People I respect keep telling me there is marvelous work being done in the area of theology. I have never encountered it, and not for lack of looking. Sometimes I wonder, though, whether perhaps I am just reading the wrong things. The religion and theology section of my university's library is quite large, and the percentage of it I have read is quite small. So I am open to the possibility that the really good arguments are in the books I have not read.
As it happens, Jerry Coyne has been wondering the same thing. A philosopher correspondent of his selected two books for him to read. The correspondent suggested that theology is the wrong place to look for discussions regarding God's existence, and that the analytical philosophy of religion provides more nourishing fare. The two books were The Existence of God, by Richard Swinburne, and Perceiving God: The Epistemology of Religious Experience by William Alston. Coyne's correspondent writes,
Either work may well irritate you or leave you dissatisfied in various ways. But I predict that if you approach it with an open mind, it won't elicit the sort of commentary that other believers' works have occasioned in your posts. Instead, I expect you will manage more respect -- thereby achieving (in my opinion at least!) a deeper respectability.
I have not read Alston's book, so I will have to add it to my list. I am familiar with Alston from his writing on the problem of evil. He argues that human epistemic capabilities are so limited, especially when compared to God's infinite wisdom, that we should not expect to comprehend why God would allow intense pain and suffering. The question is simply too complex for our limited understanding, in rather the same way that quantum physics is too complex for a dog. Suffice it to say, I do not believe atheists have much to fear from such arguments.
I have read my share of Swinburne, however, including The Existence of God. I fear he had the opposite effect on me from what Coyne's correspondent described. It is not anything I learned from the fundamentalists that has driven me to my generally negative opinion of theology and the philosophy of religion. It is people like Swinburne who did that.
On the other hand, I read The Existence of God several years ago, so perhaps it is time to revisit it. So here is the first of what will be a very occasional series analyzing Swinburne's arguments. I warn you in advance that I may lose interest in this part way through, but let's have a go at it anyway.
Chapter One is mostly a groundlaying chapter. It's most important statement is probably this:
I take the proposition `God exists' (and the equivalent proposition `There is a God') to be logically equivalent to `there exists a person without a body (i.e. a spirit) who is eternal, is perfectly free, omnipotent, omniscient, perfectly good, and the creator of all things.'. I use `God' as the name of the person picked out by this description. I understand by God's being eternal that he always has existed and always will exist. (p. 8)
Swinburne goes on to elaborate further regarding what he means by this terminology, but I will take his intention as sufficiently clear.
Much of the chapter is given over to logical preliminaries. Swinburne distinguishes first between deductive and inductive arguments. He notes that his case for the existence of God is primarily inductive. He then draws a further distinction between P-inductive and C-inductive arguments. The former are those whose premises render the conclusion probable, while the latter are those whose premises render the conclusion more likely than it would have been otherwise.
His main thesis is that the classical theological arguments, such as the cosmological and ontological arguments, are good C-inductive arguments. In each case the evidence cited makes God's existence more likely than it would be without that evidence. He implies, but does state explicitly, that he regards the sum total of these C-inductive arguments to be a strong P-inductive argument. He also considers the argument from evil and concludes that it does not constitute a strong C-inductive argument against God.
For me, the big red flag goes up when he starts using the notation of probability theory to express his ideas, as he does here. I will withhold final judgment until I see Swinburne's specific application of it later in the book, but I know from sad experience that such formalism often serves as a way of lending phony precision to bad arguments.
The problem comes when we try to apply Bayes' Theorem to a quantity like, ``The probability that God exists, given certain evidence and background knowledge.'' Swinburne notes that in using probabilistic formalism he does not mean to suggest that we can necessarily assign specific numbers to any particular probability. It is often enough to be able to compare probabilities without assigning specific numbers. That is fine as far as it goes.
But part of applying Bayes' Theorem to a statement such as the one above involves assigning some prior probability to God's existence. I am not sure what basis we have for making such an assignment, but I would note that it is not even clear that God as Swinburne describes Him is even within the realm of possibility. That is, a notion like “Mind without body,” might well be, so far as we know, a contradiction in terms. We certainly have no experience with such a thing. Every mind with which we are familiar is embodied. Moreover, God is said to be able to alter natural laws with an act of His will, and to have perfect foreknowledge of all that is to occur. Can such things be? No intelligence with which we are familiar can do anything remotely like what Swinburne describes.
For these reasons, and others I could mention, the God hypothesis should be given such a low prior probability that truly extraordinary evidence is needed to render it plausible. When you then factor in the evidence against God (as Swinburne defines Him) from the problems of evil and divine hiddeness, I'd say Swinburne as his evidential work cut out for him.
A theist might reply that if I am comparing theism to the atheistic alternative then it is not enough just to say theism has a low prior probability. After all, perhaps atheism ought to be given an even lower prior probability. That is, rival explanations must compete with one another. That one might seem to have a low probability is not a reason for rejecting it if all the alternatives have even lower probabilities.
It is a fair point, but I would note that non-God based understandings of the universe have been adequate for explaining virtually everything in our daily lives. The protestations of theists notwithstanding, there just does not seem to be anything in our regular experience that cries out for an explanation as extraordinary as the one Swinburne advocates. Even the fine-tuning of the fundamental constants now seems like a complete non-issue in light of the many plausible multiverse theories coming from physics. You might argue that the multiverse is speculative, and indeed it is. But it has a solid grounding in physics and does not ask me to hypothesize into existence anything that is utterly contrary to all experience. It asks only that I look around me and suppose that there is far more of the same.
Is Swinburne up to the challenge? I doubt it, but let us wait and see...
- Log in to post comments
I've been enjoying the discussion of this over at WEIT.
This is the first time I've seen Swinburne's suppositions, however. Can it be too soon to pass judgment?
At the risk of stating the obvious, his presuppositional bias is right there on display.
If he believes that there is a god, then it is his obligation to approach the problem by attacking the null hypothesis. He should not be piling improbability onto improbability to prove god, he should be disproving the hypothesis, "no god exists."
Second, and perhaps more importantly, he goes on apparently at length about the attributes of this god thingy. How the hell does he know this? Under what authority can he claim that god has any of those attributes? By virtue of what set of experimental data? Seems to me that he's using source material that is in dispute with regard to authenticity and authority (ie, a specific "holy" book without regard to the hundreds of other claimants to that appellation.)
Seems to me that he's already stepped deep in the mire and hasn't even gotten past page 8.
Unless he takes a serious backwards step in future chapters to describe how he comes by the certain knowledge of god's attributes, I think we can declare him to be no better than someone claiming the existence of the Icelandic huldenfolke.
This is sophisticated? Really?
I often play a little game with books of this nature. I will pick it up and start reading, stopping when I reach an obvious suppositional bias, factual error, or logical fault.
Swinburne: Page 8. Really. Francis Collins did much better, and he's a matchbook theologian.
I would be interested in seeing why Swinburne considers the ontological argument a strong C-inductive argument. All versions of the ontological argument I'm familiar with are pure deductive, with the conclusion necessary rather than probabilistic.
DiPietro's comments also apply to most cosmological arguments. There are probabilistic arguments for a god; the fine-tuning argument is the first that comes to mind. But the classical ones proceed deductively, from what the proponents view as solid metaphysical axioms, and what the skeptics view as philosophical fantasy.
Importantly, Pascal's wager is not a probabilistic argument for god, but a pragmatic argument for belief.
