It wasn’t clever when Gödel did it either
Stranger Fruit is a blog on ScienceBlogs I don’t usually read, but this particular post was brought to my attention just now.
Today in 1978, the logician Kurt Gödel died in Princeton, New Jersey. Gödel, of course, is remembered for his incompleteness theorems but also took the ontological proof for the existence of God serious enough to express his own version of it in modal logic.
Which he then follows with the proof itself. You can look at it if you like, but it’s essentially Anselm of Canterbury’s with more maths to confuse the layperson.
The general form of the ontological argument (and a simplification of Anselm’s version), of course, is this:
1. God is, by definition, a being than which nothing greater can be conceived/imagined.
2. Existence in reality is greater than existence in the mind.
3. God must exist in reality, if God did not then God would not be that which nothing greater can be conceived/imagined.
Analogously, it is possible to conceive of a perfect ham sandwich. A ham sandwich which, by definition, is better than any other such sandwich. For such a ham sandwich to exist in reality is greater than for it to simply exist in the mind. Therefore somewhere must exist the perfect ham sandwich.
I’m expecting that sandwich to arrive any moment now.
Anyway, Lynch then goes on to say that because Dawkins refuted the Anselm of Canterbury version but not the Gödel version, clearly The God Delusion is an infantile piece of crap.
John Lynch is apparently an atheist, and he’s bringing this up because he feels the chapter in TGD dealing with the ontological argument is ignorant of sophisticated theology (never mind that all of the various version of it have at its base the same single flaw), and that this weakens the book as a whole.
It’s a textbook example of the Courtier’s Reply, and he reinforces it in the comments.
“I freely admit to not understanding Godel’s argument - I was merely using it to make a broader point about Dawkins’ willingness to caricature the ontological argument.”
“Thanks for playing. Come back when you have actually read something substantial on the ontological argument. And that excludes Dawkins.”
“I’ll stand by my claim that Chapter 3 greatly weakens the overall work and that it does not reflect anything interesting that’s happening in philosophy of religion.”
The ontological argument doesn’t need caricaturing, and Dawkins doesn’t. Essentially saying “other people have written more! I don’t understand it but they can’t all be wrong!” is intellectually void. “Philosophy of religion” is a euphemism for “theology”, and arguably nothing interesting has happened in that area in many, many centuries.
I hope for the sake of ScienceBlogs in general that this is an attempt at satire, but I doubt it. We (and secularism everywhere) have about as much to fear from appeasers as we do from fundamentalists themselves.
Edit: Larry Moran takes apart another one of Lynch’s posts. This is relevant.
clem said,
January 18th, 2007 at 8:51 pm
As a vegan (and I am sure vegetarians, jews, muslims and many hindus and buddhists will agree) I can assure you that the perfect ham sandwich than which no greater can be conceived is one that does not exist. Therefore, by the Anselm argument, that perfect ham sandwich must exist. But when I order a nenexistent ham sandwich in restaurants, I get funny looks for some reason. Or does that just mean that the ham of that perfect sandwich is still attached to the pig?
Cairnarvon said,
January 18th, 2007 at 8:58 pm
Ask your pastor~
J. J. Ramsey said,
January 20th, 2007 at 5:28 am
Cairnarvon: “Philosophy of religion” is a euphemism for “theology”
That statement reveals just how badly you’ve read Lynch’s posts. Do you really think that Hume and Kant are theologians?
J. J. Ramsey said,
January 20th, 2007 at 3:21 pm
I’ll probably regret trying to respond to a strawman with facts, but here goes.
Cairnarvon: “never mind that all of the various version of it have at its base the same single flaw”
Reading the IEP entry on the ontological argument shows this to be false. For example, the second version of Anselm’s argument is not refutable by Kant’s “existence is not a predicate” argument, though as the IEP shows, it has other flaws. BTW, the IEP entry also shows that your “ham sandwich” argument, which is a trivial variant of Gaunilo’s argument, doesn’t work.
Cairnarvon: “Lynch then goes on to say that because Dawkins refuted the Anselm of Canterbury version but not the Gödel version, clearly The God Delusion is an infantile piece of crap.”
Actually, Lynch doesn’t think that Dawkins did a good job with Anselm’s (first) ontological argument. Dawkins snarks, and he rambles about Bertrand Russell’s take on the argument. He quotes a parody of the argument that contains a flaw not present in the original, which makes it useless as a refutation. The only substantive thing that Dawkins wrote on it was a quote from Norman Malcolm, “but what could it mean to say that it will be a better house if it exists than if it does not?”, which as a standalone statement is vulnerable to the obvious reply, “A house that exists at least stands a chance of keeping rain off of one’s head, unlike a house that doesn’t.” Now the fuller quote from Malcolm, which adds “One might say, with some intelligibility, that it would be better (for oneself or for mankind) if God exists than if He does not-but that is a different matter,” would be a partial counter to the obvious reply. However, if Dawkins is going to attempt to refute the ontological argument, it is his responsibility to anticipate the obvious reply. He can’t expect a possibly hostile reader to make his argument for him.
