Sunday, July 29, 2007

Reply to exapologist on cosmological arguments

EA: 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."

VR: I’m wondering if this is enough to block an argument in defense of something like theism against naturalism. The very idea of “indestructible-in-alpha” strikes me as incoherent. “Indestructible” is a claim with modal force, therefore it can’t be a world-indexed property. Do you just mean “not actually destroyed in alpha?” So how contingent is this being, really?

Of course even at best cosmological argument doesn’t prove the existence of a being omniscient, omnipotent and perfectly good, or a personal God, etc. However, if successful it does refute the naturalistic view that the physical world is all that exists.

Cosmological arguments have to deal with the termination problem. What characteristics are needed to terminate the chain of explanations? The Unmoved Mover argument holds that explanations cannot terminate in that which is in motion, but must rather terminate in an ummoved mover. The argument from contingency holds that the physical world is contingent, and therefore depends on something that is not contingent—that is, necessary. Perhaps we can develop the “argument from destructibility.” Whatever is destructible must depend for its existence on what is not destructible.

Of course, naturalistic cosmological argument critics maintain that there if we extend causal reasoning beyond the space-time manifold, there are no “termination-making characteristics” that can be applied to God but not to the physical universe. That’s the claim to be found in Parsons’ atheist manifesto. There are two ways of casting this issue. One is to suppose that a naturalistic world is beginningless. Does that beginningless world possess all the termination-making characteristics?

The matter gets more complicated when we accept the Standard Big Bang theory. With this picture the universe had a temporal beginning. Now that beginning is the beginning of time itself, so unlike the case of a Bengal tiger popping into existence, there is no time before the universe exists, when time exists and the universe does not. I have a strong intuition that nevertheless this does not dissipate the sense that there must be a cause of the physical universe. Does it make sense to say that the universe had a beginning moment but cannot possibly have a final moment?


exapologist said...

Hi Victor,

I think there's a perfectly intelligible sense in which a bearer x could have the property of indestructibility in such a way that it's indexed to a possible world W, viz., in the sense that there is nothing in W that has what it takes to destroy x. This fact then grounds the counterfactuals about x's indestructibility in W, so that (using Lewis-Stalnaker possible worlds semantics for counterfactuals) x is indestructible in all the worlds "closest" to W.

Now my question is, why in the world should we think that such a thing needs an explanation in terms of something beyond it? I think it's perfectly reasonable to say that it wouldn't be surprising in the least if it had no such explanation. But PSR will have none of that. No terminus of explanation will do unless it's a being that exists in all possible worlds. But again, surely this is explanatory overkill. For as my epistemically possible case points out, there are less modally extreme candidates that can end the terminus of explanation in a satisfactory manner. But if so, then the motivation for PSR is undercut. To push the debate forward, then, we need an argument for why my sort of candidate explanatory terminus is somehow inadequate -- one that doesn't beg the question against an agnostic like myself.

Now you point to the standard Big Bang model as evidence that my candidate is off the mark. My main reply is that there are a number of plausible models in play that explain the origin of our universe in terms of physical antecedents; such models seem to me at least as plausible as the hypothesis of an immaterial tripersonal creator out of nothing. In addition, my uniform experience is that originating causes for material objects involve antecedent material objects. In light of these considerations, the naturalistic and theistic hypotheses seem, at the very least, evidentially counterbalanced.

At any rate, these are my initial reactions.



Anonymous said...


We've interacted on the PSR before, and I think we're pretty much in agreement on that, so I won't add anything further here.

I am, however, vaguely intrigued by your mention of "physical antecedants" of the big bang.

Are these antecedants "in time"?

If so, you simply haven't gone back far enough. The real question is about those antecedants. If not, then:

(1) In what sense are these things "physical"?
(2) How can atemporal states plausibly described as "physical" cause temporal becoming? There's an extra stage to the Kalam argument which is supposed to show that the cause must be personal. How do you avoid that argument?


exapologist said...

Hi Steve,

There are several models, but here's one. According to M-theory (the theory that unifies the five versions of superstring theory), there are entities called 'branes', or multi-dimensional membranes (ranging from 0 (for point-particles) to 9 dimensions, and our universe is just one 4-dimensional brane among many branes existing within a larger 11-dimensional spacetime. Thus, according to M-theory, the beginning of our universe is not the absolute beginning of time -- time may well have no beginning. On M-theory, then, Craig's point about the need for an agent cause of an absolute temporal beginning doesn't arise.