Tyler:
Ah, but the ontological argument is only probably logically sound, dontcha know. ;)
I certainly don't see it as legitimate to use purely deductive arguments as though they were evidence, even if you only regard them as having small effects on the probability.
Anyone read Swinburne's writings on substance dualism? Are his arguments for that as bad as they seem to be to me?
Actually, if you'd care to discuss it in more detail, I'd love to know whether Swinburne's terminology is sufficiently well-defined to avoid paradox. For instance, I've seen him define "omniscient" more or less as "knows the truth value of every proposition," which obviously runs afoul of propositions like "God knows that this sentence is false." Has he since patched that one up?
Hey, could be. It seems pretty clear that nobody knows what the prior probability for each hypothesis ought to be, nor even if they should have numerical prior probabilities at all. (What if the subset of possible worlds with a God is non-measurable?) Which makes Bayesian arguments on the subject more or less doomed from the start.
Jonathan West--not to be confused with John West!--has a nice series of blog posts on Swinburne. Personally, I read a couple of his papers and then concluded that his arguments were too badly flawed to even be very educational. I found Plantinga more worthwhile reading than Swinburne, and that's saying something.
'He argues that human epistemic capabilities are so limited, especially when compared to God's infinite wisdom, that we should not expect to comprehend why God would allow intense pain and suffering. The question is simply too complex for our limited understanding, in rather the same way that quantum physics is too complex for a dog.'
I think we should hit Alston over the head with a piece of 2 by 4 theological debating wood.
I have absolutely no idea why that would be a good thing to do. How could a puny earthling understand such things?
But Alston's god would allow it, so Alston would have to agree that it is a really good idea that he be hit over the head repeatedly while reading out his book on how we can't understand why his alleged god allows suffering.
There must be a good reason why god allows suffering and who knows, perhaps Alston will work it out before he slides into unconsciousness. Even if he doesn't , that hardly refutes the fact that there really is a good reason why Alston should be bludgeoned senseless.
Kevin: "Second, and perhaps more importantly, he goes on apparently at length about the attributes of this god thingy. How the hell does he know this? Under what authority can he claim that god has any of those attributes?"
I don't think this is a valid argument against Swinburne. He's proposing a hypothesis for consideration, and he can make his proposed hypothesis as specific as he likes. Being more specific makes the hypothesis more subject to disconfirmation. Each additional attribute he adds to God means one more respect in which the hypothesis could be wrong. It therefore reduces the probability of its being true.
A better objection would be to question whether these attributes are even coherent. Anton has already mentioned the possible incoherence of omniscience. I would say that "perfectly good" is an incoherent concept.
More of a problem still is that, to apply Bayes's Theorem, Swinburne needs to sum over an exhaustive set of hypotheses. I can't imagine how he will be able to do that. I suspect he's going to consider just his own hypothesis (call it G) and its complement (~G). But note that ~G isn't just atheism. It includes all the possible alternatives to G, including those involving gods that lack one or more of Swinburne's chosen attributes, e.g. gods that aren't "perfectly good".
From what little I've read of Swinburne's argument in the past, it's just an exercise in making up numbers to get the result he wants.
P.S. I should have said that adding more attributes to God reduces the probability of the hypothesis unless the evidence supports those attributes. (If adding an attribute increases the probability of the evidence given the hypothesis, this could offset the lowering of the prior probability.)
Russell
A belief in who or what? There are a lot of deities out there to choose from.
How is it "marvelous work" to apply a mathematical method invented in 1764 to some probabilities the author admits are purely speculative? In 1765 it might have been marvelous. Today...not so much.
What would have been theologically marvelous is if theologians had made some progress - heck, any progress - in the last 245 years in validating their speculative premises. You know, while chemistry and physics took us from horse-drawn carriages to space flight, and medicine took us from leeches and curses to genetic manipulation and antibiotics. THAT is marvelous progress. Ending up 245 years later with basically the same valid yet indeterminantly sound arguments you started with...again, not so much.
This is a good place to start - a summary of a dissertation on Swinburne's probability arguments: http://philosophicaldisquisitions.blogspot.com/2010/07/gwiazda-on-swinb…
It really weirds me out how a lot of otherwise intelligent people seem to accept the existence of spirits and souls and other bodiless intelligences as a given. I mean, did I miss that lecture in Physics 101 or something? Where exactly do they get the idea that this is true, or worth accepting a priori?
I mean heck, when I was commenting on Peter Doumit's much-lacking article on BioLogos, the only piece of evidence he could propose for bodiless intelligences were out of body experiences. And this is a man with a masters in Earth Science. What gives?
Richard: I understand your position. However, I think it's worse than that.
Swinburne assigns his deity attributes that are contradictory and anti-logical. Hey fine, it's his deity.
My problem is, why assume that a deity has THOSE attributes? Under what set of conditions have those assumptions been tested?
Even in the holy book of his choice, the Old Testament deity is clearly not omniscient. Right there in the second chapter of Genesis, he goes hunting for Adam and Eve, is incredulous that they have clothed themselves, and seems completely oblivious to the fact that they ate the forbidden fruit. Even if you take the story as metaphor, under what metaphorical sense can this be anything other than a display of non-omniscience?
Same with omnipotence. Iron chariots and all that. Heck, even the Adam and Eve fable carries with it a clear implication that there are limits to his power -- after all, couldn't he "take back" the knowledge of good and evil? Why not? Surely, for an omnipotent being, this would be trivial. But no -- once eaten, the "fall" has occurred and god is powerless to do anything other than just punish the couple and all of their offspring for eternity. And in order to atone for this "sin" has no other option than to send himself down to earth in human form for a bloody, showy human sacrifice? That's a fairly hefty limitation of a deity's power, if you ask me.
So, Swinburne's assumptions about the attributes of his deity are untested, and unfounded; they do not meet the clearly stated attributes of the deity he seems to be trying to defend.
It's presuppositional bias on top of presuppositional bias.
And yet, according to his numerology (and it can't be seen as anything other than that at this point), there's a 93% likelihood that his god exists? The deity he made up out of attributes that don't exist in the holy book where the set of attributes of his deity are defined?
Call me unsophisticated, but this is a mess.
I don't have enough expertise in Bayes to see if this is accounted for in the mathematics when done by true experts in applying the theorem. If it's not, then I see no way around calling "shenanigans".
BTW: He also seems to have assigned a 50-50 probability to the existence of god. In other words, it's Pascal's wager all other again and we all know how fallacious that position is.
It may be too soon to judge completely, and he may redeem himself in later chapters, but this is so unpromising, even absent the critiques I've seen of his misuse of Bayes, that I'm not sure it's worth the effort to slog through this.
Better to move on to Pantinga, who is praised on all sides for coherence if nothing else.
Even the fine-tuning of the fundamental constants now seems like a complete non-issue in light of the many plausible multiverse theories coming from physics.
I don't wish to ignore the quite reasonable possibility that singular explanations may be found for various fine tunings, rather than a multiverse/landscape/weak-anthropomorphic one. I note for example that one of the fine tuning "mysteries," why there was more matter than antimatter in the early universe, is now beginning to yield to investigation on both theoretical and experimental levels.
Is there anything in that "Bayesian" line of reasoning that computes a favorable probability for a two-god situation? I've always preferred a Nick Nolte/Eddie Murphy type of partnership to govern the cosmos.
Pascal Boyer, Todd Tremlin and Steadman and Palmer (among others, I'm sure) quite satisfactorily expose Swinburne's "there exists a person without a body, etc." It's all in his head!
