright's introductory book on the philosophy of religion helpful.

Cosmological arguments seem to have two critical elements. One notes the contingency of the physical world. Every part of it is such that we can conceive of it not existing. Most people believe that if God were to exist, God would have the power to annihilate the universe completely. So there is some sense to be made of the idea that the universe is something that exists contingently, not necessarily.

But why, if the universe is contingent in this sense, couldn't the universe be contingent on nothing. That is the question Steven Carr is asking. We need explanatory principles, or what in philosophy are called principles of sufficient reason, in order to justify the claim that the universe needs something other than itself to explain it.

PSR1: For every contingent fact F some other fact F' obtains such that, given F', F must obtain.

This principle is incompatible with classical theism, for reasons which are similar to the ones Steven mentioned. It is a contingent fact that God freely chose to create a universe, according to classical theism. Or, if God had chosen to exist alone, that would be contingent.

So PSR needs to be revised. Wainwright offers some alternatives:

PSR2: There is a sufficient reason for the existence of every contingent being.

This doesn't entail tha there has to be an explanation of every property of that being, just the existence of the being. So, for example, that being freely choosing to do something can be fully nd completely contingent.

PSR3: Every contingent fact that requires a sufficient reason has one.

A contingent fact "requires" a sufficient erason if and only if 1) it is logically possible for it to have a sufficient reason and 2) it is unintelligible if it doesn't have one.

PSR4: There is at least some reason for every contingent fact.

I have seen this referred to as the principle of necessary reason, for every contingent fact there are necessary conditions for it.

Wainwright goes on:

"The weaker principles are strong enough to generate the conclusion that contingent being is caused by a self-exisent being."

The upshot of Wainwright's subsequent discussion is that at least PSR4 is supported by the success of human inquiry, and that therefore the weaker forms of PSR are more plausible than their denials. Hence, he does find some legitimacy in forms of the cosmological argument that use some of the weaker versions of PSR4, and he thinks they do lend significant support to the claim that the physical universe depends on something other than itself, which is self-existent.

15 comments:

Steven Carr said...

'PSR1: For every contingent fact F some other fact F' obtains such that, given F', F must obtain.'

Is that true?

Plantinga claims that it is not true for his theory of counterfactuals of freedom.

Whatever you think of his theory, Plantinga clearly denies that contingent counterfactuals of freedom need other facts to explain them.

Anonymous said...

If you'd even bothered to read the post properly Carr, then you would have noticed that Wainwright, and by implication Reppert, argue that it is not very plausible, for reasons involving freedom, like the ones you mentioned!

Deary me.

Steven Carr said...

'PSR4: There is at least some reason for every contingent fact.'

And Plantinga's FWD defense depends upon that not being true.

It does seem dubious to me.

Is there any reason why a uranium atom decays at 6:17 pm rather than 6:18 pm?

Nobel Prize to the first person who finds one!

Victor Reppert said...

The fact that it decays at all has many enabling conditions, such as the existence of the uranium. The argument might be construed as saying that there has to be something that is a necessary condition for the existence of any object that exists contingently.

Anonymous said...

I'm not convinced by any of the versions of PSR that refer to facts rather than "beings".

The idea that for every contingent fact F there is some other fact F' which is a sufficient condition for F seems to entail that there are no contingent facts or that there are infinite regress of them.

If we say that the explaining fact F' need not be logically sufficent for F, then the larger fact (F and F') is a contingent fact and will require some further fact F'' to explain it. Again, we either have no contingent facts or an infinite regress of them.

The more plausible principle is the one Victor has as

PSR2: There is a sufficient reason for the existence of every contingent being.

This will suffice for the Argument from Contingency for the existence of a necessary being (or beings) whose existence is a condition of the possibility of the existence of the material world.

The question then is how to motivate this principle. It certainly doesn't have the same intuitive punch as the equivalent in the Kalam argument of "Everything that begins to exist has a cause of its existence".

exapologist said...

Hi Victor,

Aside from some questions about some of the principles themselves, I'm not sure I see how these principles, plus the fact that there are contingent beings or contingent facts, entail that at least one self-existent being exists. How does the derivation go?

Best,

EA

Anonymous said...

I'd like to reword some things I put in my previous post and attempt to answer exapologists question about the derivation.

I said that the principle that all contingent facts have some explanation committed one to either the non-existence of contingent facts or to an infinite regress of them. I still want to say that, but I was trying to get last night's post written rather more quickly than I should have.

The problem with this principle is not merely that it entails that for every contingent fact F that requires explanation there is some other contingent fact (F & F') that requires explanation, but that the explanatory relationship between F and F' is itself contingent and therefore consistutes a further contingent fact requiring explanation. This is more problematic that the regress I pointed out in my previous post because F' on it's own may not be a contingent fact and if it isn't then the 'larger' fact (F & F') may be considered identical to the fact F.

Anyways ...