Apart from this, though, I think Wes Morriston has nicely shown that Craig's arguments about the necessity of agent causation for absolute temporal beginnings fail on independent philosophical grounds. Indeed, I think his large series of articles on the kalam argument show that there is little about the argument that is right. You can access his articles online by going to his department webpage at University of Colorado, Boulder. You can find it by just Googling "Wes Morriston department of Philosophy".



stunney said...

I think a cosmological argumnt can be re-cast as an abductive inference to a best explanation.

If some impersonal universe-creating thing——---something that doesn't possess a free and rational will——---has always existed, or exists eternally or timlessly (like a timeless computer program, or timeless mathematical structure, or some universal quantum law or field ,or simply some universe-generating mechanism—–---we don't know what), then either universes should have always existed, or this universe should have always existed. But there is zero scientific evidence for either proposition. So that's one problem.

It would not be 'up to' any impersonal, non-mindlike universe-generating thing whether to generate universes. Its generation of universes follows impersonally from its nature as a universe-generator. It necessarily generates a universe, or multiple universes. And it always does what its nature dictates because it always has the nature that it has (whatever that nature is). And the universe-generator has always–—-for an infinite past duration--–—had its nature.

By contrast, if some free rational mind capable of universe-creation has always existed, then there's no necessity that there be a universe at all, since whether there is or not depends on that mind's free choice. And if that mind chooses to create a universe at all, there'd be no reason that such a universe would have to have (though it could have) an infinite past. Quite the contrary. Such a created universe could easily have a beginning along with time itself. And this is what, in fact, we actually observe to be the case scientifically with regard to our universe. (We know of no others, though there could be others.)

Now, an alternative might be that either this universe, or a universe-generator, simply popped into existence, with its laws and constants and initial conditions, all of which are fine-tuned for life at least in this universe, out of absolute nothingness, or what I sometimes refer to as the Acme of Ontological Zilchness. But how plausible and how probable is that?

Even if one asserted such a thing, there would no possible scientific evidence for it, since there logically cannot be a science that's about absolute nothingness. And one shouldn't not confuse quantum vacua or other fields with absolute nothingness. Such vacua and/or fields are not 'nothings' but 'somethings'––namely, fields or structures that are governed either by quantum mechanical laws or whatever laws the ultimate physical theory or ToE may specify.

And it's evident that we can quickly go from one universe-generator to a universe-generator-Generator-GENERATOR et cetera ad infinitum. But no matter. If there's always been such a thing or such a series, it should always have been GENERATING universe-generator-Generators...

I suggest that theism is a better explanation than naturalism for the following reasons:

1) It fits better with observational data pointing to a finite past

2) It fits better with observational data pointing to the universe's uniqueness

3) It fits better with the appearance of the universe's logical/metaphysical contingency

4) It terminates the explanatory regress with the fewest number of universes and/or universe-generating mechanisms.

5) It doesn't postulate an implausible transition from an Acme of Ontological Zilchness to a universe, since there is never any such Acme given the eternal existence of God.

Anonymous said...


Thanks for your response. Although I can't say it leaves me any the wiser.

I've come to think that at the frontiers of theoretical science, the "scientists" can get away with saying anything and be taken very seriously despite the fact that what is being said is so contrary to reason and common sense as to seem utterly without sense. It's a very curious thing for a discipline supposedly rooted in experience and experiment. I'd quite like to see how string theory would cope with Hume's fork (as in the passage which ends "commit it then to the flames!"). Not that I endorse the fork, it would just be good fun to see that discussion.

Despite these doubts about theoretical science, I do tend to favour realism as a model of the scientific endevour, but for reasons allied to the Argument from Reason, I think such realism is at home in a theistic framework and not at home in an atheist framework, especially when we stray to such theoretical extremes. (My essay on the Argument from Reason at, gives my reasons for thinking this way.) But to offer such thoughts as a way of avoiding the Kalam argument seems to require scientific realism. By my lights this makes the atheists position look very awkward, if not actually incoherent.

All that said, if someone decides to endorse such a theory as an alternative to theism (although it isn't necessarily an either/or choice) I can't see how I would refute him, but to paraphrase a Chesterton quote I have recently used elsewhere on this blog:

"It is absurd for the atheist to complain that it is unthinkable for an admittedly unthinkable God to make everything out of nothing, and then pretend that it is more thinkable that "strings" existing in 0- to 9-dimensional branes themselves existing in an 11-dimensional spacetime should turn themselves into everything."