Very nice post. I look forward to more in this series.
I'm having red flags pop up from the start. "Analytical philosophy of religion." The analytical philosophy of religion is theology. So we achieve a greater analysis of god by setting aside reason and using faith to justify faith? Evangelicals have a better insight into--and make a better argument about--the divine than scholars? I wonder if this "philosopher" even knows what s/he is talking about.
"Modern" means instead of tap-dancing we use hip-hop?
First, as someone inclined to defend Swinburne, I think this review has gotten off on the right foot. At the very least, it's good to have an idea about where Swinburne is coming from, so one can see whether he succeeds by his own lights. We can also ask, then, if his starting points could be reasonably accepted. I think one virtue of Swinburne's work is that he constructs his arguments in such a way that each step is not totally ridiculous, or can at least be seductive to a reasonable person.
Second and lastly, I feel obliged to point out that Swinburne has another volume, The Coherence of Theism, where he addresses more fully some of the issues raised here regarding the concept of God. In fact, there is still another volume called The Christian God. I do not think it is incumbent upon anyone to read these volumes alongside The Existence of God (btw, I assume you have the much more recent second edition), but it's important not to have significant expectations that these issues will be dealt with satisfactorily in this work.
One feature of Swinburne's work, to his intellectual credit but to the detriment of full engagement by anyone else, is that it is painstakingly systematic. His corpus, including both the general works in philosophy and the specific works on Christian doctrines form a cohesive whole. Because of this, there are all sorts of discussions in The Existence of God which readers will find unsatisfying but are developed elsewhere - Swinburne's mind/body dualism, his argument from miracles, his views on simplicity and probability, his response to the problem of evil, and so on. On these grounds I would question Schellenberg's recommendation. However, as one can't sensibly recommend Swinburne's whole corpus, and as I can't think of any good single-volume alternative, maybe it is still the best option.
If, as I recall, Swinburne both gives a "% likely" output, and a "we're introducing probabilities for comparative purposes", I call shenanigans on that. It clearly cannot be both.
Anton Mates: Patrick Grim has also argued that there is no "all true propositions" even without egocentric (reflexive, indexical, etc.) ones, at least if you allow formal and factual propositions to both count. (Basically: use cantor's theorem.)
Jud: I don't wish to ignore the quite reasonable possibility that singular explanations may be found for various fine tunings, rather than a multiverse/landscape/weak-anthropomorphic one.
I agree. There's been a 200+ year trend in science of discovering phenomena we thought were independent really aren't. The sets of "fundamental" rules and constants has grown steadily smaller. It would be historically naive to think that this process has ended and that our current set of constants are it. This impacts the fine tuning argument because every time we find a constant or law that is dependent on another, we reduce the number of dials that could be wiggled - the probability of the universe looking the way it looks goes up.
But, pace Kevin, whatever problems the intellectual might have with the idea of a supposedly omniscient etc deity who seems oblivious to what's been going on in his garden, as Boyer et al have shown, the ordinary un-intellectual believer will if pressed happily mouth theological niceties about divine attributes that he or she has picked up while showing (by means of some ingenious experiments) that in fact that is not at all how he or she conceives of the deity: rather, the deity is thought of in largely human terms, as a person who, though supernatural, can only concentrate on one thing at a time (eg he has to answer that prayer there before getting round to saving this person here from perishing in a tsunami or whatever... And I would suggest that this same contradiction between what is held in the head and what is held in the heart (in Pascal's sense)in fact exists in the mind of someone like Swinburne, who after ascribing that impressive list of attributes to God (omnipotent, omniscient, perfectly good, etc) goes ahead and speaks of God as being concerned, as a person might be, for us and our morals and therefore so arranging the world and presenting his existence in so ambiguous and riddling a way that suffering - in the form of, say, a tsunami, or an atomic bombing, or cancer - is in itself a stimulus to disinterested moral action on the part of those who do not suffer (by 'disinterested moral action' I mean moral action that is not undertaken out of a - cynical - desire for a heavenly reward but for its own sake, out of pure compassion); I'm not going to get into what I think about Swinburne's way of justifying the ways of God to man here, but how God's impassibility may be reconciled with a curious and 'personal' concern for his creatures (even a concern that plays itself out in such disgusting ways) is surely a question that a theologian or apologetic philosopher should be aware of and address.
Neither creationism nor intelligent design belong in the science classrooms of our children. Please see...
http://mainereason.blogspot.com/2010/09/test.html
Tim: I think you're spot-on.
Ask someone what the attributes of their god are, and you'll get the usual set. Ask them why their god doesn't behave as if it has those attributes, and you'll get a whole laundry list of reasons that have nothing to do with the previously mentioned attributes of their god. So, one of the possible answers of a prayer to an omniscient, omnipotent, omnibenevolent god is "no" (the most common, I would hypothesize). Why? It's a mystery, or it's The Plan(TM), or insufficient faith.
In addition, I have to go back to the attributes themselves. Where did we decide that Yahweh/Allah had these attributes? I've read the bible cover-to-cover and I have great difficulty reconciling the attributes assigned to it and the actual attributes as demonstrated.
Seems to me that someone else (or a collection of someones ala the Council of Nicea) decided that god *had* to have those attributes in order to *be* god, and then defended those attributes against the clear evidence of the holy book.
To use the weakest of analogies, it's like watching a child come up with excuses for why Santa is real and the fable is true, even though you live in an apartment with no chimney.
Fascinating, to say the least. Infuriating when one encounters it in real life.
@ Joshua:
I'm wondering if you would be willing to spend some time summarizing Swinburne's message and why you think it deserves your support.
You state that you are willing to defend him, and state the reasons. But you never actually defend any of his positions, or clearly state what they are.
Clearly, you're at odds with most of those who I've read who have summarized his work. If those who criticize him can provide a synopsis of his work and where they think it's gone off the rails, why can't a supporter provide a synopsis or a counter-argument as to what he's gotten right?
Or at at least where the critic has gotten Swinburne wrong?
All of the defenses of Swinburne (and not just, but let's keep the discussion narrow) conveniently avoid stating any positive assertion about an argument that the defender finds compelling.
If a "Swinburnian" would please be so kind as to jump in and tell us this crucial bit of information, I think many of us would be less inclined to reply "why bother?".
@ Kevin:
I take it that Dr. Rosenhouse is planning to (re)read Swinburne's book, The Existence of God. My comment was just meant to clarify that Swinburne doesn't substantially develop certain positions which are the topics of entire volumes of his work. For myself, I haven't even read The Coherence of Theism, but that's where he covers some topics addressed in this post and thread.
If Dr. Rosenhouse for some mysterious reason wants someone other than Swinburne to defend Swinburne, he can certainly find people more qualified than me to do so. Alternatively he (or anyone else in this thread) can email Swinburne himself, through the mighty powers of the Internet.
So I think your comment is a bit off-base. I'm not sure I have any particular stake in Swinburne's views, much less views criticized in this post. The content of my original comment is pretty clear.
So if god's wisdom is infinite and we should not expect to comprehend god, why is anybody writing about god?
I owe you for pointing me to that: Grim's "Logic and Limits of Knowledge and Truth" is probably the most thorough and elegant discussion of this issue I've ever read.