(1) Something exists
(2) Whatever exists does so either dependantly or independently.
(3) Not everything can exist dependently.
(4) Therefore something exists independently.
(5) Contingent beings can only exist dependently.
(6) Therefore, something exists necessarily which is the ground of contingent existences.

The PSR, in whatever form, is supposed to ground (5) and/or (3).

The kind of dependence in question is a dependence for current existence. If whatever form of PSR is in play is correct, then were the grounding/independent being somehow cease to exist so too would the grounded/dependent/contingent. All contingent existence is thought of as in some sense "borrowed", but not all existence can be borrowed ... who would it be borrowed from?

There is a very nice formulation of the argument in Beyond the New Theism by Germain Grisez, I thought I had a photocopy of the relevant passages but I can't seem to find that. There's also a nice formulation in Keith Yandell's Philosophy of Religion: A Comprehensive Introduction. I can definitely lay my hands on the latter, and will post a version of this when I get time.

Steve

exapologist said...

Hi Steve,

Thanks for your nice comment. Looking at the argument as you put it, I'm wondering what sort of evidence one could have for the version of PSR that's supposed to entail (5). It seems to me that it would have to be something along the lines that it's a synthetic a priori proposition or, what, an inductive "track record" argument, or perhaps that it's a presupposition of reason? Whatever it may be, is it strong enough to undermine the intuition of the seeming possibility of a logically contingent, yet metaphysically independent ("free-standing"), being (something along the lines of Swinburne's conception of God)? My intuition is that if any version of PSR entails that such a being is metaphysically impossible, then so much the worse for that version of PSR. For it seems to me to be explanatory overkill to require an explanation of such a being -- i.e., a logically (even metaphysically) contingent, yet eternal being that has, say, indestructibility (or at least the world-indexed essential property of being indestructible-in-alpha, the actual world) as an essential property.

In any case, that's how it seems to me at the moment. What do you think?

Best,

EA

Anonymous said...

Exapologist,

I agree. While I find such versions of the PSR relatively plausible, I can't see why an atheist interlocutor would accept them and I can't think of any good arguments for them either.

The arguments I have seen tend to jump back and forth between different conceptions of necessity/contingency (metaphysical vs logical) or to use the Aristotelian/Thomistic metaphysics mentioned above by Shulamite. I can't get my head round that potency/composite stuff.

Norm Geisler argues in this potency/composite fashion, and I remember studying his books rather hard during my PhD in an attempt to understand this and finally sending him a long email in which I asked what sort of necessity was in play. His reply was rather insultingly short: "God is actually necessary".

Does anyone out there understand these things? Shulamite, can you expand on your comments?

Steve

exapologist said...

Steve,

I remember trying hard to get my head around Geisler's Thomistic version from existential causality, too (the one in his Christian Apologetics text and his Philosophy of Religion text, etc., right?). I remember asking WLC what he thought of it, and he said that (i) he couldn't make sense of the notion of God as a being whose existence and essence are idential, and (ii) he couldn't find a persuasive reason to think that limited beings are composites of being and essence.

The versions of the cosmological argument that I find most interesting are Swinburne's version, and Craig's Big Bang version.

Best,

EA

Anonymous said...

EA,

Yes those are the Geisler texts I had in mind. I agree with WLC on beings whose essence and existence are identical. It's a tantalising and exciting idea, but I just can't seem to get my head round it. I'm not sure about the other criticism from WLC as on one reading it seems to conflict with his first criticism. Still, I think there are plenty of places to find trouble with the argument.

I've never been particularly interested in Swinburne on the Cosmological Argument, though admittedly I haven't read him much on that topic. This is mostly because he seems to assume that existence is more puzzling than nonexistence and that although he thinks God is contingent his existence would not require an explanation. It is, admittedly impossible for God to have an explanation, which may prevent us from looking for one; but it doesn't do anything to remove puzzlement over why He should exist rather than not exist. If God can "just be" then the world could too.

I like Craig's argument. I find the discussion of infinity fascinating. However, I'm no longer convinced by the philosophical arguments for the impossibility of a beginless past. As far as I can see Craig himself seems to have moved away from these arguments somewhat. My problem with them is that he want's them to prove that an infinite past is "metaphysically impossible" while admitting that they do not prove that it is "logically impossible". To me it's always been unclear how you prove the former without proving the latter. I still think there's something to be said for those arguments, but I don't think that Craig's presentation works in it's usual forms. The scientific version of the argument has more weight, I think, but is always open the possibility that the Big Bang will be reinterpreted as something other than an absolute beginning. I think the evidence for the Big Bang itself is very unlikely to be overturned, but reinterpretation is not so easy to dismiss.

Steve

Anonymous said...

