We have two kinds of beliefs: Beliefs that ought to be the same for everyone, and beliefs that do not need to be the same for everyone. We have perfectly good words which denote these categories: truths and facts in the first sense, and opinions, attitudes, and feelings in the second. It is precisely because we have a moral duty to believe the truth that religions have placed their claims firmly in that category.
Every church, every religion--including the Catholic Church--has for millennia insisted that their claims have the moral status of truth: that you ought to believe them.
The insistence that religious beliefs are truth is always toxic. Either we take seriously (and this, I think, is the sense that Harris originally meant) the notion that our beliefs are true, that everyone ought to believe them, or we vacate our notion that there are any beliefs at all that everyone ought to believe, that there is no moral duty to believe the truth.
That religion might somehow be true but "incomprehensible" or unknowable doesn't help. The rational response to an unknown or incomprehensible truth is agnosticism and skepticism, not arbitrary belief. I don't know whether there's non-terrestrial life. Arbitrarily deciding to believe or disbelieve the claim would be equally irrational, so I remain skeptical and agnostic.
Even that religion might not be known with "certainty" doesn't help. We don't know anything with certainty. (Although, according to Sullivan, the beliefs we know with the highest confidence, mathematics, "may even be the closest to divine truth that human beings can achieve.") If certainty is the standard, then nothing is true, and we're back at epistemic nihilism.
In one sense, no, nobody ever has to justify any idea at all to anyone else's satisfaction ever, at least not in the United States. You can believe the Holocaust never happened, that Kennedy was assassinated by the Bavarian Illuminati, that the moon landings were a hoax. You can believe in ESP, tinfoil hats, timecubes, shape-changing lizard beings, a 6,000 year-old universe and even the causal efficacy of prayer. You can believe--and even publicly discuss--just about anything at all and you can rest assured that neither I, nor Sam Harris, nor Andrew Sullivan will punch you in the nose because of your belief. More importantly, you can rest assured that neither the FBI, the ATF, the Secret Service nor your local police department will come and arrest you and put you in jail. No matter how weird, misguided, insane or just plain stupid I think your ideas are, you have a political right to believe them and a political right to freely discuss them.
But if you're petitioning the community of rational human beings to accept the legitimacy of your beliefs, if you think you have something to say about the truth, then yes, you do have an obligation to justify your beliefs. If that's "intolerance", so be it: I'm not going to "tolerate" Gene Ray, Paul Rassinier, Uri Geller, Kent Hovind or Sylvia Browne as members in good standing of the community of rational human beings. Not because I happen to disagree with their conclusions, but because they do not use anything even remotely resembling reason in coming to their conclusions. I don't want them arrested or assaulted--I'll tolerate them to that extent--but I'm not going to tolerate giving them the slightest bit of standing, legitimacy or seriousness in the forum of rational discourse.
Sure, we'd love to have strong scientific, empirical justification of religious truth claims (I'm not holding my breath). But hey, we rational people are open minded, we'll settle for something else, so long as it's at least within spittin' distance of scientific or even historical rigor. We want something more than just Making Stuff Up.
But Sullivan gives us nothing. His entire argument is nothing but an elaborate version of the "deconstructionist" theme of, "Science doesn't explain everything, therefore anything I say is just as truthful and rational as science."
My point here is to say that once you have conceded the possibility of a truth that is not reducible to empirical proof, you have allowed for the validity of religious faith as a form of legitimate truth-seeking in a different mode.This is a pure non sequitur. The incompleteness of empirical proof, even accepted arguendo, does nothing to establish or even "allow" the validity of religious faith.
Sullivan actually admits that religious faith is irrational; quoting Hobbes, "For the nature of God is incomprehensible;" and quoting himself, "God must, by definition, surpass our human understanding." It is entirely absurd to believe one can speak the truth about what one doesn't understand. By that standard, I'm an expert on medieval French poetry.
Sullivan gives us another laundry list, "We have Scripture; we have reason; we have religious authority; we have our own spiritual experiences of the divine," but yet again he fails to explain how any of these items do anything at all to establish truth in any sort of rational manner. He might as well list dance, song, sitting in a circle and holding hands, chanting 'om mani padme om', speaking in tongues or handling snakes. Even the UFO nuts and crop circle faithful have a more explicit and detailed (although still entirely irrational) methodology.