Plantinga has an exchange with him on the Cantorian argument here, and does not (I think) acquit himself well. Basically Plantinga seems to admit the soundness of the argument, but says that his intuitions about "all true propositions" are so strong that he's willing to sacrifice the premises of almost any such such argument rather than accept the conclusion. Which kind of leaves him saying that he doesn't know what "omniscience" actually means, he's just confident that it does mean something and he believes in it, whatever it is.
drdave: So if god's wisdom is infinite and we should not expect to comprehend god, why is anybody writing about god?
You're making the mistake of thinking its a premise believers take seriously. They don't - if they seriously thought they didn't know god's will, there could be no religious doctrine.
The incomprehensibility claim is more of a rhetorical device meant to downplay personal responsibility for a claim. The speaker is about to tell you (or just told you) exactly what they think god's will is. But since that's both arrogant and unsubstantiated, they're going to add on a weasel phrase to change the implication from "I know what god's plan is" to the much blander "I'm giving you some unnamed person's opinion of what god's plan is."
Journalists do the same thing. That whole "...but some X believe..." thing is a way of introducing an opinion they want to present without taking responsibility for it. Same structure - 'we can't know gods will, but...' is just another way of doing it.
@ Joshua...
Critics have had no trouble providing their synopsis of Swinburne. They're champing at the bit to tell us why we should discount his philosophical positions, where he errs in logic, in mathematics, and all the rest.
Yet, when given an open door in an open forum and a direct question, "what arguments of Swinburne do YOU think are most compelling" you demur.
Why? Don't you understand them? Or can't you come up with a synopsis that doesn't, in retrospect, sound as lame as the synopsis already provided by critics?
You said "I'm inclined to defend" Swinburne, and then decline to do so.
I'm at a loss to understand why, unless you can't find the wheat in all that chaff.
Have you considered that it might all be just chaff? At this point, that's the impression I'm left with.
@ eric...
Well done. I hereby award you two internets for today's contribution.
@Kevin:
I think I don't understand what is going on here. Hopefully Dr. Rosenhouse was able to comprehend my first, quite simple comment.
You seem caught up in my "inclined to defend" clause. That was just to say: While I find myself on Swinburne's side on several issues, I nevertheless think that this review (by Rosenhouse) was interesting and fair-minded. If you are still confused, think of it this way: I was complimenting Rosenhouse's first installment, not challenging someone named Kevin's negative view of Swinburne.
I also pointed out a feature of Swinburne's work (its multi-volume and systematic nature) that hinders the ability of readers to get an adequate treatment of certain specialized issues in one work. I hope that note is helpful, at least for Rosenhouse, in understanding why some issues might not be fully developed in The Existence of God. I can't imagine any grounds for disputing this. Also, I don't think it is incumbent upon Rosenhouse or anyone else to read Swinburne's (or anyone else's) whole work. But it may be misguided to suggest that Swinburne's has left, say, the nature of God undefended, when he has produced works on this topic in isolation.
I don't understand why you (or commenters at Coyne's blog) think "Swinburnians" should explain or defend Swinburne to you in comment threads. First, note that Swinburne defends himself in books and articles, and in responses in journals, if you have access. If that doesn't satisfy you and you think further challenges go un-addressed, I happen to know from experience that he answers emails. To make this easy:
http://users.ox.ac.uk/~orie0087/framesetcontacts.shtml
If someone thinks Swinburne has misunderstood the application of Bayesian calculus, for example, send him an email and explain the issues to him.
If Swinburne's responses to your favorite challenges are still not satisfying, I don't see a problem. Many people are unconvinced by Swinburne, including otherwise allied thinkers, like Plantinga. Why Internet Commenter's opinion should be relevant remains mysterious to me.
Maybe because this is the internet? And you're a commenter?
Seriously, you just told me to shut the fuck up for asking you a serious question -- to whit: What part of Swinburne's religious philosophy do YOU find compelling? Not what part Swinburne finds compelling. Not whether or not he's gone on at great lengths to "defend himself".
But in an internet forum that YOU joined voluntarily, where you declared by my reading that Swinburne had some ideas worth considering, and where I asked you very politely, "which ones might those be, because all I see so far are the negatives," you tell me that I'm being a dick for asking an unfair question. And then declare that I don't matter since I'm asking the question on an internet forum where I was seeking someone's opinions on the MERITS of Swinburne.
Color me gobsmacked.
Kevin, as another Kevin I have to give you an internet pat on the head for not cursing this idiot out. He made a comment that there was much much more to see in further developments and that our good JR should not read TOO MUCH into the current work...
and you were like "Really? Do pray tell..." And he was like "Geez your an ass..." and you were like "Pardon?" and this Blanchard was all "Look you stupid twit if you have questions ask someone who knows something about them becuase I CERTAINLY DON'T!!"
so I'll say it for you... "Fuck You Blanchy and the ass you rode in on.."
@ the other Kevin...
It took every bit of restraint I could muster, I'll tell you that.
Actually, I don't think Joshua was saying that you don't matter; I think he was saying that he doesn't matter. He's calling himself a random Internet Commenter, whose opinion on Swinburne's arguments therefore shouldn't be very relevant to your judgment of Swinburne.
Not that he wasn't snarky elsewhere in his post, but that particular line doesn't read (to me) as an insult. Just self-deprecation.
Well, if someone has that low of an opinion of themselves, then I rescind my previous offer.
I now don't care to know what he thinks are the merits of Swinburne. Not that I'm discounting that Swinburne has merits, mind you, just that I cannot possibly be compelled to care any more what HIS opinion is with regard to any potential merits that Swinburne might have.
But it's been like this all over the internets. The critics of Swinburne have been chiming in one atop another. The supporters offer the weakest hand-raising and then quit there.
If someone OTHER than Joshua can please offer me what they think are the merits of Swinburne, I would honestly like to know. Other than the fact that he apparently likes to write in excruciating detail. That's not a merit, in my opinion.
What is the Compelling Argument that Swinburne makes which makes his writings something to be studied. We're told that he should be respected. OK. Why? Other than the fact that it's the way he earns a living and apparently has published a LOT of material on the subject.
Still waiting on an answer to why.
From anyone other than Joshua.
@ the first Kevin:
Yikes. The single clause where I indicated I supported Swinburne had one purpose: just to heighten my compliment of Rosenhouse's post. I have absolutely no interest, need, or desire, to defend Swinburne. In fact, if you read my comment again, I suggested some disadvantages to reading Swinburne - namely, that his multiple volumes constitute a single systematic picture, which lessens the force of any one volume, and who has time for that?
But, if you want to know what Swinburne thinks, then like Rosenhouse, I can't think of a better way of finding out than reading a book by him. If you want to know what he thinks about concerns raised in this particular post, I can't think of a better way of finding out than emailing Swinburne himself.
You write, "The supporters offer the weakest hand-raising and then quit there." If I wanted to waste time defending someone from a blog post, I surely would do so. I don't, so I didn't.
Lastly, Anton is correct that I was generally citing my own unimportance, not insulting you.
Contrary to every single instinct within me, but because you've made me feel bad, I'll say that some of the reasons why Swinburne has respect in phil religion include:
(1) He has done independently respectable work in probability theory, epistemology, and philosophy of science, which inform his religious arguments.
(2) He has a thorough grasp of the Christian theological tradition.
(3) He is reasonably literate (for a philosopher) in the natural sciences.