As to Wes Morriston on agent causation, I'll have to go away and do some reading, but the bulk of your response rather suggests that you don't want to lean on this line of thought as you intend to deny that the (multi-)universe has a beginning.

That has always seemed the atheists best course, but given that this now depends on such peculiar theories as those you have sketched, I think the Kalam inference certainly can't be thought unreasonable.


exapologist said...

Hi Steve,

I must say that I was surprised to see you raise concerns of peculiarity and lack of direct experimental support regarding M-theory, given that your explanation of the same data is in terms of the hypothesis of an immaterial tripersonal creator-out-of-nothing.

I'm not sure if you were serious about the Rortian line you took against theoretical physicists -- that was only half-serious, right? M-theory isn't just pulled out of thin air. It accrues at least some justification via embodying various theoretical virtues, such as it's incredible explanatory scope and power, as it unifies all of the versions of superstring theory, and most importantly, it holds serious promise of reconciling the extremely well-supported but hitherto seemingly incompatible theories of General Relativity and quantum mechanics.

At any rate, I think M-theory, with its eternal 11-dimensional spacetime manifold and its many embedded membrane universes, is at least as plausible a hypothesis as theism regarding the data of the existence and apparent fine-tuning of our universe. Throw in the data of massive (and apparently random and pointless) animal and human suffering, massive religious diversity, religious ambiguity, divine hiddenness, studies indicating the ineffectiveness of prayer, etc., etc., and the scales start tipping in favor of a non-theistic hypothesis.

Of course, we need to add the pro-theistic data of religious experience, the apparent irreducibility of consciousness, etc., but when mixed in with the above-mentioned data, I think the evidence is going to be pretty much counterbalanced
(Pr(Theism)=Pr(naturalism)). In any case, that's my overall assessment in a nutshell.

I think Victor's suggested Bayesian approach to evaluating these matters is the way to go. It would be interesting to see everyone on this blog take Reppert's advice and try to see if we can push the debate forward. Thus, suppose we list naturalism and theism as the live hypotheses, and then make a big list of data -- some pro-theism, some pro-naturalism. Something like this:

H1: Theism
H2: Naturalism

D1: Apparent contingency
D2: Apparent fine-tuning
D3: Apparent irreducibility of consciousness
D4: Religious experience
D5: Massive amounts of human and animal suffering
D6: Radical religious diversity
D7: Divine hiddenness
D8: Apparent ineffectiveness of prayer

etc., etc.

What's the posterior probability of theism on D1-D8? What's the posterior probability of naturalism on D1-D8? Can we add to this list? If so, does it significantly change our assessment of the respective probabilities of our hypotheses?

So what do you think? Is this kind of what you had in mind, Victor? If so, then I'm very interested in seeing this sort of assessment.



Anonymous said...


Well, I'm not entirely serious about the Rortian attack on scientifc realism. But I think on the assumption of atheism those arguments raise some really serious problems to which I cannot see a decent response.

On the Humean line and lack of verification for string theory, my point is simply that the atheist too often wants to have things both ways. The unreasonable atheism I am opposing here does things like this:

(1) Complain that theism is incoherent because it is unverifiable/unfalsifiable.
(1*) Escape arguments for theism by endorsing theories which are unverifiable/unfalsifiable.
(2) Endorse these theories under a Realist interpretation.
(2*) Endorse an atheism which doesn't support realism about science.

By the way, as far as I understand them, many of the current interpretations of Quantum Theory and Relativity Theory are heavily based in a verificationism which is now completely discredited. I've often wondered what those theories would look like had they not been formulated in the hay day of positivism.


exapologist said...

Hi Steve,

Yep, I'm with you about the untenable double-standard that is unfortunately adopted by some atheists in the way you mention: they complain about the lack of direct verifiability and direct experimental support re: the theistic hypothesis, but then go on and endorse a naturalistic alternative that's in the same boat in these respects. I couldn't agree more with you here. I think this is where Victor's recent post on Bayesian methods comes in nicely. *Any* hypothesis-- I don't care whether it appeals to gods or not -- is capable of accruing justification if it leads us to expect the data.

Re: verificationism and relativity theory: I've heard/read people like Moreland and Craig say things like this, but I think Craig's point was in reference to the Special Theory of Relativity, not the General Theory of Relativity, no?