The Non-God Delusion

Here is basically what I think about rationality applied to religion, in the form of a (casual) book review.

I read "The God Delusion" by Richard Dawkins last June. It's essentially an extended proselytizing tract for atheism. As one might expect from a book of that type, it doesn't advance a lot of new arguments (although Dawkins does offer his unique insights), but is rather an attempt to make existing arguments accessible for the masses. And on that front it has been successful; it is generally well written and has sold very well.

As with so many books of its type, it's mostly quite sensible. The problem is that its core conclusion -- that God is so unlikely to exist that belief in God is just silly -- is directly based on the two things it gets wrong. And I am not talking about differences of opinion -- I am talking about categorical errors in Dawkins' reasoning that seemed quite obvious to me (and Dawkins' apparent ignorance of which, given the book's rallying cry of rational scientific thought, I found to be fairly ironic).

The two errors are related, but they're worth dividing conceptually. First is that Dawkins sets up a straw man: he relies on one of the most well-documented fallacies of logic in order to argue that only a certain conception of God can possibly work for theists (a conception which he then more sensibly argues does not work). Second is that he misuses probabilistic reasoning.

Dawkins' Ontological Error

Dawkins' main argument is based on a core premise: if God exists, God must have created the world in a supernatural manner that did not involve natural selection. Dawkins then argues that all evidence (of which there is a lot) indicates that the world was not created in such a manner, and it is thus so unlikely that God exists that one would be irresponsible to believe in God.

But there is a problem with this line of reasoning, and let's just say it how it is: Dawkins' premise is inane. It begs the question. It can't possibly be a starting point for a rational argument about the existence or non-existence of God.

To see why, let's first look at the line of reasoning which Dawkins' ironically mirrors.

The Ontological Argument

The Ontological Argument (the OA) was(/is, in some circles) an attempt to prove the existence of God rationally; Descartes' version is perhaps the most famous. I'll boil the OA down into what I think is a fair summary:

God is the greatest entity that can be conceived.
It is greater to exist than not to exist.
Therefore, God exists.

There are two problems with the OA.

OA problem one: "great" is subjective

The first problem is simple: who is to say that existence is greater than non-existence? A number of logicians have, in parody (usually), argued exactly the opposite (e.g., creating the universe while not existing is greater than creating it while existing) to illustrate the simple principle that the validity of a logical conclusion is only as good as that of the premises on which the conclusion is based. If the premises are subjective value statements, the conclusion won't be worth much. Consider the following argument:

The best NFL team in the world uses a dolphin as its mascot.
The Miami Dolphins use a dolphin as their mascot.
Therefore, the Miami Dolphins are the best NFL team in the world.

Obviously, the conclusion of this argument is very debatable -- as is the underlying premise in the first sentence. Now, you might think, "wait a minute; the conclusion for this argument is subjective, while the conclusion for the OA is objective." But the conclusion for the OA only appears to be objective; the conclusion is only as strong as its weakest premise, which is a subjective one. This becomes clearer if you transform the OA slightly:

X is the greatest NBA team that can be conceived.
It is greater for an NBA team to win every game it plays than to lose any games.
It is greater for an NBA team to exist than not to exist.
Therefore, X exists and has won every game it ever played.

Here, you have a seemingly objective conclusion that obviously doesn't work because it is based on a flawed premise.

OA problem two: assuming existence

As you read the NBA argument above, you no doubt asked yourself (probably so quickly that it was purely reflexive): is there an NBA basketball team that has won every game it ever played? You realized that there isn't, so X doesn't exist. The argument didn't prove any valid point; it merely raised the question it was trying to prove, and you answered the question in your own mind.

This illustrates the even deeper problem with the OA: it assumes what it is trying to prove. It defines God as the entity that has all the greatest conceivable qualities, one of which is existence. So, at its core, the OA argues that the entity that exists (and has all the other greatest qualities) exists (and presumably has all the other greatest qualities). That's fine if there is an entity that exists and has all the other greatest qualities, but what if there isn't? This type of argument begs the question.

To look at it another way, consider the following arguments, in the form of "X was the first person to do Y, therefore, at least one person has done Y":

1) X was the first person to orbit the Earth. Therefore, at least one person has orbited the Earth.

2) X was the first person to land on Mars. Therefore, at least one person has landed on Mars.

3) X was the first person to give birth to more than 70 children. Therefore, at least one person has given birth to more than 70 children.

We know that there is a single person who fits X in the first example: X is Yuri Gagarin, and so the conclusion is true. However, with regards to the second example, there is no person who fits the bill of X. X is an "empty name" -- it doesn't refer to anything -- and so the conclusion is false (if there was no first person to do it, no other person could have done it, either). Turning to the third example, the highest documented number of children a person has given birth to is 69. However, records on that sort of thing don't go back very far and cover only limited geography, so it's possible someone had 70+ children at some point in human history. Thus, we can't really know whether X refers to somebody, and so we can neither prove nor disprove the conclusion.

In other words, the conclusions above that a person did Y are true only to the extent that X refers to something. If X doesn't refer to anything -- if it is an empty name -- then the conclusion isn't true. The existence of X is a key underlying question, the answer to which affects the validity of the conclusion.