One of my reasons for appreciating Swinburne is that, unlike Alvin Plantinga, he does not invoke mysterious faculties (e.g. Plantinga's "Sensus Divinitatis.") Swinburne approaches his arguments with the methodological assumption that religious beliefs should be justified by the same (mundane) procedures as any other belief. Another reason I like Swinburne is that his defense of theism is inductive and not deductive. Someone like William Lane Craig will say they can demonstrate the existence of God from indisputable premises. Swinburne's claim is just that Christian theism has (on an internally plausible construal of its content) some explanatory value with respect to a set of features of the world and the nature of persons, plus that there is a world at all. He thinks (1) that this makes Christian theism probable to some degree with respect to the evidence and - more controversially - (2) that Christian theism provides a better explanation than otherwise plausible alternatives.
So those are some "merits" of Swinburne, which might motivate someone to read him. Swinburne's actual arguments, of course, are contained in his books, articles, and brain. His brain can be accessed via email.
I think Joshua has been unfairly treated here. I don't think a commenter is under an obligation to give an argument in support of any position (even his own) unless he has undertaken to do so. Perhaps the whole kerfuffle was a result of Joshua's phrase, "as someone inclined to defend Swinburne". I suspect some people mistakenly took this as an undertaking to defend Swinburne's argument.
Having said that, I hope you, Joshua, will stick around to confirm or correct Jason's interpretation of Swinburne, as far as you can. I have no inclination to read a book-length account of an argument which (from the little I've read about it) seems seriously misguided, and all the more so when you tell me that I need to read additional books in order to fully understand that account. Like others here, I would like to see a brief account which I can consider reasonably reliable, and perhaps the best I'm going to get is Jason's account with possibly some confirmation/correction by you.
In response to your last post, I would say it's certainly a good thing that Swinburne is addressing the question of God's existence by means of an inductive argument. But that's only a good thing by the outdated standards of philosophy of religion, where addressing such a question by purely deductive arguments is still widely considered a reasonable option. To me, your commendation is a bit like commending a chemist for not invoking the theory of phlogiston!
You comment that Swinburne's work on epistemology (etc) is "respectable". But from my point of view the standard for respectability in philosophy is disappointingly low, so that's not much of a recommendation as far as I'm concerned.
I note that your commendation is tentative--you even put "merits" in scare quotes--so I don't intend this as a criticism of your comment, but rather as an explanation of why I (and others) are not much impressed by such "merits".
Richard: Just a quick clarification, the quotes around merits were actually quotes proper, as "merits" was the term used by Kevin. The reason for quoting his term was because I take it that the "merits" are supposed to constitute initial motivations for reading, and not just a repetition of Swinburne's arguments.
@Joshua. Thanks for clarifying that.
Richard Wein: from my point of view the standard for respectability in philosophy is disappointingly low, so that's not much of a recommendation as far as I'm concerned.
I agree with you Richard. As a scientist I put a high value on usefulness; novel research findings or techniques that will help me succeed at future research gets respect. Research results that go nowhere may be intellectually interesting, but at the end of the day I'd rather have a useful approximation than an inapplicable truth.
Was Swinburne the first theologian to apply Bayesian logic to theology? If so, I could see how introducing a novel technique to the field could be exciting and interesting. But if the technique is known and all he did was plug in his own hypothetical values, I'm not sure how that is really impressive. A student could do that. At the very least, what would be more useful would be a sensitivity analysis - i.e. determining how (much) varying each input affects the output value. That would at least provide future philosophers with info useful for deciding what to study.
@Eric:
Swinburne is part of a wider tradition of natural theology, which seeks to see what religious claims might be inferable from accepted facts about the world. Generally, however, natural theology has been deductive. So from facts about contingency or causation, one might reason that there must be an uncaused cause. This sort of argument is unsatisfying for a number of reasons. One reason is that it's hard to identify the end of deductive theological arguments with any specificity. Second, these arguments work on a level of abstraction that makes their radical conclusions seem quite implausible.
Swinburne studied under another philosopher at Oxford, Basil Mitchell, who advocated for "cumulative case" arguments, which have several virtues. First, they closely resemble how we reason generally about things in the real world. Second, they don't work from high levels of abstraction. Third, they are much better at accommodating specific hypotheses. Swinburne is notable for taking natural theology in a particularly empirical direction. It's not quite right to say that he's modeling his approach after natural science per se. But, as I think he says himself, he is modeling his approach after empirical inquiry generally - he compares himself to a set of inquirers: scientist, detective, and historian.
On this Bayesian thing, I think it's important to note that the quantifications of his arguments is a separate matter from the overall project of the work. There are (of course) less mathematical ways of going about inductive inquiry. In fact, very rarely does anyone quantify the probability of inductive conclusions, especially in history or law. But Swinburne fits his arguments into the calculus, usually after the completion of the main text, just in order to illustrate how the structure of the argument works and how (according to him) sometimes even modest probability assignments will render his argument successful.
Perhaps you could expound a bit on what this entails. From some of the things I've seen Swinburne quoted as saying, it seems he believes that while individual lines of evidence for the existence of god remain unconvincing, when looked at in total they can add up to a decent case. We often see this same fallacious thinking from believers in UFOs or ghosts. For instance, no one can point to a single good piece of evidence for UFO visitation, but many believers maintain that the sheer number of reports add up to a convincing case for their existence nonetheless. But a thousand pieces of unreliable or questionable evidence can never add up to one piece of good evidence. It just doesn't work that way. Similarly, if all the individual arguments for god's existence fail, they can't add up to one good justification. But perhaps I'm misinterpreting what's meant by a "cumulative case" for god.
@HH:
That's surely a concern. It is true that cumulative-style cases can rhetorically mask the weakness of each piece of evidence. It's up to the reader to decide if Swinburne is using the strategy in this seedy conspiracy-theorist-like fashion.
On the virtues of the approach, I'll quote Swinburne:
"That arguments may support or weaken each other is even more evident, when we are dealing with inductive arguments. That Smith has blood on his hands hardly makes it probable that Smith murdered Mrs Jones, nor (by itself) does the fact that Smith stood to gain from Mrs Jones's death, nor (by itself) does the fact that Smith was near the scene of the murder at the time of its being committed, but all these phenomena together (perhaps with other phenomena as well) may indeed make the conclusion probably" (Existence of God, 2nd edition, 12).
"The crucial issue ... is whether all the arguments taken together make it probable that God exists, whether the balance of all the relevant evidence favours the claim of theism or not. For clearly, in so far as the probability of a hypothesis is relevant to whether or not we ought to act on it, we ought to act on a hypothesis in so far as it is rendered probable by the total evidence available to us - all we know about the world, not just some limited piece of knowledge. The religious person claims that his religious viewpoint makes sense of the whole of his experience; and his atheistic rival is liable to make a similar claim" (Ibid, 13-14).
In case you haven't fallen asleep yet, I think that's quite reasonable. In fact, I can't think of a better alternative to investigate such a question.
*'probably' in first excerpt should read 'probable'
Using Bayesian reasoning, one not only needs to assign a prior to each hypothesis but also a means of assigning the conditional probability of a hypothesis given the data that one has observed. Without a quantitative method for doing such a thing, the posterior judgments will be very subjective and unreliable. This makes me very suspicious of Swinburne's approach.
In case you haven't fallen asleep yet, I think that's quite reasonable. In fact, I can't think of a better alternative to investigate such a question.
I agree. Nothing beats stacking the ol' probability deck. They've been doing it for millennia. I can't think of a better alternative either. Neither can anybody else. Better than nothing I guess.
PZ is making another mistake here. He is assuming that O'keefe's boat isn't always stocked like this. More than likely, O'keefe is a not unusual GOP operative, someone like Limbaugh who is into some weird BDSM stuff.