If Sullivan's admirable secularism and pluralism do not allow him to "force or even rig laws to encourage others to share my [truthful] faith", if the truth of his faith deals "with matters that cannot be subject to common consensus," then one has to ask: If not on truth, on what basis should we create our laws? Maybe there are two kinds of truth, truthful truths and non-truthful or less-truthful truths? It's absurd.
The epistemic relativism, subjectivism and deconstructionism of this latest epistle (including the irritating "spaces" trope) might earn him a graduate degree in "leftist academia". But without any sort of explanation of his supposed methodology, a rational person is entirely justified in considering religion to be inadmissible in (rational) public debate, ludicrous, irrational, and--while the charge of "lying" is unjustified--most definitely deluded.
After all, this is precisely the sort of rational "intolerance" that Sullivan applies to liberalism.
I want to separate out this section because the points here peripheral to the main point above.
Sullivan appears to consider some controversial philosophical points as being settled and obvious.
The historical sciences, including the study of recorded human history, archeology, paleontology, criminal forensics and physical cosmology, are fully scientific and emprical. Both rely on the same fundamental methodology: to construct the most parsimonious falsifiable logical explanation for a given set of empirically verifiable facts. In the case of the the terrestrial historical sciences, the facts are most often written texts and artifacts; cosmology relies on the facts of astronomical observation.
Historical and forensic sciences seek to establish a different kind of truth than "universal" sciences, such many physical sciences. Universal sciences are concerned with finding universal relationships, relationships which are true independent of any particular state of the universe. (Controlled experiments are a particularly effective technique of finding these sorts of truths, but they're not a metaphysical necessity.) Historical and forensic sciences, on the other hand, are concerned with finding a particular past state of the world, usually assuming arguendo our current best understanding of scientific universals.
The goals are different, there are a few techniques peculiar to one class of sciences or the other, but the fundamental methodology is the same.
There's an enormous philosophical controversy about idea that mathematics are or are not truth-apt (capable of being true or false), beyond the trivial observation that it is true that some theorem is derivable from some set of axioms.
Materialism is not a necessary metaphysical presupposition for any empirical science. Most scientists take one or another kind of materialism for granted, but only on the massive weight of evidence of the efficacy of materialistic explanations.
Sullivan also throws in an outright howler. He quotes Harris:
I do not think that the utter reducibility of consciousness to matter has been established. It may be that the very concepts of mind and matter are fundamentally misleading us.and concludes that Harris has conceded a big point:
So you allow for a space where the logic of science and of materialism does not lead us toward truth, but may even mislead us about it, and lead us away from it.
It boggles the mind that Sullivan draws this irrational conclusion from Harris's remark. Harris himself does not draw any conclusion at all in his comment; he expresses only skepticism. Even if one reads in some implied conclusion, Harris does not at all state that he has come to that conclusion by any method other than scientific, empirical investigation. Even Sullivan's confusion about the relationship of empiricism to materialism fails to excuse this failure of rational analysis.
Moreover, why would Sullivan consider Harris's comment to represent such a big concession? Very few people--even religious people--consider religion to be specifically scientifically and empirically justified. The whole point of the debate is to challenge Sullivan to offer an alternative justification for religious truth (with empiricism serving as a template); just by participating Harris evidences open-mindedness about at least the possibility of such an alternative. But the burden is still on Sullivan to explain and justify his alternative. It is precisely because Sullivan considers this comment such a big concession that I'm rationally justified in interpreting Sullivan's argument in the logically invalid sense.
 The historical method is entirely scientific and empirical. See the section after the break.
 Sullivan uses weaselly phrasing here: He might be taken literally as uncontroversially allowing only the possibility that religion might be true; this sense, however, would be more clearly phrased as, "... allowed for the possibility..." I think the rest of the essay easily substantiates my stronger interpretation.
 To be fair, I don't think that any philosophical controversy at all has ever been resolved. Philosophers are still arguing over ideas that Socrates introduced almost three thousand years ago.
 See The Scientific Method (part 1) for more information.