Sigh. You may build another strawman now.
Cairnarvon said,
January 20th, 2007 at 4:24 pm
My ham sandwich analogy stands. It’s precisely as incoherent as the God version. Saying that it’s possible to conceive of an objectively perfect being is exactly as nonsensical as saying that it’s possible for a ham sandwich or piland.
Anselm’s second version is as ridiculous as the first, for the same reasons.
Rereading the chapter in question, Dawkins doesn’t even quote a parody. He quotes the original argument, and then does proceed to make fun of it for a bit.
He considered its flaws to be mostly self-evident, which, apparently, is giving his critics too much credit, that’s true. The main point of the chapter, though, wasn’t to engage in in-depth refutation of every single version of the ontological argument (which, I maintain, depends on the same ridiculous flaws every time), but to show why a priori arguments in general are bullshit. He did this quite adequately, I think.
You can maintain the ontological argument is a respectable bit of philosophising all you want, but the fact is that even a four-year-old can refute it.
This really is the Courtier’s reply in its purest form, and I agree with PZ Myers when he says it’s akin to saying that because Lynch doesn’t deal with a translation of The God Delusion into Chinese, his entire premise is flawed.
(And yes, I would say Kant engaged in theology, for accepting the circular reasoning of religion as a priori truth. Hume, less so, perhaps.
This has nothing to do with reading Lynch’s post, incidentally, as all he talked about was the ontological argument, which is very firmly in the camp of pure theology. If you want, though, I’ll qualify my statement: theology is masturbatory philosophy of religion.)
J. J. Ramsey said,
January 20th, 2007 at 4:53 pm
Cairnarvon: “My ham sandwich analogy stands. It’s precisely as incoherent as the God version.”
Actually, it’s as incoherent as Gaunilo’s perfect island argument:
The above quote does not concede that the ontological argument works, but rather that it cannot be applied to any supposed “perfect thing,” such as a perfect ham sandwich.
Cairnarvon: “Rereading the chapter in question, Dawkins doesn’t even quote a parody.”
See page 83 in TGD, where Dawkins describes Douglas Gasking’s parody. Note that the incoherence of premise 5 is independent of any problems in the original ontological argument.
Cairnarvon: “You can maintain the ontological argument is a respectable bit of philosophising all you want, but the fact is that even a four-year-old can refute it.”
Funny that this comes from someone whose refutation of said argument is flawed.
Cairnarvon: “I would say Kant engaged in theology, for accepting the circular reasoning of religion as a priori truth.”
He didn’t do that in his argument against the ontological argument. From the bits and pieces of Kant, that I’ve seen, he deliberately avoided leaning on purported religious revelation in his philosophy.
Cairnarvon: “This has nothing to do with reading Lynch’s post, incidentally, as all he talked about was the ontological argument, which is very firmly in the camp of pure theology.”
The legitimate criticisms of the argument are from the camp of philosophy of religion, which treats religion as the object of critical study, not as something treated with kid gloves.
Cairnarvon said,
January 20th, 2007 at 5:02 pm
Or a perfect being. That was my goddamn point.
Then see page 80, where he describes Anselm’s argument.
You were more fun when you weren’t just outright trolling~
J. J. Ramsey said,
January 20th, 2007 at 5:26 pm
Me: “The above quote does not concede that the ontological argument works, but rather that it cannot be applied to any supposed ‘perfect thing,’ such as a perfect ham sandwich.”
Cairnarvon: “Or a perfect being. That was my goddamn point.”
I think you misread me, though I wasn’t as clear as I could be. What I was trying to say is that the ontological argument cannot be applied to every supposed “perfect thing,” like a piland or a perfect ham sandwich, but only to “concepts that are entirely defined in terms of properties that admit of some sort of intrinsic maximum.”
Cairnarvon: “Then see page 80, where he describes Anselm’s argument.”
Which doesn’t contain the flaw in premise 5 of Gaskin’s parody argument.
Cairnarvon said,
January 20th, 2007 at 5:56 pm
And my point was that beings cannot be defined like that either.
How doesn’t it?
But before Dawkins even gets to Gaskin, he first spends three and a half of the five-and-a-bit pages of the subchapter dealing with Anselm himself. Gaskin isn’t his main point, it’s just a demonstration of the same tortured logic that supported Anselm.