And where in the hell did he get a huge boat anyway, so young? I didn't know being a GOP slime mold paid so well.
@Tyler
I think the problem lies more with the priors (P(H)) than with the conditional probabilities (P(E|H)). It's generally not difficult to construct a hypothesis that confers a high probability on the evidence. For example, after a lottery has been won, the hypothesis "the draw was rigged to ensure that particular person won" gives a high P(E|H). But in the absence of further evidence we shouldn't arrive at a high posterior probability (P(H|E)) because our prior knowledge about lotteries leads us to assign a low prior probability to the hypothesis. As Jason noted above, it's the prior probabilities which are the biggest problem for Bayesian reasoning, and I think that's where Swinburne most seriously falls down.
@Joshua
You wrote: "...(according to [Swinburne]) sometimes even modest probability assignments will render his argument successful."
The argument cannot be rendered successful merely by establishing P(E|H), no matter how well that is done. From what I've read, Swinburne doesn't assign a probability to P(H). He makes an argument about the relative size of P(H) and the divisor in Bayes' theorem, based on an argument regarding the simplicity of his hypothesis relative to the alternatives. Atheists tend to argue that God cannot be simple, so this seems to be the crux of the matter.
At least Swinburne doesn't go down the path followed by many apologists, insisting that God is simple by definition. I suppose I can grant Swinburne that he avoids some of the most egregious errors of other apologists, and that might be a good reason for reading his book rather than some other book-length argument for God, if I was inclined to read any such book.
Anyway, this is getting a bit ahead of myself, since I'd like to wait for Jason's account of how Swinburne applies Bayes theorem.
You shouldn't need to assign a low prior probability in order to defeat the Lottery Paradox when using Bayes' theorem.
Here's why: the probability of a particular person winning is dependent only on the lottery space size, whereas the prior, to make any sense, must depend only on the administrative arrangements for securing the draw against tampering. Since these are independent, there must exist possible lotteries where the chance of winning is much less than the prior probability of tampering, but nonetheless we would still conclude that any given win was honest.
This implies that P(E|~H) is not merely equal to the chance of winning the lottery; allowance must be made for the fact that you didn't identify the winner until after the draw. But I'm not convinced there's a sane value you can assign in this case.
Andrew: "the probability of a particular person winning is dependent only on the lottery space size"
If I've understood you correctly, Andrew, I think you're mistaken. You can't calculate the probability of a particular person winning, P(E), without taking into account the priors. You can't just say that P(E) = 1/N, where N is the number of tickets sold, because that's just P(E|"the lottery was fair"), not P(E). P(E) involves a summation over all possible hypotheses.
P(H|E) = P(E|H).P(H) / {P(E|H).P(H) + P(E|~H).P(~H)}
If P(E|H) is close to 1 and P(H) > 1/2, then P(E|H).P(H) > P(E|~H).P(~H), whatever the value of P(E|~H). And then P(H|E) > 1/2. In other words, if the prior probability of H ("rigged in favour of person A, who actually won") is high enough, this hypothesis will have a high posterior probability.
However, unless you had a prior reason to think the lottery was rigged in favour of A in particular, you should give H equal prior probability with the alternative hypotheses that the lottery was rigged in favour of B, C, D, etc. So the prior probability of each of these hypotheses, including P(H), must be low.
In principle it makes no difference to Bayes theorem whether you pick your hypothesis before or after observing the evidence. It's just that picking your hypothesis first helps to keep you honest in your assignment of priors.
I should add that I'm no expert on Bayes theorem. I could be wrong. But this seems to make sense.
Actually, I find Swinburne's arguments for God particulary good compared to most others. I think his notion of C-induction piled upon one another makes the beilif in God P-inductively true very close to most theologians notion of evidence for God (of the Judeo-Christian variety).
Having said that, it is remarkable how weak this argument is. It conceded much to the atheist and really boils down to whether you can accept the "piling on" notion of evidence for. I find this notion weak and rather trivial: if this is all they have, then consider me an atheist about the emperor's clothes: there is no THERE there. But this is to disagree: obviously Swinburne accepts the piling on notion of evidence for. So be it. But woe, have the mighty fallen from Augustine.
Richard,
I think you're understating the problem I pointed out. You may make superficial observations of apparent "design" in nature and thus assign a high P(E|H) to what you observe given the "argument from design", but there are plenty of counterexamples that would make you assign a lower conditional probability, such as a multitude of design imperfections. Without explicit quantitative reasoning, your conditional probabilities, and thus your posterior judgments, are pretty much entirely subjective.
@ Richard:
Hmm. I think you're right. So arguably our prior P(H) is dependent, not on the chances of winning the lottery, but on the number of players; if H is the hypothesis "the lottery was successfully rigged in favour of player A", our estimate of P(H) should be on the lines of P(H0)/N, where H0 is "the lottery was successfully rigged in favour of some player" and N is the number of players. (For simplicity I'll assume an unsuccessfully rigged lottery is still fair.)
Real-world lotteries tend to have fixed odds of winning that don't depend directly on N, but in fact are chosen to keep the expected number of winners within some reasonable bracket. So let's try it with some real-world numbers:
Suppose player A enters 1 line in the UK national lottery and wins or shares in a jackpot. The probability of this is about 1 in 14 million. The number of players is, we'll assume, something in the region of 25 million. (They don't all play 1 line, the average seems to be about 2.)
Let's further assume that P(H0), the probability that the draw was rigged, is something like one in a million.
So P(H), the prior probability of the draw being rigged in A's favour, should be something like 1 in 25x10^12.
P(E|H) is 1; if we base P(E|~H) on the winning odds alone then it is 1 in 14 million (the correction required for the one-in-a-million chance that the draw was rigged in someone else's favour can be neglected).
Plugging in the numbers gives us P(H|E) as 1 in 1.8 million, which is interestingly lower than our assumption about the overall probability of rigging.
@Andrew
On reflection, I think it was remiss of me to simply say that 1/N is just P(E|"the lottery was fair"), not P(E). If you take the priors to be equal for all players, then P(E) will work out to 1/N. Alternatively you could get P(E) = 1/N just by taking the scenario to be symmetrical with respect to all players. But then it follows that the prior probabilities are symmetrical for all players.
Your calculation seems correct to me. One minor proviso is the asymmetricality between players: not all players have the same number of lines. In principle A's having only one line might not be independent of which hypothesis is true, and that would affect the calculation. Perhaps a player is more likely to buy only one line if he knows that the lottery will be rigged in his favour, because one line is all he then needs. But I think that's pretty implausible, and we can ignore it.
Incidentally, I find the following rearrangement of the formula for Bayes' theorem much easier for calculation:
O(H|E) / O(H) = P(E|H) / P(E|~H)
where O(x) is the "betting odds" of x. Conveniently, for small probabilities O(x) ~= P(x).
Joshua Blanchard wrote:
It's seemingly reasonable, but is in fact quite bad thinking. Again, I'm reminded of ghosts hunters who mistakenly group several unrelated phenomena (EMF fluctuations, pareidolia in indistinct audio recordings, photographic artifacts, unintuitive temperature convection, human psychology) into a single phenomenon with a single explanation--ghost hauntings. They often maintain--erroneously--that Occam's razor favors ghosts as being the "simpler" explanation, forgetting that invoking supernatural explanations with unknown qualities multiplies entities beyond necessity. When each ghostly phenomenon is examined individually, more prosaic explanations are invariably uncovered, and what was amassed as collective evidence for the existence of ghosts is found to be unrelated and unpersuasive.