Similarly, the OA is true only if God refers to an actual entity; it's not true if God is an empty name. The existence of God is a key underlying question, the answer to which affects the validity of the conclusion. So the OA works only if God exists, which gets us nowhere. We have to look elsewhere for any rational argument for or against the existence of God. (I tend to think there isn't one.)

Dawkins' use of the OA

Despite the fact that Dawkins himself devotes a few (somewhat long-winded) pages to discussing the OA, he seems to be entirely ignorant of his own adaptation of it as the only support for the flawed premise that underlies his core argument.

Remember, Dawkins' premise is: if God exists, God must have created the world in a supernatural manner that did not involve natural selection. Here is what he says in his sole attempt to defend it:

"I am continually astonished by those theists who . . . seem to rejoice in natural selection as 'God's way of achieving his creation'. They note that evolution by natural selection would be a very easy and neat way to achieve a world full of life. God wouldn't need to do anything at all! Peter Atkins . . . takes this line of thought to a sensibly godless conclusion when he postulates a hypothetically lazy God who tries to get away with as little as possible in order to make a universe containing life. Atkins's lazy God is even lazier than the deist God of the eighteenth-century Enlightenment: deus otiosus -- literally God at leisure, unoccupied, unemployed, superfluous, useless. Step by step, Atkins succeeds in reducing the amount of work the lazy God has to do until he finally ends up doing nothing at all: he might as well not bother to exist."

Does this sound familiar? I noted above that most logicians who have made this argument do so in parody. But Dawkins, shockingly, seems to take it seriously, despite it being such an obvious form of the OA:

God is the laziest entity that can be conceived.
Not existing is lazier than existing.
Therefore, God doesn't exist.

Obviously, as a form of the OA, this argument is subject to the same problems set out above: first, it is based on a subjective premise, and second, it is only true if God refers to an empty name, so it assumes the non-existence of God in the first place.

Dawkins' use of this argument to develop his core premise is embarrassing on its own. But it is all the more galling given his harsh criticism of the OA: "The very idea that grand conclusions could follow from such logomachist trickery offends me aesthetically, so I must take care to refrain from bandying words like 'fool'." I will also take care to so restrain myself, given the intellectual hypocrisy of Dawkins' attempt to spoonfeed unthinking masses the very sort of stuff he so ardently decries.

In sum, Dawkins' premise that God, if he exists, must have created the world in a supernatural manner that did not involve natural selection, is a gross error. It assumes what it purports to show. Given the fundamental flaw in Dawkins' premise, one can already be certain that any further conclusion based on it will be equally flawed; as mentioned above, the validity of a logical conclusion is only as good as that of the premises on which the conclusion is based.

Dawkins' Probability Error

Notwithstanding that his premise about the necessary nature of God erroneously relies on a false (and intellectually corrupt) proof of God's non-existence, Dawkins admits: "[t]hat you cannot prove God's non-existence is accepted and trivial." But Dawkins sees this as a technical point -- it is true "only in the sense that we can never absolutely prove the non-existence of anything." He goes on to say -- and this is the thrust of his argument -- that "[w]hat matters is not whether God is disprovable (he isn't) but whether his existence is probable."

So, to summarize Dawkins' position again for ease of reference, Dawkins first sets up his (faulty) premise that if God exists, God must have created the world in a supernatural manner that did not involve natural selection. He then goes on to make the argument that existing evidence (of which there is a lot) indicates that the world was not created in such a manner, and it is thus so unlikely that God exists that one would be irresponsible to believe in God.

As noted above, this argument is prima facie inconclusive since its premise is faulty. It is like asserting that 2+3>6 or 5+3>6 because of the underlying premise 3>6. Dawkins may, in fact, be right, or he may not be, but his argument gets him nowhere unless we have some other method of verifying the correctness of his argument (as we do with the simple mathematical examples).

But there is another problem with Dawkins' probabilistic reasoning -- one that would still exist even if he altered his premise.

Let's look at where Dawkins begins his probabilistic line of reasoning. To illustrate the alleged improbability of God, Dawkins starts with Bertand Russell's (whose history of Western philosophy I'm currently reading) teapot-in-space analogy. Russell said:

"If I were to suggest that between the Earth and Mars there is a china teapot revolving about the sun in an elliptical orbit, nobody would be able to disprove my assertion provided I were careful to add that the teapot is too small to be revealed even by our most powerful telescopes. But if I were to go on to say that, since my assertion cannot be disproved, it is an intolerable presumption on the part of human reason to doubt it, I should rightly be thought to be talking nonsense. If, however, the existence of such a teapot were affirmed in ancient books, taught as the sacred truth every Sunday, and instilled into the minds of children at school, hesitation to believe in its existence would become a mark of eccentricity and entitle the doubter to the attentions of the psychiatrist in an enlightened age or of the Inquisitor in an earlier time."

Dawkins applies the teapot principle to various hypothetical entities. Presumably, he says, you don't believe in the tooth fairy. But although you can't prove that the tooth fairy doesn't exist, you wouldn't describe yourself as a tooth fairy agnostic. You would be an a-fairyist (like a-theist) because you think the probability of the tooth fairy existing is so low as to be unimportant. The same would go for the Flying Spaghetti Monster or any number of other potential entities.

