It occurs to me that the whole Harris/Sullivan debate is going to hinge on what Sullivan means by the word "doubt".
Doubt has a very rigorous, precise and unequivocal meaning to the rational scientific mind. It's more than a just vague admission that one might be wrong. Because scientific truth is established by evidence, there's always in principle the possibility that some future evidence might falsify what we believe to be true. More importantly, because evidence is central to scientific doubt, it has physical, perceptual meaning.
It is the clarity and physicality of scientific doubt which allows us to have a rational basis for belief in scientific statements. We can test any particular doubt, and by examining the evidence, decisively dispel it. General relativity predicts that a clock on the ground will run more slowly--by a definite amount--than a clock in orbit. If we doubt General Relativity in this way, we need merely to put a clock in orbit and compare it to a clock on a ground. We can do so, we have done so, the clocks are different, this doubt is rationally dispelled, and we have yet another rational justification to have confidence in General Relativity.
We can collect only a finite amount of evidence, but one can imagine an infinity of doubt. We can doubt even the rock-solid scientific principle that the laws of physics are the same at all times, a principle that has been scientifically tested a quintillion times: Every time anyone turns on a light switch, expecting the room to become illuminated today just the same as it was yesterday, and confirms that the room has indeed become illuminated, they are testing and overcoming a doubt. But tomorrow, eh? Tomorrow the light might not come on, the kettle might not get hot, not because the bulb will have burned out or the gas will have been turned off, but just because the laws of physics might be different tomorrow than they were today, and yesterday. It's logically possible.
So yes, we can always doubt in principle any scientific belief. On the other hand, precisely because scientific doubt is both unequivocal and physically meaningful, every time we do overcome a doubt, we gain rational confidence in the statement we are doubting. After we've tested a principle a quintillion times, we achieve a level of confidence that indistinguishable in practice (although not in principle) from certainty.
The whole point of scientific doubt is to establish a rational basis for decisiveness even given the lack of absolute certainty. Scientific doubt is all about evidence that we can, by definition, look at. It is precisely to because doubt has rigorous physical meaning that scientists can insist that any statement must have doubtable implications to have scientific meaning. And it is precisely because doubt has rigorous physical meaning that overcoming doubt can establish rational confidence any decision based on science.
What does "doubt" mean, though, to the religious moderate? It's not possible to doubt religious statements in the same sense as a scientist doubts scientific truth. Religious statements aren't established by evidence, so it's not possible that the actual evidence we see with our eyes might not match what the religious statements entail.
How does a religious moderate construct "doubt", not just to achieve a superficial feeling of humility, but to establish a rational basis for making decisions?
I think it's completely uncontroversial to say that Andrew Sullivan--as well as most self-described religious moderates--would reject the label of "moral relativism": Moral truth, in their view, really is truth, not merely opinion. And religious belief is the foundation of that moral truth.
Even absent perfect certainty, how specifically does their construction of "doubt" provide a rational basis for confidence and decisiveness in any way at all? "Truthiness" just doesn't cut the mustard: It is not a rational basis to believe a moral proposition actually is true just because it somehow "feels" true--even if one admits with faux humility that one might be "wrong" in some vague, unspecified manner.
The "room" for doubt in any scientific statement is in its epistemic basis: the connection between the scientific statement and the evidence it entails. Without this connection, there is nothing to doubt in the scientific sense. Where's the room to doubt for religious statements? You can't doubt the epistemic basis of a definition ("God is definitionally the Creator of such a universe" [emphasis added]) per se; a definition (by definition) doesn't have an epistemic basis.
Finding room to doubt religious statements, on some basis more rational and principled than arbitrary truthiness, is the challenge that Andrew Sullivan will have to rise to. To be honest, I don't think he will, because I don't think he can.