The "cumulative case" for god is fraught with identical errors. There is no reason to believe that the god hypothesis--which consistently fails to be the best explanation for such unrelated phenomena as the origin of the Universe, the diversity of life, the wellspring of human morality and spontaneous remission of cancer, among other things--should be considered the best explanation for them collectively, especially since the majority of the "evidence" relies on logical fallacies like arguments from ignorance. The "cumulative case" method can be accurately summarized as piling up bullshit and hoping to impress by the sheer quantity of it.
This strategy is not limited to fringe conspiracy-theorists. It's just arguing like a lawyer (trying to make a plausible case) and not investigating like a scientist on an honest search for truth. A scientist, because he seeks the best possible explanation, tries to falsify his hypothesis and ensures that it accounts for all of the data. A lawyer seeks only a seemingly plausible explanation and regularly ignores data that undermines his case. His goal to persuade and convince, not investigate honestly.
This is why I find apologists such as Swinburn intellectually dishonest--because they argue like lawyers. They don't seek disconfirming data. They routinely ignore better explanations. They rationalize bad conclusions and prop up failed hypotheses with ad hoc excuses. They're committed to a predetermined conclusion and don't follow the evidence wherever it might lead them. There is something quite sordid and slimy about such a fundamentally dishonest enterprise, and I find it quite difficult to respect those who engage in it.
Joshua Blanchard:
You could use repeatable, empirical observations to determine whether/the extent to which some entity is intervening in the world. Its called science.
However, the results of scientific investigation have been overwhelmingly negative. No such entity has been found. So I think what you really mean here is: if we ignore those methods of investigation that have already yielded negative theological results, this method is the best of what remains.
That may be true, but its also a pretty clear sign of some serious confirmation bias among natural theologians.
@Eric
I don't think Swinburne takes God to be an object of the natural sciences. Rather, he takes Christian theism to be, in part, a system of explanations. If it is a system of explanations, then in a (broadly) empirical inquiry, it can be evaluated accordingly.
It is very true that "those methods of investigation that have already yielded negative theological results," if there are such methods, don't confirm theism. What methods (and results!) do you have in mind?
If it is a system of explanations, then in a (broadly) empirical inquiry, it can be evaluated accordingly.
You can say that again. I'm talking broad. ReaLLY broad. Like Brooklyn Bridge broad. There's an old saying about bridges and people selling bridges. If I knew what the heck you're talking about. Which I don't. It all sounds very "airy faery". But nobody knows for sure. Nobody knows what you're talking about.
@386sx:
Empirical inquiry covers a variety of intellectual activity, not only the natural sciences. The point is just that Swinburne's project is meant to fall under this umbrella, not that it is supposed to mirror the natural sciences. To pick an easy example, when I want to find the best Mexican food in town, I investigate according to broadly empirical principles; yet I'm not doing physics, chemistry, or biology.
Okay. He uses a lot of fancy logic and math to find the best Mexican food. We're already pretty sure it exists. Pretty easy example, thanks. I have a broadly empirical bridge for sale if anybody interested.
To pick an easy example, when I want to find the best Mexican food in town, I investigate according to broadly empirical principles; yet I'm not doing physics, chemistry, or biology.
The broadly empirical thing would be the best way to find the best Mexican food if there isn't any. So I guess I don't blame him. If he weren't broadly empirical then he would fail miserably. I guess broadly empirical is something like taking a poll or like "riffing". Maybe he's "riffing" off of math and science. You can get some good things from "riffing". IIRC Lennon's best song Instant Karma was from "riffing". The best song of all time.
Things you can get from broadly empirical:
Best Mexican food.
Best song of all time.
Best Christian god exists.
eric @ 22:
Not to mention that we don't even know whether or how "dial-wiggling" did or even could have happened. Which makes the "Look at these Big Numbers" part of the cosmological argument kinda silly.
@ 30:
YES. The really annoying part is people's failure to think in terms of probability/certainty distribution, except where it's convenient. So when I say "God probably allowed for that hurricane because of X", why, that's just some reasonable and humble thinking. But when you say "God probably does not exit", I'm allowed to stamp my feet and cry "But howww do you knoowwww?" No matter what, my opponents are speaking from a position of self-assured, absolute, unfounded certainty, whereas I'm "just telling you what I believe."
Joshua @ 64:
No, you're not exactly doing those things, but your empirical findings are "tethered" to the findings in those fields, such that it is conceivably/logically possible for your gastronomical experiences to clash with the science (for example, a salt-free food tasting salty), prompting more revisions. Theology enjoys no such intertwining with cosmology, but merely a sort of humming to its tune.
Empirical inquiry covers a variety of intellectual activity, not only the natural sciences. The point is just that Swinburne's project is meant to fall under this umbrella, not that it is supposed to mirror the natural sciences.
I see. The broadly empirical umbrella is like the "What, me worry" of empirical inquiry.
I've found these two reviews of "The Existence of God":
http://www.infidels.org/library/modern/gabe_czobel/swinburne.html
http://www.essays.se/essay/306f54fea1/
The second of these in particular is very detailed, and I think I now have a fairly good idea of what Swinburne's argument is.
Swinburne makes several arguments for God, his own take on traditional arguments such as the cosmological and teleological arguments. He acknowledges that none of these arguments individually is "P-inductive" (making the God hypothesis, H, more probable than not), but argues that they are each "C-inductive" (making H more probable than it was before the argument), and that they add up to a P-inductive argument when taken together (and when taken with a couple of contrary arguments, like the argument from evil, which count against H). Since he acknowledges that none of the arguments individually is sufficient to establish his conclusion, it makes sense to start by looking at how he combines the results of the arguments. If that's a failure, then his whole argument is a failure, and there's no need to evaluate the individual arguments.
The conclusion of each individual argument is that the evidence presented in that argument makes H more probable, i.e. P(H|Ei) > P(H), or P(H|Ei)/P(H) > 1. (Swinburne conditions every probability on background knowledge, K. But that's taken for granted in Bayesian reasoning, and can be omitted for the sake of brevity. I write "Ei" to clarify that this is just one of several items of evidence.) So you might expect him to attempt to multiply these ratios together over successive arguments to reach a cumulative P(H|E), where E=E1&E2&E3... I think that's the standard Bayesian approach to accumulating evidence. As each additional piece of evidence is considered, the posterior probability from one inference becomes the prior probability for the next. But Swinburne doesn't attempt to do this, probably because he didn't actually estimate any such ratios. This means that the question of whether the individual arguments are good C-inductive arguments plays no part in his overall argument! Instead, the content of those indivual arguments seems to inform his overall argument in a much vaguer way.
Swinburne makes an extremely subjective judgement of the conditional probability of all the evidence combined, P(E|H) = 1/3. This extreme subjectivity itself casts a cloud on his final conclusion. But I'm more concerned with his treatment of priors and of the conditional probabilities for alternative hypotheses, since these are even more problematic in calculations of this sort. He doesn't even put specific figures on these, but relies on much vaguer hand-waving arguments, as described on p. 76 of the second review. Put simply, his overall argument turns on the claim that the God hypothesis is simpler than the alternatives! I think it's incumbent on both writers and those who recommend them to clearly identify such a crucial and controversial claim from the outset. Then if I consider the claim absurd, I know I need not read the book, or I can go straight to the section of the book which purports to support the claim. I've never yet come across a supporter of Swinburne who has identified this central claim, and this is probably the fault of Swinburne himself for failing to make it clear.