Dawkins then applies the teapot principle to the definition of God he (erroneously) presumes must be the necessary one. He explains at great length why that version of God is very unlikely to exist (especially responding to intelligent design). Which is fine. But what it doesn't say is much, if anything, about the probability of the existence of God generally (just as my rational conclusions about hypothetical likely non-candidates to be George Bush or the tooth fairy do not say much about the likelihood of their existence or non-existence).

Much more interestingly, in transforming Russell's argument from one about possibility to one about probability, Dawkins (seemingly inadvertently) exposes the flaw in Russell's (and Dawkins' adopted) analogy: the degree to which one can be "thought to be talking nonsense" by asserting the existence of X is a corollary of the specificity with which X is defined. If, in Russell's example, X were "an object that our most powerful telescopes cannot reveal," the probability of X's existence would approach 100%. Where X is "a china teapot revolving about the sun in an elliptical orbit," the probability of X's existence would approach 0%. And there are infinite potential Xs somewhere inbetween.

So Russell's teapot principle only holds true when the hypothetical object whose existence is questioned is defined very specifically. The exact reverse holds true when the object is defined very broadly.

Dawkins' application of the teapot principle doesn't work because he defines God narrowly (as discussed above) -- in fact, he uses a definition of God against which he has some strong arguments up his sleeve. This is not strong probabilistic reasoning.

Admittedly, Dawkins rightfully says that if a definition of God becomes too broad, it loses meaning. But there is plenty of room for a conception of God somewhere between the 0% and 100% ends of the probability spectrum. (Personally, I think the Mormon conception fits the bill quite well, but I'll leave that for another post.)

Fixing Dawkins' Errors

For the sake of completeness, let's discuss what Dawkins should have said had he put his thinking cap on.

Hume noted that "the only way to prove anything a priori is through an opposite contradiction." In other words, you can define the properties of something by showing what it cannot be conceived of being -- for instance, "I am a married bachelor." You can conceive of God existing. You can conceive of God not existing. So you're not going to get anywhere vis-a-vis the existence of God through a priori reasoning.

Now, you might be able to rule out (or come close to ruling out) certain immediate characteristics of a hypothetical God via a priori reasoning. For instance, if God exists, it seems incredibly unlikely that he hurls lightning bolts from the sky. Or, in other words, a conception of God that involves him hurling lightning bolts from the sky is almost certainly wrong. (We can figure this out a priori not because we know anything about God, but because we know where lightning bolts come from, at least at some level of causality.)

So what Dawkins could have said is that it is unlikely that God (if he exists) created the world in a manner that did not involve what we call natural selection. Or, in other words, a conception of God that involves him creating the world in that manner is probably not right.

That would have been fine. But Dawkins overreached in his open zeal to spread his atheist dogma and ended up using the very logomachist trickery he found so offensive in an attempt to advance an inadvancable argument. One can rationally narrow down the likely characteristics of God (if he exists), but one cannot argue a priori that God does not exist. Yet Dawkins employs exactly such a priori reasoning in order to rule out every conception of God except the one that makes his probability argument look good. Skillful legerdemain.

If one relies only on rationality, I think that one must be agnostic (which is why I view pseudo-rational atheists like Dawkins quite like I view pseudo-rational theists). We can't rationally know whether God exists. Admittedly, some belief systems in God may run contrary to rational thought; but, as we have seen, so can some belief systems in the non-existence of God. Those belief systems can, and in my opinion, should, be subject to critical thought. On the other hand, other belief systems in the existence of God (such as mine) as well as in the non-existence of God comprise beliefs about what lies outside the reach of the rational world.*

"The God Delusion" and its scriptorial acceptance in some circles are fantastic examples of how atheism is so very much like theism. The atheist belief system propounded in those circles is cult-like, built on the backs of charismatic figures such as Dawkins, whose assertion of beliefs as truth is taken as gospel by unthinking masses. There is more than a little irony in Dawkins' acrimonious attitude towards theists, since his acrimony derives from perceived theist attributes that he embodies all too well. Dawkins, the dogmatist, refuses to recognize his God belief for what it is -- a belief in an unknowable area, a hypothesis in relation to which probabilistic (or any other kind of) reasoning says very little. (While I, a theist, am happy to recognize exactly that.)

* Saying these beliefs are "outside the reach of the rational world" is not to say that I think the existence of God is not a question that could, theoretically, be verified by scientific analysis; I would be the first to promote such an analysis, but I just don't think a conclusive one is possible. Saying these beliefs are "outside the reach of the rational world" is also not to say that I think rationality cannot inform religion; it tells us what is improbable on one end of the spectrum of probabilities and meaningless on the other. But there is a lot of area between the near-0% teapot and the near-100% generic object.

P.S. I'm aware that I use some terms perhaps slightly ambiguously -- e.g., "rationality" where I mean both rationality and empiricism. This is more or less a stream-of-consciousness post in its earliest form and, to the extent I can, I will clean things up over time.