J. J. Ramsey said,
January 20th, 2007 at 7:42 pm
Cairnarvon: “And my point was that beings cannot be defined like that.”
No, your point was that one can use Anselm’s flawed logic to argue that a perfect ham sandwich (or a piland) exists, and this doesn’t work because a ham sandwich (or a piland) is not defined wholly in terms that admit of an intrinsic maximum.
Actually, the real problem with Anselm’s first ontological argument is that though one can define a being entirely in terms of properties that admit of some sort of intrinsic maximum, and one can even make as part of the definition that this being exists, making existence part of the definition of the concept of God cannot make God exist. That was more or less Kant’s point, and a younger atheistic Antony Flew, IIRC, put it explicitly in these terms in God : A Critical Enquiry. Noting this, BTW, requires more abstract reasoning than a four-year old is likely to have.
Me: “Which doesn’t contain the flaw in premise 5 of Gaskin’s parody argument.”
Cairnarvon: “How doesn’t it?”
Premise 5 supposes a creator that created everything while not existing, which is an incoherence that Anselm’s argument lacks.
Cairnarvon: “But before Dawkins even gets to Gaskin, he first spends three and a half of the five-and-a-bit pages of the subchapter dealing with Anselm himself.”
And three out of those three and a half pages contain no argument against Anselm at all. Dawkins does some mockery and some musing about Zeno and Bertrand Russell, and that’s about it. It’s not until that last half-page until he has something resembling an argument, and even then it’s incomplete for the reasons I mentioned above. After that half-page is Gasking’s cute but unillustrative parody argument, following by more argument-free mocking and musing. Those five-and-a-bit pages are mostly tangential verbiage, and amongst the verbiage is some very anemic argument.
Cairnarvon said,
January 20th, 2007 at 8:29 pm
Just repeating that you can define a being in terms that admit an intrinsic maximum doesn’t make it true, y’know.
That is the incoherent part? I think you’re slightly disconnected from parsimony. Which, indeed, was also Dawkins’ point.
Anyway, I’m through with this argument now. We can keep going back and forth on niggling little details without saying anything forever, as “Why I am opposed to religion” proved.
J. J. Ramsey said,
January 20th, 2007 at 9:29 pm
Cairnarvon: “Just repeating that you can define a being in terms that admit an intrinsic maximum doesn’t make it true, y’know.”
Ahem, Anselm already did define a being in terms that admit an intrinsic maximum when defining God as “a being than which nothing greater can be conceived/imagined,” so obviously, such a definition is possible. Whether there exists such a being that fits the definition is a whole other matter. Anselm, of course, did not show that such a being existed.
Me: “Premise 5 supposes a creator that created everything while not existing, which is an incoherence that Anselm’s argument lacks.”
Cairnarvon: “That is the incoherent part?”
Of Gasking’s argument, yes. More to the point, it is what makes it an ineffective critique of Anselm’s argument. If the only faults in Gasking’s parody argument were the ones in Anselm’s original, then the parody argument would have been a good illustration of the problems in Anselm’s argument. As it stands, this is not the case.
Cairnarvon said,
January 20th, 2007 at 10:08 pm
“Great” is not an objective quality. It isn’t for a being, it isn’t for an island, it isn’t for a ham sandwich.
If you’re going to argue that it’s permissible in the God argument, I want my damn sandwich.
J. J. Ramsey said,
January 20th, 2007 at 10:53 pm
Cairnarvon: “‘Great’ is not an objective quality. It isn’t for a being, it isn’t for an island, it isn’t for a ham sandwich. If you’re going to argue that it’s permissible in the God argument, I want my damn sandwich.”
I think I see what you are trying to say, that “great” is so ill-defined that there is no way to define “greater can be conceived/imagined.” The catch is that this isn’t quite true. Anselm has at least a rough working idea of what makes one being greater than another and what constitutes maximal greatness, and it is a working idea which others can find reasonable. With islands and ham sandwiches, you lack a working idea of maximal greatness.
J. J. Ramsey said,
January 21st, 2007 at 12:07 am
I should clarify a bit. As I see it, an ontological argument for a perfect X that follows the form of Anselm’s first ontological argument–which includes Gaunilo’s argument–has two weaknesses:
1) The difficulty in determining a working definition of “maximal greatness” of X.
2) The objection which Kant articulated as “existence is not a predicate” (though it may be articulated in other, better ways).
The second weakness is always fatal, period. The first weakness is fatal if there is no reasonable way to define what would make X maximally great–and this is a bigger “if” than you appear to grant.