As far as I could tell from these reviews, Swinburne's case for the simplicity of God rests mainly on the idea that hypotheses involving infinities are simpler than others! But even if we accepted that a hypothesis which sets a parameter at infinity is simpler than an otherwise identical hypothesis with that parameter set to a finite value, that would at best make an infinite god simpler than a finite god, other things being equal. It would do nothing to establish the simplicity of the God hypothesis relative to quite different sorts of hypothesis. That simple observation demolishes Swinburne's argument, as far as I've seen it. Of course, there could be something significant that I haven't seen. But establishing the simplicity of his hypothesis over the alternatives is crucial to Swinburne's argument, and anyone wanting to defend Swinburne as having a decent argument needs to adress that point in particular.
The final stage of Swinburne's argument involves an astonishing bit of mathematical legerdemain. He excludes one of his pieces of evidence (religious experience) from his overall calculation and then uses it to modify the result of that calculation afterwards. This is convenient for his conclusion, because the simplistic nature of his overall calculation means it would be difficult for him to come up with any specific figure other than 1/2, whether he includes the evidence of religious experience or not. But he wants his hypothesis to have a probability > 1/2. Saving this piece of evidence until later gives him the chance to argue that the probability is increased above 1/2 once you add the extra evidence. I don't mean to accuse Swinburne of deliberate deceipt, but this error indicates great mathematical naivety. He seems unfamiliar with the concept of "spurious accuracy".
Overall, Swinburne's book seems to be a collection of old arguments for God, with all their well-known problems, plus a few new wrinkles that fail to address these problems, adding a number of new problems, and dressing up the whole in gratuitous probabilistic terminology that obfuscates the real issues.
Joshua @64: Empirical inquiry covers a variety of intellectual activity, not only the natural sciences. The point is just that Swinburne's project is meant to fall under this umbrella, not that it is supposed to mirror the natural sciences.
Chemists don't ignore data coming out of physics experiments. Biologists don't ignore data collected by geologists. If natural theology is claiming to be empirical, it can't ignore the wealth of empirical data already collected by other disciplines. Empirical data does not stay within a discipline: every other discipline that also claims to be empirical is going to have to deal with it one way or another.
If natural theology is a non-chemistry, non-physics empirical method, it can't ignore the data collected via chemistry and physics. And that data says we have run countless controlled experiments and in no case has anyone ever found any evidence for the force/cause/entity Swinburne claim exists.
So, you haven't changed my mind. I think its a pretty strong form of confirmation bias going on here, when Swinburne claims natural theology is empirical but then ignores all the data collected by other empirical disciplines that doesn't agree with his preferred conclusion.
And so far as I can tell--also from these reviews, so someone correct me if I'm wrong--Swinburne actually thinks it makes sense to set a parameter "at infinity," full stop. But of course that's mathematically absurd; "infinity" is not a number. There are lots of infinite ordinals and cardinals, and lots of infinite limits of various functions/sets/sequences, and you could set a parameter to correspond to one of those--but Swinburne doesn't seem to do that work, and even if he did that would make hash of the claim that infinite values are "simpler" than finite ones.
I see that Quentin Smith touches on this issue in a review of Swinburne.
I also note that, in "The Beginning of the Universe and of Time," Swinburne directly contradicts this infinite-is-simpler position. He argues that if current physics tells us nothing about the universe before the Big Bang, then we ought to assume that the universe began at that point instead of extending infinitely into the past:
"But I suggest that in the absence of positive evidence, it is better justified to postulate nothing than something; and if so, the best justified conclusion would be that the Universe began a finite time ago."
But by that same argument, it is unjustified to postulate that God has abilities or knowledge for which we have no positive evidence, and the best justified conclusion would be that God's power and wisdom are finite!
@Anton. Good point about infinite values. I agree with Quentin Smith that Swinburne is conflating different senses of "infinite".
This is the best they've got? Honestly?
So much here that's just massively off base, not even bothering to mention the egregious misuse of pseudo-statistics. My favorite, though, has to be Swinburne's easy dismissal of the problem of evil. Seriously, where does one even begin? First, you have to assume that this god creature engages in some sort of twisted caricature of utilitarianism, maximizing our opportunities for moral advancement (whatever that actually might entail) by thoughtfully providing so many of us with suffering. This is apparently a simple calculus, with a definably "best" fit solution (simple for god, anyway) so I suppose at least we can assume we aren't stuck at some local maximum in the suffering/moral self-improvement landscape. It sounds monstrous, but I suppose you can't argue against the truth of something from its consequences. So maybe it's... possible. Maybe. For the sake of argument.
But then, one might have to inquire a bit further about the optimal *distribution* of that suffering. Presumably, in order for suffering to provide us opportunity to increase our "perfect being" karma, or whatever he imagines this sort of benefit to actually fucking be, we must actually have the capacity to do something about that suffering. Or, if merely feeling kinda bad about it is itself a moral good, then we must at least assume that our opportunity to be aware of that suffering must likewise be maximized. In fact, I suppose *both* would have to be maximized, in this best of all possible worlds.
Now, a serious person might wonder whether there is any evidence *whatsoever* that this is the condition in which we actually find ourselves? But let's just say we don't bother to wonder about that. The fact is, if Swinburne's god is real, then this must all be true in our world. Swinburne's god's universe must contain suffering to be the best possible universe, our universe contains suffering, therefore the balance of that suffering must be exactly that balance necessary to provide evidence for Swinburne's god. In much the same way, a perfect world must contain some heat sources, our world does indeed contain heat sources, so clearly the number, distribution, and magnitude of heat sources in our world much be exactly equal to the necessarily optimal balance of heat sources required to provide evidence for the existance of Swinburne's god. Substitute any other possible quantity you like.
So I'll wonder about something else instead.
What about the distribution *between people*? Now, maybe "perfectly just/fair/whatever" isn't one of the conditions Swinburne assigns to his imagined deity. But I might be forgiven for thinking that it sorta might be considered a subset of his general omnibenevolence, i.e. "perfectly good," criterion. Maybe, maybe not. But assuming it is, then how is it that there is such an uneven distribution between the necessarily inflicted suffering and generously granted opportunity to feel kinda bad about it? People in the midst of massive suffering are, almost without exception, awfully damned self involved. Not to mention relatively poor in time, energy, and other resources to apply to karma enhancement. They have a significantly diminished capacity to feel kinda bad about the suffering of others, not to mention do anything about it. Maybe, of course, his god simply assigns karma based on the relative employment of one's actual opportunity for moral exercise. Perhaps, if your moral capacity only has a 10% unemployment rate, you are doing fairly well, regardless of your absolute capacity for good. With that in mind, perhaps someone with massive, uncontrollable murderous impulses is actually being very, very good each time they fail to kill the people they come across every day, while your average saint is heavily deducted each time he or she fails to say "god bless you" when someone in earshot sneezes. Of course, one might wonder whether, if that were the case, it might be even further maximized if were instead a world where we were *all* moral cripples, for whom maximizing our moral capacity would require hardly any resources at all.
This is all very convincing.
Shorter Swinburne/Liebnitz (minus the pretend statistics):
1)God would only create the best of all possible worlds, therefore this is the best of all possible worlds, therefore god.
2) Simple is more probable, and I arbitrarily assert that god is simple, therefore god.
Those are the only two actual arguments I've seen so far. Maybe I need to read the book myself to find out where there any more of these precious gems of enlightenment.