A few posts up I promised a version of Keith Yandell's forumulation of the cosmological argument. Here goes:

(1) If it is logically possible that the truth of a logically contingent existential proposition be explained, then there actually is an explanation of its truth
(2)"There exists things whose existence is logically possible to explain" is a true logically contingent existential proposition.
(3) Therefore, there is an explanation of the truth of "There exists things whose existence it is logically possible to explain".
(4) The truth of "There exist things whose existence it is logicall possible to explain" cannot be explained by there being things whose existence it is logically possible to explain (the existence of those things is just what is to be explained).
(5) That a logically contingent existential proposition is true can only be explanation by some other existential proposition being true.
(6) So, some existential proposition conerning something whose existence it is logically impossible to explain, and whose existence can explain the existence of things whose existence it is logically possible to explain, is true.
(7) Something exists whose existence it is logically impossible to explain and whose existence can explain the existence of things whose existence can be explained.
(8) This something is God.
(9) So, God exists.

(8) and (9) are considerably shorter than Yandell's own argument. I'd already skipped a few intermediate inferences, but Yandell's numbered propositions go up to 24, so I think that can be excused in the context.

(1) is Yandell's version of the PSR. It evades many of the problems that usually beset such principles: It does not extain an infinite regress of contingent facts, and nor does it entail that all facts are necessary facts. He allows that it cannot be demonstrated, but also rightly points out:

Yandell: "A standard objection to [other] formulations of the cosmological argument is that if one infers from the world to God, and it is logically possible that God not exist, then one might as well have stopped with the world. A Cosmological Argument with [(1)] as an essential premise ... will be subject to no such objection. Further, if one rejects [(1)] one is left with an ultimate mystery, an intelligble and basic question to which there might have been an answer, but is not. Reject [(1)] and the mystery likes on your side of the fence, not on the monotheists' side."

Steve

exapologist said...

Steve,

It seems that our views about Craig's arguments are pretty much the same: the philosophical arguments about a finite past don't seem to work, and the Big Bang argument, while having some force, has some contemporary rival explanations that don't entail a singularity that precludes prior physical antecedents.

I like Yandell's version. Van Inwagen has some interesting things to say in reply to this sort of modal version in his textbook, Metaphysics (the one from Westview Press; not the "Big Questions" text).

Best,

EA

exapologist said...

Whoops! Looking back at Yandell's argument, Its not really a modal cosmological argument in the sense that Van Inwagen means in his text -- never mind about that last comment.

-EA

exapologist said...

Hi Vic,

I don't yet see any advance over my objection about the case I can't rule out as impossible. Here it is again:

" I'm wondering what sort of evidence one could have for the version of PSR that's supposed to entail (5). It seems to me that it would have to be something along the lines that it's a synthetic a priori proposition or, what, an inductive "track record" argument, or perhaps that it's a presupposition of reason? Whatever it may be, is it strong enough to undermine the intuition of the seeming possibility of a logically contingent, yet metaphysically independent ("free-standing"), being (something along the lines of Swinburne's conception of God)? My intuition is that if any version of PSR entails that such a being is metaphysically impossible, then so much the worse for that version of PSR. For it seems to me to be explanatory overkill to require an explanation of such a being -- i.e., a logically (even metaphysically) contingent, yet eternal being that has, say, indestructibility (or at least the world-indexed essential property of being indestructible-in-alpha, the actual world) as an essential property."

As far as I can tell, my case above is also an undercutting defeater for Wainwright's ammended versions of PSR, or their bases.

As I look back to Yandell's version that Steve Lovell was kind enough to present, I think his version of PSR (which is his premise (1)) is pretty dubious. For starters, it relies on logical possibility as a reliable (or reliable enough) guide to possibility, and as someone who's writing their doctoral dissertation on thought experiments and the epistemology of modality, I hope my opinion will carry at least non-negligible weight in saying that it's harder to find a worse guide to metaphysical possibility than logical possibility.

What about broadly logical possibility in Plantinga's sense? Yes, that would help. The problem, though, is that we need an account of the epistemology of broadly logical possibility -- we need to know when or whether something *is*, in fact, broadly logically possible. What are the candidates? Here are some biggies:

-Humean imaginability
-Yablo-style imaginability
-Chalmers/Jackson-style two-dimensionalist imaginability (other "epistemic possibility" accounts go here as well)
-Cartesian conceivability
-Bealer/Van Cleve/Tidman-style rational intuition accounts
-Heimir Geirsson's revised Yablo-style account
-Peter Kung's account
-believabiliity accounts
-entertainability accounts

The problem with all such accounts is that they leave us with a nasty dilemma: either (i) they admit of straight counterexamples, or, (ii) while plausible, they're unable to justify possibility claims remote from ordinary experience. If so, then there is no account of conceivability-possibility inferences on offer that both (i) are plausible, and which (ii) justify "high-flying" possibility claims. But of course this is bad news for Leibnizian cosmological arguments.

So those are my main criticisms:

(i) My epistemically possible scenario of the logically contingent yet metaphysically independent being is an undercutting defeater for all versions of PSR strong enough to play a role in Leibnizian cosmological arguments (including Wainwright's),

and

(ii) Leibnizian cosmological arguments turn on possibility claims beyond the reach of prima facie justification.

Best,

EA