I will pick out two things for discussion (but feel free to bring in other stuff).
FIRST THING: Dawkins’ attack on Aquinas’ first three ways is really twofold:
(1) First, if we are going to halt the regress with something, why not just stop at e.g. the Big Bang? Why add God and then play the “Oh, and this is the exception to the rule that everything requires a cause” card? Rather than just play the card at the Big Bang?
This first objection can be explained by analogy. The ancient Hindus, struck by the fact that things that are not supported fall, wondered what prevented the Earth from falling. If all things fall that are not supported, then the Earth must have a support. But what? They posited a big elephant. But then what supports the elephant? They posited a big turtle to support the elephant. But what supported the turtle? It’s here, it seems, that they played the “exception to the rule card” – the cosmic turtle is the one exception to the rule that unsupported things fall. But if the “exception to the rule” card is going to be played at some point, it should surely be played as early as possible. There really is no justification for first introducing a cosmic elephant and turtle, and then playing the card.
(2) Dawkins points out that the most these arguments would establish, at best, is a first cause – but not that this first cause is the omnipotent, omniscient, supremely good, etc, God of traditional monotheism. Compare the elephant and turtle. Even if the Earth does require a support, even an animal support – why an elephant and a turtle? Why not a cosmic gerbil sitting on an enormous rhino?
The first objection looks strong. But of course theists have responses to it. For example, if they could show that, while there must be an exception to the rule that everything has a cause, it cannot plausibly be the Big Bang, but must be God, then the argument would still work. And of course, many theists believe they can indeed show that.
They may say, for example, that a (non-dependently) necessary being is an obvious an exception to the rule that everything has a cause. A being that exists necessarily requires no cause to bring it into existence. The Big Bang, on the other hand, is not a necessary thing – yes, it happened, but it did not have to happen. There might have been no bang. Therefore, it's unreasonable to make the Big Bang the exception to the rule, but not unreasonable to make God the exception to the rule.
In fact, something along these lines is suggested by Aquinas’ third way (though this is lost in Dawkins' presentation of it).
I don’t say this modified version of the first cause argument is strong, but it does not have the fault that Dawkins finds with the first three of Aquinas’ five ways.
So of course the theists will say that Dawkins has merely knocked down a straw man. He hasn’t really engaged with the strongest and most sophisticated arguments for the existence of God. He's dealt merely with overly simplistic versions of those arguments.
Would that be a fair objection, though? After all, we cannot expect Dawkins to cover every version of every theistic argument in a book like this. The most we can ask for is an overview of the main arguments, which is what we get.
But then, the theist may again object, how can Dawkins be justified in rubbishing all these arguments on the basis of such a cursory treatment? [I don’t really agree with this objection, in fact, but I raise it for discussion].
On page 2, Dawkins does raise a well-known objection to God’s being both omnipotent and omniscient.
"If he is omniscient, he must already know how he is going to intervene to change the course of history using omnipotence. But that means he can’t change his mind about his intervention, which means he is not omnipotent."
Actually, I am not yet convinced by this kind of objection. Like Dawkins, I think that the very idea of God makes little sense, but not for this particular reason.
Foreknowledge of something, including your own actions, does not prevent you from changing your mind. The suggestion that it does may turn on a famous modal fallacy.
Compare (for any person a and proposition P [e.g. a = Bert, and P = his keys are on the table):
(i) Necessarily: If a knows P, then P is true.
(ii) If a knows P, then necessarily: P is true (i.e. is a necessary truth).
Note that the only difference between (i) and (ii) is that the “scope” of the modal operator “necessarily” shifts. But as a result, very different claims are made. Claim (i) is true. Necessarily, you cannot know what's not true (you can at best, only believe it). But claim (ii) is false – we can know things that aren't necessary truths.
Get these two claims confused, however, and you may end up thinking that if someone knows something will happen, then what they know couldn't have been otherwise.
But of course, that’s wrong. Compare:
If I know that I am going to Paris tomorrow, then it must be true that I will go to Paris tomorrow. But it’s not a necessary truth that I’ll go to Paris tomorrow. In fact, I am able to change my mind about going. So – despite the fact that I know I'm going, I can still change my mind.
NB the same applies to things I have done. I know I went to the bank yesterday. Necessarily: if I know I went to the bank, then I did. Still, I could have done otherwise.
Ditto God’s knowledge of what he’ll do tomorrow.
(Or is there something special about God’s foreknowledge that entails he cannot change his mind – e.g. his infallibility. I don't think God's infallibility makes any difference, actually, but you may disagree)?
[POSTSCRIPT: Here's yet another illustration of the last point. Consider my knowledge of where a subatomic particle landed in an experiment yesterday. Necesarily: if I know it landed at position x, then it landed at position x. It doesn't follow that it couldn't have landed somewhere else (indeed, if thaw laws governing such particles allow indeterminacy, it needn't even be determined by the laws of nature that it should land at position x).]